LinuxPPC-Dev Archive on lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH 14/20] powerpc: Add memory management headers for new 64-bit BookE
From: Benjamin Herrenschmidt @ 2009-07-23  5:59 UTC (permalink / raw)
  To: linuxppc-dev

This adds the PTE and pgtable format definitions, along with changes
to the kernel memory map and other definitions related to implementing
support for 64-bit Book3E. This also shields some asm-offset bits that
are currently only relevant on 32-bit

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---

 arch/powerpc/include/asm/mmu-book3e.h    |   50 ++++++++++++++++++++++
 arch/powerpc/include/asm/mmu.h           |    1 
 arch/powerpc/include/asm/page.h          |    4 +
 arch/powerpc/include/asm/page_64.h       |   10 ++++
 arch/powerpc/include/asm/pgtable-ppc64.h |   61 +++++++++++++++++++--------
 arch/powerpc/include/asm/pte-book3e.h    |   70 +++++++++++++++++++++++++++++++
 arch/powerpc/include/asm/pte-common.h    |    3 +
 arch/powerpc/kernel/asm-offsets.c        |    5 +-
 8 files changed, 186 insertions(+), 18 deletions(-)

--- /dev/null	1970-01-01 00:00:00.000000000 +0000
+++ linux-work/arch/powerpc/include/asm/pte-book3e.h	2009-07-22 16:40:20.000000000 +1000
@@ -0,0 +1,70 @@
+#ifndef _ASM_POWERPC_PTE_BOOK3E_H
+#define _ASM_POWERPC_PTE_BOOK3E_H
+#ifdef __KERNEL__
+
+/* PTE bit definitions for processors compliant to the Book3E
+ * architecture 2.06 or later. The position of the PTE bits
+ * matches the HW definition of the optional Embedded Page Table
+ * category.
+ */
+
+/* Architected bits */
+#define _PAGE_PRESENT	0x000001 /* software: pte contains a translation */
+#define _PAGE_FILE	0x000002 /* (!present only) software: pte holds file offset */
+#define _PAGE_SW1	0x000002
+#define _PAGE_BAP_SR	0x000004
+#define _PAGE_BAP_UR	0x000008
+#define _PAGE_BAP_SW	0x000010
+#define _PAGE_BAP_UW	0x000020
+#define _PAGE_BAP_SX	0x000040
+#define _PAGE_BAP_UX	0x000080
+#define _PAGE_PSIZE_MSK	0x000f00
+#define _PAGE_PSIZE_4K	0x000200
+#define _PAGE_PSIZE_64K	0x000600
+#define _PAGE_PSIZE_1M	0x000a00
+#define _PAGE_PSIZE_16M	0x000e00
+#define _PAGE_DIRTY	0x001000 /* C: page changed */
+#define _PAGE_SW0	0x002000
+#define _PAGE_U3	0x004000
+#define _PAGE_U2	0x008000
+#define _PAGE_U1	0x010000
+#define _PAGE_U0	0x020000
+#define _PAGE_ACCESSED	0x040000
+#define _PAGE_LENDIAN	0x080000
+#define _PAGE_GUARDED	0x100000
+#define _PAGE_COHERENT	0x200000 /* M: enforce memory coherence */
+#define _PAGE_NO_CACHE	0x400000 /* I: cache inhibit */
+#define _PAGE_WRITETHRU	0x800000 /* W: cache write-through */
+
+/* "Higher level" linux bit combinations */
+#define _PAGE_EXEC	_PAGE_BAP_SX /* Can be executed from potentially */
+#define _PAGE_HWEXEC	_PAGE_BAP_UX /* .. and was cache cleaned */
+#define _PAGE_RW	(_PAGE_BAP_SW | _PAGE_BAP_UW) /* User write permission */
+#define _PAGE_KERNEL_RW	(_PAGE_BAP_SW | _PAGE_BAP_SR | _PAGE_DIRTY)
+#define _PAGE_KERNEL_RO	(_PAGE_BAP_SR)
+#define _PAGE_USER	(_PAGE_BAP_UR | _PAGE_BAP_SR) /* Can be read */
+
+#define _PAGE_HASHPTE	0
+#define _PAGE_BUSY	0
+
+#define _PAGE_SPECIAL	_PAGE_SW0
+
+/* Flags to be preserved on PTE modifications */
+#define _PAGE_HPTEFLAGS	_PAGE_BUSY
+
+/* Base page size */
+#ifdef CONFIG_PPC_64K_PAGES
+#define _PAGE_PSIZE	_PAGE_PSIZE_64K
+#define PTE_RPN_SHIFT	(28)
+#else
+#define _PAGE_PSIZE	_PAGE_PSIZE_4K
+#define	PTE_RPN_SHIFT	(24)
+#endif
+
+/* On 32-bit, we never clear the top part of the PTE */
+#ifdef CONFIG_PPC32
+#define _PTE_NONE_MASK	0xffffffff00000000ULL
+#endif
+
+#endif /* __KERNEL__ */
+#endif /*  _ASM_POWERPC_PTE_FSL_BOOKE_H */
Index: linux-work/arch/powerpc/include/asm/pgtable-ppc64.h
===================================================================
--- linux-work.orig/arch/powerpc/include/asm/pgtable-ppc64.h	2009-07-22 16:37:24.000000000 +1000
+++ linux-work/arch/powerpc/include/asm/pgtable-ppc64.h	2009-07-22 16:40:20.000000000 +1000
@@ -5,11 +5,6 @@
  * the ppc64 hashed page table.
  */
 
-#ifndef __ASSEMBLY__
-#include <linux/stddef.h>
-#include <asm/tlbflush.h>
-#endif /* __ASSEMBLY__ */
-
 #ifdef CONFIG_PPC_64K_PAGES
 #include <asm/pgtable-ppc64-64k.h>
 #else
@@ -38,26 +33,46 @@
 #endif
 
 /*
- * Define the address range of the vmalloc VM area.
+ * Define the address range of the kernel non-linear virtual area
+ */
+
+#ifdef CONFIG_PPC_BOOK3E
+#define KERN_VIRT_START ASM_CONST(0x8000000000000000)
+#else
+#define KERN_VIRT_START ASM_CONST(0xD000000000000000)
+#endif
+#define KERN_VIRT_SIZE	PGTABLE_RANGE
+
+/*
+ * The vmalloc space starts at the beginning of that region, and
+ * occupies half of it on hash CPUs and a quarter of it on Book3E
  */
-#define VMALLOC_START ASM_CONST(0xD000000000000000)
-#define VMALLOC_SIZE  (PGTABLE_RANGE >> 1)
-#define VMALLOC_END   (VMALLOC_START + VMALLOC_SIZE)
+#define VMALLOC_START	KERN_VIRT_START
+#ifdef CONFIG_PPC_BOOK3E
+#define VMALLOC_SIZE	(KERN_VIRT_SIZE >> 2)
+#else
+#define VMALLOC_SIZE	(KERN_VIRT_SIZE >> 1)
+#endif
+#define VMALLOC_END	(VMALLOC_START + VMALLOC_SIZE)
 
 /*
- * Define the address ranges for MMIO and IO space :
+ * The second half of the kernel virtual space is used for IO mappings,
+ * it's itself carved into the PIO region (ISA and PHB IO space) and
+ * the ioremap space
  *
- *  ISA_IO_BASE = VMALLOC_END, 64K reserved area
+ *  ISA_IO_BASE = KERN_IO_START, 64K reserved area
  *  PHB_IO_BASE = ISA_IO_BASE + 64K to ISA_IO_BASE + 2G, PHB IO spaces
  * IOREMAP_BASE = ISA_IO_BASE + 2G to VMALLOC_START + PGTABLE_RANGE
  */
+#define KERN_IO_START	(KERN_VIRT_START + (KERN_VIRT_SIZE >> 1))
 #define FULL_IO_SIZE	0x80000000ul
-#define  ISA_IO_BASE	(VMALLOC_END)
-#define  ISA_IO_END	(VMALLOC_END + 0x10000ul)
+#define  ISA_IO_BASE	(KERN_IO_START)
+#define  ISA_IO_END	(KERN_IO_START + 0x10000ul)
 #define  PHB_IO_BASE	(ISA_IO_END)
-#define  PHB_IO_END	(VMALLOC_END + FULL_IO_SIZE)
+#define  PHB_IO_END	(KERN_IO_START + FULL_IO_SIZE)
 #define IOREMAP_BASE	(PHB_IO_END)
-#define IOREMAP_END	(VMALLOC_START + PGTABLE_RANGE)
+#define IOREMAP_END	(KERN_VIRT_START + KERN_VIRT_SIZE)
+
 
 /*
  * Region IDs
@@ -72,19 +87,28 @@
 #define USER_REGION_ID		(0UL)
 
 /*
- * Defines the address of the vmemap area, in its own region
+ * Defines the address of the vmemap area, in its own region on
+ * hash table CPUs and after the vmalloc space on Book3E
  */
+#ifdef CONFIG_PPC_BOOK3E
+#define VMEMMAP_BASE		VMALLOC_END
+#define VMEMMAP_END		KERN_IO_START
+#else
 #define VMEMMAP_BASE		(VMEMMAP_REGION_ID << REGION_SHIFT)
+#endif
 #define vmemmap			((struct page *)VMEMMAP_BASE)
 
 
 /*
  * Include the PTE bits definitions
  */
+#ifdef CONFIG_PPC_BOOK3S
 #include <asm/pte-hash64.h>
+#else
+#include <asm/pte-book3e.h>
+#endif
 #include <asm/pte-common.h>
 
-
 #ifdef CONFIG_PPC_MM_SLICES
 #define HAVE_ARCH_UNMAPPED_AREA
 #define HAVE_ARCH_UNMAPPED_AREA_TOPDOWN
@@ -92,6 +116,9 @@
 
 #ifndef __ASSEMBLY__
 
+#include <linux/stddef.h>
+#include <asm/tlbflush.h>
+
 /*
  * This is the default implementation of various PTE accessors, it's
  * used in all cases except Book3S with 64K pages where we have a
Index: linux-work/arch/powerpc/include/asm/pte-common.h
===================================================================
--- linux-work.orig/arch/powerpc/include/asm/pte-common.h	2009-07-22 16:37:24.000000000 +1000
+++ linux-work/arch/powerpc/include/asm/pte-common.h	2009-07-22 16:40:20.000000000 +1000
@@ -34,6 +34,9 @@
 #ifndef _PAGE_4K_PFN
 #define _PAGE_4K_PFN		0
 #endif
+#ifndef _PAGE_SAO
+#define _PAGE_SAO	0
+#endif
 #ifndef _PAGE_PSIZE
 #define _PAGE_PSIZE		0
 #endif
Index: linux-work/arch/powerpc/include/asm/mmu-book3e.h
===================================================================
--- linux-work.orig/arch/powerpc/include/asm/mmu-book3e.h	2009-07-22 16:37:24.000000000 +1000
+++ linux-work/arch/powerpc/include/asm/mmu-book3e.h	2009-07-22 16:40:20.000000000 +1000
@@ -161,6 +161,29 @@
 #define TLBILX_T_CLASS2			6
 #define TLBILX_T_CLASS3			7
 
+/*
+ * The kernel use the constants below to index in the page sizes array.
+ * The use of fixed constants for this purpose is better for performances
+ * of the low level hash refill handlers.
+ *
+ * A non supported page size has a "shift" field set to 0
+ *
+ * Any new page size being implemented can get a new entry in here. Whether
+ * the kernel will use it or not is a different matter though. The actual page
+ * size used by hugetlbfs is not defined here and may be made variable
+ *
+ * Note: This is identical to the definition in mmu-hash64.h and possibly
+ * to be moved to a common file. This is also only used on 64-bit for now
+ */
+
+#define MMU_PAGE_4K		0
+#define MMU_PAGE_64K		1
+#define MMU_PAGE_1M		2
+#define MMU_PAGE_16M		3
+#define MMU_PAGE_256M		4
+#define MMU_PAGE_1G		5
+#define MMU_PAGE_COUNT		6
+
 #ifndef __ASSEMBLY__
 
 extern unsigned int tlbcam_index;
@@ -170,6 +193,33 @@ typedef struct {
 	unsigned int	active;
 	unsigned long	vdso_base;
 } mm_context_t;
+
+/* Page size definitions, common between 32 and 64-bit
+ *
+ *    shift : is the "PAGE_SHIFT" value for that page size
+ *    penc  : is the pte encoding mask
+ *
+ */
+struct mmu_psize_def
+{
+	unsigned int	shift;	/* number of bits */
+	unsigned int	enc;	/* PTE encoding */
+};
+extern struct mmu_psize_def mmu_psize_defs[MMU_PAGE_COUNT];
+
+/* The page sizes use the same names as 64-bit hash but are
+ * constants
+ */
+#if defined(CONFIG_PPC_4K_PAGES)
+#define mmu_virtual_psize	MMU_PAGE_4K
+#elif defined(CONFIG_PPC_64K_PAGES)
+#define mmu_virtual_psize	MMU_PAGE_64K
+#else
+#error Unsupported page size
+#endif
+
+extern int mmu_linear_psize;
+
 #endif /* !__ASSEMBLY__ */
 
 #endif /* _ASM_POWERPC_MMU_BOOK3E_H_ */
Index: linux-work/arch/powerpc/include/asm/page_64.h
===================================================================
--- linux-work.orig/arch/powerpc/include/asm/page_64.h	2009-07-22 16:37:24.000000000 +1000
+++ linux-work/arch/powerpc/include/asm/page_64.h	2009-07-22 16:40:20.000000000 +1000
@@ -135,12 +135,22 @@ extern void slice_set_range_psize(struct
 #endif /* __ASSEMBLY__ */
 #else
 #define slice_init()
+#ifdef CONFIG_PPC_STD_MMU_64
 #define get_slice_psize(mm, addr)	((mm)->context.user_psize)
 #define slice_set_user_psize(mm, psize)		\
 do {						\
 	(mm)->context.user_psize = (psize);	\
 	(mm)->context.sllp = SLB_VSID_USER | mmu_psize_defs[(psize)].sllp; \
 } while (0)
+#else /* CONFIG_PPC_STD_MMU_64 */
+#ifdef CONFIG_PPC_64K_PAGES
+#define get_slice_psize(mm, addr)	MMU_PAGE_64K
+#else /* CONFIG_PPC_64K_PAGES */
+#define get_slice_psize(mm, addr)	MMU_PAGE_4K
+#endif /* !CONFIG_PPC_64K_PAGES */
+#define slice_set_user_psize(mm, psize)	do { BUG(); } while(0)
+#endif /* !CONFIG_PPC_STD_MMU_64 */
+
 #define slice_set_range_psize(mm, start, len, psize)	\
 	slice_set_user_psize((mm), (psize))
 #define slice_mm_new_context(mm)	1
Index: linux-work/arch/powerpc/include/asm/mmu.h
===================================================================
--- linux-work.orig/arch/powerpc/include/asm/mmu.h	2009-07-22 16:46:59.000000000 +1000
+++ linux-work/arch/powerpc/include/asm/mmu.h	2009-07-22 16:56:22.000000000 +1000
@@ -17,6 +17,7 @@
 #define MMU_FTR_TYPE_40x		ASM_CONST(0x00000004)
 #define MMU_FTR_TYPE_44x		ASM_CONST(0x00000008)
 #define MMU_FTR_TYPE_FSL_E		ASM_CONST(0x00000010)
+#define MMU_FTR_TYPE_3E			ASM_CONST(0x00000020)
 
 /*
  * This is individual features
Index: linux-work/arch/powerpc/kernel/asm-offsets.c
===================================================================
--- linux-work.orig/arch/powerpc/kernel/asm-offsets.c	2009-07-22 16:57:40.000000000 +1000
+++ linux-work/arch/powerpc/kernel/asm-offsets.c	2009-07-22 16:59:51.000000000 +1000
@@ -52,9 +52,11 @@
 #include <linux/kvm_host.h>
 #endif
 
+#ifdef CONFIG_PPC32
 #if defined(CONFIG_BOOKE) || defined(CONFIG_40x)
 #include "head_booke.h"
 #endif
+#endif
 
 #if defined(CONFIG_FSL_BOOKE)
 #include "../mm/mmu_decl.h"
@@ -260,6 +276,7 @@ int main(void)
 	DEFINE(_SRR1, STACK_FRAME_OVERHEAD+sizeof(struct pt_regs)+8);
 #endif /* CONFIG_PPC64 */
 
+#if defined(CONFIG_PPC32)
 #if defined(CONFIG_BOOKE) || defined(CONFIG_40x)
 	DEFINE(EXC_LVL_SIZE, STACK_EXC_LVL_FRAME_SIZE);
 	DEFINE(MAS0, STACK_INT_FRAME_SIZE+offsetof(struct exception_regs, mas0));
@@ -278,7 +295,7 @@ int main(void)
 	DEFINE(_DSRR1, STACK_INT_FRAME_SIZE+offsetof(struct exception_regs, dsrr1));
 	DEFINE(SAVED_KSP_LIMIT, STACK_INT_FRAME_SIZE+offsetof(struct exception_regs, saved_ksp_limit));
 #endif
-
+#endif
 	DEFINE(CLONE_VM, CLONE_VM);
 	DEFINE(CLONE_UNTRACED, CLONE_UNTRACED);
 
Index: linux-work/arch/powerpc/include/asm/page.h
===================================================================
--- linux-work.orig/arch/powerpc/include/asm/page.h	2009-07-22 16:57:39.000000000 +1000
+++ linux-work/arch/powerpc/include/asm/page.h	2009-07-22 16:59:51.000000000 +1000
@@ -139,7 +139,11 @@ extern phys_addr_t kernstart_addr;
  * Don't compare things with KERNELBASE or PAGE_OFFSET to test for
  * "kernelness", use is_kernel_addr() - it should do what you want.
  */
+#ifdef CONFIG_PPC_BOOK3E_64
+#define is_kernel_addr(x)	((x) >= 0x8000000000000000ul)
+#else
 #define is_kernel_addr(x)	((x) >= PAGE_OFFSET)
+#endif
 
 #ifndef __ASSEMBLY__
 

^ permalink raw reply

* [PATCH 13/20] powerpc: Add SPR definitions for new 64-bit BookE
From: Benjamin Herrenschmidt @ 2009-07-23  5:59 UTC (permalink / raw)
  To: linuxppc-dev

This adds various SPRs defined on 64-bit BookE, along with changes
to the definition of the base MSR values to add the values needed
for 64-bit Book3E.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---

 arch/powerpc/include/asm/reg.h       |   10 +++-------
 arch/powerpc/include/asm/reg_booke.h |   35 ++++++++++++++++++++++++++++++++---
 2 files changed, 35 insertions(+), 10 deletions(-)

--- linux-work.orig/arch/powerpc/include/asm/reg.h	2009-07-22 16:37:24.000000000 +1000
+++ linux-work/arch/powerpc/include/asm/reg.h	2009-07-22 16:40:20.000000000 +1000
@@ -98,19 +98,15 @@
 #define MSR_RI		__MASK(MSR_RI_LG)	/* Recoverable Exception */
 #define MSR_LE		__MASK(MSR_LE_LG)	/* Little Endian */
 
-#ifdef CONFIG_PPC64
+#if defined(CONFIG_PPC_BOOK3S_64)
+/* Server variant */
 #define MSR_		MSR_ME | MSR_RI | MSR_IR | MSR_DR | MSR_ISF |MSR_HV
 #define MSR_KERNEL      MSR_ | MSR_SF
-
 #define MSR_USER32	MSR_ | MSR_PR | MSR_EE
 #define MSR_USER64	MSR_USER32 | MSR_SF
-
-#else /* 32-bit */
+#elif defined(CONFIG_PPC_BOOK3S_32)
 /* Default MSR for kernel mode. */
-#ifndef MSR_KERNEL	/* reg_booke.h also defines this */
 #define MSR_KERNEL	(MSR_ME|MSR_RI|MSR_IR|MSR_DR)
-#endif
-
 #define MSR_USER	(MSR_KERNEL|MSR_PR|MSR_EE)
 #endif
 
Index: linux-work/arch/powerpc/include/asm/reg_booke.h
===================================================================
--- linux-work.orig/arch/powerpc/include/asm/reg_booke.h	2009-07-22 16:37:24.000000000 +1000
+++ linux-work/arch/powerpc/include/asm/reg_booke.h	2009-07-22 16:40:20.000000000 +1000
@@ -18,18 +18,26 @@
 #define MSR_IS		MSR_IR	/* Instruction Space */
 #define MSR_DS		MSR_DR	/* Data Space */
 #define MSR_PMM		(1<<2)	/* Performance monitor mark bit */
+#define MSR_CM		(1<<31) /* Computation Mode (0=32-bit, 1=64-bit) */
 
-/* Default MSR for kernel mode. */
-#if defined (CONFIG_40x)
+#if defined(CONFIG_PPC_BOOK3E_64)
+#define MSR_		MSR_ME | MSR_CE
+#define MSR_KERNEL      MSR_ | MSR_CM
+#define MSR_USER32	MSR_ | MSR_PR | MSR_EE
+#define MSR_USER64	MSR_USER32 | MSR_CM
+#elif defined (CONFIG_40x)
 #define MSR_KERNEL	(MSR_ME|MSR_RI|MSR_IR|MSR_DR|MSR_CE)
-#elif defined(CONFIG_BOOKE)
+#define MSR_USER	(MSR_KERNEL|MSR_PR|MSR_EE)
+#else
 #define MSR_KERNEL	(MSR_ME|MSR_RI|MSR_CE)
+#define MSR_USER	(MSR_KERNEL|MSR_PR|MSR_EE)
 #endif
 
 /* Special Purpose Registers (SPRNs)*/
 #define SPRN_DECAR	0x036	/* Decrementer Auto Reload Register */
 #define SPRN_IVPR	0x03F	/* Interrupt Vector Prefix Register */
 #define SPRN_USPRG0	0x100	/* User Special Purpose Register General 0 */
+#define SPRN_SPRG3R	0x103	/* Special Purpose Register General 3 Read */
 #define SPRN_SPRG4R	0x104	/* Special Purpose Register General 4 Read */
 #define SPRN_SPRG5R	0x105	/* Special Purpose Register General 5 Read */
 #define SPRN_SPRG6R	0x106	/* Special Purpose Register General 6 Read */
@@ -38,11 +46,18 @@
 #define SPRN_SPRG5W	0x115	/* Special Purpose Register General 5 Write */
 #define SPRN_SPRG6W	0x116	/* Special Purpose Register General 6 Write */
 #define SPRN_SPRG7W	0x117	/* Special Purpose Register General 7 Write */
+#define SPRN_EPCR	0x133	/* Embedded Processor Control Register */
 #define SPRN_DBCR2	0x136	/* Debug Control Register 2 */
 #define SPRN_IAC3	0x13A	/* Instruction Address Compare 3 */
 #define SPRN_IAC4	0x13B	/* Instruction Address Compare 4 */
 #define SPRN_DVC1	0x13E	/* Data Value Compare Register 1 */
 #define SPRN_DVC2	0x13F	/* Data Value Compare Register 2 */
+#define SPRN_MAS8	0x155	/* MMU Assist Register 8 */
+#define SPRN_TLB0PS	0x158	/* TLB 0 Page Size Register */
+#define SPRN_MAS5_MAS6	0x15c	/* MMU Assist Register 5 || 6 */
+#define SPRN_MAS8_MAS1	0x15d	/* MMU Assist Register 8 || 1 */
+#define SPRN_MAS7_MAS3	0x174	/* MMU Assist Register 7 || 3 */
+#define SPRN_MAS0_MAS1	0x175	/* MMU Assist Register 0 || 1 */
 #define SPRN_IVOR0	0x190	/* Interrupt Vector Offset Register 0 */
 #define SPRN_IVOR1	0x191	/* Interrupt Vector Offset Register 1 */
 #define SPRN_IVOR2	0x192	/* Interrupt Vector Offset Register 2 */
@@ -425,6 +440,19 @@
 #define SGR_NORMAL	0		/* Speculative fetching allowed. */
 #define SGR_GUARDED	1		/* Speculative fetching disallowed. */
 
+/* Bit definitions for EPCR */
+#define SPRN_EPCR_EXTGS		0x80000000
+#define SPRN_EPCR_DTLBGS	0x40000000
+#define SPRN_EPCR_ITLBGS	0x20000000
+#define SPRN_EPCR_DSIGS		0x10000000
+#define SPRN_EPCR_ISIGS		0x08000000
+#define SPRN_EPCR_DUVD		0x04000000
+#define SPRN_EPCR_ICM		0x02000000
+#define SPRN_EPCR_GICM		0x01000000
+#define SPRN_EPCR_DGTMI		0x00800000
+#define SPRN_EPCR_DMIUH		0x00400000
+
+
 /*
  * The IBM-403 is an even more odd special case, as it is much
  * older than the IBM-405 series.  We put these down here incase someone
@@ -480,3 +508,4 @@
 #endif /* 403GCX */
 #endif /* __ASM_POWERPC_REG_BOOKE_H__ */
 #endif /* __KERNEL__ */
+

^ permalink raw reply

* [PATCH 12/20] powerpc/mm: Add hook to flush pgtables when a PTE page is freed
From: Benjamin Herrenschmidt @ 2009-07-23  5:59 UTC (permalink / raw)
  To: linuxppc-dev

64-bit BookE will use that hook to maintain the virtually linear
page tables or the indirect entries in the TLB when using the
HW loader.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---

 arch/powerpc/include/asm/pgalloc.h |   13 ++++++++++++-
 1 file changed, 12 insertions(+), 1 deletion(-)

--- linux-work.orig/arch/powerpc/include/asm/pgalloc.h	2009-07-22 16:47:46.000000000 +1000
+++ linux-work/arch/powerpc/include/asm/pgalloc.h	2009-07-22 16:47:49.000000000 +1000
@@ -4,6 +4,12 @@
 
 #include <linux/mm.h>
 
+#ifdef CONFIG_PPC_BOOK3E
+extern void tlb_flush_pgtable(struct mmu_gather *tlb, unsigned long address);
+#else
+#define tlb_flush_pgtable(tlb, address)	do { } while(0)
+#endif
+
 static inline void pte_free_kernel(struct mm_struct *mm, pte_t *pte)
 {
 	free_page((unsigned long)pte);
@@ -41,11 +47,16 @@ extern void pgtable_free_tlb(struct mmu_
 #define __pte_free_tlb(tlb,ptepage,address)		\
 do { \
 	pgtable_page_dtor(ptepage); \
+ 	tlb_flush_pgtable(tlb, address);				\
 	pgtable_free_tlb(tlb, pgtable_free_cache(page_address(ptepage), \
 					PTE_NONCACHE_NUM, PTE_TABLE_SIZE-1)); \
 } while (0)
 #else
-#define __pte_free_tlb(tlb, pte, address)	pte_free((tlb)->mm, (pte))
+#define __pte_free_tlb(tlb,pte,address)			\
+do { 							\
+	tlb_flush_pgtable(tlb, address);		\
+	pte_free((tlb)->mm, (pte));  			\
+} while (0)
 #endif
 
 

^ permalink raw reply

* [PATCH 11/20] powerpc: Move definitions of secondary CPU spinloop to header file
From: Benjamin Herrenschmidt @ 2009-07-23  5:59 UTC (permalink / raw)
  To: linuxppc-dev

Those definitions are currently declared extern in the .c file where
they are used, move them to a header file instead.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
--

 arch/powerpc/include/asm/smp.h |    5 +++++
 arch/powerpc/kernel/setup_64.c |    3 ---
 2 files changed, 5 insertions(+), 3 deletions(-)

--- linux-work.orig/arch/powerpc/include/asm/smp.h	2009-07-22 16:38:32.000000000 +1000
+++ linux-work/arch/powerpc/include/asm/smp.h	2009-07-22 16:46:59.000000000 +1000
@@ -148,6 +148,11 @@ extern struct smp_ops_t *smp_ops;
 extern void arch_send_call_function_single_ipi(int cpu);
 extern void arch_send_call_function_ipi(cpumask_t mask);
 
+#ifdef CONFIG_PPC64
+extern void generic_secondary_smp_init(void);
+extern unsigned long __secondary_hold_spinloop;
+#endif /* CONFIG_PPC64 */
+
 #endif /* __ASSEMBLY__ */
 
 #endif /* __KERNEL__ */
Index: linux-work/arch/powerpc/kernel/setup_64.c
===================================================================
--- linux-work.orig/arch/powerpc/kernel/setup_64.c	2009-07-22 16:38:12.000000000 +1000
+++ linux-work/arch/powerpc/kernel/setup_64.c	2009-07-22 16:46:59.000000000 +1000
@@ -230,9 +230,6 @@ void early_setup_secondary(void)
 #endif /* CONFIG_SMP */
 
 #if defined(CONFIG_SMP) || defined(CONFIG_KEXEC)
-extern unsigned long __secondary_hold_spinloop;
-extern void generic_secondary_smp_init(void);
-
 void smp_release_cpus(void)
 {
 	unsigned long *ptr;

^ permalink raw reply

* [PATCH 10/20] powerpc: Clean ifdef usage in copy_thread()
From: Benjamin Herrenschmidt @ 2009-07-23  5:59 UTC (permalink / raw)
  To: linuxppc-dev

Currently, a single ifdef covers SLB related bits and more generic ppc64
related bits, split this in two separate ifdef's since 64-bit BookE will
need one but not the other.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---

 arch/powerpc/kernel/process.c |    2 ++
 1 file changed, 2 insertions(+)

--- linux-work.orig/arch/powerpc/kernel/process.c	2009-07-22 16:30:49.000000000 +1000
+++ linux-work/arch/powerpc/kernel/process.c	2009-07-22 16:31:02.000000000 +1000
@@ -664,6 +664,7 @@ int copy_thread(unsigned long clone_flag
 		sp_vsid |= SLB_VSID_KERNEL | llp;
 		p->thread.ksp_vsid = sp_vsid;
 	}
+#endif /* CONFIG_PPC_STD_MMU_64 */
 
 	/*
 	 * The PPC64 ABI makes use of a TOC to contain function 
@@ -671,6 +672,7 @@ int copy_thread(unsigned long clone_flag
 	 * to the TOC entry.  The first entry is a pointer to the actual
 	 * function.
  	 */
+#ifdef CONFIG_PPC64
 	kregs->nip = *((unsigned long *)ret_from_fork);
 #else
 	kregs->nip = (unsigned long)ret_from_fork;

^ permalink raw reply

* [PATCH 9/20] powerpc/mm: Call mmu_context_init() from ppc64
From: Benjamin Herrenschmidt @ 2009-07-23  5:59 UTC (permalink / raw)
  To: linuxppc-dev

Our 64-bit hash context handling has no init function, but 64-bit Book3E
will use the common mmu_context_nohash.c code which does, so define an
empty inline mmu_context_init() for 64-bit server and call it from
our 64-bit setup_arch()

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---

 arch/powerpc/include/asm/mmu_context.h |    7 ++++++-
 arch/powerpc/kernel/setup_64.c         |    4 ++++
 arch/powerpc/mm/mmu_context_hash64.c   |    1 +
 3 files changed, 11 insertions(+), 1 deletion(-)

--- linux-work.orig/arch/powerpc/include/asm/mmu_context.h	2009-07-22 16:25:25.000000000 +1000
+++ linux-work/arch/powerpc/include/asm/mmu_context.h	2009-07-22 16:25:50.000000000 +1000
@@ -14,7 +14,6 @@
 /*
  * Most if the context management is out of line
  */
-extern void mmu_context_init(void);
 extern int init_new_context(struct task_struct *tsk, struct mm_struct *mm);
 extern void destroy_context(struct mm_struct *mm);
 
@@ -23,6 +22,12 @@ extern void switch_stab(struct task_stru
 extern void switch_slb(struct task_struct *tsk, struct mm_struct *mm);
 extern void set_context(unsigned long id, pgd_t *pgd);
 
+#ifdef CONFIG_PPC_BOOK3S_64
+static inline void mmu_context_init(void) { }
+#else
+extern void mmu_context_init(void);
+#endif
+
 /*
  * switch_mm is the entry point called from the architecture independent
  * code in kernel/sched.c
Index: linux-work/arch/powerpc/kernel/setup_64.c
===================================================================
--- linux-work.orig/arch/powerpc/kernel/setup_64.c	2009-07-22 16:26:23.000000000 +1000
+++ linux-work/arch/powerpc/kernel/setup_64.c	2009-07-22 16:26:31.000000000 +1000
@@ -534,6 +534,10 @@ void __init setup_arch(char **cmdline_p)
 #endif
 
 	paging_init();
+
+	/* Initialize the MMU context management stuff */
+	mmu_context_init();
+
 	ppc64_boot_msg(0x15, "Setup Done");
 }
 
Index: linux-work/arch/powerpc/mm/mmu_context_hash64.c
===================================================================
--- linux-work.orig/arch/powerpc/mm/mmu_context_hash64.c	2009-07-22 16:24:43.000000000 +1000
+++ linux-work/arch/powerpc/mm/mmu_context_hash64.c	2009-07-22 16:24:56.000000000 +1000
@@ -76,3 +76,4 @@ void destroy_context(struct mm_struct *m
 
 	mm->context.id = NO_CONTEXT;
 }
+

^ permalink raw reply

* [PATCH 8/20] powerpc/mm: Make low level TLB flush ops on BookE take additional arguments
From: Benjamin Herrenschmidt @ 2009-07-23  5:59 UTC (permalink / raw)
  To: linuxppc-dev

We need to pass down whether the page is direct or indirect and we'll
need to pass the page size to _tlbil_va and _tlbivax_bcast

We also add a new low level _tlbil_pid_noind() which does a TLB flush
by PID but avoids flushing indirect entries if possible

This implements those new prototypes but defines them with inlines
or macros so that no additional arguments are actually passed on current
processors.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/include/asm/tlbflush.h |   11 +++++++--
 arch/powerpc/mm/mmu_decl.h          |   17 ++++++++++++--
 arch/powerpc/mm/tlb_nohash.c        |   41 +++++++++++++++++++++++++++---------
 arch/powerpc/mm/tlb_nohash_low.S    |    6 ++---
 4 files changed, 57 insertions(+), 18 deletions(-)

--- linux-work.orig/arch/powerpc/include/asm/tlbflush.h	2009-07-22 16:09:19.000000000 +1000
+++ linux-work/arch/powerpc/include/asm/tlbflush.h	2009-07-22 16:11:32.000000000 +1000
@@ -6,7 +6,7 @@
  *
  *  - flush_tlb_mm(mm) flushes the specified mm context TLB's
  *  - flush_tlb_page(vma, vmaddr) flushes one page
- *  - local_flush_tlb_mm(mm) flushes the specified mm context on
+ *  - local_flush_tlb_mm(mm, full) flushes the specified mm context on
  *                           the local processor
  *  - local_flush_tlb_page(vma, vmaddr) flushes one page on the local processor
  *  - flush_tlb_page_nohash(vma, vmaddr) flushes one page if SW loaded TLB
@@ -29,7 +29,8 @@
  * specific tlbie's
  */
 
-#include <linux/mm.h>
+struct vm_area_struct;
+struct mm_struct;
 
 #define MMU_NO_CONTEXT      	((unsigned int)-1)
 
@@ -40,12 +41,18 @@ extern void flush_tlb_kernel_range(unsig
 extern void local_flush_tlb_mm(struct mm_struct *mm);
 extern void local_flush_tlb_page(struct vm_area_struct *vma, unsigned long vmaddr);
 
+extern void __local_flush_tlb_page(struct mm_struct *mm, unsigned long vmaddr,
+				   int tsize, int ind);
+
 #ifdef CONFIG_SMP
 extern void flush_tlb_mm(struct mm_struct *mm);
 extern void flush_tlb_page(struct vm_area_struct *vma, unsigned long vmaddr);
+extern void __flush_tlb_page(struct mm_struct *mm, unsigned long vmaddr,
+			     int tsize, int ind);
 #else
 #define flush_tlb_mm(mm)		local_flush_tlb_mm(mm)
 #define flush_tlb_page(vma,addr)	local_flush_tlb_page(vma,addr)
+#define __flush_tlb_page(mm,addr,p,i)	__local_flush_tlb_page(mm,addr,p,i)
 #endif
 #define flush_tlb_page_nohash(vma,addr)	flush_tlb_page(vma,addr)
 
Index: linux-work/arch/powerpc/mm/mmu_decl.h
===================================================================
--- linux-work.orig/arch/powerpc/mm/mmu_decl.h	2009-07-22 16:09:19.000000000 +1000
+++ linux-work/arch/powerpc/mm/mmu_decl.h	2009-07-22 16:16:56.000000000 +1000
@@ -36,21 +36,31 @@ static inline void _tlbil_pid(unsigned i
 {
 	asm volatile ("sync; tlbia; isync" : : : "memory");
 }
+#define _tlbil_pid_noind(pid)	_tlbil_pid(pid)
+
 #else /* CONFIG_40x || CONFIG_8xx */
 extern void _tlbil_all(void);
 extern void _tlbil_pid(unsigned int pid);
+#define _tlbil_pid_noind(pid)	_tlbil_pid(pid)
 #endif /* !(CONFIG_40x || CONFIG_8xx) */
 
 /*
  * On 8xx, we directly inline tlbie, on others, it's extern
  */
 #ifdef CONFIG_8xx
-static inline void _tlbil_va(unsigned long address, unsigned int pid)
+static inline void _tlbil_va(unsigned long address, unsigned int pid,
+			     unsigned int tsize, unsigned int ind)
 {
 	asm volatile ("tlbie %0; sync" : : "r" (address) : "memory");
 }
 #else /* CONFIG_8xx */
-extern void _tlbil_va(unsigned long address, unsigned int pid);
+extern void __tlbil_va(unsigned long address, unsigned int pid,
+		       unsigned int tsize, unsigned int ind);
+static inline void _tlbil_va(unsigned long address, unsigned int pid,
+			     unsigned int tsize, unsigned int ind)
+{
+	__tlbil_va(address, pid);
+}
 #endif /* CONIFG_8xx */
 
 /*
@@ -58,7 +68,8 @@ extern void _tlbil_va(unsigned long addr
  * implementation. When that becomes the case, this will be
  * an extern.
  */
-static inline void _tlbivax_bcast(unsigned long address, unsigned int pid)
+static inline void _tlbivax_bcast(unsigned long address, unsigned int pid,
+				   unsigned int tsize, unsigned int ind);
 {
 	BUG();
 }
Index: linux-work/arch/powerpc/mm/tlb_nohash.c
===================================================================
--- linux-work.orig/arch/powerpc/mm/tlb_nohash.c	2009-07-22 15:52:22.000000000 +1000
+++ linux-work/arch/powerpc/mm/tlb_nohash.c	2009-07-22 16:16:56.000000000 +1000
@@ -67,16 +67,23 @@ void local_flush_tlb_mm(struct mm_struct
 }
 EXPORT_SYMBOL(local_flush_tlb_mm);
 
-void local_flush_tlb_page(struct vm_area_struct *vma, unsigned long vmaddr)
+void __local_flush_tlb_page(struct mm_struct *mm, unsigned long vmaddr,
+			    int tsize, int ind)
 {
 	unsigned int pid;
 
 	preempt_disable();
-	pid = vma ? vma->vm_mm->context.id : 0;
+	pid = mm ? mm->context.id : 0;
 	if (pid != MMU_NO_CONTEXT)
-		_tlbil_va(vmaddr, pid);
+		_tlbil_va(vmaddr, pid, tsize, ind);
 	preempt_enable();
 }
+
+void local_flush_tlb_page(struct vm_area_struct *vma, unsigned long vmaddr)
+{
+	__local_flush_tlb_page(vma ? vma->vm_mm : NULL, vmaddr,
+			       mmu_psize_defs[mmu_virtual_psize].enc, 0);
+}
 EXPORT_SYMBOL(local_flush_tlb_page);
 
 
@@ -96,6 +103,8 @@ static int mm_is_core_local(struct mm_st
 struct tlb_flush_param {
 	unsigned long addr;
 	unsigned int pid;
+	unsigned int tsize;
+	unsigned int ind;
 };
 
 static void do_flush_tlb_mm_ipi(void *param)
@@ -109,7 +118,7 @@ static void do_flush_tlb_page_ipi(void *
 {
 	struct tlb_flush_param *p = param;
 
-	_tlbil_va(p->addr, p->pid);
+	_tlbil_va(p->addr, p->pid, p->tsize, p->ind);
 }
 
 
@@ -149,37 +158,49 @@ void flush_tlb_mm(struct mm_struct *mm)
 }
 EXPORT_SYMBOL(flush_tlb_mm);
 
-void flush_tlb_page(struct vm_area_struct *vma, unsigned long vmaddr)
+void __flush_tlb_page(struct mm_struct *mm, unsigned long vmaddr,
+		      int tsize, int ind)
 {
 	struct cpumask *cpu_mask;
 	unsigned int pid;
 
 	preempt_disable();
-	pid = vma ? vma->vm_mm->context.id : 0;
+	pid = mm ? mm->context.id : 0;
 	if (unlikely(pid == MMU_NO_CONTEXT))
 		goto bail;
-	cpu_mask = mm_cpumask(vma->vm_mm);
+	cpu_mask = mm_cpumask(mm);
 	if (!mm_is_core_local(mm)) {
 		/* If broadcast tlbivax is supported, use it */
 		if (mmu_has_feature(MMU_FTR_USE_TLBIVAX_BCAST)) {
 			int lock = mmu_has_feature(MMU_FTR_LOCK_BCAST_INVAL);
 			if (lock)
 				spin_lock(&tlbivax_lock);
-			_tlbivax_bcast(vmaddr, pid);
+			_tlbivax_bcast(vmaddr, pid, tsize, ind);
 			if (lock)
 				spin_unlock(&tlbivax_lock);
 			goto bail;
 		} else {
-			struct tlb_flush_param p = { .pid = pid, .addr = vmaddr };
+			struct tlb_flush_param p = {
+				.pid = pid,
+				.addr = vmaddr,
+				.tsize = tsize,
+				.ind = ind,
+			};
 			/* Ignores smp_processor_id() even if set in cpu_mask */
 			smp_call_function_many(cpu_mask,
 					       do_flush_tlb_page_ipi, &p, 1);
 		}
 	}
-	_tlbil_va(vmaddr, pid);
+	_tlbil_va(vmaddr, pid, tsize, ind);
  bail:
 	preempt_enable();
 }
+
+void flush_tlb_page(struct vm_area_struct *vma, unsigned long vmaddr)
+{
+	__flush_tlb_page(vma ? vma->vm_mm : NULL, vmaddr,
+			 mmu_psize_defs[mmu_virtual_psize].enc, 0);
+}
 EXPORT_SYMBOL(flush_tlb_page);
 
 #endif /* CONFIG_SMP */
Index: linux-work/arch/powerpc/mm/tlb_nohash_low.S
===================================================================
--- linux-work.orig/arch/powerpc/mm/tlb_nohash_low.S	2009-07-22 16:17:13.000000000 +1000
+++ linux-work/arch/powerpc/mm/tlb_nohash_low.S	2009-07-22 16:17:21.000000000 +1000
@@ -39,7 +39,7 @@
 /*
  * 40x implementation needs only tlbil_va
  */
-_GLOBAL(_tlbil_va)
+_GLOBAL(__tlbil_va)
 	/* We run the search with interrupts disabled because we have to change
 	 * the PID and I don't want to preempt when that happens.
 	 */
@@ -71,7 +71,7 @@ _GLOBAL(_tlbil_va)
  * 440 implementation uses tlbsx/we for tlbil_va and a full sweep
  * of the TLB for everything else.
  */
-_GLOBAL(_tlbil_va)
+_GLOBAL(__tlbil_va)
 	mfspr	r5,SPRN_MMUCR
 	rlwimi	r5,r4,0,24,31			/* Set TID */
 
@@ -170,7 +170,7 @@ ALT_MMU_FTR_SECTION_END_IFSET(MMU_FTR_US
  * Flush MMU TLB for a particular address, but only on the local processor
  * (no broadcast)
  */
-_GLOBAL(_tlbil_va)
+_GLOBAL(__tlbil_va)
 	mfmsr	r10
 	wrteei	0
 	slwi	r4,r4,16

^ permalink raw reply

* [PATCH 7/20] powerpc: Modify some ppc_asm.h macros to accomodate 64-bits Book3E
From: Benjamin Herrenschmidt @ 2009-07-23  5:59 UTC (permalink / raw)
  To: linuxppc-dev

The way I intend to use tophys/tovirt on 64-bit BookE is different
from the "trick" that we currently play for 32-bit BookE so change
the condition of definition of these macros to make it so.

Also, make sure we only use rfid and mtmsrd instead of rfi and mtmsr
for 64-bit server processors, not all 64-bit processors.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---

 arch/powerpc/include/asm/ppc_asm.h |   14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)

--- linux-work.orig/arch/powerpc/include/asm/ppc_asm.h	2009-07-22 15:53:06.000000000 +1000
+++ linux-work/arch/powerpc/include/asm/ppc_asm.h	2009-07-22 15:58:24.000000000 +1000
@@ -375,8 +375,15 @@ END_FTR_SECTION_IFCLR(CPU_FTR_601)
 #define PPC440EP_ERR42
 #endif
 
-
-#if defined(CONFIG_BOOKE)
+/*
+ * toreal/fromreal/tophys/tovirt macros. 32-bit BookE makes them
+ * keep the address intact to be compatible with code shared with
+ * 32-bit classic.
+ *
+ * On the other hand, I find it useful to have them behave as expected
+ * by their name (ie always do the addition) on 64-bit BookE
+ */
+#if defined(CONFIG_BOOKE) && !defined(CONFIG_PPC64)
 #define toreal(rd)
 #define fromreal(rd)
 
@@ -426,10 +433,9 @@ END_FTR_SECTION_IFCLR(CPU_FTR_601)
 	.previous
 #endif
 
-#ifdef CONFIG_PPC64
+#ifdef CONFIG_PPC_BOOK3S_64
 #define RFI		rfid
 #define MTMSRD(r)	mtmsrd	r
-
 #else
 #define FIX_SRR1(ra, rb)
 #ifndef CONFIG_40x

^ permalink raw reply

* [PATCH 6/20] powerpc/mm: Add support for early ioremap on non-hash 64-bit processors
From: Benjamin Herrenschmidt @ 2009-07-23  5:59 UTC (permalink / raw)
  To: linuxppc-dev

This adds some code to do early ioremap's using page tables instead of
bolting entries in the hash table. This will be used by the upcoming
64-bits BookE port.

The patch also changes the test for early vs. late ioremap to use
slab_is_available() instead of our old hackish mem_init_done.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---

 arch/powerpc/mm/pgtable_64.c |   59 +++++++++++++++++++++++++++++++++++++++----
 1 file changed, 54 insertions(+), 5 deletions(-)

--- linux-work.orig/arch/powerpc/mm/pgtable_64.c	2009-07-22 15:49:09.000000000 +1000
+++ linux-work/arch/powerpc/mm/pgtable_64.c	2009-07-23 14:56:39.000000000 +1000
@@ -33,6 +33,8 @@
 #include <linux/stddef.h>
 #include <linux/vmalloc.h>
 #include <linux/init.h>
+#include <linux/bootmem.h>
+#include <linux/lmb.h>
 
 #include <asm/pgalloc.h>
 #include <asm/page.h>
@@ -55,19 +57,36 @@
 
 unsigned long ioremap_bot = IOREMAP_BASE;
 
+
+#ifdef CONFIG_PPC_MMU_NOHASH
+static void *early_alloc_pgtable(unsigned long size)
+{
+	void *pt;
+
+	if (init_bootmem_done)
+		pt = __alloc_bootmem(size, size, __pa(MAX_DMA_ADDRESS));
+	else
+		pt = __va(lmb_alloc_base(size, size,
+					 __pa(MAX_DMA_ADDRESS)));
+	memset(pt, 0, size);
+
+	return pt;
+}
+#endif /* CONFIG_PPC_MMU_NOHASH */
+
 /*
- * map_io_page currently only called by __ioremap
- * map_io_page adds an entry to the ioremap page table
+ * map_kernel_page currently only called by __ioremap
+ * map_kernel_page adds an entry to the ioremap page table
  * and adds an entry to the HPT, possibly bolting it
  */
-static int map_io_page(unsigned long ea, unsigned long pa, int flags)
+static int map_kernel_page(unsigned long ea, unsigned long pa, int flags)
 {
 	pgd_t *pgdp;
 	pud_t *pudp;
 	pmd_t *pmdp;
 	pte_t *ptep;
 
-	if (mem_init_done) {
+	if (slab_is_available()) {
 		pgdp = pgd_offset_k(ea);
 		pudp = pud_alloc(&init_mm, pgdp, ea);
 		if (!pudp)
@@ -81,6 +100,35 @@ static int map_io_page(unsigned long ea,
 		set_pte_at(&init_mm, ea, ptep, pfn_pte(pa >> PAGE_SHIFT,
 							  __pgprot(flags)));
 	} else {
+#ifdef CONFIG_PPC_MMU_NOHASH
+		/* Warning ! This will blow up if bootmem is not initialized
+		 * which our ppc64 code is keen to do that, we'll need to
+		 * fix it and/or be more careful
+		 */
+		pgdp = pgd_offset_k(ea);
+#ifdef PUD_TABLE_SIZE
+		if (pgd_none(*pgdp)) {
+			pudp = early_alloc_pgtable(PUD_TABLE_SIZE);
+			BUG_ON(pudp == NULL);
+			pgd_populate(&init_mm, pgdp, pudp);
+		}
+#endif /* PUD_TABLE_SIZE */
+		pudp = pud_offset(pgdp, ea);
+		if (pud_none(*pudp)) {
+			pmdp = early_alloc_pgtable(PMD_TABLE_SIZE);
+			BUG_ON(pmdp == NULL);
+			pud_populate(&init_mm, pudp, pmdp);
+		}
+		pmdp = pmd_offset(pudp, ea);
+		if (!pmd_present(*pmdp)) {
+			ptep = early_alloc_pgtable(PAGE_SIZE);
+			BUG_ON(ptep == NULL);
+			pmd_populate_kernel(&init_mm, pmdp, ptep);
+		}
+		ptep = pte_offset_kernel(pmdp, ea);
+		set_pte_at(&init_mm, ea, ptep, pfn_pte(pa >> PAGE_SHIFT,
+							  __pgprot(flags)));
+#else /* CONFIG_PPC_MMU_NOHASH */
 		/*
 		 * If the mm subsystem is not fully up, we cannot create a
 		 * linux page table entry for this mapping.  Simply bolt an
@@ -93,6 +141,7 @@ static int map_io_page(unsigned long ea,
 			       "memory at %016lx !\n", pa);
 			return -ENOMEM;
 		}
+#endif /* !CONFIG_PPC_MMU_NOHASH */
 	}
 	return 0;
 }
@@ -124,7 +173,7 @@ void __iomem * __ioremap_at(phys_addr_t 
 	WARN_ON(size & ~PAGE_MASK);
 
 	for (i = 0; i < size; i += PAGE_SIZE)
-		if (map_io_page((unsigned long)ea+i, pa+i, flags))
+		if (map_kernel_page((unsigned long)ea+i, pa+i, flags))
 			return NULL;
 
 	return (void __iomem *)ea;

^ permalink raw reply

* [PATCH 5/20] powerpc/mm: Add more bit definitions for Book3E MMU registers
From: Benjamin Herrenschmidt @ 2009-07-23  5:59 UTC (permalink / raw)
  To: linuxppc-dev

This adds various additional bit definitions for various MMU related
SPRs used on Book3E.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---

 arch/powerpc/include/asm/mmu-book3e.h |  168 ++++++++++++++++++++++++----------
 1 file changed, 119 insertions(+), 49 deletions(-)

--- linux-work.orig/arch/powerpc/include/asm/mmu-book3e.h	2009-07-22 15:19:22.000000000 +1000
+++ linux-work/arch/powerpc/include/asm/mmu-book3e.h	2009-07-22 15:30:58.000000000 +1000
@@ -38,58 +38,128 @@
 #define BOOK3E_PAGESZ_1TB	30
 #define BOOK3E_PAGESZ_2TB	31
 
-#define MAS0_TLBSEL(x)	((x << 28) & 0x30000000)
-#define MAS0_ESEL(x)	((x << 16) & 0x0FFF0000)
-#define MAS0_NV(x)	((x) & 0x00000FFF)
-
-#define MAS1_VALID 	0x80000000
-#define MAS1_IPROT	0x40000000
-#define MAS1_TID(x)	((x << 16) & 0x3FFF0000)
-#define MAS1_IND	0x00002000
-#define MAS1_TS		0x00001000
-#define MAS1_TSIZE(x)	((x << 7) & 0x00000F80)
-
-#define MAS2_EPN	0xFFFFF000
-#define MAS2_X0		0x00000040
-#define MAS2_X1		0x00000020
-#define MAS2_W		0x00000010
-#define MAS2_I		0x00000008
-#define MAS2_M		0x00000004
-#define MAS2_G		0x00000002
-#define MAS2_E		0x00000001
+/* MAS registers bit definitions */
+
+#define MAS0_TLBSEL(x)		((x << 28) & 0x30000000)
+#define MAS0_ESEL(x)		((x << 16) & 0x0FFF0000)
+#define MAS0_NV(x)		((x) & 0x00000FFF)
+#define MAS0_HES		0x00004000
+#define MAS0_WQ_ALLWAYS		0x00000000
+#define MAS0_WQ_COND		0x00001000
+#define MAS0_WQ_CLR_RSRV       	0x00002000
+
+#define MAS1_VALID		0x80000000
+#define MAS1_IPROT		0x40000000
+#define MAS1_TID(x)		((x << 16) & 0x3FFF0000)
+#define MAS1_IND		0x00002000
+#define MAS1_TS			0x00001000
+#define MAS1_TSIZE_MASK		0x00000f80
+#define MAS1_TSIZE_SHIFT	7
+#define MAS1_TSIZE(x)		((x << MAS1_TSIZE_SHIFT) & MAS1_TSIZE_MASK)
+
+#define MAS2_EPN		0xFFFFF000
+#define MAS2_X0			0x00000040
+#define MAS2_X1			0x00000020
+#define MAS2_W			0x00000010
+#define MAS2_I			0x00000008
+#define MAS2_M			0x00000004
+#define MAS2_G			0x00000002
+#define MAS2_E			0x00000001
 #define MAS2_EPN_MASK(size)		(~0 << (size + 10))
 #define MAS2_VAL(addr, size, flags)	((addr) & MAS2_EPN_MASK(size) | (flags))
 
-#define MAS3_RPN	0xFFFFF000
-#define MAS3_U0		0x00000200
-#define MAS3_U1		0x00000100
-#define MAS3_U2		0x00000080
-#define MAS3_U3		0x00000040
-#define MAS3_UX		0x00000020
-#define MAS3_SX		0x00000010
-#define MAS3_UW		0x00000008
-#define MAS3_SW		0x00000004
-#define MAS3_UR		0x00000002
-#define MAS3_SR		0x00000001
-
-#define MAS4_TLBSELD(x) MAS0_TLBSEL(x)
-#define MAS4_INDD	0x00008000
-#define MAS4_TSIZED(x)	MAS1_TSIZE(x)
-#define MAS4_X0D	0x00000040
-#define MAS4_X1D	0x00000020
-#define MAS4_WD		0x00000010
-#define MAS4_ID		0x00000008
-#define MAS4_MD		0x00000004
-#define MAS4_GD		0x00000002
-#define MAS4_ED		0x00000001
-
-#define MAS6_SPID0	0x3FFF0000
-#define MAS6_SPID1	0x00007FFE
-#define MAS6_ISIZE(x)	MAS1_TSIZE(x)
-#define MAS6_SAS	0x00000001
-#define MAS6_SPID	MAS6_SPID0
-
-#define MAS7_RPN	0xFFFFFFFF
+#define MAS3_RPN		0xFFFFF000
+#define MAS3_U0			0x00000200
+#define MAS3_U1			0x00000100
+#define MAS3_U2			0x00000080
+#define MAS3_U3			0x00000040
+#define MAS3_UX			0x00000020
+#define MAS3_SX			0x00000010
+#define MAS3_UW			0x00000008
+#define MAS3_SW			0x00000004
+#define MAS3_UR			0x00000002
+#define MAS3_SR			0x00000001
+#define MAS3_SPSIZE		0x0000003e
+#define MAS3_SPSIZE_SHIFT	1
+
+#define MAS4_TLBSELD(x) 	MAS0_TLBSEL(x)
+#define MAS4_INDD		0x00008000	/* Default IND */
+#define MAS4_TSIZED(x)		MAS1_TSIZE(x)
+#define MAS4_X0D		0x00000040
+#define MAS4_X1D		0x00000020
+#define MAS4_WD			0x00000010
+#define MAS4_ID			0x00000008
+#define MAS4_MD			0x00000004
+#define MAS4_GD			0x00000002
+#define MAS4_ED			0x00000001
+#define MAS4_WIMGED_MASK	0x0000001f	/* Default WIMGE */
+#define MAS4_WIMGED_SHIFT	0
+#define MAS4_VLED		MAS4_X1D	/* Default VLE */
+#define MAS4_ACMD		0x000000c0	/* Default ACM */
+#define MAS4_ACMD_SHIFT		6
+#define MAS4_TSIZED_MASK	0x00000f80	/* Default TSIZE */
+#define MAS4_TSIZED_SHIFT	7
+
+#define MAS6_SPID0		0x3FFF0000
+#define MAS6_SPID1		0x00007FFE
+#define MAS6_ISIZE(x)		MAS1_TSIZE(x)
+#define MAS6_SAS		0x00000001
+#define MAS6_SPID		MAS6_SPID0
+#define MAS6_SIND 		0x00000002	/* Indirect page */
+#define MAS6_SIND_SHIFT		1
+#define MAS6_SPID_MASK		0x3fff0000
+#define MAS6_SPID_SHIFT		16
+#define MAS6_ISIZE_MASK		0x00000f80
+#define MAS6_ISIZE_SHIFT	7
+
+#define MAS7_RPN		0xFFFFFFFF
+
+/* TLBnCFG encoding */
+#define TLBnCFG_N_ENTRY		0x00000fff	/* number of entries */
+#define TLBnCFG_HES		0x00002000	/* HW select supported */
+#define TLBnCFG_IPROT		0x00008000	/* IPROT supported */
+#define TLBnCFG_GTWE		0x00010000	/* Guest can write */
+#define TLBnCFG_IND		0x00020000	/* IND entries supported */
+#define TLBnCFG_PT		0x00040000	/* Can load from page table */
+#define TLBnCFG_ASSOC		0xff000000	/* Associativity */
+
+/* TLBnPS encoding */
+#define TLBnPS_4K		0x00000004
+#define TLBnPS_8K		0x00000008
+#define TLBnPS_16K		0x00000010
+#define TLBnPS_32K		0x00000020
+#define TLBnPS_64K		0x00000040
+#define TLBnPS_128K		0x00000080
+#define TLBnPS_256K		0x00000100
+#define TLBnPS_512K		0x00000200
+#define TLBnPS_1M 		0x00000400
+#define TLBnPS_2M 		0x00000800
+#define TLBnPS_4M 		0x00001000
+#define TLBnPS_8M 		0x00002000
+#define TLBnPS_16M		0x00004000
+#define TLBnPS_32M		0x00008000
+#define TLBnPS_64M		0x00010000
+#define TLBnPS_128M		0x00020000
+#define TLBnPS_256M		0x00040000
+#define TLBnPS_512M		0x00080000
+#define TLBnPS_1G		0x00100000
+#define TLBnPS_2G		0x00200000
+#define TLBnPS_4G		0x00400000
+#define TLBnPS_8G		0x00800000
+#define TLBnPS_16G		0x01000000
+#define TLBnPS_32G		0x02000000
+#define TLBnPS_64G		0x04000000
+#define TLBnPS_128G		0x08000000
+#define TLBnPS_256G		0x10000000
+
+/* tlbilx action encoding */
+#define TLBILX_T_ALL			0
+#define TLBILX_T_TID			1
+#define TLBILX_T_FULLMATCH		3
+#define TLBILX_T_CLASS0			4
+#define TLBILX_T_CLASS1			5
+#define TLBILX_T_CLASS2			6
+#define TLBILX_T_CLASS3			7
 
 #ifndef __ASSEMBLY__
 

^ permalink raw reply

* [PATCH 4/20] powerpc/mm: Add opcode definitions for tlbivax and tlbsrx.
From: Benjamin Herrenschmidt @ 2009-07-23  5:59 UTC (permalink / raw)
  To: linuxppc-dev

This adds the opcode definitions to ppc-opcode.h for the two instructions
tlbivax and tlbsrx. as defined by Book3E 2.06

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---

 arch/powerpc/include/asm/ppc-opcode.h |    6 ++++++
 1 file changed, 6 insertions(+)

--- linux-work.orig/arch/powerpc/include/asm/ppc-opcode.h	2009-07-22 15:25:45.000000000 +1000
+++ linux-work/arch/powerpc/include/asm/ppc-opcode.h	2009-07-22 15:26:05.000000000 +1000
@@ -48,6 +48,8 @@
 #define PPC_INST_TLBIE			0x7c000264
 #define PPC_INST_TLBILX			0x7c000024
 #define PPC_INST_WAIT			0x7c00007c
+#define PPC_INST_TLBIVAX		0x7c000624
+#define PPC_INST_TLBSRX_DOT		0x7c0006a5
 
 /* macros to insert fields into opcodes */
 #define __PPC_RA(a)	(((a) & 0x1f) << 16)
@@ -76,6 +78,10 @@
 					__PPC_WC(w))
 #define PPC_TLBIE(lp,a) 	stringify_in_c(.long PPC_INST_TLBIE | \
 					       __PPC_RB(a) | __PPC_RS(lp))
+#define PPC_TLBSRX_DOT(a,b)	stringify_in_c(.long PPC_INST_TLBSRX_DOT | \
+					__PPC_RA(a) | __PPC_RB(b))
+#define PPC_TLBIVAX(a,b)	stringify_in_c(.long PPC_INST_TLBIVAX | \
+					__PPC_RA(a) | __PPC_RB(b))
 
 /*
  * Define what the VSX XX1 form instructions will look like, then add

^ permalink raw reply

* [PATCH 3/20] powerpc/mm: Add HW threads support to no_hash TLB management
From: Benjamin Herrenschmidt @ 2009-07-23  5:59 UTC (permalink / raw)
  To: linuxppc-dev

The current "no hash" MMU context management code is written with
the assumption that one CPU == one TLB. This is not the case on
implementations that support HW multithreading, where several
linux CPUs can share the same TLB.

This adds some basic support for this to our context management
and our TLB flushing code.

It also cleans up the optional debugging output a bit

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---

 arch/powerpc/include/asm/cputhreads.h |   16 +++++
 arch/powerpc/mm/mmu_context_nohash.c  |   93 ++++++++++++++++++++++------------
 arch/powerpc/mm/tlb_nohash.c          |   10 ++-
 3 files changed, 86 insertions(+), 33 deletions(-)

--- linux-work.orig/arch/powerpc/mm/mmu_context_nohash.c	2009-07-21 12:43:27.000000000 +1000
+++ linux-work/arch/powerpc/mm/mmu_context_nohash.c	2009-07-21 12:56:16.000000000 +1000
@@ -25,10 +25,20 @@
  *     also clear mm->cpu_vm_mask bits when processes are migrated
  */
 
-#undef DEBUG
-#define DEBUG_STEAL_ONLY
-#undef DEBUG_MAP_CONSISTENCY
-/*#define DEBUG_CLAMP_LAST_CONTEXT   15 */
+#define DEBUG_MAP_CONSISTENCY
+#define DEBUG_CLAMP_LAST_CONTEXT   31
+//#define DEBUG_HARDER
+
+/* We don't use DEBUG because it tends to be compiled in always nowadays
+ * and this would generate way too much output
+ */
+#ifdef DEBUG_HARDER
+#define pr_hard(args...)	printk(KERN_DEBUG args)
+#define pr_hardcont(args...)	printk(KERN_CONT args)
+#else
+#define pr_hard(args...)	do { } while(0)
+#define pr_hardcont(args...)	do { } while(0)
+#endif
 
 #include <linux/kernel.h>
 #include <linux/mm.h>
@@ -71,7 +81,7 @@ static DEFINE_SPINLOCK(context_lock);
 static unsigned int steal_context_smp(unsigned int id)
 {
 	struct mm_struct *mm;
-	unsigned int cpu, max;
+	unsigned int cpu, max, i;
 
 	max = last_context - first_context;
 
@@ -89,15 +99,22 @@ static unsigned int steal_context_smp(un
 				id = first_context;
 			continue;
 		}
-		pr_devel("[%d] steal context %d from mm @%p\n",
-			 smp_processor_id(), id, mm);
+		pr_hardcont(" | steal %d from 0x%p", id, mm);
 
 		/* Mark this mm has having no context anymore */
 		mm->context.id = MMU_NO_CONTEXT;
 
-		/* Mark it stale on all CPUs that used this mm */
-		for_each_cpu(cpu, mm_cpumask(mm))
-			__set_bit(id, stale_map[cpu]);
+		/* Mark it stale on all CPUs that used this mm. For threaded
+		 * implementations, we set it on all threads on each core
+		 * represented in the mask. A future implementation will use
+		 * a core map instead but this will do for now.
+		 */
+		for_each_cpu(cpu, mm_cpumask(mm)) {
+			for (i = cpu_first_thread_in_core(cpu);
+			     i <= cpu_last_thread_in_core(cpu); i++)
+				__set_bit(id, stale_map[i]);
+			cpu = i - 1;
+		}
 		return id;
 	}
 
@@ -126,7 +143,7 @@ static unsigned int steal_context_up(uns
 	/* Pick up the victim mm */
 	mm = context_mm[id];
 
-	pr_devel("[%d] steal context %d from mm @%p\n", cpu, id, mm);
+	pr_hardcont(" | steal %d from 0x%p", id, mm);
 
 	/* Flush the TLB for that context */
 	local_flush_tlb_mm(mm);
@@ -179,19 +196,14 @@ void switch_mmu_context(struct mm_struct
 	/* No lockless fast path .. yet */
 	spin_lock(&context_lock);
 
-#ifndef DEBUG_STEAL_ONLY
-	pr_devel("[%d] activating context for mm @%p, active=%d, id=%d\n",
-		 cpu, next, next->context.active, next->context.id);
-#endif
+	pr_hard("[%d] activating context for mm @%p, active=%d, id=%d",
+		cpu, next, next->context.active, next->context.id);
 
 #ifdef CONFIG_SMP
 	/* Mark us active and the previous one not anymore */
 	next->context.active++;
 	if (prev) {
-#ifndef DEBUG_STEAL_ONLY
-		pr_devel(" old context %p active was: %d\n",
-			 prev, prev->context.active);
-#endif
+		pr_hardcont(" (old=0x%p a=%d)", prev, prev->context.active);
 		WARN_ON(prev->context.active < 1);
 		prev->context.active--;
 	}
@@ -201,8 +213,14 @@ void switch_mmu_context(struct mm_struct
 
 	/* If we already have a valid assigned context, skip all that */
 	id = next->context.id;
-	if (likely(id != MMU_NO_CONTEXT))
+	if (likely(id != MMU_NO_CONTEXT)) {
+#ifdef DEBUG_MAP_CONSISTENCY
+		if (context_mm[id] != next)
+			pr_err("MMU: mm 0x%p has id %d but context_mm[%d] says 0x%p\n",
+			       next, id, id, context_mm[id]);
+#endif
 		goto ctxt_ok;
+	}
 
 	/* We really don't have a context, let's try to acquire one */
 	id = next_context;
@@ -234,11 +252,7 @@ void switch_mmu_context(struct mm_struct
 	next_context = id + 1;
 	context_mm[id] = next;
 	next->context.id = id;
-
-#ifndef DEBUG_STEAL_ONLY
-	pr_devel("[%d] picked up new id %d, nrf is now %d\n",
-		 cpu, id, nr_free_contexts);
-#endif
+	pr_hardcont(" | new id=%d,nrf=%d", id, nr_free_contexts);
 
 	context_check_map();
  ctxt_ok:
@@ -247,15 +261,20 @@ void switch_mmu_context(struct mm_struct
 	 * local TLB for it and unmark it before we use it
 	 */
 	if (test_bit(id, stale_map[cpu])) {
-		pr_devel("[%d] flushing stale context %d for mm @%p !\n",
-			 cpu, id, next);
+		pr_hardcont(" | stale flush %d [%d..%d]",
+			    id, cpu_first_thread_in_core(cpu),
+			    cpu_last_thread_in_core(cpu));
+
 		local_flush_tlb_mm(next);
 
 		/* XXX This clear should ultimately be part of local_flush_tlb_mm */
-		__clear_bit(id, stale_map[cpu]);
+		for (cpu = cpu_first_thread_in_core(cpu);
+		     cpu <= cpu_last_thread_in_core(cpu); cpu++)
+			__clear_bit(id, stale_map[cpu]);
 	}
 
 	/* Flick the MMU and release lock */
+	pr_hardcont(" -> %d\n", id);
 	set_context(id, next->pgd);
 	spin_unlock(&context_lock);
 }
@@ -265,6 +284,8 @@ void switch_mmu_context(struct mm_struct
  */
 int init_new_context(struct task_struct *t, struct mm_struct *mm)
 {
+	pr_hard("initing context for mm @%p\n", mm);
+
 	mm->context.id = MMU_NO_CONTEXT;
 	mm->context.active = 0;
 
@@ -304,7 +325,9 @@ static int __cpuinit mmu_context_cpu_not
 					    unsigned long action, void *hcpu)
 {
 	unsigned int cpu = (unsigned int)(long)hcpu;
-
+#ifdef CONFIG_HOTPLUG_CPU
+	struct task_struct *p;
+#endif
 	/* We don't touch CPU 0 map, it's allocated at aboot and kept
 	 * around forever
 	 */
@@ -323,8 +346,16 @@ static int __cpuinit mmu_context_cpu_not
 		pr_devel("MMU: Freeing stale context map for CPU %d\n", cpu);
 		kfree(stale_map[cpu]);
 		stale_map[cpu] = NULL;
-		break;
-#endif
+
+		/* We also clear the cpu_vm_mask bits of CPUs going away */
+		read_lock(&tasklist_lock);
+		for_each_process(p) {
+			if (p->mm)
+				cpu_mask_clear_cpu(cpu, mm_cpumask(p->mm));
+		}
+		read_unlock(&tasklist_lock);
+	break;
+#endif /* CONFIG_HOTPLUG_CPU */
 	}
 	return NOTIFY_OK;
 }
Index: linux-work/arch/powerpc/include/asm/cputhreads.h
===================================================================
--- linux-work.orig/arch/powerpc/include/asm/cputhreads.h	2009-07-21 12:43:27.000000000 +1000
+++ linux-work/arch/powerpc/include/asm/cputhreads.h	2009-07-21 12:56:16.000000000 +1000
@@ -5,6 +5,15 @@
 
 /*
  * Mapping of threads to cores
+ *
+ * Note: This implementation is limited to a power of 2 number of
+ * threads per core and the same number for each core in the system
+ * (though it would work if some processors had less threads as long
+ * as the CPU numbers are still allocated, just not brought offline).
+ *
+ * However, the API allows for a different implementation in the future
+ * if needed, as long as you only use the functions and not the variables
+ * directly.
  */
 
 #ifdef CONFIG_SMP
@@ -67,5 +76,12 @@ static inline int cpu_first_thread_in_co
 	return cpu & ~(threads_per_core - 1);
 }
 
+static inline int cpu_last_thread_in_core(int cpu)
+{
+	return cpu | (threads_per_core - 1);
+}
+
+
+
 #endif /* _ASM_POWERPC_CPUTHREADS_H */
 
Index: linux-work/arch/powerpc/mm/tlb_nohash.c
===================================================================
--- linux-work.orig/arch/powerpc/mm/tlb_nohash.c	2009-07-21 12:43:31.000000000 +1000
+++ linux-work/arch/powerpc/mm/tlb_nohash.c	2009-07-21 12:57:21.000000000 +1000
@@ -87,6 +87,12 @@ EXPORT_SYMBOL(local_flush_tlb_page);
 
 static DEFINE_SPINLOCK(tlbivax_lock);
 
+static int mm_is_core_local(struct mm_struct *mm)
+{
+	return cpumask_subset(mm_cpumask(mm),
+			      topology_thread_cpumask(smp_processor_id()));
+}
+
 struct tlb_flush_param {
 	unsigned long addr;
 	unsigned int pid;
@@ -131,7 +137,7 @@ void flush_tlb_mm(struct mm_struct *mm)
 	pid = mm->context.id;
 	if (unlikely(pid == MMU_NO_CONTEXT))
 		goto no_context;
-	if (!cpumask_equal(mm_cpumask(mm), cpumask_of(smp_processor_id()))) {
+	if (!mm_is_core_local(mm)) {
 		struct tlb_flush_param p = { .pid = pid };
 		/* Ignores smp_processor_id() even if set. */
 		smp_call_function_many(mm_cpumask(mm),
@@ -153,7 +159,7 @@ void flush_tlb_page(struct vm_area_struc
 	if (unlikely(pid == MMU_NO_CONTEXT))
 		goto bail;
 	cpu_mask = mm_cpumask(vma->vm_mm);
-	if (!cpumask_equal(cpu_mask, cpumask_of(smp_processor_id()))) {
+	if (!mm_is_core_local(mm)) {
 		/* If broadcast tlbivax is supported, use it */
 		if (mmu_has_feature(MMU_FTR_USE_TLBIVAX_BCAST)) {
 			int lock = mmu_has_feature(MMU_FTR_LOCK_BCAST_INVAL);

^ permalink raw reply

* [PATCH 2/20] powerpc/of: Remove useless register save/restore when calling OF back
From: Benjamin Herrenschmidt @ 2009-07-23  5:59 UTC (permalink / raw)
  To: linuxppc-dev

enter_prom() used to save and restore registers such as CTR, XER etc..
which are volatile, or SRR0,1... which we don't care about. This
removes a bunch of useless code and while at it turns an mtmsrd into
an MTMSRD macro which will be useful to Book3E.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---

 arch/powerpc/kernel/entry_64.S |   38 ++++++--------------------------------
 1 file changed, 6 insertions(+), 32 deletions(-)

--- linux-work.orig/arch/powerpc/kernel/entry_64.S	2009-07-22 15:20:26.000000000 +1000
+++ linux-work/arch/powerpc/kernel/entry_64.S	2009-07-22 15:22:44.000000000 +1000
@@ -823,30 +823,17 @@ _GLOBAL(enter_prom)
 	 * of all registers that it saves.  We therefore save those registers
 	 * PROM might touch to the stack.  (r0, r3-r13 are caller saved)
    	 */
-	SAVE_8GPRS(2, r1)
+	SAVE_GPR(2, r1)
 	SAVE_GPR(13, r1)
 	SAVE_8GPRS(14, r1)
 	SAVE_10GPRS(22, r1)
-	mfcr	r4
-	std	r4,_CCR(r1)
-	mfctr	r5
-	std	r5,_CTR(r1)
-	mfspr	r6,SPRN_XER
-	std	r6,_XER(r1)
-	mfdar	r7
-	std	r7,_DAR(r1)
-	mfdsisr	r8
-	std	r8,_DSISR(r1)
-	mfsrr0	r9
-	std	r9,_SRR0(r1)
-	mfsrr1	r10
-	std	r10,_SRR1(r1)
+	mfcr	r10
 	mfmsr	r11
+	std	r10,_CCR(r1)
 	std	r11,_MSR(r1)
 
 	/* Get the PROM entrypoint */
-	ld	r0,GPR4(r1)
-	mtlr	r0
+	mtlr	r4
 
 	/* Switch MSR to 32 bits mode
 	 */
@@ -860,8 +847,7 @@ _GLOBAL(enter_prom)
         mtmsrd  r11
         isync
 
-	/* Restore arguments & enter PROM here... */
-	ld	r3,GPR3(r1)
+	/* Enter PROM here... */
 	blrl
 
 	/* Just make sure that r1 top 32 bits didn't get
@@ -871,7 +857,7 @@ _GLOBAL(enter_prom)
 
 	/* Restore the MSR (back to 64 bits) */
 	ld	r0,_MSR(r1)
-	mtmsrd	r0
+	MTMSRD(r0)
         isync
 
 	/* Restore other registers */
@@ -881,18 +867,6 @@ _GLOBAL(enter_prom)
 	REST_10GPRS(22, r1)
 	ld	r4,_CCR(r1)
 	mtcr	r4
-	ld	r5,_CTR(r1)
-	mtctr	r5
-	ld	r6,_XER(r1)
-	mtspr	SPRN_XER,r6
-	ld	r7,_DAR(r1)
-	mtdar	r7
-	ld	r8,_DSISR(r1)
-	mtdsisr	r8
-	ld	r9,_SRR0(r1)
-	mtsrr0	r9
-	ld	r10,_SRR1(r1)
-	mtsrr1	r10
 	
         addi	r1,r1,PROM_FRAME_SIZE
 	ld	r0,16(r1)

^ permalink raw reply

* [PATCH 1/20] powerpc/mm: Fix misplaced #endif in pgtable-ppc64-64k.h
From: Benjamin Herrenschmidt @ 2009-07-23  5:59 UTC (permalink / raw)
  To: linuxppc-dev

A misplaced #endif causes more definitions than intended to be
protected by #ifndef __ASSEMBLY__. This breaks upcoming 64-bit
BookE support patch when using 64k pages.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---

 arch/powerpc/include/asm/pgtable-ppc64-64k.h |    4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

--- linux-work.orig/arch/powerpc/include/asm/pgtable-ppc64-64k.h	2009-07-22 11:45:34.000000000 +1000
+++ linux-work/arch/powerpc/include/asm/pgtable-ppc64-64k.h	2009-07-22 11:45:47.000000000 +1000
@@ -10,10 +10,10 @@
 #define PGD_INDEX_SIZE  4
 
 #ifndef __ASSEMBLY__
-
 #define PTE_TABLE_SIZE	(sizeof(real_pte_t) << PTE_INDEX_SIZE)
 #define PMD_TABLE_SIZE	(sizeof(pmd_t) << PMD_INDEX_SIZE)
 #define PGD_TABLE_SIZE	(sizeof(pgd_t) << PGD_INDEX_SIZE)
+#endif	/* __ASSEMBLY__ */
 
 #define PTRS_PER_PTE	(1 << PTE_INDEX_SIZE)
 #define PTRS_PER_PMD	(1 << PMD_INDEX_SIZE)
@@ -32,8 +32,6 @@
 #define PGDIR_SIZE	(1UL << PGDIR_SHIFT)
 #define PGDIR_MASK	(~(PGDIR_SIZE-1))
 
-#endif	/* __ASSEMBLY__ */
-
 /* Bits to mask out from a PMD to get to the PTE page */
 #define PMD_MASKED_BITS		0x1ff
 /* Bits to mask out from a PGD/PUD to get to the PMD page */

^ permalink raw reply

* Re: [PATCH] Hold reference to device_node during EEH event handling
From: Michael Ellerman @ 2009-07-23  1:21 UTC (permalink / raw)
  To: Mike Mason; +Cc: linuxppc-dev, linasvepstas, Paul Mackerras
In-Reply-To: <4A67A3A2.4050508@us.ibm.com>

[-- Attachment #1: Type: text/plain, Size: 2562 bytes --]

On Wed, 2009-07-22 at 16:41 -0700, Mike Mason wrote:
> Michael Ellerman wrote:
> > On Thu, 2009-07-16 at 09:33 -0700, Mike Mason wrote:
> >> Michael Ellerman wrote:
> >>> On Wed, 2009-07-15 at 14:43 -0700, Mike Mason wrote:
> >>>> This patch increments the device_node reference counter when an EEH
> >>>> error occurs and decrements the counter when the event has been
> >>>> handled.  This is to prevent the device_node from being released until
> >>>> eeh_event_handler() has had a chance to deal with the event.  We've
> >>>> seen cases where the device_node is released too soon when an EEH
> >>>> event occurs during a dlpar remove, causing the event handler to
> >>>> attempt to access bad memory locations.
> >>>>
> >>>> Please review and let me know of any concerns.
> >>> Taking a reference sounds sane, but ...
> >>>
> >>>> Signed-off-by: Mike Mason <mmlnx@us.ibm.com> 
> >>>>
> >>>> --- a/arch/powerpc/platforms/pseries/eeh_event.c	2008-10-09 15:13:53.000000000 -0700
> >>>> +++ b/arch/powerpc/platforms/pseries/eeh_event.c	2009-07-14 14:14:00.000000000 -0700
> >>>> @@ -75,6 +75,14 @@ static int eeh_event_handler(void * dumm
> >>>>  	if (event == NULL)
> >>>>  		return 0;
> >>>>  
> >>>> +	/* EEH holds a reference to the device_node, so if it
> >>>> +	 * equals 1 it's no longer valid and the event should
> >>>> +	 * be ignored */
> >>>> +	if (atomic_read(&event->dn->kref.refcount) == 1) {
> >>>> +		of_node_put(event->dn);
> >>>> +		return 0;
> >>>> +	}
> >>> That's really gross :)
> >> Agreed.  I'll look for another way to determine if device is gone and
> >> the event should be ignored.  Suggestions are welcome :-)
> 
> Actually, it turns out the atomic_read() isn't necessary.  I just need
> to take the reference to the device_node when the EEH error is
> detected and let EEH try to handle the error.  EEH detects the fact
> that the device is no longer valid, aborts the recovery attempt, then
> gives the device_node reference back.  Works as expected.

How does it detect that the device is no longer valid?

> I'll resubmit the patch without the atomic_read().
> 
> > 
> > Benh and I had a quick chat about it, and were wondering whether what
> > you really should be doing is taking a reference to the pci device
> > (perhaps as well as the device node).
> 
> EEH already does that 3 lines before the of_node_get (see below).

Ah right, while you're touching the code, mind changing it to the
simpler and more obvious:

>         event->dev = pci_dev_get(dev);

cheers


[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 197 bytes --]

^ permalink raw reply

* Re: [LTP] msgctl10 fails on Powerpc Linux-2.6.29.6
From: Kumar Gala @ 2009-07-23  1:03 UTC (permalink / raw)
  To: srikanth krishnakar; +Cc: Linuxppc-dev, Ltp-list
In-Reply-To: <6213bc560907220932u3b771af2q743ae55dda5a2198@mail.gmail.com>


On Jul 22, 2009, at 11:32 AM, srikanth krishnakar wrote:

>
>
> On Wed, Jul 22, 2009 at 9:14 PM, Kumar Gala  
> <galak@kernel.crashing.org> wrote:
>
> On Jul 22, 2009, at 10:38 AM, srikanth krishnakar wrote:
>
>
>
> On Wed, Jul 22, 2009 at 8:52 PM, Kumar Gala  
> <galak@kernel.crashing.org> wrote:
> I'm not seeing any BUG* in traps.c @ line 904.
>
>
> On Jul 21, 2009, at 4:33 AM, srikanth krishnakar wrote:
>
> The LTP test case msgctl10.c fails on linux-2.6.29.6 for PowerPC  
> architecture (ppc440)
>
>
> msgctl10    1  B------------[ cut here ]------------
> kernel BUG at arch/powerpc/kernel/traps.c:904!
> Oops: Exception in kernel mode, sig: 5 [#9]
>
> I'm not seeing any BUG* in traps.c @ line 904.  Do you have some  
> other patches on top of 2.6.29.6?
>
> - k
>
>
> I have LTTng patches on top of linux-2.6.29.6.
>
> Does it modify traps.c?  Can you see what the code around line 904  
> looks like and post that.
>
> - k
>  896
>  897 #if defined(CONFIG_XILINX_VIRTEX_5_FXT) &&  
> defined(CONFIG_PPC_FPU)
>  898         if (reason & REASON_ILLEGAL) {
>  899                 if (excep_state < 1) {
>  900                         excep_state++;
>  901                         return;
>  902                 }
>  903                 /* should never get here */
>  904                 BUG();
>  905         }
>  906 #endif
>  907
>  908         /* Try to emulate it if we should. */
>

Are you sure this is coming from the LTTng patches?  Which actual  
patch makes this change?

- k

^ permalink raw reply

* Re: [RFC/PATCH] mm: Pass virtual address to [__]p{te, ud, md}_free_tlb()
From: Kumar Gala @ 2009-07-23  0:59 UTC (permalink / raw)
  To: Benjamin Herrenschmidt
  Cc: Nick Piggin, Linux-Arch, Linux Memory Management, Hugh Dickins,
	linux-kernel, linuxppc-dev, Linus Torvalds
In-Reply-To: <1248310415.3367.22.camel@pasglop>


On Jul 22, 2009, at 7:53 PM, Benjamin Herrenschmidt wrote:

> On Wed, 2009-07-22 at 09:31 -0700, Linus Torvalds wrote:
>>> CC'ing Linus here. How do you want to proceed with that merge ?  
>>> (IE. so
>>> far nobody objected to the patch itself)
>>
>> Maybe you can put it as a separate branch in -next, and have it  
>> merged
>> before the stuff that depends on it, and then just sending it to me  
>> (as a
>> git branch or patch or whatever) in the first day of the merge  
>> window?
>
> Hrm... my powerpc-next branch will contain stuff that depend on it, so
> I'll probably have to pull it in though, unless I tell all my
> sub-maintainers to also pull from that other branch first :-)

Can you not cherry pick it into powerpc-next to 'pull it through'?

- k

^ permalink raw reply

* Re: [RFC/PATCH] mm: Pass virtual address to [__]p{te,ud,md}_free_tlb()
From: Benjamin Herrenschmidt @ 2009-07-23  0:53 UTC (permalink / raw)
  To: Linus Torvalds
  Cc: Nick Piggin, Linux-Arch, linuxppc-dev, Hugh Dickins, linux-kernel,
	Linux Memory Management
In-Reply-To: <alpine.LFD.2.01.0907220930320.19335@localhost.localdomain>

On Wed, 2009-07-22 at 09:31 -0700, Linus Torvalds wrote:
> > CC'ing Linus here. How do you want to proceed with that merge ? (IE. so
> > far nobody objected to the patch itself)
> 
> Maybe you can put it as a separate branch in -next, and have it merged 
> before the stuff that depends on it, and then just sending it to me (as a 
> git branch or patch or whatever) in the first day of the merge window?

Hrm... my powerpc-next branch will contain stuff that depend on it, so
I'll probably have to pull it in though, unless I tell all my
sub-maintainers to also pull from that other branch first :-)

Cheers,
Ben.

^ permalink raw reply

* Re: [PATCH v3] net: Rework mdio-ofgpio driver to use of_mdio infrastructure
From: David Miller @ 2009-07-23  0:11 UTC (permalink / raw)
  To: mware; +Cc: netdev, linuxppc-dev
In-Reply-To: <16861.1248307810@internode.on.net>

From: Mark Ware <mware@elphinstone.net>
Date: Thu, 23 Jul 2009 09:40:10 +0930

> 
> The patch is actually attached this time.  (I hope - did I mention my dislike for this webmail interface?)

No it isn't!

And anyways you have to properly provide the full commit message
and signoffs when you resubmit patches.  You can't just resubmit
the patch because all of the context is lost in patchwork.

Grrr...

^ permalink raw reply

* Re: [PATCH v3] net: Rework mdio-ofgpio driver to use of_mdioinfrastructure
From: David Miller @ 2009-07-23  0:10 UTC (permalink / raw)
  To: mware; +Cc: netdev, linuxppc-dev
In-Reply-To: <16811.1248307327@internode.on.net>

From: Mark Ware <mware@elphinstone.net>
Date: Thu, 23 Jul 2009 09:02:07 +0900

> Patch is attached, hopefully with correct MIME type - this webmail
> client is worse than Outlook.

Where is it attached? :-/

^ permalink raw reply

* Re: Re: [PATCH v3] net: Rework mdio-ofgpio driver to use of_mdio infrastructure
From: Mark Ware @ 2009-07-23  0:10 UTC (permalink / raw)
  To: davem; +Cc: netdev, linuxppc-dev

 On Thu 23/07/09  2:10 AM , David Miller davem@davemloft.net sent:
> Your email client has massively corrupted this patch, I even tried
> to fix it up but it's so bad that I just gave up half-way through.
>=20
> Please use an attachment or whatever is necessary to get this
> patch submitted cleanly and without it getting damaged.  Read:
>=20
> linux/Documentation/email-clients.txt
>=20
> for tips.
>=20
> Thanks.

Dave,

The patch is actually attached this time.  (I hope - did I mention my disli=
ke for this webmail interface?)

Sorry for the further noise.

Regards,
Mark

^ permalink raw reply

* Re: [PATCH v3] net: Rework mdio-ofgpio driver to use of_mdioinfrastructure
From: Mark Ware @ 2009-07-23  0:02 UTC (permalink / raw)
  To: davem; +Cc: netdev, linuxppc-dev

 On Thu 23/07/09  2:10 AM , David Miller davem@davemloft.net sent:
> Your email client has massively corrupted this patch, I even tried
> to fix it up but it's so bad that I just gave up half-way through.
>=20
> Please use an attachment or whatever is necessary to get this
> patch submitted cleanly and without it getting damaged.  Read:
>=20
> linux/Documentation/email-clients.txt for tips.
>=20
> Thanks.

Sorry for the waste of time.  I've obviously somehow broken my thunderbird =
config.

Patch is attached, hopefully with correct MIME type - this webmail client i=
s worse than Outlook.

(Incidently, I was able to apply the corrupted patch after the following tw=
o substitutions: s/^  / / and s/^$/ /)

Regards,
Mark

^ permalink raw reply

* Re: [PATCH] Hold reference to device_node during EEH event handling
From: Mike Mason @ 2009-07-22 23:41 UTC (permalink / raw)
  To: michael; +Cc: linuxppc-dev, linasvepstas, Paul Mackerras
In-Reply-To: <1247790973.16836.11.camel@concordia>

Michael Ellerman wrote:
> On Thu, 2009-07-16 at 09:33 -0700, Mike Mason wrote:
>> Michael Ellerman wrote:
>>> On Wed, 2009-07-15 at 14:43 -0700, Mike Mason wrote:
>>>> This patch increments the device_node reference counter when an EEH
>>>> error occurs and decrements the counter when the event has been
>>>> handled.  This is to prevent the device_node from being released until
>>>> eeh_event_handler() has had a chance to deal with the event.  We've
>>>> seen cases where the device_node is released too soon when an EEH
>>>> event occurs during a dlpar remove, causing the event handler to
>>>> attempt to access bad memory locations.
>>>>
>>>> Please review and let me know of any concerns.
>>> Taking a reference sounds sane, but ...
>>>
>>>> Signed-off-by: Mike Mason <mmlnx@us.ibm.com> 
>>>>
>>>> --- a/arch/powerpc/platforms/pseries/eeh_event.c	2008-10-09 15:13:53.000000000 -0700
>>>> +++ b/arch/powerpc/platforms/pseries/eeh_event.c	2009-07-14 14:14:00.000000000 -0700
>>>> @@ -75,6 +75,14 @@ static int eeh_event_handler(void * dumm
>>>>  	if (event == NULL)
>>>>  		return 0;
>>>>  
>>>> +	/* EEH holds a reference to the device_node, so if it
>>>> +	 * equals 1 it's no longer valid and the event should
>>>> +	 * be ignored */
>>>> +	if (atomic_read(&event->dn->kref.refcount) == 1) {
>>>> +		of_node_put(event->dn);
>>>> +		return 0;
>>>> +	}
>>> That's really gross :)
>> Agreed.  I'll look for another way to determine if device is gone and
>> the event should be ignored.  Suggestions are welcome :-)

Actually, it turns out the atomic_read() isn't necessary.  I just need to take the reference to the device_node when the EEH error is detected and let EEH try to handle the error.  EEH detects the fact that the device is no longer valid, aborts the recovery attempt, then gives the device_node reference back.  Works as expected.

I'll resubmit the patch without the atomic_read().

> 
> Benh and I had a quick chat about it, and were wondering whether what
> you really should be doing is taking a reference to the pci device
> (perhaps as well as the device node).

EEH already does that 3 lines before the of_node_get (see below).

> 
> @@ -140,7 +149,7 @@ int eeh_send_failure_event (struct devic
>         if (dev)
>                 pci_dev_get(dev);
>  
> -       event->dn = dn;
> +       event->dn = of_node_get(dn);
>         event->dev = dev;
> 

Thanks,
Mike

^ permalink raw reply

* Re: Best hardware platform for native compiling...
From: Olof Johansson @ 2009-07-22 20:18 UTC (permalink / raw)
  To: Benjamin Herrenschmidt; +Cc: linuxppc-dev, David Jander
In-Reply-To: <1248180459.31121.9.camel@pasglop>

On Tue, Jul 21, 2009 at 10:47:39PM +1000, Benjamin Herrenschmidt wrote:
> On Tue, 2009-07-21 at 14:00 +0200, Gabriel Paubert wrote:
> > Well, I got one recently here in Spain. Shipping charges are fairly 
> > large (it's not exactly a light and compact machine). But the current 
> > dollar exchange rate helps ;-)
> > 
> > Now I have not yet found the way to install Debian on it
> > (it refuses to boot Debian's CDROM), but I have not had
> > time to investigate either.
> 
> It uses SLOF which should be capable of decent netbooting, so it should
> be possible to either netboot yaboot and have it load the kernel and
> initramfs, or maybe make a zImage with both included and netboot that.

Placing the debian (netinst) installer ramdisk and kernel on the /boot
partition and booting it from there should work too (after adding it to
the yaboot config).


-Olof

^ permalink raw reply

* Re: [PATCH v3] net: Rework mdio-ofgpio driver to use of_mdio infrastructure
From: David Miller @ 2009-07-22 16:40 UTC (permalink / raw)
  To: mware; +Cc: netdev, linuxppc-dev
In-Reply-To: <4A66400F.3090309@elphinstone.net>

From: Mark Ware <mware@elphinstone.net>
Date: Wed, 22 Jul 2009 08:24:15 +1000

> Changes to the fs_enet driver
> (aa73832c5a80d6c52c69b18af858d88fa595dd3c) cause kernel crashes when
> using the mdio-ofgpio driver.
> 
> This patch replicates similar changes made to the fs_enet mii-bitbang
> drivers.  It has been tested on a custom mpc8280 based board using an
> NFS mounted root.
> 
> Signed-off-by: Mark Ware <mware@elphinstone.net>
> Acked-by: Grant Likely <grant.likely@secretlab.ca>

Your email client has massively corrupted this patch, I even tried
to fix it up but it's so bad that I just gave up half-way through.

Please use an attachment or whatever is necessary to get this
patch submitted cleanly and without it getting damaged.  Read:

linux/Documentation/email-clients.txt

for tips.

Thanks.

^ permalink raw reply


This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox