* [PATCH v3 01/24] fix default __strnlen_user macro
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
@ 2011-09-27 20:29 ` Mark Salter
2011-09-28 13:20 ` Arnd Bergmann
2011-09-27 20:29 ` [PATCH v3 02/24] fixed generic page.h for non-zero PAGE_OFFSET Mark Salter
` (24 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:29 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
The existing __strnlen_user macro simply resolved to strnlen. However, the
count returned by strnlen_user should include the NULL byte. This patch
fixes the __strnlen_user macro to include the NULL byte in the count.
Signed-off-by: Mark Salter <msalter@redhat.com>
---
include/asm-generic/uaccess.h | 7 ++++++-
1 files changed, 6 insertions(+), 1 deletions(-)
diff --git a/include/asm-generic/uaccess.h b/include/asm-generic/uaccess.h
index ac68c99..9788568 100644
--- a/include/asm-generic/uaccess.h
+++ b/include/asm-generic/uaccess.h
@@ -289,9 +289,14 @@ strncpy_from_user(char *dst, const char __user *src, long count)
* Return 0 on exception, a value greater than N if too long
*/
#ifndef __strnlen_user
-#define __strnlen_user strnlen
+#define __strnlen_user(s, n) (strnlen((s), (n)) + 1)
#endif
+/*
+ * Unlike strnlen, strnlen_user includes the nul terminator in
+ * its returned count. Callers should check for a returned value
+ * greater than N as an indication the string is too long.
+ */
static inline long strnlen_user(const char __user *src, long n)
{
if (!access_ok(VERIFY_READ, src, 1))
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v3 02/24] fixed generic page.h for non-zero PAGE_OFFSET
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
2011-09-27 20:29 ` [PATCH v3 01/24] fix default __strnlen_user macro Mark Salter
@ 2011-09-27 20:29 ` Mark Salter
2011-09-27 20:29 ` [PATCH v3 03/24] add ELF machine define for TI C6X DSPs Mark Salter
` (23 subsequent siblings)
25 siblings, 0 replies; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:29 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
asm-generic/page.h had several problems when used with a non-zero PAGE_OFFSET.
This patch adds a default ARCH_PFN_OFFSET and fixes the __va, __pa, and
pfn_valid macros to work with non-zero PAGE_OFFSETs.
Signed-off-by: Mark Salter <msalter@redhat.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
---
include/asm-generic/page.h | 10 +++++++---
1 files changed, 7 insertions(+), 3 deletions(-)
diff --git a/include/asm-generic/page.h b/include/asm-generic/page.h
index 75fec18..f376db2 100644
--- a/include/asm-generic/page.h
+++ b/include/asm-generic/page.h
@@ -71,10 +71,14 @@ extern unsigned long memory_end;
#define PAGE_OFFSET (0)
#endif
+#ifndef ARCH_PFN_OFFSET
+#define ARCH_PFN_OFFSET (PAGE_OFFSET >> PAGE_SHIFT)
+#endif
+
#ifndef __ASSEMBLY__
-#define __va(x) ((void *)((unsigned long)(x) + PAGE_OFFSET))
-#define __pa(x) ((unsigned long) (x) - PAGE_OFFSET)
+#define __va(x) ((void *)((unsigned long) (x)))
+#define __pa(x) ((unsigned long) (x))
#define virt_to_pfn(kaddr) (__pa(kaddr) >> PAGE_SHIFT)
#define pfn_to_virt(pfn) __va((pfn) << PAGE_SHIFT)
@@ -86,7 +90,7 @@ extern unsigned long memory_end;
#define page_to_phys(page) ((dma_addr_t)page_to_pfn(page) << PAGE_SHIFT)
#endif
-#define pfn_valid(pfn) ((pfn) < max_mapnr)
+#define pfn_valid(pfn) ((pfn) >= ARCH_PFN_OFFSET && ((pfn) - ARCH_PFN_OFFSET) < max_mapnr)
#define virt_addr_valid(kaddr) (((void *)(kaddr) >= (void *)PAGE_OFFSET) && \
((void *)(kaddr) < (void *)memory_end))
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v3 03/24] add ELF machine define for TI C6X DSPs
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
2011-09-27 20:29 ` [PATCH v3 01/24] fix default __strnlen_user macro Mark Salter
2011-09-27 20:29 ` [PATCH v3 02/24] fixed generic page.h for non-zero PAGE_OFFSET Mark Salter
@ 2011-09-27 20:29 ` Mark Salter
2011-09-27 20:29 ` [PATCH v3 04/24] add missing __iomem to generic iounmap declaration Mark Salter
` (22 subsequent siblings)
25 siblings, 0 replies; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:29 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
Signed-off-by: Mark Salter <msalter@redhat.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
---
include/linux/elf-em.h | 1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff --git a/include/linux/elf-em.h b/include/linux/elf-em.h
index 18bea78..8e2b7ba 100644
--- a/include/linux/elf-em.h
+++ b/include/linux/elf-em.h
@@ -33,6 +33,7 @@
#define EM_H8_300 46 /* Renesas H8/300,300H,H8S */
#define EM_MN10300 89 /* Panasonic/MEI MN10300, AM33 */
#define EM_BLACKFIN 106 /* ADI Blackfin Processor */
+#define EM_TI_C6000 140 /* TI C6X DSPs */
#define EM_FRV 0x5441 /* Fujitsu FR-V */
#define EM_AVR32 0x18ad /* Atmel AVR32 */
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v3 04/24] add missing __iomem to generic iounmap declaration
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (2 preceding siblings ...)
2011-09-27 20:29 ` [PATCH v3 03/24] add ELF machine define for TI C6X DSPs Mark Salter
@ 2011-09-27 20:29 ` Mark Salter
2011-09-28 13:19 ` Arnd Bergmann
2011-09-27 20:29 ` [PATCH v3 05/24] C6X: build infrastructure Mark Salter
` (21 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:29 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
Signed-off-by: Mark Salter <msalter@redhat.com>
---
include/asm-generic/io.h | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/include/asm-generic/io.h b/include/asm-generic/io.h
index 9120887..c2cf2ed 100644
--- a/include/asm-generic/io.h
+++ b/include/asm-generic/io.h
@@ -327,7 +327,7 @@ static inline void __iomem *ioremap(phys_addr_t offset, unsigned long size)
#define ioremap_wc ioremap_nocache
#endif
-static inline void iounmap(void *addr)
+static inline void iounmap(void __iomem *addr)
{
}
#endif /* CONFIG_MMU */
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v3 05/24] C6X: build infrastructure
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (3 preceding siblings ...)
2011-09-27 20:29 ` [PATCH v3 04/24] add missing __iomem to generic iounmap declaration Mark Salter
@ 2011-09-27 20:29 ` Mark Salter
2011-09-28 13:23 ` Arnd Bergmann
2011-09-27 20:29 ` [PATCH v3 06/24] C6X: early boot code Mark Salter
` (20 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:29 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
Signed-off-by: Mark Salter <msalter@redhat.com>
---
arch/c6x/Kconfig | 173 +++++++++++++++++++++++++++++++++++
arch/c6x/Makefile | 52 +++++++++++
arch/c6x/boot/Makefile | 22 +++++
arch/c6x/configs/dsk6455_defconfig | 43 +++++++++
arch/c6x/configs/evmc6457_defconfig | 40 ++++++++
arch/c6x/configs/evmc6472_defconfig | 41 ++++++++
arch/c6x/configs/evmc6474_defconfig | 41 ++++++++
arch/c6x/include/asm/Kbuild | 55 +++++++++++
arch/c6x/kernel/Makefile | 12 +++
arch/c6x/kernel/vmlinux.lds.S | 162 ++++++++++++++++++++++++++++++++
arch/c6x/lib/Makefile | 7 ++
arch/c6x/mm/Makefile | 5 +
arch/c6x/platforms/Kconfig | 41 ++++++++
arch/c6x/platforms/Makefile | 12 +++
14 files changed, 706 insertions(+), 0 deletions(-)
create mode 100644 arch/c6x/Kconfig
create mode 100644 arch/c6x/Makefile
create mode 100644 arch/c6x/boot/Makefile
create mode 100644 arch/c6x/configs/dsk6455_defconfig
create mode 100644 arch/c6x/configs/evmc6457_defconfig
create mode 100644 arch/c6x/configs/evmc6472_defconfig
create mode 100644 arch/c6x/configs/evmc6474_defconfig
create mode 100644 arch/c6x/include/asm/Kbuild
create mode 100644 arch/c6x/kernel/Makefile
create mode 100644 arch/c6x/kernel/vmlinux.lds.S
create mode 100644 arch/c6x/lib/Makefile
create mode 100644 arch/c6x/mm/Makefile
create mode 100644 arch/c6x/platforms/Kconfig
create mode 100644 arch/c6x/platforms/Makefile
diff --git a/arch/c6x/Kconfig b/arch/c6x/Kconfig
new file mode 100644
index 0000000..3f3dfc7
--- /dev/null
+++ b/arch/c6x/Kconfig
@@ -0,0 +1,173 @@
+#
+# For a description of the syntax of this configuration file,
+# see Documentation/kbuild/kconfig-language.txt.
+#
+
+config TMS320C6X
+ def_bool y
+ select CLKDEV_LOOKUP
+ select GENERIC_IRQ_SHOW
+ select HAVE_DMA_API_DEBUG
+ select HAVE_GENERIC_HARDIRQS
+ select HAVE_MEMBLOCK
+ select HAVE_SPARSE_IRQ
+ select OF
+ select OF_EARLY_FLATTREE
+
+config MMU
+ def_bool n
+
+config ZONE_DMA
+ def_bool y
+
+config FPU
+ def_bool n
+
+config HIGHMEM
+ def_bool n
+
+config NUMA
+ def_bool n
+
+config RWSEM_GENERIC_SPINLOCK
+ def_bool y
+
+config RWSEM_XCHGADD_ALGORITHM
+ def_bool n
+
+config GENERIC_CALIBRATE_DELAY
+ def_bool y
+
+config GENERIC_HWEIGHT
+ def_bool y
+
+config GENERIC_CLOCKEVENTS
+ def_bool y
+
+config GENERIC_CLOCKEVENTS_BROADCAST
+ bool
+
+config GENERIC_BUG
+ def_bool y
+
+config COMMON_CLKDEV
+ def_bool y
+
+config C6X_BIG_KERNEL
+ bool "Build a big kernel"
+ help
+ The C6X function call instruction has a limited range of +/- 2MiB.
+ This is sufficient for most kernels, but some kernel configurations
+ with lots of compiled-in functionality may require a larger range
+ for function calls. Use this option to have the compiler generate
+ function calls with 32-bit range. This will make the kernel both
+ larger and slower.
+
+ If unsure, say N.
+
+source "init/Kconfig"
+
+# Use the generic interrupt handling code in kernel/irq/
+
+source "kernel/Kconfig.freezer"
+
+config CMDLINE_BOOL
+ bool "Default bootloader kernel arguments"
+
+config CMDLINE
+ string "Kernel command line"
+ depends on CMDLINE_BOOL
+ default "console=ttyS0,57600"
+ help
+ On some architectures there is currently no way for the boot loader
+ to pass arguments to the kernel. For these architectures, you should
+ supply some command-line options at build time by entering them
+ here.
+
+config CMDLINE_FORCE
+ bool "Force default kernel command string"
+ depends on CMDLINE_BOOL
+ default n
+ help
+ Set this to have arguments from the default kernel command string
+ override those passed by the boot loader.
+
+config CPU_BIG_ENDIAN
+ bool "Build big-endian kernel"
+ default n
+ help
+ Say Y if you plan on running a kernel in big-endian mode.
+ Note that your board must be properly built and your board
+ port must properly enable any big-endian related features
+ of your chipset/board/processor.
+
+config FORCE_MAX_ZONEORDER
+ int "Maximum zone order"
+ default "13"
+ help
+ The kernel memory allocator divides physically contiguous memory
+ blocks into "zones", where each zone is a power of two number of
+ pages. This option selects the largest power of two that the kernel
+ keeps in the memory allocator. If you need to allocate very large
+ blocks of physically contiguous memory, then you may need to
+ increase this value.
+
+ This config option is actually maximum order plus one. For example,
+ a value of 11 means that the largest free memory block is 2^10 pages.
+
+menu "Processor type and features"
+
+source "arch/c6x/platforms/Kconfig"
+
+config TMS320C6X_CACHES_ON
+ bool "L2 cache support"
+ default y
+
+config KERNEL_RAM_BASE_ADDRESS
+ hex "Virtual address of memory base"
+ default 0xe0000000 if SOC_TMS320C6455
+ default 0xe0000000 if SOC_TMS320C6457
+ default 0xe0000000 if SOC_TMS320C6472
+ default 0x80000000
+
+source "mm/Kconfig"
+
+source "kernel/Kconfig.preempt"
+
+source "kernel/Kconfig.hz"
+source "kernel/time/Kconfig"
+
+endmenu
+
+menu "Executable file formats"
+
+source "fs/Kconfig.binfmt"
+
+endmenu
+
+source "net/Kconfig"
+
+source "drivers/Kconfig"
+
+source "fs/Kconfig"
+
+source "security/Kconfig"
+
+source "crypto/Kconfig"
+
+source "lib/Kconfig"
+
+menu "Kernel hacking"
+
+source "lib/Kconfig.debug"
+
+config ACCESS_CHECK
+ bool "Check the user pointer address"
+ default y
+ help
+ Usually the pointer transfer from user space is checked to see if its
+ address is in the kernel space.
+
+ Say N here to disable that check to improve the performance.
+
+endmenu
diff --git a/arch/c6x/Makefile b/arch/c6x/Makefile
new file mode 100644
index 0000000..7186297
--- /dev/null
+++ b/arch/c6x/Makefile
@@ -0,0 +1,52 @@
+#
+# linux/arch/c6x/Makefile
+#
+# This file is subject to the terms and conditions of the GNU General Public
+# License. See the file "COPYING" in the main directory of this archive
+# for more details.
+#
+
+cflags-y += -mno-dsbt -msdata=none
+
+cflags-$(CONFIG_C6X_BIG_KERNEL) += -mlong-calls
+
+CFLAGS_MODULE += -mlong-calls -mno-dsbt -msdata=none
+
+CHECKFLAGS +=
+
+KBUILD_CFLAGS += $(cflags-y)
+KBUILD_AFLAGS += $(cflags-y)
+
+ifdef CONFIG_CPU_BIG_ENDIAN
+KBUILD_CFLAGS += -mbig-endian
+KBUILD_AFLAGS += -mbig-endian
+LINKFLAGS += -mbig-endian
+KBUILD_LDFLAGS += -mbig-endian
+LDFLAGS += -EB
+endif
+
+head-y := arch/c6x/kernel/head.o
+core-y += arch/c6x/kernel/ arch/c6x/mm/ arch/c6x/platforms/
+libs-y += arch/c6x/lib/
+
+# Default to vmlinux.bin, override when needed
+all: vmlinux.bin
+
+boot := arch/$(ARCH)/boot
+
+# With make 3.82 we cannot mix normal and wildcard targets
+
+vmlinux.bin: vmlinux
+ $(Q)$(MAKE) $(build)=$(boot) $(patsubst %,$(boot)/%,$@)
+
+dtbImage.%: vmlinux
+ $(Q)$(MAKE) $(build)=$(boot) $(patsubst %,$(boot)/%,$@)
+
+archclean:
+ $(Q)$(MAKE) $(clean)=$(boot)
+
+define archhelp
+ @echo ' vmlinux.bin - Binary kernel image (arch/$(ARCH)/boot/vmlinux.bin)'
+ @echo ' dtbImage.<dt> - ELF image with $(arch)/boot/dts/<dt>.dts linked in'
+ @echo ' - stripped elf with fdt blob'
+endef
diff --git a/arch/c6x/boot/Makefile b/arch/c6x/boot/Makefile
new file mode 100644
index 0000000..75ed02f
--- /dev/null
+++ b/arch/c6x/boot/Makefile
@@ -0,0 +1,22 @@
+#
+# Makefile for bootable kernel images
+#
+
+hostprogs-y := install-dtb
+
+OBJCOPYFLAGS_vmlinux.bin := -O binary
+$(obj)/vmlinux.bin: vmlinux FORCE
+ $(call if_changed,objcopy)
+
+quiet_cmd_install_dtb = INSTALL_DTB $@
+cmd_install_dtb = $(obj)/install-dtb $(obj)/$*.dtb vmlinux $@
+
+$(obj)/dtbImage.%: vmlinux $(obj)/install-dtb $(obj)/%.dtb
+ $(call cmd,install_dtb)
+
+clean-files := $(obj)/*.dtb
+
+DTC_FLAGS ?= -p 1024
+
+$(obj)/%.dtb: $(src)/dts/%.dts FORCE
+ $(call cmd,dtc)
diff --git a/arch/c6x/configs/dsk6455_defconfig b/arch/c6x/configs/dsk6455_defconfig
new file mode 100644
index 0000000..cf8bc10
--- /dev/null
+++ b/arch/c6x/configs/dsk6455_defconfig
@@ -0,0 +1,43 @@
+CONFIG_EXPERIMENTAL=y
+# CONFIG_LOCALVERSION_AUTO is not set
+CONFIG_SYSVIPC=y
+CONFIG_SPARSE_IRQ=y
+CONFIG_LOG_BUF_SHIFT=14
+CONFIG_NAMESPACES=y
+# CONFIG_UTS_NS is not set
+# CONFIG_USER_NS is not set
+# CONFIG_PID_NS is not set
+CONFIG_BLK_DEV_INITRD=y
+CONFIG_CC_OPTIMIZE_FOR_SIZE=y
+CONFIG_EXPERT=y
+# CONFIG_FUTEX is not set
+# CONFIG_SLUB_DEBUG is not set
+CONFIG_MODULES=y
+CONFIG_MODULE_FORCE_LOAD=y
+CONFIG_MODULE_UNLOAD=y
+CONFIG_MODULE_FORCE_UNLOAD=y
+CONFIG_CMDLINE_BOOL=y
+CONFIG_CMDLINE=""
+CONFIG_NO_HZ=y
+CONFIG_HIGH_RES_TIMERS=y
+CONFIG_BLK_DEV_LOOP=y
+CONFIG_BLK_DEV_RAM=y
+CONFIG_BLK_DEV_RAM_COUNT=2
+CONFIG_BLK_DEV_RAM_SIZE=17000
+CONFIG_MISC_DEVICES=y
+# CONFIG_INPUT is not set
+# CONFIG_SERIO is not set
+# CONFIG_VT is not set
+# CONFIG_HW_RANDOM is not set
+# CONFIG_HWMON is not set
+# CONFIG_USB_SUPPORT is not set
+# CONFIG_IOMMU_SUPPORT is not set
+# CONFIG_MISC_FILESYSTEMS is not set
+CONFIG_CRC16=y
+# CONFIG_ENABLE_MUST_CHECK is not set
+# CONFIG_SCHED_DEBUG is not set
+# CONFIG_DEBUG_BUGVERBOSE is not set
+CONFIG_MTD=y
+CONFIG_MTD_CFI=y
+CONFIG_MTD_CFI_AMDSTD=y
+CONFIG_MTD_PHYSMAP_OF=y
diff --git a/arch/c6x/configs/evmc6457_defconfig b/arch/c6x/configs/evmc6457_defconfig
new file mode 100644
index 0000000..d2eff63
--- /dev/null
+++ b/arch/c6x/configs/evmc6457_defconfig
@@ -0,0 +1,40 @@
+CONFIG_EXPERIMENTAL=y
+# CONFIG_LOCALVERSION_AUTO is not set
+CONFIG_SYSVIPC=y
+CONFIG_SPARSE_IRQ=y
+CONFIG_LOG_BUF_SHIFT=14
+CONFIG_NAMESPACES=y
+# CONFIG_UTS_NS is not set
+# CONFIG_USER_NS is not set
+# CONFIG_PID_NS is not set
+CONFIG_BLK_DEV_INITRD=y
+CONFIG_CC_OPTIMIZE_FOR_SIZE=y
+CONFIG_EXPERT=y
+# CONFIG_FUTEX is not set
+# CONFIG_SLUB_DEBUG is not set
+CONFIG_MODULES=y
+CONFIG_MODULE_FORCE_LOAD=y
+CONFIG_MODULE_UNLOAD=y
+CONFIG_MODULE_FORCE_UNLOAD=y
+CONFIG_CMDLINE_BOOL=y
+CONFIG_CMDLINE=""
+CONFIG_BOARD_EVM6457=y
+CONFIG_NO_HZ=y
+CONFIG_HIGH_RES_TIMERS=y
+CONFIG_BLK_DEV_LOOP=y
+CONFIG_BLK_DEV_RAM=y
+CONFIG_BLK_DEV_RAM_COUNT=2
+CONFIG_BLK_DEV_RAM_SIZE=17000
+CONFIG_MISC_DEVICES=y
+# CONFIG_INPUT is not set
+# CONFIG_SERIO is not set
+# CONFIG_VT is not set
+# CONFIG_HW_RANDOM is not set
+# CONFIG_HWMON is not set
+# CONFIG_USB_SUPPORT is not set
+# CONFIG_IOMMU_SUPPORT is not set
+# CONFIG_MISC_FILESYSTEMS is not set
+CONFIG_CRC16=y
+# CONFIG_ENABLE_MUST_CHECK is not set
+# CONFIG_SCHED_DEBUG is not set
+# CONFIG_DEBUG_BUGVERBOSE is not set
diff --git a/arch/c6x/configs/evmc6472_defconfig b/arch/c6x/configs/evmc6472_defconfig
new file mode 100644
index 0000000..aa81c58
--- /dev/null
+++ b/arch/c6x/configs/evmc6472_defconfig
@@ -0,0 +1,41 @@
+CONFIG_EXPERIMENTAL=y
+# CONFIG_LOCALVERSION_AUTO is not set
+CONFIG_SYSVIPC=y
+CONFIG_SPARSE_IRQ=y
+CONFIG_LOG_BUF_SHIFT=14
+CONFIG_NAMESPACES=y
+# CONFIG_UTS_NS is not set
+# CONFIG_USER_NS is not set
+# CONFIG_PID_NS is not set
+CONFIG_BLK_DEV_INITRD=y
+CONFIG_CC_OPTIMIZE_FOR_SIZE=y
+CONFIG_EXPERT=y
+# CONFIG_FUTEX is not set
+# CONFIG_SLUB_DEBUG is not set
+CONFIG_MODULES=y
+CONFIG_MODULE_FORCE_LOAD=y
+CONFIG_MODULE_UNLOAD=y
+CONFIG_MODULE_FORCE_UNLOAD=y
+CONFIG_CMDLINE_BOOL=y
+CONFIG_CMDLINE=""
+# CONFIG_CMDLINE_FORCE is not set
+CONFIG_BOARD_EVM6472=y
+CONFIG_NO_HZ=y
+CONFIG_HIGH_RES_TIMERS=y
+CONFIG_BLK_DEV_LOOP=y
+CONFIG_BLK_DEV_RAM=y
+CONFIG_BLK_DEV_RAM_COUNT=2
+CONFIG_BLK_DEV_RAM_SIZE=17000
+CONFIG_MISC_DEVICES=y
+# CONFIG_INPUT is not set
+# CONFIG_SERIO is not set
+# CONFIG_VT is not set
+# CONFIG_HW_RANDOM is not set
+# CONFIG_HWMON is not set
+# CONFIG_USB_SUPPORT is not set
+# CONFIG_IOMMU_SUPPORT is not set
+# CONFIG_MISC_FILESYSTEMS is not set
+CONFIG_CRC16=y
+# CONFIG_ENABLE_MUST_CHECK is not set
+# CONFIG_SCHED_DEBUG is not set
+# CONFIG_DEBUG_BUGVERBOSE is not set
diff --git a/arch/c6x/configs/evmc6474_defconfig b/arch/c6x/configs/evmc6474_defconfig
new file mode 100644
index 0000000..ceee6fc
--- /dev/null
+++ b/arch/c6x/configs/evmc6474_defconfig
@@ -0,0 +1,41 @@
+CONFIG_EXPERIMENTAL=y
+# CONFIG_LOCALVERSION_AUTO is not set
+CONFIG_SYSVIPC=y
+CONFIG_SPARSE_IRQ=y
+CONFIG_LOG_BUF_SHIFT=14
+CONFIG_NAMESPACES=y
+# CONFIG_UTS_NS is not set
+# CONFIG_USER_NS is not set
+# CONFIG_PID_NS is not set
+CONFIG_BLK_DEV_INITRD=y
+CONFIG_CC_OPTIMIZE_FOR_SIZE=y
+CONFIG_EXPERT=y
+# CONFIG_FUTEX is not set
+# CONFIG_SLUB_DEBUG is not set
+CONFIG_MODULES=y
+CONFIG_MODULE_FORCE_LOAD=y
+CONFIG_MODULE_UNLOAD=y
+CONFIG_MODULE_FORCE_UNLOAD=y
+CONFIG_CMDLINE_BOOL=y
+CONFIG_CMDLINE=""
+# CONFIG_CMDLINE_FORCE is not set
+CONFIG_BOARD_EVM6474=y
+CONFIG_NO_HZ=y
+CONFIG_HIGH_RES_TIMERS=y
+CONFIG_BLK_DEV_LOOP=y
+CONFIG_BLK_DEV_RAM=y
+CONFIG_BLK_DEV_RAM_COUNT=2
+CONFIG_BLK_DEV_RAM_SIZE=17000
+CONFIG_MISC_DEVICES=y
+# CONFIG_INPUT is not set
+# CONFIG_SERIO is not set
+# CONFIG_VT is not set
+# CONFIG_HW_RANDOM is not set
+# CONFIG_HWMON is not set
+# CONFIG_USB_SUPPORT is not set
+# CONFIG_IOMMU_SUPPORT is not set
+# CONFIG_MISC_FILESYSTEMS is not set
+CONFIG_CRC16=y
+# CONFIG_ENABLE_MUST_CHECK is not set
+# CONFIG_SCHED_DEBUG is not set
+# CONFIG_DEBUG_BUGVERBOSE is not set
diff --git a/arch/c6x/include/asm/Kbuild b/arch/c6x/include/asm/Kbuild
new file mode 100644
index 0000000..87815f8
--- /dev/null
+++ b/arch/c6x/include/asm/Kbuild
@@ -0,0 +1,55 @@
+include include/asm-generic/Kbuild.asm
+
+generic-y += atomic.h
+generic-y += auxvec.h
+generic-y += bitsperlong.h
+generic-y += bug.h
+generic-y += bugs.h
+generic-y += cputime.h
+generic-y += current.h
+generic-y += device.h
+generic-y += div64.h
+generic-y += dma.h
+generic-y += emergency-restart.h
+generic-y += errno.h
+generic-y += fb.h
+generic-y += fcntl.h
+generic-y += futex.h
+generic-y += hw_irq.h
+generic-y += io.h
+generic-y += ioctl.h
+generic-y += ioctls.h
+generic-y += ipcbuf.h
+generic-y += irq_regs.h
+generic-y += kdebug.h
+generic-y += kmap_types.h
+generic-y += local.h
+generic-y += mman.h
+generic-y += mmu_context.h
+generic-y += msgbuf.h
+generic-y += param.h
+generic-y += pci.h
+generic-y += percpu.h
+generic-y += pgalloc.h
+generic-y += poll.h
+generic-y += posix_types.h
+generic-y += resource.h
+generic-y += scatterlist.h
+generic-y += segment.h
+generic-y += sembuf.h
+generic-y += shmbuf.h
+generic-y += shmparam.h
+generic-y += siginfo.h
+generic-y += socket.h
+generic-y += sockios.h
+generic-y += stat.h
+generic-y += statfs.h
+generic-y += swab.h
+generic-y += termbits.h
+generic-y += termios.h
+generic-y += tlbflush.h
+generic-y += topology.h
+generic-y += types.h
+generic-y += ucontext.h
+generic-y += user.h
+generic-y += vga.h
diff --git a/arch/c6x/kernel/Makefile b/arch/c6x/kernel/Makefile
new file mode 100644
index 0000000..580a515
--- /dev/null
+++ b/arch/c6x/kernel/Makefile
@@ -0,0 +1,12 @@
+#
+# Makefile for arch/c6x/kernel/
+#
+
+extra-y := head.o vmlinux.lds
+
+obj-y := process.o traps.o irq.o signal.o ptrace.o
+obj-y += setup.o sys_c6x.o time.o devicetree.o
+obj-y += switch_to.o entry.o vectors.o c6x_ksyms.o
+obj-y += soc.o dma.o
+
+obj-$(CONFIG_MODULES) += module.o
diff --git a/arch/c6x/kernel/vmlinux.lds.S b/arch/c6x/kernel/vmlinux.lds.S
new file mode 100644
index 0000000..1d81c4c
--- /dev/null
+++ b/arch/c6x/kernel/vmlinux.lds.S
@@ -0,0 +1,162 @@
+/*
+ * ld script for the c6x kernel
+ *
+ * Copyright (C) 2010, 2011 Texas Instruments Incorporated
+ * Mark Salter <msalter@redhat.com>
+ */
+#include <asm-generic/vmlinux.lds.h>
+#include <asm/thread_info.h>
+#include <asm/page.h>
+
+ENTRY(_c_int00)
+
+#if defined(CONFIG_CPU_BIG_ENDIAN)
+jiffies = jiffies_64 + 4;
+#else
+jiffies = jiffies_64;
+#endif
+
+#define READONLY_SEGMENT_START \
+ . = PAGE_OFFSET;
+#define READWRITE_SEGMENT_START \
+ . = ALIGN(128); \
+ _data_lma = .;
+
+SECTIONS
+{
+ /*
+ * Start kernel read only segment
+ */
+ READONLY_SEGMENT_START
+
+ .vectors :
+ {
+ _vectors_start = .;
+ *(.vectors)
+ . = ALIGN(0x400);
+ _vectors_end = .;
+ }
+
+ . = ALIGN(0x1000);
+ .cmdline :
+ {
+ *(.cmdline)
+ }
+
+ /*
+ * This section contains data which may be shared with other
+ * cores. It needs to be a fixed offset from PAGE_OFFSET
+ * regardless of kernel configuration.
+ */
+ .virtio_ipc_dev :
+ {
+ *(.virtio_ipc_dev)
+ }
+
+ . = ALIGN(PAGE_SIZE);
+ .init :
+ {
+ _stext = .;
+ _sinittext = .;
+ HEAD_TEXT
+ INIT_TEXT
+ _einittext = .;
+ }
+
+ __init_begin = _stext;
+ INIT_DATA_SECTION(16)
+
+ PERCPU_SECTION(128)
+
+ . = ALIGN(PAGE_SIZE);
+ __init_end = .;
+
+ .text :
+ {
+ _text = .;
+ TEXT_TEXT
+ SCHED_TEXT
+ LOCK_TEXT
+ IRQENTRY_TEXT
+ KPROBES_TEXT
+ *(.fixup)
+ *(.gnu.warning)
+ }
+
+ EXCEPTION_TABLE(16)
+ NOTES
+
+ RO_DATA_SECTION(PAGE_SIZE)
+ .const :
+ {
+ *(.const .const.* .gnu.linkonce.r.*)
+ *(.switch)
+ }
+
+ . = ALIGN (8) ;
+ __fdt_blob : AT(ADDR(__fdt_blob) - LOAD_OFFSET)
+ {
+ _fdt_start = . ; /* place for fdt blob */
+ *(__fdt_blob) ; /* Any link-placed DTB */
+ BYTE(0); /* section always has contents */
+ . = _fdt_start + 0x4000; /* Pad up to 16kbyte */
+ _fdt_end = . ;
+ }
+
+ _etext = .;
+
+ /*
+ * Start kernel read-write segment.
+ */
+ READWRITE_SEGMENT_START
+ _sdata = .;
+
+ .fardata : AT(ADDR(.fardata) - LOAD_OFFSET)
+ {
+ INIT_TASK_DATA(THREAD_SIZE)
+ NOSAVE_DATA
+ PAGE_ALIGNED_DATA(PAGE_SIZE)
+ CACHELINE_ALIGNED_DATA(128)
+ READ_MOSTLY_DATA(128)
+ DATA_DATA
+ CONSTRUCTORS
+ *(.data1)
+ *(.fardata .fardata.*)
+ *(.data.debug_bpt)
+ }
+
+ .neardata ALIGN(8) : AT(ADDR(.neardata) - LOAD_OFFSET)
+ {
+ *(.neardata2 .neardata2.* .gnu.linkonce.s2.*)
+ *(.neardata .neardata.* .gnu.linkonce.s.*)
+ . = ALIGN(8);
+ }
+
+ _edata = .;
+
+ __bss_start = .;
+ SBSS(8)
+ BSS(8)
+ .far :
+ {
+ . = ALIGN(8);
+ *(.dynfar)
+ *(.far .far.* .gnu.linkonce.b.*)
+ . = ALIGN(8);
+ }
+ __bss_stop = .;
+
+ _end = .;
+
+ DWARF_DEBUG
+
+ /DISCARD/ :
+ {
+ EXIT_TEXT
+ EXIT_DATA
+ EXIT_CALL
+ *(.discard)
+ *(.discard.*)
+ *(.interp)
+ }
+}
diff --git a/arch/c6x/lib/Makefile b/arch/c6x/lib/Makefile
new file mode 100644
index 0000000..f39d4f0
--- /dev/null
+++ b/arch/c6x/lib/Makefile
@@ -0,0 +1,7 @@
+#
+# Makefile for arch/c6x/lib/
+#
+
+lib-y := divu.o divi.o pop_rts.o push_rts.o remi.o remu.o strasgi.o llshru.o
+lib-y += llshr.o llshl.o negll.o mpyll.o divremi.o divremu.o
+lib-y += csum_64plus.o memcpy_64plus.o strasgi_64plus.o
diff --git a/arch/c6x/mm/Makefile b/arch/c6x/mm/Makefile
new file mode 100644
index 0000000..136a975
--- /dev/null
+++ b/arch/c6x/mm/Makefile
@@ -0,0 +1,5 @@
+#
+# Makefile for the linux c6x-specific parts of the memory manager.
+#
+
+obj-y := init.o dma-coherent.o
diff --git a/arch/c6x/platforms/Kconfig b/arch/c6x/platforms/Kconfig
new file mode 100644
index 0000000..6e36375
--- /dev/null
+++ b/arch/c6x/platforms/Kconfig
@@ -0,0 +1,41 @@
+
+config SOC_TMS320C6455
+ bool "TMS320C6455"
+ default n
+
+config SOC_TMS320C6457
+ bool "TMS320C6457"
+ default n
+
+config SOC_TMS320C6472
+ bool "TMS320C6472"
+ default n
+
+config SOC_TMS320C6474
+ bool "TMS320C6474"
+ default n
+
+comment "Board Selection"
+choice
+ prompt "Board"
+ help
+ This option specifies the specific board for which the kernel will be
+ compiled.
+
+config BOARD_DSK6455
+ bool "DSK6455"
+ select SOC_TMS320C6455
+
+config BOARD_EVM6457
+ bool "EVM6472"
+ select SOC_TMS320C6457
+
+config BOARD_EVM6472
+ bool "EVM6472"
+ select SOC_TMS320C6472
+
+config BOARD_EVM6474
+ bool "EVM6474"
+ select SOC_TMS320C6474
+
+endchoice
diff --git a/arch/c6x/platforms/Makefile b/arch/c6x/platforms/Makefile
new file mode 100644
index 0000000..9a95b9b
--- /dev/null
+++ b/arch/c6x/platforms/Makefile
@@ -0,0 +1,12 @@
+#
+# Makefile for arch/c6x/platforms
+#
+# Copyright 2010, 2011 Texas Instruments Incorporated
+#
+
+obj-y = platform.o cache.o megamod-pic.o pll.o plldata.o timer64.o
+obj-y += dscr.o
+
+# SoC objects
+obj-$(CONFIG_SOC_TMS320C6455) += emif.o
+obj-$(CONFIG_SOC_TMS320C6457) += emif.o
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v3 05/24] C6X: build infrastructure
2011-09-27 20:29 ` [PATCH v3 05/24] C6X: build infrastructure Mark Salter
@ 2011-09-28 13:23 ` Arnd Bergmann
2011-09-28 14:32 ` Mark Salter
0 siblings, 1 reply; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-28 13:23 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch
On Tuesday 27 September 2011, Mark Salter wrote:
> +comment "Board Selection"
> +choice
> + prompt "Board"
> + help
> + This option specifies the specific board for which the kernel will be
> + compiled.
> +
> +config BOARD_DSK6455
> + bool "DSK6455"
> + select SOC_TMS320C6455
> +
> +config BOARD_EVM6457
> + bool "EVM6472"
> + select SOC_TMS320C6457
> +
> +config BOARD_EVM6472
> + bool "EVM6472"
> + select SOC_TMS320C6472
> +
> +config BOARD_EVM6474
> + bool "EVM6474"
> + select SOC_TMS320C6474
> +
> +endchoice
This really wants to be a flat list instead of a "choice" statement,
so you can build a kernel for multiple boards at once. Is there a
reason why this is not possible right now?
Also, how different are these boards and socs still? Since you are
based on device trees for hardware configuration, do you actually
need the options at all?
Arnd
^ permalink raw reply [flat|nested] 66+ messages in thread
* Re: [PATCH v3 05/24] C6X: build infrastructure
2011-09-28 13:23 ` Arnd Bergmann
@ 2011-09-28 14:32 ` Mark Salter
2011-09-28 14:57 ` Arnd Bergmann
0 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-28 14:32 UTC (permalink / raw)
To: Arnd Bergmann; +Cc: linux-kernel, linux-arch
On Wed, 2011-09-28 at 15:23 +0200, Arnd Bergmann wrote:
> On Tuesday 27 September 2011, Mark Salter wrote:
> > +comment "Board Selection"
> > +choice
> > + prompt "Board"
> > + help
> > + This option specifies the specific board for which the kernel will be
> > + compiled.
> > +
> > +config BOARD_DSK6455
> > + bool "DSK6455"
> > + select SOC_TMS320C6455
> > +
> > +config BOARD_EVM6457
> > + bool "EVM6472"
> > + select SOC_TMS320C6457
> > +
> > +config BOARD_EVM6472
> > + bool "EVM6472"
> > + select SOC_TMS320C6472
> > +
> > +config BOARD_EVM6474
> > + bool "EVM6474"
> > + select SOC_TMS320C6474
> > +
> > +endchoice
>
> This really wants to be a flat list instead of a "choice" statement,
> so you can build a kernel for multiple boards at once. Is there a
> reason why this is not possible right now?
This is just left over from the time before device trees. I'll take
the BOARD_foo configs out completely. Its just a matter of selecting
the SOC in the defconfig rather than the board.
>
> Also, how different are these boards and socs still? Since you are
> based on device trees for hardware configuration, do you actually
> need the options at all?
The BOARD configs can go. The SOC configs will go out when clkdev
stuff gets reworked when generic clock bindings are available.
--Mark
^ permalink raw reply [flat|nested] 66+ messages in thread
* Re: [PATCH v3 05/24] C6X: build infrastructure
2011-09-28 14:32 ` Mark Salter
@ 2011-09-28 14:57 ` Arnd Bergmann
0 siblings, 0 replies; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-28 14:57 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch
On Wednesday 28 September 2011, Mark Salter wrote:
> > Also, how different are these boards and socs still? Since you are
> > based on device trees for hardware configuration, do you actually
> > need the options at all?
>
> The BOARD configs can go. The SOC configs will go out when clkdev
> stuff gets reworked when generic clock bindings are available.
>
Ok, sounds good.
Arnd
^ permalink raw reply [flat|nested] 66+ messages in thread
* [PATCH v3 06/24] C6X: early boot code
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (4 preceding siblings ...)
2011-09-27 20:29 ` [PATCH v3 05/24] C6X: build infrastructure Mark Salter
@ 2011-09-27 20:29 ` Mark Salter
2011-09-28 13:26 ` Arnd Bergmann
2011-09-27 20:29 ` [PATCH v3 07/24] C6X: devicetree support Mark Salter
` (19 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:29 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
This patch provides the early boot code for C6X architecture. There is a
16 entry vector table which is used to direct reset and interrupt events. The
vector table entries contain a small amount of code (maximum of 8 opcodes)
which simply branches to the actual event handling code.
The head.S code simply clears BSS, setups up a few control registers, and calls
machine_init followed by start_kernel. The machine_init code in setup.c does
the early flat tree parsing (memory, commandline, etc). At setup_arch time, the
code does the usual memory setup and minimally scans the devicetree for any
needed information.
Signed-off-by: Mark Salter <msalter@redhat.com>
---
arch/c6x/kernel/head.S | 84 ++++++++
arch/c6x/kernel/setup.c | 499 +++++++++++++++++++++++++++++++++++++++++++++
arch/c6x/kernel/vectors.S | 81 ++++++++
3 files changed, 664 insertions(+), 0 deletions(-)
create mode 100644 arch/c6x/kernel/head.S
create mode 100644 arch/c6x/kernel/setup.c
create mode 100644 arch/c6x/kernel/vectors.S
diff --git a/arch/c6x/kernel/head.S b/arch/c6x/kernel/head.S
new file mode 100644
index 0000000..133eab6
--- /dev/null
+++ b/arch/c6x/kernel/head.S
@@ -0,0 +1,84 @@
+;
+; Port on Texas Instruments TMS320C6x architecture
+;
+; Copyright (C) 2004, 2009, 2010, 2011 Texas Instruments Incorporated
+; Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+;
+; This program is free software; you can redistribute it and/or modify
+; it under the terms of the GNU General Public License version 2 as
+; published by the Free Software Foundation.
+;
+#include <linux/linkage.h>
+#include <linux/of_fdt.h>
+#include <asm/asm-offsets.h>
+
+ __HEAD
+ENTRY(_c_int00)
+ ;; Save magic and pointer
+ MV .S1 A4,A10
+ MV .S2 B4,B10
+ MVKL .S2 __bss_start,B5
+ MVKH .S2 __bss_start,B5
+ MVKL .S2 __bss_stop,B6
+ MVKH .S2 __bss_stop,B6
+ SUB .L2 B6,B5,B6 ; bss size
+
+ ;; Set the stack pointer
+ MVKL .S2 current_ksp,B0
+ MVKH .S2 current_ksp,B0
+ LDW .D2T2 *B0,B15
+
+ ;; clear bss
+ SHR .S2 B6,3,B0 ; number of dwords to clear
+ ZERO .L2 B13
+ ZERO .L2 B12
+bss_loop:
+ BDEC .S2 bss_loop,B0
+ NOP 3
+ CMPLT .L2 B0,0,B1
+ [!B1] STDW .D2T2 B13:B12,*B5++[1]
+
+ NOP 4
+ AND .D2 ~7,B15,B15
+
+ ;; Clear GIE and PGIE
+ MVC .S2 CSR,B2
+ CLR .S2 B2,0,1,B2
+ MVC .S2 B2,CSR
+ MVC .S2 TSR,B2
+ CLR .S2 B2,0,1,B2
+ MVC .S2 B2,TSR
+ MVC .S2 ITSR,B2
+ CLR .S2 B2,0,1,B2
+ MVC .S2 B2,ITSR
+ MVC .S2 NTSR,B2
+ CLR .S2 B2,0,1,B2
+ MVC .S2 B2,NTSR
+
+ ;; pass DTB pointer to machine_init (or zero if none)
+ MVKL .S1 OF_DT_HEADER,A0
+ MVKH .S1 OF_DT_HEADER,A0
+ CMPEQ .L1 A10,A0,A0
+ [A0] MV .S1X B10,A4
+ [!A0] MVK .S1 0,A4
+
+#ifdef CONFIG_C6X_BIG_KERNEL
+ MVKL .S1 machine_init,A0
+ MVKH .S1 machine_init,A0
+ B .S2X A0
+ ADDKPC .S2 0f,B3,4
+0:
+#else
+ CALLP .S2 machine_init,B3
+#endif
+
+ ;; Jump to Linux init
+#ifdef CONFIG_C6X_BIG_KERNEL
+ MVKL .S1 start_kernel,A0
+ MVKH .S1 start_kernel,A0
+ B .S2X A0
+#else
+ B .S2 start_kernel
+#endif
+ NOP 5
+L1: BNOP .S2 L1,5
diff --git a/arch/c6x/kernel/setup.c b/arch/c6x/kernel/setup.c
new file mode 100644
index 0000000..b6041c4
--- /dev/null
+++ b/arch/c6x/kernel/setup.c
@@ -0,0 +1,499 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2006, 2009, 2010, 2011 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/dma-mapping.h>
+#include <linux/memblock.h>
+#include <linux/seq_file.h>
+#include <linux/bootmem.h>
+#include <linux/clkdev.h>
+#include <linux/initrd.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/of_fdt.h>
+#include <linux/string.h>
+#include <linux/errno.h>
+#include <linux/cache.h>
+#include <linux/delay.h>
+#include <linux/sched.h>
+#include <linux/clk.h>
+#include <linux/fs.h>
+#include <linux/of.h>
+
+
+#include <asm/sections.h>
+#include <asm/div64.h>
+#include <asm/setup.h>
+#include <asm/dscr.h>
+#include <asm/clock.h>
+#include <asm/soc.h>
+
+static const char *c6x_soc_name;
+
+int c6x_num_cores;
+EXPORT_SYMBOL(c6x_num_cores);
+
+unsigned int c6x_silicon_rev;
+EXPORT_SYMBOL(c6x_silicon_rev);
+
+/*
+ * Device status register. This holds information
+ * about device configuration needed by some drivers.
+ */
+unsigned int c6x_devstat;
+EXPORT_SYMBOL(c6x_devstat);
+
+/*
+ * Some SoCs have fuse registers holding a unique MAC
+ * address. This is parsed out of the device tree with
+ * the resulting MAC being held here.
+ */
+unsigned char c6x_fuse_mac[6];
+
+unsigned long memory_start;
+unsigned long memory_end;
+
+unsigned long ram_start;
+unsigned long ram_end;
+
+/* Uncached memory for DMA consistent use (memdma=) */
+static unsigned long dma_start __initdata;
+static unsigned long dma_size __initdata;
+
+char c6x_command_line[COMMAND_LINE_SIZE];
+
+#if defined(CONFIG_CMDLINE_BOOL)
+static const char default_command_line[COMMAND_LINE_SIZE] __section(.cmdline) =
+ CONFIG_CMDLINE;
+#endif
+
+struct cpuinfo_c6x {
+ const char *cpu_name;
+ const char *cpu_voltage;
+ const char *mmu;
+ const char *fpu;
+ char *cpu_rev;
+ unsigned int core_id;
+ char __cpu_rev[5];
+};
+
+static DEFINE_PER_CPU(struct cpuinfo_c6x, cpu_data);
+
+unsigned int ticks_per_ns_scaled;
+EXPORT_SYMBOL(ticks_per_ns_scaled);
+
+unsigned int c6x_core_freq;
+EXPORT_SYMBOL(c6x_core_freq);
+
+static void __init get_cpuinfo(void)
+{
+ unsigned cpu_id, rev_id, csr;
+ struct clk *coreclk = clk_get_sys(NULL, "core");
+ unsigned long core_khz;
+ u64 tmp;
+ struct cpuinfo_c6x *p;
+ struct device_node *node, *np;
+
+ p = &per_cpu(cpu_data, smp_processor_id());
+
+ if (!IS_ERR(coreclk))
+ c6x_core_freq = clk_get_rate(coreclk);
+ else {
+ printk(KERN_WARNING
+ "Cannot find core clock frequency. Using 700MHz\n");
+ c6x_core_freq = 700000000;
+ }
+
+ core_khz = c6x_core_freq / 1000;
+
+ tmp = (uint64_t)core_khz << C6X_NDELAY_SCALE;
+ do_div(tmp, 1000000);
+ ticks_per_ns_scaled = tmp;
+
+ csr = get_creg(CSR);
+ cpu_id = csr >> 24;
+ rev_id = (csr >> 16) & 0xff;
+
+ p->mmu = "none";
+ p->fpu = "none";
+ p->cpu_voltage = "unknown";
+
+ switch (cpu_id) {
+ case 0:
+ p->cpu_name = "C67x";
+ p->fpu = "yes";
+ break;
+ case 2:
+ p->cpu_name = "C62x";
+ break;
+ case 8:
+ p->cpu_name = "C64x";
+ break;
+ case 12:
+ p->cpu_name = "C64x";
+ break;
+ case 16:
+ p->cpu_name = "C64x+";
+ p->cpu_voltage = "1.2";
+ break;
+ default:
+ p->cpu_name = "unknown";
+ break;
+ }
+
+ if (cpu_id < 16) {
+ switch (rev_id) {
+ case 0x1:
+ if (cpu_id > 8) {
+ p->cpu_rev = "DM640/DM641/DM642/DM643";
+ p->cpu_voltage = "1.2 - 1.4";
+ } else {
+ p->cpu_rev = "C6201";
+ p->cpu_voltage = "2.5";
+ }
+ break;
+ case 0x2:
+ p->cpu_rev = "C6201B/C6202/C6211";
+ p->cpu_voltage = "1.8";
+ break;
+ case 0x3:
+ p->cpu_rev = "C6202B/C6203/C6204/C6205";
+ p->cpu_voltage = "1.5";
+ break;
+ case 0x201:
+ p->cpu_rev = "C6701 revision 0 (early CPU)";
+ p->cpu_voltage = "1.8";
+ break;
+ case 0x202:
+ p->cpu_rev = "C6701/C6711/C6712";
+ p->cpu_voltage = "1.8";
+ break;
+ case 0x801:
+ p->cpu_rev = "C64x";
+ p->cpu_voltage = "1.5";
+ break;
+ default:
+ p->cpu_rev = "unknown";
+ }
+ } else {
+ p->cpu_rev = p->__cpu_rev;
+ snprintf(p->__cpu_rev, sizeof(p->__cpu_rev), "0x%x", cpu_id);
+ }
+
+ p->core_id = get_coreid();
+
+ node = of_find_node_by_name(NULL, "cpus");
+ if (node) {
+ for_each_child_of_node(node, np)
+ if (!strcmp("cpu", np->name))
+ ++c6x_num_cores;
+ of_node_put(node);
+ }
+
+ node = of_find_node_by_name(NULL, "soc");
+ if (node) {
+ if (of_property_read_string(node, "model", &c6x_soc_name))
+ c6x_soc_name = "unknown";
+ of_node_put(node);
+ } else
+ c6x_soc_name = "unknown";
+
+ printk(KERN_INFO "CPU%d: %s rev %s, %s volts, %uMHz\n",
+ p->core_id, p->cpu_name, p->cpu_rev,
+ p->cpu_voltage, c6x_core_freq / 1000000);
+}
+
+/*
+ * Early parsing of the command line
+ */
+static u32 mem_size __initdata;
+
+/* "mem=" parsing. */
+static int __init early_mem(char *p)
+{
+ if (!p)
+ return -EINVAL;
+
+ mem_size = memparse(p, &p);
+ /* don't remove all of memory when handling "mem={invalid}" */
+ if (mem_size == 0)
+ return -EINVAL;
+
+ return 0;
+}
+early_param("mem", early_mem);
+
+/* "memdma=<size>[@<address>]" parsing. */
+static int __init early_memdma(char *p)
+{
+ if (!p)
+ return -EINVAL;
+
+ dma_size = memparse(p, &p);
+ if (*p == '@')
+ dma_start = memparse(p, &p);
+
+ return 0;
+}
+early_param("memdma", early_memdma);
+
+int __init c6x_add_memory(phys_addr_t start, unsigned long size)
+{
+ static int ram_found __initdata;
+
+ /* We only handle one bank (the one with PAGE_OFFSET) for now */
+ if (ram_found)
+ return -EINVAL;
+
+ if (start > PAGE_OFFSET || PAGE_OFFSET >= (start + size))
+ return 0;
+
+ ram_start = start;
+ ram_end = start + size;
+
+ ram_found = 1;
+ return 0;
+}
+
+/*
+ * Do early machine setup and device tree parsing. This is called very
+ * early on the boot process.
+ */
+notrace void __init machine_init(unsigned long dt_ptr)
+{
+ struct boot_param_header *dtb = __va(dt_ptr);
+ struct boot_param_header *fdt = (struct boot_param_header *)_fdt_start;
+
+ /* interrupts must be masked */
+ set_creg(IER, 2);
+
+ /*
+ * Set the Interrupt Service Table (IST) to the beginning of the
+ * vector table.
+ */
+ set_ist(_vectors_start);
+
+ lockdep_init();
+
+ /*
+ * dtb is passed in from bootloader.
+ * fdt is linked in blob.
+ */
+ if (dtb && dtb != fdt)
+ fdt = dtb;
+
+ /* Do some early initialization based on the flat device tree */
+ early_init_devtree(fdt);
+
+ /* parse_early_param needs a boot_command_line */
+ strlcpy(boot_command_line, c6x_command_line, COMMAND_LINE_SIZE);
+ parse_early_param();
+}
+
+void __init setup_arch(char **cmdline_p)
+{
+ int bootmap_size;
+ struct memblock_region *reg;
+
+ printk(KERN_INFO "Initializing kernel\n");
+
+ /* Initialize command line */
+ *cmdline_p = c6x_command_line;
+
+ memblock_init();
+
+ memory_end = ram_end;
+ memory_end &= ~(PAGE_SIZE - 1);
+
+ if (mem_size && (PAGE_OFFSET + PAGE_ALIGN(mem_size)) < memory_end)
+ memory_end = PAGE_OFFSET + PAGE_ALIGN(mem_size);
+
+ /* add block that this kernel can use */
+ memblock_add(PAGE_OFFSET, memory_end - PAGE_OFFSET);
+
+ /* reserve kernel text/data/bss */
+ memblock_reserve(PAGE_OFFSET,
+ PAGE_ALIGN((unsigned long)&_end - PAGE_OFFSET));
+
+ if (dma_size) {
+ /* align to cacheability granularity */
+ dma_size = CACHE_REGION_END(dma_size);
+
+ if (!dma_start)
+ dma_start = memory_end - dma_size;
+
+ /* align to cacheability granularity */
+ dma_start = CACHE_REGION_START(dma_start);
+
+ /* reserve DMA memory taken from kernel memory */
+ if (memblock_is_region_memory(dma_start, dma_size))
+ memblock_reserve(dma_start, dma_size);
+ }
+
+ memory_start = PAGE_ALIGN((unsigned int) &_end);
+
+ printk(KERN_INFO "Memory Start=%08lx, Memory End=%08lx\n",
+ memory_start, memory_end);
+
+#ifdef CONFIG_BLK_DEV_INITRD
+ /*
+ * Reserve initrd memory if in kernel memory.
+ */
+ if (initrd_start < initrd_end)
+ if (memblock_is_region_memory(initrd_start,
+ initrd_end - initrd_start))
+ memblock_reserve(initrd_start,
+ initrd_end - initrd_start);
+#endif
+
+ init_mm.start_code = (unsigned long) &_stext;
+ init_mm.end_code = (unsigned long) &_etext;
+ init_mm.end_data = memory_start;
+ init_mm.brk = memory_start;
+
+ /*
+ * Give all the memory to the bootmap allocator, tell it to put the
+ * boot mem_map at the start of memory
+ */
+ bootmap_size = init_bootmem_node(NODE_DATA(0),
+ memory_start >> PAGE_SHIFT,
+ PAGE_OFFSET >> PAGE_SHIFT,
+ memory_end >> PAGE_SHIFT);
+ memblock_reserve(memory_start, bootmap_size);
+
+ memblock_analyze();
+ unflatten_device_tree();
+
+ c6x_cache_init();
+
+ /* Set the whole external memory as non-cacheable */
+ disable_caching(ram_start, ram_end - 1);
+
+ /* Set caching of external RAM used by Linux */
+ for_each_memblock(memory, reg)
+ enable_caching(CACHE_REGION_START(reg->base),
+ CACHE_REGION_START(reg->base + reg->size - 1));
+
+#ifdef CONFIG_BLK_DEV_INITRD
+ /*
+ * Enable caching for initrd which falls outside kernel memory.
+ */
+ if (initrd_start < initrd_end) {
+ if (!memblock_is_region_memory(initrd_start,
+ initrd_end - initrd_start))
+ enable_caching(CACHE_REGION_START(initrd_start),
+ CACHE_REGION_START(initrd_end - 1));
+ }
+#endif
+
+ /*
+ * Disable caching for dma coherent memory taken from kernel memory.
+ */
+ if (dma_size && memblock_is_region_memory(dma_start, dma_size))
+ disable_caching(dma_start,
+ CACHE_REGION_START(dma_start + dma_size - 1));
+
+ /* Initialize the coherent memory allocator */
+ coherent_mem_init(dma_start, dma_size);
+
+ /*
+ * Free all memory as a starting point.
+ */
+ free_bootmem(PAGE_OFFSET, memory_end - PAGE_OFFSET);
+
+ /*
+ * Then reserve memory which is already being used.
+ */
+ for_each_memblock(reserved, reg) {
+ pr_debug("reserved - 0x%08x-0x%08x\n",
+ (u32) reg->base, (u32) reg->size);
+ reserve_bootmem(reg->base, reg->size, BOOTMEM_DEFAULT);
+ }
+
+ max_low_pfn = PFN_DOWN(memory_end);
+ min_low_pfn = PFN_UP(memory_start);
+ max_mapnr = max_low_pfn - min_low_pfn;
+
+ /* Get kmalloc into gear */
+ paging_init();
+
+ /*
+ * Probe for Device State Configuration Registers.
+ * We have to do this early in case timer needs to be enabled
+ * through DSCR.
+ */
+ dscr_probe();
+
+ /* We do this early for timer and core clock frequency */
+ c64x_setup_clocks();
+
+ /* Get CPU info */
+ get_cpuinfo();
+
+#if defined(CONFIG_VT) && defined(CONFIG_DUMMY_CONSOLE)
+ conswitchp = &dummy_con;
+#endif
+}
+
+#define cpu_to_ptr(n) ((void *)((long)(n)+1))
+#define ptr_to_cpu(p) ((long)(p) - 1)
+
+static int show_cpuinfo(struct seq_file *m, void *v)
+{
+ int n = ptr_to_cpu(v);
+ struct cpuinfo_c6x *p = &per_cpu(cpu_data, n);
+
+ if (n == 0) {
+ seq_printf(m,
+ "soc\t\t: %s\n"
+ "soc revision\t: 0x%x\n"
+ "soc cores\t: %d\n",
+ c6x_soc_name, c6x_silicon_rev, c6x_num_cores);
+ }
+
+ seq_printf(m,
+ "\n"
+ "processor\t: %d\n"
+ "cpu\t\t: %s\n"
+ "core revision\t: %s\n"
+ "core voltage\t: %s\n"
+ "core id\t\t: %d\n"
+ "mmu\t\t: %s\n"
+ "fpu\t\t: %s\n"
+ "cpu MHz\t\t: %u\n"
+ "bogomips\t: %lu.%02lu\n\n",
+ n,
+ p->cpu_name, p->cpu_rev, p->cpu_voltage,
+ p->core_id, p->mmu, p->fpu,
+ (c6x_core_freq + 500000) / 1000000,
+ (loops_per_jiffy/(500000/HZ)),
+ (loops_per_jiffy/(5000/HZ))%100);
+
+ return 0;
+}
+
+static void *c_start(struct seq_file *m, loff_t *pos)
+{
+ return *pos < nr_cpu_ids ? cpu_to_ptr(*pos) : NULL;
+}
+static void *c_next(struct seq_file *m, void *v, loff_t *pos)
+{
+ ++*pos;
+ return NULL;
+}
+static void c_stop(struct seq_file *m, void *v)
+{
+}
+
+const struct seq_operations cpuinfo_op = {
+ c_start,
+ c_stop,
+ c_next,
+ show_cpuinfo
+};
diff --git a/arch/c6x/kernel/vectors.S b/arch/c6x/kernel/vectors.S
new file mode 100644
index 0000000..c95c66f
--- /dev/null
+++ b/arch/c6x/kernel/vectors.S
@@ -0,0 +1,81 @@
+;
+; Port on Texas Instruments TMS320C6x architecture
+;
+; Copyright (C) 2004, 2006, 2009, 2010, 2011 Texas Instruments Incorporated
+; Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+;
+; This program is free software; you can redistribute it and/or modify
+; it under the terms of the GNU General Public License version 2 as
+; published by the Free Software Foundation.
+;
+; This section handles all the interrupt vector routines.
+; At RESET the processor sets up the DRAM timing parameters and
+; branches to the label _c_int00 which handles initialization for the C code.
+;
+
+#define ALIGNMENT 5
+
+ .macro IRQVEC name, handler
+ .align ALIGNMENT
+ .hidden \name
+ .global \name
+\name:
+#ifdef CONFIG_C6X_BIG_KERNEL
+ STW .D2T1 A0,*B15--[2]
+ || MVKL .S1 \handler,A0
+ MVKH .S1 \handler,A0
+ B .S2X A0
+ LDW .D2T1 *++B15[2],A0
+ NOP 4
+ NOP
+ NOP
+ .endm
+#else /* CONFIG_C6X_BIG_KERNEL */
+ B .S2 \handler
+ NOP
+ NOP
+ NOP
+ NOP
+ NOP
+ NOP
+ NOP
+ .endm
+#endif /* CONFIG_C6X_BIG_KERNEL */
+
+ .sect ".vectors","ax"
+ .align ALIGNMENT
+ .global RESET
+ .hidden RESET
+RESET:
+#ifdef CONFIG_C6X_BIG_KERNEL
+ MVKL .S1 _c_int00,A0 ; branch to _c_int00
+ MVKH .S1 _c_int00,A0
+ B .S2X A0
+#else
+ B .S2 _c_int00
+ NOP
+ NOP
+#endif
+ NOP
+ NOP
+ NOP
+ NOP
+ NOP
+
+
+ IRQVEC NMI,_nmi_handler ; NMI interrupt
+ IRQVEC AINT,_bad_interrupt ; reserved
+ IRQVEC MSGINT,_bad_interrupt ; reserved
+
+ IRQVEC INT4,_int4_handler
+ IRQVEC INT5,_int5_handler
+ IRQVEC INT6,_int6_handler
+ IRQVEC INT7,_int7_handler
+ IRQVEC INT8,_int8_handler
+ IRQVEC INT9,_int9_handler
+ IRQVEC INT10,_int10_handler
+ IRQVEC INT11,_int11_handler
+ IRQVEC INT12,_int12_handler
+ IRQVEC INT13,_int13_handler
+ IRQVEC INT14,_int14_handler
+ IRQVEC INT15,_int15_handler
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v3 06/24] C6X: early boot code
2011-09-27 20:29 ` [PATCH v3 06/24] C6X: early boot code Mark Salter
@ 2011-09-28 13:26 ` Arnd Bergmann
2011-09-28 14:06 ` Mark Salter
0 siblings, 1 reply; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-28 13:26 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch
On Tuesday 27 September 2011, Mark Salter wrote:
> +int c6x_num_cores;
> +EXPORT_SYMBOL(c6x_num_cores);
> +
> +unsigned int c6x_silicon_rev;
> +EXPORT_SYMBOL(c6x_silicon_rev);
> +
> +/*
> + * Device status register. This holds information
> + * about device configuration needed by some drivers.
> + */
> +unsigned int c6x_devstat;
> +EXPORT_SYMBOL(c6x_devstat);
> +unsigned int ticks_per_ns_scaled;
> +EXPORT_SYMBOL(ticks_per_ns_scaled);
> +
> +unsigned int c6x_core_freq;
> +EXPORT_SYMBOL(c6x_core_freq);
These all look like they are used internally (maybe not
ticks_per_ns_scaled) so it would be more appropriate to make
them EXPORT_SYMBOL_GPL.
Arnd
^ permalink raw reply [flat|nested] 66+ messages in thread
* Re: [PATCH v3 06/24] C6X: early boot code
2011-09-28 13:26 ` Arnd Bergmann
@ 2011-09-28 14:06 ` Mark Salter
2011-09-28 15:00 ` Arnd Bergmann
0 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-28 14:06 UTC (permalink / raw)
To: Arnd Bergmann; +Cc: linux-kernel, linux-arch
On Wed, 2011-09-28 at 15:26 +0200, Arnd Bergmann wrote:
> On Tuesday 27 September 2011, Mark Salter wrote:
>
> > +int c6x_num_cores;
> > +EXPORT_SYMBOL(c6x_num_cores);
> > +
> > +unsigned int c6x_silicon_rev;
> > +EXPORT_SYMBOL(c6x_silicon_rev);
> > +
> > +/*
> > + * Device status register. This holds information
> > + * about device configuration needed by some drivers.
> > + */
> > +unsigned int c6x_devstat;
> > +EXPORT_SYMBOL(c6x_devstat);
>
> > +unsigned int ticks_per_ns_scaled;
> > +EXPORT_SYMBOL(ticks_per_ns_scaled);
> > +
> > +unsigned int c6x_core_freq;
> > +EXPORT_SYMBOL(c6x_core_freq);
>
> These all look like they are used internally (maybe not
> ticks_per_ns_scaled) so it would be more appropriate to make
> them EXPORT_SYMBOL_GPL.
Hmm, we don't need to export c6x_core_freq at all.
c6x_silicon_rev is needed by drivers (though none in the current
patch series) to deal with silicon errata.
c6x_devstat has pin-strap config info some drivers may need.
ticks_per_ns_scale is used in udelay/mdelay.
--Mark
^ permalink raw reply [flat|nested] 66+ messages in thread
* Re: [PATCH v3 06/24] C6X: early boot code
2011-09-28 14:06 ` Mark Salter
@ 2011-09-28 15:00 ` Arnd Bergmann
0 siblings, 0 replies; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-28 15:00 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch
On Wednesday 28 September 2011, Mark Salter wrote:
> c6x_silicon_rev is needed by drivers (though none in the current
> patch series) to deal with silicon errata.
>
> c6x_devstat has pin-strap config info some drivers may need.
Ok, I guess these can be EXPORT_SYMBOL_GPL then, or you could
remove them entirely by evaluating them in the boot loader and
changing the device tree accordingly.
For device drivers, it's usually better to make decisions based
on the revision of the device they talk to, not the revision of
the entire SoC, since that requires knowledge of all SoCs that
contain each version of the hardware macro.
Arnd
^ permalink raw reply [flat|nested] 66+ messages in thread
* [PATCH v3 07/24] C6X: devicetree support
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (5 preceding siblings ...)
2011-09-27 20:29 ` [PATCH v3 06/24] C6X: early boot code Mark Salter
@ 2011-09-27 20:29 ` Mark Salter
2011-09-28 13:31 ` Arnd Bergmann
2011-09-27 20:29 ` [PATCH v3 08/24] C6X: memory management and DMA support Mark Salter
` (18 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:29 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter, devicetree-discuss
This is the basic devicetree support for C6X. Currently, four boards are
supported. Each one uses a different SoC part. Two of the four supported
SoCs are multicore. One with 3 cores and the other with 6 cores. There is
no coherency between the core-level caches, so SMP is not an option. It is
possible to run separate kernel instances on the various cores. There is
currently no C6X bootloader support for device trees so we build in the
DTB for now.
There are some interesting twists to the hardware which are of note for device
tree support. Each core has its own interrupt controller which is controlled
by special purpose core registers. This core controller provides 12 general
purpose prioritized interrupt sources. Each core is contained within a
hardware "module" which provides L1 and L2 caches, power control, and another
interrupt controller which cascades into the core interrupt controller. These
core module functions are controlled by memory mapped registers. The addresses
for these registers are the same for each core. That is, when coreN accesses
a module-level MMIO register at a given address, it accesses the register for
coreN even though other cores would use the same address to access the register
in the module containing those cores. Other hardware modules (timers, enet, etc)
which are memory mapped can be accessed by all cores.
The timers need some further explanation for multicore SoCs. Even though all
timer control registers are visible to all cores, interrupt routing or other
considerations may make a given timer more suitable for use by a core than
some other timer. Because of this and the desire to have the same image run
on more than one core, the timer nodes have a "ti,core-mask" property which
is used by the driver to scan for a suitable timer to use.
Signed-off-by: Mark Salter <msalter@redhat.com>
Cc: devicetree-discuss@lists.ozlabs.org
---
Documentation/devicetree/bindings/c6x/clocks.txt | 40 +++
Documentation/devicetree/bindings/c6x/dscr.txt | 127 ++++++++
Documentation/devicetree/bindings/c6x/emifa.txt | 62 ++++
.../devicetree/bindings/c6x/interrupt.txt | 104 ++++++
Documentation/devicetree/bindings/c6x/soc.txt | 28 ++
Documentation/devicetree/bindings/c6x/timer64.txt | 26 ++
arch/c6x/boot/dts/dsk6455.dts | 62 ++++
arch/c6x/boot/dts/evmc6457.dts | 48 +++
arch/c6x/boot/dts/evmc6472.dts | 73 +++++
arch/c6x/boot/dts/evmc6474.dts | 58 ++++
arch/c6x/boot/dts/tms320c6455.dtsi | 96 ++++++
arch/c6x/boot/dts/tms320c6457.dtsi | 68 ++++
arch/c6x/boot/dts/tms320c6472.dtsi | 134 ++++++++
arch/c6x/boot/dts/tms320c6474.dtsi | 89 ++++++
arch/c6x/boot/install-dtb.c | 333 ++++++++++++++++++++
arch/c6x/kernel/devicetree.c | 53 +++
arch/c6x/platforms/platform.c | 17 +
17 files changed, 1418 insertions(+), 0 deletions(-)
create mode 100644 Documentation/devicetree/bindings/c6x/clocks.txt
create mode 100644 Documentation/devicetree/bindings/c6x/dscr.txt
create mode 100644 Documentation/devicetree/bindings/c6x/emifa.txt
create mode 100644 Documentation/devicetree/bindings/c6x/interrupt.txt
create mode 100644 Documentation/devicetree/bindings/c6x/soc.txt
create mode 100644 Documentation/devicetree/bindings/c6x/timer64.txt
create mode 100644 arch/c6x/boot/dts/dsk6455.dts
create mode 100644 arch/c6x/boot/dts/evmc6457.dts
create mode 100644 arch/c6x/boot/dts/evmc6472.dts
create mode 100644 arch/c6x/boot/dts/evmc6474.dts
create mode 100644 arch/c6x/boot/dts/tms320c6455.dtsi
create mode 100644 arch/c6x/boot/dts/tms320c6457.dtsi
create mode 100644 arch/c6x/boot/dts/tms320c6472.dtsi
create mode 100644 arch/c6x/boot/dts/tms320c6474.dtsi
create mode 100644 arch/c6x/boot/install-dtb.c
create mode 100644 arch/c6x/kernel/devicetree.c
create mode 100644 arch/c6x/platforms/platform.c
diff --git a/Documentation/devicetree/bindings/c6x/clocks.txt b/Documentation/devicetree/bindings/c6x/clocks.txt
new file mode 100644
index 0000000..a04f5fd
--- /dev/null
+++ b/Documentation/devicetree/bindings/c6x/clocks.txt
@@ -0,0 +1,40 @@
+C6X PLL Clock Controllers
+-------------------------
+
+This is a first-cut support for the SoC clock controllers. This is still
+under development and will probably change as the common device tree
+clock support is added to the kernel.
+
+Required properties:
+
+- compatible: "ti,c64x+pll"
+ May also have SoC-specific value to support SoC-specific initialization
+ in the driver. One of:
+ "ti,c6455-pll"
+ "ti,c6457-pll"
+ "ti,c6472-pll"
+ "ti,c6474-pll"
+
+- reg: base address and size of register area
+- clock-frequency: input clock frequency in hz
+
+
+Optional properties:
+
+- ti,c64x+pll-bypass-delay: CPU cycles to delay when entering bypass mode
+
+- ti,c64x+pll-reset-delay: CPU cycles to delay after PLL reset
+
+- ti,c64x+pll-lock-delay: CPU cycles to delay after PLL frequency change
+
+Example:
+
+ clock-controller@29a0000 {
+ compatible = "ti,c6472-pll", "ti,c64x+pll";
+ reg = <0x029a0000 0x200>;
+ clock-frequency = <25000000>;
+
+ ti,c64x+pll-bypass-delay = <200>;
+ ti,c64x+pll-reset-delay = <12000>;
+ ti,c64x+pll-lock-delay = <80000>;
+ };
diff --git a/Documentation/devicetree/bindings/c6x/dscr.txt b/Documentation/devicetree/bindings/c6x/dscr.txt
new file mode 100644
index 0000000..d847758
--- /dev/null
+++ b/Documentation/devicetree/bindings/c6x/dscr.txt
@@ -0,0 +1,127 @@
+Device State Configuration Registers
+------------------------------------
+
+TI C6X SoCs contain a region of miscellaneous registers which provide various
+function for SoC control or status. Details vary considerably among from SoC
+to SoC with no two being alike.
+
+In general, the Device State Configuraion Registers (DSCR) will provide one or
+more configuration registers often protected by a lock register where one or
+more key values must be written to a lock register in order to unlock the
+configuration register for writes. These configuration register may be used to
+enable (and disable in some cases) SoC pin drivers, select peripheral clock
+sources (internal or pin), etc. In some cases, a configuration register is
+write once or the individual bits are write once. In addition to device config,
+the DSCR block may provide registers which which are used to reset peripherals,
+provide device ID information, provide ethernet MAC addresses, as well as other
+miscellaneous functions.
+
+For device state control (enable/disable), each device control is assigned an
+id which is used by individual device drivers to control the state as needed.
+
+Required properties:
+
+- compatible: must be "ti,c64x+dscr"
+- reg: register area base and size
+
+Optional properties:
+
+ NOTE: These are optional in that not all SoCs will have all properties. For
+ SoCs which do support a given property, leaving the property out of the
+ device tree will result in reduced functionality or possibly driver
+ failure.
+
+- ti,dscr-devstat
+ offset of the devstat register
+
+- ti,dscr-silicon-rev
+ offset, start bit, and bitsize of silicon revision field
+
+- ti,dscr-rmii-resets
+ offset and bitmask of RMII reset field. May have multiple tuples if more
+ than one ethernet port is available.
+
+- ti,dscr-locked-regs
+ possibly multiple tuples describing registers which are write protected by
+ a lock register. Each tuple consists of the register offset, lock register
+ offsset, and the key value used to unlock the register.
+
+- ti,dscr-kick-regs
+ offset and key values of two "kick" registers used to write protect other
+ registers in DSCR. On SoCs using kick registers, the first key must be
+ written to the first kick register and the second key must be written to
+ the second register before other registers in the area are write-enabled.
+
+- ti,dscr-mac-fuse-regs
+ MAC addresses are contained in two registers. Each element of a MAC address
+ is contained in a single byte. This property has two tuples. Each tuple has
+ a register offset and four cells representing bytes in the register from
+ most significant to least. The value of these four cells is the MAC byte
+ index (1-6) of the byte within the register. A value of 0 means the byte
+ is unused in the MAC address.
+
+- ti,dscr-devstate-ctl-regs
+ This property describes the bitfields used to control the state of devices.
+ Each tuple describes a range of identical bitfields used to control one or
+ more devices (one bitfield per device). The layout of each tuple is:
+
+ start_id num_ids reg enable disable start_bit nbits
+
+ Where:
+ start_id is device id for the first device control in the range
+ num_ids is the number of device controls in the range
+ reg is the offset of the register holding the control bits
+ enable is the value to enable a device
+ disable is the value to disable a device (0xffffffff if cannot disable)
+ start_bit is the bit number of the first bit in the range
+ nbits is the number of bits per device control
+
+- ti,dscr-devstate-stat-regs
+ This property describes the bitfields used to provide device state status
+ for device states controlled by the DSCR. Each tuple describes a range of
+ identical bitfields used to provide status for one or more devices (one
+ bitfield per device). The layout of each tuple is:
+
+ start_id num_ids reg enable disable start_bit nbits
+
+ Where:
+ start_id is device id for the first device status in the range
+ num_ids is the number of devices covered by the range
+ reg is the offset of the register holding the status bits
+ enable is the value indicating device is enabled
+ disable is the value indicating device is disabled
+ start_bit is the bit number of the first bit in the range
+ nbits is the number of bits per device status
+
+- ti,dscr-privperm
+ Offset and default value for register used to set access privilege for
+ some SoC devices.
+
+
+Example:
+
+ device-state-config-regs@2a80000 {
+ compatible = "ti,c64x+dscr";
+ reg = <0x02a80000 0x41000>;
+
+ ti,dscr-devstat = <0>;
+ ti,dscr-silicon-rev = <8 28 0xf>;
+ ti,dscr-rmii-resets = <0x40020 0x00040000>;
+
+ ti,dscr-locked-regs = <0x40008 0x40004 0x0f0a0b00>;
+ ti,dscr-devstate-ctl-regs =
+ <0 12 0x40008 1 0 0 2
+ 12 1 0x40008 3 0 30 2
+ 13 2 0x4002c 1 0xffffffff 0 1>;
+ ti,dscr-devstate-stat-regs =
+ <0 10 0x40014 1 0 0 3
+ 10 2 0x40018 1 0 0 3>;
+
+ ti,dscr-mac-fuse-regs = <0x700 1 2 3 4
+ 0x704 5 6 0 0>;
+
+ ti,dscr-privperm = <0x41c 0xaaaaaaaa>;
+
+ ti,dscr-kick-regs = <0x38 0x83E70B13
+ 0x3c 0x95A4F1E0>;
+ };
diff --git a/Documentation/devicetree/bindings/c6x/emifa.txt b/Documentation/devicetree/bindings/c6x/emifa.txt
new file mode 100644
index 0000000..0ff6e9b
--- /dev/null
+++ b/Documentation/devicetree/bindings/c6x/emifa.txt
@@ -0,0 +1,62 @@
+External Memory Interface
+-------------------------
+
+The emifa node describes a simple external bus controller found on some C6X
+SoCs. This interface provides external busses with a number of chip selects.
+
+Required properties:
+
+- compatible: must be "ti,c64x+emifa", "simple-bus"
+- reg: register area base and size
+- #address-cells: must be 2 (chip-select + offset)
+- #size-cells: must be 1
+- ranges: mapping from EMIFA space to parent space
+
+
+Optional properties:
+
+- ti,dscr-dev-enable: Device ID if EMIF is enabled/disabled from DSCR
+
+- ti,emifa-burst-priority:
+ Number of memory transfers after which the EMIF will elevate the priority
+ of the oldest command in the command FIFO. Setting this field to 255
+ disables this feature, thereby allowing old commands to stay in the FIFO
+ indefinitely.
+
+- ti,emifa-ce-config:
+ Configuration values for each of the supported chip selects.
+
+Example:
+
+ emifa@70000000 {
+ compatible = "ti,c64x+emifa", "simple-bus";
+ #address-cells = <2>;
+ #size-cells = <1>;
+ reg = <0x70000000 0x100>;
+ ranges = <0x2 0x0 0xa0000000 0x00000008
+ 0x3 0x0 0xb0000000 0x00400000
+ 0x4 0x0 0xc0000000 0x10000000
+ 0x5 0x0 0xD0000000 0x10000000>;
+
+ ti,dscr-dev-enable = <13>;
+ ti,emifa-burst-priority = <255>;
+ ti,emifa-ce-config = <0x00240120
+ 0x00240120
+ 0x00240122
+ 0x00240122>;
+
+ flash@3,0 {
+ #address-cells = <1>;
+ #size-cells = <1>;
+ compatible = "cfi-flash";
+ reg = <0x3 0x0 0x400000>;
+ bank-width = <1>;
+ device-width = <1>;
+ partition@0 {
+ reg = <0x0 0x400000>;
+ label = "NOR";
+ };
+ };
+ };
+
+This shows a flash chip attached to chip select 3.
diff --git a/Documentation/devicetree/bindings/c6x/interrupt.txt b/Documentation/devicetree/bindings/c6x/interrupt.txt
new file mode 100644
index 0000000..42bb796
--- /dev/null
+++ b/Documentation/devicetree/bindings/c6x/interrupt.txt
@@ -0,0 +1,104 @@
+C6X Interrupt Chips
+-------------------
+
+* C64X+ Core Interrupt Controller
+
+ The core interrupt controller provides 16 prioritized interrupts to the
+ C64X+ core. Priority 0 and 1 are used for reset and NMI respectively.
+ Priority 2 and 3 are reserved. Priority 4-15 are used for interrupt
+ sources coming from outside the core.
+
+ Required properties:
+ --------------------
+ - compatible: Should be "ti,c64x+core-pic";
+ - #interrupt-cells: <1>
+
+ Interrupt Specifier Definition
+ ------------------------------
+ Single cell specifying the core interrupt priority level (4-15) where
+ 4 is highest priority and 15 is lowest priority.
+
+ Example
+ -------
+ core_pic: interrupt-controller@0 {
+ interrupt-controller;
+ #interrupt-cells = <1>;
+ compatible = "ti,c64x+core-pic";
+ };
+
+
+
+* C64x+ Megamodule Interrupt Controller
+
+ The megamodule PIC consists of four interrupt mupliplexers each of which
+ combine up to 32 interrupt inputs into a single interrupt output which
+ may be cascaded into the core interrupt controller. The megamodule PIC
+ has a total of 12 outputs cascading into the core interrupt controller.
+ One for each core interrupt priority level. In addition to the combined
+ interrupt sources, individual megamodule interrupts may be cascaded to
+ the core interrupt controller. When an individual interrupt is cascaded,
+ it is no longer handled through a megamodule interrupt combiner and is
+ considered to have the core interrupt controller as the parent.
+
+ Required properties:
+ --------------------
+ - compatible: "ti,c64x+megamod-pic"
+ - interrupt-controller
+ - #interrupt-cells: <1>
+ - reg: base address and size of register area
+ - interrupt-parent: must be core interrupt controller
+ - interrupts: This should have four cells; one for each interrupt combiner.
+ The cells contain the core priority interrupt to which the
+ corresponding combiner output is wired.
+
+ Optional properties:
+ --------------------
+ - ti,c64x+megamod-pic-mux: Array of 12 cells correspnding to the 12 core
+ priority interrupts. The first cell corresponds to
+ core priority 4 and the last cell corresponds to
+ core priority 15. The value of each cell is the
+ megamodule interrupt source which is MUXed to
+ the core interrupt corresponding to the cell
+ position. Allowed values are 4 - 127. Mapping for
+ interrupts 0 - 3 (combined interrupt sources) are
+ ignored.
+
+ Interrupt Specifier Definition
+ ------------------------------
+ Single cell specifying the megamodule interrupt source (4-127). Note that
+ interrupts mapped directly to the core with "ti,c64x+megamod-pic-mux" will
+ use the core interrupt controller as their parent and the specifier will
+ be the core priority level, not the megamodule interrupt number.
+
+ Examples
+ --------
+ megamod_pic: interrupt-controller@1800000 {
+ compatible = "ti,c64x+megamod-pic";
+ interrupt-controller;
+ #interrupt-cells = <1>;
+ reg = <0x1800000 0x1000>;
+ interrupt-parent = <&core_pic>;
+ interrupts = < 12 13 14 15 >;
+ };
+
+ This is a minimal example where all individual interrupts go through a
+ combiner. Combiner-0 is mapped to core interrupt 12, combiner-1 is mapped
+ to interrupt 13, etc.
+
+
+ megamod_pic: interrupt-controller@1800000 {
+ compatible = "ti,c64x+megamod-pic";
+ interrupt-controller;
+ #interrupt-cells = <1>;
+ reg = <0x1800000 0x1000>;
+ interrupt-parent = <&core_pic>;
+ interrupts = < 12 13 14 15 >;
+ ti,c64x+megamod-pic-mux = < 0 0 0 0
+ 32 0 0 0
+ 0 0 0 0 >;
+ };
+
+ This the same as the first example except that megamodule interrupt 32 is
+ mapped directly to core priority interrupt 8. The node using this interrupt
+ must set the core controller as its interrupt parent and use 8 in the
+ interrupt specifier value.
diff --git a/Documentation/devicetree/bindings/c6x/soc.txt b/Documentation/devicetree/bindings/c6x/soc.txt
new file mode 100644
index 0000000..b1e4973
--- /dev/null
+++ b/Documentation/devicetree/bindings/c6x/soc.txt
@@ -0,0 +1,28 @@
+C6X System-on-Chip
+------------------
+
+Required properties:
+
+- compatible: "simple-bus"
+- #address-cells: must be 1
+- #size-cells: must be 1
+- ranges
+
+Optional properties:
+
+- model: specific SoC model
+
+- nodes for IP blocks within SoC
+
+
+Example:
+
+ soc {
+ compatible = "simple-bus";
+ model = "tms320c6455";
+ #address-cells = <1>;
+ #size-cells = <1>;
+ ranges;
+
+ ...
+ };
diff --git a/Documentation/devicetree/bindings/c6x/timer64.txt b/Documentation/devicetree/bindings/c6x/timer64.txt
new file mode 100644
index 0000000..95911fe
--- /dev/null
+++ b/Documentation/devicetree/bindings/c6x/timer64.txt
@@ -0,0 +1,26 @@
+Timer64
+-------
+
+The timer64 node describes C6X event timers.
+
+Required properties:
+
+- compatible: must be "ti,c64x+timer64"
+- reg: base address and size of register region
+- interrupt-parent: interrupt controller
+- interrupts: interrupt id
+
+Optional properties:
+
+- ti,dscr-dev-enable: Device ID used to enable timer IP through DSCR interface.
+
+- ti,core-mask: on multi-core SoCs, bitmask of cores allowed to use this timer.
+
+Example:
+ timer0: timer@25e0000 {
+ compatible = "ti,c64x+timer64";
+ ti,core-mask = < 0x01 >;
+ reg = <0x25e0000 0x40>;
+ interrupt-parent = <&megamod_pic>;
+ interrupts = < 16 >;
+ };
diff --git a/arch/c6x/boot/dts/dsk6455.dts b/arch/c6x/boot/dts/dsk6455.dts
new file mode 100644
index 0000000..2b71f80
--- /dev/null
+++ b/arch/c6x/boot/dts/dsk6455.dts
@@ -0,0 +1,62 @@
+/*
+ * arch/c6x/boot/dts/dsk6455.dts
+ *
+ * DSK6455 Evaluation Platform For TMS320C6455
+ * Copyright (C) 2011 Texas Instruments Incorporated
+ *
+ * Author: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ */
+
+/dts-v1/;
+
+/include/ "tms320c6455.dtsi"
+
+/ {
+ model = "Spectrum Digital DSK6455";
+ compatible = "spectrum-digital,dsk6455";
+
+ chosen {
+ bootargs = "root=/dev/nfs ip=dhcp rw";
+ };
+
+ memory {
+ device_type = "memory";
+ reg = <0xE0000000 0x08000000>;
+ };
+
+ soc {
+ megamod_pic: interrupt-controller@1800000 {
+ interrupts = < 12 13 14 15 >;
+ };
+
+ emifa@70000000 {
+ flash@3,0 {
+ #address-cells = <1>;
+ #size-cells = <1>;
+ compatible = "cfi-flash";
+ reg = <0x3 0x0 0x400000>;
+ bank-width = <1>;
+ device-width = <1>;
+ partition@0 {
+ reg = <0x0 0x400000>;
+ label = "NOR";
+ };
+ };
+ };
+
+ timer1: timer@2980000 {
+ interrupt-parent = <&megamod_pic>;
+ interrupts = < 69 >;
+ };
+
+ clock-controller@029a0000 {
+ clock-frequency = <50000000>;
+ };
+ };
+};
diff --git a/arch/c6x/boot/dts/evmc6457.dts b/arch/c6x/boot/dts/evmc6457.dts
new file mode 100644
index 0000000..0301eb9a
--- /dev/null
+++ b/arch/c6x/boot/dts/evmc6457.dts
@@ -0,0 +1,48 @@
+/*
+ * arch/c6x/boot/dts/evmc6457.dts
+ *
+ * EVMC6457 Evaluation Platform For TMS320C6457
+ *
+ * Copyright (C) 2011 Texas Instruments Incorporated
+ *
+ * Author: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ */
+
+/dts-v1/;
+
+/include/ "tms320c6457.dtsi"
+
+/ {
+ model = "eInfochips EVMC6457";
+ compatible = "einfochips,evmc6457";
+
+ chosen {
+ bootargs = "console=hvc root=/dev/nfs ip=dhcp rw";
+ };
+
+ memory {
+ device_type = "memory";
+ reg = <0xE0000000 0x10000000>;
+ };
+
+ soc {
+ megamod_pic: interrupt-controller@1800000 {
+ interrupts = < 12 13 14 15 >;
+ };
+
+ timer0: timer@2940000 {
+ interrupt-parent = <&megamod_pic>;
+ interrupts = < 67 >;
+ };
+
+ clock-controller@29a0000 {
+ clock-frequency = <60000000>;
+ };
+ };
+};
diff --git a/arch/c6x/boot/dts/evmc6472.dts b/arch/c6x/boot/dts/evmc6472.dts
new file mode 100644
index 0000000..3e207b4
--- /dev/null
+++ b/arch/c6x/boot/dts/evmc6472.dts
@@ -0,0 +1,73 @@
+/*
+ * arch/c6x/boot/dts/evmc6472.dts
+ *
+ * EVMC6472 Evaluation Platform For TMS320C6472
+ *
+ * Copyright (C) 2011 Texas Instruments Incorporated
+ *
+ * Author: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ */
+
+/dts-v1/;
+
+/include/ "tms320c6472.dtsi"
+
+/ {
+ model = "eInfochips EVMC6472";
+ compatible = "einfochips,evmc6472";
+
+ chosen {
+ bootargs = "console=hvc root=/dev/nfs ip=dhcp rw";
+ };
+
+ memory {
+ device_type = "memory";
+ reg = <0xE0000000 0x10000000>;
+ };
+
+ soc {
+ megamod_pic: interrupt-controller@1800000 {
+ interrupts = < 12 13 14 15 >;
+ };
+
+ timer0: timer@25e0000 {
+ interrupt-parent = <&megamod_pic>;
+ interrupts = < 16 >;
+ };
+
+ timer1: timer@25f0000 {
+ interrupt-parent = <&megamod_pic>;
+ interrupts = < 16 >;
+ };
+
+ timer2: timer@2600000 {
+ interrupt-parent = <&megamod_pic>;
+ interrupts = < 16 >;
+ };
+
+ timer3: timer@2610000 {
+ interrupt-parent = <&megamod_pic>;
+ interrupts = < 16 >;
+ };
+
+ timer4: timer@2620000 {
+ interrupt-parent = <&megamod_pic>;
+ interrupts = < 16 >;
+ };
+
+ timer5: timer@2630000 {
+ interrupt-parent = <&megamod_pic>;
+ interrupts = < 16 >;
+ };
+
+ clock-controller@29a0000 {
+ clock-frequency = <25000000>;
+ };
+ };
+};
diff --git a/arch/c6x/boot/dts/evmc6474.dts b/arch/c6x/boot/dts/evmc6474.dts
new file mode 100644
index 0000000..4dc2912
--- /dev/null
+++ b/arch/c6x/boot/dts/evmc6474.dts
@@ -0,0 +1,58 @@
+/*
+ * arch/c6x/boot/dts/evmc6474.dts
+ *
+ * EVMC6474 Evaluation Platform For TMS320C6474
+ *
+ * Copyright (C) 2011 Texas Instruments Incorporated
+ *
+ * Author: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ */
+
+/dts-v1/;
+
+/include/ "tms320c6474.dtsi"
+
+/ {
+ model = "Spectrum Digital EVMC6474";
+ compatible = "spectrum-digital,evmc6474";
+
+ chosen {
+ bootargs = "console=hvc root=/dev/nfs ip=dhcp rw";
+ };
+
+ memory {
+ device_type = "memory";
+ reg = <0x80000000 0x08000000>;
+ };
+
+ soc {
+ megamod_pic: interrupt-controller@1800000 {
+ interrupts = < 12 13 14 15 >;
+ };
+
+ timer3: timer@2940000 {
+ interrupt-parent = <&megamod_pic>;
+ interrupts = < 39 >;
+ };
+
+ timer4: timer@2950000 {
+ interrupt-parent = <&megamod_pic>;
+ interrupts = < 41 >;
+ };
+
+ timer5: timer@2960000 {
+ interrupt-parent = <&megamod_pic>;
+ interrupts = < 43 >;
+ };
+
+ clock-controller@29a0000 {
+ clock-frequency = <50000000>;
+ };
+ };
+};
diff --git a/arch/c6x/boot/dts/tms320c6455.dtsi b/arch/c6x/boot/dts/tms320c6455.dtsi
new file mode 100644
index 0000000..a804ec1
--- /dev/null
+++ b/arch/c6x/boot/dts/tms320c6455.dtsi
@@ -0,0 +1,96 @@
+
+/ {
+ #address-cells = <1>;
+ #size-cells = <1>;
+
+ cpus {
+ #address-cells = <1>;
+ #size-cells = <0>;
+
+ cpu@0 {
+ device_type = "cpu";
+ model = "ti,c64x+";
+ reg = <0>;
+ };
+ };
+
+ soc {
+ compatible = "simple-bus";
+ model = "tms320c6455";
+ #address-cells = <1>;
+ #size-cells = <1>;
+ ranges;
+
+ core_pic: interrupt-controller {
+ interrupt-controller;
+ #interrupt-cells = <1>;
+ compatible = "ti,c64x+core-pic";
+ };
+
+ /*
+ * Megamodule interrupt controller
+ */
+ megamod_pic: interrupt-controller@1800000 {
+ compatible = "ti,c64x+megamod-pic";
+ interrupt-controller;
+ #interrupt-cells = <1>;
+ reg = <0x1800000 0x1000>;
+ interrupt-parent = <&core_pic>;
+ };
+
+ cache-controller@1840000 {
+ compatible = "ti,c64x+cache";
+ reg = <0x01840000 0x8400>;
+ };
+
+ emifa@70000000 {
+ compatible = "ti,c64x+emifa", "simple-bus";
+ #address-cells = <2>;
+ #size-cells = <1>;
+ reg = <0x70000000 0x100>;
+ ranges = <0x2 0x0 0xa0000000 0x00000008
+ 0x3 0x0 0xb0000000 0x00400000
+ 0x4 0x0 0xc0000000 0x10000000
+ 0x5 0x0 0xD0000000 0x10000000>;
+
+ ti,dscr-dev-enable = <13>;
+ ti,emifa-burst-priority = <255>;
+ ti,emifa-ce-config = <0x00240120
+ 0x00240120
+ 0x00240122
+ 0x00240122>;
+ };
+
+ timer1: timer@2980000 {
+ compatible = "ti,c64x+timer64";
+ reg = <0x2980000 0x40>;
+ ti,dscr-dev-enable = <4>;
+ };
+
+ clock-controller@029a0000 {
+ compatible = "ti,c6455-pll", "ti,c64x+pll";
+ reg = <0x029a0000 0x200>;
+ ti,c64x+pll-bypass-delay = <1440>;
+ ti,c64x+pll-reset-delay = <15360>;
+ ti,c64x+pll-lock-delay = <24000>;
+ };
+
+ device-state-config-regs@2a80000 {
+ compatible = "ti,c64x+dscr";
+ reg = <0x02a80000 0x41000>;
+
+ ti,dscr-devstat = <0>;
+ ti,dscr-silicon-rev = <8 28 0xf>;
+ ti,dscr-rmii-resets = <0 0x40020 0x00040000>;
+
+ ti,dscr-locked-regs = <0x40008 0x40004 0x0f0a0b00>;
+ ti,dscr-devstate-ctl-regs =
+ <0 12 0x40008 1 0 0 2
+ 12 1 0x40008 3 0 30 2
+ 13 2 0x4002c 1 0xffffffff 0 1>;
+ ti,dscr-devstate-stat-regs =
+ <0 10 0x40014 1 0 0 3
+ 10 2 0x40018 1 0 0 3>;
+ };
+ };
+};
diff --git a/arch/c6x/boot/dts/tms320c6457.dtsi b/arch/c6x/boot/dts/tms320c6457.dtsi
new file mode 100644
index 0000000..35f4070
--- /dev/null
+++ b/arch/c6x/boot/dts/tms320c6457.dtsi
@@ -0,0 +1,68 @@
+
+/ {
+ #address-cells = <1>;
+ #size-cells = <1>;
+
+ cpus {
+ #address-cells = <1>;
+ #size-cells = <0>;
+
+ cpu@0 {
+ device_type = "cpu";
+ model = "ti,c64x+";
+ reg = <0>;
+ };
+ };
+
+ soc {
+ compatible = "simple-bus";
+ model = "tms320c6457";
+ #address-cells = <1>;
+ #size-cells = <1>;
+ ranges;
+
+ core_pic: interrupt-controller {
+ interrupt-controller;
+ #interrupt-cells = <1>;
+ compatible = "ti,c64x+core-pic";
+ };
+
+ megamod_pic: interrupt-controller@1800000 {
+ compatible = "ti,c64x+megamod-pic";
+ interrupt-controller;
+ #interrupt-cells = <1>;
+ interrupt-parent = <&core_pic>;
+ reg = <0x1800000 0x1000>;
+ };
+
+ cache-controller@1840000 {
+ compatible = "ti,c64x+cache";
+ reg = <0x01840000 0x8400>;
+ };
+
+ device-state-controller@2880800 {
+ compatible = "ti,c64x+dscr";
+ reg = <0x02880800 0x400>;
+
+ ti,dscr-devstat = <0x20>;
+ ti,dscr-silicon-rev = <0x18 28 0xf>;
+ ti,dscr-mac-fuse-regs = <0x114 3 4 5 6
+ 0x118 0 0 1 2>;
+ ti,dscr-kick-regs = <0x38 0x83E70B13
+ 0x3c 0x95A4F1E0>;
+ };
+
+ timer0: timer@2940000 {
+ compatible = "ti,c64x+timer64";
+ reg = <0x2940000 0x40>;
+ };
+
+ clock-controller@29a0000 {
+ compatible = "ti,c6457-pll", "ti,c64x+pll";
+ reg = <0x029a0000 0x200>;
+ ti,c64x+pll-bypass-delay = <300>;
+ ti,c64x+pll-reset-delay = <24000>;
+ ti,c64x+pll-lock-delay = <50000>;
+ };
+ };
+};
diff --git a/arch/c6x/boot/dts/tms320c6472.dtsi b/arch/c6x/boot/dts/tms320c6472.dtsi
new file mode 100644
index 0000000..b488aae
--- /dev/null
+++ b/arch/c6x/boot/dts/tms320c6472.dtsi
@@ -0,0 +1,134 @@
+
+/ {
+ #address-cells = <1>;
+ #size-cells = <1>;
+
+ cpus {
+ #address-cells = <1>;
+ #size-cells = <0>;
+
+ cpu@0 {
+ device_type = "cpu";
+ reg = <0>;
+ model = "ti,c64x+";
+ };
+ cpu@1 {
+ device_type = "cpu";
+ reg = <1>;
+ model = "ti,c64x+";
+ };
+ cpu@2 {
+ device_type = "cpu";
+ reg = <2>;
+ model = "ti,c64x+";
+ };
+ cpu@3 {
+ device_type = "cpu";
+ reg = <3>;
+ model = "ti,c64x+";
+ };
+ cpu@4 {
+ device_type = "cpu";
+ reg = <4>;
+ model = "ti,c64x+";
+ };
+ cpu@5 {
+ device_type = "cpu";
+ reg = <5>;
+ model = "ti,c64x+";
+ };
+ };
+
+ soc {
+ compatible = "simple-bus";
+ model = "tms320c6472";
+ #address-cells = <1>;
+ #size-cells = <1>;
+ ranges;
+
+ core_pic: interrupt-controller {
+ compatible = "ti,c64x+core-pic";
+ interrupt-controller;
+ #interrupt-cells = <1>;
+ };
+
+ megamod_pic: interrupt-controller@1800000 {
+ compatible = "ti,c64x+megamod-pic";
+ interrupt-controller;
+ #interrupt-cells = <1>;
+ reg = <0x1800000 0x1000>;
+ interrupt-parent = <&core_pic>;
+ };
+
+ cache-controller@1840000 {
+ compatible = "ti,c64x+cache";
+ reg = <0x01840000 0x8400>;
+ };
+
+ timer0: timer@25e0000 {
+ compatible = "ti,c64x+timer64";
+ ti,core-mask = < 0x01 >;
+ reg = <0x25e0000 0x40>;
+ };
+
+ timer1: timer@25f0000 {
+ compatible = "ti,c64x+timer64";
+ ti,core-mask = < 0x02 >;
+ reg = <0x25f0000 0x40>;
+ };
+
+ timer2: timer@2600000 {
+ compatible = "ti,c64x+timer64";
+ ti,core-mask = < 0x04 >;
+ reg = <0x2600000 0x40>;
+ };
+
+ timer3: timer@2610000 {
+ compatible = "ti,c64x+timer64";
+ ti,core-mask = < 0x08 >;
+ reg = <0x2610000 0x40>;
+ };
+
+ timer4: timer@2620000 {
+ compatible = "ti,c64x+timer64";
+ ti,core-mask = < 0x10 >;
+ reg = <0x2620000 0x40>;
+ };
+
+ timer5: timer@2630000 {
+ compatible = "ti,c64x+timer64";
+ ti,core-mask = < 0x20 >;
+ reg = <0x2630000 0x40>;
+ };
+
+ clock-controller@29a0000 {
+ compatible = "ti,c6472-pll", "ti,c64x+pll";
+ reg = <0x029a0000 0x200>;
+ ti,c64x+pll-bypass-delay = <200>;
+ ti,c64x+pll-reset-delay = <12000>;
+ ti,c64x+pll-lock-delay = <80000>;
+ };
+
+ device-state-controller@2a80000 {
+ compatible = "ti,c64x+dscr";
+ reg = <0x02a80000 0x1000>;
+
+ ti,dscr-devstat = <0>;
+ ti,dscr-silicon-rev = <0x70c 16 0xff>;
+
+ ti,dscr-mac-fuse-regs = <0x700 1 2 3 4
+ 0x704 5 6 0 0>;
+
+ ti,dscr-rmii-resets = <0x208 1
+ 0x20c 1>;
+
+ ti,dscr-locked-regs = <0x200 0x204 0x0a1e183a
+ 0x40c 0x420 0xbea7
+ 0x41c 0x420 0xbea7>;
+
+ ti,dscr-privperm = <0x41c 0xaaaaaaaa>;
+
+ ti,dscr-devstate-ctl-regs = <0 13 0x200 1 0 0 1>;
+ };
+ };
+};
diff --git a/arch/c6x/boot/dts/tms320c6474.dtsi b/arch/c6x/boot/dts/tms320c6474.dtsi
new file mode 100644
index 0000000..cc601bf
--- /dev/null
+++ b/arch/c6x/boot/dts/tms320c6474.dtsi
@@ -0,0 +1,89 @@
+
+/ {
+ #address-cells = <1>;
+ #size-cells = <1>;
+
+ cpus {
+ #address-cells = <1>;
+ #size-cells = <0>;
+
+ cpu@0 {
+ device_type = "cpu";
+ reg = <0>;
+ model = "ti,c64x+";
+ };
+ cpu@1 {
+ device_type = "cpu";
+ reg = <1>;
+ model = "ti,c64x+";
+ };
+ cpu@2 {
+ device_type = "cpu";
+ reg = <2>;
+ model = "ti,c64x+";
+ };
+ };
+
+ soc {
+ compatible = "simple-bus";
+ model = "tms320c6474";
+ #address-cells = <1>;
+ #size-cells = <1>;
+ ranges;
+
+ core_pic: interrupt-controller {
+ interrupt-controller;
+ #interrupt-cells = <1>;
+ compatible = "ti,c64x+core-pic";
+ };
+
+ megamod_pic: interrupt-controller@1800000 {
+ compatible = "ti,c64x+megamod-pic";
+ interrupt-controller;
+ #interrupt-cells = <1>;
+ reg = <0x1800000 0x1000>;
+ interrupt-parent = <&core_pic>;
+ };
+
+ cache-controller@1840000 {
+ compatible = "ti,c64x+cache";
+ reg = <0x01840000 0x8400>;
+ };
+
+ timer3: timer@2940000 {
+ compatible = "ti,c64x+timer64";
+ ti,core-mask = < 0x04 >;
+ reg = <0x2940000 0x40>;
+ };
+
+ timer4: timer@2950000 {
+ compatible = "ti,c64x+timer64";
+ ti,core-mask = < 0x02 >;
+ reg = <0x2950000 0x40>;
+ };
+
+ timer5: timer@2960000 {
+ compatible = "ti,c64x+timer64";
+ ti,core-mask = < 0x01 >;
+ reg = <0x2960000 0x40>;
+ };
+
+ device-state-controller@2880800 {
+ compatible = "ti,c64x+dscr";
+ reg = <0x02880800 0x400>;
+
+ ti,dscr-devstat = <0x004>;
+ ti,dscr-silicon-rev = <0x014 28 0xf>;
+ ti,dscr-mac-fuse-regs = <0x34 3 4 5 6
+ 0x38 0 0 1 2>;
+ };
+
+ clock-controller@29a0000 {
+ compatible = "ti,c6474-pll", "ti,c64x+pll";
+ reg = <0x029a0000 0x200>;
+ ti,c64x+pll-bypass-delay = <120>;
+ ti,c64x+pll-reset-delay = <30000>;
+ ti,c64x+pll-lock-delay = <60000>;
+ };
+ };
+};
diff --git a/arch/c6x/boot/install-dtb.c b/arch/c6x/boot/install-dtb.c
new file mode 100644
index 0000000..fa979ce
--- /dev/null
+++ b/arch/c6x/boot/install-dtb.c
@@ -0,0 +1,333 @@
+/*
+ * Program to hack in a DTB to an ELF file having a placeholder
+ * section named __fst_blob.
+ *
+ * This allows for building multiple images with builtin DTBs
+ * using a single vmlinux image. This is only necessary until
+ * bootloader support exists.
+ *
+ * Copyright 2011 Texas Instruments Incorporated
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version
+ * 2 of the License, or (at your option) any later version.
+ *
+ * Usage: install-dtb <image.dtb> <kernel.elf>
+ */
+
+#include <stdio.h>
+#include <unistd.h>
+#include <stdlib.h>
+#include <string.h>
+#include <ctype.h>
+#include <errno.h>
+#include <endian.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <sys/mman.h>
+#include <fcntl.h>
+#include <getopt.h>
+#include <elf.h>
+
+#ifndef EM_TI_C6000
+#define EM_TI_C6000 140 /* TI C6X DSPs */
+#endif
+
+static void usage(void)
+{
+ fprintf(stderr, "\n");
+ fprintf(stderr, "Usage: update-dtb <dtb-file> <kernel-elf-file> [ <outputfile> ] \n");
+ fprintf(stderr, "\n");
+ exit(1);
+}
+
+static u_int32_t swab32(u_int32_t n)
+{
+ return (n >> 24) | ((n >> 8) & 0xff00) |
+ ((n & 0xff00) << 8) | (n << 24);
+}
+
+static u_int16_t swab16(u_int16_t n)
+{
+ return (n >> 8) | (n << 8);
+}
+
+static int do_swab;
+
+#define ELFWORD(x) ({ do_swab ? swab32(x) : (x); })
+#define ELFHALF(x) ({ do_swab ? swab16(x) : (x); })
+
+static int elf_read(int fd, off_t offset, void *buf, ssize_t nbytes)
+{
+ ssize_t to_read, n;
+
+ if (lseek(fd, offset, SEEK_SET) < 0)
+ return errno;
+
+ to_read = nbytes;
+ while (to_read > 0) {
+ n = read(fd, buf, to_read);
+ if (n < 0)
+ return errno;
+ if (n == 0)
+ return -EINVAL;
+ buf += n;
+ to_read -= n;
+ }
+ return 0;
+}
+
+static int elf_write(int fd, off_t offset, void *buf, ssize_t nbytes)
+{
+ ssize_t to_write, n;
+
+ if (lseek(fd, offset, SEEK_SET) < 0)
+ return errno;
+
+ to_write = nbytes;
+ while (to_write > 0) {
+ n = write(fd, buf, to_write);
+ if (n < 0)
+ return errno;
+ if (n == 0)
+ return -EINVAL;
+ buf += n;
+ to_write -= n;
+ }
+ return 0;
+}
+
+Elf32_Ehdr ehdr;
+Elf32_Shdr *shdrs;
+char *shstrtab;
+Elf32_Sym *syms;
+char *strtab;
+
+static Elf32_Shdr *find_section_by_name(const char *name, int must_have)
+{
+ int i;
+ for (i = 1; i < ELFHALF(ehdr.e_shnum); i++) {
+ if (!strcmp(name, shstrtab + shdrs[i].sh_name))
+ return &shdrs[i];
+ }
+ if (must_have) {
+ fprintf(stderr, "Can't find %s section!\n", name);
+ exit(1);
+ }
+ return NULL;
+}
+
+/* read in a copy of a section */
+static void *copy_section(int fd, Elf32_Shdr *sh, const char *name)
+{
+ int err;
+ void *buf;
+
+ buf = malloc(sh->sh_size);
+ if (buf == NULL) {
+ fprintf(stderr, "Out of memory!\n");
+ exit(1);
+ }
+ err = elf_read(fd, sh->sh_offset, buf, sh->sh_size);
+ if (err) {
+ fprintf(stderr, "Can't read %s section: %s\n",
+ name, strerror(err));
+ exit(1);
+ }
+ return buf;
+}
+
+int main(int argc, char *argv[])
+{
+ int i, dtb_fd, out_fd = -1, kernel_fd, err;
+ off_t dtb_size, kern_size;
+ size_t n;
+ void *dtb_map, *kern_map, *buf;
+ Elf32_Shdr *sh;
+ char *outfilename = NULL;
+
+ if (argc == 4) {
+ outfilename = argv[3];
+ --argc;
+ }
+
+ if (argc != 3)
+ usage();
+
+ dtb_fd = open(argv[1], O_RDONLY);
+ if (dtb_fd < 0) {
+ fprintf(stderr, "Can't open %s: %s\n",
+ argv[1], strerror(errno));
+ exit(1);
+ }
+
+ if (outfilename) {
+ out_fd = creat(outfilename, 0775);
+ if (out_fd < 0) {
+ fprintf(stderr, "Can't create %s: %s\n",
+ outfilename, strerror(errno));
+ exit(1);
+ }
+ kernel_fd = open(argv[2], O_RDONLY);
+ } else
+ kernel_fd = open(argv[2], O_RDWR);
+ if (kernel_fd < 0) {
+ fprintf(stderr, "Can't open %s: %s\n",
+ argv[2], strerror(errno));
+ exit(1);
+ }
+
+ err = elf_read(kernel_fd, 0, &ehdr, sizeof(ehdr));
+ if (err) {
+ fprintf(stderr, "Can't read ELF header: %s\n", strerror(err));
+ exit(1);
+ }
+
+ if (strncmp((char *)ehdr.e_ident, "\177ELF", 4) != 0) {
+ fprintf(stderr, "Bad ELF ident!\n");
+ exit(1);
+ }
+
+#if __BYTE_ORDER == __LITTLE_ENDIAN
+ if (ehdr.e_ident[EI_DATA] == ELFDATA2MSB)
+ do_swab = 1;
+#else
+ if (ehdr.e_ident[EI_DATA] == ELFDATA2LSB)
+ do_swab = 1;
+#endif
+
+ if (ELFHALF(ehdr.e_machine) != EM_TI_C6000) {
+ fprintf(stderr, "Bad ELF machine!\n");
+ exit(1);
+ }
+
+ if (ELFHALF(ehdr.e_shentsize) != sizeof(Elf32_Shdr)) {
+ fprintf(stderr, "Bad ELF header (e_shentsize = %d)!\n",
+ ELFHALF(ehdr.e_shentsize));
+ exit(1);
+ }
+
+ /* kernel should have _way_ fewer sections */
+ if (ELFHALF(ehdr.e_shnum) > 100) {
+ fprintf(stderr, "Bad ELF header (e_shnum = %d)!\n",
+ ELFHALF(ehdr.e_shnum));
+ exit(1);
+ }
+
+ /* get section headers */
+ n = ELFHALF(ehdr.e_shnum) * ELFHALF(ehdr.e_shentsize);
+ shdrs = malloc(n);
+ if (shdrs == NULL) {
+ fprintf(stderr, "Out of memory!\n");
+ exit(1);
+ }
+ err = elf_read(kernel_fd, ELFWORD(ehdr.e_shoff), shdrs, n);
+ if (err) {
+ fprintf(stderr, "Can't read section headers: %s\n",
+ strerror(err));
+ exit(1);
+ }
+ if (do_swab)
+ for (i = 1; i < ELFHALF(ehdr.e_shnum); i++) {
+ sh = &shdrs[i];
+ sh->sh_name = swab32(sh->sh_name);
+ sh->sh_type = swab32(sh->sh_type);
+ sh->sh_flags = swab32(sh->sh_flags);
+ sh->sh_addr = swab32(sh->sh_addr);
+ sh->sh_offset = swab32(sh->sh_offset);
+ sh->sh_size = swab32(sh->sh_size);
+ sh->sh_link = swab32(sh->sh_link);
+ sh->sh_info = swab32(sh->sh_info);
+ sh->sh_addralign = swab32(sh->sh_addralign);
+ sh->sh_entsize = swab32(sh->sh_entsize);
+ }
+ /* get shdr strings so we can search for section headers by name */
+ shstrtab = copy_section(kernel_fd, &shdrs[ELFHALF(ehdr.e_shstrndx)],
+ ".shstrtab");
+
+ /* read in symtab and associated strings */
+ sh = find_section_by_name("__fdt_blob", 1);
+
+ dtb_size = lseek(dtb_fd, 0, SEEK_END);
+ if (dtb_size == -1 || lseek(dtb_fd, 0, SEEK_SET) == -1) {
+ fprintf(stderr, "Unable to seek %s: %s\n",
+ argv[1], strerror(err));
+ exit(1);
+ }
+
+ if (dtb_size > sh->sh_size) {
+ fprintf(stderr, "dtb too large (%d) to fit in section (%d).\n",
+ (int)dtb_size, sh->sh_size);
+ exit(1);
+ }
+
+ dtb_map = mmap(NULL, dtb_size, PROT_READ, MAP_PRIVATE, dtb_fd, 0);
+ if (dtb_map == MAP_FAILED) {
+ fprintf(stderr, "Failed to map %s: %s\n",
+ argv[1], strerror(err));
+ exit(1);
+ }
+
+ buf = malloc(sh->sh_size);
+ if (buf == NULL) {
+ fprintf(stderr, "Failed to malloc %d bytes!\n", sh->sh_size);
+ exit(1);
+ }
+ memset(buf, 0, sh->sh_size);
+ memcpy(buf, dtb_map, dtb_size);
+
+ if (!outfilename) {
+ /* write the dtb */
+ err = elf_write(kernel_fd, sh->sh_offset, buf, sh->sh_size);
+ if (err) {
+ fprintf(stderr, "Unable to write DTB: %s\n",
+ strerror(err));
+ exit(1);
+ }
+ } else {
+ kern_size = lseek(kernel_fd, 0, SEEK_END);
+ if (kern_size == -1 || lseek(dtb_fd, 0, SEEK_SET) == -1) {
+ fprintf(stderr, "Unable to seek %s: %s\n",
+ argv[2], strerror(err));
+ exit(1);
+ }
+
+ kern_map = mmap(NULL, kern_size, PROT_READ, MAP_PRIVATE, kernel_fd, 0);
+ if (kern_map == MAP_FAILED) {
+ fprintf(stderr, "Failed to map %s: %s\n",
+ argv[2], strerror(err));
+ exit(1);
+ }
+
+ /* write output file in three parts */
+
+ err = elf_write(out_fd, 0, kern_map, sh->sh_offset);
+ if (err) {
+ fprintf(stderr, "Error writing to %s: %s\n",
+ outfilename, strerror(err));
+ exit(1);
+ }
+
+ err = elf_write(out_fd, sh->sh_offset, buf, sh->sh_size);
+ if (err) {
+ fprintf(stderr, "Error writing to %s: %s\n",
+ outfilename, strerror(err));
+ exit(1);
+ }
+
+ err = elf_write(out_fd, sh->sh_offset + sh->sh_size,
+ kern_map + sh->sh_offset + sh->sh_size,
+ kern_size - (sh->sh_offset + sh->sh_size));
+ if (err) {
+ fprintf(stderr, "Error writing to %s: %s\n",
+ outfilename, strerror(err));
+ exit(1);
+ }
+
+ close(out_fd);
+ }
+ close(kernel_fd);
+ close(dtb_fd);
+ return 0;
+}
diff --git a/arch/c6x/kernel/devicetree.c b/arch/c6x/kernel/devicetree.c
new file mode 100644
index 0000000..bdb56f0
--- /dev/null
+++ b/arch/c6x/kernel/devicetree.c
@@ -0,0 +1,53 @@
+/*
+ * Architecture specific OF callbacks.
+ *
+ * Copyright (C) 2011 Texas Instruments Incorporated
+ * Author: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ */
+#include <linux/init.h>
+#include <linux/of.h>
+#include <linux/of_fdt.h>
+#include <linux/initrd.h>
+#include <linux/memblock.h>
+
+void __init early_init_devtree(void *params)
+{
+ /* Setup flat device-tree pointer */
+ initial_boot_params = params;
+
+ /* Retrieve various informations from the /chosen node of the
+ * device-tree, including the platform type, initrd location and
+ * size and more ...
+ */
+ of_scan_flat_dt(early_init_dt_scan_chosen, c6x_command_line);
+
+ /* Scan memory nodes and rebuild MEMBLOCKs */
+ of_scan_flat_dt(early_init_dt_scan_root, NULL);
+ of_scan_flat_dt(early_init_dt_scan_memory, NULL);
+}
+
+
+#ifdef CONFIG_BLK_DEV_INITRD
+void __init early_init_dt_setup_initrd_arch(unsigned long start,
+ unsigned long end)
+{
+ initrd_start = (unsigned long)__va(start);
+ initrd_end = (unsigned long)__va(end);
+ initrd_below_start_ok = 1;
+}
+#endif
+
+void __init early_init_dt_add_memory_arch(u64 base, u64 size)
+{
+ c6x_add_memory(base, size);
+}
+
+void * __init early_init_dt_alloc_memory_arch(u64 size, u64 align)
+{
+ return __va(memblock_alloc(size, align));
+}
diff --git a/arch/c6x/platforms/platform.c b/arch/c6x/platforms/platform.c
new file mode 100644
index 0000000..ad4651c
--- /dev/null
+++ b/arch/c6x/platforms/platform.c
@@ -0,0 +1,17 @@
+/*
+ * Copyright 2011 Texas Instruments Incorporated
+ *
+ * This file is licensed under the terms of the GNU General Public License
+ * version 2. This program is licensed "as is" without any warranty of any
+ * kind, whether express or implied.
+ */
+
+#include <linux/init.h>
+#include <linux/of_platform.h>
+
+static int __init c6x_device_probe(void)
+{
+ of_platform_bus_probe(NULL, of_default_bus_match_table, NULL);
+ return 0;
+}
+device_initcall(c6x_device_probe);
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v3 07/24] C6X: devicetree support
2011-09-27 20:29 ` [PATCH v3 07/24] C6X: devicetree support Mark Salter
@ 2011-09-28 13:31 ` Arnd Bergmann
2011-09-28 14:44 ` Mark Salter
2011-09-28 23:11 ` Grant Likely
0 siblings, 2 replies; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-28 13:31 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch, devicetree-discuss, Grant Likely
On Tuesday 27 September 2011, Mark Salter wrote:
> diff --git a/arch/c6x/boot/install-dtb.c b/arch/c6x/boot/install-dtb.c
> new file mode 100644
> index 0000000..fa979ce
> --- /dev/null
> +++ b/arch/c6x/boot/install-dtb.c
> @@ -0,0 +1,333 @@
> +/*
> + * Program to hack in a DTB to an ELF file having a placeholder
> + * section named __fst_blob.
> + *
> + * This allows for building multiple images with builtin DTBs
> + * using a single vmlinux image. This is only necessary until
> + * bootloader support exists.
> + *
> + * Copyright 2011 Texas Instruments Incorporated
> + *
> + * This program is free software; you can redistribute it and/or
> + * modify it under the terms of the GNU General Public License
> + * as published by the Free Software Foundation; either version
> + * 2 of the License, or (at your option) any later version.
> + *
> + * Usage: install-dtb <image.dtb> <kernel.elf>
> + */
Other architectures allow simply appending the device tree to the
kernel image file, which is more convenient for users and does
not require much kernel support. Have you tried this?
> diff --git a/arch/c6x/platforms/platform.c b/arch/c6x/platforms/platform.c
> new file mode 100644
> index 0000000..ad4651c
> --- /dev/null
> +++ b/arch/c6x/platforms/platform.c
> @@ -0,0 +1,17 @@
> +
> +static int __init c6x_device_probe(void)
> +{
> + of_platform_bus_probe(NULL, of_default_bus_match_table, NULL);
> + return 0;
> +}
> +device_initcall(c6x_device_probe);
device_initcall is run rather late. I think it would be more useful
to run this as early as possible, in order to make it possible to
have more of the core services probed through the device tree.
I also think that for probing the device tree root, you should be
using of_platform_populate instead of of_platform_bus_probe
nowadays. Maybe Grant can provide more information on this.
Arnd
^ permalink raw reply [flat|nested] 66+ messages in thread* Re: [PATCH v3 07/24] C6X: devicetree support
2011-09-28 13:31 ` Arnd Bergmann
@ 2011-09-28 14:44 ` Mark Salter
2011-09-28 14:57 ` Arnd Bergmann
2011-09-28 23:11 ` Grant Likely
1 sibling, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-28 14:44 UTC (permalink / raw)
To: Arnd Bergmann; +Cc: linux-kernel, linux-arch, devicetree-discuss, Grant Likely
On Wed, 2011-09-28 at 15:31 +0200, Arnd Bergmann wrote:
> On Tuesday 27 September 2011, Mark Salter wrote:
> > diff --git a/arch/c6x/boot/install-dtb.c b/arch/c6x/boot/install-dtb.c
> > new file mode 100644
> > index 0000000..fa979ce
> > --- /dev/null
> > +++ b/arch/c6x/boot/install-dtb.c
> > @@ -0,0 +1,333 @@
> > +/*
> > + * Program to hack in a DTB to an ELF file having a placeholder
> > + * section named __fst_blob.
> > + *
> > + * This allows for building multiple images with builtin DTBs
> > + * using a single vmlinux image. This is only necessary until
> > + * bootloader support exists.
> > + *
> > + * Copyright 2011 Texas Instruments Incorporated
> > + *
> > + * This program is free software; you can redistribute it and/or
> > + * modify it under the terms of the GNU General Public License
> > + * as published by the Free Software Foundation; either version
> > + * 2 of the License, or (at your option) any later version.
> > + *
> > + * Usage: install-dtb <image.dtb> <kernel.elf>
> > + */
>
> Other architectures allow simply appending the device tree to the
> kernel image file, which is more convenient for users and does
> not require much kernel support. Have you tried this?
Well, I considered it. My main reason for doing it this way was one of
debugging convenience. The debugger loads an ELF file and while I was
making lots of device tree changes in the tree and in the code, it was
more convenient to have one ELF file to deal with rather than an ELF
file and a dtb. And the thought was to get rid of it altogether once
bootloader support comes along. Maybe the thing to do would be to just
rip it out of the kernel. It can almost as easily be post-processing
step outside the kernel.
--Mark
^ permalink raw reply [flat|nested] 66+ messages in thread
* Re: [PATCH v3 07/24] C6X: devicetree support
2011-09-28 14:44 ` Mark Salter
@ 2011-09-28 14:57 ` Arnd Bergmann
0 siblings, 0 replies; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-28 14:57 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch, devicetree-discuss, Grant Likely
On Wednesday 28 September 2011, Mark Salter wrote:
> > Other architectures allow simply appending the device tree to the
> > kernel image file, which is more convenient for users and does
> > not require much kernel support. Have you tried this?
>
> Well, I considered it. My main reason for doing it this way was one of
> debugging convenience. The debugger loads an ELF file and while I was
> making lots of device tree changes in the tree and in the code, it was
> more convenient to have one ELF file to deal with rather than an ELF
> file and a dtb. And the thought was to get rid of it altogether once
> bootloader support comes along. Maybe the thing to do would be to just
> rip it out of the kernel. It can almost as easily be post-processing
> step outside the kernel.
Ok, if you plan to properly support it in the boot loader anyway,
that's probably the best option.
Arnd
^ permalink raw reply [flat|nested] 66+ messages in thread
* Re: [PATCH v3 07/24] C6X: devicetree support
2011-09-28 13:31 ` Arnd Bergmann
2011-09-28 14:44 ` Mark Salter
@ 2011-09-28 23:11 ` Grant Likely
1 sibling, 0 replies; 66+ messages in thread
From: Grant Likely @ 2011-09-28 23:11 UTC (permalink / raw)
To: Arnd Bergmann; +Cc: Mark Salter, linux-kernel, linux-arch, devicetree-discuss
On Wed, Sep 28, 2011 at 03:31:36PM +0200, Arnd Bergmann wrote:
> On Tuesday 27 September 2011, Mark Salter wrote:
> > diff --git a/arch/c6x/boot/install-dtb.c b/arch/c6x/boot/install-dtb.c
> > new file mode 100644
> > index 0000000..fa979ce
> > --- /dev/null
> > +++ b/arch/c6x/boot/install-dtb.c
> > @@ -0,0 +1,333 @@
> > +/*
> > + * Program to hack in a DTB to an ELF file having a placeholder
> > + * section named __fst_blob.
> > + *
> > + * This allows for building multiple images with builtin DTBs
> > + * using a single vmlinux image. This is only necessary until
> > + * bootloader support exists.
> > + *
> > + * Copyright 2011 Texas Instruments Incorporated
> > + *
> > + * This program is free software; you can redistribute it and/or
> > + * modify it under the terms of the GNU General Public License
> > + * as published by the Free Software Foundation; either version
> > + * 2 of the License, or (at your option) any later version.
> > + *
> > + * Usage: install-dtb <image.dtb> <kernel.elf>
> > + */
>
> Other architectures allow simply appending the device tree to the
> kernel image file, which is more convenient for users and does
> not require much kernel support. Have you tried this?
>
> > diff --git a/arch/c6x/platforms/platform.c b/arch/c6x/platforms/platform.c
> > new file mode 100644
> > index 0000000..ad4651c
> > --- /dev/null
> > +++ b/arch/c6x/platforms/platform.c
> > @@ -0,0 +1,17 @@
> > +
> > +static int __init c6x_device_probe(void)
> > +{
> > + of_platform_bus_probe(NULL, of_default_bus_match_table, NULL);
> > + return 0;
> > +}
> > +device_initcall(c6x_device_probe);
>
> device_initcall is run rather late. I think it would be more useful
> to run this as early as possible, in order to make it possible to
> have more of the core services probed through the device tree.
>
> I also think that for probing the device tree root, you should be
> using of_platform_populate instead of of_platform_bus_probe
> nowadays. Maybe Grant can provide more information on this.
Yes. Use of_platform_populate(). It's better.
g.
^ permalink raw reply [flat|nested] 66+ messages in thread
* [PATCH v3 08/24] C6X: memory management and DMA support
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (6 preceding siblings ...)
2011-09-27 20:29 ` [PATCH v3 07/24] C6X: devicetree support Mark Salter
@ 2011-09-27 20:29 ` Mark Salter
2011-09-28 13:46 ` Arnd Bergmann
2011-09-27 20:29 ` [PATCH v3 09/24] C6X: process management Mark Salter
` (17 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:29 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
The C6X architecture currently lacks an MMU so memory management is relatively
simple. There is no bus snooping between L2 and main memory but coherent DMA
memory is supported by making regions of main memory uncached. If such a region
is desired, it can be specified on the commandline with a "memdma=" argument.
Signed-off-by: Mark Salter <msalter@redhat.com>
---
arch/c6x/include/asm/dma-mapping.h | 91 +++++++++++++++++++
arch/c6x/kernel/dma.c | 168 ++++++++++++++++++++++++++++++++++++
arch/c6x/mm/dma-coherent.c | 142 ++++++++++++++++++++++++++++++
arch/c6x/mm/init.c | 112 ++++++++++++++++++++++++
4 files changed, 513 insertions(+), 0 deletions(-)
create mode 100644 arch/c6x/include/asm/dma-mapping.h
create mode 100644 arch/c6x/kernel/dma.c
create mode 100644 arch/c6x/mm/dma-coherent.c
create mode 100644 arch/c6x/mm/init.c
diff --git a/arch/c6x/include/asm/dma-mapping.h b/arch/c6x/include/asm/dma-mapping.h
new file mode 100644
index 0000000..03579fd
--- /dev/null
+++ b/arch/c6x/include/asm/dma-mapping.h
@@ -0,0 +1,91 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010, 2011 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot <aurelien.jacquiot@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ */
+#ifndef _ASM_C6X_DMA_MAPPING_H
+#define _ASM_C6X_DMA_MAPPING_H
+
+#include <linux/dma-debug.h>
+#include <asm-generic/dma-coherent.h>
+
+#define dma_supported(d, m) 1
+
+static inline int dma_set_mask(struct device *dev, u64 dma_mask)
+{
+ if (!dev->dma_mask || !dma_supported(dev, dma_mask))
+ return -EIO;
+
+ *dev->dma_mask = dma_mask;
+
+ return 0;
+}
+
+/*
+ * DMA errors are defined by all-bits-set in the DMA address.
+ */
+static inline int dma_mapping_error(struct device *dev, dma_addr_t dma_addr)
+{
+ return dma_addr == ~0;
+}
+
+extern dma_addr_t dma_map_single(struct device *dev, void *cpu_addr,
+ size_t size, enum dma_data_direction dir);
+
+extern void dma_unmap_single(struct device *dev, dma_addr_t handle,
+ size_t size, enum dma_data_direction dir);
+
+extern int dma_map_sg(struct device *dev, struct scatterlist *sglist,
+ int nents, enum dma_data_direction direction);
+
+extern void dma_unmap_sg(struct device *dev, struct scatterlist *sglist,
+ int nents, enum dma_data_direction direction);
+
+static inline dma_addr_t dma_map_page(struct device *dev, struct page *page,
+ unsigned long offset, size_t size,
+ enum dma_data_direction dir)
+{
+ dma_addr_t handle;
+
+ handle = dma_map_single(dev, page_address(page) + offset, size, dir);
+
+ debug_dma_map_page(dev, page, offset, size, dir, handle, false);
+
+ return handle;
+}
+
+static inline void dma_unmap_page(struct device *dev, dma_addr_t handle,
+ size_t size, enum dma_data_direction dir)
+{
+ dma_unmap_single(dev, handle, size, dir);
+
+ debug_dma_unmap_page(dev, handle, size, dir, false);
+}
+
+extern void dma_sync_single_for_cpu(struct device *dev, dma_addr_t handle,
+ size_t size, enum dma_data_direction dir);
+
+extern void dma_sync_single_for_device(struct device *dev, dma_addr_t handle,
+ size_t size,
+ enum dma_data_direction dir);
+
+extern void dma_sync_sg_for_cpu(struct device *dev, struct scatterlist *sg,
+ int nents, enum dma_data_direction dir);
+
+extern void dma_sync_sg_for_device(struct device *dev, struct scatterlist *sg,
+ int nents, enum dma_data_direction dir);
+
+extern void coherent_mem_init(u32 start, u32 size);
+extern void *dma_alloc_coherent(struct device *, size_t, dma_addr_t *, gfp_t);
+extern void dma_free_coherent(struct device *, size_t, void *, dma_addr_t);
+
+#define dma_alloc_noncoherent(d, s, h, f) dma_alloc_coherent((d), (s), (h), (f))
+#define dma_free_noncoherent(d, s, v, h) dma_free_coherent((d), (s), (v), (h))
+
+#endif /* _ASM_C6X_DMA_MAPPING_H */
diff --git a/arch/c6x/kernel/dma.c b/arch/c6x/kernel/dma.c
new file mode 100644
index 0000000..4120de0
--- /dev/null
+++ b/arch/c6x/kernel/dma.c
@@ -0,0 +1,168 @@
+/*
+ * Copyright (C) 2011 Texas Instruments Incorporated
+ * Author: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/dma-mapping.h>
+#include <linux/mm.h>
+#include <linux/mm_types.h>
+#include <linux/scatterlist.h>
+
+#include <asm/cacheflush.h>
+
+dma_addr_t dma_map_single(struct device *dev, void *ptr, size_t size,
+ enum dma_data_direction dir)
+{
+ dma_addr_t addr = virt_to_phys(ptr);
+
+ BUG_ON(!valid_dma_direction(dir));
+
+ switch (dir) {
+ case DMA_FROM_DEVICE:
+ L2_cache_block_invalidate(addr, addr + size);
+ break;
+ case DMA_TO_DEVICE:
+ L2_cache_block_writeback(addr, addr + size);
+ break;
+ case DMA_BIDIRECTIONAL:
+ /*
+ * If area will be used bidirectionally, syncing will
+ * take place manually in the driver.
+ */
+ default:
+ break;
+ }
+
+ debug_dma_map_page(dev, virt_to_page(ptr),
+ (unsigned long)ptr & ~PAGE_MASK, size,
+ dir, addr, true);
+ return addr;
+}
+EXPORT_SYMBOL(dma_map_single);
+
+
+void dma_unmap_single(struct device *dev, dma_addr_t handle,
+ size_t size, enum dma_data_direction dir)
+{
+ /* nothing to do here */
+ debug_dma_unmap_page(dev, handle, size, dir, true);
+}
+EXPORT_SYMBOL(dma_unmap_single);
+
+
+int dma_map_sg(struct device *dev, struct scatterlist *sglist,
+ int nents, enum dma_data_direction dir)
+{
+ struct scatterlist *sg;
+ int i;
+
+ for_each_sg(sglist, sg, nents, i)
+ sg->dma_address = dma_map_single(dev, sg_virt(sg), sg->length,
+ dir);
+
+ debug_dma_map_sg(dev, sglist, nents, nents, dir);
+
+ return nents;
+}
+EXPORT_SYMBOL(dma_map_sg);
+
+
+void dma_unmap_sg(struct device *dev, struct scatterlist *sglist,
+ int nents, enum dma_data_direction dir)
+{
+ struct scatterlist *sg;
+ int i;
+
+ for_each_sg(sglist, sg, nents, i)
+ dma_unmap_single(dev, sg_dma_address(sg), sg->length, dir);
+
+ debug_dma_unmap_sg(dev, sglist, nents, dir);
+}
+EXPORT_SYMBOL(dma_unmap_sg);
+
+
+void dma_sync_single_for_cpu(struct device *dev, dma_addr_t handle,
+ size_t size, enum dma_data_direction dir)
+{
+ unsigned long paddr = handle;
+
+ BUG_ON(!valid_dma_direction(dir));
+
+ switch (dir) {
+ case DMA_FROM_DEVICE:
+ L2_cache_block_invalidate(paddr, paddr + size);
+ break;
+ case DMA_TO_DEVICE:
+ L2_cache_block_writeback(paddr, paddr + size);
+ break;
+ case DMA_BIDIRECTIONAL:
+ L2_cache_block_writeback_invalidate(paddr, paddr + size);
+ break;
+ default:
+ break;
+ }
+
+ debug_dma_sync_single_for_cpu(dev, handle, size, dir);
+}
+EXPORT_SYMBOL(dma_sync_single_for_cpu);
+
+
+void dma_sync_single_for_device(struct device *dev, dma_addr_t handle,
+ size_t size, enum dma_data_direction dir)
+{
+ unsigned long paddr = handle;
+
+ BUG_ON(!valid_dma_direction(dir));
+
+ /* don't bother invalidating if DMA to device */
+ if (dir != DMA_TO_DEVICE)
+ L2_cache_block_invalidate(paddr, paddr + size);
+
+ debug_dma_sync_single_for_device(dev, handle, size, dir);
+}
+EXPORT_SYMBOL(dma_sync_single_for_device);
+
+
+void dma_sync_sg_for_cpu(struct device *dev, struct scatterlist *sglist,
+ int nents, enum dma_data_direction dir)
+{
+ struct scatterlist *sg;
+ int i;
+
+ for_each_sg(sglist, sg, nents, i)
+ dma_sync_single_for_cpu(dev, sg_dma_address(sg),
+ sg->length, dir);
+
+ debug_dma_sync_sg_for_cpu(dev, sglist, nents, dir);
+}
+EXPORT_SYMBOL(dma_sync_sg_for_cpu);
+
+
+void dma_sync_sg_for_device(struct device *dev, struct scatterlist *sglist,
+ int nents, enum dma_data_direction dir)
+{
+ struct scatterlist *sg;
+ int i;
+
+ for_each_sg(sglist, sg, nents, i)
+ dma_sync_single_for_device(dev, sg_dma_address(sg),
+ sg->length, dir);
+
+ debug_dma_sync_sg_for_device(dev, sglist, nents, dir);
+}
+EXPORT_SYMBOL(dma_sync_sg_for_device);
+
+
+/* Number of entries preallocated for DMA-API debugging */
+#define PREALLOC_DMA_DEBUG_ENTRIES (1 << 16)
+
+static int __init dma_init(void)
+{
+ dma_debug_init(PREALLOC_DMA_DEBUG_ENTRIES);
+
+ return 0;
+}
+fs_initcall(dma_init);
diff --git a/arch/c6x/mm/dma-coherent.c b/arch/c6x/mm/dma-coherent.c
new file mode 100644
index 0000000..98ef9e1
--- /dev/null
+++ b/arch/c6x/mm/dma-coherent.c
@@ -0,0 +1,142 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010, 2011 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot <aurelien.jacquiot@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * DMA uncached mapping support.
+ *
+ * Using code pulled from ARM
+ * Copyright (C) 2000-2004 Russell King
+ *
+ */
+#include <linux/slab.h>
+#include <linux/bitmap.h>
+#include <linux/bitops.h>
+#include <linux/interrupt.h>
+#include <linux/dma-mapping.h>
+#include <linux/memblock.h>
+
+#include <asm/page.h>
+
+/*
+ * DMA coherent memory management, can be redefined using the memdma=
+ * kernel command line
+ */
+
+/* none by default */
+static phys_addr_t dma_base;
+static u32 dma_size;
+static u32 dma_pages;
+
+static unsigned long *dma_bitmap;
+
+/* bitmap lock */
+static DEFINE_SPINLOCK(dma_lock);
+
+/*
+ * Return a DMA coherent and contiguous memory chunk from the DMA memory
+ */
+static inline u32 __alloc_dma_pages(int order)
+{
+ unsigned long flags;
+ u32 pos;
+
+ spin_lock_irqsave(&dma_lock, flags);
+ pos = bitmap_find_free_region(dma_bitmap, dma_pages, order);
+ spin_unlock_irqrestore(&dma_lock, flags);
+
+ return dma_base + (pos << PAGE_SHIFT);
+}
+
+static void __free_dma_pages(u32 addr, int order)
+{
+ unsigned long flags;
+ u32 pos = (addr - dma_base) >> PAGE_SHIFT;
+
+ if (addr < dma_base || (pos + (1 << order)) >= dma_pages) {
+ printk(KERN_ERR "%s: freeing outside range.\n", __func__);
+ BUG();
+ }
+
+ spin_lock_irqsave(&dma_lock, flags);
+ bitmap_release_region(dma_bitmap, pos, order);
+ spin_unlock_irqrestore(&dma_lock, flags);
+}
+
+/*
+ * Allocate DMA coherent memory space and return both the kernel
+ * virtual and DMA address for that space.
+ */
+void *dma_alloc_coherent(struct device *dev, size_t size,
+ dma_addr_t *handle, gfp_t gfp)
+{
+ u32 paddr;
+ int order;
+
+ if (!dma_size || !size)
+ return NULL;
+
+ order = get_count_order(((size - 1) >> PAGE_SHIFT) + 1);
+
+ paddr = __alloc_dma_pages(order);
+
+ if (handle)
+ *handle = paddr;
+
+ if (!paddr)
+ return NULL;
+
+ return phys_to_virt(paddr);
+}
+EXPORT_SYMBOL(dma_alloc_coherent);
+
+/*
+ * Free DMA coherent memory as defined by the above mapping.
+ */
+void dma_free_coherent(struct device *dev, size_t size, void *vaddr,
+ dma_addr_t dma_handle)
+{
+ int order;
+
+ if (!dma_size || !size)
+ return;
+
+ order = get_count_order(((size - 1) >> PAGE_SHIFT) + 1);
+
+ __free_dma_pages(virt_to_phys(vaddr), order);
+}
+EXPORT_SYMBOL(dma_free_coherent);
+
+/*
+ * Initialise the coherent DMA memory allocator using the given uncached region.
+ */
+void __init coherent_mem_init(phys_addr_t start, u32 size)
+{
+ phys_addr_t bitmap_phys;
+
+ if (!size)
+ return;
+
+ printk(KERN_INFO
+ "Coherent memory (DMA) region start=0x%x size=0x%x\n",
+ start, size);
+
+ dma_base = start;
+ dma_size = size;
+
+ /* allocate bitmap */
+ dma_pages = dma_size >> PAGE_SHIFT;
+ if (dma_size & (PAGE_SIZE - 1))
+ ++dma_pages;
+
+ bitmap_phys = memblock_alloc(BITS_TO_LONGS(dma_pages) * sizeof(long),
+ sizeof(long));
+
+ dma_bitmap = phys_to_virt(bitmap_phys);
+ memset(dma_bitmap, 0, dma_pages * PAGE_SIZE);
+}
diff --git a/arch/c6x/mm/init.c b/arch/c6x/mm/init.c
new file mode 100644
index 0000000..6b4d396
--- /dev/null
+++ b/arch/c6x/mm/init.c
@@ -0,0 +1,112 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010, 2011 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/mm.h>
+#include <linux/swap.h>
+#include <linux/bootmem.h>
+#ifdef CONFIG_BLK_DEV_RAM
+#include <linux/blkdev.h>
+#endif
+#include <linux/initrd.h>
+
+#include <asm/sections.h>
+
+/*
+ * ZERO_PAGE is a special page that is used for zero-initialized
+ * data and COW.
+ */
+unsigned long empty_zero_page;
+EXPORT_SYMBOL(empty_zero_page);
+
+/*
+ * paging_init() continues the virtual memory environment setup which
+ * was begun by the code in arch/head.S.
+ * The parameters are pointers to where to stick the starting and ending
+ * addresses of available kernel virtual memory.
+ */
+void __init paging_init(void)
+{
+ struct pglist_data *pgdat = NODE_DATA(0);
+ unsigned long zones_size[MAX_NR_ZONES] = {0, };
+
+ empty_zero_page = (unsigned long) alloc_bootmem_pages(PAGE_SIZE);
+ memset((void *)empty_zero_page, 0, PAGE_SIZE);
+
+ /*
+ * Set up user data space
+ */
+ set_fs(KERNEL_DS);
+
+ /*
+ * Define zones
+ */
+ zones_size[ZONE_NORMAL] = (memory_end - PAGE_OFFSET) >> PAGE_SHIFT;
+ pgdat->node_zones[ZONE_NORMAL].zone_start_pfn =
+ __pa(PAGE_OFFSET) >> PAGE_SHIFT;
+
+ free_area_init(zones_size);
+}
+
+void __init mem_init(void)
+{
+ int codek, datak;
+ unsigned long tmp;
+ unsigned long len = memory_end - memory_start;
+
+ high_memory = (void *)(memory_end & PAGE_MASK);
+
+ /* this will put all memory onto the freelists */
+ totalram_pages = free_all_bootmem();
+
+ codek = (_etext - _stext) >> 10;
+ datak = (_end - _sdata) >> 10;
+
+ tmp = nr_free_pages() << PAGE_SHIFT;
+ printk(KERN_INFO "Memory: %luk/%luk RAM (%dk kernel code, %dk data)\n",
+ tmp >> 10, len >> 10, codek, datak);
+}
+
+#ifdef CONFIG_BLK_DEV_INITRD
+void __init free_initrd_mem(unsigned long start, unsigned long end)
+{
+ int pages = 0;
+ for (; start < end; start += PAGE_SIZE) {
+ ClearPageReserved(virt_to_page(start));
+ init_page_count(virt_to_page(start));
+ free_page(start);
+ totalram_pages++;
+ pages++;
+ }
+ printk(KERN_INFO "Freeing initrd memory: %luk freed\n",
+ (pages * PAGE_SIZE) >> 10);
+}
+#endif
+
+void __init free_initmem(void)
+{
+ unsigned long addr;
+
+ /*
+ * The following code should be cool even if these sections
+ * are not page aligned.
+ */
+ addr = PAGE_ALIGN((unsigned long)(__init_begin));
+
+ /* next to check that the page we free is not a partial page */
+ for (; addr + PAGE_SIZE < (unsigned long)(__init_end);
+ addr += PAGE_SIZE) {
+ ClearPageReserved(virt_to_page(addr));
+ init_page_count(virt_to_page(addr));
+ free_page(addr);
+ totalram_pages++;
+ }
+ printk(KERN_INFO "Freeing unused kernel memory: %dK freed\n",
+ (int) ((addr - PAGE_ALIGN((long) &__init_begin)) >> 10));
+}
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v3 08/24] C6X: memory management and DMA support
2011-09-27 20:29 ` [PATCH v3 08/24] C6X: memory management and DMA support Mark Salter
@ 2011-09-28 13:46 ` Arnd Bergmann
0 siblings, 0 replies; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-28 13:46 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch
On Tuesday 27 September 2011, Mark Salter wrote:
> +dma_addr_t dma_map_single(struct device *dev, void *ptr, size_t size,
> + enum dma_data_direction dir)
> +{
> + dma_addr_t addr = virt_to_phys(ptr);
> +
> + BUG_ON(!valid_dma_direction(dir));
> +
> + switch (dir) {
> + case DMA_FROM_DEVICE:
> + L2_cache_block_invalidate(addr, addr + size);
> + break;
> + case DMA_TO_DEVICE:
> + L2_cache_block_writeback(addr, addr + size);
> + break;
> + case DMA_BIDIRECTIONAL:
> + /*
> + * If area will be used bidirectionally, syncing will
> + * take place manually in the driver.
> + */
I think this comment is not strictly correct and you have to do a
writeback+invalidate here. When you map an area, it initially
belongs to the device, until you either do an unmap or sync_*_for_cpu.
> +void dma_unmap_single(struct device *dev, dma_addr_t handle,
> + size_t size, enum dma_data_direction dir)
> +{
> + /* nothing to do here */
> + debug_dma_unmap_page(dev, handle, size, dir, true);
> +}
> +EXPORT_SYMBOL(dma_unmap_single);
> +void dma_sync_single_for_cpu(struct device *dev, dma_addr_t handle,
> + size_t size, enum dma_data_direction dir)
> +{
> + unsigned long paddr = handle;
> +
> + BUG_ON(!valid_dma_direction(dir));
> +
> + switch (dir) {
> + case DMA_FROM_DEVICE:
> + L2_cache_block_invalidate(paddr, paddr + size);
> + break;
> + case DMA_TO_DEVICE:
> + L2_cache_block_writeback(paddr, paddr + size);
> + break;
> + case DMA_BIDIRECTIONAL:
> + L2_cache_block_writeback_invalidate(paddr, paddr + size);
> + break;
> + default:
> + break;
> + }
> +
> + debug_dma_sync_single_for_cpu(dev, handle, size, dir);
> +}
> +EXPORT_SYMBOL(dma_sync_single_for_cpu);
Normally, sync_*_for_cpu is the same as unmap regarding the cache.
> +void dma_sync_single_for_device(struct device *dev, dma_addr_t handle,
> + size_t size, enum dma_data_direction dir)
> +{
> + unsigned long paddr = handle;
> +
> + BUG_ON(!valid_dma_direction(dir));
> +
> + /* don't bother invalidating if DMA to device */
> + if (dir != DMA_TO_DEVICE)
> + L2_cache_block_invalidate(paddr, paddr + size);
> +
> + debug_dma_sync_single_for_device(dev, handle, size, dir);
> +}
> +EXPORT_SYMBOL(dma_sync_single_for_device);
And this should do the same as map(). In particular it needs
to writeback the cache when direction is TO_DEVICE or BIDIRECTIONAL.
Arnd
^ permalink raw reply [flat|nested] 66+ messages in thread
* [PATCH v3 09/24] C6X: process management
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (7 preceding siblings ...)
2011-09-27 20:29 ` [PATCH v3 08/24] C6X: memory management and DMA support Mark Salter
@ 2011-09-27 20:29 ` Mark Salter
2011-09-28 13:47 ` Arnd Bergmann
2011-09-27 20:29 ` [PATCH v3 10/24] C6X: signal management Mark Salter
` (16 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:29 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
Signed-off-by: Mark Salter <msalter@redhat.com>
---
arch/c6x/include/asm/processor.h | 132 ++++++++++++++++++
arch/c6x/include/asm/thread_info.h | 121 +++++++++++++++++
arch/c6x/kernel/process.c | 263 ++++++++++++++++++++++++++++++++++++
arch/c6x/kernel/switch_to.S | 74 ++++++++++
4 files changed, 590 insertions(+), 0 deletions(-)
create mode 100644 arch/c6x/include/asm/processor.h
create mode 100644 arch/c6x/include/asm/thread_info.h
create mode 100644 arch/c6x/kernel/process.c
create mode 100644 arch/c6x/kernel/switch_to.S
diff --git a/arch/c6x/include/asm/processor.h b/arch/c6x/include/asm/processor.h
new file mode 100644
index 0000000..8154c4e
--- /dev/null
+++ b/arch/c6x/include/asm/processor.h
@@ -0,0 +1,132 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010, 2011 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * Updated for 2.6.34: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_PROCESSOR_H
+#define _ASM_C6X_PROCESSOR_H
+
+#include <asm/ptrace.h>
+#include <asm/page.h>
+#include <asm/current.h>
+
+/*
+ * Default implementation of macro that returns current
+ * instruction pointer ("program counter").
+ */
+#define current_text_addr() \
+({ \
+ void *__pc; \
+ asm("mvc .S2 pce1,%0\n" : "=b"(__pc)); \
+ __pc; \
+})
+
+/*
+ * User space process size. This is mostly meaningless for NOMMU
+ * but some C6X processors may have RAM addresses up to 0xFFFFFFFF.
+ * Since calls like mmap() can return an address or an error, we
+ * have to allow room for error returns when code does something
+ * like:
+ *
+ * addr = do_mmap(...)
+ * if ((unsigned long)addr >= TASK_SIZE)
+ * ... its an error code, not an address ...
+ *
+ * Here, we allow for 4096 error codes which means we really can't
+ * use the last 4K page on systems with RAM extending all the way
+ * to the end of the 32-bit address space.
+ */
+#define TASK_SIZE 0xFFFFF000
+
+/*
+ * This decides where the kernel will search for a free chunk of vm
+ * space during mmap's. We won't be using it
+ */
+#define TASK_UNMAPPED_BASE 0
+
+struct thread_struct {
+ unsigned long long b15_14;
+ unsigned long long a15_14;
+ unsigned long long b13_12;
+ unsigned long long a13_12;
+ unsigned long long b11_10;
+ unsigned long long a11_10;
+ unsigned long long ricl_icl;
+ unsigned long usp; /* user stack pointer */
+ unsigned long pc; /* kernel pc */
+ unsigned long wchan;
+};
+
+#define INIT_THREAD \
+{ \
+ .usp = 0, \
+ .wchan = 0, \
+}
+
+#define INIT_MMAP { \
+ &init_mm, 0, 0, NULL, PAGE_SHARED, VM_READ | VM_WRITE | VM_EXEC, 1, \
+ NULL, NULL }
+
+#define task_pt_regs(task) \
+ ((struct pt_regs *)(THREAD_START_SP + task_stack_page(task)) - 1)
+
+#define alloc_kernel_stack() __get_free_page(GFP_KERNEL)
+#define free_kernel_stack(page) free_page((page))
+
+
+/* Forward declaration, a strange C thing */
+struct task_struct;
+
+extern void start_thread(struct pt_regs *regs, unsigned int pc,
+ unsigned long usp);
+
+/* Free all resources held by a thread. */
+static inline void release_thread(struct task_struct *dead_task)
+{
+}
+
+/* Prepare to copy thread state - unlazy all lazy status */
+#define prepare_to_copy(tsk) do { } while (0)
+
+extern int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags);
+
+#define copy_segments(tsk, mm) do { } while (0)
+#define release_segments(mm) do { } while (0)
+
+/*
+ * saved PC of a blocked thread.
+ */
+#define thread_saved_pc(tsk) (task_pt_regs(tsk)->pc)
+
+/*
+ * saved kernel SP and DP of a blocked thread.
+ */
+#ifdef _BIG_ENDIAN
+#define thread_saved_ksp(tsk) \
+ (*(unsigned long *)&(tsk)->thread.b15_14)
+#define thread_saved_dp(tsk) \
+ (*(((unsigned long *)&(tsk)->thread.b15_14) + 1))
+#else
+#define thread_saved_ksp(tsk) \
+ (*(((unsigned long *)&(tsk)->thread.b15_14) + 1))
+#define thread_saved_dp(tsk) \
+ (*(unsigned long *)&(tsk)->thread.b15_14)
+#endif
+
+extern unsigned long get_wchan(struct task_struct *p);
+
+#define KSTK_EIP(tsk) (task_pt_regs(task)->pc)
+#define KSTK_ESP(tsk) (task_pt_regs(task)->sp)
+
+#define cpu_relax() do { } while (0)
+
+extern const struct seq_operations cpuinfo_op;
+
+#endif /* ASM_C6X_PROCESSOR_H */
diff --git a/arch/c6x/include/asm/thread_info.h b/arch/c6x/include/asm/thread_info.h
new file mode 100644
index 0000000..fd99148
--- /dev/null
+++ b/arch/c6x/include/asm/thread_info.h
@@ -0,0 +1,121 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010, 2011 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * Updated for 2.6.3x: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_THREAD_INFO_H
+#define _ASM_C6X_THREAD_INFO_H
+
+#ifdef __KERNEL__
+
+#include <asm/page.h>
+
+#ifdef CONFIG_4KSTACKS
+#define THREAD_SIZE 4096
+#define THREAD_SHIFT 12
+#define THREAD_ORDER 0
+#else
+#define THREAD_SIZE 8192
+#define THREAD_SHIFT 13
+#define THREAD_ORDER 1
+#endif
+
+#define THREAD_START_SP (THREAD_SIZE - 8)
+
+#ifndef __ASSEMBLY__
+
+typedef struct {
+ unsigned long seg;
+} mm_segment_t;
+
+/*
+ * low level task data.
+ */
+struct thread_info {
+ struct task_struct *task; /* main task structure */
+ struct exec_domain *exec_domain; /* execution domain */
+ unsigned long flags; /* low level flags */
+ int cpu; /* cpu we're on */
+ int preempt_count; /* 0 = preemptable, <0 = BUG */
+ mm_segment_t addr_limit; /* thread address space */
+ struct restart_block restart_block;
+};
+
+/*
+ * macros/functions for gaining access to the thread information structure
+ *
+ * preempt_count needs to be 1 initially, until the scheduler is functional.
+ */
+#define INIT_THREAD_INFO(tsk) \
+{ \
+ .task = &tsk, \
+ .exec_domain = &default_exec_domain, \
+ .flags = 0, \
+ .cpu = 0, \
+ .preempt_count = INIT_PREEMPT_COUNT, \
+ .addr_limit = KERNEL_DS, \
+ .restart_block = { \
+ .fn = do_no_restart_syscall, \
+ }, \
+}
+
+#define init_thread_info (init_thread_union.thread_info)
+#define init_stack (init_thread_union.stack)
+
+/* get the thread information struct of current task */
+static inline __attribute__((const))
+struct thread_info *current_thread_info(void)
+{
+ struct thread_info *ti;
+ asm volatile (" clr .s2 B15,0,%1,%0\n"
+ : "=b" (ti)
+ : "Iu5" (THREAD_SHIFT - 1));
+ return ti;
+}
+
+#define __HAVE_ARCH_THREAD_INFO_ALLOCATOR
+
+/* thread information allocation */
+#ifdef CONFIG_DEBUG_STACK_USAGE
+#define THREAD_FLAGS (GFP_KERNEL | __GFP_NOTRACK | __GFP_ZERO)
+#else
+#define THREAD_FLAGS (GFP_KERNEL | __GFP_NOTRACK)
+#endif
+
+#define alloc_thread_info_node(tsk, node) \
+ ((struct thread_info *)__get_free_pages(THREAD_FLAGS, THREAD_ORDER))
+
+#define free_thread_info(ti) free_pages((unsigned long) (ti), THREAD_ORDER)
+#define get_thread_info(ti) get_task_struct((ti)->task)
+#define put_thread_info(ti) put_task_struct((ti)->task)
+#endif /* __ASSEMBLY__ */
+
+#define PREEMPT_ACTIVE 0x10000000
+
+/*
+ * thread information flag bit numbers
+ * - pending work-to-be-done flags are in LSW
+ * - other flags in MSW
+ */
+#define TIF_SYSCALL_TRACE 0 /* syscall trace active */
+#define TIF_NOTIFY_RESUME 1 /* resumption notification requested */
+#define TIF_SIGPENDING 2 /* signal pending */
+#define TIF_NEED_RESCHED 3 /* rescheduling necessary */
+#define TIF_RESTORE_SIGMASK 4 /* restore signal mask in do_signal() */
+
+#define TIF_POLLING_NRFLAG 16 /* true if polling TIF_NEED_RESCHED */
+#define TIF_MEMDIE 17 /* OOM killer killed process */
+
+#define TIF_WORK_MASK 0x00007FFE /* work on irq/exception return */
+#define TIF_ALLWORK_MASK 0x00007FFF /* work on any return to u-space */
+
+#endif /* __KERNEL__ */
+
+#endif /* _ASM_C6X_THREAD_INFO_H */
diff --git a/arch/c6x/kernel/process.c b/arch/c6x/kernel/process.c
new file mode 100644
index 0000000..aa65c87
--- /dev/null
+++ b/arch/c6x/kernel/process.c
@@ -0,0 +1,263 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2006, 2009, 2010, 2011 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ */
+#include <linux/module.h>
+#include <linux/unistd.h>
+#include <linux/ptrace.h>
+#include <linux/init_task.h>
+#include <linux/tick.h>
+#include <linux/mqueue.h>
+#include <linux/syscalls.h>
+#include <linux/reboot.h>
+
+#include <asm/syscalls.h>
+
+/* hooks for board specific support */
+void (*c6x_restart)(void);
+void (*c6x_halt)(void);
+
+extern asmlinkage void ret_from_fork(void);
+
+static struct signal_struct init_signals = INIT_SIGNALS(init_signals);
+static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand);
+
+/*
+ * Initial thread structure.
+ */
+union thread_union init_thread_union __init_task_data = {
+ INIT_THREAD_INFO(init_task)
+};
+
+/*
+ * Initial task structure.
+ */
+struct task_struct init_task = INIT_TASK(init_task);
+EXPORT_SYMBOL(init_task);
+
+/*
+ * power off function, if any
+ */
+void (*pm_power_off)(void);
+EXPORT_SYMBOL(pm_power_off);
+
+static void c6x_idle(void)
+{
+ unsigned long tmp;
+
+ /*
+ * Put local_irq_enable and idle in same execute packet
+ * to make them atomic and avoid race to idle with
+ * interrupts enabled.
+ */
+ asm volatile (" mvc .s2 CSR,%0\n"
+ " or .d2 1,%0,%0\n"
+ " mvc .s2 %0,CSR\n"
+ "|| idle\n"
+ : "=b"(tmp));
+}
+
+/*
+ * The idle loop for C64x
+ */
+void cpu_idle(void)
+{
+ /* endless idle loop with no priority at all */
+ while (1) {
+ tick_nohz_stop_sched_tick(1);
+ while (1) {
+ local_irq_disable();
+ if (need_resched()) {
+ local_irq_enable();
+ break;
+ }
+ c6x_idle(); /* enables local irqs */
+ }
+ tick_nohz_restart_sched_tick();
+
+ preempt_enable_no_resched();
+ schedule();
+ preempt_disable();
+ }
+}
+
+static void halt_loop(void)
+{
+ printk(KERN_EMERG "System Halted, OK to turn off power\n");
+ local_irq_disable();
+ while (1)
+ asm volatile("idle\n");
+}
+
+void machine_restart(char *__unused)
+{
+ if (c6x_restart)
+ c6x_restart();
+ halt_loop();
+}
+
+void machine_halt(void)
+{
+ if (c6x_halt)
+ c6x_halt();
+ halt_loop();
+}
+
+void machine_power_off(void)
+{
+ if (pm_power_off)
+ pm_power_off();
+ halt_loop();
+}
+
+static void kernel_thread_helper(int dummy, void *arg, int (*fn)(void *))
+{
+ do_exit(fn(arg));
+}
+
+/*
+ * Create a kernel thread
+ */
+int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags)
+{
+ struct pt_regs regs;
+
+ /*
+ * copy_thread sets a4 to zero (child return from fork)
+ * so we can't just set things up to directly return to
+ * fn.
+ */
+ memset(®s, 0, sizeof(regs));
+ regs.b4 = (unsigned long) arg;
+ regs.a6 = (unsigned long) fn;
+ regs.pc = (unsigned long) kernel_thread_helper;
+ local_save_flags(regs.csr);
+ regs.csr |= 1;
+ regs.tsr = 5; /* Set GEE and GIE in TSR */
+
+ /* Ok, create the new process.. */
+ return do_fork(flags | CLONE_VM | CLONE_UNTRACED, -1, ®s,
+ 0, NULL, NULL);
+}
+EXPORT_SYMBOL(kernel_thread);
+
+void flush_thread(void)
+{
+}
+
+void exit_thread(void)
+{
+}
+
+SYSCALL_DEFINE1(c6x_clone, struct pt_regs *, regs)
+{
+ unsigned long clone_flags;
+ unsigned long newsp;
+
+ /* syscall puts clone_flags in A4 and usp in B4 */
+ clone_flags = regs->orig_a4;
+ if (regs->b4)
+ newsp = regs->b4;
+ else
+ newsp = regs->sp;
+
+ return do_fork(clone_flags, newsp, regs, 0, (int __user *)regs->a6,
+ (int __user *)regs->b6);
+}
+
+/*
+ * Do necessary setup to start up a newly executed thread.
+ */
+void start_thread(struct pt_regs *regs, unsigned int pc, unsigned long usp)
+{
+ /*
+ * The binfmt loader will setup a "full" stack, but the C6X
+ * operates an "empty" stack. So we adjust the usp so that
+ * argc doesn't get destroyed if an interrupt is taken before
+ * it is read from the stack.
+ *
+ * NB: Library startup code needs to match this.
+ */
+ usp -= 8;
+
+ set_fs(USER_DS);
+ regs->pc = pc;
+ regs->sp = usp;
+ regs->tsr |= 0x40; /* set user mode */
+ current->thread.usp = usp;
+}
+
+/*
+ * Copy a new thread context in its stack.
+ */
+int copy_thread(unsigned long clone_flags, unsigned long usp,
+ unsigned long ustk_size,
+ struct task_struct *p, struct pt_regs *regs)
+{
+ struct pt_regs *childregs;
+
+ childregs = task_pt_regs(p);
+
+ *childregs = *regs;
+ childregs->a4 = 0;
+
+ if (usp == -1)
+ /* case of __kernel_thread: we return to supervisor space */
+ childregs->sp = (unsigned long)(childregs + 1);
+ else
+ /* Otherwise use the given stack */
+ childregs->sp = usp;
+
+ /* Set usp/ksp */
+ p->thread.usp = childregs->sp;
+ /* switch_to uses stack to save/restore 14 callee-saved regs */
+ thread_saved_ksp(p) = (unsigned long)childregs - 8;
+ p->thread.pc = (unsigned int) ret_from_fork;
+ p->thread.wchan = (unsigned long) ret_from_fork;
+#ifdef __DSBT__
+ {
+ unsigned long dp;
+
+ asm volatile ("mv .S2 b14,%0\n" : "=b"(dp));
+
+ thread_saved_dp(p) = dp;
+ if (usp == -1)
+ childregs->dp = dp;
+ }
+#endif
+ return 0;
+}
+
+/*
+ * c6x_execve() executes a new program.
+ */
+SYSCALL_DEFINE4(c6x_execve, const char __user *, name,
+ const char __user *const __user *, argv,
+ const char __user *const __user *, envp,
+ struct pt_regs *, regs)
+{
+ int error;
+ char *filename;
+
+ filename = getname(name);
+ error = PTR_ERR(filename);
+ if (IS_ERR(filename))
+ goto out;
+
+ error = do_execve(filename, argv, envp, regs);
+ putname(filename);
+out:
+ return error;
+}
+
+unsigned long get_wchan(struct task_struct *p)
+{
+ return p->thread.wchan;
+}
diff --git a/arch/c6x/kernel/switch_to.S b/arch/c6x/kernel/switch_to.S
new file mode 100644
index 0000000..09177ed
--- /dev/null
+++ b/arch/c6x/kernel/switch_to.S
@@ -0,0 +1,74 @@
+/*
+ * Copyright (C) 2011 Texas Instruments Incorporated
+ * Author: Mark Salter (msalter@redhat.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/linkage.h>
+#include <asm/asm-offsets.h>
+
+#define SP B15
+
+ /*
+ * void __switch_to(struct thread_info *prev,
+ * struct thread_info *next,
+ * struct task_struct *tsk) ;
+ */
+ENTRY(__switch_to)
+ LDDW .D2T2 *+B4(THREAD_B15_14),B7:B6
+ || MV .L2X A4,B5 ; prev
+ || MV .L1X B4,A5 ; next
+ || MVC .S2 RILC,B1
+
+ STW .D2T2 B3,*+B5(THREAD_PC)
+ || STDW .D1T1 A13:A12,*+A4(THREAD_A13_12)
+ || MVC .S2 ILC,B0
+
+ LDW .D2T2 *+B4(THREAD_PC),B3
+ || LDDW .D1T1 *+A5(THREAD_A13_12),A13:A12
+
+ STDW .D1T1 A11:A10,*+A4(THREAD_A11_10)
+ || STDW .D2T2 B1:B0,*+B5(THREAD_RICL_ICL)
+#ifndef __DSBT__
+ || MVKL .S2 current_ksp,B1
+#endif
+
+ STDW .D2T2 B15:B14,*+B5(THREAD_B15_14)
+ || STDW .D1T1 A15:A14,*+A4(THREAD_A15_14)
+#ifndef __DSBT__
+ || MVKH .S2 current_ksp,B1
+#endif
+
+ ;; Switch to next SP
+ MV .S2 B7,SP
+#ifdef __DSBT__
+ || STW .D2T2 B7,*+B14(current_ksp)
+#else
+ || STW .D2T2 B7,*B1
+ || MV .L2 B6,B14
+#endif
+ || LDDW .D1T1 *+A5(THREAD_RICL_ICL),A1:A0
+
+ STDW .D2T2 B11:B10,*+B5(THREAD_B11_10)
+ || LDDW .D1T1 *+A5(THREAD_A15_14),A15:A14
+
+ STDW .D2T2 B13:B12,*+B5(THREAD_B13_12)
+ || LDDW .D1T1 *+A5(THREAD_A11_10),A11:A10
+
+ B .S2 B3 ; return in next E1
+ || LDDW .D2T2 *+B4(THREAD_B13_12),B13:B12
+
+ LDDW .D2T2 *+B4(THREAD_B11_10),B11:B10
+ NOP
+
+ MV .L2X A0,B0
+ || MV .S1 A6,A4
+
+ MVC .S2 B0,ILC
+ || MV .L2X A1,B1
+
+ MVC .S2 B1,RILC
+ENDPROC(__switch_to)
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v3 09/24] C6X: process management
2011-09-27 20:29 ` [PATCH v3 09/24] C6X: process management Mark Salter
@ 2011-09-28 13:47 ` Arnd Bergmann
0 siblings, 0 replies; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-28 13:47 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch
On Tuesday 27 September 2011, Mark Salter wrote:
> Signed-off-by: Mark Salter <msalter@redhat.com>
> ---
> arch/c6x/include/asm/processor.h | 132 ++++++++++++++++++
> arch/c6x/include/asm/thread_info.h | 121 +++++++++++++++++
> arch/c6x/kernel/process.c | 263 ++++++++++++++++++++++++++++++++++++
> arch/c6x/kernel/switch_to.S | 74 ++++++++++
> 4 files changed, 590 insertions(+), 0 deletions(-)
> create mode 100644 arch/c6x/include/asm/processor.h
> create mode 100644 arch/c6x/include/asm/thread_info.h
> create mode 100644 arch/c6x/kernel/process.c
> create mode 100644 arch/c6x/kernel/switch_to.S
Acked-by: Arnd Bergmann <arnd@arndb.de>
^ permalink raw reply [flat|nested] 66+ messages in thread
* [PATCH v3 10/24] C6X: signal management
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (8 preceding siblings ...)
2011-09-27 20:29 ` [PATCH v3 09/24] C6X: process management Mark Salter
@ 2011-09-27 20:29 ` Mark Salter
2011-09-28 13:48 ` Arnd Bergmann
2011-09-27 20:29 ` [PATCH v3 11/24] C6X: time management Mark Salter
` (15 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:29 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
Signed-off-by: Mark Salter <msalter@redhat.com>
---
arch/c6x/include/asm/sigcontext.h | 80 ++++++++
arch/c6x/include/asm/signal.h | 17 ++
arch/c6x/kernel/signal.c | 377 +++++++++++++++++++++++++++++++++++++
3 files changed, 474 insertions(+), 0 deletions(-)
create mode 100644 arch/c6x/include/asm/sigcontext.h
create mode 100644 arch/c6x/include/asm/signal.h
create mode 100644 arch/c6x/kernel/signal.c
diff --git a/arch/c6x/include/asm/sigcontext.h b/arch/c6x/include/asm/sigcontext.h
new file mode 100644
index 0000000..eb702f3
--- /dev/null
+++ b/arch/c6x/include/asm/sigcontext.h
@@ -0,0 +1,80 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_SIGCONTEXT_H
+#define _ASM_C6X_SIGCONTEXT_H
+
+
+struct sigcontext {
+ unsigned long sc_mask; /* old sigmask */
+ unsigned long sc_sp; /* old user stack pointer */
+
+ unsigned long sc_a4;
+ unsigned long sc_b4;
+ unsigned long sc_a6;
+ unsigned long sc_b6;
+ unsigned long sc_a8;
+ unsigned long sc_b8;
+
+ unsigned long sc_a0;
+ unsigned long sc_a1;
+ unsigned long sc_a2;
+ unsigned long sc_a3;
+ unsigned long sc_a5;
+ unsigned long sc_a7;
+ unsigned long sc_a9;
+
+ unsigned long sc_b0;
+ unsigned long sc_b1;
+ unsigned long sc_b2;
+ unsigned long sc_b3;
+ unsigned long sc_b5;
+ unsigned long sc_b7;
+ unsigned long sc_b9;
+
+ unsigned long sc_a16;
+ unsigned long sc_a17;
+ unsigned long sc_a18;
+ unsigned long sc_a19;
+ unsigned long sc_a20;
+ unsigned long sc_a21;
+ unsigned long sc_a22;
+ unsigned long sc_a23;
+ unsigned long sc_a24;
+ unsigned long sc_a25;
+ unsigned long sc_a26;
+ unsigned long sc_a27;
+ unsigned long sc_a28;
+ unsigned long sc_a29;
+ unsigned long sc_a30;
+ unsigned long sc_a31;
+
+ unsigned long sc_b16;
+ unsigned long sc_b17;
+ unsigned long sc_b18;
+ unsigned long sc_b19;
+ unsigned long sc_b20;
+ unsigned long sc_b21;
+ unsigned long sc_b22;
+ unsigned long sc_b23;
+ unsigned long sc_b24;
+ unsigned long sc_b25;
+ unsigned long sc_b26;
+ unsigned long sc_b27;
+ unsigned long sc_b28;
+ unsigned long sc_b29;
+ unsigned long sc_b30;
+ unsigned long sc_b31;
+
+ unsigned long sc_csr;
+ unsigned long sc_pc;
+};
+
+#endif /* _ASM_C6X_SIGCONTEXT_H */
diff --git a/arch/c6x/include/asm/signal.h b/arch/c6x/include/asm/signal.h
new file mode 100644
index 0000000..f1cd870
--- /dev/null
+++ b/arch/c6x/include/asm/signal.h
@@ -0,0 +1,17 @@
+#ifndef _ASM_C6X_SIGNAL_H
+#define _ASM_C6X_SIGNAL_H
+
+#include <asm-generic/signal.h>
+
+#ifndef __ASSEMBLY__
+#include <linux/linkage.h>
+
+struct pt_regs;
+
+extern asmlinkage int do_rt_sigreturn(struct pt_regs *regs);
+extern asmlinkage void do_notify_resume(struct pt_regs *regs,
+ u32 thread_info_flags,
+ int syscall);
+#endif
+
+#endif /* _ASM_C6X_SIGNAL_H */
diff --git a/arch/c6x/kernel/signal.c b/arch/c6x/kernel/signal.c
new file mode 100644
index 0000000..304f675
--- /dev/null
+++ b/arch/c6x/kernel/signal.c
@@ -0,0 +1,377 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2006, 2009, 2010, 2011 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * Updated for 2.6.34: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/kernel.h>
+#include <linux/uaccess.h>
+#include <linux/syscalls.h>
+#include <linux/tracehook.h>
+
+#include <asm/ucontext.h>
+#include <asm/cacheflush.h>
+
+
+#define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP)))
+
+/*
+ * Do a signal return, undo the signal stack.
+ */
+
+#define RETCODE_SIZE (9 << 2) /* 9 instructions = 36 bytes */
+
+struct rt_sigframe {
+ struct siginfo __user *pinfo;
+ void __user *puc;
+ struct siginfo info;
+ struct ucontext uc;
+ unsigned long retcode[RETCODE_SIZE >> 2];
+};
+
+static int restore_sigcontext(struct pt_regs *regs,
+ struct sigcontext __user *sc)
+{
+ int err = 0;
+
+ /* The access_ok check was done by caller, so use __get_user here */
+#define COPY(x) (err |= __get_user(regs->x, &sc->sc_##x))
+
+ COPY(sp); COPY(a4); COPY(b4); COPY(a6); COPY(b6); COPY(a8); COPY(b8);
+ COPY(a0); COPY(a1); COPY(a2); COPY(a3); COPY(a5); COPY(a7); COPY(a9);
+ COPY(b0); COPY(b1); COPY(b2); COPY(b3); COPY(b5); COPY(b7); COPY(b9);
+
+ COPY(a16); COPY(a17); COPY(a18); COPY(a19);
+ COPY(a20); COPY(a21); COPY(a22); COPY(a23);
+ COPY(a24); COPY(a25); COPY(a26); COPY(a27);
+ COPY(a28); COPY(a29); COPY(a30); COPY(a31);
+ COPY(b16); COPY(b17); COPY(b18); COPY(b19);
+ COPY(b20); COPY(b21); COPY(b22); COPY(b23);
+ COPY(b24); COPY(b25); COPY(b26); COPY(b27);
+ COPY(b28); COPY(b29); COPY(b30); COPY(b31);
+
+ COPY(csr); COPY(pc);
+
+#undef COPY
+
+ return err;
+}
+
+asmlinkage int do_rt_sigreturn(struct pt_regs *regs)
+{
+ struct rt_sigframe __user *frame;
+ sigset_t set;
+
+ /*
+ * Since we stacked the signal on a dword boundary,
+ * 'sp' should be dword aligned here. If it's
+ * not, then the user is trying to mess with us.
+ */
+ if (regs->sp & 7)
+ goto badframe;
+
+ frame = (struct rt_sigframe __user *) ((unsigned long) regs->sp + 8);
+
+ if (!access_ok(VERIFY_READ, frame, sizeof(*frame)))
+ goto badframe;
+ if (__copy_from_user(&set, &frame->uc.uc_sigmask, sizeof(set)))
+ goto badframe;
+
+ sigdelsetmask(&set, ~_BLOCKABLE);
+ spin_lock_irq(¤t->sighand->siglock);
+ current->blocked = set;
+ recalc_sigpending();
+ spin_unlock_irq(¤t->sighand->siglock);
+
+ if (restore_sigcontext(regs, &frame->uc.uc_mcontext))
+ goto badframe;
+
+ return regs->a4;
+
+badframe:
+ force_sig(SIGSEGV, current);
+ return 0;
+}
+
+static int setup_sigcontext(struct sigcontext __user *sc, struct pt_regs *regs,
+ unsigned long mask)
+{
+ int err = 0;
+
+ err |= __put_user(mask, &sc->sc_mask);
+
+ /* The access_ok check was done by caller, so use __put_user here */
+#define COPY(x) (err |= __put_user(regs->x, &sc->sc_##x))
+
+ COPY(sp); COPY(a4); COPY(b4); COPY(a6); COPY(b6); COPY(a8); COPY(b8);
+ COPY(a0); COPY(a1); COPY(a2); COPY(a3); COPY(a5); COPY(a7); COPY(a9);
+ COPY(b0); COPY(b1); COPY(b2); COPY(b3); COPY(b5); COPY(b7); COPY(b9);
+
+ COPY(a16); COPY(a17); COPY(a18); COPY(a19);
+ COPY(a20); COPY(a21); COPY(a22); COPY(a23);
+ COPY(a24); COPY(a25); COPY(a26); COPY(a27);
+ COPY(a28); COPY(a29); COPY(a30); COPY(a31);
+ COPY(b16); COPY(b17); COPY(b18); COPY(b19);
+ COPY(b20); COPY(b21); COPY(b22); COPY(b23);
+ COPY(b24); COPY(b25); COPY(b26); COPY(b27);
+ COPY(b28); COPY(b29); COPY(b30); COPY(b31);
+
+ COPY(csr); COPY(pc);
+
+#undef COPY
+
+ return err;
+}
+
+static inline void __user *get_sigframe(struct k_sigaction *ka,
+ struct pt_regs *regs,
+ unsigned long framesize)
+{
+ unsigned long sp = regs->sp;
+
+ /*
+ * This is the X/Open sanctioned signal stack switching.
+ */
+ if ((ka->sa.sa_flags & SA_ONSTACK) && sas_ss_flags(sp) == 0)
+ sp = current->sas_ss_sp + current->sas_ss_size;
+
+ /*
+ * No matter what happens, 'sp' must be dword
+ * aligned. Otherwise, nasty things will happen
+ */
+ return (void __user *)((sp - framesize) & ~7);
+}
+
+static int setup_rt_frame(int signr, struct k_sigaction *ka, siginfo_t *info,
+ sigset_t *set, struct pt_regs *regs)
+{
+ struct rt_sigframe __user *frame;
+ unsigned long __user *retcode;
+ int err = 0;
+
+ frame = get_sigframe(ka, regs, sizeof(*frame));
+
+ if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame)))
+ goto segv_and_exit;
+
+ err |= __put_user(&frame->info, &frame->pinfo);
+ err |= __put_user(&frame->uc, &frame->puc);
+ err |= copy_siginfo_to_user(&frame->info, info);
+
+ /* Clear all the bits of the ucontext we don't use. */
+ err |= __clear_user(&frame->uc, offsetof(struct ucontext, uc_mcontext));
+
+ err |= setup_sigcontext(&frame->uc.uc_mcontext, regs, set->sig[0]);
+ err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set));
+
+ /* Set up to return from userspace */
+ retcode = (unsigned long __user *) &frame->retcode;
+
+ /* The access_ok check was done above, so use __put_user here */
+#define COPY(x) (err |= __put_user(x, retcode++))
+
+ COPY(0x0000002AUL | (__NR_rt_sigreturn << 7));
+ /* MVK __NR_rt_sigreturn,B0 */
+ COPY(0x10000000UL); /* SWE */
+ COPY(0x00006000UL); /* NOP 4 */
+ COPY(0x00006000UL); /* NOP 4 */
+ COPY(0x00006000UL); /* NOP 4 */
+ COPY(0x00006000UL); /* NOP 4 */
+ COPY(0x00006000UL); /* NOP 4 */
+ COPY(0x00006000UL); /* NOP 4 */
+ COPY(0x00006000UL); /* NOP 4 */
+
+#undef COPY
+
+ if (err)
+ goto segv_and_exit;
+
+ flush_icache_range((unsigned long) &frame->retcode,
+ (unsigned long) &frame->retcode + RETCODE_SIZE);
+
+ retcode = (unsigned long __user *) &frame->retcode;
+
+ /* Change user context to branch to signal handler */
+ regs->sp = (unsigned long) frame - 8;
+ regs->b3 = (unsigned long) retcode;
+ regs->pc = (unsigned long) ka->sa.sa_handler;
+
+ /* Give the signal number to the handler */
+ regs->a4 = signr;
+
+ /*
+ * For realtime signals we must also set the second and third
+ * arguments for the signal handler.
+ * -- Peter Maydell <pmaydell@chiark.greenend.org.uk> 2000-12-06
+ */
+ regs->b4 = (unsigned long)&frame->info;
+ regs->a6 = (unsigned long)&frame->uc;
+
+ return 0;
+
+segv_and_exit:
+ force_sigsegv(signr, current);
+ return -EFAULT;
+}
+
+static inline void
+handle_restart(struct pt_regs *regs, struct k_sigaction *ka, int has_handler)
+{
+ switch (regs->a4) {
+ case -ERESTARTNOHAND:
+ if (!has_handler)
+ goto do_restart;
+ regs->a4 = -EINTR;
+ break;
+
+ case -ERESTARTSYS:
+ if (has_handler && !(ka->sa.sa_flags & SA_RESTART)) {
+ regs->a4 = -EINTR;
+ break;
+ }
+ /* fallthrough */
+ case -ERESTARTNOINTR:
+do_restart:
+ regs->a4 = regs->orig_a4;
+ regs->pc -= 4;
+ break;
+ }
+}
+
+/*
+ * handle the actual delivery of a signal to userspace
+ */
+static int handle_signal(int sig,
+ siginfo_t *info, struct k_sigaction *ka,
+ sigset_t *oldset, struct pt_regs *regs,
+ int syscall)
+{
+ int ret;
+
+ /* Are we from a system call? */
+ if (syscall) {
+ /* If so, check system call restarting.. */
+ switch (regs->a4) {
+ case -ERESTART_RESTARTBLOCK:
+ case -ERESTARTNOHAND:
+ regs->a4 = -EINTR;
+ break;
+
+ case -ERESTARTSYS:
+ if (!(ka->sa.sa_flags & SA_RESTART)) {
+ regs->a4 = -EINTR;
+ break;
+ }
+
+ /* fallthrough */
+ case -ERESTARTNOINTR:
+ regs->a4 = regs->orig_a4;
+ regs->pc -= 4;
+ }
+ }
+
+ /* Set up the stack frame */
+ ret = setup_rt_frame(sig, ka, info, oldset, regs);
+ if (ret == 0) {
+ spin_lock_irq(¤t->sighand->siglock);
+ sigorsets(¤t->blocked, ¤t->blocked,
+ &ka->sa.sa_mask);
+ if (!(ka->sa.sa_flags & SA_NODEFER))
+ sigaddset(¤t->blocked, sig);
+ recalc_sigpending();
+ spin_unlock_irq(¤t->sighand->siglock);
+ }
+
+ return ret;
+}
+
+/*
+ * handle a potential signal
+ */
+static void do_signal(struct pt_regs *regs, int syscall)
+{
+ struct k_sigaction ka;
+ siginfo_t info;
+ sigset_t *oldset;
+ int signr;
+
+ /* we want the common case to go fast, which is why we may in certain
+ * cases get here from kernel mode */
+ if (!user_mode(regs))
+ return;
+
+ if (test_thread_flag(TIF_RESTORE_SIGMASK))
+ oldset = ¤t->saved_sigmask;
+ else
+ oldset = ¤t->blocked;
+
+ signr = get_signal_to_deliver(&info, &ka, regs, NULL);
+ if (signr > 0) {
+ if (handle_signal(signr, &info, &ka, oldset,
+ regs, syscall) == 0) {
+ /* a signal was successfully delivered; the saved
+ * sigmask will have been stored in the signal frame,
+ * and will be restored by sigreturn, so we can simply
+ * clear the TIF_RESTORE_SIGMASK flag */
+ if (test_thread_flag(TIF_RESTORE_SIGMASK))
+ clear_thread_flag(TIF_RESTORE_SIGMASK);
+
+ tracehook_signal_handler(signr, &info, &ka, regs, 0);
+ }
+
+ return;
+ }
+
+ /* did we come from a system call? */
+ if (syscall) {
+ /* restart the system call - no handlers present */
+ switch (regs->a4) {
+ case -ERESTARTNOHAND:
+ case -ERESTARTSYS:
+ case -ERESTARTNOINTR:
+ regs->a4 = regs->orig_a4;
+ regs->pc -= 4;
+ break;
+
+ case -ERESTART_RESTARTBLOCK:
+ regs->a4 = regs->orig_a4;
+ regs->b0 = __NR_restart_syscall;
+ regs->pc -= 4;
+ break;
+ }
+ }
+
+ /* if there's no signal to deliver, we just put the saved sigmask
+ * back */
+ if (test_thread_flag(TIF_RESTORE_SIGMASK)) {
+ clear_thread_flag(TIF_RESTORE_SIGMASK);
+ sigprocmask(SIG_SETMASK, ¤t->saved_sigmask, NULL);
+ }
+}
+
+/*
+ * notification of userspace execution resumption
+ * - triggered by current->work.notify_resume
+ */
+asmlinkage void do_notify_resume(struct pt_regs *regs, u32 thread_info_flags,
+ int syscall)
+{
+ /* deal with pending signal delivery */
+ if (thread_info_flags & ((1 << TIF_SIGPENDING) |
+ (1 << TIF_RESTORE_SIGMASK)))
+ do_signal(regs, syscall);
+
+ if (thread_info_flags & (1 << TIF_NOTIFY_RESUME)) {
+ clear_thread_flag(TIF_NOTIFY_RESUME);
+ tracehook_notify_resume(regs);
+ if (current->replacement_session_keyring)
+ key_replace_session_keyring();
+ }
+}
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v3 10/24] C6X: signal management
2011-09-27 20:29 ` [PATCH v3 10/24] C6X: signal management Mark Salter
@ 2011-09-28 13:48 ` Arnd Bergmann
0 siblings, 0 replies; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-28 13:48 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch
On Tuesday 27 September 2011, Mark Salter wrote:
>
> Signed-off-by: Mark Salter <msalter@redhat.com>
> ---
> arch/c6x/include/asm/sigcontext.h | 80 ++++++++
> arch/c6x/include/asm/signal.h | 17 ++
> arch/c6x/kernel/signal.c | 377 +++++++++++++++++++++++++++++++++++++
> 3 files changed, 474 insertions(+), 0 deletions(-)
> create mode 100644 arch/c6x/include/asm/sigcontext.h
> create mode 100644 arch/c6x/include/asm/signal.h
> create mode 100644 arch/c6x/kernel/signal.c
>
Acked-by: Arnd Bergmann <arnd@arndb.de>
^ permalink raw reply [flat|nested] 66+ messages in thread
* [PATCH v3 11/24] C6X: time management
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (9 preceding siblings ...)
2011-09-27 20:29 ` [PATCH v3 10/24] C6X: signal management Mark Salter
@ 2011-09-27 20:29 ` Mark Salter
2011-09-27 23:41 ` Thomas Gleixner
2011-09-27 20:29 ` [PATCH v3 12/24] C6X: interrupt handling Mark Salter
` (14 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:29 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
The C6X architecture includes a 64-bit free running core clock counter which
is used as the clocksource. The SoCs have a number of 64-bit programmable
timers. One of these is used as the clockevent timer.
Signed-off-by: Mark Salter <msalter@redhat.com>
---
arch/c6x/include/asm/timer64.h | 6 ++
arch/c6x/include/asm/timex.h | 38 ++++++++
arch/c6x/kernel/time.c | 65 ++++++++++++++
arch/c6x/platforms/timer64.c | 188 ++++++++++++++++++++++++++++++++++++++++
4 files changed, 297 insertions(+), 0 deletions(-)
create mode 100644 arch/c6x/include/asm/timer64.h
create mode 100644 arch/c6x/include/asm/timex.h
create mode 100644 arch/c6x/kernel/time.c
create mode 100644 arch/c6x/platforms/timer64.c
diff --git a/arch/c6x/include/asm/timer64.h b/arch/c6x/include/asm/timer64.h
new file mode 100644
index 0000000..bbe27bb
--- /dev/null
+++ b/arch/c6x/include/asm/timer64.h
@@ -0,0 +1,6 @@
+#ifndef _C6X_TIMER64_H
+#define _C6X_TIMER64_H
+
+extern void __init timer64_init(void);
+
+#endif /* _C6X_TIMER64_H */
diff --git a/arch/c6x/include/asm/timex.h b/arch/c6x/include/asm/timex.h
new file mode 100644
index 0000000..8739a23
--- /dev/null
+++ b/arch/c6x/include/asm/timex.h
@@ -0,0 +1,38 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010, 2011 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * Modified for 2.6.34: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_TIMEX_H
+#define _ASM_C6X_TIMEX_H
+
+/*
+ * This should be close enough...
+ */
+#define CLOCK_TICK_RATE ((1000 * 1000000UL) / 6)
+
+/* 64-bit timestamp */
+typedef unsigned long long cycles_t;
+
+extern cycles_t cacheflush_time;
+
+static inline cycles_t get_cycles(void)
+{
+ unsigned l, h;
+
+ asm volatile (" dint\n"
+ " mvc .s2 TSCL,%0\n"
+ " mvc .s2 TSCH,%1\n"
+ " rint\n"
+ : "=b"(l), "=b"(h));
+ return ((cycles_t)h << 32) | l;
+}
+
+#endif /* _ASM_C6X_TIMEX_H */
diff --git a/arch/c6x/kernel/time.c b/arch/c6x/kernel/time.c
new file mode 100644
index 0000000..4c9f136
--- /dev/null
+++ b/arch/c6x/kernel/time.c
@@ -0,0 +1,65 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010, 2011 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/kernel.h>
+#include <linux/clocksource.h>
+#include <linux/errno.h>
+#include <linux/sched.h>
+#include <linux/param.h>
+#include <linux/string.h>
+#include <linux/mm.h>
+#include <linux/interrupt.h>
+#include <linux/timex.h>
+#include <linux/profile.h>
+
+#include <asm/timer64.h>
+
+static u32 sched_clock_multiplier;
+#define SCHED_CLOCK_SHIFT 16
+
+static cycle_t tsc_read(struct clocksource *cs)
+{
+ return get_cycles();
+}
+
+static struct clocksource clocksource_tsc = {
+ .name = "timestamp",
+ .rating = 300,
+ .read = tsc_read,
+ .mask = CLOCKSOURCE_MASK(64),
+ .flags = CLOCK_SOURCE_IS_CONTINUOUS,
+};
+
+/*
+ * scheduler clock - returns current time in nanoseconds.
+ */
+u64 sched_clock(void)
+{
+ u64 tsc = get_cycles();
+
+ return (tsc * sched_clock_multiplier) >> SCHED_CLOCK_SHIFT;
+}
+
+void time_init(void)
+{
+ u64 tmp = (u64)NSEC_PER_SEC << SCHED_CLOCK_SHIFT;
+
+ do_div(tmp, c6x_core_freq);
+ sched_clock_multiplier = tmp;
+
+ clocksource_register_hz(&clocksource_tsc, c6x_core_freq);
+
+ /* write anything into TSCL to enable counting */
+ set_creg(TSCL, 0);
+
+ /* probe for timer64 event timer */
+ timer64_init();
+}
diff --git a/arch/c6x/platforms/timer64.c b/arch/c6x/platforms/timer64.c
new file mode 100644
index 0000000..1894eb0
--- /dev/null
+++ b/arch/c6x/platforms/timer64.c
@@ -0,0 +1,188 @@
+/*
+ * Copyright (C) 2010, 2011 Texas Instruments Incorporated
+ * Contributed by: Mark Salter (msalter@redhat.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/clockchips.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/of.h>
+#include <linux/of_irq.h>
+#include <linux/of_address.h>
+#include <asm/soc.h>
+#include <asm/dscr.h>
+#include <asm/timer64.h>
+
+struct timer_regs {
+ u32 reserved0;
+ u32 emumgt;
+ u32 reserved1;
+ u32 reserved2;
+ u32 cntlo;
+ u32 cnthi;
+ u32 prdlo;
+ u32 prdhi;
+ u32 tcr;
+ u32 tgcr;
+ u32 wdtcr;
+};
+
+static struct timer_regs __iomem *timer;
+
+#define TCR_TSTATLO 0x001
+#define TCR_INVOUTPLO 0x002
+#define TCR_INVINPLO 0x004
+#define TCR_CPLO 0x008
+#define TCR_ENAMODELO_ONCE 0x040
+#define TCR_ENAMODELO_CONT 0x080
+#define TCR_ENAMODELO_MASK 0x0c0
+#define TCR_PWIDLO_MASK 0x030
+#define TCR_CLKSRCLO 0x100
+#define TCR_TIENLO 0x200
+#define TCR_TSTATHI (0x001 << 16)
+#define TCR_INVOUTPHI (0x002 << 16)
+#define TCR_CPHI (0x008 << 16)
+#define TCR_PWIDHI_MASK (0x030 << 16)
+#define TCR_ENAMODEHI_ONCE (0x040 << 16)
+#define TCR_ENAMODEHI_CONT (0x080 << 16)
+#define TCR_ENAMODEHI_MASK (0x0c0 << 16)
+
+#define TGCR_TIMLORS 0x001
+#define TGCR_TIMHIRS 0x002
+#define TGCR_TIMMODE_UD32 0x004
+#define TGCR_TIMMODE_WDT64 0x008
+#define TGCR_TIMMODE_CD32 0x00c
+#define TGCR_TIMMODE_MASK 0x00c
+#define TGCR_PSCHI_MASK (0x00f << 8)
+#define TGCR_TDDRHI_MASK (0x00f << 12)
+
+/*
+ * Timer clocks are divided down from the CPU clock
+ * The divisor is in the EMUMGTCLKSPD register
+ */
+#define TIMER_DIVISOR \
+ ((soc_readl(&timer->emumgt) & (0xf << 16)) >> 16)
+
+#define timer_period(f, d) (((f) * 1000000) / ((d) * HZ))
+#define ticks2usecs(f, d, x) (((x) * (d)) / (f))
+
+
+static int next_event(unsigned long delta,
+ struct clock_event_device *evt)
+{
+ soc_writel(soc_readl(&timer->tcr) & ~TCR_ENAMODELO_MASK, &timer->tcr);
+ soc_writel(delta - 1, &timer->prdlo);
+ soc_writel(0, &timer->cntlo);
+ soc_writel(soc_readl(&timer->tcr) | TCR_ENAMODELO_ONCE, &timer->tcr);
+
+ return 0;
+}
+
+static void set_clock_mode(enum clock_event_mode mode,
+ struct clock_event_device *evt)
+{
+}
+
+static struct clock_event_device t64_clockevent_device = {
+ .name = "TIMER64_EVT32_TIMER",
+ .features = CLOCK_EVT_FEAT_ONESHOT,
+ .rating = 200,
+ .set_mode = set_clock_mode,
+ .set_next_event = next_event,
+};
+
+static irqreturn_t timer_interrupt(int irq, void *dev_id)
+{
+ struct clock_event_device *cd = &t64_clockevent_device;
+
+ cd->event_handler(cd);
+
+ return IRQ_HANDLED;
+}
+
+static struct irqaction timer_iact = {
+ .name = "timer",
+ .flags = IRQF_TIMER,
+ .handler = timer_interrupt,
+ .dev_id = &t64_clockevent_device,
+};
+
+void __init timer64_init(void)
+{
+ struct clock_event_device *cd = &t64_clockevent_device;
+ struct device_node *np, *first = NULL;
+ u32 val;
+ int err, found = 0;
+
+ for_each_compatible_node(np, NULL, "ti,c64x+timer64") {
+ err = of_property_read_u32_array(np, "ti,core-mask", &val, 1);
+ if (!err) {
+ if (val & (1 << get_coreid())) {
+ found = 1;
+ break;
+ }
+ } else if (!first)
+ first = np;
+ }
+ if (!found) {
+ /* try first one with no core-mask */
+ if (first)
+ np = of_node_get(first);
+ else {
+ pr_debug("Cannot find ti,c64x+timer64 timer.\n");
+ return;
+ }
+ }
+
+ timer = of_iomap(np, 0);
+ if (!timer) {
+ pr_debug("%s: Cannot map timer registers.\n", np->full_name);
+ goto out;
+ }
+ pr_debug("%s: Timer registers=%p.\n", np->full_name, timer);
+
+ cd->irq = irq_of_parse_and_map(np, 0);
+ if (cd->irq == NO_IRQ) {
+ pr_debug("%s: Cannot find interrupt.\n", np->full_name);
+ iounmap(timer);
+ goto out;
+ }
+
+ /* If there is a device state control, enable it. */
+ err = of_property_read_u32_array(np, "ti,dscr-dev-enable", &val, 1);
+ if (!err)
+ dscr_set_devstate(val, DSCR_DEVSTATE_ENABLED);
+
+ /* disable timer, reset count */
+ soc_writel(soc_readl(&timer->tcr) & ~TCR_ENAMODELO_MASK, &timer->tcr);
+ soc_writel(0, &timer->prdlo);
+
+ /* use internal clock and 1 cycle pulse width */
+ val = soc_readl(&timer->tcr);
+ soc_writel(val & ~(TCR_CLKSRCLO | TCR_PWIDLO_MASK), &timer->tcr);
+
+ /* dual 32-bit unchained mode */
+ val = soc_readl(&timer->tgcr) & ~TGCR_TIMMODE_MASK;
+ soc_writel(val, &timer->tgcr);
+ soc_writel(val | (TGCR_TIMLORS | TGCR_TIMMODE_UD32), &timer->tgcr);
+
+ pr_debug("%s: Timer irq=%d.\n", np->full_name, cd->irq);
+
+ clockevents_calc_mult_shift(cd, c6x_core_freq / TIMER_DIVISOR, 5);
+
+ cd->max_delta_ns = clockevent_delta2ns(0x7fffffff, cd);
+ cd->min_delta_ns = clockevent_delta2ns(250, cd);
+
+ cd->cpumask = cpumask_of(smp_processor_id());
+
+ clockevents_register_device(cd);
+ setup_irq(cd->irq, &timer_iact);
+
+out:
+ of_node_put(np);
+ return;
+}
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v3 11/24] C6X: time management
2011-09-27 20:29 ` [PATCH v3 11/24] C6X: time management Mark Salter
@ 2011-09-27 23:41 ` Thomas Gleixner
2011-09-28 12:48 ` Mark Salter
0 siblings, 1 reply; 66+ messages in thread
From: Thomas Gleixner @ 2011-09-27 23:41 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch
On Tue, 27 Sep 2011, Mark Salter wrote:
> The C6X architecture includes a 64-bit free running core clock counter which
> is used as the clocksource. The SoCs have a number of 64-bit programmable
> timers. One of these is used as the clockevent timer.
>
> + * Port on Texas Instruments TMS320C6x architecture
> + *
> + * Copyright (C) 2004, 2009, 2010, 2011 Texas Instruments Incorporated
> + * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
> + *
> + * Modified for 2.6.34: Mark Salter <msalter@redhat.com>
Hmm. Who wrote that code and who is assigned the authorship? If the
original Author is Aurelian, then this patch is missing:
- An extra From: ....
- A SOB from the orginal author or an explnation why it's not
available.
If not, then this want's to be documented in a proper form.
Please do not attribute other peoples code to yourself w/o a fcking
good reason. And porting forward itself is not considered to be one.
> +/*
> + * This should be close enough...
What's the point of that comment ?
> + */
> +#define CLOCK_TICK_RATE ((1000 * 1000000UL) / 6)
> +
> +/* 64-bit timestamp */
> +typedef unsigned long long cycles_t;
...
> +static int next_event(unsigned long delta,
> + struct clock_event_device *evt)
> +{
> + soc_writel(soc_readl(&timer->tcr) & ~TCR_ENAMODELO_MASK, &timer->tcr);
> + soc_writel(delta - 1, &timer->prdlo);
> + soc_writel(0, &timer->cntlo);
> + soc_writel(soc_readl(&timer->tcr) | TCR_ENAMODELO_ONCE, &timer->tcr);
> +
> + return 0;
> +}
> +
> +static void set_clock_mode(enum clock_event_mode mode,
> + struct clock_event_device *evt)
> +{
So there is no way to disable that thing ?
> +}
> +
> +static struct clock_event_device t64_clockevent_device = {
> + .name = "TIMER64_EVT32_TIMER",
> + .features = CLOCK_EVT_FEAT_ONESHOT,
> + .rating = 200,
> + .set_mode = set_clock_mode,
> + .set_next_event = next_event,
> +};
> +
> +static irqreturn_t timer_interrupt(int irq, void *dev_id)
> +{
> + struct clock_event_device *cd = &t64_clockevent_device;
If proper set up, then you can just hand in dev_id to the handler.
> +
> + cd->event_handler(cd);
> +
> + return IRQ_HANDLED;
> +}
> +
> +static struct irqaction timer_iact = {
> + .name = "timer",
> + .flags = IRQF_TIMER,
> + .handler = timer_interrupt,
> + .dev_id = &t64_clockevent_device,
Which seems to be the case :)
Thanks,
tglx
^ permalink raw reply [flat|nested] 66+ messages in thread* Re: [PATCH v3 11/24] C6X: time management
2011-09-27 23:41 ` Thomas Gleixner
@ 2011-09-28 12:48 ` Mark Salter
0 siblings, 0 replies; 66+ messages in thread
From: Mark Salter @ 2011-09-28 12:48 UTC (permalink / raw)
To: Thomas Gleixner; +Cc: linux-kernel, linux-arch
On Wed, 2011-09-28 at 01:41 +0200, Thomas Gleixner wrote:
> On Tue, 27 Sep 2011, Mark Salter wrote:
>
> > The C6X architecture includes a 64-bit free running core clock counter which
> > is used as the clocksource. The SoCs have a number of 64-bit programmable
> > timers. One of these is used as the clockevent timer.
> >
> > + * Port on Texas Instruments TMS320C6x architecture
> > + *
> > + * Copyright (C) 2004, 2009, 2010, 2011 Texas Instruments Incorporated
> > + * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
> > + *
> > + * Modified for 2.6.34: Mark Salter <msalter@redhat.com>
>
> Hmm. Who wrote that code and who is assigned the authorship? If the
> original Author is Aurelian, then this patch is missing:
>
> - An extra From: ....
>
> - A SOB from the orginal author or an explnation why it's not
> available.
>
> If not, then this want's to be documented in a proper form.
>
> Please do not attribute other peoples code to yourself w/o a fcking
> good reason. And porting forward itself is not considered to be one.
>
Indeed not. I was focused on the code in putting the series together and
completely overlooked the proper attribution and CCs.
> > +/*
> > + * This should be close enough...
>
> What's the point of that comment ?
None. It shouldn't be there.
>
> > + */
> > +#define CLOCK_TICK_RATE ((1000 * 1000000UL) / 6)
> > +
> > +/* 64-bit timestamp */
> > +typedef unsigned long long cycles_t;
>
> ...
>
> > +static int next_event(unsigned long delta,
> > + struct clock_event_device *evt)
> > +{
> > + soc_writel(soc_readl(&timer->tcr) & ~TCR_ENAMODELO_MASK, &timer->tcr);
> > + soc_writel(delta - 1, &timer->prdlo);
> > + soc_writel(0, &timer->cntlo);
> > + soc_writel(soc_readl(&timer->tcr) | TCR_ENAMODELO_ONCE, &timer->tcr);
> > +
> > + return 0;
> > +}
> > +
> > +static void set_clock_mode(enum clock_event_mode mode,
> > + struct clock_event_device *evt)
> > +{
>
> So there is no way to disable that thing ?
I will check...
--Mark
^ permalink raw reply [flat|nested] 66+ messages in thread
* [PATCH v3 12/24] C6X: interrupt handling
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (10 preceding siblings ...)
2011-09-27 20:29 ` [PATCH v3 11/24] C6X: time management Mark Salter
@ 2011-09-27 20:29 ` Mark Salter
2011-09-27 23:30 ` Thomas Gleixner
2011-09-27 20:29 ` [PATCH v3 13/24] C6X: syscalls Mark Salter
` (13 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:29 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
The C6X core includes an interrupt controller with 12 inputs of fixed priority.
Each core is contained within a core "megamodule" which includes another
interrupt controller which cascades into the prioritized core controller. The
megamodule PIC has four interrupt combiners; each with up to 32 inputs which
are combined into one interrupt. The megamodule contains a set of MUX registers
which are used map megamodule interrupt sources to one of 12 core interrupts.
Any one of these 12 MUX outputs feeding into the core interrupt controller may
be the output from one of the four interrupt combiners or an individual
megamodule interrupt source. An individual megamodule interrupt source will
bypass the combiners if the MUX registers are set up such that the individual
interrupt feeds directly to one of the 12 core interrupts. This priority
mapping of interrupts is supported in the devicetree. For details, see
Documentation/devicetree/bindings/c6x/interrupt.txt
The code in arch/c6x/kernel/irq.c which supports mapping of hardware to virtual
interrupts was taken as-is from powerpc. There is ongoing work to generalize
the virtual to hardware mapping and plans are to convert powerpc to use this.
C6X will also need to be converted over when the generic code is ready.
Signed-off-by: Mark Salter <msalter@redhat.com>
---
arch/c6x/include/asm/hardirq.h | 20 +
arch/c6x/include/asm/irq.h | 302 +++++++++++++++
arch/c6x/include/asm/irqflags.h | 72 ++++
arch/c6x/include/asm/megamod-pic.h | 9 +
arch/c6x/kernel/irq.c | 730 ++++++++++++++++++++++++++++++++++++
arch/c6x/platforms/megamod-pic.c | 349 +++++++++++++++++
6 files changed, 1482 insertions(+), 0 deletions(-)
create mode 100644 arch/c6x/include/asm/hardirq.h
create mode 100644 arch/c6x/include/asm/irq.h
create mode 100644 arch/c6x/include/asm/irqflags.h
create mode 100644 arch/c6x/include/asm/megamod-pic.h
create mode 100644 arch/c6x/kernel/irq.c
create mode 100644 arch/c6x/platforms/megamod-pic.c
diff --git a/arch/c6x/include/asm/hardirq.h b/arch/c6x/include/asm/hardirq.h
new file mode 100644
index 0000000..9621954
--- /dev/null
+++ b/arch/c6x/include/asm/hardirq.h
@@ -0,0 +1,20 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#ifndef _ASM_C6X_HARDIRQ_H
+#define _ASM_C6X_HARDIRQ_H
+
+extern void ack_bad_irq(int irq);
+#define ack_bad_irq ack_bad_irq
+
+#include <asm-generic/hardirq.h>
+
+#endif /* _ASM_C6X_HARDIRQ_H */
diff --git a/arch/c6x/include/asm/irq.h b/arch/c6x/include/asm/irq.h
new file mode 100644
index 0000000..a6ae3c9
--- /dev/null
+++ b/arch/c6x/include/asm/irq.h
@@ -0,0 +1,302 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2006, 2009, 2010, 2011 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * Large parts taken directly from powerpc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_IRQ_H
+#define _ASM_C6X_IRQ_H
+
+#include <linux/threads.h>
+#include <linux/list.h>
+#include <linux/radix-tree.h>
+#include <asm/percpu.h>
+
+#define irq_canonicalize(irq) (irq)
+
+/*
+ * The C64X+ core has 16 IRQ vectors. One each is used by Reset and NMI. Two
+ * are reserved. The remaining 12 vectors are used to route SoC interrupts.
+ * These interrupt vectors are prioritized with IRQ 4 having the highest
+ * priority and IRQ 15 having the lowest.
+ *
+ * The C64x+ megamodule provides a PIC which combines SoC IRQ sources into a
+ * single core IRQ vector. There are four combined sources, each of which
+ * feed into one of the 12 general interrupt vectors. The remaining 8 vectors
+ * can each route a single SoC interrupt directly.
+ */
+#define NR_PRIORITY_IRQS 16
+
+#define NR_IRQS_LEGACY NR_PRIORITY_IRQS
+
+/* Total number of virq in the platform */
+#define NR_IRQS 256
+
+/* This number is used when no interrupt has been assigned */
+#define NO_IRQ 0
+
+/* This type is the placeholder for a hardware interrupt number. It has to
+ * be big enough to enclose whatever representation is used by a given
+ * platform.
+ */
+typedef unsigned long irq_hw_number_t;
+
+/* Interrupt controller "host" data structure. This could be defined as a
+ * irq domain controller. That is, it handles the mapping between hardware
+ * and virtual interrupt numbers for a given interrupt domain. The host
+ * structure is generally created by the PIC code for a given PIC instance
+ * (though a host can cover more than one PIC if they have a flat number
+ * model). It's the host callbacks that are responsible for setting the
+ * irq_chip on a given irq_desc after it's been mapped.
+ *
+ * The host code and data structures are fairly agnostic to the fact that
+ * we use an open firmware device-tree. We do have references to struct
+ * device_node in two places: in irq_find_host() to find the host matching
+ * a given interrupt controller node, and of course as an argument to its
+ * counterpart host->ops->match() callback. However, those are treated as
+ * generic pointers by the core and the fact that it's actually a device-node
+ * pointer is purely a convention between callers and implementation. This
+ * code could thus be used on other architectures by replacing those two
+ * by some sort of arch-specific void * "token" used to identify interrupt
+ * controllers.
+ */
+struct irq_host;
+struct radix_tree_root;
+struct device_node;
+
+/* Functions below are provided by the host and called whenever a new mapping
+ * is created or an old mapping is disposed. The host can then proceed to
+ * whatever internal data structures management is required. It also needs
+ * to setup the irq_desc when returning from map().
+ */
+struct irq_host_ops {
+ /* Match an interrupt controller device node to a host, returns
+ * 1 on a match
+ */
+ int (*match)(struct irq_host *h, struct device_node *node);
+
+ /* Create or update a mapping between a virtual irq number and a hw
+ * irq number. This is called only once for a given mapping.
+ */
+ int (*map)(struct irq_host *h, unsigned int virq, irq_hw_number_t hw);
+
+ /* Dispose of such a mapping */
+ void (*unmap)(struct irq_host *h, unsigned int virq);
+
+ /* Translate device-tree interrupt specifier from raw format coming
+ * from the firmware to a irq_hw_number_t (interrupt line number) and
+ * type (sense) that can be passed to set_irq_type(). In the absence
+ * of this callback, irq_create_of_mapping() and irq_of_parse_and_map()
+ * will return the hw number in the first cell and IRQ_TYPE_NONE for
+ * the type (which amount to keeping whatever default value the
+ * interrupt controller has for that line)
+ */
+ int (*xlate)(struct irq_host *h, struct device_node *ctrler,
+ const u32 *intspec, unsigned int intsize,
+ irq_hw_number_t *out_hwirq, unsigned int *out_type);
+};
+
+struct irq_host {
+ struct list_head link;
+
+ /* type of reverse mapping technique */
+ unsigned int revmap_type;
+#define IRQ_HOST_MAP_PRIORITY 0 /* core priority irqs, get irqs 1..15 */
+#define IRQ_HOST_MAP_NOMAP 1 /* no fast reverse mapping */
+#define IRQ_HOST_MAP_LINEAR 2 /* linear map of interrupts */
+#define IRQ_HOST_MAP_TREE 3 /* radix tree */
+ union {
+ struct {
+ unsigned int size;
+ unsigned int *revmap;
+ } linear;
+ struct radix_tree_root tree;
+ } revmap_data;
+ struct irq_host_ops *ops;
+ void *host_data;
+ irq_hw_number_t inval_irq;
+
+ /* Optional device node pointer */
+ struct device_node *of_node;
+};
+
+struct irq_data;
+extern irq_hw_number_t irqd_to_hwirq(struct irq_data *d);
+extern irq_hw_number_t virq_to_hw(unsigned int virq);
+extern bool virq_is_host(unsigned int virq, struct irq_host *host);
+
+/**
+ * irq_alloc_host - Allocate a new irq_host data structure
+ * @of_node: optional device-tree node of the interrupt controller
+ * @revmap_type: type of reverse mapping to use
+ * @revmap_arg: for IRQ_HOST_MAP_LINEAR linear only: size of the map
+ * @ops: map/unmap host callbacks
+ * @inval_irq: provide a hw number in that host space that is always invalid
+ *
+ * Allocates and initialize and irq_host structure. Note that in the case of
+ * IRQ_HOST_MAP_LEGACY, the map() callback will be called before this returns
+ * for all legacy interrupts except 0 (which is always the invalid irq for
+ * a legacy controller). For a IRQ_HOST_MAP_LINEAR, the map is allocated by
+ * this call as well. For a IRQ_HOST_MAP_TREE, the radix tree will be allocated
+ * later during boot automatically (the reverse mapping will use the slow path
+ * until that happens).
+ */
+extern struct irq_host *irq_alloc_host(struct device_node *of_node,
+ unsigned int revmap_type,
+ unsigned int revmap_arg,
+ struct irq_host_ops *ops,
+ irq_hw_number_t inval_irq);
+
+
+/**
+ * irq_find_host - Locates a host for a given device node
+ * @node: device-tree node of the interrupt controller
+ */
+extern struct irq_host *irq_find_host(struct device_node *node);
+
+
+/**
+ * irq_set_default_host - Set a "default" host
+ * @host: default host pointer
+ *
+ * For convenience, it's possible to set a "default" host that will be used
+ * whenever NULL is passed to irq_create_mapping(). It makes life easier for
+ * platforms that want to manipulate a few hard coded interrupt numbers that
+ * aren't properly represented in the device-tree.
+ */
+extern void irq_set_default_host(struct irq_host *host);
+
+
+/**
+ * irq_set_virq_count - Set the maximum number of virt irqs
+ * @count: number of linux virtual irqs, capped with NR_IRQS
+ *
+ * This is mainly for use by platforms like iSeries who want to program
+ * the virtual irq number in the controller to avoid the reverse mapping
+ */
+extern void irq_set_virq_count(unsigned int count);
+
+
+/**
+ * irq_create_mapping - Map a hardware interrupt into linux virq space
+ * @host: host owning this hardware interrupt or NULL for default host
+ * @hwirq: hardware irq number in that host space
+ *
+ * Only one mapping per hardware interrupt is permitted. Returns a linux
+ * virq number.
+ * If the sense/trigger is to be specified, set_irq_type() should be called
+ * on the number returned from that call.
+ */
+extern unsigned int irq_create_mapping(struct irq_host *host,
+ irq_hw_number_t hwirq);
+
+
+/**
+ * irq_dispose_mapping - Unmap an interrupt
+ * @virq: linux virq number of the interrupt to unmap
+ */
+extern void irq_dispose_mapping(unsigned int virq);
+
+/**
+ * irq_find_mapping - Find a linux virq from an hw irq number.
+ * @host: host owning this hardware interrupt
+ * @hwirq: hardware irq number in that host space
+ *
+ * This is a slow path, for use by generic code. It's expected that an
+ * irq controller implementation directly calls the appropriate low level
+ * mapping function.
+ */
+extern unsigned int irq_find_mapping(struct irq_host *host,
+ irq_hw_number_t hwirq);
+
+/**
+ * irq_create_direct_mapping - Allocate a virq for direct mapping
+ * @host: host to allocate the virq for or NULL for default host
+ *
+ * This routine is used for irq controllers which can choose the hardware
+ * interrupt numbers they generate. In such a case it's simplest to use
+ * the linux virq as the hardware interrupt number.
+ */
+extern unsigned int irq_create_direct_mapping(struct irq_host *host);
+
+/**
+ * irq_radix_revmap_insert - Insert a hw irq to linux virq number mapping.
+ * @host: host owning this hardware interrupt
+ * @virq: linux irq number
+ * @hwirq: hardware irq number in that host space
+ *
+ * This is for use by irq controllers that use a radix tree reverse
+ * mapping for fast lookup.
+ */
+extern void irq_radix_revmap_insert(struct irq_host *host, unsigned int virq,
+ irq_hw_number_t hwirq);
+
+/**
+ * irq_radix_revmap_lookup - Find a linux virq from a hw irq number.
+ * @host: host owning this hardware interrupt
+ * @hwirq: hardware irq number in that host space
+ *
+ * This is a fast path, for use by irq controller code that uses radix tree
+ * revmaps
+ */
+extern unsigned int irq_radix_revmap_lookup(struct irq_host *host,
+ irq_hw_number_t hwirq);
+
+/**
+ * irq_linear_revmap - Find a linux virq from a hw irq number.
+ * @host: host owning this hardware interrupt
+ * @hwirq: hardware irq number in that host space
+ *
+ * This is a fast path, for use by irq controller code that uses linear
+ * revmaps. It does fallback to the slow path if the revmap doesn't exist
+ * yet and will create the revmap entry with appropriate locking
+ */
+
+extern unsigned int irq_linear_revmap(struct irq_host *host,
+ irq_hw_number_t hwirq);
+
+
+
+/**
+ * irq_alloc_virt - Allocate virtual irq numbers
+ * @host: host owning these new virtual irqs
+ * @count: number of consecutive numbers to allocate
+ * @hint: pass a hint number, the allocator will try to use a 1:1 mapping
+ *
+ * This is a low level function that is used internally by irq_create_mapping()
+ * and that can be used by some irq controllers implementations for things
+ * like allocating ranges of numbers for MSIs. The revmaps are left untouched.
+ */
+extern unsigned int irq_alloc_virt(struct irq_host *host,
+ unsigned int count,
+ unsigned int hint);
+
+/**
+ * irq_free_virt - Free virtual irq numbers
+ * @virq: virtual irq number of the first interrupt to free
+ * @count: number of interrupts to free
+ *
+ * This function is the opposite of irq_alloc_virt. It will not clear reverse
+ * maps, this should be done previously by unmap'ing the interrupt. In fact,
+ * all interrupts covered by the range being freed should have been unmapped
+ * prior to calling this.
+ */
+extern void irq_free_virt(unsigned int virq, unsigned int count);
+
+extern void __init init_pic_c64xplus(void);
+
+extern void init_IRQ(void);
+
+struct pt_regs;
+
+extern asmlinkage void c6x_do_IRQ(unsigned int prio, struct pt_regs *regs);
+
+extern unsigned long irq_err_count;
+
+#endif /* _ASM_C6X_IRQ_H */
diff --git a/arch/c6x/include/asm/irqflags.h b/arch/c6x/include/asm/irqflags.h
new file mode 100644
index 0000000..cf78e09
--- /dev/null
+++ b/arch/c6x/include/asm/irqflags.h
@@ -0,0 +1,72 @@
+/*
+ * C6X IRQ flag handling
+ *
+ * Copyright (C) 2010 Texas Instruments Incorporated
+ * Written by Mark Salter (msalter@redhat.com)
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public Licence
+ * as published by the Free Software Foundation; either version
+ * 2 of the Licence, or (at your option) any later version.
+ */
+
+#ifndef _ASM_IRQFLAGS_H
+#define _ASM_IRQFLAGS_H
+
+#ifndef __ASSEMBLY__
+
+/* read interrupt enabled status */
+static inline unsigned long arch_local_save_flags(void)
+{
+ unsigned long flags;
+
+ asm volatile (" mvc .s2 CSR,%0\n" : "=b"(flags));
+ return flags;
+}
+
+/* set interrupt enabled status */
+static inline void arch_local_irq_restore(unsigned long flags)
+{
+ asm volatile (" mvc .s2 %0,CSR\n" : : "b"(flags));
+}
+
+/* unconditionally enable interrupts */
+static inline void arch_local_irq_enable(void)
+{
+ unsigned long flags = arch_local_save_flags();
+ flags |= 1;
+ arch_local_irq_restore(flags);
+}
+
+/* unconditionally disable interrupts */
+static inline void arch_local_irq_disable(void)
+{
+ unsigned long flags = arch_local_save_flags();
+ flags &= ~1;
+ arch_local_irq_restore(flags);
+}
+
+/* get status and disable interrupts */
+static inline unsigned long arch_local_irq_save(void)
+{
+ unsigned long flags;
+
+ flags = arch_local_save_flags();
+ arch_local_irq_restore(flags & ~1);
+ return flags;
+}
+
+/* test flags */
+static inline int arch_irqs_disabled_flags(unsigned long flags)
+{
+ return (flags & 1) == 0;
+}
+
+/* test hardware interrupt enable bit */
+static inline int arch_irqs_disabled(void)
+{
+ return arch_irqs_disabled_flags(arch_local_save_flags());
+}
+
+#endif /* __ASSEMBLY__ */
+#endif /* __ASM_IRQFLAGS_H */
diff --git a/arch/c6x/include/asm/megamod-pic.h b/arch/c6x/include/asm/megamod-pic.h
new file mode 100644
index 0000000..eca0a86
--- /dev/null
+++ b/arch/c6x/include/asm/megamod-pic.h
@@ -0,0 +1,9 @@
+#ifndef _C6X_MEGAMOD_PIC_H
+#define _C6X_MEGAMOD_PIC_H
+
+#ifdef __KERNEL__
+
+extern void __init megamod_pic_init(void);
+
+#endif /* __KERNEL__ */
+#endif /* _C6X_MEGAMOD_PIC_H */
diff --git a/arch/c6x/kernel/irq.c b/arch/c6x/kernel/irq.c
new file mode 100644
index 0000000..832e02d
--- /dev/null
+++ b/arch/c6x/kernel/irq.c
@@ -0,0 +1,730 @@
+/*
+ * Copyright (C) 2011 Texas Instruments Incorporated
+ *
+ * This borrows heavily from powerpc version, which is:
+ *
+ * Derived from arch/i386/kernel/irq.c
+ * Copyright (C) 1992 Linus Torvalds
+ * Adapted from arch/i386 by Gary Thomas
+ * Copyright (C) 1995-1996 Gary Thomas (gdt@linuxppc.org)
+ * Updated and modified by Cort Dougan <cort@fsmlabs.com>
+ * Copyright (C) 1996-2001 Cort Dougan
+ * Adapted for Power Macintosh by Paul Mackerras
+ * Copyright (C) 1996 Paul Mackerras (paulus@cs.anu.edu.au)
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version
+ * 2 of the License, or (at your option) any later version.
+ */
+#include <linux/slab.h>
+#include <linux/seq_file.h>
+#include <linux/radix-tree.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_irq.h>
+#include <linux/interrupt.h>
+#include <linux/kernel_stat.h>
+
+#include <asm/megamod-pic.h>
+
+unsigned long irq_err_count;
+
+static raw_spinlock_t core_irq_lock;
+
+static void mask_core_irq(struct irq_data *data)
+{
+ unsigned int prio = data->irq;
+
+ BUG_ON(prio < 4 || prio >= NR_PRIORITY_IRQS);
+
+ raw_spin_lock(&core_irq_lock);
+ and_creg(IER, ~(1 << prio));
+ raw_spin_unlock(&core_irq_lock);
+}
+
+static void unmask_core_irq(struct irq_data *data)
+{
+ unsigned int prio = data->irq;
+
+ raw_spin_lock(&core_irq_lock);
+ or_creg(IER, 1 << prio);
+ raw_spin_unlock(&core_irq_lock);
+}
+
+static struct irq_chip core_chip = {
+ .name = "core",
+ .irq_mask = mask_core_irq,
+ .irq_unmask = unmask_core_irq,
+};
+
+asmlinkage void c6x_do_IRQ(unsigned int prio, struct pt_regs *regs)
+{
+ struct pt_regs *old_regs = set_irq_regs(regs);
+
+ irq_enter();
+
+ BUG_ON(prio < 4 || prio >= NR_PRIORITY_IRQS);
+
+ generic_handle_irq(prio);
+
+ irq_exit();
+
+ set_irq_regs(old_regs);
+}
+
+static struct irq_host *core_host;
+
+static int core_host_map(struct irq_host *h, unsigned int virq,
+ irq_hw_number_t hw)
+{
+ if (hw < 4 || hw >= NR_PRIORITY_IRQS)
+ return -1;
+
+ irq_set_status_flags(virq, IRQ_LEVEL);
+ irq_set_chip_and_handler(virq, &core_chip, handle_level_irq);
+ return 0;
+}
+
+static struct irq_host_ops core_host_ops = {
+ .map = core_host_map,
+};
+
+void __init init_IRQ(void)
+{
+ struct device_node *np;
+
+ raw_spin_lock_init(&core_irq_lock);
+
+ /* Mask all priority IRQs */
+ and_creg(IER, ~0xfff0);
+
+ np = of_find_compatible_node(NULL, NULL, "ti,c64x+core-pic");
+ if (np != NULL) {
+ /* create the core host */
+ core_host = irq_alloc_host(np, IRQ_HOST_MAP_PRIORITY, 0,
+ &core_host_ops, 0);
+ if (core_host)
+ irq_set_default_host(core_host);
+ of_node_put(np);
+ }
+
+ printk(KERN_INFO "Core interrupt controller initialized\n");
+
+ /* now we're ready for other SoC controllers */
+ megamod_pic_init();
+
+ /* Clear all general IRQ flags */
+ set_creg(ICR, 0xfff0);
+}
+
+void ack_bad_irq(int irq)
+{
+ printk(KERN_ERR "IRQ: spurious interrupt %d\n", irq);
+ irq_err_count++;
+}
+
+int arch_show_interrupts(struct seq_file *p, int prec)
+{
+ seq_printf(p, "%*s: %10lu\n", prec, "Err", irq_err_count);
+ return 0;
+}
+
+/*
+ * IRQ controller and virtual interrupts
+ */
+
+/* The main irq map itself is an array of NR_IRQ entries containing the
+ * associate host and irq number. An entry with a host of NULL is free.
+ * An entry can be allocated if it's free, the allocator always then sets
+ * hwirq first to the host's invalid irq number and then fills ops.
+ */
+struct irq_map_entry {
+ irq_hw_number_t hwirq;
+ struct irq_host *host;
+};
+
+static LIST_HEAD(irq_hosts);
+static DEFINE_RAW_SPINLOCK(irq_big_lock);
+static DEFINE_MUTEX(revmap_trees_mutex);
+static struct irq_map_entry irq_map[NR_IRQS];
+static unsigned int irq_virq_count = NR_IRQS;
+static struct irq_host *irq_default_host;
+
+irq_hw_number_t irqd_to_hwirq(struct irq_data *d)
+{
+ return irq_map[d->irq].hwirq;
+}
+EXPORT_SYMBOL_GPL(irqd_to_hwirq);
+
+irq_hw_number_t virq_to_hw(unsigned int virq)
+{
+ return irq_map[virq].hwirq;
+}
+EXPORT_SYMBOL_GPL(virq_to_hw);
+
+bool virq_is_host(unsigned int virq, struct irq_host *host)
+{
+ return irq_map[virq].host == host;
+}
+EXPORT_SYMBOL_GPL(virq_is_host);
+
+static int default_irq_host_match(struct irq_host *h, struct device_node *np)
+{
+ return h->of_node != NULL && h->of_node == np;
+}
+
+struct irq_host *irq_alloc_host(struct device_node *of_node,
+ unsigned int revmap_type,
+ unsigned int revmap_arg,
+ struct irq_host_ops *ops,
+ irq_hw_number_t inval_irq)
+{
+ struct irq_host *host;
+ unsigned int size = sizeof(struct irq_host);
+ unsigned int i;
+ unsigned int *rmap;
+ unsigned long flags;
+
+ /* Allocate structure and revmap table if using linear mapping */
+ if (revmap_type == IRQ_HOST_MAP_LINEAR)
+ size += revmap_arg * sizeof(unsigned int);
+ host = kzalloc(size, GFP_KERNEL);
+ if (host == NULL)
+ return NULL;
+
+ /* Fill structure */
+ host->revmap_type = revmap_type;
+ host->inval_irq = inval_irq;
+ host->ops = ops;
+ host->of_node = of_node_get(of_node);
+
+ if (host->ops->match == NULL)
+ host->ops->match = default_irq_host_match;
+
+ raw_spin_lock_irqsave(&irq_big_lock, flags);
+
+ /* Check for the priority controller. */
+ if (revmap_type == IRQ_HOST_MAP_PRIORITY) {
+ if (irq_map[0].host != NULL) {
+ raw_spin_unlock_irqrestore(&irq_big_lock, flags);
+ of_node_put(host->of_node);
+ kfree(host);
+ return NULL;
+ }
+ irq_map[0].host = host;
+ }
+
+ list_add(&host->link, &irq_hosts);
+ raw_spin_unlock_irqrestore(&irq_big_lock, flags);
+
+ /* Additional setups per revmap type */
+ switch (revmap_type) {
+ case IRQ_HOST_MAP_PRIORITY:
+ /* 0 is always the invalid number for priority */
+ host->inval_irq = 0;
+ /* setup us as the host for all priority interrupts */
+ for (i = 1; i < NR_PRIORITY_IRQS; i++) {
+ irq_map[i].hwirq = i;
+ smp_wmb();
+ irq_map[i].host = host;
+ smp_wmb();
+
+ ops->map(host, i, i);
+ }
+ break;
+ case IRQ_HOST_MAP_LINEAR:
+ rmap = (unsigned int *)(host + 1);
+ for (i = 0; i < revmap_arg; i++)
+ rmap[i] = NO_IRQ;
+ host->revmap_data.linear.size = revmap_arg;
+ smp_wmb();
+ host->revmap_data.linear.revmap = rmap;
+ break;
+ case IRQ_HOST_MAP_TREE:
+ INIT_RADIX_TREE(&host->revmap_data.tree, GFP_KERNEL);
+ break;
+ default:
+ break;
+ }
+
+ pr_debug("irq: Allocated host of type %d @0x%p\n", revmap_type, host);
+
+ return host;
+}
+
+struct irq_host *irq_find_host(struct device_node *node)
+{
+ struct irq_host *h, *found = NULL;
+ unsigned long flags;
+
+ /* We might want to match the legacy controller last since
+ * it might potentially be set to match all interrupts in
+ * the absence of a device node. This isn't a problem so far
+ * yet though...
+ */
+ raw_spin_lock_irqsave(&irq_big_lock, flags);
+ list_for_each_entry(h, &irq_hosts, link)
+ if (h->ops->match(h, node)) {
+ found = h;
+ break;
+ }
+ raw_spin_unlock_irqrestore(&irq_big_lock, flags);
+ return found;
+}
+EXPORT_SYMBOL_GPL(irq_find_host);
+
+void irq_set_default_host(struct irq_host *host)
+{
+ pr_debug("irq: Default host set to @0x%p\n", host);
+
+ irq_default_host = host;
+}
+
+void irq_set_virq_count(unsigned int count)
+{
+ pr_debug("irq: Trying to set virq count to %d\n", count);
+
+ BUG_ON(count < NR_PRIORITY_IRQS);
+ if (count < NR_IRQS)
+ irq_virq_count = count;
+}
+
+static int irq_setup_virq(struct irq_host *host, unsigned int virq,
+ irq_hw_number_t hwirq)
+{
+ int res;
+
+ res = irq_alloc_desc_at(virq, 0);
+ if (res != virq) {
+ pr_debug("irq: -> allocating desc failed\n");
+ goto error;
+ }
+
+ /* map it */
+ smp_wmb();
+ irq_map[virq].hwirq = hwirq;
+ smp_mb();
+
+ if (host->ops->map(host, virq, hwirq)) {
+ pr_debug("irq: -> mapping failed, freeing\n");
+ goto errdesc;
+ }
+
+ irq_clear_status_flags(virq, IRQ_NOREQUEST);
+
+ return 0;
+
+errdesc:
+ irq_free_descs(virq, 1);
+error:
+ irq_free_virt(virq, 1);
+ return -1;
+}
+
+unsigned int irq_create_direct_mapping(struct irq_host *host)
+{
+ unsigned int virq;
+
+ if (host == NULL)
+ host = irq_default_host;
+
+ BUG_ON(host == NULL);
+ WARN_ON(host->revmap_type != IRQ_HOST_MAP_NOMAP);
+
+ virq = irq_alloc_virt(host, 1, 0);
+ if (virq == NO_IRQ) {
+ pr_debug("irq: create_direct virq allocation failed\n");
+ return NO_IRQ;
+ }
+
+ pr_debug("irq: create_direct obtained virq %d\n", virq);
+
+ if (irq_setup_virq(host, virq, virq))
+ return NO_IRQ;
+
+ return virq;
+}
+
+unsigned int irq_create_mapping(struct irq_host *host,
+ irq_hw_number_t hwirq)
+{
+ unsigned int virq, hint;
+
+ pr_debug("irq: irq_create_mapping(0x%p, 0x%lx)\n", host, hwirq);
+
+ /* Look for default host if nececssary */
+ if (host == NULL)
+ host = irq_default_host;
+ if (host == NULL) {
+ printk(KERN_WARNING "irq_create_mapping called for"
+ " NULL host, hwirq=%lx\n", hwirq);
+ WARN_ON(1);
+ return NO_IRQ;
+ }
+ pr_debug("irq: -> using host @%p\n", host);
+
+ /* Check if mapping already exists */
+ virq = irq_find_mapping(host, hwirq);
+ if (virq != NO_IRQ) {
+ pr_debug("irq: -> existing mapping on virq %d\n", virq);
+ return virq;
+ }
+
+ /* Allocate a virtual interrupt number */
+ hint = hwirq % irq_virq_count;
+ virq = irq_alloc_virt(host, 1, hint);
+ if (virq == NO_IRQ) {
+ pr_debug("irq: -> virq allocation failed\n");
+ return NO_IRQ;
+ }
+
+ if (irq_setup_virq(host, virq, hwirq))
+ return NO_IRQ;
+
+ pr_debug("irq: irq %lu on host %s mapped to virtual irq %u\n",
+ hwirq, host->of_node ? host->of_node->full_name : "null", virq);
+
+ return virq;
+}
+EXPORT_SYMBOL_GPL(irq_create_mapping);
+
+unsigned int irq_create_of_mapping(struct device_node *controller,
+ const u32 *intspec, unsigned int intsize)
+{
+ struct irq_host *host;
+ irq_hw_number_t hwirq;
+ unsigned int type = IRQ_TYPE_NONE;
+ unsigned int virq;
+
+ if (controller == NULL)
+ host = irq_default_host;
+ else
+ host = irq_find_host(controller);
+ if (host == NULL) {
+ printk(KERN_WARNING "irq: no irq host found for %s !\n",
+ controller->full_name);
+ return NO_IRQ;
+ }
+
+ /* If host has no translation, then we assume interrupt line */
+ if (host->ops->xlate == NULL)
+ hwirq = intspec[0];
+ else {
+ if (host->ops->xlate(host, controller, intspec, intsize,
+ &hwirq, &type))
+ return NO_IRQ;
+ }
+
+ /* Create mapping */
+ virq = irq_create_mapping(host, hwirq);
+ if (virq == NO_IRQ)
+ return virq;
+
+ /* Set type if specified and different than the current one */
+ if (type != IRQ_TYPE_NONE &&
+ type != (irqd_get_trigger_type(irq_get_irq_data(virq))))
+ irq_set_irq_type(virq, type);
+ return virq;
+}
+EXPORT_SYMBOL_GPL(irq_create_of_mapping);
+
+void irq_dispose_mapping(unsigned int virq)
+{
+ struct irq_host *host;
+ irq_hw_number_t hwirq;
+
+ if (virq == NO_IRQ)
+ return;
+
+ /* Never unmap priority interrupts */
+ if (virq < NR_PRIORITY_IRQS)
+ return;
+
+ host = irq_map[virq].host;
+ if (WARN_ON(host == NULL))
+ return;
+
+ irq_set_status_flags(virq, IRQ_NOREQUEST);
+
+ /* remove chip and handler */
+ irq_set_chip_and_handler(virq, NULL, NULL);
+
+ /* Make sure it's completed */
+ synchronize_irq(virq);
+
+ /* Tell the PIC about it */
+ if (host->ops->unmap)
+ host->ops->unmap(host, virq);
+ smp_mb();
+
+ /* Clear reverse map */
+ hwirq = irq_map[virq].hwirq;
+ switch (host->revmap_type) {
+ case IRQ_HOST_MAP_LINEAR:
+ if (hwirq < host->revmap_data.linear.size)
+ host->revmap_data.linear.revmap[hwirq] = NO_IRQ;
+ break;
+ case IRQ_HOST_MAP_TREE:
+ mutex_lock(&revmap_trees_mutex);
+ radix_tree_delete(&host->revmap_data.tree, hwirq);
+ mutex_unlock(&revmap_trees_mutex);
+ break;
+ }
+
+ /* Destroy map */
+ smp_mb();
+ irq_map[virq].hwirq = host->inval_irq;
+
+ irq_free_descs(virq, 1);
+ /* Free it */
+ irq_free_virt(virq, 1);
+}
+EXPORT_SYMBOL_GPL(irq_dispose_mapping);
+
+unsigned int irq_find_mapping(struct irq_host *host,
+ irq_hw_number_t hwirq)
+{
+ unsigned int i;
+ unsigned int hint = hwirq % irq_virq_count;
+
+ /* Look for default host if nececssary */
+ if (host == NULL)
+ host = irq_default_host;
+ if (host == NULL)
+ return NO_IRQ;
+
+ /* Slow path does a linear search of the map */
+ i = hint;
+ do {
+ if (irq_map[i].host == host &&
+ irq_map[i].hwirq == hwirq)
+ return i;
+ i++;
+ if (i >= irq_virq_count)
+ i = 4;
+ } while (i != hint);
+ return NO_IRQ;
+}
+EXPORT_SYMBOL_GPL(irq_find_mapping);
+
+unsigned int irq_radix_revmap_lookup(struct irq_host *host,
+ irq_hw_number_t hwirq)
+{
+ struct irq_map_entry *ptr;
+ unsigned int virq;
+
+ if (WARN_ON_ONCE(host->revmap_type != IRQ_HOST_MAP_TREE))
+ return irq_find_mapping(host, hwirq);
+
+ /*
+ * The ptr returned references the static global irq_map.
+ * but freeing an irq can delete nodes along the path to
+ * do the lookup via call_rcu.
+ */
+ rcu_read_lock();
+ ptr = radix_tree_lookup(&host->revmap_data.tree, hwirq);
+ rcu_read_unlock();
+
+ /*
+ * If found in radix tree, then fine.
+ * Else fallback to linear lookup - this should not happen in practice
+ * as it means that we failed to insert the node in the radix tree.
+ */
+ if (ptr)
+ virq = ptr - irq_map;
+ else
+ virq = irq_find_mapping(host, hwirq);
+
+ return virq;
+}
+
+void irq_radix_revmap_insert(struct irq_host *host, unsigned int virq,
+ irq_hw_number_t hwirq)
+{
+ if (WARN_ON(host->revmap_type != IRQ_HOST_MAP_TREE))
+ return;
+
+ if (virq != NO_IRQ) {
+ mutex_lock(&revmap_trees_mutex);
+ radix_tree_insert(&host->revmap_data.tree, hwirq,
+ &irq_map[virq]);
+ mutex_unlock(&revmap_trees_mutex);
+ }
+}
+
+unsigned int irq_linear_revmap(struct irq_host *host,
+ irq_hw_number_t hwirq)
+{
+ unsigned int *revmap;
+
+ if (WARN_ON_ONCE(host->revmap_type != IRQ_HOST_MAP_LINEAR))
+ return irq_find_mapping(host, hwirq);
+
+ /* Check revmap bounds */
+ if (unlikely(hwirq >= host->revmap_data.linear.size))
+ return irq_find_mapping(host, hwirq);
+
+ /* Check if revmap was allocated */
+ revmap = host->revmap_data.linear.revmap;
+ if (unlikely(revmap == NULL))
+ return irq_find_mapping(host, hwirq);
+
+ /* Fill up revmap with slow path if no mapping found */
+ if (unlikely(revmap[hwirq] == NO_IRQ))
+ revmap[hwirq] = irq_find_mapping(host, hwirq);
+
+ return revmap[hwirq];
+}
+
+unsigned int irq_alloc_virt(struct irq_host *host,
+ unsigned int count,
+ unsigned int hint)
+{
+ unsigned long flags;
+ unsigned int i, j, found = NO_IRQ;
+
+ if (count == 0 || count > (irq_virq_count - NR_PRIORITY_IRQS))
+ return NO_IRQ;
+
+ raw_spin_lock_irqsave(&irq_big_lock, flags);
+
+ /* Use hint for 1 interrupt if any */
+ if (count == 1 && hint >= NR_PRIORITY_IRQS &&
+ hint < irq_virq_count && irq_map[hint].host == NULL) {
+ found = hint;
+ goto hint_found;
+ }
+
+ /* Look for count consecutive numbers in the allocatable
+ * (non-legacy) space
+ */
+ for (i = NR_PRIORITY_IRQS, j = 0; i < irq_virq_count; i++) {
+ if (irq_map[i].host != NULL)
+ j = 0;
+ else
+ j++;
+
+ if (j == count) {
+ found = i - count + 1;
+ break;
+ }
+ }
+ if (found == NO_IRQ) {
+ raw_spin_unlock_irqrestore(&irq_big_lock, flags);
+ return NO_IRQ;
+ }
+ hint_found:
+ for (i = found; i < (found + count); i++) {
+ irq_map[i].hwirq = host->inval_irq;
+ smp_wmb();
+ irq_map[i].host = host;
+ }
+ raw_spin_unlock_irqrestore(&irq_big_lock, flags);
+ return found;
+}
+
+void irq_free_virt(unsigned int virq, unsigned int count)
+{
+ unsigned long flags;
+ unsigned int i;
+
+ WARN_ON(virq < NR_PRIORITY_IRQS);
+ WARN_ON(count == 0 || (virq + count) > irq_virq_count);
+
+ if (virq < NR_PRIORITY_IRQS) {
+ if (virq + count < NR_PRIORITY_IRQS)
+ return;
+ count -= NR_PRIORITY_IRQS - virq;
+ virq = NR_PRIORITY_IRQS;
+ }
+
+ if (count > irq_virq_count || virq > irq_virq_count - count) {
+ if (virq > irq_virq_count)
+ return;
+ count = irq_virq_count - virq;
+ }
+
+ raw_spin_lock_irqsave(&irq_big_lock, flags);
+ for (i = virq; i < (virq + count); i++) {
+ struct irq_host *host;
+
+ host = irq_map[i].host;
+ irq_map[i].hwirq = host->inval_irq;
+ smp_wmb();
+ irq_map[i].host = NULL;
+ }
+ raw_spin_unlock_irqrestore(&irq_big_lock, flags);
+}
+
+#ifdef CONFIG_VIRQ_DEBUG
+static int virq_debug_show(struct seq_file *m, void *private)
+{
+ unsigned long flags;
+ struct irq_desc *desc;
+ const char *p;
+ static const char none[] = "none";
+ void *data;
+ int i;
+
+ seq_printf(m, "%-5s %-7s %-15s %-18s %s\n", "virq", "hwirq",
+ "chip name", "chip data", "host name");
+
+ for (i = 1; i < nr_irqs; i++) {
+ desc = irq_to_desc(i);
+ if (!desc)
+ continue;
+
+ raw_spin_lock_irqsave(&desc->lock, flags);
+
+ if (desc->action && desc->action->handler) {
+ struct irq_chip *chip;
+
+ seq_printf(m, "%5d ", i);
+ seq_printf(m, "0x%05lx ", irq_map[i].hwirq);
+
+ chip = irq_desc_get_chip(desc);
+ if (chip && chip->name)
+ p = chip->name;
+ else
+ p = none;
+ seq_printf(m, "%-15s ", p);
+
+ data = irq_desc_get_chip_data(desc);
+ seq_printf(m, "0x%16p ", data);
+
+ if (irq_map[i].host && irq_map[i].host->of_node)
+ p = irq_map[i].host->of_node->full_name;
+ else
+ p = none;
+ seq_printf(m, "%s\n", p);
+ }
+
+ raw_spin_unlock_irqrestore(&desc->lock, flags);
+ }
+
+ return 0;
+}
+
+static int virq_debug_open(struct inode *inode, struct file *file)
+{
+ return single_open(file, virq_debug_show, inode->i_private);
+}
+
+static const struct file_operations virq_debug_fops = {
+ .open = virq_debug_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+ .release = single_release,
+};
+
+static int __init irq_debugfs_init(void)
+{
+ if (debugfs_create_file("virq_mapping", S_IRUGO, powerpc_debugfs_root,
+ NULL, &virq_debug_fops) == NULL)
+ return -ENOMEM;
+
+ return 0;
+}
+device_initcall(irq_debugfs_init);
+#endif /* CONFIG_VIRQ_DEBUG */
diff --git a/arch/c6x/platforms/megamod-pic.c b/arch/c6x/platforms/megamod-pic.c
new file mode 100644
index 0000000..7c37a94
--- /dev/null
+++ b/arch/c6x/platforms/megamod-pic.c
@@ -0,0 +1,349 @@
+/*
+ * Support for C64x+ Megamodule Interrupt Controller
+ *
+ * Copyright (C) 2010, 2011 Texas Instruments Incorporated
+ * Contributed by: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/module.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/of.h>
+#include <linux/of_irq.h>
+#include <linux/of_address.h>
+#include <linux/slab.h>
+#include <asm/soc.h>
+#include <asm/megamod-pic.h>
+
+#define NR_COMBINERS 4
+#define NR_MUX_OUTPUTS 12
+
+#define IRQ_UNMAPPED 0xffff
+
+/*
+ * Megamodule Interrupt Controller register layout
+ */
+struct megamod_regs {
+ u32 evtflag[8];
+ u32 evtset[8];
+ u32 evtclr[8];
+ u32 reserved0[8];
+ u32 evtmask[8];
+ u32 mevtflag[8];
+ u32 expmask[8];
+ u32 mexpflag[8];
+ u32 intmux_unused;
+ u32 intmux[7];
+ u32 reserved1[8];
+ u32 aegmux[2];
+ u32 reserved2[14];
+ u32 intxstat;
+ u32 intxclr;
+ u32 intdmask;
+ u32 reserved3[13];
+ u32 evtasrt;
+};
+
+struct megamod_pic {
+ struct irq_host *irqhost;
+ struct megamod_regs __iomem *regs;
+ raw_spinlock_t lock;
+
+ /* hw mux mapping */
+ unsigned int output_to_irq[NR_MUX_OUTPUTS];
+};
+
+static struct megamod_pic *mm_pic;
+
+struct megamod_cascade_data {
+ struct megamod_pic *pic;
+ int index;
+};
+
+static struct megamod_cascade_data cascade_data[NR_COMBINERS];
+
+static void mask_megamod(struct irq_data *data)
+{
+ struct megamod_pic *pic = irq_data_get_irq_chip_data(data);
+ irq_hw_number_t src = irqd_to_hwirq(data);
+ u32 __iomem *evtmask = &pic->regs->evtmask[src / 32];
+
+ raw_spin_lock(&pic->lock);
+ soc_writel(soc_readl(evtmask) | (1 << (src & 31)), evtmask);
+ raw_spin_unlock(&pic->lock);
+}
+
+static void unmask_megamod(struct irq_data *data)
+{
+ struct megamod_pic *pic = irq_data_get_irq_chip_data(data);
+ irq_hw_number_t src = irqd_to_hwirq(data);
+ u32 __iomem *evtmask = &pic->regs->evtmask[src / 32];
+
+ raw_spin_lock(&pic->lock);
+ soc_writel(soc_readl(evtmask) & ~(1 << (src & 31)), evtmask);
+ raw_spin_unlock(&pic->lock);
+}
+
+static struct irq_chip megamod_chip = {
+ .name = "megamod",
+ .irq_mask = mask_megamod,
+ .irq_unmask = unmask_megamod,
+};
+
+static void megamod_irq_cascade(unsigned int irq, struct irq_desc *desc)
+{
+ struct megamod_cascade_data *cascade;
+ struct megamod_pic *pic;
+ u32 events;
+ int n, idx;
+
+ cascade = irq_desc_get_handler_data(desc);
+
+ pic = cascade->pic;
+ idx = cascade->index;
+
+ while ((events = soc_readl(&pic->regs->mevtflag[idx])) != 0) {
+ n = __ffs(events);
+
+ irq = irq_linear_revmap(pic->irqhost, idx * 32 + n);
+
+ soc_writel(1 << n, &pic->regs->evtclr[idx]);
+
+ generic_handle_irq(irq);
+ }
+}
+
+static int megamod_map(struct irq_host *h, unsigned int virq,
+ irq_hw_number_t hw)
+{
+ struct megamod_pic *pic = h->host_data;
+ int i;
+
+ /* We shouldn't see a hwirq which is muxed to core controller */
+ for (i = 0; i < NR_MUX_OUTPUTS; i++)
+ if (pic->output_to_irq[i] == hw)
+ return -1;
+
+ irq_set_chip_data(virq, pic);
+ irq_set_chip_and_handler(virq, &megamod_chip, handle_level_irq);
+
+ /* Set default irq type */
+ irq_set_irq_type(virq, IRQ_TYPE_NONE);
+
+ return 0;
+}
+
+static int megamod_xlate(struct irq_host *h, struct device_node *ct,
+ const u32 *intspec, unsigned int intsize,
+ irq_hw_number_t *out_hwirq, unsigned int *out_type)
+
+{
+ /* megamod intspecs must have 1 cell */
+ BUG_ON(intsize != 1);
+ *out_hwirq = intspec[0];
+ *out_type = IRQ_TYPE_NONE;
+ return 0;
+}
+
+static struct irq_host_ops megamod_host_ops = {
+ .map = megamod_map,
+ .xlate = megamod_xlate,
+};
+
+static void __init set_megamod_mux(struct megamod_pic *pic, int src, int output)
+{
+ int index, offset;
+ u32 val;
+
+ if (src < 0 || src >= (NR_COMBINERS * 32)) {
+ pic->output_to_irq[output] = IRQ_UNMAPPED;
+ return;
+ }
+
+ /* four mappings per mux register */
+ index = output / 4;
+ offset = (output & 3) * 8;
+
+ val = soc_readl(&pic->regs->intmux[index]);
+ val &= ~(0xff << offset);
+ val |= src << offset;
+ soc_writel(val, &pic->regs->intmux[index]);
+}
+
+/*
+ * Parse the MUX mapping, if one exists.
+ *
+ * The MUX map is an array of up to 12 cells; one for each usable core priority
+ * interrupt. The value of a given cell is the megamodule interrupt source
+ * which is to me MUXed to the output corresponding to the cell position
+ * withing the array. The first cell in the array corresponds to priority
+ * 4 and the last (12th) cell corresponds to priority 15. The allowed
+ * values are 4 - ((NR_COMBINERS * 32) - 1). Note that the combined interrupt
+ * sources (0 - 3) are not allowed to be mapped through this property. They
+ * are handled through the "interrupts" property. This allows us to use a
+ * value of zero as a "do not map" placeholder.
+ */
+static void __init parse_priority_map(struct megamod_pic *pic,
+ int *mapping, int size)
+{
+ struct device_node *np = pic->irqhost->of_node;
+ const __be32 *map;
+ int i, maplen;
+ u32 val;
+
+ map = of_get_property(np, "ti,c64x+megamod-pic-mux", &maplen);
+ if (map) {
+ maplen /= 4;
+ if (maplen > size)
+ maplen = size;
+
+ for (i = 0; i < maplen; i++) {
+ val = be32_to_cpup(map);
+ if (val && val >= 4)
+ mapping[i] = val;
+ ++map;
+ }
+ }
+}
+
+static struct megamod_pic * __init init_megamod_pic(struct device_node *np)
+{
+ struct megamod_pic *pic;
+ int i, irq;
+ int mapping[NR_MUX_OUTPUTS];
+
+ pr_info("Initializing C64x+ Megamodule PIC\n");
+
+ pic = kzalloc(sizeof(struct megamod_pic), GFP_KERNEL);
+ if (!pic) {
+ pr_err("%s: Could not alloc PIC structure.\n", np->full_name);
+ return NULL;
+ }
+
+ pic->irqhost = irq_alloc_host(np, IRQ_HOST_MAP_LINEAR,
+ NR_COMBINERS * 32, &megamod_host_ops,
+ IRQ_UNMAPPED);
+ if (!pic->irqhost) {
+ pr_err("%s: Could not alloc host.\n", np->full_name);
+ goto error_free;
+ }
+
+ pic->irqhost->host_data = pic;
+
+ raw_spin_lock_init(&pic->lock);
+
+ pic->regs = of_iomap(np, 0);
+ if (!pic->regs) {
+ pr_err("%s: Could not map registers.\n", np->full_name);
+ goto error_free;
+ }
+
+ /* Initialize MUX map */
+ for (i = 0; i < ARRAY_SIZE(mapping); i++)
+ mapping[i] = IRQ_UNMAPPED;
+
+ parse_priority_map(pic, mapping, ARRAY_SIZE(mapping));
+
+ /*
+ * We can have up to 12 interrupts cascading to the core controller.
+ * These cascades can be from the combined interrupt sources or for
+ * individual interrupt sources. The "interrupts" property only
+ * deals with the cascaded combined interrupts. The individual
+ * interrupts muxed to the core controller use the core controller
+ * as their interrupt parent.
+ */
+ for (i = 0; i < NR_COMBINERS; i++) {
+
+ irq = irq_of_parse_and_map(np, i);
+ if (irq == NO_IRQ)
+ continue;
+
+ /*
+ * We count on the core priority interrupts (4 - 15) being
+ * direct mapped. Check that device tree provided something
+ * in that range.
+ */
+ if (irq < 4 || irq >= NR_PRIORITY_IRQS) {
+ pr_err("%s: combiner-%d virq %d out of range!\n",
+ np->full_name, i, irq);
+ continue;
+ }
+
+ /* record the mapping */
+ mapping[irq - 4] = i;
+
+ pr_debug("%s: combiner-%d cascading to virq %d\n",
+ np->full_name, i, irq);
+
+ cascade_data[i].pic = pic;
+ cascade_data[i].index = i;
+
+ /* mask and clear all events in combiner */
+ soc_writel(~0, &pic->regs->evtmask[i]);
+ soc_writel(~0, &pic->regs->evtclr[i]);
+
+ irq_set_handler_data(irq, &cascade_data[i]);
+ irq_set_chained_handler(irq, megamod_irq_cascade);
+ }
+
+ /* Finally, set up the MUX registers */
+ for (i = 0; i < NR_MUX_OUTPUTS; i++) {
+ if (mapping[i] != IRQ_UNMAPPED) {
+ pr_debug("%s: setting mux %d to priority %d\n",
+ np->full_name, mapping[i], i + 4);
+ set_megamod_mux(pic, mapping[i], i);
+ }
+ }
+
+ return pic;
+
+error_free:
+ kfree(pic);
+
+ return NULL;
+}
+
+/*
+ * Return next active event after ACK'ing it.
+ * Return -1 if no events active.
+ */
+static int get_exception(void)
+{
+ int i, bit;
+ u32 mask;
+
+ for (i = 0; i < NR_COMBINERS; i++) {
+ mask = soc_readl(&mm_pic->regs->mexpflag[i]);
+ if (mask) {
+ bit = __ffs(mask);
+ soc_writel(1 << bit, &mm_pic->regs->evtclr[i]);
+ return (i * 32) + bit;
+ }
+ }
+ return -1;
+}
+
+static void assert_event(unsigned int val)
+{
+ soc_writel(val, &mm_pic->regs->evtasrt);
+}
+
+void __init megamod_pic_init(void)
+{
+ struct device_node *np;
+
+ np = of_find_compatible_node(NULL, NULL, "ti,c64x+megamod-pic");
+ if (!np)
+ return;
+
+ mm_pic = init_megamod_pic(np);
+ of_node_put(np);
+
+ soc_ops.get_exception = get_exception;
+ soc_ops.assert_event = assert_event;
+
+ return;
+}
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v3 12/24] C6X: interrupt handling
2011-09-27 20:29 ` [PATCH v3 12/24] C6X: interrupt handling Mark Salter
@ 2011-09-27 23:30 ` Thomas Gleixner
0 siblings, 0 replies; 66+ messages in thread
From: Thomas Gleixner @ 2011-09-27 23:30 UTC (permalink / raw)
To: Mark Salter; +Cc: LKML, linux-arch, Grant Likely
B1;2601;0cOn Tue, 27 Sep 2011, Mark Salter wrote:
> diff --git a/arch/c6x/kernel/irq.c b/arch/c6x/kernel/irq.c
> +static raw_spinlock_t core_irq_lock;
static DEFINE_RAW_SPINLOCK(...) please, so you can get rid of the
raw_spin_lock_init(...) below.
> +static int core_host_map(struct irq_host *h, unsigned int virq,
> + irq_hw_number_t hw)
> +{
> + if (hw < 4 || hw >= NR_PRIORITY_IRQS)
> + return -1;
Proper error code please, i.e -EINVAL or such
> +
> + irq_set_status_flags(virq, IRQ_LEVEL);
> + irq_set_chip_and_handler(virq, &core_chip, handle_level_irq);
> + return 0;
> +}
> +
> +static struct irq_host_ops core_host_ops = {
> + .map = core_host_map,
> +};
> +
> +void __init init_IRQ(void)
> +{
> + struct device_node *np;
> +
> + raw_spin_lock_init(&core_irq_lock);
See above.
Aside of that I'm still not too happy about the of_/irqhost/domain
code replication and the cleanup which needs to be done after that has
been solved. Though if you and Grant have a plan for that, I'm not in
the way.
Btw, keeping people who have been added to previous review comments of
a particular patch in the CC makes everyones life easier.
Thanks,
tglx
^ permalink raw reply [flat|nested] 66+ messages in thread
* [PATCH v3 13/24] C6X: syscalls
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (11 preceding siblings ...)
2011-09-27 20:29 ` [PATCH v3 12/24] C6X: interrupt handling Mark Salter
@ 2011-09-27 20:29 ` Mark Salter
2011-09-28 13:49 ` Arnd Bergmann
2011-09-27 20:29 ` [PATCH v3 14/24] C6X: traps Mark Salter
` (12 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:29 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
Signed-off-by: Mark Salter <msalter@redhat.com>
---
arch/c6x/include/asm/syscalls.h | 55 +++++++++++++++++++++++++++++
arch/c6x/include/asm/unistd.h | 26 ++++++++++++++
arch/c6x/kernel/sys_c6x.c | 73 +++++++++++++++++++++++++++++++++++++++
3 files changed, 154 insertions(+), 0 deletions(-)
create mode 100644 arch/c6x/include/asm/syscalls.h
create mode 100644 arch/c6x/include/asm/unistd.h
create mode 100644 arch/c6x/kernel/sys_c6x.c
diff --git a/arch/c6x/include/asm/syscalls.h b/arch/c6x/include/asm/syscalls.h
new file mode 100644
index 0000000..aed53da
--- /dev/null
+++ b/arch/c6x/include/asm/syscalls.h
@@ -0,0 +1,55 @@
+/*
+ * Copyright (C) 2011 Texas Instruments Incorporated
+ * Author: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation, version 2.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or
+ * NON INFRINGEMENT. See the GNU General Public License for
+ * more details.
+ */
+
+#ifndef __ASM_C6X_SYSCALLS_H
+#define __ASM_C6X_SYSCALLS_H
+
+#include <linux/compiler.h>
+#include <linux/linkage.h>
+#include <linux/types.h>
+
+/* The array of function pointers for syscalls. */
+extern void *sys_call_table[];
+
+/* The following are trampolines in entry.S to handle 64-bit arguments */
+extern long sys_pread_c6x(unsigned int fd, char __user *buf,
+ size_t count, off_t pos_low, off_t pos_high);
+extern long sys_pwrite_c6x(unsigned int fd, const char __user *buf,
+ size_t count, off_t pos_low, off_t pos_high);
+extern long sys_truncate64_c6x(const char __user *path,
+ off_t length_low, off_t length_high);
+extern long sys_ftruncate64_c6x(unsigned int fd,
+ off_t length_low, off_t length_high);
+extern long sys_fadvise64_c6x(int fd, u32 offset_lo, u32 offset_hi,
+ u32 len, int advice);
+extern long sys_fadvise64_64_c6x(int fd, u32 offset_lo, u32 offset_hi,
+ u32 len_lo, u32 len_hi, int advice);
+extern long sys_fallocate_c6x(int fd, int mode,
+ u32 offset_lo, u32 offset_hi,
+ u32 len_lo, u32 len_hi);
+extern int sys_cache_sync(unsigned long s, unsigned long e);
+
+struct pt_regs;
+
+extern asmlinkage long sys_c6x_clone(struct pt_regs *regs);
+extern asmlinkage long sys_c6x_execve(const char __user *name,
+ const char __user *const __user *argv,
+ const char __user *const __user *envp,
+ struct pt_regs *regs);
+
+
+#include <asm-generic/syscalls.h>
+
+#endif /* __ASM_C6X_SYSCALLS_H */
diff --git a/arch/c6x/include/asm/unistd.h b/arch/c6x/include/asm/unistd.h
new file mode 100644
index 0000000..6d54ea4
--- /dev/null
+++ b/arch/c6x/include/asm/unistd.h
@@ -0,0 +1,26 @@
+/*
+ * Copyright (C) 2011 Texas Instruments Incorporated
+ *
+ * Based on arch/tile version.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation, version 2.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or
+ * NON INFRINGEMENT. See the GNU General Public License for
+ * more details.
+ */
+#if !defined(_ASM_C6X_UNISTD_H) || defined(__SYSCALL)
+#define _ASM_C6X_UNISTD_H
+
+/* Use the standard ABI for syscalls. */
+#include <asm-generic/unistd.h>
+
+/* C6X-specific syscalls. */
+#define __NR_cache_sync (__NR_arch_specific_syscall + 0)
+__SYSCALL(__NR_cache_sync, sys_cache_sync)
+
+#endif /* _ASM_C6X_UNISTD_H */
diff --git a/arch/c6x/kernel/sys_c6x.c b/arch/c6x/kernel/sys_c6x.c
new file mode 100644
index 0000000..f597d26
--- /dev/null
+++ b/arch/c6x/kernel/sys_c6x.c
@@ -0,0 +1,73 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010, 2011 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/syscalls.h>
+#include <linux/uaccess.h>
+
+#include <asm/syscalls.h>
+
+#ifdef CONFIG_ACCESS_CHECK
+int _access_ok(unsigned long addr, unsigned long size)
+{
+ if (!size)
+ return 1;
+
+ if (!addr || addr > (0xffffffffUL - (size - 1)))
+ goto _bad_access;
+
+ if (segment_eq(get_fs(), KERNEL_DS))
+ return 1;
+
+ if (memory_start <= addr && (addr + size - 1) < memory_end)
+ return 1;
+
+_bad_access:
+ pr_debug("Bad access attempt: pid[%d] addr[%08lx] size[0x%lx]\n",
+ current->pid, addr, size);
+ return 0;
+}
+EXPORT_SYMBOL(_access_ok);
+#endif
+
+/* sys_cache_sync -- sync caches over given range */
+asmlinkage int sys_cache_sync(unsigned long s, unsigned long e)
+{
+ L1D_cache_block_writeback_invalidate(s, e);
+ L1P_cache_block_invalidate(s, e);
+
+ return 0;
+}
+
+/* Provide the actual syscall number to call mapping. */
+#undef __SYSCALL
+#define __SYSCALL(nr, call) [nr] = (call),
+
+/*
+ * Use trampolines
+ */
+#define sys_pread64 sys_pread_c6x
+#define sys_pwrite64 sys_pwrite_c6x
+#define sys_truncate64 sys_truncate64_c6x
+#define sys_ftruncate64 sys_ftruncate64_c6x
+#define sys_fadvise64 sys_fadvise64_c6x
+#define sys_fadvise64_64 sys_fadvise64_64_c6x
+#define sys_fallocate sys_fallocate_c6x
+
+/* Use sys_mmap_pgoff directly */
+#define sys_mmap2 sys_mmap_pgoff
+
+/*
+ * Note that we can't include <linux/unistd.h> here since the header
+ * guard will defeat us; <asm/unistd.h> checks for __SYSCALL as well.
+ */
+void *sys_call_table[__NR_syscalls] = {
+ [0 ... __NR_syscalls-1] = sys_ni_syscall,
+#include <asm/unistd.h>
+};
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v3 13/24] C6X: syscalls
2011-09-27 20:29 ` [PATCH v3 13/24] C6X: syscalls Mark Salter
@ 2011-09-28 13:49 ` Arnd Bergmann
0 siblings, 0 replies; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-28 13:49 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch
On Tuesday 27 September 2011, Mark Salter wrote:
> Signed-off-by: Mark Salter <msalter@redhat.com>
> ---
> arch/c6x/include/asm/syscalls.h | 55 +++++++++++++++++++++++++++++
> arch/c6x/include/asm/unistd.h | 26 ++++++++++++++
> arch/c6x/kernel/sys_c6x.c | 73 +++++++++++++++++++++++++++++++++++++++
> 3 files changed, 154 insertions(+), 0 deletions(-)
> create mode 100644 arch/c6x/include/asm/syscalls.h
> create mode 100644 arch/c6x/include/asm/unistd.h
> create mode 100644 arch/c6x/kernel/sys_c6x.c
Acked-by: Arnd Bergmann <arnd@arndb.de>
^ permalink raw reply [flat|nested] 66+ messages in thread
* [PATCH v3 14/24] C6X: traps
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (12 preceding siblings ...)
2011-09-27 20:29 ` [PATCH v3 13/24] C6X: syscalls Mark Salter
@ 2011-09-27 20:29 ` Mark Salter
2011-09-28 13:50 ` Arnd Bergmann
2011-09-27 20:29 ` [PATCH v3 15/24] C6X: clocks Mark Salter
` (11 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:29 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
This patch provides the kernel entry points for traps, syscalls and
interrupts.
Signed-off-by: Mark Salter <msalter@redhat.com>
---
arch/c6x/include/asm/traps.h | 36 ++
arch/c6x/kernel/asm-offsets.c | 123 +++++++
arch/c6x/kernel/entry.S | 803 +++++++++++++++++++++++++++++++++++++++++
arch/c6x/kernel/traps.c | 423 ++++++++++++++++++++++
4 files changed, 1385 insertions(+), 0 deletions(-)
create mode 100644 arch/c6x/include/asm/traps.h
create mode 100644 arch/c6x/kernel/asm-offsets.c
create mode 100644 arch/c6x/kernel/entry.S
create mode 100644 arch/c6x/kernel/traps.c
diff --git a/arch/c6x/include/asm/traps.h b/arch/c6x/include/asm/traps.h
new file mode 100644
index 0000000..62124d7
--- /dev/null
+++ b/arch/c6x/include/asm/traps.h
@@ -0,0 +1,36 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2011 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_TRAPS_H
+#define _ASM_C6X_TRAPS_H
+
+#define EXCEPT_TYPE_NXF 31 /* NMI */
+#define EXCEPT_TYPE_EXC 30 /* external exception */
+#define EXCEPT_TYPE_IXF 1 /* internal exception */
+#define EXCEPT_TYPE_SXF 0 /* software exception */
+
+#define EXCEPT_CAUSE_LBX (1 << 7) /* loop buffer exception */
+#define EXCEPT_CAUSE_PRX (1 << 6) /* privilege exception */
+#define EXCEPT_CAUSE_RAX (1 << 5) /* resource access exception */
+#define EXCEPT_CAUSE_RCX (1 << 4) /* resource conflict exception */
+#define EXCEPT_CAUSE_OPX (1 << 3) /* opcode exception */
+#define EXCEPT_CAUSE_EPX (1 << 2) /* execute packet exception */
+#define EXCEPT_CAUSE_FPX (1 << 1) /* fetch packet exception */
+#define EXCEPT_CAUSE_IFX (1 << 0) /* instruction fetch exception */
+
+struct exception_info {
+ char *kernel_str;
+ int signo;
+ int code;
+};
+
+extern int (*c6x_nmi_handler)(struct pt_regs *regs);
+
+#endif /* _ASM_C6X_TRAPS_H */
diff --git a/arch/c6x/kernel/asm-offsets.c b/arch/c6x/kernel/asm-offsets.c
new file mode 100644
index 0000000..759ad6d
--- /dev/null
+++ b/arch/c6x/kernel/asm-offsets.c
@@ -0,0 +1,123 @@
+/*
+ * Generate definitions needed by assembly language modules.
+ * This code generates raw asm output which is post-processed
+ * to extract and format the required data.
+ */
+
+#include <linux/sched.h>
+#include <linux/thread_info.h>
+#include <asm/procinfo.h>
+#include <linux/kbuild.h>
+#include <linux/unistd.h>
+
+void foo(void)
+{
+ OFFSET(REGS_A16, pt_regs, a16);
+ OFFSET(REGS_A17, pt_regs, a17);
+ OFFSET(REGS_A18, pt_regs, a18);
+ OFFSET(REGS_A19, pt_regs, a19);
+ OFFSET(REGS_A20, pt_regs, a20);
+ OFFSET(REGS_A21, pt_regs, a21);
+ OFFSET(REGS_A22, pt_regs, a22);
+ OFFSET(REGS_A23, pt_regs, a23);
+ OFFSET(REGS_A24, pt_regs, a24);
+ OFFSET(REGS_A25, pt_regs, a25);
+ OFFSET(REGS_A26, pt_regs, a26);
+ OFFSET(REGS_A27, pt_regs, a27);
+ OFFSET(REGS_A28, pt_regs, a28);
+ OFFSET(REGS_A29, pt_regs, a29);
+ OFFSET(REGS_A30, pt_regs, a30);
+ OFFSET(REGS_A31, pt_regs, a31);
+
+ OFFSET(REGS_B16, pt_regs, b16);
+ OFFSET(REGS_B17, pt_regs, b17);
+ OFFSET(REGS_B18, pt_regs, b18);
+ OFFSET(REGS_B19, pt_regs, b19);
+ OFFSET(REGS_B20, pt_regs, b20);
+ OFFSET(REGS_B21, pt_regs, b21);
+ OFFSET(REGS_B22, pt_regs, b22);
+ OFFSET(REGS_B23, pt_regs, b23);
+ OFFSET(REGS_B24, pt_regs, b24);
+ OFFSET(REGS_B25, pt_regs, b25);
+ OFFSET(REGS_B26, pt_regs, b26);
+ OFFSET(REGS_B27, pt_regs, b27);
+ OFFSET(REGS_B28, pt_regs, b28);
+ OFFSET(REGS_B29, pt_regs, b29);
+ OFFSET(REGS_B30, pt_regs, b30);
+ OFFSET(REGS_B31, pt_regs, b31);
+
+ OFFSET(REGS_A0, pt_regs, a0);
+ OFFSET(REGS_A1, pt_regs, a1);
+ OFFSET(REGS_A2, pt_regs, a2);
+ OFFSET(REGS_A3, pt_regs, a3);
+ OFFSET(REGS_A4, pt_regs, a4);
+ OFFSET(REGS_A5, pt_regs, a5);
+ OFFSET(REGS_A6, pt_regs, a6);
+ OFFSET(REGS_A7, pt_regs, a7);
+ OFFSET(REGS_A8, pt_regs, a8);
+ OFFSET(REGS_A9, pt_regs, a9);
+ OFFSET(REGS_A10, pt_regs, a10);
+ OFFSET(REGS_A11, pt_regs, a11);
+ OFFSET(REGS_A12, pt_regs, a12);
+ OFFSET(REGS_A13, pt_regs, a13);
+ OFFSET(REGS_A14, pt_regs, a14);
+ OFFSET(REGS_A15, pt_regs, a15);
+
+ OFFSET(REGS_B0, pt_regs, b0);
+ OFFSET(REGS_B1, pt_regs, b1);
+ OFFSET(REGS_B2, pt_regs, b2);
+ OFFSET(REGS_B3, pt_regs, b3);
+ OFFSET(REGS_B4, pt_regs, b4);
+ OFFSET(REGS_B5, pt_regs, b5);
+ OFFSET(REGS_B6, pt_regs, b6);
+ OFFSET(REGS_B7, pt_regs, b7);
+ OFFSET(REGS_B8, pt_regs, b8);
+ OFFSET(REGS_B9, pt_regs, b9);
+ OFFSET(REGS_B10, pt_regs, b10);
+ OFFSET(REGS_B11, pt_regs, b11);
+ OFFSET(REGS_B12, pt_regs, b12);
+ OFFSET(REGS_B13, pt_regs, b13);
+ OFFSET(REGS_DP, pt_regs, dp);
+ OFFSET(REGS_SP, pt_regs, sp);
+
+ OFFSET(REGS_TSR, pt_regs, tsr);
+ OFFSET(REGS_ORIG_A4, pt_regs, orig_a4);
+
+ DEFINE(REGS__END, sizeof(struct pt_regs));
+ BLANK();
+
+ OFFSET(THREAD_PC, thread_struct, pc);
+ OFFSET(THREAD_B15_14, thread_struct, b15_14);
+ OFFSET(THREAD_A15_14, thread_struct, a15_14);
+ OFFSET(THREAD_B13_12, thread_struct, b13_12);
+ OFFSET(THREAD_A13_12, thread_struct, a13_12);
+ OFFSET(THREAD_B11_10, thread_struct, b11_10);
+ OFFSET(THREAD_A11_10, thread_struct, a11_10);
+ OFFSET(THREAD_RICL_ICL, thread_struct, ricl_icl);
+ BLANK();
+
+ OFFSET(TASK_STATE, task_struct, state);
+ BLANK();
+
+ OFFSET(THREAD_INFO_FLAGS, thread_info, flags);
+ OFFSET(THREAD_INFO_PREEMPT_COUNT, thread_info, preempt_count);
+ BLANK();
+
+ /* These would be unneccessary if we ran asm files
+ * through the preprocessor.
+ */
+ DEFINE(KTHREAD_SIZE, THREAD_SIZE);
+ DEFINE(KTHREAD_SHIFT, THREAD_SHIFT);
+ DEFINE(KTHREAD_START_SP, THREAD_START_SP);
+ DEFINE(ENOSYS_, ENOSYS);
+ DEFINE(NR_SYSCALLS_, __NR_syscalls);
+
+ DEFINE(_TIF_SYSCALL_TRACE, (1<<TIF_SYSCALL_TRACE));
+ DEFINE(_TIF_NOTIFY_RESUME, (1<<TIF_NOTIFY_RESUME));
+ DEFINE(_TIF_SIGPENDING, (1<<TIF_SIGPENDING));
+ DEFINE(_TIF_NEED_RESCHED, (1<<TIF_NEED_RESCHED));
+ DEFINE(_TIF_POLLING_NRFLAG, (1<<TIF_POLLING_NRFLAG));
+
+ DEFINE(_TIF_ALLWORK_MASK, TIF_ALLWORK_MASK);
+ DEFINE(_TIF_WORK_MASK, TIF_WORK_MASK);
+}
diff --git a/arch/c6x/kernel/entry.S b/arch/c6x/kernel/entry.S
new file mode 100644
index 0000000..3e977cc
--- /dev/null
+++ b/arch/c6x/kernel/entry.S
@@ -0,0 +1,803 @@
+;
+; Port on Texas Instruments TMS320C6x architecture
+;
+; Copyright (C) 2004-2011 Texas Instruments Incorporated
+; Author: Aurelien Jacquiot (aurelien.jacquiot@virtuallogix.com)
+; Updated for 2.6.34: Mark Salter <msalter@redhat.com>
+;
+; This program is free software; you can redistribute it and/or modify
+; it under the terms of the GNU General Public License version 2 as
+; published by the Free Software Foundation.
+;
+
+#include <linux/sys.h>
+#include <linux/linkage.h>
+#include <asm/thread_info.h>
+#include <asm/asm-offsets.h>
+#include <asm/unistd.h>
+#include <asm/errno.h>
+
+; Registers naming
+#define DP B14
+#define SP B15
+
+#ifndef CONFIG_PREEMPT
+#define resume_kernel restore_all
+#endif
+
+ .altmacro
+
+ .macro MASK_INT reg
+ MVC .S2 CSR,reg
+ CLR .S2 reg,0,0,reg
+ MVC .S2 reg,CSR
+ .endm
+
+ .macro UNMASK_INT reg
+ MVC .S2 CSR,reg
+ SET .S2 reg,0,0,reg
+ MVC .S2 reg,CSR
+ .endm
+
+ .macro GET_THREAD_INFO reg
+ SHR .S1X SP,THREAD_SHIFT,reg
+ SHL .S1 reg,THREAD_SHIFT,reg
+ .endm
+
+ ;;
+ ;; This defines the normal kernel pt_regs layout.
+ ;;
+ .macro SAVE_ALL __rp __tsr
+ STW .D2T2 B0,*SP--[2] ; save original B0
+ MVKL .S2 current_ksp,B0
+ MVKH .S2 current_ksp,B0
+ LDW .D2T2 *B0,B1 ; KSP
+
+ NOP 3
+ STW .D2T2 B1,*+SP[1] ; save original B1
+ XOR .D2 SP,B1,B0 ; (SP ^ KSP)
+ LDW .D2T2 *+SP[1],B1 ; restore B0/B1
+ LDW .D2T2 *++SP[2],B0
+ SHR .S2 B0,THREAD_SHIFT,B0 ; 0 if already using kstack
+ [B0] STDW .D2T2 SP:DP,*--B1[1] ; user: save user sp/dp kstack
+ [B0] MV .S2 B1,SP ; and switch to kstack
+||[!B0] STDW .D2T2 SP:DP,*--SP[1] ; kernel: save on current stack
+
+ SUBAW .D2 SP,2,SP
+
+ ADD .D1X SP,-8,A15
+ || STDW .D2T1 A15:A14,*SP--[16] ; save A15:A14
+
+ STDW .D2T2 B13:B12,*SP--[1]
+ || STDW .D1T1 A13:A12,*A15--[1]
+ || MVC .S2 __rp,B13
+
+ STDW .D2T2 B11:B10,*SP--[1]
+ || STDW .D1T1 A11:A10,*A15--[1]
+ || MVC .S2 CSR,B12
+
+ STDW .D2T2 B9:B8,*SP--[1]
+ || STDW .D1T1 A9:A8,*A15--[1]
+ || MVC .S2 RILC,B11
+ STDW .D2T2 B7:B6,*SP--[1]
+ || STDW .D1T1 A7:A6,*A15--[1]
+ || MVC .S2 ILC,B10
+
+ STDW .D2T2 B5:B4,*SP--[1]
+ || STDW .D1T1 A5:A4,*A15--[1]
+
+ STDW .D2T2 B3:B2,*SP--[1]
+ || STDW .D1T1 A3:A2,*A15--[1]
+ || MVC .S2 __tsr,B5
+
+ STDW .D2T2 B1:B0,*SP--[1]
+ || STDW .D1T1 A1:A0,*A15--[1]
+ || MV .S1X B5,A5
+
+ STDW .D2T2 B31:B30,*SP--[1]
+ || STDW .D1T1 A31:A30,*A15--[1]
+ STDW .D2T2 B29:B28,*SP--[1]
+ || STDW .D1T1 A29:A28,*A15--[1]
+ STDW .D2T2 B27:B26,*SP--[1]
+ || STDW .D1T1 A27:A26,*A15--[1]
+ STDW .D2T2 B25:B24,*SP--[1]
+ || STDW .D1T1 A25:A24,*A15--[1]
+ STDW .D2T2 B23:B22,*SP--[1]
+ || STDW .D1T1 A23:A22,*A15--[1]
+ STDW .D2T2 B21:B20,*SP--[1]
+ || STDW .D1T1 A21:A20,*A15--[1]
+ STDW .D2T2 B19:B18,*SP--[1]
+ || STDW .D1T1 A19:A18,*A15--[1]
+ STDW .D2T2 B17:B16,*SP--[1]
+ || STDW .D1T1 A17:A16,*A15--[1]
+
+ STDW .D2T2 B13:B12,*SP--[1] ; save PC and CSR
+
+ STDW .D2T2 B11:B10,*SP--[1] ; save RILC and ILC
+ STDW .D2T1 A5:A4,*SP--[1] ; save TSR and orig A4
+
+ ;; We left an unused word on the stack just above pt_regs.
+ ;; It is used to save whether or not this frame is due to
+ ;; a syscall. It is cleared here, but the syscall handler
+ ;; sets it to a non-zero value.
+ MVK .L2 0,B1
+ STW .D2T2 B1,*+SP(REGS__END+8) ; clear syscall flag
+ .endm
+
+ .macro RESTORE_ALL __rp __tsr
+ LDDW .D2T2 *++SP[1],B9:B8 ; get TSR (B9)
+ LDDW .D2T2 *++SP[1],B11:B10 ; get RILC (B11) and ILC (B10)
+ LDDW .D2T2 *++SP[1],B13:B12 ; get PC (B13) and CSR (B12)
+
+ ADDAW .D1X SP,30,A15
+
+ LDDW .D1T1 *++A15[1],A17:A16
+ || LDDW .D2T2 *++SP[1],B17:B16
+ LDDW .D1T1 *++A15[1],A19:A18
+ || LDDW .D2T2 *++SP[1],B19:B18
+ LDDW .D1T1 *++A15[1],A21:A20
+ || LDDW .D2T2 *++SP[1],B21:B20
+ LDDW .D1T1 *++A15[1],A23:A22
+ || LDDW .D2T2 *++SP[1],B23:B22
+ LDDW .D1T1 *++A15[1],A25:A24
+ || LDDW .D2T2 *++SP[1],B25:B24
+ LDDW .D1T1 *++A15[1],A27:A26
+ || LDDW .D2T2 *++SP[1],B27:B26
+ LDDW .D1T1 *++A15[1],A29:A28
+ || LDDW .D2T2 *++SP[1],B29:B28
+ LDDW .D1T1 *++A15[1],A31:A30
+ || LDDW .D2T2 *++SP[1],B31:B30
+
+ LDDW .D1T1 *++A15[1],A1:A0
+ || LDDW .D2T2 *++SP[1],B1:B0
+
+ LDDW .D1T1 *++A15[1],A3:A2
+ || LDDW .D2T2 *++SP[1],B3:B2
+ || MVC .S2 B9,__tsr
+ LDDW .D1T1 *++A15[1],A5:A4
+ || LDDW .D2T2 *++SP[1],B5:B4
+ || MVC .S2 B11,RILC
+ LDDW .D1T1 *++A15[1],A7:A6
+ || LDDW .D2T2 *++SP[1],B7:B6
+ || MVC .S2 B10,ILC
+
+ LDDW .D1T1 *++A15[1],A9:A8
+ || LDDW .D2T2 *++SP[1],B9:B8
+ || MVC .S2 B13,__rp
+
+ LDDW .D1T1 *++A15[1],A11:A10
+ || LDDW .D2T2 *++SP[1],B11:B10
+ || MVC .S2 B12,CSR
+
+ LDDW .D1T1 *++A15[1],A13:A12
+ || LDDW .D2T2 *++SP[1],B13:B12
+
+ MV .D2X A15,SP
+ || MVKL .S1 current_ksp,A15
+ MVKH .S1 current_ksp,A15
+ || ADDAW .D1X SP,6,A14
+ STW .D1T1 A14,*A15 ; save kernel stack pointer
+
+ LDDW .D2T1 *++SP[1],A15:A14
+
+ B .S2 __rp ; return from interruption
+ LDDW .D2T2 *+SP[1],SP:DP
+ NOP 4
+ .endm
+
+ .section .text
+
+ ;;
+ ;; Jump to schedule() then return to ret_from_exception
+ ;;
+_reschedule:
+#ifdef CONFIG_C6X_BIG_KERNEL
+ MVKL .S1 schedule,A0
+ MVKH .S1 schedule,A0
+ B .S2X A0
+#else
+ B .S1 schedule
+#endif
+ ADDKPC .S2 ret_from_exception,B3,4
+
+ ;;
+ ;; Called before syscall handler when process is being debugged
+ ;;
+tracesys_on:
+#ifdef CONFIG_C6X_BIG_KERNEL
+ MVKL .S1 syscall_trace_entry,A0
+ MVKH .S1 syscall_trace_entry,A0
+ B .S2X A0
+#else
+ B .S1 syscall_trace_entry
+#endif
+ ADDKPC .S2 ret_from_syscall_trace,B3,3
+ ADD .S1X 8,SP,A4
+
+ret_from_syscall_trace:
+ ;; tracing returns (possibly new) syscall number
+ MV .D2X A4,B0
+ || MVK .S2 __NR_syscalls,B1
+ CMPLTU .L2 B0,B1,B1
+
+ [!B1] BNOP .S2 ret_from_syscall_function,5
+ || MVK .S1 -ENOSYS,A4
+
+ ;; reload syscall args from (possibly modified) stack frame
+ ;; and get syscall handler addr from sys_call_table:
+ LDW .D2T2 *+SP(REGS_B4+8),B4
+ || MVKL .S2 sys_call_table,B1
+ LDW .D2T1 *+SP(REGS_A6+8),A6
+ || MVKH .S2 sys_call_table,B1
+ LDW .D2T2 *+B1[B0],B0
+ || MVKL .S2 ret_from_syscall_function,B3
+ LDW .D2T2 *+SP(REGS_B6+8),B6
+ || MVKH .S2 ret_from_syscall_function,B3
+ LDW .D2T1 *+SP(REGS_A8+8),A8
+ LDW .D2T2 *+SP(REGS_B8+8),B8
+ NOP
+ ; B0 = sys_call_table[__NR_*]
+ BNOP .S2 B0,5 ; branch to syscall handler
+ || LDW .D2T1 *+SP(REGS_ORIG_A4+8),A4
+
+syscall_exit_work:
+ AND .D1 _TIF_SYSCALL_TRACE,A2,A0
+ [!A0] BNOP .S1 work_pending,5
+ [A0] B .S2 syscall_trace_exit
+ ADDKPC .S2 resume_userspace,B3,1
+ MVC .S2 CSR,B1
+ SET .S2 B1,0,0,B1
+ MVC .S2 B1,CSR ; enable ints
+
+work_pending:
+ AND .D1 _TIF_NEED_RESCHED,A2,A0
+ [!A0] BNOP .S1 work_notifysig,5
+
+work_resched:
+#ifdef CONFIG_C6X_BIG_KERNEL
+ MVKL .S1 schedule,A1
+ MVKH .S1 schedule,A1
+ B .S2X A1
+#else
+ B .S2 schedule
+#endif
+ ADDKPC .S2 work_rescheduled,B3,4
+work_rescheduled:
+ ;; make sure we don't miss an interrupt setting need_resched or
+ ;; sigpending between sampling and the rti
+ MASK_INT B2
+ GET_THREAD_INFO A12
+ LDW .D1T1 *+A12(THREAD_INFO_FLAGS),A2
+ MVK .S1 _TIF_WORK_MASK,A1
+ MVK .S1 _TIF_NEED_RESCHED,A3
+ NOP 2
+ AND .D1 A1,A2,A0
+ || AND .S1 A3,A2,A1
+ [!A0] BNOP .S1 restore_all,5
+ [A1] BNOP .S1 work_resched,5
+
+work_notifysig:
+ B .S2 do_notify_resume
+ LDW .D2T1 *+SP(REGS__END+8),A6 ; syscall flag
+ ADDKPC .S2 resume_userspace,B3,1
+ ADD .S1X 8,SP,A4 ; pt_regs pointer is first arg
+ MV .D2X A2,B4 ; thread_info flags is second arg
+
+ ;;
+ ;; On C64x+, the return way from exception and interrupt
+ ;; is a little bit different
+ ;;
+ENTRY(ret_from_exception)
+#ifdef CONFIG_PREEMPT
+ MASK_INT B2
+#endif
+
+ENTRY(ret_from_interrupt)
+ ;;
+ ;; Check if we are comming from user mode.
+ ;;
+ LDW .D2T2 *+SP(REGS_TSR+8),B0
+ MVK .S2 0x40,B1
+ NOP 3
+ AND .D2 B0,B1,B0
+ [!B0] BNOP .S2 resume_kernel,5
+
+resume_userspace:
+ ;; make sure we don't miss an interrupt setting need_resched or
+ ;; sigpending between sampling and the rti
+ MASK_INT B2
+ GET_THREAD_INFO A12
+ LDW .D1T1 *+A12(THREAD_INFO_FLAGS),A2
+ MVK .S1 _TIF_WORK_MASK,A1
+ MVK .S1 _TIF_NEED_RESCHED,A3
+ NOP 2
+ AND .D1 A1,A2,A0
+ [A0] BNOP .S1 work_pending,5
+ BNOP .S1 restore_all,5
+
+ ;;
+ ;; System call handling
+ ;; B0 = syscall number (in sys_call_table)
+ ;; A4,B4,A6,B6,A8,B8 = arguments of the syscall function
+ ;; A4 is the return value register
+ ;;
+system_call_saved:
+ MVK .L2 1,B2
+ STW .D2T2 B2,*+SP(REGS__END+8) ; set syscall flag
+ MVC .S2 B2,ECR ; ack the software exception
+
+ UNMASK_INT B2 ; re-enable global IT
+
+system_call_saved_noack:
+ ;; Check system call number
+ MVK .S2 __NR_syscalls,B1
+#ifdef CONFIG_C6X_BIG_KERNEL
+ || MVKL .S1 sys_ni_syscall,A0
+#endif
+ CMPLTU .L2 B0,B1,B1
+#ifdef CONFIG_C6X_BIG_KERNEL
+ || MVKH .S1 sys_ni_syscall,A0
+#endif
+
+ ;; Check for ptrace
+ GET_THREAD_INFO A12
+
+#ifdef CONFIG_C6X_BIG_KERNEL
+ [!B1] B .S2X A0
+#else
+ [!B1] B .S2 sys_ni_syscall
+#endif
+ [!B1] ADDKPC .S2 ret_from_syscall_function,B3,4
+
+ ;; Get syscall handler addr from sys_call_table
+ ;; call tracesys_on or call syscall handler
+ LDW .D1T1 *+A12(THREAD_INFO_FLAGS),A2
+ || MVKL .S2 sys_call_table,B1
+ MVKH .S2 sys_call_table,B1
+ LDW .D2T2 *+B1[B0],B0
+ NOP 2
+ ; A2 = thread_info flags
+ AND .D1 _TIF_SYSCALL_TRACE,A2,A2
+ [A2] BNOP .S1 tracesys_on,5
+ ;; B0 = _sys_call_table[__NR_*]
+ B .S2 B0
+ ADDKPC .S2 ret_from_syscall_function,B3,4
+
+ret_from_syscall_function:
+ STW .D2T1 A4,*+SP(REGS_A4+8) ; save return value in A4
+ ; original A4 is in orig_A4
+syscall_exit:
+ ;; make sure we don't miss an interrupt setting need_resched or
+ ;; sigpending between sampling and the rti
+ MASK_INT B2
+ LDW .D1T1 *+A12(THREAD_INFO_FLAGS),A2
+ MVK .S1 _TIF_ALLWORK_MASK,A1
+ NOP 3
+ AND .D1 A1,A2,A2 ; check for work to do
+ [A2] BNOP .S1 syscall_exit_work,5
+
+restore_all:
+ RESTORE_ALL NRP,NTSR
+
+ ;;
+ ;; After a fork we jump here directly from resume,
+ ;; so that A4 contains the previous task structure.
+ ;;
+ENTRY(ret_from_fork)
+#ifdef CONFIG_C6X_BIG_KERNEL
+ MVKL .S1 schedule_tail,A0
+ MVKH .S1 schedule_tail,A0
+ B .S2X A0
+#else
+ B .S2 schedule_tail
+#endif
+ ADDKPC .S2 ret_from_fork_2,B3,4
+ret_from_fork_2:
+ ;; return 0 in A4 for child process
+ GET_THREAD_INFO A12
+ BNOP .S2 syscall_exit,3
+ MVK .L2 0,B0
+ STW .D2T2 B0,*+SP(REGS_A4+8)
+ENDPROC(ret_from_fork)
+
+ ;;
+ ;; These are the interrupt handlers, responsible for calling __do_IRQ()
+ ;; int6 is used for syscalls (see _system_call entry)
+ ;;
+ .macro SAVE_ALL_INT
+ SAVE_ALL IRP,ITSR
+ .endm
+
+ .macro CALL_INT int
+#ifdef CONFIG_C6X_BIG_KERNEL
+ MVKL .S1 c6x_do_IRQ,A0
+ MVKH .S1 c6x_do_IRQ,A0
+ BNOP .S2X A0,1
+ MVK .S1 int,A4
+ ADDAW .D2 SP,2,B4
+ MVKL .S2 ret_from_interrupt,B3
+ MVKH .S2 ret_from_interrupt,B3
+#else
+ CALLP .S2 c6x_do_IRQ,B3
+ || MVK .S1 int,A4
+ || ADDAW .D2 SP,2,B4
+ B .S1 ret_from_interrupt
+ NOP 5
+#endif
+ .endm
+
+ENTRY(_int4_handler)
+ SAVE_ALL_INT
+ CALL_INT 4
+ENDPROC(_int4_handler)
+
+ENTRY(_int5_handler)
+ SAVE_ALL_INT
+ CALL_INT 5
+ENDPROC(_int5_handler)
+
+ENTRY(_int6_handler)
+ SAVE_ALL_INT
+ CALL_INT 6
+ENDPROC(_int6_handler)
+
+ENTRY(_int7_handler)
+ SAVE_ALL_INT
+ CALL_INT 7
+ENDPROC(_int7_handler)
+
+ENTRY(_int8_handler)
+ SAVE_ALL_INT
+ CALL_INT 8
+ENDPROC(_int8_handler)
+
+ENTRY(_int9_handler)
+ SAVE_ALL_INT
+ CALL_INT 9
+ENDPROC(_int9_handler)
+
+ENTRY(_int10_handler)
+ SAVE_ALL_INT
+ CALL_INT 10
+ENDPROC(_int10_handler)
+
+ENTRY(_int11_handler)
+ SAVE_ALL_INT
+ CALL_INT 11
+ENDPROC(_int11_handler)
+
+ENTRY(_int12_handler)
+ SAVE_ALL_INT
+ CALL_INT 12
+ENDPROC(_int12_handler)
+
+ENTRY(_int13_handler)
+ SAVE_ALL_INT
+ CALL_INT 13
+ENDPROC(_int13_handler)
+
+ENTRY(_int14_handler)
+ SAVE_ALL_INT
+ CALL_INT 14
+ENDPROC(_int14_handler)
+
+ENTRY(_int15_handler)
+ SAVE_ALL_INT
+ CALL_INT 15
+ENDPROC(_int15_handler)
+
+ ;;
+ ;; Handler for uninitialized and spurious interrupts
+ ;;
+ENTRY(_bad_interrupt)
+ B .S2 IRP
+ NOP 5
+ENDPROC(_bad_interrupt)
+
+ ;;
+ ;; Entry for NMI/exceptions/syscall
+ ;;
+ENTRY(_nmi_handler)
+ SAVE_ALL NRP,NTSR
+
+ MVC .S2 EFR,B2
+ CMPEQ .L2 1,B2,B2
+ || MVC .S2 TSR,B1
+ CLR .S2 B1,10,10,B1
+ MVC .S2 B1,TSR
+#ifdef CONFIG_C6X_BIG_KERNEL
+ [!B2] MVKL .S1 process_exception,A0
+ [!B2] MVKH .S1 process_exception,A0
+ [!B2] B .S2X A0
+#else
+ [!B2] B .S2 process_exception
+#endif
+ [B2] B .S2 system_call_saved
+ [!B2] ADDAW .D2 SP,2,B1
+ [!B2] MV .D1X B1,A4
+ ADDKPC .S2 ret_from_trap,B3,2
+
+ret_from_trap:
+ MV .D2X A4,B0
+ [!B0] BNOP .S2 ret_from_exception,5
+
+#ifdef CONFIG_C6X_BIG_KERNEL
+ MVKL .S2 system_call_saved_noack,B3
+ MVKH .S2 system_call_saved_noack,B3
+#endif
+ LDW .D2T2 *+SP(REGS_B0+8),B0
+ LDW .D2T1 *+SP(REGS_A4+8),A4
+ LDW .D2T2 *+SP(REGS_B4+8),B4
+ LDW .D2T1 *+SP(REGS_A6+8),A6
+ LDW .D2T2 *+SP(REGS_B6+8),B6
+ LDW .D2T1 *+SP(REGS_A8+8),A8
+#ifdef CONFIG_C6X_BIG_KERNEL
+ || B .S2 B3
+#else
+ || B .S2 system_call_saved_noack
+#endif
+ LDW .D2T2 *+SP(REGS_B8+8),B8
+ NOP 4
+ENDPROC(_nmi_handler)
+
+ ;;
+ ;; Jump to schedule() then return to ret_from_isr
+ ;;
+#ifdef CONFIG_PREEMPT
+resume_kernel:
+ GET_THREAD_INFO A12
+ LDW .D1T1 *+A12(THREAD_INFO_PREEMPT_COUNT),A1
+ NOP 4
+ [A1] BNOP .S2 restore_all,5
+
+preempt_schedule:
+ GET_THREAD_INFO A2
+ LDW .D1T1 *+A2(THREAD_INFO_FLAGS),A1
+#ifdef CONFIG_C6X_BIG_KERNEL
+ MVKL .S2 preempt_schedule_irq,B0
+ MVKH .S2 preempt_schedule_irq,B0
+ NOP 2
+#else
+ NOP 4
+#endif
+ AND .D1 _TIF_NEED_RESCHED,A1,A1
+ [!A1] BNOP .S2 restore_all,5
+#ifdef CONFIG_C6X_BIG_KERNEL
+ B .S2 B0
+#else
+ B .S2 preempt_schedule_irq
+#endif
+ ADDKPC .S2 preempt_schedule,B3,4
+#endif /* CONFIG_PREEMPT */
+
+ENTRY(enable_exception)
+ DINT
+ MVC .S2 TSR,B0
+ MVC .S2 B3,NRP
+ MVK .L2 0xc,B1
+ OR .D2 B0,B1,B0
+ MVC .S2 B0,TSR ; Set GEE and XEN in TSR
+ B .S2 NRP
+ NOP 5
+ENDPROC(enable_exception)
+
+ENTRY(sys_sigaltstack)
+#ifdef CONFIG_C6X_BIG_KERNEL
+ MVKL .S1 do_sigaltstack,A0 ; branch to do_sigaltstack
+ MVKH .S1 do_sigaltstack,A0
+ B .S2X A0
+#else
+ B .S2 do_sigaltstack
+#endif
+ LDW .D2T1 *+SP(REGS_SP+8),A6
+ NOP 4
+ENDPROC(sys_sigaltstack)
+
+ ;; kernel_execve
+ENTRY(kernel_execve)
+ MVK .S2 __NR_execve,B0
+ SWE
+ BNOP .S2 B3,5
+ENDPROC(kernel_execve)
+
+ ;;
+ ;; Special system calls
+ ;; return address is in B3
+ ;;
+ENTRY(sys_clone)
+ ADD .D1X SP,8,A4
+#ifdef CONFIG_C6X_BIG_KERNEL
+ || MVKL .S1 sys_c6x_clone,A0
+ MVKH .S1 sys_c6x_clone,A0
+ BNOP .S2X A0,5
+#else
+ || B .S2 sys_c6x_clone
+ NOP 5
+#endif
+ENDPROC(sys_clone)
+
+ENTRY(sys_rt_sigreturn)
+ ADD .D1X SP,8,A4
+#ifdef CONFIG_C6X_BIG_KERNEL
+ || MVKL .S1 do_rt_sigreturn,A0
+ MVKH .S1 do_rt_sigreturn,A0
+ BNOP .S2X A0,5
+#else
+ || B .S2 do_rt_sigreturn
+ NOP 5
+#endif
+ENDPROC(sys_rt_sigreturn)
+
+ENTRY(sys_execve)
+ ADDAW .D2 SP,2,B6 ; put regs addr in 4th parameter
+ ; & adjust regs stack addr
+ LDW .D2T2 *+SP(REGS_B4+8),B4
+
+ ;; c6x_execve(char *name, char **argv,
+ ;; char **envp, struct pt_regs *regs)
+#ifdef CONFIG_C6X_BIG_KERNEL
+ || MVKL .S1 sys_c6x_execve,A0
+ MVKH .S1 sys_c6x_execve,A0
+ B .S2X A0
+#else
+ || B .S2 sys_c6x_execve
+#endif
+ STW .D2T2 B3,*SP--[2]
+ ADDKPC .S2 ret_from_c6x_execve,B3,3
+
+ret_from_c6x_execve:
+ LDW .D2T2 *++SP[2],B3
+ NOP 4
+ BNOP .S2 B3,5
+ENDPROC(sys_execve)
+
+ENTRY(sys_pread_c6x)
+ MV .D2X A8,B7
+#ifdef CONFIG_C6X_BIG_KERNEL
+ || MVKL .S1 sys_pread64,A0
+ MVKH .S1 sys_pread64,A0
+ BNOP .S2X A0,5
+#else
+ || B .S2 sys_pread64
+ NOP 5
+#endif
+ENDPROC(sys_pread_c6x)
+
+ENTRY(sys_pwrite_c6x)
+ MV .D2X A8,B7
+#ifdef CONFIG_C6X_BIG_KERNEL
+ || MVKL .S1 sys_pwrite64,A0
+ MVKH .S1 sys_pwrite64,A0
+ BNOP .S2X A0,5
+#else
+ || B .S2 sys_pwrite64
+ NOP 5
+#endif
+ENDPROC(sys_pwrite_c6x)
+
+;; On Entry
+;; A4 - path
+;; B4 - offset_lo (LE), offset_hi (BE)
+;; A6 - offset_lo (BE), offset_hi (LE)
+ENTRY(sys_truncate64_c6x)
+#ifdef CONFIG_CPU_BIG_ENDIAN
+ MV .S2 B4,B5
+ MV .D2X A6,B4
+#else
+ MV .D2X A6,B5
+#endif
+#ifdef CONFIG_C6X_BIG_KERNEL
+ || MVKL .S1 sys_truncate64,A0
+ MVKH .S1 sys_truncate64,A0
+ BNOP .S2X A0,5
+#else
+ || B .S2 sys_truncate64
+ NOP 5
+#endif
+ENDPROC(sys_truncate64_c6x)
+
+;; On Entry
+;; A4 - fd
+;; B4 - offset_lo (LE), offset_hi (BE)
+;; A6 - offset_lo (BE), offset_hi (LE)
+ENTRY(sys_ftruncate64_c6x)
+#ifdef CONFIG_CPU_BIG_ENDIAN
+ MV .S2 B4,B5
+ MV .D2X A6,B4
+#else
+ MV .D2X A6,B5
+#endif
+#ifdef CONFIG_C6X_BIG_KERNEL
+ || MVKL .S1 sys_ftruncate64,A0
+ MVKH .S1 sys_ftruncate64,A0
+ BNOP .S2X A0,5
+#else
+ || B .S2 sys_ftruncate64
+ NOP 5
+#endif
+ENDPROC(sys_ftruncate64_c6x)
+
+#ifdef __ARCH_WANT_SYSCALL_OFF_T
+;; On Entry
+;; A4 - fd
+;; B4 - offset_lo (LE), offset_hi (BE)
+;; A6 - offset_lo (BE), offset_hi (LE)
+;; B6 - len
+;; A8 - advice
+ENTRY(sys_fadvise64_c6x)
+#ifdef CONFIG_C6X_BIG_KERNEL
+ MVKL .S1 sys_fadvise64,A0
+ MVKH .S1 sys_fadvise64,A0
+ BNOP .S2X A0,2
+#else
+ B .S2 sys_fadvise64
+ NOP 2
+#endif
+#ifdef CONFIG_CPU_BIG_ENDIAN
+ MV .L2 B4,B5
+ || MV .D2X A6,B4
+#else
+ MV .D2X A6,B5
+#endif
+ MV .D1X B6,A6
+ MV .D2X A8,B6
+#endif
+ENDPROC(sys_fadvise64_c6x)
+
+;; On Entry
+;; A4 - fd
+;; B4 - offset_lo (LE), offset_hi (BE)
+;; A6 - offset_lo (BE), offset_hi (LE)
+;; B6 - len_lo (LE), len_hi (BE)
+;; A8 - len_lo (BE), len_hi (LE)
+;; B8 - advice
+ENTRY(sys_fadvise64_64_c6x)
+#ifdef CONFIG_C6X_BIG_KERNEL
+ MVKL .S1 sys_fadvise64_64,A0
+ MVKH .S1 sys_fadvise64_64,A0
+ BNOP .S2X A0,2
+#else
+ B .S2 sys_fadvise64_64
+ NOP 2
+#endif
+#ifdef CONFIG_CPU_BIG_ENDIAN
+ MV .L2 B4,B5
+ || MV .D2X A6,B4
+ MV .L1 A8,A6
+ || MV .D1X B6,A7
+#else
+ MV .D2X A6,B5
+ MV .L1 A8,A7
+ || MV .D1X B6,A6
+#endif
+ MV .L2 B8,B6
+ENDPROC(sys_fadvise64_64_c6x)
+
+;; On Entry
+;; A4 - fd
+;; B4 - mode
+;; A6 - offset_hi
+;; B6 - offset_lo
+;; A8 - len_hi
+;; B8 - len_lo
+ENTRY(sys_fallocate_c6x)
+#ifdef CONFIG_C6X_BIG_KERNEL
+ MVKL .S1 sys_fallocate,A0
+ MVKH .S1 sys_fallocate,A0
+ BNOP .S2X A0,1
+#else
+ B .S2 sys_fallocate
+ NOP
+#endif
+ MV .D1 A6,A7
+ MV .D1X B6,A6
+ MV .D2X A8,B7
+ MV .D2 B8,B6
+ENDPROC(sys_fallocate_c6x)
+
+ ;; put this in .neardata for faster access when using DSBT mode
+ .section .neardata,"aw",@progbits
+ .global current_ksp
+ .hidden current_ksp
+current_ksp:
+ .word init_thread_union + THREAD_START_SP
diff --git a/arch/c6x/kernel/traps.c b/arch/c6x/kernel/traps.c
new file mode 100644
index 0000000..f50e3ed
--- /dev/null
+++ b/arch/c6x/kernel/traps.c
@@ -0,0 +1,423 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2006, 2009, 2010, 2011 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/module.h>
+#include <linux/ptrace.h>
+#include <linux/kallsyms.h>
+#include <linux/bug.h>
+
+#include <asm/soc.h>
+#include <asm/traps.h>
+
+int (*c6x_nmi_handler)(struct pt_regs *regs);
+
+void __init trap_init(void)
+{
+ ack_exception(EXCEPT_TYPE_NXF);
+ ack_exception(EXCEPT_TYPE_EXC);
+ ack_exception(EXCEPT_TYPE_IXF);
+ ack_exception(EXCEPT_TYPE_SXF);
+ enable_exception();
+}
+
+void show_regs(struct pt_regs *regs)
+{
+ pr_err("\n");
+ pr_err("PC: %08lx SP: %08lx\n", regs->pc, regs->sp);
+ pr_err("Status: %08lx ORIG_A4: %08lx\n", regs->csr, regs->orig_a4);
+ pr_err("A0: %08lx B0: %08lx\n", regs->a0, regs->b0);
+ pr_err("A1: %08lx B1: %08lx\n", regs->a1, regs->b1);
+ pr_err("A2: %08lx B2: %08lx\n", regs->a2, regs->b2);
+ pr_err("A3: %08lx B3: %08lx\n", regs->a3, regs->b3);
+ pr_err("A4: %08lx B4: %08lx\n", regs->a4, regs->b4);
+ pr_err("A5: %08lx B5: %08lx\n", regs->a5, regs->b5);
+ pr_err("A6: %08lx B6: %08lx\n", regs->a6, regs->b6);
+ pr_err("A7: %08lx B7: %08lx\n", regs->a7, regs->b7);
+ pr_err("A8: %08lx B8: %08lx\n", regs->a8, regs->b8);
+ pr_err("A9: %08lx B9: %08lx\n", regs->a9, regs->b9);
+ pr_err("A10: %08lx B10: %08lx\n", regs->a10, regs->b10);
+ pr_err("A11: %08lx B11: %08lx\n", regs->a11, regs->b11);
+ pr_err("A12: %08lx B12: %08lx\n", regs->a12, regs->b12);
+ pr_err("A13: %08lx B13: %08lx\n", regs->a13, regs->b13);
+ pr_err("A14: %08lx B14: %08lx\n", regs->a14, regs->dp);
+ pr_err("A15: %08lx B15: %08lx\n", regs->a15, regs->sp);
+ pr_err("A16: %08lx B16: %08lx\n", regs->a16, regs->b16);
+ pr_err("A17: %08lx B17: %08lx\n", regs->a17, regs->b17);
+ pr_err("A18: %08lx B18: %08lx\n", regs->a18, regs->b18);
+ pr_err("A19: %08lx B19: %08lx\n", regs->a19, regs->b19);
+ pr_err("A20: %08lx B20: %08lx\n", regs->a20, regs->b20);
+ pr_err("A21: %08lx B21: %08lx\n", regs->a21, regs->b21);
+ pr_err("A22: %08lx B22: %08lx\n", regs->a22, regs->b22);
+ pr_err("A23: %08lx B23: %08lx\n", regs->a23, regs->b23);
+ pr_err("A24: %08lx B24: %08lx\n", regs->a24, regs->b24);
+ pr_err("A25: %08lx B25: %08lx\n", regs->a25, regs->b25);
+ pr_err("A26: %08lx B26: %08lx\n", regs->a26, regs->b26);
+ pr_err("A27: %08lx B27: %08lx\n", regs->a27, regs->b27);
+ pr_err("A28: %08lx B28: %08lx\n", regs->a28, regs->b28);
+ pr_err("A29: %08lx B29: %08lx\n", regs->a29, regs->b29);
+ pr_err("A30: %08lx B30: %08lx\n", regs->a30, regs->b30);
+ pr_err("A31: %08lx B31: %08lx\n", regs->a31, regs->b31);
+}
+
+void dump_stack(void)
+{
+ unsigned long stack;
+
+ show_stack(current, &stack);
+}
+EXPORT_SYMBOL(dump_stack);
+
+
+void die(char *str, struct pt_regs *fp, int nr)
+{
+ console_verbose();
+ pr_err("%s: %08x\n", str, nr);
+ show_regs(fp);
+
+ pr_err("Process %s (pid: %d, stackpage=%08lx)\n",
+ current->comm, current->pid, (PAGE_SIZE +
+ (unsigned long) current));
+
+ dump_stack();
+ while (1)
+ ;
+}
+
+static void die_if_kernel(char *str, struct pt_regs *fp, int nr)
+{
+ if (user_mode(fp))
+ return;
+
+ die(str, fp, nr);
+}
+
+
+/* Internal exceptions */
+static struct exception_info iexcept_table[10] = {
+ { "Oops - instruction fetch", SIGBUS, BUS_ADRERR },
+ { "Oops - fetch packet", SIGBUS, BUS_ADRERR },
+ { "Oops - execute packet", SIGILL, ILL_ILLOPC },
+ { "Oops - undefined instruction", SIGILL, ILL_ILLOPC },
+ { "Oops - resource conflict", SIGILL, ILL_ILLOPC },
+ { "Oops - resource access", SIGILL, ILL_PRVREG },
+ { "Oops - privilege", SIGILL, ILL_PRVOPC },
+ { "Oops - loops buffer", SIGILL, ILL_ILLOPC },
+ { "Oops - software exception", SIGILL, ILL_ILLTRP },
+ { "Oops - unknown exception", SIGILL, ILL_ILLOPC }
+};
+
+/* External exceptions */
+static struct exception_info eexcept_table[128] = {
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - external exception", SIGBUS, BUS_ADRERR },
+ { "Oops - CPU memory protection fault", SIGSEGV, SEGV_ACCERR },
+ { "Oops - CPU memory protection fault in L1P", SIGSEGV, SEGV_ACCERR },
+ { "Oops - DMA memory protection fault in L1P", SIGSEGV, SEGV_ACCERR },
+ { "Oops - CPU memory protection fault in L1D", SIGSEGV, SEGV_ACCERR },
+ { "Oops - DMA memory protection fault in L1D", SIGSEGV, SEGV_ACCERR },
+ { "Oops - CPU memory protection fault in L2", SIGSEGV, SEGV_ACCERR },
+ { "Oops - DMA memory protection fault in L2", SIGSEGV, SEGV_ACCERR },
+ { "Oops - EMC CPU memory protection fault", SIGSEGV, SEGV_ACCERR },
+ { "Oops - EMC bus error", SIGBUS, BUS_ADRERR }
+};
+
+static void do_trap(struct exception_info *except_info, struct pt_regs *regs)
+{
+ unsigned long addr = instruction_pointer(regs);
+ siginfo_t info;
+
+ if (except_info->code != TRAP_BRKPT)
+ pr_err("TRAP: %s PC[0x%lx] signo[%d] code[%d]\n",
+ except_info->kernel_str, regs->pc,
+ except_info->signo, except_info->code);
+
+ die_if_kernel(except_info->kernel_str, regs, addr);
+
+ info.si_signo = except_info->signo;
+ info.si_errno = 0;
+ info.si_code = except_info->code;
+ info.si_addr = (void __user *)addr;
+
+ force_sig_info(except_info->signo, &info, current);
+}
+
+/*
+ * Process an internal exception (non maskable)
+ */
+static int process_iexcept(struct pt_regs *regs)
+{
+ unsigned int iexcept_report = get_iexcept();
+ unsigned int iexcept_num;
+
+ ack_exception(EXCEPT_TYPE_IXF);
+
+ pr_err("IEXCEPT: PC[0x%lx]\n", regs->pc);
+
+ while (iexcept_report) {
+ iexcept_num = __ffs(iexcept_report);
+ iexcept_report &= ~(1 << iexcept_num);
+ set_iexcept(iexcept_report);
+ if (*(unsigned int *)regs->pc == BKPT_OPCODE) {
+ /* This is a breakpoint */
+ struct exception_info bkpt_exception = {
+ "Oops - undefined instruction",
+ SIGTRAP, TRAP_BRKPT
+ };
+ do_trap(&bkpt_exception, regs);
+ iexcept_report &= ~(0xFF);
+ set_iexcept(iexcept_report);
+ continue;
+ }
+
+ do_trap(&iexcept_table[iexcept_num], regs);
+ }
+ return 0;
+}
+
+/*
+ * Process an external exception (maskable)
+ */
+static void process_eexcept(struct pt_regs *regs)
+{
+ int evt;
+
+ pr_err("EEXCEPT: PC[0x%lx]\n", regs->pc);
+
+ while ((evt = soc_get_exception()) >= 0)
+ do_trap(&eexcept_table[evt], regs);
+
+ ack_exception(EXCEPT_TYPE_EXC);
+}
+
+/*
+ * Main exception processing
+ */
+asmlinkage int process_exception(struct pt_regs *regs)
+{
+ unsigned int type;
+ unsigned int type_num;
+ unsigned int ie_num = 9; /* default is unknown exception */
+
+ while ((type = get_except_type()) != 0) {
+ type_num = fls(type) - 1;
+
+ switch (type_num) {
+ case EXCEPT_TYPE_NXF:
+ ack_exception(EXCEPT_TYPE_NXF);
+ if (c6x_nmi_handler)
+ (c6x_nmi_handler)(regs);
+ else
+ pr_alert("NMI interrupt!\n");
+ break;
+
+ case EXCEPT_TYPE_IXF:
+ if (process_iexcept(regs))
+ return 1;
+ break;
+
+ case EXCEPT_TYPE_EXC:
+ process_eexcept(regs);
+ break;
+
+ case EXCEPT_TYPE_SXF:
+ ie_num = 8;
+ default:
+ ack_exception(type_num);
+ do_trap(&iexcept_table[ie_num], regs);
+ break;
+ }
+ }
+ return 0;
+}
+
+static int kstack_depth_to_print = 48;
+
+static void show_trace(unsigned long *stack, unsigned long *endstack)
+{
+ unsigned long addr;
+ int i;
+
+ pr_debug("Call trace:");
+ i = 0;
+ while (stack + 1 <= endstack) {
+ addr = *stack++;
+ /*
+ * If the address is either in the text segment of the
+ * kernel, or in the region which contains vmalloc'ed
+ * memory, it *may* be the address of a calling
+ * routine; if so, print it so that someone tracing
+ * down the cause of the crash will be able to figure
+ * out the call path that was taken.
+ */
+ if (__kernel_text_address(addr)) {
+#ifndef CONFIG_KALLSYMS
+ if (i % 5 == 0)
+ pr_debug("\n ");
+#endif
+ pr_debug(" [<%08lx>]", addr);
+ print_symbol(" %s\n", addr);
+ i++;
+ }
+ }
+ pr_debug("\n");
+}
+
+void show_stack(struct task_struct *task, unsigned long *stack)
+{
+ unsigned long *p, *endstack;
+ int i;
+
+ if (!stack) {
+ if (task && task != current)
+ /* We know this is a kernel stack,
+ so this is the start/end */
+ stack = (unsigned long *)thread_saved_ksp(task);
+ else
+ stack = (unsigned long *)&stack;
+ }
+ endstack = (unsigned long *)(((unsigned long)stack + THREAD_SIZE - 1)
+ & -THREAD_SIZE);
+
+ pr_debug("Stack from %08lx:", (unsigned long)stack);
+ for (i = 0, p = stack; i < kstack_depth_to_print; i++) {
+ if (p + 1 > endstack)
+ break;
+ if (i % 8 == 0)
+ pr_cont("\n ");
+ pr_cont(" %08lx", *p++);
+ }
+ pr_cont("\n");
+ show_trace(stack, endstack);
+}
+
+int is_valid_bugaddr(unsigned long addr)
+{
+ return __kernel_text_address(addr);
+}
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v3 15/24] C6X: clocks
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (13 preceding siblings ...)
2011-09-27 20:29 ` [PATCH v3 14/24] C6X: traps Mark Salter
@ 2011-09-27 20:29 ` Mark Salter
2011-09-28 13:51 ` Arnd Bergmann
2011-09-27 20:29 ` [PATCH v3 16/24] C6X: cache control Mark Salter
` (10 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:29 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
The C6X SoCs contain several PLL controllers each with up to 16 clock outputs
feeding into the cores or peripheral clock domains. The hardware is very similar
to arm/mach-davinci clocks. This is still a work in progress which needs to be
updated once device tree clock binding changes shake out.
Signed-off-by: Mark Salter <msalter@redhat.com>
---
arch/c6x/include/asm/clkdev.h | 22 ++
arch/c6x/include/asm/clock.h | 148 ++++++++++++++
arch/c6x/platforms/pll.c | 444 +++++++++++++++++++++++++++++++++++++++++
arch/c6x/platforms/plldata.c | 404 +++++++++++++++++++++++++++++++++++++
4 files changed, 1018 insertions(+), 0 deletions(-)
create mode 100644 arch/c6x/include/asm/clkdev.h
create mode 100644 arch/c6x/include/asm/clock.h
create mode 100644 arch/c6x/platforms/pll.c
create mode 100644 arch/c6x/platforms/plldata.c
diff --git a/arch/c6x/include/asm/clkdev.h b/arch/c6x/include/asm/clkdev.h
new file mode 100644
index 0000000..76a070b
--- /dev/null
+++ b/arch/c6x/include/asm/clkdev.h
@@ -0,0 +1,22 @@
+#ifndef _ASM_CLKDEV_H
+#define _ASM_CLKDEV_H
+
+#include <linux/slab.h>
+
+struct clk;
+
+static inline int __clk_get(struct clk *clk)
+{
+ return 1;
+}
+
+static inline void __clk_put(struct clk *clk)
+{
+}
+
+static inline struct clk_lookup_alloc *__clkdev_alloc(size_t size)
+{
+ return kzalloc(size, GFP_KERNEL);
+}
+
+#endif /* _ASM_CLKDEV_H */
diff --git a/arch/c6x/include/asm/clock.h b/arch/c6x/include/asm/clock.h
new file mode 100644
index 0000000..bcf42b2
--- /dev/null
+++ b/arch/c6x/include/asm/clock.h
@@ -0,0 +1,148 @@
+/*
+ * TI C64X clock definitions
+ *
+ * Copyright (C) 2010, 2011 Texas Instruments.
+ * Contributed by: Mark Salter <msalter@redhat.com>
+ *
+ * Copied heavily from arm/mach-davinci/clock.h, so:
+ *
+ * Copyright (C) 2006-2007 Texas Instruments.
+ * Copyright (C) 2008-2009 Deep Root Systems, LLC
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#ifndef _ASM_C6X_CLOCK_H
+#define _ASM_C6X_CLOCK_H
+
+#ifndef __ASSEMBLER__
+
+#include <linux/list.h>
+
+/* PLL/Reset register offsets */
+#define PLLCTL 0x100
+#define PLLM 0x110
+#define PLLPRE 0x114
+#define PLLDIV1 0x118
+#define PLLDIV2 0x11c
+#define PLLDIV3 0x120
+#define PLLPOST 0x128
+#define PLLCMD 0x138
+#define PLLSTAT 0x13c
+#define PLLALNCTL 0x140
+#define PLLDCHANGE 0x144
+#define PLLCKEN 0x148
+#define PLLCKSTAT 0x14c
+#define PLLSYSTAT 0x150
+#define PLLDIV4 0x160
+#define PLLDIV5 0x164
+#define PLLDIV6 0x168
+#define PLLDIV7 0x16c
+#define PLLDIV8 0x170
+#define PLLDIV9 0x174
+#define PLLDIV10 0x178
+#define PLLDIV11 0x17c
+#define PLLDIV12 0x180
+#define PLLDIV13 0x184
+#define PLLDIV14 0x188
+#define PLLDIV15 0x18c
+#define PLLDIV16 0x190
+
+/* PLLM register bits */
+#define PLLM_PLLM_MASK 0xff
+#define PLLM_VAL(x) ((x) - 1)
+
+/* PREDIV register bits */
+#define PLLPREDIV_EN BIT(15)
+#define PLLPREDIV_VAL(x) ((x) - 1)
+
+/* PLLCTL register bits */
+#define PLLCTL_PLLEN BIT(0)
+#define PLLCTL_PLLPWRDN BIT(1)
+#define PLLCTL_PLLRST BIT(3)
+#define PLLCTL_PLLDIS BIT(4)
+#define PLLCTL_PLLENSRC BIT(5)
+#define PLLCTL_CLKMODE BIT(8)
+
+/* PLLCMD register bits */
+#define PLLCMD_GOSTAT BIT(0)
+
+/* PLLSTAT register bits */
+#define PLLSTAT_GOSTAT BIT(0)
+
+/* PLLDIV register bits */
+#define PLLDIV_EN BIT(15)
+#define PLLDIV_RATIO_MASK 0x1f
+#define PLLDIV_RATIO(x) ((x) - 1)
+
+struct pll_data;
+
+struct clk {
+ struct list_head node;
+ struct module *owner;
+ const char *name;
+ unsigned long rate;
+ int usecount;
+ u32 flags;
+ struct clk *parent;
+ struct list_head children; /* list of children */
+ struct list_head childnode; /* parent's child list node */
+ struct pll_data *pll_data;
+ u32 div;
+ unsigned long (*recalc) (struct clk *);
+ int (*set_rate) (struct clk *clk, unsigned long rate);
+ int (*round_rate) (struct clk *clk, unsigned long rate);
+};
+
+/* Clock flags: SoC-specific flags start at BIT(16) */
+#define ALWAYS_ENABLED BIT(1)
+#define CLK_PLL BIT(2) /* PLL-derived clock */
+#define PRE_PLL BIT(3) /* source is before PLL mult/div */
+#define FIXED_DIV_PLL BIT(4) /* fixed divisor from PLL */
+#define FIXED_RATE_PLL BIT(5) /* fixed ouput rate PLL */
+
+#define MAX_PLL_SYSCLKS 16
+
+struct pll_data {
+ void __iomem *base;
+ u32 num;
+ u32 flags;
+ u32 input_rate;
+ u32 bypass_delay; /* in loops */
+ u32 reset_delay; /* in loops */
+ u32 lock_delay; /* in loops */
+ struct clk sysclks[MAX_PLL_SYSCLKS + 1];
+};
+
+/* pll_data flag bit */
+#define PLL_HAS_PRE BIT(0)
+#define PLL_HAS_MUL BIT(1)
+#define PLL_HAS_POST BIT(2)
+
+#define CLK(dev, con, ck) \
+ { \
+ .dev_id = dev, \
+ .con_id = con, \
+ .clk = ck, \
+ } \
+
+extern void c6x_clks_init(struct clk_lookup *clocks);
+extern int clk_register(struct clk *clk);
+extern void clk_unregister(struct clk *clk);
+extern void c64x_setup_clocks(void);
+
+extern struct pll_data c6x_soc_pll1;
+
+extern struct clk clkin1;
+extern struct clk c6x_core_clk;
+extern struct clk c6x_i2c_clk;
+extern struct clk c6x_watchdog_clk;
+extern struct clk c6x_mcbsp1_clk;
+extern struct clk c6x_mcbsp2_clk;
+extern struct clk c6x_mdio_clk;
+
+#endif
+
+#endif /* _ASM_C6X_CLOCK_H */
diff --git a/arch/c6x/platforms/pll.c b/arch/c6x/platforms/pll.c
new file mode 100644
index 0000000..3aa898f
--- /dev/null
+++ b/arch/c6x/platforms/pll.c
@@ -0,0 +1,444 @@
+/*
+ * Clock and PLL control for C64x+ devices
+ *
+ * Copyright (C) 2010, 2011 Texas Instruments.
+ * Contributed by: Mark Salter <msalter@redhat.com>
+ *
+ * Copied heavily from arm/mach-davinci/clock.c, so:
+ *
+ * Copyright (C) 2006-2007 Texas Instruments.
+ * Copyright (C) 2008-2009 Deep Root Systems, LLC
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <linux/module.h>
+#include <linux/clkdev.h>
+#include <linux/clk.h>
+#include <linux/io.h>
+#include <linux/err.h>
+
+#include <asm/clock.h>
+#include <asm/soc.h>
+
+static LIST_HEAD(clocks);
+static DEFINE_MUTEX(clocks_mutex);
+static DEFINE_SPINLOCK(clockfw_lock);
+
+static void __clk_enable(struct clk *clk)
+{
+ if (clk->parent)
+ __clk_enable(clk->parent);
+ clk->usecount++;
+}
+
+static void __clk_disable(struct clk *clk)
+{
+ if (WARN_ON(clk->usecount == 0))
+ return;
+ --clk->usecount;
+
+ if (clk->parent)
+ __clk_disable(clk->parent);
+}
+
+int clk_enable(struct clk *clk)
+{
+ unsigned long flags;
+
+ if (clk == NULL || IS_ERR(clk))
+ return -EINVAL;
+
+ spin_lock_irqsave(&clockfw_lock, flags);
+ __clk_enable(clk);
+ spin_unlock_irqrestore(&clockfw_lock, flags);
+
+ return 0;
+}
+EXPORT_SYMBOL(clk_enable);
+
+void clk_disable(struct clk *clk)
+{
+ unsigned long flags;
+
+ if (clk == NULL || IS_ERR(clk))
+ return;
+
+ spin_lock_irqsave(&clockfw_lock, flags);
+ __clk_disable(clk);
+ spin_unlock_irqrestore(&clockfw_lock, flags);
+}
+EXPORT_SYMBOL(clk_disable);
+
+unsigned long clk_get_rate(struct clk *clk)
+{
+ if (clk == NULL || IS_ERR(clk))
+ return -EINVAL;
+
+ return clk->rate;
+}
+EXPORT_SYMBOL(clk_get_rate);
+
+long clk_round_rate(struct clk *clk, unsigned long rate)
+{
+ if (clk == NULL || IS_ERR(clk))
+ return -EINVAL;
+
+ if (clk->round_rate)
+ return clk->round_rate(clk, rate);
+
+ return clk->rate;
+}
+EXPORT_SYMBOL(clk_round_rate);
+
+/* Propagate rate to children */
+static void propagate_rate(struct clk *root)
+{
+ struct clk *clk;
+
+ list_for_each_entry(clk, &root->children, childnode) {
+ if (clk->recalc)
+ clk->rate = clk->recalc(clk);
+ propagate_rate(clk);
+ }
+}
+
+int clk_set_rate(struct clk *clk, unsigned long rate)
+{
+ unsigned long flags;
+ int ret = -EINVAL;
+
+ if (clk == NULL || IS_ERR(clk))
+ return ret;
+
+ if (clk->set_rate)
+ ret = clk->set_rate(clk, rate);
+
+ spin_lock_irqsave(&clockfw_lock, flags);
+ if (ret == 0) {
+ if (clk->recalc)
+ clk->rate = clk->recalc(clk);
+ propagate_rate(clk);
+ }
+ spin_unlock_irqrestore(&clockfw_lock, flags);
+
+ return ret;
+}
+EXPORT_SYMBOL(clk_set_rate);
+
+int clk_set_parent(struct clk *clk, struct clk *parent)
+{
+ unsigned long flags;
+
+ if (clk == NULL || IS_ERR(clk))
+ return -EINVAL;
+
+ /* Cannot change parent on enabled clock */
+ if (WARN_ON(clk->usecount))
+ return -EINVAL;
+
+ mutex_lock(&clocks_mutex);
+ clk->parent = parent;
+ list_del_init(&clk->childnode);
+ list_add(&clk->childnode, &clk->parent->children);
+ mutex_unlock(&clocks_mutex);
+
+ spin_lock_irqsave(&clockfw_lock, flags);
+ if (clk->recalc)
+ clk->rate = clk->recalc(clk);
+ propagate_rate(clk);
+ spin_unlock_irqrestore(&clockfw_lock, flags);
+
+ return 0;
+}
+EXPORT_SYMBOL(clk_set_parent);
+
+int clk_register(struct clk *clk)
+{
+ if (clk == NULL || IS_ERR(clk))
+ return -EINVAL;
+
+ if (WARN(clk->parent && !clk->parent->rate,
+ "CLK: %s parent %s has no rate!\n",
+ clk->name, clk->parent->name))
+ return -EINVAL;
+
+ mutex_lock(&clocks_mutex);
+ list_add_tail(&clk->node, &clocks);
+ if (clk->parent)
+ list_add_tail(&clk->childnode, &clk->parent->children);
+ mutex_unlock(&clocks_mutex);
+
+ /* If rate is already set, use it */
+ if (clk->rate)
+ return 0;
+
+ /* Else, see if there is a way to calculate it */
+ if (clk->recalc)
+ clk->rate = clk->recalc(clk);
+
+ /* Otherwise, default to parent rate */
+ else if (clk->parent)
+ clk->rate = clk->parent->rate;
+
+ return 0;
+}
+EXPORT_SYMBOL(clk_register);
+
+void clk_unregister(struct clk *clk)
+{
+ if (clk == NULL || IS_ERR(clk))
+ return;
+
+ mutex_lock(&clocks_mutex);
+ list_del(&clk->node);
+ list_del(&clk->childnode);
+ mutex_unlock(&clocks_mutex);
+}
+EXPORT_SYMBOL(clk_unregister);
+
+
+static u32 pll_read(struct pll_data *pll, int reg)
+{
+ return soc_readl(pll->base + reg);
+}
+
+static unsigned long clk_sysclk_recalc(struct clk *clk)
+{
+ u32 v, plldiv = 0;
+ struct pll_data *pll;
+ unsigned long rate = clk->rate;
+
+ if (WARN_ON(!clk->parent))
+ return rate;
+
+ rate = clk->parent->rate;
+
+ /* the parent must be a PLL */
+ if (WARN_ON(!clk->parent->pll_data))
+ return rate;
+
+ pll = clk->parent->pll_data;
+
+ /* If pre-PLL, source clock is before the multiplier and divider(s) */
+ if (clk->flags & PRE_PLL)
+ rate = pll->input_rate;
+
+ if (!clk->div) {
+ pr_debug("%s: (no divider) rate = %lu KHz\n",
+ clk->name, rate / 1000);
+ return rate;
+ }
+
+ if (clk->flags & FIXED_DIV_PLL) {
+ rate /= clk->div;
+ pr_debug("%s: (fixed divide by %d) rate = %lu KHz\n",
+ clk->name, clk->div, rate / 1000);
+ return rate;
+ }
+
+ v = pll_read(pll, clk->div);
+ if (v & PLLDIV_EN)
+ plldiv = (v & PLLDIV_RATIO_MASK) + 1;
+
+ if (plldiv == 0)
+ plldiv = 1;
+
+ rate /= plldiv;
+
+ pr_debug("%s: (divide by %d) rate = %lu KHz\n",
+ clk->name, plldiv, rate / 1000);
+
+ return rate;
+}
+
+static unsigned long clk_leafclk_recalc(struct clk *clk)
+{
+ if (WARN_ON(!clk->parent))
+ return clk->rate;
+
+ pr_debug("%s: (parent %s) rate = %lu KHz\n",
+ clk->name, clk->parent->name, clk->parent->rate / 1000);
+
+ return clk->parent->rate;
+}
+
+static unsigned long clk_pllclk_recalc(struct clk *clk)
+{
+ u32 ctrl, mult = 0, prediv = 0, postdiv = 0;
+ u8 bypass;
+ struct pll_data *pll = clk->pll_data;
+ unsigned long rate = clk->rate;
+
+ if (clk->flags & FIXED_RATE_PLL)
+ return rate;
+
+ ctrl = pll_read(pll, PLLCTL);
+ rate = pll->input_rate = clk->parent->rate;
+
+ if (ctrl & PLLCTL_PLLEN)
+ bypass = 0;
+ else
+ bypass = 1;
+
+ if (pll->flags & PLL_HAS_MUL) {
+ mult = pll_read(pll, PLLM);
+ mult = (mult & PLLM_PLLM_MASK) + 1;
+ }
+ if (pll->flags & PLL_HAS_PRE) {
+ prediv = pll_read(pll, PLLPRE);
+ if (prediv & PLLDIV_EN)
+ prediv = (prediv & PLLDIV_RATIO_MASK) + 1;
+ else
+ prediv = 0;
+ }
+ if (pll->flags & PLL_HAS_POST) {
+ postdiv = pll_read(pll, PLLPOST);
+ if (postdiv & PLLDIV_EN)
+ postdiv = (postdiv & PLLDIV_RATIO_MASK) + 1;
+ else
+ postdiv = 1;
+ }
+
+ if (!bypass) {
+ if (prediv)
+ rate /= prediv;
+ if (mult)
+ rate *= mult;
+ if (postdiv)
+ rate /= postdiv;
+
+ pr_debug("PLL%d: input = %luMHz, pre[%d] mul[%d] post[%d] "
+ "--> %luMHz output.\n",
+ pll->num, clk->parent->rate / 1000000,
+ prediv, mult, postdiv, rate / 1000000);
+ } else
+ pr_debug("PLL%d: input = %luMHz, bypass mode.\n",
+ pll->num, clk->parent->rate / 1000000);
+
+ return rate;
+}
+
+
+static void __init __init_clk(struct clk *clk)
+{
+ INIT_LIST_HEAD(&clk->node);
+ INIT_LIST_HEAD(&clk->children);
+ INIT_LIST_HEAD(&clk->childnode);
+
+ if (!clk->recalc) {
+
+ /* Check if clock is a PLL */
+ if (clk->pll_data)
+ clk->recalc = clk_pllclk_recalc;
+
+ /* Else, if it is a PLL-derived clock */
+ else if (clk->flags & CLK_PLL)
+ clk->recalc = clk_sysclk_recalc;
+
+ /* Otherwise, it is a leaf clock (PSC clock) */
+ else if (clk->parent)
+ clk->recalc = clk_leafclk_recalc;
+ }
+}
+
+void __init c6x_clks_init(struct clk_lookup *clocks)
+{
+ struct clk_lookup *c;
+ struct clk *clk;
+ size_t num_clocks = 0;
+
+ for (c = clocks; c->clk; c++) {
+ clk = c->clk;
+
+ __init_clk(clk);
+ clk_register(clk);
+ num_clocks++;
+
+ /* Turn on clocks that Linux doesn't otherwise manage */
+ if (clk->flags & ALWAYS_ENABLED)
+ clk_enable(clk);
+ }
+
+ clkdev_add_table(clocks, num_clocks);
+}
+
+#ifdef CONFIG_DEBUG_FS
+
+#include <linux/debugfs.h>
+#include <linux/seq_file.h>
+
+#define CLKNAME_MAX 10 /* longest clock name */
+#define NEST_DELTA 2
+#define NEST_MAX 4
+
+static void
+dump_clock(struct seq_file *s, unsigned nest, struct clk *parent)
+{
+ char *state;
+ char buf[CLKNAME_MAX + NEST_DELTA * NEST_MAX];
+ struct clk *clk;
+ unsigned i;
+
+ if (parent->flags & CLK_PLL)
+ state = "pll";
+ else
+ state = "";
+
+ /* <nest spaces> name <pad to end> */
+ memset(buf, ' ', sizeof(buf) - 1);
+ buf[sizeof(buf) - 1] = 0;
+ i = strlen(parent->name);
+ memcpy(buf + nest, parent->name,
+ min(i, (unsigned)(sizeof(buf) - 1 - nest)));
+
+ seq_printf(s, "%s users=%2d %-3s %9ld Hz\n",
+ buf, parent->usecount, state, clk_get_rate(parent));
+ /* REVISIT show device associations too */
+
+ /* cost is now small, but not linear... */
+ list_for_each_entry(clk, &parent->children, childnode) {
+ dump_clock(s, nest + NEST_DELTA, clk);
+ }
+}
+
+static int c6x_ck_show(struct seq_file *m, void *v)
+{
+ struct clk *clk;
+
+ /*
+ * Show clock tree; We trust nonzero usecounts equate to PSC enables...
+ */
+ mutex_lock(&clocks_mutex);
+ list_for_each_entry(clk, &clocks, node)
+ if (!clk->parent)
+ dump_clock(m, 0, clk);
+ mutex_unlock(&clocks_mutex);
+
+ return 0;
+}
+
+static int c6x_ck_open(struct inode *inode, struct file *file)
+{
+ return single_open(file, c6x_ck_show, NULL);
+}
+
+static const struct file_operations c6x_ck_operations = {
+ .open = c6x_ck_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+ .release = single_release,
+};
+
+static int __init c6x_clk_debugfs_init(void)
+{
+ debugfs_create_file("c6x_clocks", S_IFREG | S_IRUGO, NULL, NULL,
+ &c6x_ck_operations);
+
+ return 0;
+}
+device_initcall(c6x_clk_debugfs_init);
+#endif /* CONFIG_DEBUG_FS */
diff --git a/arch/c6x/platforms/plldata.c b/arch/c6x/platforms/plldata.c
new file mode 100644
index 0000000..2cfd6f4
--- /dev/null
+++ b/arch/c6x/platforms/plldata.c
@@ -0,0 +1,404 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2011 Texas Instruments Incorporated
+ * Author: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/kernel.h>
+#include <linux/delay.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <linux/ioport.h>
+#include <linux/clkdev.h>
+#include <linux/of.h>
+#include <linux/of_address.h>
+
+#include <asm/clock.h>
+#include <asm/setup.h>
+#include <asm/irq.h>
+
+/*
+ * Common SoC clock support.
+ */
+
+/* Default input for PLL1 */
+struct clk clkin1 = {
+ .name = "clkin1",
+ .node = LIST_HEAD_INIT(clkin1.node),
+ .children = LIST_HEAD_INIT(clkin1.children),
+ .childnode = LIST_HEAD_INIT(clkin1.childnode),
+};
+
+struct pll_data c6x_soc_pll1 = {
+ .num = 1,
+ .sysclks = {
+ {
+ .name = "pll1",
+ .parent = &clkin1,
+ .pll_data = &c6x_soc_pll1,
+ .flags = CLK_PLL,
+ },
+ {
+ .name = "pll1_sysclk1",
+ .parent = &c6x_soc_pll1.sysclks[0],
+ .flags = CLK_PLL,
+ },
+ {
+ .name = "pll1_sysclk2",
+ .parent = &c6x_soc_pll1.sysclks[0],
+ .flags = CLK_PLL,
+ },
+ {
+ .name = "pll1_sysclk3",
+ .parent = &c6x_soc_pll1.sysclks[0],
+ .flags = CLK_PLL,
+ },
+ {
+ .name = "pll1_sysclk4",
+ .parent = &c6x_soc_pll1.sysclks[0],
+ .flags = CLK_PLL,
+ },
+ {
+ .name = "pll1_sysclk5",
+ .parent = &c6x_soc_pll1.sysclks[0],
+ .flags = CLK_PLL,
+ },
+ {
+ .name = "pll1_sysclk6",
+ .parent = &c6x_soc_pll1.sysclks[0],
+ .flags = CLK_PLL,
+ },
+ {
+ .name = "pll1_sysclk7",
+ .parent = &c6x_soc_pll1.sysclks[0],
+ .flags = CLK_PLL,
+ },
+ {
+ .name = "pll1_sysclk8",
+ .parent = &c6x_soc_pll1.sysclks[0],
+ .flags = CLK_PLL,
+ },
+ {
+ .name = "pll1_sysclk9",
+ .parent = &c6x_soc_pll1.sysclks[0],
+ .flags = CLK_PLL,
+ },
+ {
+ .name = "pll1_sysclk10",
+ .parent = &c6x_soc_pll1.sysclks[0],
+ .flags = CLK_PLL,
+ },
+ {
+ .name = "pll1_sysclk11",
+ .parent = &c6x_soc_pll1.sysclks[0],
+ .flags = CLK_PLL,
+ },
+ {
+ .name = "pll1_sysclk12",
+ .parent = &c6x_soc_pll1.sysclks[0],
+ .flags = CLK_PLL,
+ },
+ {
+ .name = "pll1_sysclk13",
+ .parent = &c6x_soc_pll1.sysclks[0],
+ .flags = CLK_PLL,
+ },
+ {
+ .name = "pll1_sysclk14",
+ .parent = &c6x_soc_pll1.sysclks[0],
+ .flags = CLK_PLL,
+ },
+ {
+ .name = "pll1_sysclk15",
+ .parent = &c6x_soc_pll1.sysclks[0],
+ .flags = CLK_PLL,
+ },
+ {
+ .name = "pll1_sysclk16",
+ .parent = &c6x_soc_pll1.sysclks[0],
+ .flags = CLK_PLL,
+ },
+ },
+};
+
+/* CPU core clock */
+struct clk c6x_core_clk = {
+ .name = "core",
+};
+
+/* miscellaneous IO clocks */
+struct clk c6x_i2c_clk = {
+ .name = "i2c",
+};
+
+struct clk c6x_watchdog_clk = {
+ .name = "watchdog",
+};
+
+struct clk c6x_mcbsp1_clk = {
+ .name = "mcbsp1",
+};
+
+struct clk c6x_mcbsp2_clk = {
+ .name = "mcbsp2",
+};
+
+struct clk c6x_mdio_clk = {
+ .name = "mdio",
+};
+
+
+#ifdef CONFIG_SOC_TMS320C6455
+static struct clk_lookup c6455_clks[] = {
+ CLK(NULL, "pll1", &c6x_soc_pll1.sysclks[0]),
+ CLK(NULL, "pll1_sysclk2", &c6x_soc_pll1.sysclks[2]),
+ CLK(NULL, "pll1_sysclk3", &c6x_soc_pll1.sysclks[3]),
+ CLK(NULL, "pll1_sysclk4", &c6x_soc_pll1.sysclks[4]),
+ CLK(NULL, "pll1_sysclk5", &c6x_soc_pll1.sysclks[5]),
+ CLK(NULL, "core", &c6x_core_clk),
+ CLK("i2c_davinci.1", NULL, &c6x_i2c_clk),
+ CLK("watchdog", NULL, &c6x_watchdog_clk),
+ CLK("2c81800.mdio", NULL, &c6x_mdio_clk),
+ CLK("", NULL, NULL)
+};
+
+
+static void __init c6455_setup_clocks(struct device_node *node)
+{
+ struct pll_data *pll = &c6x_soc_pll1;
+ struct clk *sysclks = pll->sysclks;
+
+ pll->flags = PLL_HAS_PRE | PLL_HAS_MUL;
+
+ sysclks[2].flags |= FIXED_DIV_PLL;
+ sysclks[2].div = 3;
+ sysclks[3].flags |= FIXED_DIV_PLL;
+ sysclks[3].div = 6;
+ sysclks[4].div = PLLDIV4;
+ sysclks[5].div = PLLDIV5;
+
+ c6x_core_clk.parent = &sysclks[0];
+ c6x_i2c_clk.parent = &sysclks[3];
+ c6x_watchdog_clk.parent = &sysclks[3];
+ c6x_mdio_clk.parent = &sysclks[3];
+
+ c6x_clks_init(c6455_clks);
+}
+#endif /* CONFIG_SOC_TMS320C6455 */
+
+#ifdef CONFIG_SOC_TMS320C6457
+static struct clk_lookup c6457_clks[] = {
+ CLK(NULL, "pll1", &c6x_soc_pll1.sysclks[0]),
+ CLK(NULL, "pll1_sysclk1", &c6x_soc_pll1.sysclks[1]),
+ CLK(NULL, "pll1_sysclk2", &c6x_soc_pll1.sysclks[2]),
+ CLK(NULL, "pll1_sysclk3", &c6x_soc_pll1.sysclks[3]),
+ CLK(NULL, "pll1_sysclk4", &c6x_soc_pll1.sysclks[4]),
+ CLK(NULL, "pll1_sysclk5", &c6x_soc_pll1.sysclks[5]),
+ CLK(NULL, "core", &c6x_core_clk),
+ CLK("i2c_davinci.1", NULL, &c6x_i2c_clk),
+ CLK("watchdog", NULL, &c6x_watchdog_clk),
+ CLK("2c81800.mdio", NULL, &c6x_mdio_clk),
+ CLK("", NULL, NULL)
+};
+
+static void __init c6457_setup_clocks(struct device_node *node)
+{
+ struct pll_data *pll = &c6x_soc_pll1;
+ struct clk *sysclks = pll->sysclks;
+
+ pll->flags = PLL_HAS_MUL | PLL_HAS_POST;
+
+ sysclks[1].flags |= FIXED_DIV_PLL;
+ sysclks[1].div = 1;
+ sysclks[2].flags |= FIXED_DIV_PLL;
+ sysclks[2].div = 3;
+ sysclks[3].flags |= FIXED_DIV_PLL;
+ sysclks[3].div = 6;
+ sysclks[4].div = PLLDIV4;
+ sysclks[5].div = PLLDIV5;
+
+ c6x_core_clk.parent = &sysclks[1];
+ c6x_i2c_clk.parent = &sysclks[3];
+ c6x_watchdog_clk.parent = &sysclks[5];
+ c6x_mdio_clk.parent = &sysclks[5];
+
+ c6x_clks_init(c6457_clks);
+}
+#endif /* CONFIG_SOC_TMS320C6455 */
+
+#ifdef CONFIG_SOC_TMS320C6472
+static struct clk_lookup c6472_clks[] = {
+ CLK(NULL, "pll1", &c6x_soc_pll1.sysclks[0]),
+ CLK(NULL, "pll1_sysclk1", &c6x_soc_pll1.sysclks[1]),
+ CLK(NULL, "pll1_sysclk2", &c6x_soc_pll1.sysclks[2]),
+ CLK(NULL, "pll1_sysclk3", &c6x_soc_pll1.sysclks[3]),
+ CLK(NULL, "pll1_sysclk4", &c6x_soc_pll1.sysclks[4]),
+ CLK(NULL, "pll1_sysclk5", &c6x_soc_pll1.sysclks[5]),
+ CLK(NULL, "pll1_sysclk6", &c6x_soc_pll1.sysclks[6]),
+ CLK(NULL, "pll1_sysclk7", &c6x_soc_pll1.sysclks[7]),
+ CLK(NULL, "pll1_sysclk8", &c6x_soc_pll1.sysclks[8]),
+ CLK(NULL, "pll1_sysclk9", &c6x_soc_pll1.sysclks[9]),
+ CLK(NULL, "pll1_sysclk10", &c6x_soc_pll1.sysclks[10]),
+ CLK(NULL, "core", &c6x_core_clk),
+ CLK("i2c_davinci.1", NULL, &c6x_i2c_clk),
+ CLK("watchdog", NULL, &c6x_watchdog_clk),
+ CLK("2c81800.mdio", NULL, &c6x_mdio_clk),
+ CLK("", NULL, NULL)
+};
+
+/* assumptions used for delay loop calculations */
+#define MIN_CLKIN1_KHz 15625
+#define MAX_CORE_KHz 700000
+#define MIN_PLLOUT_KHz MIN_CLKIN1_KHz
+
+static void __init c6472_setup_clocks(struct device_node *node)
+{
+ struct pll_data *pll = &c6x_soc_pll1;
+ struct clk *sysclks = pll->sysclks;
+ int i;
+
+ pll->flags = PLL_HAS_MUL;
+
+ for (i = 1; i <= 6; i++) {
+ sysclks[i].flags |= FIXED_DIV_PLL;
+ sysclks[i].div = 1;
+ }
+
+ sysclks[7].flags |= FIXED_DIV_PLL;
+ sysclks[7].div = 3;
+ sysclks[8].flags |= FIXED_DIV_PLL;
+ sysclks[8].div = 6;
+ sysclks[9].flags |= FIXED_DIV_PLL;
+ sysclks[9].div = 2;
+ sysclks[10].div = PLLDIV10;
+
+ c6x_core_clk.parent = &sysclks[get_coreid() + 1];
+ c6x_i2c_clk.parent = &sysclks[8];
+ c6x_watchdog_clk.parent = &sysclks[8];
+ c6x_mdio_clk.parent = &sysclks[5];
+
+ c6x_clks_init(c6472_clks);
+}
+#endif /* CONFIG_SOC_TMS320C6472 */
+
+
+#ifdef CONFIG_SOC_TMS320C6474
+static struct clk_lookup c6474_clks[] = {
+ CLK(NULL, "pll1", &c6x_soc_pll1.sysclks[0]),
+ CLK(NULL, "pll1_sysclk7", &c6x_soc_pll1.sysclks[7]),
+ CLK(NULL, "pll1_sysclk9", &c6x_soc_pll1.sysclks[9]),
+ CLK(NULL, "pll1_sysclk10", &c6x_soc_pll1.sysclks[10]),
+ CLK(NULL, "pll1_sysclk11", &c6x_soc_pll1.sysclks[11]),
+ CLK(NULL, "pll1_sysclk12", &c6x_soc_pll1.sysclks[12]),
+ CLK(NULL, "pll1_sysclk13", &c6x_soc_pll1.sysclks[13]),
+ CLK(NULL, "core", &c6x_core_clk),
+ CLK("i2c_davinci.1", NULL, &c6x_i2c_clk),
+ CLK("mcbsp.1", NULL, &c6x_mcbsp1_clk),
+ CLK("mcbsp.2", NULL, &c6x_mcbsp2_clk),
+ CLK("watchdog", NULL, &c6x_watchdog_clk),
+ CLK("2c81800.mdio", NULL, &c6x_mdio_clk),
+ CLK("", NULL, NULL)
+};
+
+static void __init c6474_setup_clocks(struct device_node *node)
+{
+ struct pll_data *pll = &c6x_soc_pll1;
+ struct clk *sysclks = pll->sysclks;
+
+ pll->flags = PLL_HAS_MUL;
+
+ sysclks[7].flags |= FIXED_DIV_PLL;
+ sysclks[7].div = 1;
+ sysclks[9].flags |= FIXED_DIV_PLL;
+ sysclks[9].div = 3;
+ sysclks[10].flags |= FIXED_DIV_PLL;
+ sysclks[10].div = 6;
+
+ sysclks[11].div = PLLDIV11;
+
+ sysclks[12].flags |= FIXED_DIV_PLL;
+ sysclks[12].div = 2;
+
+ sysclks[13].div = PLLDIV13;
+
+ c6x_core_clk.parent = &sysclks[7];
+ c6x_i2c_clk.parent = &sysclks[10];
+ c6x_watchdog_clk.parent = &sysclks[10];
+ c6x_mcbsp1_clk.parent = &sysclks[10];
+ c6x_mcbsp2_clk.parent = &sysclks[10];
+
+ c6x_clks_init(c6474_clks);
+}
+#endif /* CONFIG_SOC_TMS320C6474 */
+
+static struct of_device_id c6x_clkc_match[] __initdata = {
+#ifdef CONFIG_SOC_TMS320C6455
+ { .compatible = "ti,c6455-pll", .data = c6455_setup_clocks },
+#endif
+#ifdef CONFIG_SOC_TMS320C6457
+ { .compatible = "ti,c6457-pll", .data = c6457_setup_clocks },
+#endif
+#ifdef CONFIG_SOC_TMS320C6472
+ { .compatible = "ti,c6472-pll", .data = c6472_setup_clocks },
+#endif
+#ifdef CONFIG_SOC_TMS320C6474
+ { .compatible = "ti,c6474-pll", .data = c6474_setup_clocks },
+#endif
+ { .compatible = "ti,c64x+pll" },
+ {}
+};
+
+void __init c64x_setup_clocks(void)
+{
+ void (*__setup_clocks)(struct device_node *np);
+ struct pll_data *pll = &c6x_soc_pll1;
+ struct device_node *node;
+ const struct of_device_id *id;
+ int err;
+ u32 val;
+
+ node = of_find_matching_node(NULL, c6x_clkc_match);
+ if (!node)
+ return;
+
+ pll->base = of_iomap(node, 0);
+ if (!pll->base)
+ goto out;
+
+ err = of_property_read_u32(node, "clock-frequency", &val);
+ if (err || val == 0) {
+ pr_err("%s: no clock-frequency found! Using %dMHz\n",
+ node->full_name, (int)val / 1000000);
+ val = 25000000;
+ }
+ clkin1.rate = val;
+
+ err = of_property_read_u32(node, "ti,c64x+pll-bypass-delay", &val);
+ if (err)
+ val = 5000;
+ pll->bypass_delay = val;
+
+ err = of_property_read_u32(node, "ti,c64x+pll-reset-delay", &val);
+ if (err)
+ val = 30000;
+ pll->reset_delay = val;
+
+ err = of_property_read_u32(node, "ti,c64x+pll-lock-delay", &val);
+ if (err)
+ val = 30000;
+ pll->lock_delay = val;
+
+ /* id->data is a pointer to SoC-specific setup */
+ id = of_match_node(c6x_clkc_match, node);
+ if (id && id->data) {
+ __setup_clocks = id->data;
+ __setup_clocks(node);
+ }
+
+out:
+ of_node_put(node);
+}
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v3 15/24] C6X: clocks
2011-09-27 20:29 ` [PATCH v3 15/24] C6X: clocks Mark Salter
@ 2011-09-28 13:51 ` Arnd Bergmann
0 siblings, 0 replies; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-28 13:51 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch
On Tuesday 27 September 2011, Mark Salter wrote:
> The C6X SoCs contain several PLL controllers each with up to 16 clock outputs
> feeding into the cores or peripheral clock domains. The hardware is very similar
> to arm/mach-davinci clocks. This is still a work in progress which needs to be
> updated once device tree clock binding changes shake out.
>
> Signed-off-by: Mark Salter <msalter@redhat.com>
Yes, this will have to be reworked at some point, but it's ok for now
as long as there is no generic code to help you do this better.
Acked-by: Arnd Bergmann <arnd@arndb.de>
^ permalink raw reply [flat|nested] 66+ messages in thread
* [PATCH v3 16/24] C6X: cache control
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (14 preceding siblings ...)
2011-09-27 20:29 ` [PATCH v3 15/24] C6X: clocks Mark Salter
@ 2011-09-27 20:29 ` Mark Salter
2011-09-28 13:52 ` Arnd Bergmann
2011-09-27 20:29 ` [PATCH v3 17/24] C6X: loadable module support Mark Salter
` (9 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:29 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
Signed-off-by: Mark Salter <msalter@redhat.com>
---
arch/c6x/include/asm/cache.h | 90 ++++++++
arch/c6x/include/asm/cacheflush.h | 65 ++++++
arch/c6x/platforms/cache.c | 445 +++++++++++++++++++++++++++++++++++++
3 files changed, 600 insertions(+), 0 deletions(-)
create mode 100644 arch/c6x/include/asm/cache.h
create mode 100644 arch/c6x/include/asm/cacheflush.h
create mode 100644 arch/c6x/platforms/cache.c
diff --git a/arch/c6x/include/asm/cache.h b/arch/c6x/include/asm/cache.h
new file mode 100644
index 0000000..6d521d9
--- /dev/null
+++ b/arch/c6x/include/asm/cache.h
@@ -0,0 +1,90 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2005, 2006, 2009, 2010 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_CACHE_H
+#define _ASM_C6X_CACHE_H
+
+#include <linux/irqflags.h>
+
+/*
+ * Cache line size
+ */
+#define L1D_CACHE_BYTES 64
+#define L1P_CACHE_BYTES 32
+#define L2_CACHE_BYTES 128
+
+/*
+ * L2 used as cache
+ */
+#define L2MODE_SIZE L2MODE_256K_CACHE
+
+/*
+ * For practical reasons the L1_CACHE_BYTES defines should not be smaller than
+ * the L2 line size
+ */
+#define L1_CACHE_BYTES L2_CACHE_BYTES
+
+#define L2_CACHE_ALIGN_LOW(x) \
+ (((x) & ~(L2_CACHE_BYTES - 1)))
+#define L2_CACHE_ALIGN_UP(x) \
+ (((x) + (L2_CACHE_BYTES - 1)) & ~(L2_CACHE_BYTES - 1))
+#define L2_CACHE_ALIGN_CNT(x) \
+ (((x) + (sizeof(int) - 1)) & ~(sizeof(int) - 1))
+
+#define ARCH_DMA_MINALIGN L1_CACHE_BYTES
+#define ARCH_SLAB_MINALIGN L1_CACHE_BYTES
+
+/*
+ * This is the granularity of hardware cacheability control.
+ */
+#define CACHEABILITY_ALIGN 0x01000000
+
+/*
+ * Align a physical address to MAR regions
+ */
+#define CACHE_REGION_START(v) \
+ (((u32) (v)) & ~(CACHEABILITY_ALIGN - 1))
+#define CACHE_REGION_END(v) \
+ (((u32) (v) + (CACHEABILITY_ALIGN - 1)) & ~(CACHEABILITY_ALIGN - 1))
+
+extern void __init c6x_cache_init(void);
+
+extern void enable_caching(unsigned long start, unsigned long end);
+extern void disable_caching(unsigned long start, unsigned long end);
+
+extern void L1_cache_off(void);
+extern void L1_cache_on(void);
+
+extern void L1P_cache_global_invalidate(void);
+extern void L1D_cache_global_invalidate(void);
+extern void L1D_cache_global_writeback(void);
+extern void L1D_cache_global_writeback_invalidate(void);
+extern void L2_cache_set_mode(unsigned int mode);
+extern void L2_cache_global_writeback_invalidate(void);
+extern void L2_cache_global_writeback(void);
+
+extern void L1P_cache_block_invalidate(unsigned int start, unsigned int end);
+extern void L1D_cache_block_invalidate(unsigned int start, unsigned int end);
+extern void L1D_cache_block_writeback_invalidate(unsigned int start,
+ unsigned int end);
+extern void L1D_cache_block_writeback(unsigned int start, unsigned int end);
+extern void L2_cache_block_invalidate(unsigned int start, unsigned int end);
+extern void L2_cache_block_writeback(unsigned int start, unsigned int end);
+extern void L2_cache_block_writeback_invalidate(unsigned int start,
+ unsigned int end);
+extern void L2_cache_block_invalidate_nowait(unsigned int start,
+ unsigned int end);
+extern void L2_cache_block_writeback_nowait(unsigned int start,
+ unsigned int end);
+
+extern void L2_cache_block_writeback_invalidate_nowait(unsigned int start,
+ unsigned int end);
+
+#endif /* _ASM_C6X_CACHE_H */
diff --git a/arch/c6x/include/asm/cacheflush.h b/arch/c6x/include/asm/cacheflush.h
new file mode 100644
index 0000000..df5db90
--- /dev/null
+++ b/arch/c6x/include/asm/cacheflush.h
@@ -0,0 +1,65 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_CACHEFLUSH_H
+#define _ASM_C6X_CACHEFLUSH_H
+
+#include <linux/spinlock.h>
+
+#include <asm/setup.h>
+#include <asm/cache.h>
+#include <asm/mman.h>
+#include <asm/page.h>
+#include <asm/string.h>
+
+/*
+ * virtually-indexed cache management (our cache is physically indexed)
+ */
+#define flush_cache_all() do {} while (0)
+#define flush_cache_mm(mm) do {} while (0)
+#define flush_cache_dup_mm(mm) do {} while (0)
+#define flush_cache_range(mm, start, end) do {} while (0)
+#define flush_cache_page(vma, vmaddr, pfn) do {} while (0)
+#define flush_cache_vmap(start, end) do {} while (0)
+#define flush_cache_vunmap(start, end) do {} while (0)
+#define ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE 0
+#define flush_dcache_page(page) do {} while (0)
+#define flush_dcache_mmap_lock(mapping) do {} while (0)
+#define flush_dcache_mmap_unlock(mapping) do {} while (0)
+
+/*
+ * physically-indexed cache management
+ */
+#define flush_icache_range(s, e) \
+do { \
+ L1D_cache_block_writeback((s), (e)); \
+ L1P_cache_block_invalidate((s), (e)); \
+} while (0)
+
+#define flush_icache_page(vma, page) \
+do { \
+ if ((vma)->vm_flags & PROT_EXEC) \
+ L1D_cache_block_writeback_invalidate(page_address(page), \
+ (unsigned long) page_address(page) + PAGE_SIZE)); \
+ L1P_cache_block_invalidate(page_address(page), \
+ (unsigned long) page_address(page) + PAGE_SIZE)); \
+} while (0)
+
+
+#define copy_to_user_page(vma, page, vaddr, dst, src, len) \
+do { \
+ memcpy(dst, src, len); \
+ flush_icache_range((unsigned) (dst), (unsigned) (dst) + (len)); \
+} while (0)
+
+#define copy_from_user_page(vma, page, vaddr, dst, src, len) \
+ memcpy(dst, src, len)
+
+#endif /* _ASM_C6X_CACHEFLUSH_H */
diff --git a/arch/c6x/platforms/cache.c b/arch/c6x/platforms/cache.c
new file mode 100644
index 0000000..86318a1
--- /dev/null
+++ b/arch/c6x/platforms/cache.c
@@ -0,0 +1,445 @@
+/*
+ * Copyright (C) 2011 Texas Instruments Incorporated
+ * Author: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/of.h>
+#include <linux/of_address.h>
+#include <linux/io.h>
+
+#include <asm/cache.h>
+#include <asm/soc.h>
+
+/*
+ * Internal Memory Control Registers for caches
+ */
+#define IMCR_CCFG 0x0000
+#define IMCR_L1PCFG 0x0020
+#define IMCR_L1PCC 0x0024
+#define IMCR_L1DCFG 0x0040
+#define IMCR_L1DCC 0x0044
+#define IMCR_L2ALLOC0 0x2000
+#define IMCR_L2ALLOC1 0x2004
+#define IMCR_L2ALLOC2 0x2008
+#define IMCR_L2ALLOC3 0x200c
+#define IMCR_L2WBAR 0x4000
+#define IMCR_L2WWC 0x4004
+#define IMCR_L2WIBAR 0x4010
+#define IMCR_L2WIWC 0x4014
+#define IMCR_L2IBAR 0x4018
+#define IMCR_L2IWC 0x401c
+#define IMCR_L1PIBAR 0x4020
+#define IMCR_L1PIWC 0x4024
+#define IMCR_L1DWIBAR 0x4030
+#define IMCR_L1DWIWC 0x4034
+#define IMCR_L1DWBAR 0x4040
+#define IMCR_L1DWWC 0x4044
+#define IMCR_L1DIBAR 0x4048
+#define IMCR_L1DIWC 0x404c
+#define IMCR_L2WB 0x5000
+#define IMCR_L2WBINV 0x5004
+#define IMCR_L2INV 0x5008
+#define IMCR_L1PINV 0x5028
+#define IMCR_L1DWB 0x5040
+#define IMCR_L1DWBINV 0x5044
+#define IMCR_L1DINV 0x5048
+#define IMCR_MAR_BASE 0x8000
+#define IMCR_MAR96_111 0x8180
+#define IMCR_MAR128_191 0x8200
+#define IMCR_MAR224_239 0x8380
+#define IMCR_L2MPFAR 0xa000
+#define IMCR_L2MPFSR 0xa004
+#define IMCR_L2MPFCR 0xa008
+#define IMCR_L2MPLK0 0xa100
+#define IMCR_L2MPLK1 0xa104
+#define IMCR_L2MPLK2 0xa108
+#define IMCR_L2MPLK3 0xa10c
+#define IMCR_L2MPLKCMD 0xa110
+#define IMCR_L2MPLKSTAT 0xa114
+#define IMCR_L2MPPA_BASE 0xa200
+#define IMCR_L1PMPFAR 0xa400
+#define IMCR_L1PMPFSR 0xa404
+#define IMCR_L1PMPFCR 0xa408
+#define IMCR_L1PMPLK0 0xa500
+#define IMCR_L1PMPLK1 0xa504
+#define IMCR_L1PMPLK2 0xa508
+#define IMCR_L1PMPLK3 0xa50c
+#define IMCR_L1PMPLKCMD 0xa510
+#define IMCR_L1PMPLKSTAT 0xa514
+#define IMCR_L1PMPPA_BASE 0xa600
+#define IMCR_L1DMPFAR 0xac00
+#define IMCR_L1DMPFSR 0xac04
+#define IMCR_L1DMPFCR 0xac08
+#define IMCR_L1DMPLK0 0xad00
+#define IMCR_L1DMPLK1 0xad04
+#define IMCR_L1DMPLK2 0xad08
+#define IMCR_L1DMPLK3 0xad0c
+#define IMCR_L1DMPLKCMD 0xad10
+#define IMCR_L1DMPLKSTAT 0xad14
+#define IMCR_L1DMPPA_BASE 0xae00
+#define IMCR_L2PDWAKE0 0xc040
+#define IMCR_L2PDWAKE1 0xc044
+#define IMCR_L2PDSLEEP0 0xc050
+#define IMCR_L2PDSLEEP1 0xc054
+#define IMCR_L2PDSTAT0 0xc060
+#define IMCR_L2PDSTAT1 0xc064
+
+/*
+ * CCFG register values and bits
+ */
+#define L2MODE_0K_CACHE 0x0
+#define L2MODE_32K_CACHE 0x1
+#define L2MODE_64K_CACHE 0x2
+#define L2MODE_128K_CACHE 0x3
+#define L2MODE_256K_CACHE 0x7
+
+#define L2PRIO_URGENT 0x0
+#define L2PRIO_HIGH 0x1
+#define L2PRIO_MEDIUM 0x2
+#define L2PRIO_LOW 0x3
+
+#define CCFG_ID 0x100 /* Invalidate L1P bit */
+#define CCFG_IP 0x200 /* Invalidate L1D bit */
+
+static void __iomem *cache_base;
+
+/*
+ * L1 & L2 caches generic functions
+ */
+#define imcr_get(reg) soc_readl(cache_base + (reg))
+#define imcr_set(reg, value) \
+do { \
+ soc_writel((value), cache_base + (reg)); \
+ soc_readl(cache_base + (reg)); \
+} while (0)
+
+static void cache_block_operation_wait(unsigned int wc_reg)
+{
+ /* Wait for completion */
+ while (imcr_get(wc_reg))
+ cpu_relax();
+}
+
+static DEFINE_SPINLOCK(cache_lock);
+
+/*
+ * Generic function to perform a block cache operation as
+ * invalidate or writeback/invalidate
+ */
+static void cache_block_operation(unsigned int *start,
+ unsigned int *end,
+ unsigned int bar_reg,
+ unsigned int wc_reg)
+{
+ unsigned long flags;
+ unsigned int wcnt =
+ (L2_CACHE_ALIGN_CNT((unsigned int) end)
+ - L2_CACHE_ALIGN_LOW((unsigned int) start)) >> 2;
+ unsigned int wc = 0;
+
+ for (; wcnt; wcnt -= wc, start += wc) {
+loop:
+ spin_lock_irqsave(&cache_lock, flags);
+
+ /*
+ * If another cache operation is occuring
+ */
+ if (unlikely(imcr_get(wc_reg))) {
+ spin_unlock_irqrestore(&cache_lock, flags);
+
+ /* Wait for previous operation completion */
+ cache_block_operation_wait(wc_reg);
+
+ /* Try again */
+ goto loop;
+ }
+
+ imcr_set(bar_reg, L2_CACHE_ALIGN_LOW((unsigned int) start));
+
+ if (wcnt > 0xffff)
+ wc = 0xffff;
+ else
+ wc = wcnt;
+
+ /* Set word count value in the WC register */
+ imcr_set(wc_reg, wc & 0xffff);
+
+ spin_unlock_irqrestore(&cache_lock, flags);
+
+ /* Wait for completion */
+ cache_block_operation_wait(wc_reg);
+ }
+}
+
+static void cache_block_operation_nowait(unsigned int *start,
+ unsigned int *end,
+ unsigned int bar_reg,
+ unsigned int wc_reg)
+{
+ unsigned long flags;
+ unsigned int wcnt =
+ (L2_CACHE_ALIGN_CNT((unsigned int) end)
+ - L2_CACHE_ALIGN_LOW((unsigned int) start)) >> 2;
+ unsigned int wc = 0;
+
+ for (; wcnt; wcnt -= wc, start += wc) {
+
+ spin_lock_irqsave(&cache_lock, flags);
+
+ imcr_set(bar_reg, L2_CACHE_ALIGN_LOW((unsigned int) start));
+
+ if (wcnt > 0xffff)
+ wc = 0xffff;
+ else
+ wc = wcnt;
+
+ /* Set word count value in the WC register */
+ imcr_set(wc_reg, wc & 0xffff);
+
+ spin_unlock_irqrestore(&cache_lock, flags);
+
+ /* Don't wait for completion on last cache operation */
+ if (wcnt > 0xffff)
+ cache_block_operation_wait(wc_reg);
+ }
+}
+
+/*
+ * L1 caches management
+ */
+
+/*
+ * Disable L1 caches
+ */
+void L1_cache_off(void)
+{
+ unsigned int dummy;
+
+ imcr_set(IMCR_L1PCFG, 0);
+ dummy = imcr_get(IMCR_L1PCFG);
+
+ imcr_set(IMCR_L1DCFG, 0);
+ dummy = imcr_get(IMCR_L1DCFG);
+}
+
+/*
+ * Enable L1 caches
+ */
+void L1_cache_on(void)
+{
+ unsigned int dummy;
+
+ imcr_set(IMCR_L1PCFG, 7);
+ dummy = imcr_get(IMCR_L1PCFG);
+
+ imcr_set(IMCR_L1DCFG, 7);
+ dummy = imcr_get(IMCR_L1DCFG);
+}
+
+/*
+ * L1P global-invalidate all
+ */
+void L1P_cache_global_invalidate(void)
+{
+ unsigned int set = 1;
+ imcr_set(IMCR_L1PINV, set);
+ while (imcr_get(IMCR_L1PINV) & 1)
+ cpu_relax();
+}
+
+/*
+ * L1D global-invalidate all
+ *
+ * Warning: this operation causes all updated data in L1D to
+ * be discarded rather than written back to the lower levels of
+ * memory
+ */
+void L1D_cache_global_invalidate(void)
+{
+ unsigned int set = 1;
+ imcr_set(IMCR_L1DINV, set);
+ while (imcr_get(IMCR_L1DINV) & 1)
+ cpu_relax();
+}
+
+void L1D_cache_global_writeback(void)
+{
+ unsigned int set = 1;
+ imcr_set(IMCR_L1DWB, set);
+ while (imcr_get(IMCR_L1DWB) & 1)
+ cpu_relax();
+}
+
+void L1D_cache_global_writeback_invalidate(void)
+{
+ unsigned int set = 1;
+ imcr_set(IMCR_L1DWBINV, set);
+ while (imcr_get(IMCR_L1DWBINV) & 1)
+ cpu_relax();
+}
+
+/*
+ * L2 caches management
+ */
+
+/*
+ * Set L2 operation mode
+ */
+void L2_cache_set_mode(unsigned int mode)
+{
+ unsigned int ccfg = imcr_get(IMCR_CCFG);
+
+ /* Clear and set the L2MODE bits in CCFG */
+ ccfg &= ~7;
+ ccfg |= (mode & 7);
+ imcr_set(IMCR_CCFG, ccfg);
+ ccfg = imcr_get(IMCR_CCFG);
+}
+
+/*
+ * L2 global-writeback and global-invalidate all
+ */
+void L2_cache_global_writeback_invalidate(void)
+{
+ imcr_set(IMCR_L2WBINV, 1);
+ while (imcr_get(IMCR_L2WBINV))
+ cpu_relax();
+}
+
+/*
+ * L2 global-writeback all
+ */
+void L2_cache_global_writeback(void)
+{
+ imcr_set(IMCR_L2WB, 1);
+ while (imcr_get(IMCR_L2WB))
+ cpu_relax();
+}
+
+/*
+ * Cacheability controls
+ */
+void enable_caching(unsigned long start, unsigned long end)
+{
+ unsigned int mar = IMCR_MAR_BASE + ((start >> 24) << 2);
+ unsigned int mar_e = IMCR_MAR_BASE + ((end >> 24) << 2);
+
+ for (; mar <= mar_e; mar += 4)
+ imcr_set(mar, imcr_get(mar) | 1);
+}
+
+void disable_caching(unsigned long start, unsigned long end)
+{
+ unsigned int mar = IMCR_MAR_BASE + ((start >> 24) << 2);
+ unsigned int mar_e = IMCR_MAR_BASE + ((end >> 24) << 2);
+
+ for (; mar <= mar_e; mar += 4)
+ imcr_set(mar, imcr_get(mar) & ~1);
+}
+
+
+/*
+ * L1 block operations
+ */
+void L1P_cache_block_invalidate(unsigned int start, unsigned int end)
+{
+ cache_block_operation((unsigned int *) start,
+ (unsigned int *) end,
+ IMCR_L1PIBAR, IMCR_L1PIWC);
+}
+
+void L1D_cache_block_invalidate(unsigned int start, unsigned int end)
+{
+ cache_block_operation((unsigned int *) start,
+ (unsigned int *) end,
+ IMCR_L1DIBAR, IMCR_L1DIWC);
+}
+
+void L1D_cache_block_writeback_invalidate(unsigned int start, unsigned int end)
+{
+ cache_block_operation((unsigned int *) start,
+ (unsigned int *) end,
+ IMCR_L1DWIBAR, IMCR_L1DWIWC);
+}
+
+void L1D_cache_block_writeback(unsigned int start, unsigned int end)
+{
+ cache_block_operation((unsigned int *) start,
+ (unsigned int *) end,
+ IMCR_L1DWBAR, IMCR_L1DWWC);
+}
+
+/*
+ * L2 block operations
+ */
+void L2_cache_block_invalidate(unsigned int start, unsigned int end)
+{
+ cache_block_operation((unsigned int *) start,
+ (unsigned int *) end,
+ IMCR_L2IBAR, IMCR_L2IWC);
+}
+
+void L2_cache_block_writeback(unsigned int start, unsigned int end)
+{
+ cache_block_operation((unsigned int *) start,
+ (unsigned int *) end,
+ IMCR_L2WBAR, IMCR_L2WWC);
+}
+
+void L2_cache_block_writeback_invalidate(unsigned int start, unsigned int end)
+{
+ cache_block_operation((unsigned int *) start,
+ (unsigned int *) end,
+ IMCR_L2WIBAR, IMCR_L2WIWC);
+}
+
+void L2_cache_block_invalidate_nowait(unsigned int start, unsigned int end)
+{
+ cache_block_operation_nowait((unsigned int *) start,
+ (unsigned int *) end,
+ IMCR_L2IBAR, IMCR_L2IWC);
+}
+
+void L2_cache_block_writeback_nowait(unsigned int start, unsigned int end)
+{
+ cache_block_operation_nowait((unsigned int *) start,
+ (unsigned int *) end,
+ IMCR_L2WBAR, IMCR_L2WWC);
+}
+
+void L2_cache_block_writeback_invalidate_nowait(unsigned int start,
+ unsigned int end)
+{
+ cache_block_operation_nowait((unsigned int *) start,
+ (unsigned int *) end,
+ IMCR_L2WIBAR, IMCR_L2WIWC);
+}
+
+
+/*
+ * L1 and L2 caches configuration
+ */
+void __init c6x_cache_init(void)
+{
+ struct device_node *node;
+
+ node = of_find_compatible_node(NULL, NULL, "ti,c64x+cache");
+ if (!node)
+ return;
+
+ cache_base = of_iomap(node, 0);
+
+ of_node_put(node);
+
+ if (!cache_base)
+ return;
+
+ /* Set L2 caches on the the whole L2 SRAM memory */
+ L2_cache_set_mode(L2MODE_SIZE);
+
+ /* Enable L1 */
+ L1_cache_on();
+}
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v3 16/24] C6X: cache control
2011-09-27 20:29 ` [PATCH v3 16/24] C6X: cache control Mark Salter
@ 2011-09-28 13:52 ` Arnd Bergmann
0 siblings, 0 replies; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-28 13:52 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch
On Tuesday 27 September 2011, Mark Salter wrote:
>
> Signed-off-by: Mark Salter <msalter@redhat.com>
> ---
> arch/c6x/include/asm/cache.h | 90 ++++++++
> arch/c6x/include/asm/cacheflush.h | 65 ++++++
> arch/c6x/platforms/cache.c | 445 +++++++++++++++++++++++++++++++++++++
> 3 files changed, 600 insertions(+), 0 deletions(-)
> create mode 100644 arch/c6x/include/asm/cache.h
> create mode 100644 arch/c6x/include/asm/cacheflush.h
> create mode 100644 arch/c6x/platforms/cache.c
Acked-by: Arnd Bergmann <arnd@arndb.de>
^ permalink raw reply [flat|nested] 66+ messages in thread
* [PATCH v3 17/24] C6X: loadable module support
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (15 preceding siblings ...)
2011-09-27 20:29 ` [PATCH v3 16/24] C6X: cache control Mark Salter
@ 2011-09-27 20:29 ` Mark Salter
2011-09-28 13:54 ` Arnd Bergmann
2011-09-27 20:29 ` [PATCH v3 18/24] C6X: ptrace support Mark Salter
` (8 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:29 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
Signed-off-by: Mark Salter <msalter@redhat.com>
---
arch/c6x/include/asm/module.h | 33 +++++++++++
arch/c6x/kernel/c6x_ksyms.c | 68 +++++++++++++++++++++
arch/c6x/kernel/module.c | 129 +++++++++++++++++++++++++++++++++++++++++
3 files changed, 230 insertions(+), 0 deletions(-)
create mode 100644 arch/c6x/include/asm/module.h
create mode 100644 arch/c6x/kernel/c6x_ksyms.c
create mode 100644 arch/c6x/kernel/module.c
diff --git a/arch/c6x/include/asm/module.h b/arch/c6x/include/asm/module.h
new file mode 100644
index 0000000..a453f97
--- /dev/null
+++ b/arch/c6x/include/asm/module.h
@@ -0,0 +1,33 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * Updated for 2.6.34 by: Mark Salter (msalter@redhat.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_MODULE_H
+#define _ASM_C6X_MODULE_H
+
+#define Elf_Shdr Elf32_Shdr
+#define Elf_Sym Elf32_Sym
+#define Elf_Ehdr Elf32_Ehdr
+#define Elf_Addr Elf32_Addr
+#define Elf_Word Elf32_Word
+
+/*
+ * This file contains the C6x architecture specific module code.
+ */
+struct mod_arch_specific {
+};
+
+struct loaded_sections {
+ unsigned int new_vaddr;
+ unsigned int loaded;
+};
+
+#endif /* _ASM_C6X_MODULE_H */
diff --git a/arch/c6x/kernel/c6x_ksyms.c b/arch/c6x/kernel/c6x_ksyms.c
new file mode 100644
index 0000000..358d097
--- /dev/null
+++ b/arch/c6x/kernel/c6x_ksyms.c
@@ -0,0 +1,68 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010, 2011 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ */
+#include <linux/module.h>
+#include <asm/checksum.h>
+#include <linux/io.h>
+
+/*
+ * libgcc functions - used internally by the compiler...
+ */
+extern int __c6xabi_divi(int dividend, int divisor);
+EXPORT_SYMBOL(__c6xabi_divi);
+
+extern unsigned __c6xabi_divu(unsigned dividend, unsigned divisor);
+EXPORT_SYMBOL(__c6xabi_divu);
+
+extern int __c6xabi_remi(int dividend, int divisor);
+EXPORT_SYMBOL(__c6xabi_remi);
+
+extern unsigned __c6xabi_remu(unsigned dividend, unsigned divisor);
+EXPORT_SYMBOL(__c6xabi_remu);
+
+extern int __c6xabi_divremi(int dividend, int divisor);
+EXPORT_SYMBOL(__c6xabi_divremi);
+
+extern unsigned __c6xabi_divremu(unsigned dividend, unsigned divisor);
+EXPORT_SYMBOL(__c6xabi_divremu);
+
+extern unsigned long long __c6xabi_mpyll(unsigned long long src1,
+ unsigned long long src2);
+EXPORT_SYMBOL(__c6xabi_mpyll);
+
+extern long long __c6xabi_negll(long long src);
+EXPORT_SYMBOL(__c6xabi_negll);
+
+extern unsigned long long __c6xabi_llshl(unsigned long long src1, uint src2);
+EXPORT_SYMBOL(__c6xabi_llshl);
+
+extern long long __c6xabi_llshr(long long src1, uint src2);
+EXPORT_SYMBOL(__c6xabi_llshr);
+
+extern unsigned long long __c6xabi_llshru(unsigned long long src1, uint src2);
+EXPORT_SYMBOL(__c6xabi_llshru);
+
+extern void __c6xabi_strasgi(int *dst, const int *src, unsigned cnt);
+EXPORT_SYMBOL(__c6xabi_strasgi);
+
+extern void __c6xabi_push_rts(void);
+EXPORT_SYMBOL(__c6xabi_push_rts);
+
+extern void __c6xabi_pop_rts(void);
+EXPORT_SYMBOL(__c6xabi_pop_rts);
+
+extern void __c6xabi_strasgi_64plus(int *dst, const int *src, unsigned cnt);
+EXPORT_SYMBOL(__c6xabi_strasgi_64plus);
+
+/* lib functions */
+EXPORT_SYMBOL(memcpy);
+EXPORT_SYMBOL(csum_partial);
+EXPORT_SYMBOL(ip_fast_csum);
diff --git a/arch/c6x/kernel/module.c b/arch/c6x/kernel/module.c
new file mode 100644
index 0000000..ba7159d
--- /dev/null
+++ b/arch/c6x/kernel/module.c
@@ -0,0 +1,129 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2005, 2009, 2010, 2011 Texas Instruments Incorporated
+ * Author: Thomas Charleux (thomas.charleux@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ */
+#include <linux/moduleloader.h>
+#include <linux/elf.h>
+#include <linux/vmalloc.h>
+#include <linux/kernel.h>
+
+#ifdef DEBUG
+#define DBG(fmt...) pr_debug(fmt)
+#else
+#define DBG(fmt...)
+#endif
+
+static inline int fixup_pcr(u32 *ip, Elf32_Addr dest, u32 maskbits, int shift)
+{
+ u32 opcode;
+ long ep = (long)ip & ~31;
+ long delta = ((long)dest - ep) >> 2;
+ long mask = (1 << maskbits) - 1;
+
+ if ((delta >> (maskbits - 1)) == 0 ||
+ (delta >> (maskbits - 1)) == -1) {
+ opcode = *ip;
+ opcode &= ~(mask << shift);
+ opcode |= ((delta & mask) << shift);
+ *ip = opcode;
+
+ DBG("REL PCR_S%d[%p] dest[0p] opcode[%08x]\n",
+ maskbits, ip, (void *)dest, opcode);
+
+ return 0;
+ }
+ pr_err("PCR_S%d reloc %p -> %p out of range!\n",
+ maskbits, ip, (void *)dest);
+
+ return -1;
+}
+
+/*
+ * apply a RELA relocation
+ */
+int apply_relocate_add(Elf32_Shdr *sechdrs,
+ const char *strtab,
+ unsigned int symindex,
+ unsigned int relsec,
+ struct module *me)
+{
+ Elf32_Rela *rel = (void *) sechdrs[relsec].sh_addr;
+ Elf_Sym *sym;
+ u32 *location, opcode;
+ unsigned int i;
+ Elf32_Addr v;
+ Elf_Addr offset = 0;
+
+ DBG("Applying relocate section %u to %u with offset 0x%x\n", relsec,
+ sechdrs[relsec].sh_info, offset);
+
+ for (i = 0; i < sechdrs[relsec].sh_size / sizeof(*rel); i++) {
+ /* This is where to make the change */
+ location = (void *)sechdrs[sechdrs[relsec].sh_info].sh_addr
+ + rel[i].r_offset - offset;
+
+ /* This is the symbol it is referring to. Note that all
+ undefined symbols have been resolved. */
+ sym = (Elf_Sym *)sechdrs[symindex].sh_addr
+ + ELF32_R_SYM(rel[i].r_info);
+
+ /* this is the adjustment to be made */
+ v = sym->st_value + rel[i].r_addend;
+
+ switch (ELF32_R_TYPE(rel[i].r_info)) {
+ case R_C6000_ABS32:
+ DBG("RELA ABS32: [%p] = 0x%x\n", location, v);
+ *location = v;
+ break;
+ case R_C6000_ABS16:
+ DBG("RELA ABS16: [%p] = 0x%x\n", location, v);
+ *(u16 *)location = v;
+ break;
+ case R_C6000_ABS8:
+ DBG("RELA ABS8: [%p] = 0x%x\n", location, v);
+ *(u8 *)location = v;
+ break;
+ case R_C6000_ABS_L16:
+ opcode = *location;
+ opcode &= ~0x7fff80;
+ opcode |= ((v & 0xffff) << 7);
+ DBG("RELA ABS_L16[%p] v[0x%x] opcode[0x%x]\n",
+ location, v, opcode);
+ *location = opcode;
+ break;
+ case R_C6000_ABS_H16:
+ opcode = *location;
+ opcode &= ~0x7fff80;
+ opcode |= ((v >> 9) & 0x7fff80);
+ DBG("RELA ABS_H16[%p] v[0x%x] opcode[0x%x]\n",
+ location, v, opcode);
+ *location = opcode;
+ break;
+ case R_C6000_PCR_S21:
+ if (fixup_pcr(location, v, 21, 7))
+ return -ENOEXEC;
+ break;
+ case R_C6000_PCR_S12:
+ if (fixup_pcr(location, v, 12, 16))
+ return -ENOEXEC;
+ break;
+ case R_C6000_PCR_S10:
+ if (fixup_pcr(location, v, 10, 13))
+ return -ENOEXEC;
+ break;
+ default:
+ pr_err("module %s: Unknown RELA relocation: %u\n",
+ me->name, ELF32_R_TYPE(rel[i].r_info));
+ return -ENOEXEC;
+ }
+ }
+
+ return 0;
+}
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v3 17/24] C6X: loadable module support
2011-09-27 20:29 ` [PATCH v3 17/24] C6X: loadable module support Mark Salter
@ 2011-09-28 13:54 ` Arnd Bergmann
0 siblings, 0 replies; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-28 13:54 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch
On Tuesday 27 September 2011, Mark Salter wrote:
>
> Signed-off-by: Mark Salter <msalter@redhat.com>
> ---
> arch/c6x/include/asm/module.h | 33 +++++++++++
> arch/c6x/kernel/c6x_ksyms.c | 68 +++++++++++++++++++++
> arch/c6x/kernel/module.c | 129 +++++++++++++++++++++++++++++++++++++++++
> 3 files changed, 230 insertions(+), 0 deletions(-)
> create mode 100644 arch/c6x/include/asm/module.h
> create mode 100644 arch/c6x/kernel/c6x_ksyms.c
> create mode 100644 arch/c6x/kernel/module.c
Acked-by: Arnd Bergmann <arnd@arndb.de>
> +
> +#ifdef DEBUG
> +#define DBG(fmt...) pr_debug(fmt)
> +#else
> +#define DBG(fmt...)
> +#endif
Note that this is unnecesary. pr_debug is already conditional.
Arnd
^ permalink raw reply [flat|nested] 66+ messages in thread
* [PATCH v3 18/24] C6X: ptrace support
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (16 preceding siblings ...)
2011-09-27 20:29 ` [PATCH v3 17/24] C6X: loadable module support Mark Salter
@ 2011-09-27 20:29 ` Mark Salter
2011-09-28 13:57 ` Arnd Bergmann
2011-09-27 20:30 ` [PATCH v3 19/24] C6X: headers Mark Salter
` (7 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:29 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
Signed-off-by: Mark Salter <msalter@redhat.com>
---
arch/c6x/include/asm/ptrace.h | 177 +++++++++++++++++++++++++++++++++++
arch/c6x/kernel/ptrace.c | 207 +++++++++++++++++++++++++++++++++++++++++
2 files changed, 384 insertions(+), 0 deletions(-)
create mode 100644 arch/c6x/include/asm/ptrace.h
create mode 100644 arch/c6x/kernel/ptrace.c
diff --git a/arch/c6x/include/asm/ptrace.h b/arch/c6x/include/asm/ptrace.h
new file mode 100644
index 0000000..e637a33
--- /dev/null
+++ b/arch/c6x/include/asm/ptrace.h
@@ -0,0 +1,177 @@
+/*
+ * Copyright (C) 2004, 2006, 2009, 2010 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * Updated for 2.6.34: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_PTRACE_H
+#define _ASM_C6X_PTRACE_H
+
+#define BKPT_OPCODE 0x56454314 /* illegal opcode */
+
+#ifdef _BIG_ENDIAN
+#define PT_LO(odd, even) odd
+#define PT_HI(odd, even) even
+#else
+#define PT_LO(odd, even) even
+#define PT_HI(odd, even) odd
+#endif
+
+#define PT_A4_ORG PT_LO(1, 0)
+#define PT_TSR PT_HI(1, 0)
+#define PT_ILC PT_LO(3, 2)
+#define PT_RILC PT_HI(3, 2)
+#define PT_CSR PT_LO(5, 4)
+#define PT_PC PT_HI(5, 4)
+#define PT_B16 PT_LO(7, 6)
+#define PT_B17 PT_HI(7, 6)
+#define PT_B18 PT_LO(9, 8)
+#define PT_B19 PT_HI(9, 8)
+#define PT_B20 PT_LO(11, 10)
+#define PT_B21 PT_HI(11, 10)
+#define PT_B22 PT_LO(13, 12)
+#define PT_B23 PT_HI(13, 12)
+#define PT_B24 PT_LO(15, 14)
+#define PT_B25 PT_HI(15, 14)
+#define PT_B26 PT_LO(17, 16)
+#define PT_B27 PT_HI(17, 16)
+#define PT_B28 PT_LO(19, 18)
+#define PT_B29 PT_HI(19, 18)
+#define PT_B30 PT_LO(21, 20)
+#define PT_B31 PT_HI(21, 20)
+#define PT_B0 PT_LO(23, 22)
+#define PT_B1 PT_HI(23, 22)
+#define PT_B2 PT_LO(25, 24)
+#define PT_B3 PT_HI(25, 24)
+#define PT_B4 PT_LO(27, 26)
+#define PT_B5 PT_HI(27, 26)
+#define PT_B6 PT_LO(29, 28)
+#define PT_B7 PT_HI(29, 28)
+#define PT_B8 PT_LO(31, 30)
+#define PT_B9 PT_HI(31, 30)
+#define PT_B10 PT_LO(33, 32)
+#define PT_B11 PT_HI(33, 32)
+#define PT_B12 PT_LO(35, 34)
+#define PT_B13 PT_HI(35, 34)
+#define PT_A16 PT_LO(37, 36)
+#define PT_A17 PT_HI(37, 36)
+#define PT_A18 PT_LO(39, 38)
+#define PT_A19 PT_HI(39, 38)
+#define PT_A20 PT_LO(41, 40)
+#define PT_A21 PT_HI(41, 40)
+#define PT_A22 PT_LO(43, 42)
+#define PT_A23 PT_HI(43, 42)
+#define PT_A24 PT_LO(45, 44)
+#define PT_A25 PT_HI(45, 44)
+#define PT_A26 PT_LO(47, 46)
+#define PT_A27 PT_HI(47, 46)
+#define PT_A28 PT_LO(49, 48)
+#define PT_A29 PT_HI(49, 48)
+#define PT_A30 PT_LO(51, 50)
+#define PT_A31 PT_HI(51, 50)
+#define PT_A0 PT_LO(53, 52)
+#define PT_A1 PT_HI(53, 52)
+#define PT_A2 PT_LO(55, 54)
+#define PT_A3 PT_HI(55, 54)
+#define PT_A4 PT_LO(57, 56)
+#define PT_A5 PT_HI(57, 56)
+#define PT_A6 PT_LO(59, 58)
+#define PT_A7 PT_HI(59, 58)
+#define PT_A8 PT_LO(61, 60)
+#define PT_A9 PT_HI(61, 60)
+#define PT_A10 PT_LO(63, 62)
+#define PT_A11 PT_HI(63, 62)
+#define PT_A12 PT_LO(65, 64)
+#define PT_A13 PT_HI(65, 64)
+#define PT_A14 PT_LO(67, 66)
+#define PT_A15 PT_HI(67, 66)
+#define PT_B14 PT_LO(69, 68)
+#define PT_B15 PT_HI(69, 68)
+
+#define NR_PTREGS 70
+
+#define PT_DP PT_B14 /* Data Segment Pointer (B14) */
+#define PT_SP PT_B15 /* Stack Pointer (B15) */
+
+#ifndef __ASSEMBLY__
+
+#ifdef _BIG_ENDIAN
+#define REG_PAIR(odd, even) unsigned long odd; unsigned long even
+#else
+#define REG_PAIR(odd, even) unsigned long even; unsigned long odd
+#endif
+
+/*
+ * this struct defines the way the registers are stored on the
+ * stack during a system call. fields defined with REG_PAIR
+ * are saved and restored using double-word memory operations
+ * which means the word ordering of the pair depends on endianess.
+ */
+struct pt_regs {
+ REG_PAIR(tsr, orig_a4);
+ REG_PAIR(rilc, ilc);
+ REG_PAIR(pc, csr);
+
+ REG_PAIR(b17, b16);
+ REG_PAIR(b19, b18);
+ REG_PAIR(b21, b20);
+ REG_PAIR(b23, b22);
+ REG_PAIR(b25, b24);
+ REG_PAIR(b27, b26);
+ REG_PAIR(b29, b28);
+ REG_PAIR(b31, b30);
+
+ REG_PAIR(b1, b0);
+ REG_PAIR(b3, b2);
+ REG_PAIR(b5, b4);
+ REG_PAIR(b7, b6);
+ REG_PAIR(b9, b8);
+ REG_PAIR(b11, b10);
+ REG_PAIR(b13, b12);
+
+ REG_PAIR(a17, a16);
+ REG_PAIR(a19, a18);
+ REG_PAIR(a21, a20);
+ REG_PAIR(a23, a22);
+ REG_PAIR(a25, a24);
+ REG_PAIR(a27, a26);
+ REG_PAIR(a29, a28);
+ REG_PAIR(a31, a30);
+
+ REG_PAIR(a1, a0);
+ REG_PAIR(a3, a2);
+ REG_PAIR(a5, a4);
+ REG_PAIR(a7, a6);
+ REG_PAIR(a9, a8);
+ REG_PAIR(a11, a10);
+ REG_PAIR(a13, a12);
+
+ REG_PAIR(a15, a14);
+ REG_PAIR(sp, dp);
+};
+
+#define PTRACE_GETREGS 12
+#define PTRACE_SETREGS 13
+#define PTRACE_GETFPREGS 14
+#define PTRACE_SETFPREGS 15
+
+#ifdef __KERNEL__
+
+#include <linux/linkage.h>
+
+#define user_mode(regs) ((((regs)->tsr) & 0x40) != 0)
+
+#define instruction_pointer(regs) ((regs)->pc)
+#define profile_pc(regs) instruction_pointer(regs)
+extern void show_regs(struct pt_regs *);
+
+extern asmlinkage unsigned long syscall_trace_entry(struct pt_regs *regs);
+extern asmlinkage void syscall_trace_exit(struct pt_regs *regs);
+
+#endif /* __KERNEL__ */
+#endif /* __ASSEMBLY__ */
+#endif /* _ASM_C6X_PTRACE_H */
diff --git a/arch/c6x/kernel/ptrace.c b/arch/c6x/kernel/ptrace.c
new file mode 100644
index 0000000..96fd4a5
--- /dev/null
+++ b/arch/c6x/kernel/ptrace.c
@@ -0,0 +1,207 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2006, 2009, 2010, 2011 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * Updated for 2.6.34: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/ptrace.h>
+#include <linux/tracehook.h>
+#include <linux/regset.h>
+#include <linux/elf.h>
+
+#include <asm/cacheflush.h>
+
+#define PT_REG_SIZE (sizeof(struct pt_regs))
+
+/*
+ * Called by kernel/ptrace.c when detaching.
+ */
+void ptrace_disable(struct task_struct *child)
+{
+ /* nothing to do */
+}
+
+/*
+ * Get a register number from live pt_regs for the specified task.
+ */
+static inline long get_reg(struct task_struct *task, int regno)
+{
+ long *addr = (long *)task_pt_regs(task);
+
+ if (regno == PT_TSR || regno == PT_CSR)
+ return 0;
+
+ return addr[regno];
+}
+
+/*
+ * Write contents of register REGNO in task TASK.
+ */
+static inline int put_reg(struct task_struct *task,
+ int regno,
+ unsigned long data)
+{
+ unsigned long *addr = (unsigned long *)task_pt_regs(task);
+
+ if (regno != PT_TSR && regno != PT_CSR)
+ addr[regno] = data;
+
+ return 0;
+}
+
+/* regset get/set implementations */
+
+static int gpr_get(struct task_struct *target,
+ const struct user_regset *regset,
+ unsigned int pos, unsigned int count,
+ void *kbuf, void __user *ubuf)
+{
+ struct pt_regs *regs = task_pt_regs(target);
+
+ return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
+ regs,
+ 0, sizeof(*regs));
+}
+
+static int gpr_set(struct task_struct *target,
+ const struct user_regset *regset,
+ unsigned int pos, unsigned int count,
+ const void *kbuf, const void __user *ubuf)
+{
+ int ret;
+ struct pt_regs *regs = task_pt_regs(target);
+
+ /* Don't copyin TSR or CSR */
+ ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
+ ®s,
+ 0, PT_TSR * sizeof(long));
+ if (ret)
+ return ret;
+
+ ret = user_regset_copyin_ignore(&pos, &count, &kbuf, &ubuf,
+ PT_TSR * sizeof(long),
+ (PT_TSR + 1) * sizeof(long));
+ if (ret)
+ return ret;
+
+ ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
+ ®s,
+ (PT_TSR + 1) * sizeof(long),
+ PT_CSR * sizeof(long));
+ if (ret)
+ return ret;
+
+ ret = user_regset_copyin_ignore(&pos, &count, &kbuf, &ubuf,
+ PT_CSR * sizeof(long),
+ (PT_CSR + 1) * sizeof(long));
+ if (ret)
+ return ret;
+
+ ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
+ ®s,
+ (PT_CSR + 1) * sizeof(long), -1);
+ return ret;
+}
+
+enum c6x_regset {
+ REGSET_GPR,
+};
+
+static const struct user_regset c6x_regsets[] = {
+ [REGSET_GPR] = {
+ .core_note_type = NT_PRSTATUS,
+ .n = ELF_NGREG,
+ .size = sizeof(u32),
+ .align = sizeof(u32),
+ .get = gpr_get,
+ .set = gpr_set
+ },
+};
+
+static const struct user_regset_view user_c6x_native_view = {
+ .name = "tic6x",
+ .e_machine = EM_TI_C6000,
+ .regsets = c6x_regsets,
+ .n = ARRAY_SIZE(c6x_regsets),
+};
+
+const struct user_regset_view *task_user_regset_view(struct task_struct *task)
+{
+ return &user_c6x_native_view;
+}
+
+/*
+ * Perform ptrace request
+ */
+long arch_ptrace(struct task_struct *child, long request,
+ unsigned long addr, unsigned long data)
+{
+ unsigned long __user *datap = (unsigned long __user *) data;
+ int ret = 0;
+
+ switch (request) {
+ /*
+ * write the word at location addr.
+ */
+ case PTRACE_POKETEXT:
+ ret = generic_ptrace_pokedata(child, addr, data);
+ if (ret == 0 && request == PTRACE_POKETEXT)
+ flush_icache_range(addr, addr + 4);
+ break;
+
+ /*
+ * get all gp regs from the child.
+ */
+ case PTRACE_GETREGS:
+ return copy_regset_to_user(child, &user_c6x_native_view,
+ REGSET_GPR,
+ 0, NR_PTREGS * sizeof(long),
+ datap);
+
+ /*
+ * set all gp regs in the child.
+ */
+ case PTRACE_SETREGS:
+ return copy_regset_from_user(child, &user_c6x_native_view,
+ REGSET_GPR,
+ 0, NR_PTREGS * sizeof(long),
+ datap);
+
+ default:
+ ret = ptrace_request(child, request, addr, data);
+ break;
+ }
+
+ return ret;
+}
+
+/*
+ * handle tracing of system call entry
+ * - return the revised system call number or ULONG_MAX to cause ENOSYS
+ */
+asmlinkage unsigned long syscall_trace_entry(struct pt_regs *regs)
+{
+ if (tracehook_report_syscall_entry(regs))
+ /* tracing decided this syscall should not happen, so
+ * We'll return a bogus call number to get an ENOSYS
+ * error, but leave the original number in
+ * regs->orig_a4
+ */
+ return ULONG_MAX;
+
+ return regs->b0;
+}
+
+/*
+ * handle tracing of system call exit
+ */
+asmlinkage void syscall_trace_exit(struct pt_regs *regs)
+{
+ tracehook_report_syscall_exit(regs, 0);
+}
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v3 18/24] C6X: ptrace support
2011-09-27 20:29 ` [PATCH v3 18/24] C6X: ptrace support Mark Salter
@ 2011-09-28 13:57 ` Arnd Bergmann
0 siblings, 0 replies; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-28 13:57 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch
On Tuesday 27 September 2011, Mark Salter wrote:
> + case PTRACE_GETREGS:
> + return copy_regset_to_user(child, &user_c6x_native_view,
> + REGSET_GPR,
> + 0, NR_PTREGS * sizeof(long),
> + datap);
> +
> + /*
> + * set all gp regs in the child.
> + */
> + case PTRACE_SETREGS:
> + return copy_regset_from_user(child, &user_c6x_native_view,
> + REGSET_GPR,
> + 0, NR_PTREGS * sizeof(long),
> + datap);
> +
I believe that you can let this be handled by the common
ptrace_request() function if you provide the necessary code for
setting CONFIG_HAVE_ARCH_TRACEHOOK, which is preferred now.
Arnd
^ permalink raw reply [flat|nested] 66+ messages in thread
* [PATCH v3 19/24] C6X: headers
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (17 preceding siblings ...)
2011-09-27 20:29 ` [PATCH v3 18/24] C6X: ptrace support Mark Salter
@ 2011-09-27 20:30 ` Mark Salter
2011-09-28 14:07 ` Arnd Bergmann
2011-09-27 20:30 ` [PATCH v3 20/24] C6X: library code Mark Salter
` (6 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:30 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
Signed-off-by: Mark Salter <msalter@redhat.com>
---
arch/c6x/include/asm/asm-offsets.h | 1 +
arch/c6x/include/asm/bitops.h | 105 +++++++++++++
arch/c6x/include/asm/byteorder.h | 12 ++
arch/c6x/include/asm/delay.h | 67 +++++++++
arch/c6x/include/asm/elf.h | 113 ++++++++++++++
arch/c6x/include/asm/ftrace.h | 6 +
arch/c6x/include/asm/linkage.h | 30 ++++
arch/c6x/include/asm/memblock.h | 4 +
arch/c6x/include/asm/mmu.h | 18 +++
arch/c6x/include/asm/mutex.h | 6 +
arch/c6x/include/asm/page.h | 11 ++
arch/c6x/include/asm/pgtable.h | 81 ++++++++++
arch/c6x/include/asm/procinfo.h | 28 ++++
arch/c6x/include/asm/prom.h | 1 +
arch/c6x/include/asm/sections.h | 12 ++
arch/c6x/include/asm/setup.h | 32 ++++
arch/c6x/include/asm/string.h | 21 +++
arch/c6x/include/asm/system.h | 168 +++++++++++++++++++++
arch/c6x/include/asm/tlb.h | 8 +
arch/c6x/include/asm/uaccess.h | 107 +++++++++++++
arch/c6x/include/asm/unaligned.h | 288 ++++++++++++++++++++++++++++++++++++
21 files changed, 1119 insertions(+), 0 deletions(-)
create mode 100644 arch/c6x/include/asm/asm-offsets.h
create mode 100644 arch/c6x/include/asm/bitops.h
create mode 100644 arch/c6x/include/asm/byteorder.h
create mode 100644 arch/c6x/include/asm/delay.h
create mode 100644 arch/c6x/include/asm/elf.h
create mode 100644 arch/c6x/include/asm/ftrace.h
create mode 100644 arch/c6x/include/asm/linkage.h
create mode 100644 arch/c6x/include/asm/memblock.h
create mode 100644 arch/c6x/include/asm/mmu.h
create mode 100644 arch/c6x/include/asm/mutex.h
create mode 100644 arch/c6x/include/asm/page.h
create mode 100644 arch/c6x/include/asm/pgtable.h
create mode 100644 arch/c6x/include/asm/procinfo.h
create mode 100644 arch/c6x/include/asm/prom.h
create mode 100644 arch/c6x/include/asm/sections.h
create mode 100644 arch/c6x/include/asm/setup.h
create mode 100644 arch/c6x/include/asm/string.h
create mode 100644 arch/c6x/include/asm/system.h
create mode 100644 arch/c6x/include/asm/tlb.h
create mode 100644 arch/c6x/include/asm/uaccess.h
create mode 100644 arch/c6x/include/asm/unaligned.h
diff --git a/arch/c6x/include/asm/asm-offsets.h b/arch/c6x/include/asm/asm-offsets.h
new file mode 100644
index 0000000..d370ee3
--- /dev/null
+++ b/arch/c6x/include/asm/asm-offsets.h
@@ -0,0 +1 @@
+#include <generated/asm-offsets.h>
diff --git a/arch/c6x/include/asm/bitops.h b/arch/c6x/include/asm/bitops.h
new file mode 100644
index 0000000..39ab7e8
--- /dev/null
+++ b/arch/c6x/include/asm/bitops.h
@@ -0,0 +1,105 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_BITOPS_H
+#define _ASM_C6X_BITOPS_H
+
+#ifdef __KERNEL__
+
+#include <linux/bitops.h>
+
+#include <asm/system.h>
+#include <asm/byteorder.h>
+
+/*
+ * clear_bit() doesn't provide any barrier for the compiler.
+ */
+#define smp_mb__before_clear_bit() barrier()
+#define smp_mb__after_clear_bit() barrier()
+
+/*
+ * We are lucky, DSP is perfect for bitops: do it in 3 cycles
+ */
+
+/**
+ * __ffs - find first bit in word.
+ * @word: The word to search
+ *
+ * Undefined if no bit exists, so code should check against 0 first.
+ * Note __ffs(0) = undef, __ffs(1) = 0, __ffs(0x80000000) = 31.
+ *
+ */
+static inline unsigned long __ffs(unsigned long x)
+{
+ asm (" bitr .M1 %0,%0\n"
+ " nop\n"
+ " lmbd .L1 1,%0,%0\n"
+ : "+a"(x));
+
+ return x;
+}
+
+/*
+ * ffz - find first zero in word.
+ * @word: The word to search
+ *
+ * Undefined if no zero exists, so code should check against ~0UL first.
+ */
+#define ffz(x) __ffs(~(x))
+
+/**
+ * fls - find last (most-significant) bit set
+ * @x: the word to search
+ *
+ * This is defined the same way as ffs.
+ * Note fls(0) = 0, fls(1) = 1, fls(0x80000000) = 32.
+ */
+static inline int fls(int x)
+{
+ if (!x)
+ return 0;
+
+ asm (" lmbd .L1 1,%0,%0\n" : "+a"(x));
+
+ return 32 - x;
+}
+
+/**
+ * ffs - find first bit set
+ * @x: the word to search
+ *
+ * This is defined the same way as
+ * the libc and compiler builtin ffs routines, therefore
+ * differs in spirit from the above ffz (man ffs).
+ * Note ffs(0) = 0, ffs(1) = 1, ffs(0x80000000) = 32.
+ */
+static inline int ffs(int x)
+{
+ if (!x)
+ return 0;
+
+ return __ffs(x) + 1;
+}
+
+#include <asm-generic/bitops/__fls.h>
+#include <asm-generic/bitops/fls64.h>
+#include <asm-generic/bitops/find.h>
+
+#include <asm-generic/bitops/sched.h>
+#include <asm-generic/bitops/hweight.h>
+#include <asm-generic/bitops/lock.h>
+
+#include <asm-generic/bitops/atomic.h>
+#include <asm-generic/bitops/non-atomic.h>
+#include <asm-generic/bitops/le.h>
+#include <asm-generic/bitops/ext2-atomic.h>
+
+#endif /* __KERNEL__ */
+#endif /* _ASM_C6X_BITOPS_H */
diff --git a/arch/c6x/include/asm/byteorder.h b/arch/c6x/include/asm/byteorder.h
new file mode 100644
index 0000000..166038d
--- /dev/null
+++ b/arch/c6x/include/asm/byteorder.h
@@ -0,0 +1,12 @@
+#ifndef _ASM_C6X_BYTEORDER_H
+#define _ASM_C6X_BYTEORDER_H
+
+#include <asm/types.h>
+
+#ifdef _BIG_ENDIAN
+#include <linux/byteorder/big_endian.h>
+#else /* _BIG_ENDIAN */
+#include <linux/byteorder/little_endian.h>
+#endif /* _BIG_ENDIAN */
+
+#endif /* _ASM_BYTEORDER_H */
diff --git a/arch/c6x/include/asm/delay.h b/arch/c6x/include/asm/delay.h
new file mode 100644
index 0000000..f314c2e
--- /dev/null
+++ b/arch/c6x/include/asm/delay.h
@@ -0,0 +1,67 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010, 2011 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_DELAY_H
+#define _ASM_C6X_DELAY_H
+
+#include <linux/kernel.h>
+
+extern unsigned int ticks_per_ns_scaled;
+
+static inline void __delay(unsigned long loops)
+{
+ uint32_t tmp;
+
+ /* 6 cycles per loop */
+ asm volatile (" mv .s1 %0,%1\n"
+ "0: [%1] b .s1 0b\n"
+ " add .l1 -6,%0,%0\n"
+ " cmplt .l1 1,%0,%1\n"
+ " nop 3\n"
+ : "+a"(loops), "=A"(tmp));
+}
+
+static inline void _c6x_tickdelay(unsigned int x)
+{
+ uint32_t cnt, endcnt;
+
+ asm volatile (" mvc .s2 TSCL,%0\n"
+ " add .s2x %0,%1,%2\n"
+ " || mvk .l2 1,B0\n"
+ "0: [B0] b .s2 0b\n"
+ " mvc .s2 TSCL,%0\n"
+ " sub .s2 %0,%2,%0\n"
+ " cmpgt .l2 0,%0,B0\n"
+ " nop 2\n"
+ : "=b"(cnt), "+a"(x), "=b"(endcnt) : : "B0");
+}
+
+/* use scaled math to avoid slow division */
+#define C6X_NDELAY_SCALE 10
+
+static inline void _ndelay(unsigned int n)
+{
+ _c6x_tickdelay((ticks_per_ns_scaled * n) >> C6X_NDELAY_SCALE);
+}
+
+static inline void _udelay(unsigned int n)
+{
+ while (n >= 10) {
+ _ndelay(10000);
+ n -= 10;
+ }
+ while (n-- > 0)
+ _ndelay(1000);
+}
+
+#define udelay(x) _udelay((unsigned int)(x))
+#define ndelay(x) _ndelay((unsigned int)(x))
+
+#endif /* _ASM_C6X_DELAY_H */
diff --git a/arch/c6x/include/asm/elf.h b/arch/c6x/include/asm/elf.h
new file mode 100644
index 0000000..d57865b
--- /dev/null
+++ b/arch/c6x/include/asm/elf.h
@@ -0,0 +1,113 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_ELF_H
+#define _ASM_C6X_ELF_H
+
+/*
+ * ELF register definitions..
+ */
+#include <asm/ptrace.h>
+
+typedef unsigned long elf_greg_t;
+typedef unsigned long elf_fpreg_t;
+
+#define ELF_NGREG 58
+#define ELF_NFPREG 1
+
+typedef elf_greg_t elf_gregset_t[ELF_NGREG];
+typedef elf_fpreg_t elf_fpregset_t[ELF_NFPREG];
+
+/*
+ * This is used to ensure we don't load something for the wrong architecture.
+ */
+#define elf_check_arch(x) ((x)->e_machine == EM_TI_C6000)
+
+#define elf_check_const_displacement(x) (1)
+
+/*
+ * These are used to set parameters in the core dumps.
+ */
+#ifdef __LITTLE_ENDIAN__
+#define ELF_DATA ELFDATA2LSB
+#else
+#define ELF_DATA ELFDATA2MSB
+#endif
+
+#define ELF_CLASS ELFCLASS32
+#define ELF_ARCH EM_TI_C6000
+
+/* Nothing for now. Need to setup DP... */
+#define ELF_PLAT_INIT(_r)
+
+#define USE_ELF_CORE_DUMP
+#define ELF_EXEC_PAGESIZE 4096
+
+#define ELF_CORE_COPY_REGS(_dest, _regs) \
+ memcpy((char *) &_dest, (char *) _regs, \
+ sizeof(struct pt_regs));
+
+/* This yields a mask that user programs can use to figure out what
+ instruction set this cpu supports. */
+
+#define ELF_HWCAP (0)
+
+/* This yields a string that ld.so will use to load implementation
+ specific libraries for optimization. This is more specific in
+ intent than poking at uname or /proc/cpuinfo. */
+
+#define ELF_PLATFORM (NULL)
+
+#define SET_PERSONALITY(ex) set_personality(PER_LINUX)
+
+/* C6X specific section types */
+#define SHT_C6000_UNWIND 0x70000001
+#define SHT_C6000_PREEMPTMAP 0x70000002
+#define SHT_C6000_ATTRIBUTES 0x70000003
+
+/* C6X specific DT_ tags */
+#define DT_C6000_DSBT_BASE 0x70000000
+#define DT_C6000_DSBT_SIZE 0x70000001
+#define DT_C6000_PREEMPTMAP 0x70000002
+#define DT_C6000_DSBT_INDEX 0x70000003
+
+/* C6X specific relocs */
+#define R_C6000_NONE 0
+#define R_C6000_ABS32 1
+#define R_C6000_ABS16 2
+#define R_C6000_ABS8 3
+#define R_C6000_PCR_S21 4
+#define R_C6000_PCR_S12 5
+#define R_C6000_PCR_S10 6
+#define R_C6000_PCR_S7 7
+#define R_C6000_ABS_S16 8
+#define R_C6000_ABS_L16 9
+#define R_C6000_ABS_H16 10
+#define R_C6000_SBR_U15_B 11
+#define R_C6000_SBR_U15_H 12
+#define R_C6000_SBR_U15_W 13
+#define R_C6000_SBR_S16 14
+#define R_C6000_SBR_L16_B 15
+#define R_C6000_SBR_L16_H 16
+#define R_C6000_SBR_L16_W 17
+#define R_C6000_SBR_H16_B 18
+#define R_C6000_SBR_H16_H 19
+#define R_C6000_SBR_H16_W 20
+#define R_C6000_SBR_GOT_U15_W 21
+#define R_C6000_SBR_GOT_L16_W 22
+#define R_C6000_SBR_GOT_H16_W 23
+#define R_C6000_DSBT_INDEX 24
+#define R_C6000_PREL31 25
+#define R_C6000_COPY 26
+#define R_C6000_ALIGN 253
+#define R_C6000_FPHEAD 254
+#define R_C6000_NOCMP 255
+
+#endif /*_ASM_C6X_ELF_H */
diff --git a/arch/c6x/include/asm/ftrace.h b/arch/c6x/include/asm/ftrace.h
new file mode 100644
index 0000000..3701958
--- /dev/null
+++ b/arch/c6x/include/asm/ftrace.h
@@ -0,0 +1,6 @@
+#ifndef _ASM_C6X_FTRACE_H
+#define _ASM_C6X_FTRACE_H
+
+/* empty */
+
+#endif /* _ASM_C6X_FTRACE_H */
diff --git a/arch/c6x/include/asm/linkage.h b/arch/c6x/include/asm/linkage.h
new file mode 100644
index 0000000..376925c
--- /dev/null
+++ b/arch/c6x/include/asm/linkage.h
@@ -0,0 +1,30 @@
+#ifndef _ASM_C6X_LINKAGE_H
+#define _ASM_C6X_LINKAGE_H
+
+#ifdef __ASSEMBLER__
+
+#define __ALIGN .align 2
+#define __ALIGN_STR ".align 2"
+
+#ifndef __DSBT__
+#define ENTRY(name) \
+ .global name @ \
+ __ALIGN @ \
+name:
+#else
+#define ENTRY(name) \
+ .global name @ \
+ .hidden name @ \
+ __ALIGN @ \
+name:
+#endif
+
+#define ENDPROC(name) \
+ .type name, @function @ \
+ .size name, . - name
+
+#endif
+
+#include <asm-generic/linkage.h>
+
+#endif /* _ASM_C6X_LINKAGE_H */
diff --git a/arch/c6x/include/asm/memblock.h b/arch/c6x/include/asm/memblock.h
new file mode 100644
index 0000000..1181a97
--- /dev/null
+++ b/arch/c6x/include/asm/memblock.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_C6X_MEMBLOCK_H
+#define _ASM_C6X_MEMBLOCK_H
+
+#endif /* _ASM_C6X_MEMBLOCK_H */
diff --git a/arch/c6x/include/asm/mmu.h b/arch/c6x/include/asm/mmu.h
new file mode 100644
index 0000000..41592bf
--- /dev/null
+++ b/arch/c6x/include/asm/mmu.h
@@ -0,0 +1,18 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_MMU_H
+#define _ASM_C6X_MMU_H
+
+typedef struct {
+ unsigned long end_brk;
+} mm_context_t;
+
+#endif /* _ASM_C6X_MMU_H */
diff --git a/arch/c6x/include/asm/mutex.h b/arch/c6x/include/asm/mutex.h
new file mode 100644
index 0000000..7a7248e
--- /dev/null
+++ b/arch/c6x/include/asm/mutex.h
@@ -0,0 +1,6 @@
+#ifndef _ASM_C6X_MUTEX_H
+#define _ASM_C6X_MUTEX_H
+
+#include <asm-generic/mutex-null.h>
+
+#endif /* _ASM_C6X_MUTEX_H */
diff --git a/arch/c6x/include/asm/page.h b/arch/c6x/include/asm/page.h
new file mode 100644
index 0000000..d18e2b0
--- /dev/null
+++ b/arch/c6x/include/asm/page.h
@@ -0,0 +1,11 @@
+#ifndef _ASM_C6X_PAGE_H
+#define _ASM_C6X_PAGE_H
+
+#define VM_DATA_DEFAULT_FLAGS \
+ (VM_READ | VM_WRITE | \
+ ((current->personality & READ_IMPLIES_EXEC) ? VM_EXEC : 0) | \
+ VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC)
+
+#include <asm-generic/page.h>
+
+#endif /* _ASM_C6X_PAGE_H */
diff --git a/arch/c6x/include/asm/pgtable.h b/arch/c6x/include/asm/pgtable.h
new file mode 100644
index 0000000..68c8af4
--- /dev/null
+++ b/arch/c6x/include/asm/pgtable.h
@@ -0,0 +1,81 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_PGTABLE_H
+#define _ASM_C6X_PGTABLE_H
+
+#include <asm-generic/4level-fixup.h>
+
+#include <asm/setup.h>
+#include <asm/page.h>
+
+/*
+ * All 32bit addresses are effectively valid for vmalloc...
+ * Sort of meaningless for non-VM targets.
+ */
+#define VMALLOC_START 0
+#define VMALLOC_END 0xffffffff
+
+#define pgd_present(pgd) (1)
+#define pgd_none(pgd) (0)
+#define pgd_bad(pgd) (0)
+#define pgd_clear(pgdp)
+#define kern_addr_valid(addr) (1)
+
+#define pmd_offset(a, b) ((void *)0)
+#define pmd_none(x) (!pmd_val(x))
+#define pmd_present(x) (pmd_val(x))
+#define pmd_clear(xp) do { set_pmd(xp, __pmd(0)); } while (0)
+#define pmd_bad(x) (pmd_val(x) & ~PAGE_MASK)
+
+#define PAGE_NONE __pgprot(0) /* these mean nothing to NO_MM */
+#define PAGE_SHARED __pgprot(0) /* these mean nothing to NO_MM */
+#define PAGE_COPY __pgprot(0) /* these mean nothing to NO_MM */
+#define PAGE_READONLY __pgprot(0) /* these mean nothing to NO_MM */
+#define PAGE_KERNEL __pgprot(0) /* these mean nothing to NO_MM */
+#define pgprot_noncached(prot) (prot)
+
+extern void paging_init(void);
+
+#define __swp_type(x) (0)
+#define __swp_offset(x) (0)
+#define __swp_entry(typ, off) ((swp_entry_t) { ((typ) | ((off) << 7)) })
+#define __pte_to_swp_entry(pte) ((swp_entry_t) { pte_val(pte) })
+#define __swp_entry_to_pte(x) ((pte_t) { (x).val })
+
+static inline int pte_file(pte_t pte)
+{
+ return 0;
+}
+
+#define set_pte(pteptr, pteval) (*(pteptr) = pteval)
+#define set_pte_at(mm, addr, ptep, pteval) set_pte(ptep, pteval)
+
+/*
+ * ZERO_PAGE is a global shared page that is always zero: used
+ * for zero-mapped memory areas etc..
+ */
+#define ZERO_PAGE(vaddr) virt_to_page(empty_zero_page)
+extern unsigned long empty_zero_page;
+
+#define swapper_pg_dir ((pgd_t *) 0)
+
+/*
+ * No page table caches to initialise
+ */
+#define pgtable_cache_init() do { } while (0)
+#define io_remap_pfn_range remap_pfn_range
+
+#define io_remap_page_range(vma, vaddr, paddr, size, prot) \
+ remap_pfn_range(vma, vaddr, (paddr) >> PAGE_SHIFT, size, prot)
+
+#include <asm-generic/pgtable.h>
+
+#endif /* _ASM_C6X_PGTABLE_H */
diff --git a/arch/c6x/include/asm/procinfo.h b/arch/c6x/include/asm/procinfo.h
new file mode 100644
index 0000000..c139d1e
--- /dev/null
+++ b/arch/c6x/include/asm/procinfo.h
@@ -0,0 +1,28 @@
+/*
+ * Copyright (C) 2010 Texas Instruments Incorporated
+ * Author: Mark Salter (msalter@redhat.com)
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_PROCINFO_H
+#define _ASM_C6X_PROCINFO_H
+
+#ifdef __KERNEL__
+
+struct proc_info_list {
+ unsigned int cpu_val;
+ unsigned int cpu_mask;
+ const char *arch_name;
+ const char *elf_name;
+ unsigned int elf_hwcap;
+};
+
+#else /* __KERNEL__ */
+#include <asm/elf.h>
+#warning "Please include asm/elf.h instead"
+#endif /* __KERNEL__ */
+
+#endif /* _ASM_C6X_PROCINFO_H */
diff --git a/arch/c6x/include/asm/prom.h b/arch/c6x/include/asm/prom.h
new file mode 100644
index 0000000..b4ec95f
--- /dev/null
+++ b/arch/c6x/include/asm/prom.h
@@ -0,0 +1 @@
+/* dummy prom.h; here to make linux/of.h's #includes happy */
diff --git a/arch/c6x/include/asm/sections.h b/arch/c6x/include/asm/sections.h
new file mode 100644
index 0000000..f703989
--- /dev/null
+++ b/arch/c6x/include/asm/sections.h
@@ -0,0 +1,12 @@
+#ifndef _ASM_C6X_SECTIONS_H
+#define _ASM_C6X_SECTIONS_H
+
+#include <asm-generic/sections.h>
+
+extern char _vectors_start[];
+extern char _vectors_end[];
+
+extern char _data_lma[];
+extern char _fdt_start[], _fdt_end[];
+
+#endif /* _ASM_C6X_SECTIONS_H */
diff --git a/arch/c6x/include/asm/setup.h b/arch/c6x/include/asm/setup.h
new file mode 100644
index 0000000..1808f27
--- /dev/null
+++ b/arch/c6x/include/asm/setup.h
@@ -0,0 +1,32 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010 2011 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_SETUP_H
+#define _ASM_C6X_SETUP_H
+
+#define COMMAND_LINE_SIZE 1024
+
+#ifndef __ASSEMBLY__
+extern char c6x_command_line[COMMAND_LINE_SIZE];
+
+extern int c6x_add_memory(phys_addr_t start, unsigned long size);
+
+extern unsigned long ram_start;
+extern unsigned long ram_end;
+
+extern int c6x_num_cores;
+extern unsigned int c6x_silicon_rev;
+extern unsigned int c6x_devstat;
+extern unsigned char c6x_fuse_mac[6];
+
+extern void machine_init(unsigned long dt_ptr);
+
+#endif /* !__ASSEMBLY__ */
+#endif /* _ASM_C6X_SETUP_H */
diff --git a/arch/c6x/include/asm/string.h b/arch/c6x/include/asm/string.h
new file mode 100644
index 0000000..b21517c
--- /dev/null
+++ b/arch/c6x/include/asm/string.h
@@ -0,0 +1,21 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2011 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_STRING_H
+#define _ASM_C6X_STRING_H
+
+#include <asm/page.h>
+#include <linux/linkage.h>
+
+asmlinkage extern void *memcpy(void *to, const void *from, size_t n);
+
+#define __HAVE_ARCH_MEMCPY
+
+#endif /* _ASM_C6X_STRING_H */
diff --git a/arch/c6x/include/asm/system.h b/arch/c6x/include/asm/system.h
new file mode 100644
index 0000000..e076dc0
--- /dev/null
+++ b/arch/c6x/include/asm/system.h
@@ -0,0 +1,168 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010, 2011 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_SYSTEM_H
+#define _ASM_C6X_SYSTEM_H
+
+#include <linux/linkage.h>
+#include <linux/irqflags.h>
+
+#define prepare_to_switch() do { } while (0)
+
+struct task_struct;
+struct thread_struct;
+asmlinkage void *__switch_to(struct thread_struct *prev,
+ struct thread_struct *next,
+ struct task_struct *tsk);
+
+#define switch_to(prev, next, last) \
+ do { \
+ current->thread.wchan = (u_long) __builtin_return_address(0); \
+ (last) = __switch_to(&(prev)->thread, \
+ &(next)->thread, (prev)); \
+ mb(); \
+ current->thread.wchan = 0; \
+ } while (0)
+
+/* Reset the board */
+#define HARD_RESET_NOW()
+
+#define get_creg(reg) \
+ ({ unsigned int __x; \
+ asm volatile ("mvc .s2 " #reg ",%0\n" : "=b"(__x)); __x; })
+
+#define set_creg(reg, v) \
+ do { unsigned int __x = (unsigned int)(v); \
+ asm volatile ("mvc .s2 %0," #reg "\n" : : "b"(__x)); \
+ } while (0)
+
+#define or_creg(reg, n) \
+ do { unsigned __x, __n = (unsigned)(n); \
+ asm volatile ("mvc .s2 " #reg ",%0\n" \
+ "or .l2 %1,%0,%0\n" \
+ "mvc .s2 %0," #reg "\n" \
+ "nop\n" \
+ : "=&b"(__x) : "b"(__n)); \
+ } while (0)
+
+#define and_creg(reg, n) \
+ do { unsigned __x, __n = (unsigned)(n); \
+ asm volatile ("mvc .s2 " #reg ",%0\n" \
+ "and .l2 %1,%0,%0\n" \
+ "mvc .s2 %0," #reg "\n" \
+ "nop\n" \
+ : "=&b"(__x) : "b"(__n)); \
+ } while (0)
+
+#define get_coreid() (get_creg(DNUM) & 0xff)
+
+/* Set/get IST */
+#define set_ist(x) set_creg(ISTP, x)
+#define get_ist() get_creg(ISTP)
+
+/*
+ * Exception management
+ */
+asmlinkage void enable_exception(void);
+#define disable_exception()
+#define get_except_type() get_creg(EFR)
+#define ack_exception(type) set_creg(ECR, 1 << (type))
+#define get_iexcept() get_creg(IERR)
+#define set_iexcept(mask) set_creg(IERR, (mask))
+
+/*
+ * Misc. functions
+ */
+#define nop() asm("NOP\n");
+#define mb() barrier()
+#define rmb() barrier()
+#define wmb() barrier()
+#define set_mb(var, value) do { var = value; mb(); } while (0)
+#define set_wmb(var, value) do { var = value; wmb(); } while (0)
+
+#define smp_mb() barrier()
+#define smp_rmb() barrier()
+#define smp_wmb() barrier()
+#define smp_read_barrier_depends() do { } while (0)
+
+#define xchg(ptr, x) \
+ ((__typeof__(*(ptr)))__xchg((unsigned int)(x), (void *) (ptr), \
+ sizeof(*(ptr))))
+#define tas(ptr) xchg((ptr), 1)
+
+unsigned int _lmbd(unsigned int, unsigned int);
+unsigned int _bitr(unsigned int);
+
+struct __xchg_dummy { unsigned int a[100]; };
+#define __xg(x) ((volatile struct __xchg_dummy *)(x))
+
+static inline unsigned int __xchg(unsigned int x, volatile void *ptr, int size)
+{
+ unsigned int tmp;
+ unsigned long flags;
+
+ local_irq_save(flags);
+
+ switch (size) {
+ case 1:
+ tmp = 0;
+ tmp = *((unsigned char *) ptr);
+ *((unsigned char *) ptr) = (unsigned char) x;
+ break;
+ case 2:
+ tmp = 0;
+ tmp = *((unsigned short *) ptr);
+ *((unsigned short *) ptr) = x;
+ break;
+ case 4:
+ tmp = 0;
+ tmp = *((unsigned int *) ptr);
+ *((unsigned int *) ptr) = x;
+ break;
+ }
+ local_irq_restore(flags);
+ return tmp;
+}
+
+#include <asm-generic/cmpxchg-local.h>
+
+/*
+ * cmpxchg_local and cmpxchg64_local are atomic wrt current CPU. Always make
+ * them available.
+ */
+#define cmpxchg_local(ptr, o, n) \
+ ((__typeof__(*(ptr)))__cmpxchg_local_generic((ptr), \
+ (unsigned long)(o), \
+ (unsigned long)(n), \
+ sizeof(*(ptr))))
+#define cmpxchg64_local(ptr, o, n) __cmpxchg64_local_generic((ptr), (o), (n))
+
+#include <asm-generic/cmpxchg.h>
+
+#define _extu(x, s, e) \
+ ({ unsigned int __x; \
+ asm volatile ("extu .S2 %3,%1,%2,%0\n" : \
+ "=b"(__x) : "n"(s), "n"(e), "b"(x)); \
+ __x; })
+
+
+extern unsigned int c6x_core_freq;
+
+struct pt_regs;
+
+extern void die(char *str, struct pt_regs *fp, int nr);
+extern asmlinkage int process_exception(struct pt_regs *regs);
+extern void time_init(void);
+extern void free_initmem(void);
+
+extern void (*c6x_restart)(void);
+extern void (*c6x_halt)(void);
+
+#endif /* _ASM_C6X_SYSTEM_H */
diff --git a/arch/c6x/include/asm/tlb.h b/arch/c6x/include/asm/tlb.h
new file mode 100644
index 0000000..8709e5e
--- /dev/null
+++ b/arch/c6x/include/asm/tlb.h
@@ -0,0 +1,8 @@
+#ifndef _ASM_C6X_TLB_H
+#define _ASM_C6X_TLB_H
+
+#define tlb_flush(tlb) flush_tlb_mm((tlb)->mm)
+
+#include <asm-generic/tlb.h>
+
+#endif /* _ASM_C6X_TLB_H */
diff --git a/arch/c6x/include/asm/uaccess.h b/arch/c6x/include/asm/uaccess.h
new file mode 100644
index 0000000..453dd26
--- /dev/null
+++ b/arch/c6x/include/asm/uaccess.h
@@ -0,0 +1,107 @@
+/*
+ * Copyright (C) 2011 Texas Instruments Incorporated
+ * Author: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_UACCESS_H
+#define _ASM_C6X_UACCESS_H
+
+#include <linux/types.h>
+#include <linux/compiler.h>
+#include <linux/string.h>
+
+#ifdef CONFIG_ACCESS_CHECK
+#define __access_ok _access_ok
+#endif
+
+/*
+ * __copy_from_user/copy_to_user are based on ones in asm-generic/uaccess.h
+ *
+ * C6X supports unaligned 32 and 64 bit loads and stores.
+ */
+static inline __must_check long __copy_from_user(void *to,
+ const void __user *from, unsigned long n)
+{
+ u32 tmp32;
+ u64 tmp64;
+
+ if (__builtin_constant_p(n)) {
+ switch (n) {
+ case 1:
+ *(u8 *)to = *(u8 __force *)from;
+ return 0;
+ case 4:
+ asm volatile ("ldnw .d1t1 *%2,%0\n"
+ "nop 4\n"
+ "stnw .d1t1 %0,*%1\n"
+ : "=&a"(tmp32)
+ : "A"(to), "a"(from)
+ : "memory");
+ return 0;
+ case 8:
+ asm volatile ("ldndw .d1t1 *%2,%0\n"
+ "nop 4\n"
+ "stndw .d1t1 %0,*%1\n"
+ : "=&a"(tmp64)
+ : "a"(to), "a"(from)
+ : "memory");
+ return 0;
+ default:
+ break;
+ }
+ }
+
+ memcpy(to, (const void __force *)from, n);
+ return 0;
+}
+
+static inline __must_check long __copy_to_user(void __user *to,
+ const void *from, unsigned long n)
+{
+ u32 tmp32;
+ u64 tmp64;
+
+ if (__builtin_constant_p(n)) {
+ switch (n) {
+ case 1:
+ *(u8 __force *)to = *(u8 *)from;
+ return 0;
+ case 4:
+ asm volatile ("ldnw .d1t1 *%2,%0\n"
+ "nop 4\n"
+ "stnw .d1t1 %0,*%1\n"
+ : "=&a"(tmp32)
+ : "a"(to), "a"(from)
+ : "memory");
+ return 0;
+ case 8:
+ asm volatile ("ldndw .d1t1 *%2,%0\n"
+ "nop 4\n"
+ "stndw .d1t1 %0,*%1\n"
+ : "=&a"(tmp64)
+ : "a"(to), "a"(from)
+ : "memory");
+ return 0;
+ default:
+ break;
+ }
+ }
+
+ memcpy((void __force *)to, from, n);
+ return 0;
+}
+
+#define __copy_to_user __copy_to_user
+#define __copy_from_user __copy_from_user
+
+extern int _access_ok(unsigned long addr, unsigned long size);
+#ifdef CONFIG_ACCESS_CHECK
+#define __access_ok _access_ok
+#endif
+
+#include <asm-generic/uaccess.h>
+
+#endif /* _ASM_C6X_UACCESS_H */
diff --git a/arch/c6x/include/asm/unaligned.h b/arch/c6x/include/asm/unaligned.h
new file mode 100644
index 0000000..1dcb006
--- /dev/null
+++ b/arch/c6x/include/asm/unaligned.h
@@ -0,0 +1,288 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ * Rewritten for 2.6.3x: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_UNALIGNED_H
+#define _ASM_C6X_UNALIGNED_H
+
+/*
+ * The C64x+ can do unaligned word and dword accesses in hardware
+ * using special load/store instructions.
+ */
+
+static inline u16 __get_unaligned_le16(const u8 *p)
+{
+ return p[0] | p[1] << 8;
+}
+
+static inline u16 __get_unaligned_be16(const u8 *p)
+{
+ return p[0] << 8 | p[1];
+}
+
+static inline void __put_unaligned_le16(u16 val, u8 *p)
+{
+ *p++ = val;
+ *p++ = val >> 8;
+}
+
+static inline void __put_unaligned_be16(u16 val, u8 *p)
+{
+ *p++ = val >> 8;
+ *p++ = val;
+}
+
+static inline u32 __get_unaligned32_swab(const u8 *p)
+{
+ u32 val = (u32) p;
+ asm volatile (" ldnw .d1t1 *%0,%0\n"
+ " nop 4\n"
+ " swap2 .s1 %0,%0\n"
+ " swap4 .l1 %0,%0\n"
+ : "+a"(val));
+ return val;
+}
+
+static inline u32 __get_unaligned32(const u8 *p)
+{
+ u32 val = (u32) p;
+ asm volatile (" ldnw .d1t1 *%0,%0\n"
+ " nop 4\n"
+ : "+a"(val));
+ return val;
+}
+
+static inline void __put_unaligned32_swab(u32 val, u8 *p)
+{
+ asm volatile (" swap2 .s1 %0,%0\n"
+ " swap4 .l1 %0,%0\n"
+ " stnw .d2t1 %0,*%1\n"
+ : : "a"(val), "b"(p) : "memory");
+}
+
+static inline void __put_unaligned32(u32 val, u8 *p)
+{
+ asm volatile (" stnw .d2t1 %0,*%1\n"
+ : : "a"(val), "b"(p) : "memory");
+}
+
+static inline u64 __get_unaligned64_swab(const u8 *p)
+{
+ u64 val;
+
+ asm volatile (" ldndw .d2t1 *%1,%0\n"
+ " nop 4\n"
+ " swap2 .s1 %p0,%P0\n"
+ " || swap2 .l1 %P0,%p0\n"
+ " swap4 .l1 %p0,%p0\n"
+ " swap4 .l1 %P0,%P0\n"
+ : "=a"(val) : "b"(p));
+ return val;
+}
+
+static inline u64 __get_unaligned64(const u8 *p)
+{
+ u64 val;
+ asm volatile (" ldndw .d1t1 *%1,%0\n"
+ " nop 4\n"
+ : "=a"(val) : "a"(p));
+ return val;
+}
+
+static inline void __put_unaligned64_swab(u64 val, u8 *p)
+{
+ asm volatile (" swap2 .s1 %p0,%P0\n"
+ " || swap2 .l1 %P0,%p0\n"
+ " swap4 .l1 %p0,%p0\n"
+ " swap4 .l1 %P0,%P0\n"
+ " stndw .d2t1 %0,*%1\n"
+ : : "a"(val), "b"(p) : "memory");
+}
+
+static inline void __put_unaligned64(u64 val, u8 *p)
+{
+ asm volatile (" stndw .d2t1 %0,*%1\n"
+ : : "a"(val), "b"(p) : "memory");
+}
+
+
+static inline u16 get_unaligned_le16(const void *p)
+{
+ return __get_unaligned_le16((const u8 *)p);
+}
+
+static inline u16 get_unaligned_be16(const void *p)
+{
+ return __get_unaligned_be16((const u8 *)p);
+}
+
+static inline void put_unaligned_le16(u16 val, void *p)
+{
+ __put_unaligned_le16(val, p);
+}
+
+static inline void put_unaligned_be16(u16 val, void *p)
+{
+ __put_unaligned_be16(val, p);
+}
+
+
+static inline u32 get_unaligned_le32(const void *p)
+{
+#ifdef CONFIG_CPU_BIG_ENDIAN
+ return __get_unaligned32_swab((const u8 *)p);
+#else
+ return __get_unaligned32((const u8 *)p);
+#endif
+}
+
+static inline u32 get_unaligned_be32(const void *p)
+{
+#ifdef CONFIG_CPU_BIG_ENDIAN
+ return __get_unaligned32((const u8 *)p);
+#else
+ return __get_unaligned32_swab((const u8 *)p);
+#endif
+}
+
+static inline void put_unaligned_le32(u32 val, void *p)
+{
+#ifdef CONFIG_CPU_BIG_ENDIAN
+ __put_unaligned32_swab(val, p);
+#else
+ __put_unaligned32(val, p);
+#endif
+}
+
+static inline void put_unaligned_be32(u32 val, void *p)
+{
+#ifdef CONFIG_CPU_BIG_ENDIAN
+ __put_unaligned32(val, p);
+#else
+ __put_unaligned32_swab(val, p);
+#endif
+}
+
+static inline u64 get_unaligned_le64(const void *p)
+{
+#ifdef CONFIG_CPU_BIG_ENDIAN
+ return __get_unaligned64_swab((const u8 *)p);
+#else
+ return __get_unaligned64((const u8 *)p);
+#endif
+}
+
+static inline u64 get_unaligned_be64(const void *p)
+{
+#ifdef CONFIG_CPU_BIG_ENDIAN
+ return __get_unaligned64((const u8 *)p);
+#else
+ return __get_unaligned64_swab((const u8 *)p);
+#endif
+}
+
+static inline void put_unaligned_le64(u64 val, void *p)
+{
+#ifdef CONFIG_CPU_BIG_ENDIAN
+ __put_unaligned64_swab(val, p);
+#else
+ __put_unaligned64(val, p);
+#endif
+}
+
+static inline void put_unaligned_be64(u64 val, void *p)
+{
+#ifdef CONFIG_CPU_BIG_ENDIAN
+ __put_unaligned64(val, p);
+#else
+ __put_unaligned64_swab(val, p);
+#endif
+}
+
+/*
+ * Cause a link-time error if we try an unaligned access other than
+ * 1,2,4 or 8 bytes long
+ */
+extern int __bad_unaligned_access_size(void);
+
+#define __get_unaligned_le(ptr) (typeof(*(ptr)))({ \
+ sizeof(*(ptr)) == 1 ? *(ptr) : \
+ (sizeof(*(ptr)) == 2 ? get_unaligned_le16((ptr)) : \
+ (sizeof(*(ptr)) == 4 ? get_unaligned_le32((ptr)) : \
+ (sizeof(*(ptr)) == 8 ? get_unaligned_le64((ptr)) : \
+ __bad_unaligned_access_size()))); \
+ })
+
+#define __get_unaligned_be(ptr) (__force typeof(*(ptr)))({ \
+ sizeof(*(ptr)) == 1 ? *(ptr) : \
+ (sizeof(*(ptr)) == 2 ? get_unaligned_be16((ptr)) : \
+ (sizeof(*(ptr)) == 4 ? get_unaligned_be32((ptr)) : \
+ (sizeof(*(ptr)) == 8 ? get_unaligned_be64((ptr)) : \
+ __bad_unaligned_access_size()))); \
+ })
+
+#define __put_unaligned_le(val, ptr) ({ \
+ void *__gu_p = (ptr); \
+ switch (sizeof(*(ptr))) { \
+ case 1: \
+ *(u8 *)__gu_p = (__force u8)(val); \
+ break; \
+ case 2: \
+ put_unaligned_le16((__force u16)(val), __gu_p); \
+ break; \
+ case 4: \
+ put_unaligned_le32((__force u32)(val), __gu_p); \
+ break; \
+ case 8: \
+ put_unaligned_le64((__force u64)(val), __gu_p); \
+ break; \
+ default: \
+ __bad_unaligned_access_size(); \
+ break; \
+ } \
+ (void)0; })
+
+#define __put_unaligned_be(val, ptr) ({ \
+ void *__gu_p = (ptr); \
+ switch (sizeof(*(ptr))) { \
+ case 1: \
+ *(u8 *)__gu_p = (__force u8)(val); \
+ break; \
+ case 2: \
+ put_unaligned_be16((__force u16)(val), __gu_p); \
+ break; \
+ case 4: \
+ put_unaligned_be32((__force u32)(val), __gu_p); \
+ break; \
+ case 8: \
+ put_unaligned_be64((__force u64)(val), __gu_p); \
+ break; \
+ default: \
+ __bad_unaligned_access_size(); \
+ break; \
+ } \
+ (void)0; })
+
+
+#ifdef _BIG_ENDIAN
+#define get_unaligned __get_unaligned_be
+#define put_unaligned __put_unaligned_be
+#define get_unaligned16 get_unaligned_be16
+#define get_unaligned32 get_unaligned_be32
+#define get_unaligned64 get_unaligned_be64
+#else
+#define get_unaligned __get_unaligned_le
+#define put_unaligned __put_unaligned_le
+#define get_unaligned16 get_unaligned_le16
+#define get_unaligned32 get_unaligned_le32
+#define get_unaligned64 get_unaligned_le64
+#endif
+
+#endif /* _ASM_C6X_UNALIGNED_H */
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v3 19/24] C6X: headers
2011-09-27 20:30 ` [PATCH v3 19/24] C6X: headers Mark Salter
@ 2011-09-28 14:07 ` Arnd Bergmann
2011-09-29 13:32 ` Mark Salter
2011-09-29 14:24 ` Mark Salter
0 siblings, 2 replies; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-28 14:07 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch
On Tuesday 27 September 2011, Mark Salter wrote:
> +
> +static inline void __delay(unsigned long loops)
> +{
> + uint32_t tmp;
> +
> + /* 6 cycles per loop */
> + asm volatile (" mv .s1 %0,%1\n"
> + "0: [%1] b .s1 0b\n"
> + " add .l1 -6,%0,%0\n"
> + " cmplt .l1 1,%0,%1\n"
> + " nop 3\n"
> + : "+a"(loops), "=A"(tmp));
> +}
> +
> +static inline void _c6x_tickdelay(unsigned int x)
> +{
> + uint32_t cnt, endcnt;
> +
> + asm volatile (" mvc .s2 TSCL,%0\n"
> + " add .s2x %0,%1,%2\n"
> + " || mvk .l2 1,B0\n"
> + "0: [B0] b .s2 0b\n"
> + " mvc .s2 TSCL,%0\n"
> + " sub .s2 %0,%2,%0\n"
> + " cmpgt .l2 0,%0,B0\n"
> + " nop 2\n"
> + : "=b"(cnt), "+a"(x), "=b"(endcnt) : : "B0");
> +}
Do you actually need to provide the __delay function here?
I think it's really only used by the calibrate_delay() function
that you don't need.
> diff --git a/arch/c6x/include/asm/unaligned.h b/arch/c6x/include/asm/unaligned.h
> new file mode 100644
> index 0000000..1dcb006
> --- /dev/null
> +++ b/arch/c6x/include/asm/unaligned.h
> @@ -0,0 +1,288 @@
> +/*
> + * Port on Texas Instruments TMS320C6x architecture
This file basically just optimizes unaligned endianess conversion, but
you don't provide an optimized swab.h, which is the regular endianess
conversion.
I think it would less code and more useful to provide your own swab.h
and use the generic unaligned.h.
Arnd
^ permalink raw reply [flat|nested] 66+ messages in thread* Re: [PATCH v3 19/24] C6X: headers
2011-09-28 14:07 ` Arnd Bergmann
@ 2011-09-29 13:32 ` Mark Salter
2011-09-29 14:24 ` Mark Salter
1 sibling, 0 replies; 66+ messages in thread
From: Mark Salter @ 2011-09-29 13:32 UTC (permalink / raw)
To: Arnd Bergmann; +Cc: linux-kernel, linux-arch
On Wed, 2011-09-28 at 16:07 +0200, Arnd Bergmann wrote:
> On Tuesday 27 September 2011, Mark Salter wrote:
>
> > +
> > +static inline void __delay(unsigned long loops)
> > +{
> > + uint32_t tmp;
> > +
> > + /* 6 cycles per loop */
> > + asm volatile (" mv .s1 %0,%1\n"
> > + "0: [%1] b .s1 0b\n"
> > + " add .l1 -6,%0,%0\n"
> > + " cmplt .l1 1,%0,%1\n"
> > + " nop 3\n"
> > + : "+a"(loops), "=A"(tmp));
> > +}
> > +
> > +static inline void _c6x_tickdelay(unsigned int x)
> > +{
> > + uint32_t cnt, endcnt;
> > +
> > + asm volatile (" mvc .s2 TSCL,%0\n"
> > + " add .s2x %0,%1,%2\n"
> > + " || mvk .l2 1,B0\n"
> > + "0: [B0] b .s2 0b\n"
> > + " mvc .s2 TSCL,%0\n"
> > + " sub .s2 %0,%2,%0\n"
> > + " cmpgt .l2 0,%0,B0\n"
> > + " nop 2\n"
> > + : "=b"(cnt), "+a"(x), "=b"(endcnt) : : "B0");
> > +}
>
> Do you actually need to provide the __delay function here?
> I think it's really only used by the calibrate_delay() function
> that you don't need.
>
It is still used in spinlock_debug.c and a few drivers.
--Mark
^ permalink raw reply [flat|nested] 66+ messages in thread* Re: [PATCH v3 19/24] C6X: headers
2011-09-28 14:07 ` Arnd Bergmann
2011-09-29 13:32 ` Mark Salter
@ 2011-09-29 14:24 ` Mark Salter
2011-09-29 14:41 ` Arnd Bergmann
1 sibling, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-29 14:24 UTC (permalink / raw)
To: Arnd Bergmann; +Cc: linux-kernel, linux-arch
On Wed, 2011-09-28 at 16:07 +0200, Arnd Bergmann wrote:
> On Tuesday 27 September 2011, Mark Salter wrote:
> > diff --git a/arch/c6x/include/asm/unaligned.h b/arch/c6x/include/asm/unaligned.h
> > new file mode 100644
> > index 0000000..1dcb006
> > --- /dev/null
> > +++ b/arch/c6x/include/asm/unaligned.h
> > @@ -0,0 +1,288 @@
> > +/*
> > + * Port on Texas Instruments TMS320C6x architecture
>
> This file basically just optimizes unaligned endianess conversion, but
> you don't provide an optimized swab.h, which is the regular endianess
> conversion.
>
> I think it would less code and more useful to provide your own swab.h
> and use the generic unaligned.h.
>
Ah, I missed swab.h which is a trivial optimization. We can do 1 cycle
swab16, 2 cycle swab32, and 3 cycle swab64.
I think we still want the c6x unaligned.h because the hardware can do
unaligned 32 and 64 memory accesses with no performance penalty. There
are some addressing mode limitations which is why the compiler doesn't
use them generally.
--Mark
^ permalink raw reply [flat|nested] 66+ messages in thread
* Re: [PATCH v3 19/24] C6X: headers
2011-09-29 14:24 ` Mark Salter
@ 2011-09-29 14:41 ` Arnd Bergmann
0 siblings, 0 replies; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-29 14:41 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch
On Thursday 29 September 2011 10:24:43 Mark Salter wrote:
> I think we still want the c6x unaligned.h because the hardware can do
> unaligned 32 and 64 memory accesses with no performance penalty. There
> are some addressing mode limitations which is why the compiler doesn't
> use them generally.
Ok. I guess you still only need to override __cpu_to_{le,be}{16,32,64}p and
__{le,be}{16,32,64}_to_cpup then and use include/linux/unaligned/access_ok.h.
Arnd
^ permalink raw reply [flat|nested] 66+ messages in thread
* [PATCH v3 20/24] C6X: library code
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (18 preceding siblings ...)
2011-09-27 20:30 ` [PATCH v3 19/24] C6X: headers Mark Salter
@ 2011-09-27 20:30 ` Mark Salter
2011-09-28 14:09 ` Arnd Bergmann
2011-09-27 20:30 ` [PATCH v3 21/24] C6X: general SoC support Mark Salter
` (5 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:30 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
Signed-off-by: Mark Salter <msalter@redhat.com>
---
arch/c6x/include/asm/checksum.h | 115 +++++++++++
arch/c6x/lib/csum_64plus.S | 404 +++++++++++++++++++++++++++++++++++++++
arch/c6x/lib/divi.S | 53 +++++
arch/c6x/lib/divremi.S | 46 +++++
arch/c6x/lib/divremu.S | 87 +++++++++
arch/c6x/lib/divu.S | 98 ++++++++++
arch/c6x/lib/llshl.S | 37 ++++
arch/c6x/lib/llshr.S | 38 ++++
arch/c6x/lib/llshru.S | 38 ++++
arch/c6x/lib/memcpy_64plus.S | 46 +++++
arch/c6x/lib/mpyll.S | 49 +++++
arch/c6x/lib/negll.S | 31 +++
arch/c6x/lib/pop_rts.S | 32 +++
arch/c6x/lib/push_rts.S | 31 +++
arch/c6x/lib/remi.S | 64 ++++++
arch/c6x/lib/remu.S | 82 ++++++++
arch/c6x/lib/strasgi.S | 89 +++++++++
arch/c6x/lib/strasgi_64plus.S | 39 ++++
18 files changed, 1379 insertions(+), 0 deletions(-)
create mode 100644 arch/c6x/include/asm/checksum.h
create mode 100644 arch/c6x/lib/csum_64plus.S
create mode 100644 arch/c6x/lib/divi.S
create mode 100644 arch/c6x/lib/divremi.S
create mode 100644 arch/c6x/lib/divremu.S
create mode 100644 arch/c6x/lib/divu.S
create mode 100644 arch/c6x/lib/llshl.S
create mode 100644 arch/c6x/lib/llshr.S
create mode 100644 arch/c6x/lib/llshru.S
create mode 100644 arch/c6x/lib/memcpy_64plus.S
create mode 100644 arch/c6x/lib/mpyll.S
create mode 100644 arch/c6x/lib/negll.S
create mode 100644 arch/c6x/lib/pop_rts.S
create mode 100644 arch/c6x/lib/push_rts.S
create mode 100644 arch/c6x/lib/remi.S
create mode 100644 arch/c6x/lib/remu.S
create mode 100644 arch/c6x/lib/strasgi.S
create mode 100644 arch/c6x/lib/strasgi_64plus.S
diff --git a/arch/c6x/include/asm/checksum.h b/arch/c6x/include/asm/checksum.h
new file mode 100644
index 0000000..9bb2845
--- /dev/null
+++ b/arch/c6x/include/asm/checksum.h
@@ -0,0 +1,115 @@
+/*
+ * Port on Texas Instruments TMS320C6x architecture
+ *
+ * Copyright (C) 2004, 2009, 2010 Texas Instruments Incorporated
+ * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef _ASM_C6X_CHECKSUM_H
+#define _ASM_C6X_CHECKSUM_H
+
+#include <asm/byteorder.h>
+
+/*
+ * Computes the checksum of a memory block at buff, length len,
+ * and adds in "sum" (32-bit)
+ *
+ * returns a 32-bit number suitable for feeding into itself
+ * or csum_tcpudp_magic
+ *
+ * this function must be called with even lengths, except
+ * for the last fragment, which may be odd
+ *
+ * it's best to have buff aligned on a 32-bit boundary
+ */
+extern __wsum csum_partial(const void *buff, int len, __wsum sum);
+
+/*
+ * The same as csum_partial, but copies from src while it checksums
+ *
+ * here even more important to align src and dst on a 32-bit (or even
+ * better 64-bit) boundary
+ */
+extern __wsum csum_partial_copy(const void *src, void *dst,
+ int len, __wsum sum);
+
+/*
+ * This is a new version of the above that records errors it finds in *errp,
+ * but continues and zeros the rest of the buffer.
+ */
+#define csum_partial_copy_nocheck csum_partial_copy
+
+/*
+ * The same as csum_partial_copy, but copies from user space.
+ *
+ * here even more important to align src and dst on a 32-bit (or even
+ * better 64-bit) boundary
+ */
+#define csum_partial_copy_from_user(src, dst, len, sum, err_ptr) \
+ csum_partial_copy(src, dst, len, sum)
+
+/*
+ * This is a version of ip_compute_csum() optimized for IP headers,
+ * which always checksum on 4 octet boundaries.
+ *
+ */
+extern __sum16 ip_fast_csum(const void *iph, unsigned int ihl);
+
+/*
+ * Fold a partial checksum
+ */
+static inline __sum16 csum_fold(__wsum csum)
+{
+ u32 sum = (u32)csum;
+ sum = (sum & 0xffff) + (sum >> 16);
+ sum = (sum & 0xffff) + (sum >> 16);
+ return (__sum16)~sum;
+}
+
+/*
+ * Computes the checksum of the TCP/UDP pseudo-header
+ */
+static inline __wsum
+csum_tcpudp_nofold(__be32 saddr, __be32 daddr, unsigned short len,
+ unsigned short proto, __wsum sum)
+{
+ unsigned long long s = (__force u32)sum;
+
+ s += (__force u32)saddr;
+ s += (__force u32)daddr;
+#ifdef _BIG_ENDIAN
+ s += proto + len;
+#else
+ s += (proto + len) << 8;
+#endif
+ s += (s >> 32);
+
+ return (__force __wsum)s;
+}
+
+/*
+ * Computes the checksum of the TCP/UDP pseudo-header
+ * returns a 16-bit checksum, already complemented
+ */
+static inline __sum16
+csum_tcpudp_magic(__be32 saddr, __be32 daddr, unsigned short len,
+ unsigned short proto, __wsum sum)
+{
+ return csum_fold(csum_tcpudp_nofold(saddr, daddr, len, proto, sum));
+}
+
+/*
+ * This routine is used for miscellaneous IP-like checksums, mainly
+ * in icmp.c
+ */
+static inline __sum16
+ip_compute_csum(const void *buff, int len)
+{
+ extern unsigned int do_csum(const unsigned char *, size_t);
+ return (__force __sum16)~do_csum(buff, len);
+}
+
+#endif /* _ASM_C6X_CHECKSUM_H */
diff --git a/arch/c6x/lib/csum_64plus.S b/arch/c6x/lib/csum_64plus.S
new file mode 100644
index 0000000..da9b5ff
--- /dev/null
+++ b/arch/c6x/lib/csum_64plus.S
@@ -0,0 +1,404 @@
+;
+; linux/arch/c6x/lib/csum_64plus.s
+;
+; Port on Texas Instruments TMS320C6x architecture
+;
+; Copyright (C) 2006, 2009, 2010 Texas Instruments Incorporated
+; Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+;
+; This program is free software; you can redistribute it and/or modify
+; it under the terms of the GNU General Public License version 2 as
+; published by the Free Software Foundation.
+;
+#include <linux/linkage.h>
+
+;
+;unsigned int csum_partial_copy(const char *src, char * dst,
+; int len, int sum)
+;
+; A4: src
+; B4: dst
+; A6: len
+; B6: sum
+; return csum in A4
+;
+
+ .text
+ENTRY(csum_partial_copy)
+ MVC .S2 ILC,B30
+
+ MV .D1X B6,A31 ; given csum
+ ZERO .D1 A9 ; csum (a side)
+|| ZERO .D2 B9 ; csum (b side)
+|| SHRU .S2X A6,2,B5 ; len / 4
+
+ ;; Check alignment and size
+ AND .S1 3,A4,A1
+|| AND .S2 3,B4,B0
+ OR .L2X B0,A1,B0 ; non aligned condition
+|| MVC .S2 B5,ILC
+|| MVK .D2 1,B2
+|| MV .D1X B5,A1 ; words condition
+ [!A1] B .S1 L8
+ [B0] BNOP .S1 L6,5
+
+ SPLOOP 1
+
+ ;; Main loop for aligned words
+ LDW .D1T1 *A4++,A7
+ NOP 4
+ MV .S2X A7,B7
+|| EXTU .S1 A7,0,16,A16
+ STW .D2T2 B7,*B4++
+|| MPYU .M2 B7,B2,B8
+|| ADD .L1 A16,A9,A9
+ NOP
+ SPKERNEL 8,0
+|| ADD .L2 B8,B9,B9
+
+ ZERO .D1 A1
+|| ADD .L1X A9,B9,A9 ; add csum from a and b sides
+
+L6:
+ [!A1] BNOP .S1 L8,5
+
+ ;; Main loop for non-aligned words
+ SPLOOP 2
+ || MVK .L1 1,A2
+
+ LDNW .D1T1 *A4++,A7
+ NOP 3
+
+ NOP
+ MV .S2X A7,B7
+ || EXTU .S1 A7,0,16,A16
+ || MPYU .M1 A7,A2,A8
+
+ ADD .L1 A16,A9,A9
+ SPKERNEL 6,0
+ || STNW .D2T2 B7,*B4++
+ || ADD .L1 A8,A9,A9
+
+L8: AND .S2X 2,A6,B5
+ CMPGT .L2 B5,0,B0
+ [!B0] BNOP .S1 L82,4
+
+ ;; Manage half-word
+ ZERO .L1 A7
+|| ZERO .D1 A8
+
+#ifdef CONFIG_CPU_BIG_ENDIAN
+
+ LDBU .D1T1 *A4++,A7
+ LDBU .D1T1 *A4++,A8
+ NOP 3
+ SHL .S1 A7,8,A0
+ ADD .S1 A8,A9,A9
+ STB .D2T1 A7,*B4++
+|| ADD .S1 A0,A9,A9
+ STB .D2T1 A8,*B4++
+
+#else
+
+ LDBU .D1T1 *A4++,A7
+ LDBU .D1T1 *A4++,A8
+ NOP 3
+ ADD .S1 A7,A9,A9
+ SHL .S1 A8,8,A0
+
+ STB .D2T1 A7,*B4++
+|| ADD .S1 A0,A9,A9
+ STB .D2T1 A8,*B4++
+
+#endif
+
+ ;; Manage eventually the last byte
+L82: AND .S2X 1,A6,B0
+ [!B0] BNOP .S1 L9,5
+
+|| ZERO .L1 A7
+
+L83: LDBU .D1T1 *A4++,A7
+ NOP 4
+
+ MV .L2X A7,B7
+
+#ifdef CONFIG_CPU_BIG_ENDIAN
+
+ STB .D2T2 B7,*B4++
+|| SHL .S1 A7,8,A7
+ ADD .S1 A7,A9,A9
+
+#else
+
+ STB .D2T2 B7,*B4++
+|| ADD .S1 A7,A9,A9
+
+#endif
+
+ ;; Fold the csum
+L9: SHRU .S2X A9,16,B0
+ [!B0] BNOP .S1 L10,5
+
+L91: SHRU .S2X A9,16,B4
+|| EXTU .S1 A9,16,16,A3
+ ADD .D1X A3,B4,A9
+
+ SHRU .S1 A9,16,A0
+ [A0] BNOP .S1 L91,5
+
+L10: ADD .D1 A31,A9,A9
+ MV .D1 A9,A4
+
+ BNOP .S2 B3,4
+ MVC .S2 B30,ILC
+ENDPROC(csum_partial_copy)
+
+;
+;unsigned short
+;ip_fast_csum(unsigned char *iph, unsigned int ihl)
+;{
+; unsigned int checksum = 0;
+; unsigned short *tosum = (unsigned short *) iph;
+; int len;
+;
+; len = ihl*4;
+;
+; if (len <= 0)
+; return 0;
+;
+; while(len) {
+; len -= 2;
+; checksum += *tosum++;
+; }
+; if (len & 1)
+; checksum += *(unsigned char*) tosum;
+;
+; while(checksum >> 16)
+; checksum = (checksum & 0xffff) + (checksum >> 16);
+;
+; return ~checksum;
+;}
+;
+; A4: iph
+; B4: ihl
+; return checksum in A4
+;
+ .text
+
+ENTRY(ip_fast_csum)
+ ZERO .D1 A5
+ || MVC .S2 ILC,B30
+ SHL .S2 B4,2,B0
+ CMPGT .L2 B0,0,B1
+ [!B1] BNOP .S1 L15,4
+ [!B1] ZERO .D1 A3
+
+ [!B0] B .S1 L12
+ SHRU .S2 B0,1,B0
+ MVC .S2 B0,ILC
+ NOP 3
+
+ SPLOOP 1
+ LDHU .D1T1 *A4++,A3
+ NOP 3
+ NOP
+ SPKERNEL 5,0
+ || ADD .L1 A3,A5,A5
+
+L12: SHRU .S1 A5,16,A0
+ [!A0] BNOP .S1 L14,5
+
+L13: SHRU .S2X A5,16,B4
+ EXTU .S1 A5,16,16,A3
+ ADD .D1X A3,B4,A5
+ SHRU .S1 A5,16,A0
+ [A0] BNOP .S1 L13,5
+
+L14: NOT .D1 A5,A3
+ EXTU .S1 A3,16,16,A3
+
+L15: BNOP .S2 B3,3
+ MVC .S2 B30,ILC
+ MV .D1 A3,A4
+ENDPROC(ip_fast_csum)
+
+;
+;unsigned short
+;do_csum(unsigned char *buff, unsigned int len)
+;{
+; int odd, count;
+; unsigned int result = 0;
+;
+; if (len <= 0)
+; goto out;
+; odd = 1 & (unsigned long) buff;
+; if (odd) {
+;#ifdef __LITTLE_ENDIAN
+; result += (*buff << 8);
+;#else
+; result = *buff;
+;#endif
+; len--;
+; buff++;
+; }
+; count = len >> 1; /* nr of 16-bit words.. */
+; if (count) {
+; if (2 & (unsigned long) buff) {
+; result += *(unsigned short *) buff;
+; count--;
+; len -= 2;
+; buff += 2;
+; }
+; count >>= 1; /* nr of 32-bit words.. */
+; if (count) {
+; unsigned int carry = 0;
+; do {
+; unsigned int w = *(unsigned int *) buff;
+; count--;
+; buff += 4;
+; result += carry;
+; result += w;
+; carry = (w > result);
+; } while (count);
+; result += carry;
+; result = (result & 0xffff) + (result >> 16);
+; }
+; if (len & 2) {
+; result += *(unsigned short *) buff;
+; buff += 2;
+; }
+; }
+; if (len & 1)
+;#ifdef __LITTLE_ENDIAN
+; result += *buff;
+;#else
+; result += (*buff << 8);
+;#endif
+; result = (result & 0xffff) + (result >> 16);
+; /* add up carry.. */
+; result = (result & 0xffff) + (result >> 16);
+; if (odd)
+; result = ((result >> 8) & 0xff) | ((result & 0xff) << 8);
+;out:
+; return result;
+;}
+;
+; A4: buff
+; B4: len
+; return checksum in A4
+;
+
+ENTRY(do_csum)
+ CMPGT .L2 B4,0,B0
+ [!B0] BNOP .S1 L26,3
+ EXTU .S1 A4,31,31,A0
+
+ MV .L1 A0,A3
+|| MV .S1X B3,A5
+|| MV .L2 B4,B3
+|| ZERO .D1 A1
+
+#ifdef CONFIG_CPU_BIG_ENDIAN
+ [A0] SUB .L2 B3,1,B3
+|| [A0] LDBU .D1T1 *A4++,A1
+#else
+ [!A0] BNOP .S1 L21,5
+|| [A0] LDBU .D1T1 *A4++,A0
+ SUB .L2 B3,1,B3
+|| SHL .S1 A0,8,A1
+L21:
+#endif
+ SHR .S2 B3,1,B0
+ [!B0] BNOP .S1 L24,3
+ MVK .L1 2,A0
+ AND .L1 A4,A0,A0
+
+ [!A0] BNOP .S1 L22,5
+|| [A0] LDHU .D1T1 *A4++,A0
+ SUB .L2 B0,1,B0
+|| SUB .S2 B3,2,B3
+|| ADD .L1 A0,A1,A1
+L22:
+ SHR .S2 B0,1,B0
+|| ZERO .L1 A0
+
+ [!B0] BNOP .S1 L23,5
+|| [B0] MVC .S2 B0,ILC
+
+ SPLOOP 3
+ SPMASK L1
+|| MV .L1 A1,A2
+|| LDW .D1T1 *A4++,A1
+
+ NOP 4
+ ADD .L1 A0,A1,A0
+ ADD .L1 A2,A0,A2
+
+ SPKERNEL 1,2
+|| CMPGTU .L1 A1,A2,A0
+
+ ADD .L1 A0,A2,A6
+ EXTU .S1 A6,16,16,A7
+ SHRU .S2X A6,16,B0
+ NOP 1
+ ADD .L1X A7,B0,A1
+L23:
+ MVK .L2 2,B0
+ AND .L2 B3,B0,B0
+ [B0] LDHU .D1T1 *A4++,A0
+ NOP 4
+ [B0] ADD .L1 A0,A1,A1
+L24:
+ EXTU .S2 B3,31,31,B0
+#ifdef CONFIG_CPU_BIG_ENDIAN
+ [!B0] BNOP .S1 L25,4
+|| [B0] LDBU .D1T1 *A4,A0
+ SHL .S1 A0,8,A0
+ ADD .L1 A0,A1,A1
+L25:
+#else
+ [B0] LDBU .D1T1 *A4,A0
+ NOP 4
+ [B0] ADD .L1 A0,A1,A1
+#endif
+ EXTU .S1 A1,16,16,A0
+ SHRU .S2X A1,16,B0
+ NOP 1
+ ADD .L1X A0,B0,A0
+ SHRU .S1 A0,16,A1
+ ADD .L1 A0,A1,A0
+ EXTU .S1 A0,16,16,A1
+ EXTU .S1 A1,16,24,A2
+
+ EXTU .S1 A1,24,16,A0
+|| MV .L2X A3,B0
+
+ [B0] OR .L1 A0,A2,A1
+L26:
+ NOP 1
+ BNOP .S2X A5,4
+ MV .L1 A1,A4
+ENDPROC(do_csum)
+
+;__wsum csum_partial(const void *buff, int len, __wsum wsum)
+;{
+; unsigned int sum = (__force unsigned int)wsum;
+; unsigned int result = do_csum(buff, len);
+;
+; /* add in old sum, and carry.. */
+; result += sum;
+; if (sum > result)
+; result += 1;
+; return (__force __wsum)result;
+;}
+;
+ENTRY(csum_partial)
+ MV .L1X B3,A9
+|| CALLP .S2 do_csum,B3
+|| MV .S1 A6,A8
+ BNOP .S2X A9,2
+ ADD .L1 A8,A4,A1
+ CMPGTU .L1 A8,A1,A0
+ ADD .L1 A1,A0,A4
+ENDPROC(csum_partial)
diff --git a/arch/c6x/lib/divi.S b/arch/c6x/lib/divi.S
new file mode 100644
index 0000000..4bde924
--- /dev/null
+++ b/arch/c6x/lib/divi.S
@@ -0,0 +1,53 @@
+;; Copyright 2010 Free Software Foundation, Inc.
+;; Contributed by Bernd Schmidt <bernds@codesourcery.com>.
+;;
+;; This program is free software; you can redistribute it and/or modify
+;; it under the terms of the GNU General Public License as published by
+;; the Free Software Foundation; either version 2 of the License, or
+;; (at your option) any later version.
+;;
+;; This program is distributed in the hope that it will be useful,
+;; but WITHOUT ANY WARRANTY; without even the implied warranty of
+;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+;; GNU General Public License for more details.
+;;
+;; You should have received a copy of the GNU General Public License
+;; along with this program; if not, write to the Free Software
+;; Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+#include <linux/linkage.h>
+
+ ;; ABI considerations for the divide functions
+ ;; The following registers are call-used:
+ ;; __c6xabi_divi A0,A1,A2,A4,A6,B0,B1,B2,B4,B5
+ ;; __c6xabi_divu A0,A1,A2,A4,A6,B0,B1,B2,B4
+ ;; __c6xabi_remi A1,A2,A4,A5,A6,B0,B1,B2,B4
+ ;; __c6xabi_remu A1,A4,A5,A7,B0,B1,B2,B4
+ ;;
+ ;; In our implementation, divu and remu are leaf functions,
+ ;; while both divi and remi call into divu.
+ ;; A0 is not clobbered by any of the functions.
+ ;; divu does not clobber B2 either, which is taken advantage of
+ ;; in remi.
+ ;; divi uses B5 to hold the original return address during
+ ;; the call to divu.
+ ;; remi uses B2 and A5 to hold the input values during the
+ ;; call to divu. It stores B3 in on the stack.
+
+ .text
+ENTRY(__c6xabi_divi)
+ call .s2 __c6xabi_divu
+|| mv .d2 B3, B5
+|| cmpgt .l1 0, A4, A1
+|| cmpgt .l2 0, B4, B1
+
+ [A1] neg .l1 A4, A4
+|| [B1] neg .l2 B4, B4
+|| xor .s1x A1, B1, A1
+ [A1] addkpc .s2 _divu_ret, B3, 4
+_divu_ret:
+ neg .l1 A4, A4
+|| mv .l2 B3,B5
+|| ret .s2 B5
+ nop 5
+ENDPROC(__c6xabi_divi)
diff --git a/arch/c6x/lib/divremi.S b/arch/c6x/lib/divremi.S
new file mode 100644
index 0000000..64bc5aa
--- /dev/null
+++ b/arch/c6x/lib/divremi.S
@@ -0,0 +1,46 @@
+;; Copyright 2010 Free Software Foundation, Inc.
+;; Contributed by Bernd Schmidt <bernds@codesourcery.com>.
+;;
+;; This program is free software; you can redistribute it and/or modify
+;; it under the terms of the GNU General Public License as published by
+;; the Free Software Foundation; either version 2 of the License, or
+;; (at your option) any later version.
+;;
+;; This program is distributed in the hope that it will be useful,
+;; but WITHOUT ANY WARRANTY; without even the implied warranty of
+;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+;; GNU General Public License for more details.
+;;
+;; You should have received a copy of the GNU General Public License
+;; along with this program; if not, write to the Free Software
+;; Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+#include <linux/linkage.h>
+
+ .text
+ENTRY(__c6xabi_divremi)
+ stw .d2t2 B3, *B15--[2]
+|| cmpgt .l1 0, A4, A1
+|| cmpgt .l2 0, B4, B2
+|| mv .s1 A4, A5
+|| call .s2 __c6xabi_divu
+
+ [A1] neg .l1 A4, A4
+|| [B2] neg .l2 B4, B4
+|| xor .s2x B2, A1, B0
+|| mv .d2 B4, B2
+
+ [B0] addkpc .s2 _divu_ret_1, B3, 1
+ [!B0] addkpc .s2 _divu_ret_2, B3, 1
+ nop 2
+_divu_ret_1:
+ neg .l1 A4, A4
+_divu_ret_2:
+ ldw .d2t2 *++B15[2], B3
+
+ mpy32 .m1x A4, B2, A6
+ nop 3
+ ret .s2 B3
+ sub .l1 A5, A6, A5
+ nop 4
+ENDPROC(__c6xabi_divremi)
diff --git a/arch/c6x/lib/divremu.S b/arch/c6x/lib/divremu.S
new file mode 100644
index 0000000..caa9f23
--- /dev/null
+++ b/arch/c6x/lib/divremu.S
@@ -0,0 +1,87 @@
+;; Copyright 2011 Free Software Foundation, Inc.
+;; Contributed by Bernd Schmidt <bernds@codesourcery.com>.
+;;
+;; This program is free software; you can redistribute it and/or modify
+;; it under the terms of the GNU General Public License as published by
+;; the Free Software Foundation; either version 2 of the License, or
+;; (at your option) any later version.
+;;
+;; This program is distributed in the hope that it will be useful,
+;; but WITHOUT ANY WARRANTY; without even the implied warranty of
+;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+;; GNU General Public License for more details.
+;;
+;; You should have received a copy of the GNU General Public License
+;; along with this program; if not, write to the Free Software
+;; Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+#include <linux/linkage.h>
+
+ .text
+ENTRY(__c6xabi_divremu)
+ ;; We use a series of up to 31 subc instructions. First, we find
+ ;; out how many leading zero bits there are in the divisor. This
+ ;; gives us both a shift count for aligning (shifting) the divisor
+ ;; to the, and the number of times we have to execute subc.
+
+ ;; At the end, we have both the remainder and most of the quotient
+ ;; in A4. The top bit of the quotient is computed first and is
+ ;; placed in A2.
+
+ ;; Return immediately if the dividend is zero. Setting B4 to 1
+ ;; is a trick to allow us to leave the following insns in the jump
+ ;; delay slot without affecting the result.
+ mv .s2x A4, B1
+
+ [b1] lmbd .l2 1, B4, B1
+||[!b1] b .s2 B3 ; RETURN A
+||[!b1] mvk .d2 1, B4
+
+||[!b1] zero .s1 A5
+ mv .l1x B1, A6
+|| shl .s2 B4, B1, B4
+
+ ;; The loop performs a maximum of 28 steps, so we do the
+ ;; first 3 here.
+ cmpltu .l1x A4, B4, A2
+ [!A2] sub .l1x A4, B4, A4
+|| shru .s2 B4, 1, B4
+|| xor .s1 1, A2, A2
+
+ shl .s1 A2, 31, A2
+|| [b1] subc .l1x A4,B4,A4
+|| [b1] add .s2 -1, B1, B1
+ [b1] subc .l1x A4,B4,A4
+|| [b1] add .s2 -1, B1, B1
+
+ ;; RETURN A may happen here (note: must happen before the next branch)
+__divremu0:
+ cmpgt .l2 B1, 7, B0
+|| [b1] subc .l1x A4,B4,A4
+|| [b1] add .s2 -1, B1, B1
+ [b1] subc .l1x A4,B4,A4
+|| [b1] add .s2 -1, B1, B1
+|| [b0] b .s1 __divremu0
+ [b1] subc .l1x A4,B4,A4
+|| [b1] add .s2 -1, B1, B1
+ [b1] subc .l1x A4,B4,A4
+|| [b1] add .s2 -1, B1, B1
+ [b1] subc .l1x A4,B4,A4
+|| [b1] add .s2 -1, B1, B1
+ [b1] subc .l1x A4,B4,A4
+|| [b1] add .s2 -1, B1, B1
+ [b1] subc .l1x A4,B4,A4
+|| [b1] add .s2 -1, B1, B1
+ ;; loop backwards branch happens here
+
+ ret .s2 B3
+|| mvk .s1 32, A1
+ sub .l1 A1, A6, A6
+|| extu .s1 A4, A6, A5
+ shl .s1 A4, A6, A4
+ shru .s1 A4, 1, A4
+|| sub .l1 A6, 1, A6
+ or .l1 A2, A4, A4
+ shru .s1 A4, A6, A4
+ nop
+ENDPROC(__c6xabi_divremu)
diff --git a/arch/c6x/lib/divu.S b/arch/c6x/lib/divu.S
new file mode 100644
index 0000000..64af3c0
--- /dev/null
+++ b/arch/c6x/lib/divu.S
@@ -0,0 +1,98 @@
+;; Copyright 2010 Free Software Foundation, Inc.
+;; Contributed by Bernd Schmidt <bernds@codesourcery.com>.
+;;
+;; This program is free software; you can redistribute it and/or modify
+;; it under the terms of the GNU General Public License as published by
+;; the Free Software Foundation; either version 2 of the License, or
+;; (at your option) any later version.
+;;
+;; This program is distributed in the hope that it will be useful,
+;; but WITHOUT ANY WARRANTY; without even the implied warranty of
+;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+;; GNU General Public License for more details.
+;;
+;; You should have received a copy of the GNU General Public License
+;; along with this program; if not, write to the Free Software
+;; Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+#include <linux/linkage.h>
+
+ ;; ABI considerations for the divide functions
+ ;; The following registers are call-used:
+ ;; __c6xabi_divi A0,A1,A2,A4,A6,B0,B1,B2,B4,B5
+ ;; __c6xabi_divu A0,A1,A2,A4,A6,B0,B1,B2,B4
+ ;; __c6xabi_remi A1,A2,A4,A5,A6,B0,B1,B2,B4
+ ;; __c6xabi_remu A1,A4,A5,A7,B0,B1,B2,B4
+ ;;
+ ;; In our implementation, divu and remu are leaf functions,
+ ;; while both divi and remi call into divu.
+ ;; A0 is not clobbered by any of the functions.
+ ;; divu does not clobber B2 either, which is taken advantage of
+ ;; in remi.
+ ;; divi uses B5 to hold the original return address during
+ ;; the call to divu.
+ ;; remi uses B2 and A5 to hold the input values during the
+ ;; call to divu. It stores B3 in on the stack.
+
+ .text
+ENTRY(__c6xabi_divu)
+ ;; We use a series of up to 31 subc instructions. First, we find
+ ;; out how many leading zero bits there are in the divisor. This
+ ;; gives us both a shift count for aligning (shifting) the divisor
+ ;; to the, and the number of times we have to execute subc.
+
+ ;; At the end, we have both the remainder and most of the quotient
+ ;; in A4. The top bit of the quotient is computed first and is
+ ;; placed in A2.
+
+ ;; Return immediately if the dividend is zero.
+ mv .s2x A4, B1
+ [B1] lmbd .l2 1, B4, B1
+|| [!B1] b .s2 B3 ; RETURN A
+|| [!B1] mvk .d2 1, B4
+ mv .l1x B1, A6
+|| shl .s2 B4, B1, B4
+
+ ;; The loop performs a maximum of 28 steps, so we do the
+ ;; first 3 here.
+ cmpltu .l1x A4, B4, A2
+ [!A2] sub .l1x A4, B4, A4
+|| shru .s2 B4, 1, B4
+|| xor .s1 1, A2, A2
+
+ shl .s1 A2, 31, A2
+|| [B1] subc .l1x A4,B4,A4
+|| [B1] add .s2 -1, B1, B1
+ [B1] subc .l1x A4,B4,A4
+|| [B1] add .s2 -1, B1, B1
+
+ ;; RETURN A may happen here (note: must happen before the next branch)
+_divu_loop:
+ cmpgt .l2 B1, 7, B0
+|| [B1] subc .l1x A4,B4,A4
+|| [B1] add .s2 -1, B1, B1
+ [B1] subc .l1x A4,B4,A4
+|| [B1] add .s2 -1, B1, B1
+|| [B0] b .s1 _divu_loop
+ [B1] subc .l1x A4,B4,A4
+|| [B1] add .s2 -1, B1, B1
+ [B1] subc .l1x A4,B4,A4
+|| [B1] add .s2 -1, B1, B1
+ [B1] subc .l1x A4,B4,A4
+|| [B1] add .s2 -1, B1, B1
+ [B1] subc .l1x A4,B4,A4
+|| [B1] add .s2 -1, B1, B1
+ [B1] subc .l1x A4,B4,A4
+|| [B1] add .s2 -1, B1, B1
+ ;; loop backwards branch happens here
+
+ ret .s2 B3
+|| mvk .s1 32, A1
+ sub .l1 A1, A6, A6
+ shl .s1 A4, A6, A4
+ shru .s1 A4, 1, A4
+|| sub .l1 A6, 1, A6
+ or .l1 A2, A4, A4
+ shru .s1 A4, A6, A4
+ nop
+ENDPROC(__c6xabi_divu)
diff --git a/arch/c6x/lib/llshl.S b/arch/c6x/lib/llshl.S
new file mode 100644
index 0000000..7b105e2
--- /dev/null
+++ b/arch/c6x/lib/llshl.S
@@ -0,0 +1,37 @@
+;; Copyright (C) 2010 Texas Instruments Incorporated
+;; Contributed by Mark Salter <msalter@redhat.com>.
+;;
+;; This program is free software; you can redistribute it and/or modify
+;; it under the terms of the GNU General Public License as published by
+;; the Free Software Foundation; either version 2 of the License, or
+;; (at your option) any later version.
+;;
+;; This program is distributed in the hope that it will be useful,
+;; but WITHOUT ANY WARRANTY; without even the implied warranty of
+;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+;; GNU General Public License for more details.
+;;
+;; You should have received a copy of the GNU General Public License
+;; along with this program; if not, write to the Free Software
+;; Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+;; uint64_t __c6xabi_llshl(uint64_t val, uint shift)
+
+#include <linux/linkage.h>
+
+ .text
+ENTRY(__c6xabi_llshl)
+ mv .l1x B4,A1
+ [!A1] b .s2 B3 ; just return if zero shift
+ mvk .s1 32,A0
+ sub .d1 A0,A1,A0
+ cmplt .l1 0,A0,A2
+ [A2] shru .s1 A4,A0,A0
+ [!A2] neg .l1 A0,A5
+|| [A2] shl .s1 A5,A1,A5
+ [!A2] shl .s1 A4,A5,A5
+|| [A2] or .d1 A5,A0,A5
+|| [!A2] mvk .l1 0,A4
+ [A2] shl .s1 A4,A1,A4
+ bnop .s2 B3,5
+ENDPROC(__c6xabi_llshl)
diff --git a/arch/c6x/lib/llshr.S b/arch/c6x/lib/llshr.S
new file mode 100644
index 0000000..fde1bec
--- /dev/null
+++ b/arch/c6x/lib/llshr.S
@@ -0,0 +1,38 @@
+;; Copyright (C) 2010 Texas Instruments Incorporated
+;; Contributed by Mark Salter <msalter@redhat.com>.
+;;
+;; This program is free software; you can redistribute it and/or modify
+;; it under the terms of the GNU General Public License as published by
+;; the Free Software Foundation; either version 2 of the License, or
+;; (at your option) any later version.
+;;
+;; This program is distributed in the hope that it will be useful,
+;; but WITHOUT ANY WARRANTY; without even the implied warranty of
+;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+;; GNU General Public License for more details.
+;;
+;; You should have received a copy of the GNU General Public License
+;; along with this program; if not, write to the Free Software
+;; Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+;; uint64_t __c6xabi_llshr(uint64_t val, uint shift)
+
+#include <linux/linkage.h>
+
+ .text
+ENTRY(__c6xabi_llshr)
+ mv .l1x B4,A1
+ [!A1] b .s2 B3 ; return if zero shift count
+ mvk .s1 32,A0
+ sub .d1 A0,A1,A0
+ cmplt .l1 0,A0,A2
+ [A2] shl .s1 A5,A0,A0
+ nop
+ [!A2] neg .l1 A0,A4
+|| [A2] shru .s1 A4,A1,A4
+ [!A2] shr .s1 A5,A4,A4
+|| [A2] or .d1 A4,A0,A4
+ [!A2] shr .s1 A5,0x1f,A5
+ [A2] shr .s1 A5,A1,A5
+ bnop .s2 B3,5
+ENDPROC(__c6xabi_llshr)
diff --git a/arch/c6x/lib/llshru.S b/arch/c6x/lib/llshru.S
new file mode 100644
index 0000000..596ae3f
--- /dev/null
+++ b/arch/c6x/lib/llshru.S
@@ -0,0 +1,38 @@
+;; Copyright (C) 2010 Texas Instruments Incorporated
+;; Contributed by Mark Salter <msalter@redhat.com>.
+;;
+;; This program is free software; you can redistribute it and/or modify
+;; it under the terms of the GNU General Public License as published by
+;; the Free Software Foundation; either version 2 of the License, or
+;; (at your option) any later version.
+;;
+;; This program is distributed in the hope that it will be useful,
+;; but WITHOUT ANY WARRANTY; without even the implied warranty of
+;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+;; GNU General Public License for more details.
+;;
+;; You should have received a copy of the GNU General Public License
+;; along with this program; if not, write to the Free Software
+;; Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+;; uint64_t __c6xabi_llshru(uint64_t val, uint shift)
+
+#include <linux/linkage.h>
+
+ .text
+ENTRY(__c6xabi_llshru)
+ mv .l1x B4,A1
+ [!A1] b .s2 B3 ; return if zero shift count
+ mvk .s1 32,A0
+ sub .d1 A0,A1,A0
+ cmplt .l1 0,A0,A2
+ [A2] shl .s1 A5,A0,A0
+ nop
+ [!A2] neg .l1 A0,A4
+|| [A2] shru .s1 A4,A1,A4
+ [!A2] shru .s1 A5,A4,A4
+|| [A2] or .d1 A4,A0,A4
+|| [!A2] mvk .l1 0,A5
+ [A2] shru .s1 A5,A1,A5
+ bnop .s2 B3,5
+ENDPROC(__c6xabi_llshru)
diff --git a/arch/c6x/lib/memcpy_64plus.S b/arch/c6x/lib/memcpy_64plus.S
new file mode 100644
index 0000000..0bbc2cb
--- /dev/null
+++ b/arch/c6x/lib/memcpy_64plus.S
@@ -0,0 +1,46 @@
+; Port on Texas Instruments TMS320C6x architecture
+;
+; Copyright (C) 2006, 2009, 2010 Texas Instruments Incorporated
+; Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
+;
+; This program is free software; you can redistribute it and/or modify
+; it under the terms of the GNU General Public License version 2 as
+; published by the Free Software Foundation.
+;
+
+#include <linux/linkage.h>
+
+ .text
+
+ENTRY(memcpy)
+ AND .L1 0x1,A6,A0
+ || AND .S1 0x2,A6,A1
+ || AND .L2X 0x4,A6,B0
+ || MV .D1 A4,A3
+ || MVC .S2 ILC,B2
+
+ [A0] LDB .D2T1 *B4++,A5
+ [A1] LDB .D2T1 *B4++,A7
+ [A1] LDB .D2T1 *B4++,A8
+ [B0] LDNW .D2T1 *B4++,A9
+ || SHRU .S2X A6,0x3,B1
+ [!B1] BNOP .S2 B3,1
+
+ [A0] STB .D1T1 A5,*A3++
+ ||[B1] MVC .S2 B1,ILC
+ [A1] STB .D1T1 A7,*A3++
+ [A1] STB .D1T1 A8,*A3++
+ [B0] STNW .D1T1 A9,*A3++ ; return when len < 8
+
+ SPLOOP 2
+
+ LDNDW .D2T1 *B4++,A9:A8
+ NOP 3
+
+ NOP
+ SPKERNEL 0,0
+ || STNDW .D1T1 A9:A8,*A3++
+
+ BNOP .S2 B3,4
+ MVC .S2 B2,ILC
+ENDPROC(memcpy)
diff --git a/arch/c6x/lib/mpyll.S b/arch/c6x/lib/mpyll.S
new file mode 100644
index 0000000..f103441
--- /dev/null
+++ b/arch/c6x/lib/mpyll.S
@@ -0,0 +1,49 @@
+;; Copyright (C) 2010 Texas Instruments Incorporated
+;; Contributed by Mark Salter <msalter@redhat.com>.
+;;
+;; This program is free software; you can redistribute it and/or modify
+;; it under the terms of the GNU General Public License as published by
+;; the Free Software Foundation; either version 2 of the License, or
+;; (at your option) any later version.
+;;
+;; This program is distributed in the hope that it will be useful,
+;; but WITHOUT ANY WARRANTY; without even the implied warranty of
+;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+;; GNU General Public License for more details.
+;;
+;; You should have received a copy of the GNU General Public License
+;; along with this program; if not, write to the Free Software
+;; Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+#include <linux/linkage.h>
+
+ ;; uint64_t __c6xabi_mpyll(uint64_t x, uint64_t y)
+ ;;
+ ;; 64x64 multiply
+ ;; First compute partial results using 32-bit parts of x and y:
+ ;;
+ ;; b63 b32 b31 b0
+ ;; -----------------------------
+ ;; | 1 | 0 |
+ ;; -----------------------------
+ ;;
+ ;; P0 = X0*Y0
+ ;; P1 = X0*Y1 + X1*Y0
+ ;; P2 = X1*Y1
+ ;;
+ ;; result = (P2 << 64) + (P1 << 32) + P0
+ ;;
+ ;; Since the result is also 64-bit, we can skip the P2 term.
+
+ .text
+ENTRY(__c6xabi_mpyll)
+ mpy32u .m1x A4,B4,A1:A0 ; X0*Y0
+ b .s2 B3
+ || mpy32u .m2x B5,A4,B1:B0 ; X0*Y1 (don't need upper 32-bits)
+ || mpy32u .m1x A5,B4,A3:A2 ; X1*Y0 (don't need upper 32-bits)
+ nop
+ nop
+ mv .s1 A0,A4
+ add .l1x A2,B0,A5
+ add .s1 A1,A5,A5
+ENDPROC(__c6xabi_mpyll)
diff --git a/arch/c6x/lib/negll.S b/arch/c6x/lib/negll.S
new file mode 100644
index 0000000..82f4bce
--- /dev/null
+++ b/arch/c6x/lib/negll.S
@@ -0,0 +1,31 @@
+;; Copyright (C) 2010 Texas Instruments Incorporated
+;; Contributed by Mark Salter <msalter@redhat.com>.
+;;
+;; This program is free software; you can redistribute it and/or modify
+;; it under the terms of the GNU General Public License as published by
+;; the Free Software Foundation; either version 2 of the License, or
+;; (at your option) any later version.
+;;
+;; This program is distributed in the hope that it will be useful,
+;; but WITHOUT ANY WARRANTY; without even the implied warranty of
+;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+;; GNU General Public License for more details.
+;;
+;; You should have received a copy of the GNU General Public License
+;; along with this program; if not, write to the Free Software
+;; Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+;; int64_t __c6xabi_negll(int64_t val)
+
+#include <linux/linkage.h>
+
+ .text
+ENTRY(__c6xabi_negll)
+ b .s2 B3
+ mvk .l1 0,A0
+ subu .l1 A0,A4,A3:A2
+ sub .l1 A0,A5,A0
+|| ext .s1 A3,24,24,A5
+ add .l1 A5,A0,A5
+ mv .s1 A2,A4
+ENDPROC(__c6xabi_negll)
diff --git a/arch/c6x/lib/pop_rts.S b/arch/c6x/lib/pop_rts.S
new file mode 100644
index 0000000..d7d96c7
--- /dev/null
+++ b/arch/c6x/lib/pop_rts.S
@@ -0,0 +1,32 @@
+;; Copyright 2010 Free Software Foundation, Inc.
+;; Contributed by Bernd Schmidt <bernds@codesourcery.com>.
+;;
+;; This program is free software; you can redistribute it and/or modify
+;; it under the terms of the GNU General Public License as published by
+;; the Free Software Foundation; either version 2 of the License, or
+;; (at your option) any later version.
+;;
+;; This program is distributed in the hope that it will be useful,
+;; but WITHOUT ANY WARRANTY; without even the implied warranty of
+;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+;; GNU General Public License for more details.
+;;
+;; You should have received a copy of the GNU General Public License
+;; along with this program; if not, write to the Free Software
+;; Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+#include <linux/linkage.h>
+
+ .text
+
+ENTRY(__c6xabi_pop_rts)
+ lddw .d2t2 *++B15, B3:B2
+ lddw .d2t1 *++B15, A11:A10
+ lddw .d2t2 *++B15, B11:B10
+ lddw .d2t1 *++B15, A13:A12
+ lddw .d2t2 *++B15, B13:B12
+ lddw .d2t1 *++B15, A15:A14
+|| b .s2 B3
+ ldw .d2t2 *++B15[2], B14
+ nop 4
+ENDPROC(__c6xabi_pop_rts)
diff --git a/arch/c6x/lib/push_rts.S b/arch/c6x/lib/push_rts.S
new file mode 100644
index 0000000..f6e3db3
--- /dev/null
+++ b/arch/c6x/lib/push_rts.S
@@ -0,0 +1,31 @@
+;; Copyright 2010 Free Software Foundation, Inc.
+;; Contributed by Bernd Schmidt <bernds@codesourcery.com>.
+;;
+;; This program is free software; you can redistribute it and/or modify
+;; it under the terms of the GNU General Public License as published by
+;; the Free Software Foundation; either version 2 of the License, or
+;; (at your option) any later version.
+;;
+;; This program is distributed in the hope that it will be useful,
+;; but WITHOUT ANY WARRANTY; without even the implied warranty of
+;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+;; GNU General Public License for more details.
+;;
+;; You should have received a copy of the GNU General Public License
+;; along with this program; if not, write to the Free Software
+;; Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+#include <linux/linkage.h>
+
+ .text
+
+ENTRY(__c6xabi_push_rts)
+ stw .d2t2 B14, *B15--[2]
+ stdw .d2t1 A15:A14, *B15--
+|| b .s2x A3
+ stdw .d2t2 B13:B12, *B15--
+ stdw .d2t1 A13:A12, *B15--
+ stdw .d2t2 B11:B10, *B15--
+ stdw .d2t1 A11:A10, *B15--
+ stdw .d2t2 B3:B2, *B15--
+ENDPROC(__c6xabi_push_rts)
diff --git a/arch/c6x/lib/remi.S b/arch/c6x/lib/remi.S
new file mode 100644
index 0000000..6f2ca18
--- /dev/null
+++ b/arch/c6x/lib/remi.S
@@ -0,0 +1,64 @@
+;; Copyright 2010 Free Software Foundation, Inc.
+;; Contributed by Bernd Schmidt <bernds@codesourcery.com>.
+;;
+;; This program is free software; you can redistribute it and/or modify
+;; it under the terms of the GNU General Public License as published by
+;; the Free Software Foundation; either version 2 of the License, or
+;; (at your option) any later version.
+;;
+;; This program is distributed in the hope that it will be useful,
+;; but WITHOUT ANY WARRANTY; without even the implied warranty of
+;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+;; GNU General Public License for more details.
+;;
+;; You should have received a copy of the GNU General Public License
+;; along with this program; if not, write to the Free Software
+;; Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+#include <linux/linkage.h>
+
+ ;; ABI considerations for the divide functions
+ ;; The following registers are call-used:
+ ;; __c6xabi_divi A0,A1,A2,A4,A6,B0,B1,B2,B4,B5
+ ;; __c6xabi_divu A0,A1,A2,A4,A6,B0,B1,B2,B4
+ ;; __c6xabi_remi A1,A2,A4,A5,A6,B0,B1,B2,B4
+ ;; __c6xabi_remu A1,A4,A5,A7,B0,B1,B2,B4
+ ;;
+ ;; In our implementation, divu and remu are leaf functions,
+ ;; while both divi and remi call into divu.
+ ;; A0 is not clobbered by any of the functions.
+ ;; divu does not clobber B2 either, which is taken advantage of
+ ;; in remi.
+ ;; divi uses B5 to hold the original return address during
+ ;; the call to divu.
+ ;; remi uses B2 and A5 to hold the input values during the
+ ;; call to divu. It stores B3 in on the stack.
+
+ .text
+
+ENTRY(__c6xabi_remi)
+ stw .d2t2 B3, *B15--[2]
+|| cmpgt .l1 0, A4, A1
+|| cmpgt .l2 0, B4, B2
+|| mv .s1 A4, A5
+|| call .s2 __c6xabi_divu
+
+ [A1] neg .l1 A4, A4
+|| [B2] neg .l2 B4, B4
+|| xor .s2x B2, A1, B0
+|| mv .d2 B4, B2
+
+ [B0] addkpc .s2 _divu_ret_1, B3, 1
+ [!B0] addkpc .s2 _divu_ret_2, B3, 1
+ nop 2
+_divu_ret_1:
+ neg .l1 A4, A4
+_divu_ret_2:
+ ldw .d2t2 *++B15[2], B3
+
+ mpy32 .m1x A4, B2, A6
+ nop 3
+ ret .s2 B3
+ sub .l1 A5, A6, A4
+ nop 4
+ENDPROC(__c6xabi_remi)
diff --git a/arch/c6x/lib/remu.S b/arch/c6x/lib/remu.S
new file mode 100644
index 0000000..3fae719
--- /dev/null
+++ b/arch/c6x/lib/remu.S
@@ -0,0 +1,82 @@
+;; Copyright 2010 Free Software Foundation, Inc.
+;; Contributed by Bernd Schmidt <bernds@codesourcery.com>.
+;;
+;; This program is free software; you can redistribute it and/or modify
+;; it under the terms of the GNU General Public License as published by
+;; the Free Software Foundation; either version 2 of the License, or
+;; (at your option) any later version.
+;;
+;; This program is distributed in the hope that it will be useful,
+;; but WITHOUT ANY WARRANTY; without even the implied warranty of
+;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+;; GNU General Public License for more details.
+;;
+;; You should have received a copy of the GNU General Public License
+;; along with this program; if not, write to the Free Software
+;; Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+#include <linux/linkage.h>
+
+ ;; ABI considerations for the divide functions
+ ;; The following registers are call-used:
+ ;; __c6xabi_divi A0,A1,A2,A4,A6,B0,B1,B2,B4,B5
+ ;; __c6xabi_divu A0,A1,A2,A4,A6,B0,B1,B2,B4
+ ;; __c6xabi_remi A1,A2,A4,A5,A6,B0,B1,B2,B4
+ ;; __c6xabi_remu A1,A4,A5,A7,B0,B1,B2,B4
+ ;;
+ ;; In our implementation, divu and remu are leaf functions,
+ ;; while both divi and remi call into divu.
+ ;; A0 is not clobbered by any of the functions.
+ ;; divu does not clobber B2 either, which is taken advantage of
+ ;; in remi.
+ ;; divi uses B5 to hold the original return address during
+ ;; the call to divu.
+ ;; remi uses B2 and A5 to hold the input values during the
+ ;; call to divu. It stores B3 in on the stack.
+
+
+ .text
+
+ENTRY(__c6xabi_remu)
+ ;; The ABI seems designed to prevent these functions calling each other,
+ ;; so we duplicate most of the divsi3 code here.
+ mv .s2x A4, B1
+ lmbd .l2 1, B4, B1
+|| [!B1] b .s2 B3 ; RETURN A
+|| [!B1] mvk .d2 1, B4
+
+ mv .l1x B1, A7
+|| shl .s2 B4, B1, B4
+
+ cmpltu .l1x A4, B4, A1
+ [!A1] sub .l1x A4, B4, A4
+ shru .s2 B4, 1, B4
+
+_remu_loop:
+ cmpgt .l2 B1, 7, B0
+|| [B1] subc .l1x A4,B4,A4
+|| [B1] add .s2 -1, B1, B1
+ ;; RETURN A may happen here (note: must happen before the next branch)
+ [B1] subc .l1x A4,B4,A4
+|| [B1] add .s2 -1, B1, B1
+|| [B0] b .s1 _remu_loop
+ [B1] subc .l1x A4,B4,A4
+|| [B1] add .s2 -1, B1, B1
+ [B1] subc .l1x A4,B4,A4
+|| [B1] add .s2 -1, B1, B1
+ [B1] subc .l1x A4,B4,A4
+|| [B1] add .s2 -1, B1, B1
+ [B1] subc .l1x A4,B4,A4
+|| [B1] add .s2 -1, B1, B1
+ [B1] subc .l1x A4,B4,A4
+|| [B1] add .s2 -1, B1, B1
+ ;; loop backwards branch happens here
+
+ ret .s2 B3
+ [B1] subc .l1x A4,B4,A4
+|| [B1] add .s2 -1, B1, B1
+ [B1] subc .l1x A4,B4,A4
+
+ extu .s1 A4, A7, A4
+ nop 2
+ENDPROC(__c6xabi_remu)
diff --git a/arch/c6x/lib/strasgi.S b/arch/c6x/lib/strasgi.S
new file mode 100644
index 0000000..de274076
--- /dev/null
+++ b/arch/c6x/lib/strasgi.S
@@ -0,0 +1,89 @@
+;; Copyright 2010 Free Software Foundation, Inc.
+;; Contributed by Bernd Schmidt <bernds@codesourcery.com>.
+;;
+;; This program is free software; you can redistribute it and/or modify
+;; it under the terms of the GNU General Public License as published by
+;; the Free Software Foundation; either version 2 of the License, or
+;; (at your option) any later version.
+;;
+;; This program is distributed in the hope that it will be useful,
+;; but WITHOUT ANY WARRANTY; without even the implied warranty of
+;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+;; GNU General Public License for more details.
+;;
+;; You should have received a copy of the GNU General Public License
+;; along with this program; if not, write to the Free Software
+;; Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+#include <linux/linkage.h>
+
+ .text
+
+ENTRY(__c6xabi_strasgi)
+ ;; This is essentially memcpy, with alignment known to be at least
+ ;; 4, and the size a multiple of 4 greater than or equal to 28.
+ ldw .d2t1 *B4++, A0
+|| mvk .s2 16, B1
+ ldw .d2t1 *B4++, A1
+|| mvk .s2 20, B2
+|| sub .d1 A6, 24, A6
+ ldw .d2t1 *B4++, A5
+ ldw .d2t1 *B4++, A7
+|| mv .l2x A6, B7
+ ldw .d2t1 *B4++, A8
+ ldw .d2t1 *B4++, A9
+|| mv .s2x A0, B5
+|| cmpltu .l2 B2, B7, B0
+
+_strasgi_loop:
+ stw .d1t2 B5, *A4++
+|| [B0] ldw .d2t1 *B4++, A0
+|| mv .s2x A1, B5
+|| mv .l2 B7, B6
+
+ [B0] sub .d2 B6, 24, B7
+|| [B0] b .s2 _strasgi_loop
+|| cmpltu .l2 B1, B6, B0
+
+ [B0] ldw .d2t1 *B4++, A1
+|| stw .d1t2 B5, *A4++
+|| mv .s2x A5, B5
+|| cmpltu .l2 12, B6, B0
+
+ [B0] ldw .d2t1 *B4++, A5
+|| stw .d1t2 B5, *A4++
+|| mv .s2x A7, B5
+|| cmpltu .l2 8, B6, B0
+
+ [B0] ldw .d2t1 *B4++, A7
+|| stw .d1t2 B5, *A4++
+|| mv .s2x A8, B5
+|| cmpltu .l2 4, B6, B0
+
+ [B0] ldw .d2t1 *B4++, A8
+|| stw .d1t2 B5, *A4++
+|| mv .s2x A9, B5
+|| cmpltu .l2 0, B6, B0
+
+ [B0] ldw .d2t1 *B4++, A9
+|| stw .d1t2 B5, *A4++
+|| mv .s2x A0, B5
+|| cmpltu .l2 B2, B7, B0
+
+ ;; loop back branch happens here
+
+ cmpltu .l2 B1, B6, B0
+|| ret .s2 b3
+
+ [B0] stw .d1t1 A1, *A4++
+|| cmpltu .l2 12, B6, B0
+ [B0] stw .d1t1 A5, *A4++
+|| cmpltu .l2 8, B6, B0
+ [B0] stw .d1t1 A7, *A4++
+|| cmpltu .l2 4, B6, B0
+ [B0] stw .d1t1 A8, *A4++
+|| cmpltu .l2 0, B6, B0
+ [B0] stw .d1t1 A9, *A4++
+
+ ;; return happens here
+ENDPROC(__c6xabi_strasgi)
diff --git a/arch/c6x/lib/strasgi_64plus.S b/arch/c6x/lib/strasgi_64plus.S
new file mode 100644
index 0000000..c9fd159
--- /dev/null
+++ b/arch/c6x/lib/strasgi_64plus.S
@@ -0,0 +1,39 @@
+;; Copyright 2010 Free Software Foundation, Inc.
+;; Contributed by Bernd Schmidt <bernds@codesourcery.com>.
+;;
+;; This program is free software; you can redistribute it and/or modify
+;; it under the terms of the GNU General Public License as published by
+;; the Free Software Foundation; either version 2 of the License, or
+;; (at your option) any later version.
+;;
+;; This program is distributed in the hope that it will be useful,
+;; but WITHOUT ANY WARRANTY; without even the implied warranty of
+;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+;; GNU General Public License for more details.
+;;
+;; You should have received a copy of the GNU General Public License
+;; along with this program; if not, write to the Free Software
+;; Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+#include <linux/linkage.h>
+
+ .text
+
+ENTRY(__c6xabi_strasgi_64plus)
+ shru .s2x a6, 2, b31
+|| mv .s1 a4, a30
+|| mv .d2 b4, b30
+
+ add .s2 -4, b31, b31
+
+ sploopd 1
+|| mvc .s2 b31, ilc
+ ldw .d2t2 *b30++, b31
+ nop 4
+ mv .s1x b31,a31
+ spkernel 6, 0
+|| stw .d1t1 a31, *a30++
+
+ ret .s2 b3
+ nop 5
+ENDPROC(__c6xabi_strasgi_64plus)
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v3 20/24] C6X: library code
2011-09-27 20:30 ` [PATCH v3 20/24] C6X: library code Mark Salter
@ 2011-09-28 14:09 ` Arnd Bergmann
0 siblings, 0 replies; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-28 14:09 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch
On Tuesday 27 September 2011, Mark Salter wrote:
> +++ b/arch/c6x/include/asm/checksum.h
> @@ -0,0 +1,115 @@
> +/*
> + * Port on Texas Instruments TMS320C6x architecture
> + *
> + * Copyright (C) 2004, 2009, 2010 Texas Instruments Incorporated
> + * Author: Aurelien Jacquiot (aurelien.jacquiot@jaluna.com)
> + *
> + * This program is free software; you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License version 2 as
> + * published by the Free Software Foundation.
> + */
> +#ifndef _ASM_C6X_CHECKSUM_H
> +#define _ASM_C6X_CHECKSUM_H
> +
> +#include <asm/byteorder.h>
I think you should fall back on the generic checksum.h file and only
provide the declarations that are different. This file is largely
identical to the generic version.
Arnd
^ permalink raw reply [flat|nested] 66+ messages in thread
* [PATCH v3 21/24] C6X: general SoC support
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (19 preceding siblings ...)
2011-09-27 20:30 ` [PATCH v3 20/24] C6X: library code Mark Salter
@ 2011-09-27 20:30 ` Mark Salter
2011-09-28 14:19 ` Arnd Bergmann
2011-09-27 20:30 ` [PATCH v3 22/24] C6X: EMIF - External Memory Interface Mark Salter
` (4 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:30 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
This patch provides a soc_ops struct which provides hooks for SoC functionality
which doesn't fit well into other places.
Signed-off-by: Mark Salter <msalter@redhat.com>
---
arch/c6x/include/asm/soc.h | 35 +++++++++++++++++++
arch/c6x/kernel/soc.c | 78 ++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 113 insertions(+), 0 deletions(-)
create mode 100644 arch/c6x/include/asm/soc.h
create mode 100644 arch/c6x/kernel/soc.c
diff --git a/arch/c6x/include/asm/soc.h b/arch/c6x/include/asm/soc.h
new file mode 100644
index 0000000..43f5015
--- /dev/null
+++ b/arch/c6x/include/asm/soc.h
@@ -0,0 +1,35 @@
+/*
+ * Miscellaneous SoC-specific hooks.
+ *
+ * Copyright (C) 2011 Texas Instruments Incorporated
+ *
+ * Author: Mark Salter <msalter@redhat.com>
+ *
+ * This file is licensed under the terms of the GNU General Public License
+ * version 2. This program is licensed "as is" without any warranty of any
+ * kind, whether express or implied.
+ */
+#ifndef _ASM_C6X_SOC_H
+#define _ASM_C6X_SOC_H
+
+struct soc_ops {
+ /* Return active exception event or -1 if none */
+ int (*get_exception)(void);
+
+ /* Assert an event */
+ void (*assert_event)(unsigned int evt);
+};
+
+extern struct soc_ops soc_ops;
+
+extern int soc_get_exception(void);
+extern void soc_assert_event(unsigned int event);
+extern int soc_mac_addr(unsigned int index, u8 *addr);
+
+/*
+ * for mmio on SoC devices. regs are always same byte order as cpu.
+ */
+#define soc_readl(addr) __raw_readl(addr)
+#define soc_writel(b, addr) __raw_writel((b), (addr))
+
+#endif /* _ASM_C6X_SOC_H */
diff --git a/arch/c6x/kernel/soc.c b/arch/c6x/kernel/soc.c
new file mode 100644
index 0000000..2df9e02
--- /dev/null
+++ b/arch/c6x/kernel/soc.c
@@ -0,0 +1,78 @@
+/*
+ * Miscellaneous SoC-specific hooks.
+ *
+ * Copyright (C) 2011 Texas Instruments Incorporated
+ * Author: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/module.h>
+#include <linux/ctype.h>
+#include <asm/system.h>
+#include <asm/setup.h>
+#include <asm/soc.h>
+
+struct soc_ops soc_ops;
+
+int soc_get_exception(void)
+{
+ if (!soc_ops.get_exception)
+ return -1;
+ return soc_ops.get_exception();
+}
+
+void soc_assert_event(unsigned int evt)
+{
+ if (soc_ops.assert_event)
+ soc_ops.assert_event(evt);
+}
+
+static u8 cmdline_mac[6];
+
+static int __init get_mac_addr_from_cmdline(char *str)
+{
+ int count, i, val;
+
+ for (count = 0; count < 6 && *str; count++, str += 3) {
+ if (!isxdigit(str[0]) || !isxdigit(str[1]))
+ return 0;
+ if (str[2] != ((count < 5) ? ':' : '\0'))
+ return 0;
+
+ for (i = 0, val = 0; i < 2; i++) {
+ val = val << 4;
+ val |= isdigit(str[i]) ?
+ str[i] - '0' : toupper(str[i]) - 'A' + 10;
+ }
+ cmdline_mac[count] = val;
+ }
+ return 1;
+}
+__setup("emac_addr=", get_mac_addr_from_cmdline);
+
+int soc_mac_addr(unsigned int index, u8 *addr)
+{
+ int i, have_cmdline_mac = 0, have_fuse_mac = 0;
+
+ for (i = 0; i < 6; i++) {
+ if (cmdline_mac[i])
+ have_cmdline_mac = 1;
+ if (c6x_fuse_mac[i])
+ have_fuse_mac = 1;
+ }
+
+ /* cmdline overrides hardware fuse MAC */
+ if (have_cmdline_mac)
+ memcpy(addr, cmdline_mac, 6);
+ else if (have_fuse_mac)
+ memcpy(addr, c6x_fuse_mac, 6);
+ else
+ return 0;
+
+ /* adjust for specific EMAC device */
+ addr[5] += index * c6x_num_cores;
+ return 1;
+}
+EXPORT_SYMBOL(soc_mac_addr);
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v3 21/24] C6X: general SoC support
2011-09-27 20:30 ` [PATCH v3 21/24] C6X: general SoC support Mark Salter
@ 2011-09-28 14:19 ` Arnd Bergmann
0 siblings, 0 replies; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-28 14:19 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch
On Tuesday 27 September 2011, Mark Salter wrote:
> +int soc_mac_addr(unsigned int index, u8 *addr)
> +{
> + int i, have_cmdline_mac = 0, have_fuse_mac = 0;
> +
> + for (i = 0; i < 6; i++) {
> + if (cmdline_mac[i])
> + have_cmdline_mac = 1;
> + if (c6x_fuse_mac[i])
> + have_fuse_mac = 1;
> + }
> +
> + /* cmdline overrides hardware fuse MAC */
> + if (have_cmdline_mac)
> + memcpy(addr, cmdline_mac, 6);
> + else if (have_fuse_mac)
> + memcpy(addr, c6x_fuse_mac, 6);
> + else
> + return 0;
IMHO it's more important to let the user override the MAC address through
the device tree than the command line, although if you allow both, the
command line should take precedence. If you think you need to keep the
command line option, please add a way to read it from the device tree
as a fallback before falling back to the fused mac address.
Also, a final fallback on random_mac_address() would be reasonable here.
Finally, how about passing a struct device pointer from the driver, rather
than the index? That would seem more logical and make it easier to get
to the device tree properties.
> + /* adjust for specific EMAC device */
> + addr[5] += index * c6x_num_cores;
> + return 1;
> +}
> +EXPORT_SYMBOL(soc_mac_addr);
EXPORT_SYMBOL_GPL()
Arnd
^ permalink raw reply [flat|nested] 66+ messages in thread
* [PATCH v3 22/24] C6X: EMIF - External Memory Interface
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (20 preceding siblings ...)
2011-09-27 20:30 ` [PATCH v3 21/24] C6X: general SoC support Mark Salter
@ 2011-09-27 20:30 ` Mark Salter
2011-09-28 14:13 ` Arnd Bergmann
2011-09-27 20:30 ` [PATCH v3 23/24] C6X: DSCR - Device State Configuration Registers Mark Salter
` (3 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:30 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
Several SoC parts provide a simple bridge to support external memory mapped
devices. This code probes the device tree for an EMIF node and sets up the
bridge registers if such a node is found. Beyond initial set up, there is no
further need to access the bridge control registers. External devices on the
bus are accessed through their MMIO registers using suitable drivers. The
bridge hardware does provide for timeout and other error interrupts, but these
are not yet supported.
Signed-off-by: Mark Salter <msalter@redhat.com>
---
arch/c6x/platforms/emif.c | 88 +++++++++++++++++++++++++++++++++++++++++++++
1 files changed, 88 insertions(+), 0 deletions(-)
create mode 100644 arch/c6x/platforms/emif.c
diff --git a/arch/c6x/platforms/emif.c b/arch/c6x/platforms/emif.c
new file mode 100644
index 0000000..6a0d4ff
--- /dev/null
+++ b/arch/c6x/platforms/emif.c
@@ -0,0 +1,88 @@
+/*
+ * External Memory Interface
+ *
+ * Copyright (C) 2011 Texas Instruments Incorporated
+ * Author: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/of.h>
+#include <linux/of_address.h>
+#include <linux/io.h>
+#include <asm/soc.h>
+#include <asm/dscr.h>
+
+#define NUM_EMIFA_CHIP_ENABLES 4
+
+struct emifa_regs {
+ u32 midr;
+ u32 stat;
+ u32 reserved1[6];
+ u32 bprio;
+ u32 reserved2[23];
+ u32 ce_config;
+ u32 cecfg[NUM_EMIFA_CHIP_ENABLES];
+ u32 reserved3[4];
+ u32 awcc;
+ u32 reserved4[7];
+ u32 intraw;
+ u32 intmsk;
+ u32 intmskset;
+ u32 intmskclr;
+};
+
+static struct of_device_id emifa_match[] __initdata = {
+ { .compatible = "ti,c64x+emifa" },
+ {}
+};
+
+/*
+ * Parse device tree for existence of an EMIF (External Memory Interface)
+ * and initialize it if found.
+ */
+static int __init c6x_emifa_init(void)
+{
+ struct emifa_regs __iomem *regs;
+ struct device_node *node;
+ const __be32 *p;
+ u32 val;
+ int i, len, err;
+
+ node = of_find_matching_node(NULL, emifa_match);
+ if (!node)
+ return 0;
+
+ regs = of_iomap(node, 0);
+ if (!regs)
+ return 0;
+
+ /* look for a dscr-based enable for emifa pin buffers */
+ err = of_property_read_u32_array(node, "ti,dscr-dev-enable", &val, 1);
+ if (!err)
+ dscr_set_devstate(val, DSCR_DEVSTATE_ENABLED);
+
+ /* set up the chip enables */
+ p = of_get_property(node, "ti,emifa-ce-config", &len);
+ if (p) {
+ len /= sizeof(u32);
+ if (len > NUM_EMIFA_CHIP_ENABLES)
+ len = NUM_EMIFA_CHIP_ENABLES;
+ for (i = 0; i <= len; i++)
+ soc_writel(be32_to_cpup(&p[i]), ®s->cecfg[i]);
+ }
+
+ err = of_property_read_u32_array(node, "ti,emifa-burst-priority", &val, 1);
+ if (!err)
+ soc_writel(val, ®s->bprio);
+
+ err = of_property_read_u32_array(node, "ti,emifa-async-wait-control", &val, 1);
+ if (!err)
+ soc_writel(val, ®s->awcc);
+
+ iounmap(regs);
+ of_node_put(node);
+ return 0;
+}
+pure_initcall(c6x_emifa_init);
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v3 22/24] C6X: EMIF - External Memory Interface
2011-09-27 20:30 ` [PATCH v3 22/24] C6X: EMIF - External Memory Interface Mark Salter
@ 2011-09-28 14:13 ` Arnd Bergmann
0 siblings, 0 replies; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-28 14:13 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch
On Tuesday 27 September 2011, Mark Salter wrote:
> Several SoC parts provide a simple bridge to support external memory mapped
> devices. This code probes the device tree for an EMIF node and sets up the
> bridge registers if such a node is found. Beyond initial set up, there is no
> further need to access the bridge control registers. External devices on the
> bus are accessed through their MMIO registers using suitable drivers. The
> bridge hardware does provide for timeout and other error interrupts, but these
> are not yet supported.
>
> Signed-off-by: Mark Salter <msalter@redhat.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
^ permalink raw reply [flat|nested] 66+ messages in thread
* [PATCH v3 23/24] C6X: DSCR - Device State Configuration Registers
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (21 preceding siblings ...)
2011-09-27 20:30 ` [PATCH v3 22/24] C6X: EMIF - External Memory Interface Mark Salter
@ 2011-09-27 20:30 ` Mark Salter
2011-09-28 14:11 ` Arnd Bergmann
2011-09-27 20:30 ` [PATCH v3 24/24] C6X: MAINTAINERS Mark Salter
` (2 subsequent siblings)
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:30 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
All SoCs provide an area of device configuration registers called the DSCR. The
location of specific registers as well as their use varies considerably from
implementation to implementation. Rather than having to rely on additional
SoC-specific DSCR code for each new supported SoC, this code generalize things
as much as possible using device tree properties. Initialization must take
place early on (setup_arch time) in case the event timer device needs to be
enable via the DSCR.
Signed-off-by: Mark Salter <msalter@redhat.com>
---
arch/c6x/include/asm/dscr.h | 34 +++
arch/c6x/platforms/dscr.c | 598 +++++++++++++++++++++++++++++++++++++++++++
2 files changed, 632 insertions(+), 0 deletions(-)
create mode 100644 arch/c6x/include/asm/dscr.h
create mode 100644 arch/c6x/platforms/dscr.c
diff --git a/arch/c6x/include/asm/dscr.h b/arch/c6x/include/asm/dscr.h
new file mode 100644
index 0000000..561ba83
--- /dev/null
+++ b/arch/c6x/include/asm/dscr.h
@@ -0,0 +1,34 @@
+/*
+ * Copyright (C) 2011 Texas Instruments Incorporated
+ * Author: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ */
+#ifndef _ASM_C6X_DSCR_H
+#define _ASM_C6X_DSCR_H
+
+enum dscr_devstate_t {
+ DSCR_DEVSTATE_ENABLED,
+ DSCR_DEVSTATE_DISABLED,
+};
+
+/*
+ * Set the device state of the device with the given ID.
+ *
+ * Individual drivers should use this to enable or disable the
+ * hardware device. The devid used to identify the device being
+ * controlled should be a property in the device's tree node.
+ */
+extern void dscr_set_devstate(int devid, enum dscr_devstate_t state);
+
+/*
+ * Assert or de-assert an RMII reset.
+ */
+extern void dscr_rmii_reset(int id, int assert);
+
+extern void dscr_probe(void);
+
+#endif /* _ASM_C6X_DSCR_H */
diff --git a/arch/c6x/platforms/dscr.c b/arch/c6x/platforms/dscr.c
new file mode 100644
index 0000000..f848a65
--- /dev/null
+++ b/arch/c6x/platforms/dscr.c
@@ -0,0 +1,598 @@
+/*
+ * Device State Control Registers driver
+ *
+ * Copyright (C) 2011 Texas Instruments Incorporated
+ * Author: Mark Salter <msalter@redhat.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+/*
+ * The Device State Control Registers (DSCR) provide SoC level control over
+ * a number of peripherals. Details vary considerably among the various SoC
+ * parts. In general, the DSCR block will provide one or more configuration
+ * registers often protected by a lock register. One or more key values must
+ * be written to a lock register in order to unlock the configuration register.
+ * The configuration register may be used to enable (and disable in some
+ * cases) SoC pin drivers, peripheral clock sources (internal or pin), etc.
+ * In some cases, a configuration register is write once or the individual
+ * bits are write once. That is, you may be able to enable a device, but
+ * will not be able to disable it.
+ *
+ * In addition to device configuration, the DSCR block may provide registers
+ * which are used to reset SoC peripherals, provide device ID information,
+ * provide MAC addresses, and other miscellaneous functions.
+ */
+
+#include <linux/of.h>
+#include <linux/of_address.h>
+#include <linux/of_platform.h>
+#include <linux/module.h>
+#include <linux/io.h>
+#include <linux/delay.h>
+#include <asm/soc.h>
+#include <asm/dscr.h>
+
+#define MAX_DEVSTATE_IDS 32
+#define MAX_DEVCTL_REGS 8
+#define MAX_DEVSTAT_REGS 8
+#define MAX_LOCKED_REGS 4
+#define MAX_SOC_EMACS 2
+
+struct rmii_reset_reg {
+ u32 reg;
+ u32 mask;
+};
+
+/*
+ * Some registerd may be locked. In order to write to these
+ * registers, the key value must first be written to the lockreg.
+ */
+struct locked_reg {
+ u32 reg; /* offset from base */
+ u32 lockreg; /* offset from base */
+ u32 key; /* unlock key */
+};
+
+/*
+ * This describes a contiguous area of like control bits used to enable/disable
+ * SoC devices. Each controllable device is given an ID which is used by the
+ * individual device drivers to control the device state. These IDs start at
+ * zero and are assigned sequentially to the control bitfield ranges described
+ * by this structure.
+ */
+struct devstate_ctl_reg {
+ u32 reg; /* register holding the control bits */
+ u8 start_id; /* start id of this range */
+ u8 num_ids; /* number of devices in this range */
+ u8 enable_only; /* bits are write-once to enable only */
+ u8 enable; /* value used to enable device */
+ u8 disable; /* value used to disable device */
+ u8 shift; /* starting (rightmost) bit in range */
+ u8 nbits; /* number of bits per device */
+};
+
+
+/*
+ * This describes a region of status bits indicating the state of
+ * various devices. This is used internally to wait for status
+ * change completion when enabling/disabling a device. Status is
+ * optional and not all device controls will have a corresponding
+ * status.
+ */
+struct devstate_stat_reg {
+ u32 reg; /* register holding the status bits */
+ u8 start_id; /* start id of this range */
+ u8 num_ids; /* number of devices in this range */
+ u8 enable; /* value indicating enabled state */
+ u8 disable; /* value indicating disabled state */
+ u8 shift; /* starting (rightmost) bit in range */
+ u8 nbits; /* number of bits per device */
+};
+
+struct devstate_info {
+ struct devstate_ctl_reg *ctl;
+ struct devstate_stat_reg *stat;
+};
+
+/* These are callbacks to SOC-specific code. */
+struct dscr_ops {
+ void (*init)(struct device_node *node);
+};
+
+struct dscr_regs {
+ spinlock_t lock;
+ void __iomem *base;
+ u32 kick_reg[2];
+ u32 kick_key[2];
+ struct locked_reg locked[MAX_LOCKED_REGS];
+ struct devstate_info devstate_info[MAX_DEVSTATE_IDS];
+ struct rmii_reset_reg rmii_resets[MAX_SOC_EMACS];
+ struct devstate_ctl_reg devctl[MAX_DEVCTL_REGS];
+ struct devstate_stat_reg devstat[MAX_DEVSTAT_REGS];
+};
+
+static struct dscr_regs dscr;
+
+static struct locked_reg *find_locked_reg(u32 reg)
+{
+ int i;
+
+ for (i = 0; i < MAX_LOCKED_REGS; i++)
+ if (dscr.locked[i].key && reg == dscr.locked[i].reg)
+ return &dscr.locked[i];
+ return NULL;
+}
+
+/*
+ * Write to a register with one lock
+ */
+static void dscr_write_locked1(u32 reg, u32 val,
+ u32 lock, u32 key)
+{
+ void __iomem *reg_addr = dscr.base + reg;
+ void __iomem *lock_addr = dscr.base + lock;
+
+ /*
+ * For some registers, the lock is relocked after a short number
+ * of cycles. We have to put the lock write and register write in
+ * the same fetch packet to meet this timing. The .align ensures
+ * the two stw instructions are in the same fetch packet.
+ */
+ asm volatile ("b .s2 0f\n"
+ "nop 5\n"
+ " .align 5\n"
+ "0:\n"
+ "stw .D1T2 %3,*%2\n"
+ "stw .D1T2 %1,*%0\n"
+ :
+ : "a"(reg_addr), "b"(val), "a"(lock_addr), "b"(key)
+ );
+
+ /* in case the hw doesn't reset the lock */
+ soc_writel(0, lock_addr);
+}
+
+/*
+ * Write to a register protected by two lock registers
+ */
+static void dscr_write_locked2(u32 reg, u32 val,
+ u32 lock0, u32 key0,
+ u32 lock1, u32 key1)
+{
+ soc_writel(key0, dscr.base + lock0);
+ soc_writel(key1, dscr.base + lock1);
+ soc_writel(val, dscr.base + reg);
+ soc_writel(0, dscr.base + lock0);
+ soc_writel(0, dscr.base + lock1);
+}
+
+static void dscr_write(u32 reg, u32 val)
+{
+ struct locked_reg *lock;
+
+ lock = find_locked_reg(reg);
+ if (lock)
+ dscr_write_locked1(reg, val, lock->lockreg, lock->key);
+ else if (dscr.kick_key[0])
+ dscr_write_locked2(reg, val, dscr.kick_reg[0], dscr.kick_key[0],
+ dscr.kick_reg[1], dscr.kick_key[1]);
+ else
+ soc_writel(val, dscr.base + reg);
+}
+
+
+/*
+ * Drivers can use this interface to enable/disable SoC IP blocks.
+ */
+void dscr_set_devstate(int id, enum dscr_devstate_t state)
+{
+ struct devstate_ctl_reg *ctl;
+ struct devstate_stat_reg *stat;
+ struct devstate_info *info;
+ u32 ctl_val, val;
+ int ctl_shift, ctl_mask;
+ unsigned long flags;
+
+ if (!dscr.base)
+ return;
+
+ if (id < 0 || id >= MAX_DEVSTATE_IDS)
+ return;
+
+ info = &dscr.devstate_info[id];
+ ctl = info->ctl;
+ stat = info->stat;
+
+ if (ctl == NULL)
+ return;
+
+ ctl_shift = ctl->shift + ctl->nbits * (id - ctl->start_id);
+ ctl_mask = ((1 << ctl->nbits) - 1) << ctl_shift;
+
+ switch (state) {
+ case DSCR_DEVSTATE_ENABLED:
+ ctl_val = ctl->enable << ctl_shift;
+ break;
+ case DSCR_DEVSTATE_DISABLED:
+ if (ctl->enable_only)
+ return;
+ ctl_val = ctl->disable << ctl_shift;
+ break;
+ default:
+ return;
+ }
+
+ spin_lock_irqsave(&dscr.lock, flags);
+
+ val = soc_readl(dscr.base + ctl->reg);
+ val &= ~ctl_mask;
+ val |= ctl_val;
+
+ dscr_write(ctl->reg, val);
+
+ spin_unlock_irqrestore(&dscr.lock, flags);
+
+ if (!stat)
+ return;
+
+ ctl_shift = stat->shift + stat->nbits * (id - stat->start_id);
+
+ if (state == DSCR_DEVSTATE_ENABLED)
+ ctl_val = stat->enable;
+ else
+ ctl_val = stat->disable;
+
+ do {
+ val = soc_readl(dscr.base + stat->reg);
+ val >>= ctl_shift;
+ val &= ((1 << stat->nbits) - 1);
+ } while (val != ctl_val);
+}
+EXPORT_SYMBOL(dscr_set_devstate);
+
+/*
+ * Drivers can use this to reset RMII module.
+ */
+void dscr_rmii_reset(int id, int assert)
+{
+ struct rmii_reset_reg *r;
+ unsigned long flags;
+ u32 val;
+
+ if (id < 0 || id >= MAX_SOC_EMACS)
+ return;
+
+ r = &dscr.rmii_resets[id];
+ if (r->mask == 0)
+ return;
+
+ spin_lock_irqsave(&dscr.lock, flags);
+
+ val = soc_readl(dscr.base + r->reg);
+ if (assert)
+ dscr_write(r->reg, val | r->mask);
+ else
+ dscr_write(r->reg, val & ~(r->mask));
+
+ spin_unlock_irqrestore(&dscr.lock, flags);
+}
+EXPORT_SYMBOL(dscr_rmii_reset);
+
+static void __init dscr_parse_devstat(struct device_node *node,
+ void __iomem *base)
+{
+ u32 val;
+ int err;
+
+ err = of_property_read_u32_array(node, "ti,dscr-devstat", &val, 1);
+ if (!err)
+ c6x_devstat = soc_readl(base + val);
+ printk(KERN_INFO "DEVSTAT: %08x\n", c6x_devstat);
+}
+
+static void __init dscr_parse_silicon_rev(struct device_node *node,
+ void __iomem *base)
+{
+ u32 vals[3];
+ int err;
+
+ err = of_property_read_u32_array(node, "ti,dscr-silicon-rev", vals, 3);
+ if (!err) {
+ c6x_silicon_rev = soc_readl(base + vals[0]);
+ c6x_silicon_rev >>= vals[1];
+ c6x_silicon_rev &= vals[2];
+ }
+}
+
+/*
+ * Some SoCs will have a pair of fuse registers which hold
+ * an ethernet MAC address. The "ti,dscr-mac-fuse-regs"
+ * property is a mapping from fuse register bytes to MAC
+ * address bytes. The expected format is:
+ *
+ * ti,dscr-mac-fuse-regs = <reg0 b3 b2 b1 b0
+ * reg1 b3 b2 b1 b0>
+ *
+ * reg0 and reg1 are the offsets of the two fuse registers.
+ * b3-b0 positionally represent bytes within the fuse register.
+ * b3 is the most significant byte and b0 is the least.
+ * Allowable values for b3-b0 are:
+ *
+ * 0 = fuse register byte not used in MAC address
+ * 1-6 = index+1 into c6x_fuse_mac[]
+ */
+static void __init dscr_parse_mac_fuse(struct device_node *node,
+ void __iomem *base)
+{
+ u32 vals[10], fuse;
+ int f, i, j, err;
+
+ err = of_property_read_u32_array(node, "ti,dscr-mac-fuse-regs",
+ vals, 10);
+ if (err)
+ return;
+
+ for (f = 0; f < 2; f++) {
+ fuse = soc_readl(base + vals[f * 5]);
+ for (j = (f * 5) + 1, i = 24; i >= 0; i -= 8, j++)
+ if (vals[j] && vals[j] <= 6)
+ c6x_fuse_mac[vals[j] - 1] = fuse >> i;
+ }
+}
+
+static void __init dscr_parse_rmii_resets(struct device_node *node,
+ void __iomem *base)
+{
+ const __be32 *p;
+ int i, size;
+
+ /* look for RMII reset registers */
+ p = of_get_property(node, "ti,dscr-rmii-resets", &size);
+ if (p) {
+ /* parse all the reg/mask pairs we can handle */
+ size /= (sizeof(*p) * 2);
+ if (size > MAX_SOC_EMACS)
+ size = MAX_SOC_EMACS;
+
+ for (i = 0; i < size; i++) {
+ dscr.rmii_resets[i].reg = be32_to_cpup(p++);
+ dscr.rmii_resets[i].mask = be32_to_cpup(p++);
+ }
+ }
+}
+
+
+static void __init dscr_parse_privperm(struct device_node *node,
+ void __iomem *base)
+{
+ u32 vals[2];
+ int err;
+
+ err = of_property_read_u32_array(node, "ti,dscr-privperm", vals, 2);
+ if (err)
+ return;
+ dscr_write(vals[0], vals[1]);
+}
+
+/*
+ * SoCs may have "locked" DSCR registers which can only be written
+ * to only after writing a key value to a lock registers. These
+ * regisers can be described with the "ti,dscr-locked-regs" property.
+ * This property provides a list of register descriptions with each
+ * description consisting of three values.
+ *
+ * ti,dscr-locked-regs = <reg0 lockreg0 key0
+ * ...
+ * regN lockregN keyN>;
+ *
+ * reg is the offset of the locked register
+ * lockreg is the offset of the lock register
+ * key is the unlock key written to lockreg
+ *
+ */
+static void __init dscr_parse_locked_regs(struct device_node *node,
+ void __iomem *base)
+{
+ struct locked_reg *r;
+ const __be32 *p;
+ int i, size;
+
+ p = of_get_property(node, "ti,dscr-locked-regs", &size);
+ if (p) {
+ /* parse all the register descriptions we can handle */
+ size /= (sizeof(*p) * 3);
+ if (size > MAX_LOCKED_REGS)
+ size = MAX_LOCKED_REGS;
+
+ for (i = 0; i < size; i++) {
+ r = &dscr.locked[i];
+
+ r->reg = be32_to_cpup(p++);
+ r->lockreg = be32_to_cpup(p++);
+ r->key = be32_to_cpup(p++);
+ }
+ }
+}
+
+/*
+ * SoCs may have DSCR registers which are only write enabled after
+ * writing specific key values to two registers. The two key registers
+ * and the key values can be parsed from a "ti,dscr-kick-regs"
+ * propety with the following layout:
+ *
+ * ti,dscr-kick-regs = <kickreg0 key0 kickreg1 key1>
+ *
+ * kickreg is the offset of the "kick" register
+ * key is the value which unlocks writing for protected regs
+ */
+static void __init dscr_parse_kick_regs(struct device_node *node,
+ void __iomem *base)
+{
+ u32 vals[4];
+ int err;
+
+ err = of_property_read_u32_array(node, "ti,dscr-kick-regs", vals, 4);
+ if (!err) {
+ dscr.kick_reg[0] = vals[0];
+ dscr.kick_key[0] = vals[1];
+ dscr.kick_reg[1] = vals[2];
+ dscr.kick_key[1] = vals[3];
+ }
+}
+
+
+/*
+ * SoCs may provide controls to enable/disable individual IP blocks. These
+ * controls in the DSCR usually control pin drivers but also may control
+ * clocking and or resets. The device tree is used to describe the bitfields
+ * in registers used to control device state. The number of bits and their
+ * values may vary even within the same register.
+ *
+ * The layout of these bitfields is described by the ti,dscr-devstate-ctl-regs
+ * property. This property is a list where each element describes a contiguous
+ * range of control fields with like properties. Each element of the list
+ * consists of 7 cells with the following values:
+ *
+ * start_id num_ids reg enable disable start_bit nbits
+ *
+ * start_id is device id for the first device control in the range
+ * num_ids is the number of device controls in the range
+ * reg is the offset of the register holding the control bits
+ * enable is the value to enable a device
+ * disable is the value to disable a device (0xffffffff if cannot disable)
+ * start_bit is the bit number of the first bit in the range
+ * nbits is the number of bits per device control
+ */
+static void __init dscr_parse_devstate_ctl_regs(struct device_node *node,
+ void __iomem *base)
+{
+ struct devstate_ctl_reg *r;
+ const __be32 *p;
+ int i, j, size;
+
+ p = of_get_property(node, "ti,dscr-devstate-ctl-regs", &size);
+ if (p) {
+ /* parse all the ranges we can handle */
+ size /= (sizeof(*p) * 7);
+ if (size > MAX_DEVCTL_REGS)
+ size = MAX_DEVCTL_REGS;
+
+ for (i = 0; i < size; i++) {
+ r = &dscr.devctl[i];
+
+ r->start_id = be32_to_cpup(p++);
+ r->num_ids = be32_to_cpup(p++);
+ r->reg = be32_to_cpup(p++);
+ r->enable = be32_to_cpup(p++);
+ r->disable = be32_to_cpup(p++);
+ if (r->disable == 0xffffffff)
+ r->enable_only = 1;
+ r->shift = be32_to_cpup(p++);
+ r->nbits = be32_to_cpup(p++);
+
+ for (j = r->start_id;
+ j < (r->start_id + r->num_ids);
+ j++)
+ dscr.devstate_info[j].ctl = r;
+ }
+ }
+}
+
+/*
+ * SoCs may provide status registers indicating the state (enabled/disabled) of
+ * devices on the SoC. The device tree is used to describe the bitfields in
+ * registers used to provide device status. The number of bits and their
+ * values used to provide status may vary even within the same register.
+ *
+ * The layout of these bitfields is described by the ti,dscr-devstate-stat-regs
+ * property. This property is a list where each element describes a contiguous
+ * range of status fields with like properties. Each element of the list
+ * consists of 7 cells with the following values:
+ *
+ * start_id num_ids reg enable disable start_bit nbits
+ *
+ * start_id is device id for the first device status in the range
+ * num_ids is the number of devices covered by the range
+ * reg is the offset of the register holding the status bits
+ * enable is the value indicating device is enabled
+ * disable is the value indicating device is disabled
+ * start_bit is the bit number of the first bit in the range
+ * nbits is the number of bits per device status
+ */
+static void __init dscr_parse_devstate_stat_regs(struct device_node *node,
+ void __iomem *base)
+{
+ struct devstate_stat_reg *r;
+ const __be32 *p;
+ int i, j, size;
+
+ p = of_get_property(node, "ti,dscr-devstate-stat-regs", &size);
+ if (p) {
+ /* parse all the ranges we can handle */
+ size /= (sizeof(*p) * 7);
+ if (size > MAX_DEVSTAT_REGS)
+ size = MAX_DEVSTAT_REGS;
+
+ for (i = 0; i < size; i++) {
+ r = &dscr.devstat[i];
+
+ r->start_id = be32_to_cpup(p++);
+ r->num_ids = be32_to_cpup(p++);
+ r->reg = be32_to_cpup(p++);
+ r->enable = be32_to_cpup(p++);
+ r->disable = be32_to_cpup(p++);
+ r->shift = be32_to_cpup(p++);
+ r->nbits = be32_to_cpup(p++);
+
+ for (j = r->start_id;
+ j < (r->start_id + r->num_ids);
+ j++)
+ dscr.devstate_info[j].stat = r;
+ }
+ }
+}
+
+static struct of_device_id dscr_ids[] __initdata = {
+ { .compatible = "ti,c64x+dscr" },
+ {}
+};
+
+/*
+ * Probe for DSCR area.
+ *
+ * This has to be done early on in case timer or interrupt controller
+ * needs something. e.g. On C6455 SoC, timer must be enabled through
+ * DSCR before it is functional.
+ */
+void __init dscr_probe(void)
+{
+ struct device_node *node;
+ void __iomem *base;
+
+ spin_lock_init(&dscr.lock);
+
+ node = of_find_matching_node(NULL, dscr_ids);
+ if (!node)
+ return;
+
+ base = of_iomap(node, 0);
+ if (!base) {
+ of_node_put(node);
+ return;
+ }
+
+ dscr.base = base;
+
+ dscr_parse_devstat(node, base);
+ dscr_parse_silicon_rev(node, base);
+ dscr_parse_mac_fuse(node, base);
+ dscr_parse_rmii_resets(node, base);
+ dscr_parse_locked_regs(node, base);
+ dscr_parse_kick_regs(node, base);
+ dscr_parse_devstate_ctl_regs(node, base);
+ dscr_parse_devstate_stat_regs(node, base);
+ dscr_parse_privperm(node, base);
+}
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v3 23/24] C6X: DSCR - Device State Configuration Registers
2011-09-27 20:30 ` [PATCH v3 23/24] C6X: DSCR - Device State Configuration Registers Mark Salter
@ 2011-09-28 14:11 ` Arnd Bergmann
0 siblings, 0 replies; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-28 14:11 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch
On Tuesday 27 September 2011, Mark Salter wrote:
>
> All SoCs provide an area of device configuration registers called the DSCR. The
> location of specific registers as well as their use varies considerably from
> implementation to implementation. Rather than having to rely on additional
> SoC-specific DSCR code for each new supported SoC, this code generalize things
> as much as possible using device tree properties. Initialization must take
> place early on (setup_arch time) in case the event timer device needs to be
> enable via the DSCR.
>
> Signed-off-by: Mark Salter <msalter@redhat.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
^ permalink raw reply [flat|nested] 66+ messages in thread
* [PATCH v3 24/24] C6X: MAINTAINERS
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (22 preceding siblings ...)
2011-09-27 20:30 ` [PATCH v3 23/24] C6X: DSCR - Device State Configuration Registers Mark Salter
@ 2011-09-27 20:30 ` Mark Salter
2011-09-28 14:09 ` Arnd Bergmann
2011-09-27 23:15 ` [PATCH v3 00/24] C6X: New architecture Stephen Rothwell
2011-09-28 21:49 ` Valdis.Kletnieks
25 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-27 20:30 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-arch, Mark Salter
Signed-off-by: Mark Salter <msalter@redhat.com>
---
MAINTAINERS | 8 ++++++++
1 files changed, 8 insertions(+), 0 deletions(-)
diff --git a/MAINTAINERS b/MAINTAINERS
index 28f65c2..adc3e81 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -1626,6 +1626,14 @@ T: git git://git.alsa-project.org/alsa-kernel.git
S: Maintained
F: sound/pci/oxygen/
+C6X ARCHITECTURE
+M: Mark Salter <msalter@redhat.com>
+M: Aurelien Jacquiot <a-jacquiot@ti.com>
+L: linux-c6x-dev@linux-c6x.org
+W: http://www.linux-c6x.org/wiki/index.php/Main_Page
+S: Maintained
+F: arch/c6x/
+
CACHEFILES: FS-CACHE BACKEND FOR CACHING ON MOUNTED FILESYSTEMS
M: David Howells <dhowells@redhat.com>
L: linux-cachefs@redhat.com
--
1.7.6.2
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v3 00/24] C6X: New architecture
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (23 preceding siblings ...)
2011-09-27 20:30 ` [PATCH v3 24/24] C6X: MAINTAINERS Mark Salter
@ 2011-09-27 23:15 ` Stephen Rothwell
2011-09-28 14:33 ` Arnd Bergmann
2011-09-28 21:49 ` Valdis.Kletnieks
25 siblings, 1 reply; 66+ messages in thread
From: Stephen Rothwell @ 2011-09-27 23:15 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch, Arnd Bergmann
[-- Attachment #1: Type: text/plain, Size: 877 bytes --]
On Tue, 27 Sep 2011 16:29:41 -0400 Mark Salter <msalter@redhat.com> wrote:
>
> Here is v3 of the C6X architecture patch series. Patches are
> available in 'upstream-v3' branch at:
>
> git@linux-c6x.org:/git/projects/linux-c6x-upstreaming
Is it time for this to be included in linux-next (i.e. are you intending
to send this to Linus during the next merge window)? If so, please send
a request (cc'd to here, linux-next and linux-kernel) asking for
inclusion. I would prefer a publically fetchable tree (the tree above is
(implicitly) fetched over ssh and is so not generally available) and
branch name that does not change over time (like "for-next" or something)
but it can just be an alias for the above.
Arnd, do you have an opinion on this?
--
Cheers,
Stephen Rothwell sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/
[-- Attachment #2: Type: application/pgp-signature, Size: 490 bytes --]
^ permalink raw reply [flat|nested] 66+ messages in thread* Re: [PATCH v3 00/24] C6X: New architecture
2011-09-27 23:15 ` [PATCH v3 00/24] C6X: New architecture Stephen Rothwell
@ 2011-09-28 14:33 ` Arnd Bergmann
0 siblings, 0 replies; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-28 14:33 UTC (permalink / raw)
To: Stephen Rothwell; +Cc: Mark Salter, linux-kernel, linux-arch, Thomas Gleixner
On Wednesday 28 September 2011, Stephen Rothwell wrote:
> Show Details
> On Tue, 27 Sep 2011 16:29:41 -0400 Mark Salter <msalter@redhat.com> wrote:
> >
> > Here is v3 of the C6X architecture patch series. Patches are
> > available in 'upstream-v3' branch at:
> >
> > git@linux-c6x.org:/git/projects/linux-c6x-upstreaming
>
> Is it time for this to be included in linux-next (i.e. are you intending
> to send this to Linus during the next merge window)? If so, please send
> a request (cc'd to here, linux-next and linux-kernel) asking for
> inclusion. I would prefer a publically fetchable tree (the tree above is
> (implicitly) fetched over ssh and is so not generally available) and
> branch name that does not change over time (like "for-next" or something)
> but it can just be an alias for the above.
>
> Arnd, do you have an opinion on this?
>From my perspective it can go into -next now. All the issues I have
reported in version 2 have been fixed, and the comments I made this
time are relatively insignificant, so I trust that Mark can fix those
either between now and the merge window or provide updates during the
3.2 bug fix phase. The only thing that absolutely needs to get fixed
before 3.2.0 IMO is the ptrace ABI since we can no longer change that
after a release, but it should also be a trivial change.
Thomas, you also had some important comments, please complain if you
see those as show-stoppers for integration into linux-next.
Arnd
^ permalink raw reply [flat|nested] 66+ messages in thread
* Re: [PATCH v3 00/24] C6X: New architecture
2011-09-27 20:29 [PATCH v3 00/24] C6X: New architecture Mark Salter
` (24 preceding siblings ...)
2011-09-27 23:15 ` [PATCH v3 00/24] C6X: New architecture Stephen Rothwell
@ 2011-09-28 21:49 ` Valdis.Kletnieks
2011-09-29 10:33 ` Arnd Bergmann
25 siblings, 1 reply; 66+ messages in thread
From: Valdis.Kletnieks @ 2011-09-28 21:49 UTC (permalink / raw)
To: Mark Salter; +Cc: linux-kernel, linux-arch
[-- Attachment #1: Type: text/plain, Size: 449 bytes --]
On Tue, 27 Sep 2011 16:29:41 EDT, Mark Salter said:
> This architecture supports members of the Texas Instruments family
> of C6x single and multicore DSPs. The multicore DSPs do not support
> cache coherancy, so are not suitable for SMP.
Is there a usage model for the multicore? I know somebody had some patches for
"HPC dedicated compute cores" that would just basically run a userspace process
and that's it - would those be applicable here?
[-- Attachment #2: Type: application/pgp-signature, Size: 227 bytes --]
^ permalink raw reply [flat|nested] 66+ messages in thread* Re: [PATCH v3 00/24] C6X: New architecture
2011-09-28 21:49 ` Valdis.Kletnieks
@ 2011-09-29 10:33 ` Arnd Bergmann
2011-09-29 12:21 ` Mark Salter
0 siblings, 1 reply; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-29 10:33 UTC (permalink / raw)
To: Valdis.Kletnieks; +Cc: Mark Salter, linux-kernel, linux-arch
On Wednesday 28 September 2011, Valdis.Kletnieks@vt.edu wrote:
> Show Details
> On Tue, 27 Sep 2011 16:29:41 EDT, Mark Salter said:
>
> > This architecture supports members of the Texas Instruments family
> > of C6x single and multicore DSPs. The multicore DSPs do not support
> > cache coherancy, so are not suitable for SMP.
>
> Is there a usage model for the multicore? I know somebody had some patches for
> "HPC dedicated compute cores" that would just basically run a userspace process
> and that's it - would those be applicable here?
No, that's a different thing. Even with dedicated compute cores, you need
cache coherency.
Arnd
^ permalink raw reply [flat|nested] 66+ messages in thread
* Re: [PATCH v3 00/24] C6X: New architecture
2011-09-29 10:33 ` Arnd Bergmann
@ 2011-09-29 12:21 ` Mark Salter
2011-09-29 12:42 ` Valdis.Kletnieks
0 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-29 12:21 UTC (permalink / raw)
To: Arnd Bergmann; +Cc: Valdis.Kletnieks, linux-kernel, linux-arch
On Thu, 2011-09-29 at 12:33 +0200, Arnd Bergmann wrote:
> On Wednesday 28 September 2011, Valdis.Kletnieks@vt.edu wrote:
> > Show Details
> > On Tue, 27 Sep 2011 16:29:41 EDT, Mark Salter said:
> >
> > > This architecture supports members of the Texas Instruments family
> > > of C6x single and multicore DSPs. The multicore DSPs do not support
> > > cache coherancy, so are not suitable for SMP.
> >
> > Is there a usage model for the multicore? I know somebody had some patches for
> > "HPC dedicated compute cores" that would just basically run a userspace process
> > and that's it - would those be applicable here?
>
> No, that's a different thing. Even with dedicated compute cores, you need
> cache coherency.
One thing that we plan to push later is an XIP model where cores can
share a
single kernel text image but run with their own copies of data. So you
end up
with something like a loosely coupled multiprocessor system with some
hardware
support for multicore communication and peripheral sharing.
--Mark
^ permalink raw reply [flat|nested] 66+ messages in thread
* Re: [PATCH v3 00/24] C6X: New architecture
2011-09-29 12:21 ` Mark Salter
@ 2011-09-29 12:42 ` Valdis.Kletnieks
2011-09-29 13:02 ` Mark Salter
0 siblings, 1 reply; 66+ messages in thread
From: Valdis.Kletnieks @ 2011-09-29 12:42 UTC (permalink / raw)
To: Mark Salter; +Cc: Arnd Bergmann, linux-kernel, linux-arch
[-- Attachment #1: Type: text/plain, Size: 806 bytes --]
On Thu, 29 Sep 2011 08:21:58 EDT, Mark Salter said:
> One thing that we plan to push later is an XIP model where cores can share a
> single kernel text image but run with their own copies of data. So you end up
> with something like a loosely coupled multiprocessor system with some hardware
> support for multicore communication and peripheral sharing.
Ah, OK. I eventually came across a mention in one of the patches of running
a separate instance of Linux on each core, whick left me even more mystified.
But the above clarifies it, thanks...
I wonder how many currently global variables will need to be moved to per_cpu
structures to make it work, and how intrusive that will end up. Or is the plan
to have the boot loader simply plonk the entire .data/.bss in different locations
for each core?
[-- Attachment #2: Type: application/pgp-signature, Size: 227 bytes --]
^ permalink raw reply [flat|nested] 66+ messages in thread
* Re: [PATCH v3 00/24] C6X: New architecture
2011-09-29 12:42 ` Valdis.Kletnieks
@ 2011-09-29 13:02 ` Mark Salter
2011-09-29 13:18 ` Arnd Bergmann
0 siblings, 1 reply; 66+ messages in thread
From: Mark Salter @ 2011-09-29 13:02 UTC (permalink / raw)
To: Valdis.Kletnieks; +Cc: Arnd Bergmann, linux-kernel, linux-arch
On Thu, 2011-09-29 at 08:42 -0400, Valdis.Kletnieks@vt.edu wrote:
> On Thu, 29 Sep 2011 08:21:58 EDT, Mark Salter said:
>
> > One thing that we plan to push later is an XIP model where cores can share a
> > single kernel text image but run with their own copies of data. So you end up
> > with something like a loosely coupled multiprocessor system with some hardware
> > support for multicore communication and peripheral sharing.
>
> Ah, OK. I eventually came across a mention in one of the patches of running
> a separate instance of Linux on each core, whick left me even more mystified.
> But the above clarifies it, thanks...
>
> I wonder how many currently global variables will need to be moved to per_cpu
> structures to make it work, and how intrusive that will end up. Or is the plan
> to have the boot loader simply plonk the entire .data/.bss in different locations
> for each core?
>
What we have prototyped is more like the latter. The XIP image (binary
kernel block) is in ROM. Bootloader boots the first core by jumping to
kernel image. Kernel copies initialized data to its reserved chunk of
RAM and runs from there. The other kernels can be started by the boot
kernel and they too will jump to the ROM image and make their own copies
of all data. Without cache coherency, sharing any cached data is
problematic.
Other models are also possible. For instance, other cores could run
dedicated DSP applications using some other minimal OS with HW supported
communication with the Linux OS acting as the overall manager.
--Mark
^ permalink raw reply [flat|nested] 66+ messages in thread
* Re: [PATCH v3 00/24] C6X: New architecture
2011-09-29 13:02 ` Mark Salter
@ 2011-09-29 13:18 ` Arnd Bergmann
0 siblings, 0 replies; 66+ messages in thread
From: Arnd Bergmann @ 2011-09-29 13:18 UTC (permalink / raw)
To: Mark Salter; +Cc: Valdis.Kletnieks, linux-kernel, linux-arch
On Thursday 29 September 2011 09:02:11 Mark Salter wrote:
>
> Other models are also possible. For instance, other cores could run
> dedicated DSP applications using some other minimal OS with HW supported
> communication with the Linux OS acting as the overall manager.
Yes, that makes a lot of sense. I guess if you ever want to turn c6x
into something more capable, adding an MMU would be higher on the
priority list than proper SMP, which requires both cache coherency
and atomic operations.
Arnd
^ permalink raw reply [flat|nested] 66+ messages in thread