* [Adeos-main] Advice regarding MIPS port
@ 2008-05-05 16:43 Florent Audebert
2008-05-06 19:54 ` Jan Kiszka
0 siblings, 1 reply; 6+ messages in thread
From: Florent Audebert @ 2008-05-05 16:43 UTC (permalink / raw)
To: adeos-main
[-- Attachment #1: Type: text/plain, Size: 1578 bytes --]
Hi,
I am starting to work on a Xenomai port for MIPS architecture as part
of an internship at Open Wide (France). It is my first real diving
into the kernel sources so I'm far from being an expert. The goal of
this project is to be able to run it on a Broadcom BCM6448 board.
My first objective is to be able to run an Adeos patched kernel. I
have decided to work with QEMU (mips-malta board emulation) instead of
my real target since it provides me an easy way to use GDB and thus to
understand a little bit more what is going on. Furthermore, it
provides an easy way of testing it for other people since it only
requires QEMU and no specific real hardware.
I read as much as documents I can (Life with Adeos, Porting Adeos,
...) and finally started messing with the code. I managed to compile a
MIPS kernel merging common Adeos code from latest x86 patch and
adapting (see below) / commenting out. I guess my kernel is far from
functional. My current issue seems to be related to the timer (see
attached file).
Here is a list of explicit changes I have done so far :
- include/asm-mips/irqflags.h : added raw_local_*() and local_irq_*().
- include/asm-mips/irq.h : do_IRQ() calls __ipipe_handle_irq()
instead of generic_handle_irq().
- arch/mips/kernel/ipipe.c : created __ipipe_do_IRQ() (used with
__ipipe_enable_pipeline()) which calls generic_handle_irq().
Do you think my approach is correct ? Are my few changes relevant ? I
avoided to modify entry.S and genex.S, is this mandatory ?
Any help or thoughts are welcome !
Regards,
--
Florent Audebert
[-- Attachment #2: adeos-mips-qemu.txt --]
[-- Type: text/plain, Size: 3612 bytes --]
(qemu) Linux version 2.6.24.4-g6c207edf-dirty (florent@coincoin) (gcc version 3.4.5) #110 Mon May 5 14:14:43 CEST 2008
LINUX started...
console [early0] enabled
CPU revision is: 00018000 (MIPS 4Kc)
registering PCI controller with io_map_base unset
Determined physical RAM map:
memory: 00001000 @ 00000000 (reserved)
memory: 000ef000 @ 00001000 (ROM data)
memory: 00525000 @ 000f0000 (reserved)
memory: 079ea000 @ 00615000 (usable)
Wasting 49824 bytes for tracking 1557 unused pages
Initrd not found or empty - disabling initrd
Zone PFN ranges:
DMA 0 -> 4096
Normal 4096 -> 32767
Movable zone start PFN for each node
early_node_map[1] active PFN ranges
0: 0 -> 32767
Built 1 zonelists in Zone order, mobility grouping on. Total pages: 32512
Kernel command line: root=/dev/hda1 console=ttyS0
Primary instruction cache 2kB, VIPT, 2-way, linesize 16 bytes.
Primary data cache 2kB, 2-way, VIPT, no aliases, linesize 16 bytes
Synthesized clear page handler (26 instructions).
Synthesized copy page handler (46 instructions).
Synthesized TLB refill handler (20 instructions).
Synthesized TLB load handler fastpath (32 instructions).
Synthesized TLB store handler fastpath (32 instructions).
Synthesized TLB modify handler fastpath (31 instructions).
PID hash table entries: 512 (order: 9, 2048 bytes)
CPU frequency 200.00 MHz
I-pipe 1.9-01: pipeline enabled.
Console: colour dummy device 80x25
Dentry cache hash table entries: 16384 (order: 4, 65536 bytes)
Inode-cache hash table entries: 8192 (order: 3, 32768 bytes)
CPU 0 Unable to handle kernel paging request at virtual address 00000098, epc == 80185bc4, ra == 80185b98
Oops[#1]:
Cpu 0
$ 0 : 00000000 10002400 00000000 805fc250
$ 4 : 805a0000 00000000 00000000 00000000
$ 8 : 00000000 00000000 00000000 016e3b77
$12 : 8046dbe8 00000000 00000000 80470018
$16 : 80600000 00000000 00000000 00000000
$20 : 00000000 00000000 00000000 00000000
$24 : 00000000 00000000
$28 : 8059a000 8059bb78 00000000 80185b98
Hi : 00000000
Lo : 00000000
epc : 80185bc4 tick_periodic+0x144/0x170 Not tainted
ra : 80185b98 tick_periodic+0x118/0x170
Status: 10002402 KERNEL EXL
Cause : 00808008
BadVA : 00000098
PrId : 00018000 (MIPS 4Kc)
Modules linked in:
Process swapper (pid: 0, threadinfo=8059a000, task=8059cfc8)
Stack : 00000001 805fd3c0 00000000 805a6d60 00000000 00000000 00000000 01312d00
00000000 003d0900 00000000 016e3600 00000000 016e3600 80600000 805fd3c0
80185c0c 805a80c4 00000000 805fd440 00000001 00000000 8014f104 00000001
00000000 00000000 00000006 00000003 8059cfc8 8059cfc8 801024e8 00000000
00000000 8059bc90 8019d8c8 00000000 00000000 00000000 8011439c 00000000
...
Call Trace:
[<80185bc4>] tick_periodic+0x144/0x170
[<80185c0c>] tick_handle_periodic+0x1c/0x224
[<8011439c>] timer_interrupt+0x2c/0x40
[<801984b0>] handle_IRQ_event+0x4c/0x144
[<8019b2b8>] handle_level_irq+0xcc/0x15c
[<801147a8>] test_do_IRQ+0x68/0x94
[<801a0c30>] __ipipe_sync_stage+0x6fc/0xb48
[<8019d7fc>] __ipipe_unstall_root+0xf0/0x11c
[<8019d8c8>] __ipipe_restore_root+0xa0/0xac
[<801af410>] __free_pages_ok+0x50c/0x518
[<801b2608>] __free_pages+0xb0/0xbc
[<805d6570>] __free_pages_bootmem+0x278/0x288
[<805d5c74>] free_all_bootmem_core+0x238/0x3ec
[<805d5fec>] free_all_bootmem+0x18/0x28
[<805c4650>] mem_init+0x48/0x298
[<805b5ee8>] start_kernel+0x340/0x438
[<8051dc20>] kernel_entry+0x0/0x90
Code: 8f820040 afa20010 8fa20010 <8c420098> 30420018 38420010 2c420001 0c055f12 00402021
Kernel panic - not syncing: Fatal exception in interrupt
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Adeos-main] Advice regarding MIPS port
2008-05-05 16:43 [Adeos-main] Advice regarding MIPS port Florent Audebert
@ 2008-05-06 19:54 ` Jan Kiszka
2008-06-19 16:00 ` Florent Audebert
0 siblings, 1 reply; 6+ messages in thread
From: Jan Kiszka @ 2008-05-06 19:54 UTC (permalink / raw)
To: Florent Audebert; +Cc: adeos-main
[-- Attachment #1: Type: text/plain, Size: 2117 bytes --]
Florent Audebert wrote:
> Hi,
>
> I am starting to work on a Xenomai port for MIPS architecture as part
Cool!
> of an internship at Open Wide (France). It is my first real diving
> into the kernel sources so I'm far from being an expert. The goal of
> this project is to be able to run it on a Broadcom BCM6448 board.
>
> My first objective is to be able to run an Adeos patched kernel. I
> have decided to work with QEMU (mips-malta board emulation) instead of
> my real target since it provides me an easy way to use GDB and thus to
> understand a little bit more what is going on. Furthermore, it
> provides an easy way of testing it for other people since it only
> requires QEMU and no specific real hardware.
Yes, a reasonable approach for arch bring-up. Specifically when using
low-end targets, QEMU is able to provide faster debug round-trip times
than real hw.
>
> I read as much as documents I can (Life with Adeos, Porting Adeos,
> ...) and finally started messing with the code. I managed to compile a
> MIPS kernel merging common Adeos code from latest x86 patch and
> adapting (see below) / commenting out. I guess my kernel is far from
> functional. My current issue seems to be related to the timer (see
> attached file).
>
> Here is a list of explicit changes I have done so far :
>
> - include/asm-mips/irqflags.h : added raw_local_*() and local_irq_*().
> - include/asm-mips/irq.h : do_IRQ() calls __ipipe_handle_irq()
> instead of generic_handle_irq().
> - arch/mips/kernel/ipipe.c : created __ipipe_do_IRQ() (used with
> __ipipe_enable_pipeline()) which calls generic_handle_irq().
>
> Do you think my approach is correct ? Are my few changes relevant ? I
> avoided to modify entry.S and genex.S, is this mandatory ?
I would be fairly surprised if you get away without touching them on
MIPS. Maybe you should study changes to those files (or comparable ones)
on other archs first - offline, but you may use QEMU here as well. Try
to understand some typical code paths from the hardware to the delivery
to some domain.
Jan
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 254 bytes --]
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Adeos-main] Advice regarding MIPS port
2008-05-06 19:54 ` Jan Kiszka
@ 2008-06-19 16:00 ` Florent Audebert
2008-06-20 10:15 ` Gilles Chanteperdrix
0 siblings, 1 reply; 6+ messages in thread
From: Florent Audebert @ 2008-06-19 16:00 UTC (permalink / raw)
To: Jan Kiszka; +Cc: adeos-main
Hi,
I'm back on my work concerning Adeos port on MIPS. I'm a little bit lost atm.
My kernel is stuck in calibrate_delay() loop. Timer interrupts seem to
occur well but my i-pipe is stalled. There is no arch-dependent places
where I explicitely unstall root domain. I saw fast stall / unstall
calls in other archs entry.S. Are these calls mandatory or are they
here in order to reduce latency ?
In case the previous asumption is false, is there any other potential
leads ? (Something wrong in my local_irq_* ?)
I didn't post my code so far since nothing is working yet... If you
think it could worth it, I'll be glad to send you a git-diff from my
2.6.24 working directory.
--
Florent Audebert
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Adeos-main] Advice regarding MIPS port
2008-06-19 16:00 ` Florent Audebert
@ 2008-06-20 10:15 ` Gilles Chanteperdrix
2008-07-17 15:12 ` Florent Audebert
0 siblings, 1 reply; 6+ messages in thread
From: Gilles Chanteperdrix @ 2008-06-20 10:15 UTC (permalink / raw)
To: Florent Audebert; +Cc: adeos-main
Florent Audebert wrote:
> Hi,
>
> I'm back on my work concerning Adeos port on MIPS. I'm a little bit lost atm.
>
> My kernel is stuck in calibrate_delay() loop. Timer interrupts seem to
> occur well but my i-pipe is stalled. There is no arch-dependent places
> where I explicitely unstall root domain.
The root domain should be unstalled in the start_kernel function, file
init/main.c by the call to local_irq_enable(), which should result in an
ipipe_unstall_root().
> I saw fast stall / unstall
> calls in other archs entry.S. Are these calls mandatory or are they
> here in order to reduce latency ?
They are mandatory in a final I-pipe patch, but you can ignore them for
a first attempt, without them, you will get big latencies, that is all.
>
> In case the previous asumption is false, is there any other potential
> leads ? (Something wrong in my local_irq_* ?)
Well, when debugging, every ugly hack is permitted, so you can put as
much printks as you want in the interrupt handler or the calibrate delay
loop.
>
> I didn't post my code so far since nothing is working yet... If you
> think it could worth it, I'll be glad to send you a git-diff from my
> 2.6.24 working directory.
>
--
Gilles Chanteperdrix
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Adeos-main] Advice regarding MIPS port
2008-06-20 10:15 ` Gilles Chanteperdrix
@ 2008-07-17 15:12 ` Florent Audebert
2008-07-23 12:55 ` Florent Audebert
0 siblings, 1 reply; 6+ messages in thread
From: Florent Audebert @ 2008-07-17 15:12 UTC (permalink / raw)
To: Gilles Chanteperdrix; +Cc: adeos-main
Hi,
I'm still on my the way with the MIPS port. My kernel is now booting
and I'm able to register a domain. I was successful catching various
interrupts (only printk'ing in the handler) using
ipipe_virtualize_irq().
Next, I would like to test syscalls grabbing. I tried registering an
handler with something like this in my test module :
ipipe_catch_event(ipipe_root_domain,
IPIPE_EVENT_SYSCALL,
(ipipe_event_handler_t)adeos_sys_handler);
Before testing on MIPS, I'm trying my module on X86 arch and it does
not work. I searched in Xenomai sources and didn't spot anything
specific when using ipipe_catch_event(). I'm obviously missing
something... What is the common way of using ipipe_catch_event() ?
This issue aside, do you have any advice concerning other tests cases ?
Regards,
--
Florent Audebert
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Adeos-main] Advice regarding MIPS port
2008-07-17 15:12 ` Florent Audebert
@ 2008-07-23 12:55 ` Florent Audebert
0 siblings, 0 replies; 6+ messages in thread
From: Florent Audebert @ 2008-07-23 12:55 UTC (permalink / raw)
To: Gilles Chanteperdrix; +Cc: adeos-main
[-- Attachment #1: Type: text/plain, Size: 640 bytes --]
Hi,
The attached patch is for a 2.6.24.4 vanilla kernel (2.6.24 with few
hunks). It boots on Qemu 0.9.1 MIPS Malta emulation with I-pipe
enabled.
You can test it using provided kernel config and creating a Debian
MIPS disk image[1]. Here is the command line used in order to start
Qemu :
qemu-system-mips -M 'malta' -cpu '4Kc' -kernel vmlinux \
-hda hda.img -append 'root=/dev/hda1 console=ttyS0' \
-net tap -net nic -nographic
As said in my previous message, it really needs more testing. Any help
or advice is welcome.
Regards,
[1] http://www.aurel32.net/info/debian_mips_qemu.php
--
Florent Audebert
[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: adeos-ipipe-2.6.24-mips-1.0.patch --]
[-- Type: text/x-diff; name=adeos-ipipe-2.6.24-mips-1.0.patch, Size: 205607 bytes --]
diff --git a/arch/mips/Kconfig b/arch/mips/Kconfig
index b22c043..69ea020 100644
--- a/arch/mips/Kconfig
+++ b/arch/mips/Kconfig
@@ -1843,6 +1843,8 @@ config HZ
source "kernel/Kconfig.preempt"
+source "kernel/ipipe/Kconfig"
+
config MIPS_INSANE_LARGE
bool "Support for large 64-bit configurations"
depends on CPU_R10000 && 64BIT
diff --git a/arch/mips/kernel/Makefile b/arch/mips/kernel/Makefile
index ffa0836..1a46160 100644
--- a/arch/mips/kernel/Makefile
+++ b/arch/mips/kernel/Makefile
@@ -81,6 +81,9 @@ obj-$(CONFIG_PCSPEAKER) += pcspeaker.o
obj-$(CONFIG_KEXEC) += machine_kexec.o relocate_kernel.o
obj-$(CONFIG_EARLY_PRINTK) += early_printk.o
+obj-$(CONFIG_IPIPE) += ipipe.o
+obj-$(CONFIG_IPIPE_TRACE_MCOUNT) += ipipe-mcount.o
+
CFLAGS_cpu-bugs64.o = $(shell if $(CC) $(KBUILD_CFLAGS) -Wa,-mdaddi -c -o /dev/null -xc /dev/null >/dev/null 2>&1; then echo "-DHAVE_AS_SET_DADDI"; fi)
obj-$(CONFIG_HAVE_STD_PC_SERIAL_PORT) += 8250-platform.o
diff --git a/arch/mips/kernel/i8253.c b/arch/mips/kernel/i8253.c
index c2d497c..775e266 100644
--- a/arch/mips/kernel/i8253.c
+++ b/arch/mips/kernel/i8253.c
@@ -141,6 +141,12 @@ static cycle_t pit_read(void)
static int old_count;
static u32 old_jifs;
+#ifdef CONFIG_IPIPE
+ if (!__ipipe_pipeline_head_p(ipipe_root_domain))
+ /* We don't really own the PIT. */
+ return (cycle_t)(jiffies * LATCH) + (LATCH - 1) - old_count;
+#endif /* CONFIG_IPIPE */
+
spin_lock_irqsave(&i8253_lock, flags);
/*
* Although our caller may have the read side of xtime_lock,
diff --git a/arch/mips/kernel/i8259.c b/arch/mips/kernel/i8259.c
index 9158dd8..df97de1 100644
--- a/arch/mips/kernel/i8259.c
+++ b/arch/mips/kernel/i8259.c
@@ -29,7 +29,7 @@
*/
static int i8259A_auto_eoi = -1;
-DEFINE_SPINLOCK(i8259A_lock);
+IPIPE_DEFINE_SPINLOCK(i8259A_lock);
static void disable_8259A_irq(unsigned int irq);
static void enable_8259A_irq(unsigned int irq);
static void mask_and_ack_8259A(unsigned int irq);
@@ -66,6 +66,7 @@ static void disable_8259A_irq(unsigned int irq)
irq -= I8259A_IRQ_BASE;
mask = 1 << irq;
spin_lock_irqsave(&i8259A_lock, flags);
+ ipipe_irq_lock(irq);
cached_irq_mask |= mask;
if (irq & 8)
outb(cached_slave_mask, PIC_SLAVE_IMR);
@@ -80,13 +81,16 @@ static void enable_8259A_irq(unsigned int irq)
unsigned long flags;
irq -= I8259A_IRQ_BASE;
- mask = ~(1 << irq);
+ mask = (1 << irq);
spin_lock_irqsave(&i8259A_lock, flags);
- cached_irq_mask &= mask;
- if (irq & 8)
- outb(cached_slave_mask, PIC_SLAVE_IMR);
- else
- outb(cached_master_mask, PIC_MASTER_IMR);
+ if (cached_irq_mask & mask) {
+ cached_irq_mask &= ~mask;
+ if (irq & 8)
+ outb(cached_slave_mask, PIC_SLAVE_IMR);
+ else
+ outb(cached_master_mask, PIC_MASTER_IMR);
+ ipipe_irq_unlock(irq);
+ }
spin_unlock_irqrestore(&i8259A_lock, flags);
}
@@ -169,9 +173,25 @@ static void mask_and_ack_8259A(unsigned int irq)
*/
if (cached_irq_mask & irqmask)
goto spurious_8259A_irq;
+#ifdef CONFIG_IPIPE
+ if (irq == 0) {
+ /*
+ * Fast timer ack -- don't mask (unless supposedly
+ * spurious). We trace outb's in order to detect
+ * broken hardware inducing large delays.
+ */
+ ipipe_trace_special('o', irq);
+ outb(0x60, PIC_MASTER_CMD); /* Specific EOI to master. */
+ ipipe_trace_special('x', irq);
+ spin_unlock_irqrestore(&i8259A_lock, flags);
+ return;
+ }
+#endif /* CONFIG_IPIPE */
+
cached_irq_mask |= irqmask;
handle_real_irq:
+ ipipe_trace_special('o', irq);
if (irq & 8) {
inb(PIC_SLAVE_IMR); /* DUMMY - (do we need this?) */
outb(cached_slave_mask, PIC_SLAVE_IMR);
@@ -183,6 +203,7 @@ handle_real_irq:
outb(0x60+irq, PIC_MASTER_CMD); /* 'Specific EOI to master */
}
smtc_im_ack_irq(irq);
+ ipipe_trace_special('x', irq);
spin_unlock_irqrestore(&i8259A_lock, flags);
return;
diff --git a/arch/mips/kernel/ipipe.c b/arch/mips/kernel/ipipe.c
new file mode 100644
index 0000000..c1e4c3e
--- /dev/null
+++ b/arch/mips/kernel/ipipe.c
@@ -0,0 +1,364 @@
+/* -*- linux-c -*-
+ * linux/arch/mips/kernel/ipipe.c
+ *
+ * Copyright (C) 2002-2005 Philippe Gerum.
+ * Copyright (C) 2004 Wolfgang Grandegger (Adeos/arm port over 2.4).
+ * Copyright (C) 2005 Heikki Lindholm (PowerPC 970 fixes).
+ * Copyright (C) 2005 Stelian Pop.
+ * Copyright (C) 2006-2008 Gilles Chanteperdrix.
+ *
+ * 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, Inc., 675 Mass Ave, Cambridge MA 02139,
+ * USA; 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.
+ *
+ * Architecture-dependent I-PIPE support for MIPS.
+ */
+
+#include <linux/kernel.h>
+#include <linux/smp.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/bitops.h>
+#include <linux/interrupt.h>
+#include <linux/module.h>
+#include <linux/errno.h>
+#include <asm/unistd.h>
+
+#ifdef CONFIG_SMP
+#error "No SMP support"
+#endif
+
+int __ipipe_tick_irq = 0;
+
+/* Next tick date (timebase value). */
+DEFINE_PER_CPU(struct pt_regs, __ipipe_tick_regs);
+
+extern struct irq_desc irq_desc[];
+
+int ipipe_get_sysinfo(struct ipipe_sysinfo *info)
+{
+ info->ncpus = num_online_cpus();
+ info->cpufreq = ipipe_cpu_freq();
+ info->archdep.tmirq = __ipipe_tick_irq;
+ info->archdep.tmfreq = info->cpufreq;
+
+ return 0;
+}
+
+static void __ipipe_do_IRQ(unsigned int irq, struct pt_regs *regs)
+{
+ struct pt_regs *old_regs = get_irq_regs();
+
+ /* No set_irq_regs() on MIPS (?) */
+ current_thread_info()->regs = regs;
+
+ generic_handle_irq(irq);
+
+ current_thread_info()->regs = old_regs;
+}
+
+static int __ipipe_ack_irq(unsigned irq)
+{
+ irq_desc_t *desc = irq_desc + irq;
+ desc->ipipe_ack(irq, desc);
+ return 1;
+}
+
+void __ipipe_enable_irqdesc(struct ipipe_domain *ipd, unsigned irq)
+{
+ irq_desc[irq].status &= ~IRQ_DISABLED;
+}
+
+/*
+ * __ipipe_enable_pipeline() -- We are running on the boot CPU, hw
+ * interrupts are off, and secondary CPUs are still lost in space.
+ */
+void __ipipe_enable_pipeline(void)
+{
+ unsigned irq;
+
+ /* First, virtualize all interrupts from the root domain. */
+
+ for (irq = 0; irq < NR_IRQS; irq++)
+ ipipe_virtualize_irq(ipipe_root_domain,
+ irq,
+ (ipipe_irq_handler_t)&__ipipe_do_IRQ,
+ NULL,
+ &__ipipe_ack_irq,
+ IPIPE_STDROOT_MASK);
+}
+
+/*
+ * ipipe_trigger_irq() -- Push the interrupt at front of the pipeline
+ * just like if it has been actually received from a hw source. Also
+ * works for virtual interrupts.
+ */
+int ipipe_trigger_irq(unsigned irq)
+{
+ unsigned long flags;
+
+ if (irq >= IPIPE_NR_IRQS ||
+ (ipipe_virtual_irq_p(irq)
+ && !test_bit(irq - IPIPE_VIRQ_BASE, &__ipipe_virtual_irq_map)))
+ return -EINVAL;
+
+ local_irq_save_hw(flags);
+ __ipipe_handle_irq(irq, get_irq_regs());
+ local_irq_restore_hw(flags);
+
+ return 1;
+}
+
+/*
+ * ipipe_critical_enter() -- Grab the superlock excluding all CPUs
+ * but the current one from a critical section. This lock is used when
+ * we must enforce a global critical section for a single CPU in a
+ * possibly SMP system whichever context the CPUs are running.
+ */
+unsigned long ipipe_critical_enter(void (*syncfn) (void))
+{
+ unsigned long flags;
+
+ local_irq_save_hw(flags);
+
+#ifdef CONFIG_SMP
+ if (num_online_cpus() > 1) { /* We might be running a SMP-kernel on a UP box... */
+ int cpu = ipipe_processor_id();
+ cpumask_t lock_map;
+
+ if (!cpu_test_and_set(cpu, __ipipe_cpu_lock_map)) {
+ while (cpu_test_and_set(BITS_PER_LONG - 1,
+ __ipipe_cpu_lock_map)) {
+ int n = 0;
+ do {
+ cpu_relax();
+ } while (++n < cpu);
+ }
+
+ spin_lock(&__ipipe_cpu_barrier);
+
+ __ipipe_cpu_sync = syncfn;
+
+ /* Send the sync IPI to all processors but the current one. */
+ send_IPI_allbutself(IPIPE_CRITICAL_VECTOR);
+
+ cpus_andnot(lock_map, cpu_online_map,
+ __ipipe_cpu_lock_map);
+
+ while (!cpus_equal(__ipipe_cpu_sync_map, lock_map))
+ cpu_relax();
+ }
+
+ atomic_inc(&__ipipe_critical_count);
+ }
+#endif /* CONFIG_SMP */
+
+ return flags;
+}
+
+/* ipipe_critical_exit() -- Release the superlock. */
+void ipipe_critical_exit(unsigned long flags)
+{
+#ifdef CONFIG_SMP
+ if (num_online_cpus() > 1) { /* We might be running a SMP-kernel on a UP box... */
+ if (atomic_dec_and_test(&__ipipe_critical_count)) {
+ spin_unlock(&__ipipe_cpu_barrier);
+
+ while (!cpus_empty(__ipipe_cpu_sync_map))
+ cpu_relax();
+
+ cpu_clear(ipipe_processor_id(), __ipipe_cpu_lock_map);
+ cpu_clear(BITS_PER_LONG - 1, __ipipe_cpu_lock_map);
+ }
+ }
+#endif /* CONFIG_SMP */
+
+ local_irq_restore_hw(flags);
+}
+
+asmlinkage void __ipipe_fast_stall_root(void)
+{
+ __set_bit(IPIPE_STALL_FLAG, &__ipipe_root_status);
+
+#if 0
+ // __asm__ __volatile__ (
+ // "ori %0, 1 \n"
+ // : "=r" (__ipipe_root_status)
+ // : "0" (__ipipe_root_status)
+ // : "memory");
+#endif
+}
+
+asmlinkage void __ipipe_fast_unstall_root(void)
+{
+ __clear_bit(IPIPE_STALL_FLAG, &__ipipe_root_status);
+}
+
+/*
+ * __ipipe_handle_irq() -- IPIPE's generic IRQ handler. An optimistic
+ * interrupt protection log is maintained here for each domain. Hw
+ * interrupts are off on entry.
+ */
+int __ipipe_handle_irq(unsigned int irq, struct pt_regs *regs)
+{
+ struct ipipe_domain *this_domain, *next_domain;
+ struct list_head *head, *pos;
+ int m_ack;
+
+ m_ack = (regs == NULL);
+
+ if (irq >= IPIPE_NR_IRQS) {
+ printk(KERN_ERR "I-pipe: spurious interrupt %d\n", irq);
+ goto finalize_nosync;
+ }
+
+ head = __ipipe_pipeline.next;
+ next_domain = list_entry(head, struct ipipe_domain, p_link);
+ if (likely(test_bit(IPIPE_WIRED_FLAG, &next_domain->irqs[irq].control))) {
+ if (!m_ack && next_domain->irqs[irq].acknowledge != NULL)
+ next_domain->irqs[irq].acknowledge(irq);
+ if (likely(__ipipe_dispatch_wired(next_domain, irq))) {
+ goto finalize;
+ } else
+ goto finalize_nosync;
+ }
+
+ this_domain = ipipe_current_domain;
+
+ if (test_bit(IPIPE_STICKY_FLAG, &this_domain->irqs[irq].control))
+ head = &this_domain->p_link;
+
+ /* Ack the interrupt. */
+
+ pos = head;
+
+ while (pos != &__ipipe_pipeline) {
+ next_domain = list_entry(pos, struct ipipe_domain, p_link);
+
+ /*
+ * For each domain handling the incoming IRQ, mark it
+ * as pending in its log.
+ */
+ if (test_bit(IPIPE_HANDLE_FLAG, &next_domain->irqs[irq].control)) {
+ /*
+ * Domains that handle this IRQ are polled for
+ * acknowledging it by decreasing priority
+ * order. The interrupt must be made pending
+ * _first_ in the domain's status flags before
+ * the PIC is unlocked.
+ */
+ __ipipe_set_irq_pending(next_domain, irq);
+
+ if (!m_ack && next_domain->irqs[irq].acknowledge != NULL)
+ m_ack = next_domain->irqs[irq].acknowledge(irq);
+ }
+
+ /*
+ * If the domain does not want the IRQ to be passed
+ * down the interrupt pipe, exit the loop now.
+ */
+
+ if (!test_bit(IPIPE_PASS_FLAG, &next_domain->irqs[irq].control))
+ break;
+
+ pos = next_domain->p_link.next;
+ }
+
+finalize:
+
+ /* Given our deferred dispatching model for regular IRQs, we
+ * only record CPU regs for the last timer interrupt, so that
+ * the timer handler charges CPU times properly. It is assumed
+ * that other interrupt handlers don't actually care for such
+ * information. */
+
+ /* FIXME: Too much things here. */
+ if (irq == __ipipe_tick_irq) {
+ int i;
+ struct pt_regs *tick_regs = &__raw_get_cpu_var(__ipipe_tick_regs);
+
+ for (i = 0; i < 32; ++i)
+ tick_regs->regs[i] = regs->regs[i];
+
+ tick_regs->cp0_status = regs->cp0_badvaddr;
+ tick_regs->cp0_cause = regs->cp0_cause;
+ tick_regs->cp0_epc = regs->cp0_epc;
+ }
+
+ /*
+ * Now walk the pipeline, yielding control to the highest
+ * priority domain that has pending interrupt(s) or
+ * immediately to the current domain if the interrupt has been
+ * marked as 'sticky'. This search does not go beyond the
+ * current domain in the pipeline.
+ */
+
+ __ipipe_walk_pipeline(head);
+
+finalize_nosync:
+ if (!ipipe_root_domain_p || __ipipe_test_root())
+ return 0;
+
+#ifdef CONFIG_SMP
+ /*
+ * Prevent a spurious rescheduling from being triggered on
+ * preemptible kernels along the way out through
+ * ret_from_intr.
+ */
+ if (!regs)
+ __set_bit(IPIPE_STALL_FLAG, &ipipe_root_cpudom_var(status));
+#endif /* CONFIG_SMP */
+
+ return 1;
+}
+
+asmlinkage int __ipipe_syscall_root(struct pt_regs regs) /* HW interrupts off */
+{
+ unsigned long flags;
+
+ /*
+ * This routine either returns:
+ * 0 -- if the syscall is to be passed to Linux;
+ * >0 -- if the syscall should not be passed to Linux, and no
+ * tail work should be performed;
+ * <0 -- if the syscall should not be passed to Linux but the
+ * tail work has to be performed (for handling signals etc).
+ */
+
+ /* Syscall number is stored in v0 / $2 */
+ if (__ipipe_syscall_watched_p(current, regs.regs[2]) &&
+ __ipipe_event_monitored_p(IPIPE_EVENT_SYSCALL) &&
+ __ipipe_dispatch_event(IPIPE_EVENT_SYSCALL,
+ ®s) > 0) {
+ if (ipipe_root_domain_p && !in_atomic()) {
+ /*
+ * Sync pending VIRQs before _TIF_NEED_RESCHED
+ * is tested.
+ */
+ local_irq_save_hw(flags);
+ if ((ipipe_root_cpudom_var(irqpend_himask) &
+ IPIPE_IRQMASK_VIRT) != 0)
+ __ipipe_sync_pipeline(IPIPE_IRQMASK_VIRT);
+ local_irq_restore_hw(flags);
+ return -1;
+ }
+ return 1;
+ }
+ return 0;
+}
+
+int __ipipe_check_tickdev(const char *devname)
+{
+ return 1;
+}
diff --git a/arch/mips/kernel/scall32-o32.S b/arch/mips/kernel/scall32-o32.S
index 82480a1..97177f5 100644
--- a/arch/mips/kernel/scall32-o32.S
+++ b/arch/mips/kernel/scall32-o32.S
@@ -55,6 +55,20 @@ NESTED(handle_sys, PT_SIZE, sp)
bgez t3, stackargs
stack_done:
+#ifdef CONFIG_IPIPE
+ LONG_S v0, PT_R2(sp) # Backup v0 (contains syscall number)
+
+ jal __ipipe_syscall_root # Return value stored in v0
+
+ # >0 -- Syscall _not_ to be passed to Linux and no tail work
+ bgtz v0, o32_syscall_exit_work
+ # <0 -- Syscall to be passed to Linux but tail work
+ bltz v0, o32_syscall_exit
+ # =0 -- Syscall is to be passed to Linux
+
+ # Restore v0
+ LONG_L v0, PT_R2(sp)
+#endif /* CONFIG_IPIPE */
lw t0, TI_FLAGS($28) # syscall tracing enabled?
li t1, _TIF_SYSCALL_TRACE | _TIF_SYSCALL_AUDIT
and t0, t1
diff --git a/drivers/pci/htirq.c b/drivers/pci/htirq.c
index 279c940..daabf90 100644
--- a/drivers/pci/htirq.c
+++ b/drivers/pci/htirq.c
@@ -21,7 +21,7 @@
* With multiple simultaneous hypertransport irq devices it might pay
* to make this more fine grained. But start with simple, stupid, and correct.
*/
-static DEFINE_SPINLOCK(ht_irq_lock);
+static IPIPE_DEFINE_SPINLOCK(ht_irq_lock);
struct ht_irq_cfg {
struct pci_dev *dev;
diff --git a/drivers/serial/8250.c b/drivers/serial/8250.c
index f94109c..4634af5 100644
--- a/drivers/serial/8250.c
+++ b/drivers/serial/8250.c
@@ -2546,6 +2546,51 @@ static int serial8250_console_early_setup(void)
return serial8250_find_port_for_earlycon();
}
+#ifdef CONFIG_IPIPE_DEBUG
+
+#include <stdarg.h>
+
+void __ipipe_serial_debug(const char *fmt, ...)
+{
+ struct uart_8250_port *up = &serial8250_ports[0];
+ unsigned int ier, count;
+ unsigned long flags;
+ char buf[128];
+ va_list ap;
+
+ va_start(ap, fmt);
+ vsprintf(buf, fmt, ap);
+ va_end(ap);
+ count = strlen(buf);
+
+ touch_nmi_watchdog();
+
+ local_irq_save_hw(flags);
+
+ /*
+ * First save the IER then disable the interrupts
+ */
+ ier = serial_in(up, UART_IER);
+
+ if (up->capabilities & UART_CAP_UUE)
+ serial_out(up, UART_IER, UART_IER_UUE);
+ else
+ serial_out(up, UART_IER, 0);
+
+ uart_console_write(&up->port, buf, count, serial8250_console_putchar);
+
+ /*
+ * Finally, wait for transmitter to become empty
+ * and restore the IER
+ */
+ wait_for_xmitr(up, BOTH_EMPTY);
+ serial_out(up, UART_IER, ier);
+
+ local_irq_restore_hw(flags);
+}
+
+#endif
+
static struct uart_driver serial8250_reg;
static struct console serial8250_console = {
.name = "ttyS",
diff --git a/include/asm-mips/atomic.h b/include/asm-mips/atomic.h
index a798d62..053d5f6 100644
--- a/include/asm-mips/atomic.h
+++ b/include/asm-mips/atomic.h
@@ -80,9 +80,9 @@ static __inline__ void atomic_add(int i, atomic_t * v)
} else {
unsigned long flags;
- raw_local_irq_save(flags);
+ local_irq_save_hw(flags);
v->counter += i;
- raw_local_irq_restore(flags);
+ local_irq_restore_hw(flags);
}
}
@@ -125,9 +125,9 @@ static __inline__ void atomic_sub(int i, atomic_t * v)
} else {
unsigned long flags;
- raw_local_irq_save(flags);
+ local_irq_save_hw(flags);
v->counter -= i;
- raw_local_irq_restore(flags);
+ local_irq_restore_hw(flags);
}
}
@@ -174,11 +174,11 @@ static __inline__ int atomic_add_return(int i, atomic_t * v)
} else {
unsigned long flags;
- raw_local_irq_save(flags);
+ local_irq_save_hw(flags);
result = v->counter;
result += i;
v->counter = result;
- raw_local_irq_restore(flags);
+ local_irq_restore_hw(flags);
}
smp_llsc_mb();
@@ -226,11 +226,11 @@ static __inline__ int atomic_sub_return(int i, atomic_t * v)
} else {
unsigned long flags;
- raw_local_irq_save(flags);
+ local_irq_save_hw(flags);
result = v->counter;
result -= i;
v->counter = result;
- raw_local_irq_restore(flags);
+ local_irq_restore_hw(flags);
}
smp_llsc_mb();
@@ -294,12 +294,12 @@ static __inline__ int atomic_sub_if_positive(int i, atomic_t * v)
} else {
unsigned long flags;
- raw_local_irq_save(flags);
+ local_irq_save_hw(flags);
result = v->counter;
result -= i;
if (result >= 0)
v->counter = result;
- raw_local_irq_restore(flags);
+ local_irq_restore_hw(flags);
}
smp_llsc_mb();
@@ -461,9 +461,9 @@ static __inline__ void atomic64_add(long i, atomic64_t * v)
} else {
unsigned long flags;
- raw_local_irq_save(flags);
+ local_irq_save_hw(flags);
v->counter += i;
- raw_local_irq_restore(flags);
+ local_irq_restore_hw(flags);
}
}
@@ -506,9 +506,9 @@ static __inline__ void atomic64_sub(long i, atomic64_t * v)
} else {
unsigned long flags;
- raw_local_irq_save(flags);
+ local_irq_save_hw(flags);
v->counter -= i;
- raw_local_irq_restore(flags);
+ local_irq_restore_hw(flags);
}
}
@@ -555,11 +555,11 @@ static __inline__ long atomic64_add_return(long i, atomic64_t * v)
} else {
unsigned long flags;
- raw_local_irq_save(flags);
+ local_irq_save_hw(flags);
result = v->counter;
result += i;
v->counter = result;
- raw_local_irq_restore(flags);
+ local_irq_restore_hw(flags);
}
smp_llsc_mb();
@@ -607,11 +607,11 @@ static __inline__ long atomic64_sub_return(long i, atomic64_t * v)
} else {
unsigned long flags;
- raw_local_irq_save(flags);
+ local_irq_save_hw(flags);
result = v->counter;
result -= i;
v->counter = result;
- raw_local_irq_restore(flags);
+ local_irq_restore_hw(flags);
}
smp_llsc_mb();
@@ -675,12 +675,12 @@ static __inline__ long atomic64_sub_if_positive(long i, atomic64_t * v)
} else {
unsigned long flags;
- raw_local_irq_save(flags);
+ local_irq_save_hw(flags);
result = v->counter;
result -= i;
if (result >= 0)
v->counter = result;
- raw_local_irq_restore(flags);
+ local_irq_restore_hw(flags);
}
smp_llsc_mb();
diff --git a/include/asm-mips/bitops.h b/include/asm-mips/bitops.h
index ec75ce4..cfe7690 100644
--- a/include/asm-mips/bitops.h
+++ b/include/asm-mips/bitops.h
@@ -104,9 +104,9 @@ static inline void set_bit(unsigned long nr, volatile unsigned long *addr)
a += nr >> SZLONG_LOG;
mask = 1UL << bit;
- raw_local_irq_save(flags);
+ local_irq_save_hw(flags);
*a |= mask;
- raw_local_irq_restore(flags);
+ local_irq_restore_hw(flags);
}
}
@@ -169,9 +169,9 @@ static inline void clear_bit(unsigned long nr, volatile unsigned long *addr)
a += nr >> SZLONG_LOG;
mask = 1UL << bit;
- raw_local_irq_save(flags);
+ local_irq_save_hw(flags);
*a &= ~mask;
- raw_local_irq_restore(flags);
+ local_irq_restore_hw(flags);
}
}
@@ -238,9 +238,9 @@ static inline void change_bit(unsigned long nr, volatile unsigned long *addr)
a += nr >> SZLONG_LOG;
mask = 1UL << bit;
- raw_local_irq_save(flags);
+ local_irq_save_hw(flags);
*a ^= mask;
- raw_local_irq_restore(flags);
+ local_irq_restore_hw(flags);
}
}
@@ -303,10 +303,10 @@ static inline int test_and_set_bit(unsigned long nr,
a += nr >> SZLONG_LOG;
mask = 1UL << bit;
- raw_local_irq_save(flags);
+ local_irq_save_hw(flags);
res = (mask & *a);
*a |= mask;
- raw_local_irq_restore(flags);
+ local_irq_restore_hw(flags);
}
smp_llsc_mb();
@@ -371,10 +371,10 @@ static inline int test_and_set_bit_lock(unsigned long nr,
a += nr >> SZLONG_LOG;
mask = 1UL << bit;
- raw_local_irq_save(flags);
+ local_irq_save_hw(flags);
res = (mask & *a);
*a |= mask;
- raw_local_irq_restore(flags);
+ local_irq_restore_hw(flags);
}
smp_llsc_mb();
@@ -460,10 +460,10 @@ static inline int test_and_clear_bit(unsigned long nr,
a += nr >> SZLONG_LOG;
mask = 1UL << bit;
- raw_local_irq_save(flags);
+ local_irq_save_hw(flags);
res = (mask & *a);
*a &= ~mask;
- raw_local_irq_restore(flags);
+ local_irq_restore_hw(flags);
}
smp_llsc_mb();
@@ -530,10 +530,10 @@ static inline int test_and_change_bit(unsigned long nr,
a += nr >> SZLONG_LOG;
mask = 1UL << bit;
- raw_local_irq_save(flags);
+ local_irq_save_hw(flags);
res = (mask & *a);
*a ^= mask;
- raw_local_irq_restore(flags);
+ local_irq_restore_hw(flags);
}
smp_llsc_mb();
diff --git a/include/asm-mips/cmpxchg.h b/include/asm-mips/cmpxchg.h
index a5ec0e5..d58d7ed 100644
--- a/include/asm-mips/cmpxchg.h
+++ b/include/asm-mips/cmpxchg.h
@@ -56,11 +56,11 @@
} else { \
unsigned long __flags; \
\
- raw_local_irq_save(__flags); \
+ local_irq_save_hw(__flags); \
__ret = *m; \
if (__ret == old) \
*m = new; \
- raw_local_irq_restore(__flags); \
+ local_irq_restore_hw(__flags); \
} \
\
__ret; \
diff --git a/include/asm-mips/i8259.h b/include/asm-mips/i8259.h
index 8572a2d..29fc5a5 100644
--- a/include/asm-mips/i8259.h
+++ b/include/asm-mips/i8259.h
@@ -35,7 +35,7 @@
#define SLAVE_ICW4_DEFAULT 0x01
#define PIC_ICW4_AEOI 2
-extern spinlock_t i8259A_lock;
+extern ipipe_spinlock_t i8259A_lock;
extern int i8259A_irq_pending(unsigned int irq);
extern void make_8259A_irq(unsigned int irq);
diff --git a/include/asm-mips/ipipe.h b/include/asm-mips/ipipe.h
new file mode 100644
index 0000000..5a8b042
--- /dev/null
+++ b/include/asm-mips/ipipe.h
@@ -0,0 +1,154 @@
+/* -*- linux-c -*-
+ * include/asm-mips/ipipe.h
+ *
+ * Copyright (C) 2002-2005 Philippe Gerum.
+ * Copyright (C) 2005 Stelian Pop.
+ * Copyright (C) 2006-2008 Gilles Chanteperdrix.
+ *
+ * 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, Inc., 675 Mass Ave, Cambridge MA 02139,
+ * USA; 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.
+ */
+
+#ifndef __MIPS_IPIPE_H
+#define __MIPS_IPIPE_H
+
+#ifdef CONFIG_IPIPE
+
+#include <linux/ipipe_percpu.h>
+
+#define IPIPE_ARCH_STRING "1.0-01"
+#define IPIPE_MAJOR_NUMBER 1
+#define IPIPE_MINOR_NUMBER 0
+#define IPIPE_PATCH_NUMBER 1
+
+#ifdef CONFIG_SMP
+#error "I-pipe/mips: SMP not yet implemented"
+#define ipipe_processor_id() (current_thread_info()->cpu)
+#else /* !CONFIG_SMP */
+#define ipipe_processor_id() 0
+#endif /* CONFIG_SMP */
+
+#define smp_processor_id_hw() ipipe_processor_id()
+
+#define prepare_arch_switch(next) \
+do { \
+ ipipe_schedule_notify(current, next); \
+ local_irq_disable_hw(); \
+} while(0)
+
+#define task_hijacked(p) !ipipe_root_domain_p
+
+struct ipipe_sysinfo {
+
+ int ncpus; /* Number of CPUs on board */
+ u64 cpufreq; /* CPU frequency (in Hz) */
+
+ /* Arch-dependent block */
+
+ struct {
+ unsigned tmirq; /* Timer tick IRQ */
+ u64 tmfreq; /* Timer frequency */
+ } archdep;
+};
+
+int __ipipe_check_tickdev(const char *devname);
+
+#define ipipe_read_tsc(t) do { t = __ipipe_mach_get_tsc(); } while (0)
+#define __ipipe_read_timebase() __ipipe_mach_get_tsc()
+
+/* FIXME: Good value ? */
+#define ipipe_cpu_freq() (HZ /* * CLOCK_TICK_RATE */)
+#define ipipe_tsc2ns(t) \
+({ \
+ unsigned long long delta = (t)*1000; \
+ do_div(delta, ipipe_cpu_freq() / 1000000 + 1); \
+ (unsigned long)delta; \
+})
+#define ipipe_tsc2us(t) \
+({ \
+ unsigned long long delta = (t); \
+ do_div(delta, ipipe_cpu_freq() / 1000000 + 1); \
+ (unsigned long)delta; \
+})
+
+/* Private interface -- Internal use only */
+
+#define __ipipe_check_platform() do { } while(0)
+
+#define __ipipe_init_platform() do { } while(0)
+
+#define __ipipe_enable_irq(irq) irq_desc[irq].chip->enable(irq)
+
+#define __ipipe_disable_irq(irq) irq_desc[irq].chip->disable(irq)
+
+#define __ipipe_hook_critical_ipi(ipd) do { } while(0)
+
+void __ipipe_enable_irqdesc(struct ipipe_domain *ipd, unsigned irq);
+
+void __ipipe_enable_pipeline(void);
+
+void __ipipe_do_critical_sync(unsigned irq,
+ void *cookie);
+
+DECLARE_PER_CPU(struct pt_regs, __ipipe_tick_regs);
+
+int __ipipe_handle_irq(unsigned int irq, struct pt_regs *regs);
+
+static inline unsigned long __ipipe_ffnz(unsigned long ul)
+{
+ return ffs(ul) - 1;
+}
+
+/* When running handlers, enable hw interrupts for all domains but the
+ * one heading the pipeline, so that IRQs can never be significantly
+ * deferred for the latter. */
+#define __ipipe_run_isr(ipd, irq) \
+do { \
+ local_irq_enable_nohead(ipd); \
+ if (ipd == ipipe_root_domain) { \
+ if (likely(!ipipe_virtual_irq_p(irq))) \
+ { \
+ ((void (*)(unsigned, struct pt_regs *)) \
+ ipd->irqs[irq].handler) (irq, \
+ &__raw_get_cpu_var(__ipipe_tick_regs)); \
+ } \
+ else { \
+ irq_enter(); \
+ ipd->irqs[irq].handler(irq,ipd->irqs[irq].cookie); \
+ irq_exit(); \
+ } \
+ } else { \
+ __clear_bit(IPIPE_SYNC_FLAG, &ipipe_cpudom_var(ipd, status)); \
+ ipd->irqs[irq].handler(irq,ipd->irqs[irq].cookie); \
+ __set_bit(IPIPE_SYNC_FLAG, &ipipe_cpudom_var(ipd, status)); \
+ } \
+ local_irq_disable_hw(); \
+} while(0)
+
+#define __ipipe_syscall_watched_p(p, sc) \
+ (((p)->flags & PF_EVNOTIFY) || \
+ /* FIXME: Is __NR_Linux_syscalls what we really want ? */ \
+ /* (include/asm-mips/unistd.h) */ \
+ (unsigned long)sc >= __NR_Linux_syscalls + 1)
+
+#else /* !CONFIG_IPIPE */
+
+#define task_hijacked(p) 0
+
+#define smp_processor_id_hw() smp_processor_id()
+
+#endif /* CONFIG_IPIPE */
+
+#endif /* !__MIPS_IPIPE_H */
diff --git a/include/asm-mips/ipipe_base.h b/include/asm-mips/ipipe_base.h
new file mode 100644
index 0000000..7db4695
--- /dev/null
+++ b/include/asm-mips/ipipe_base.h
@@ -0,0 +1,117 @@
+/* -*- linux-c -*-
+ * include/asm-mips/ipipe_base.h
+ *
+ * Copyright (C) 2007 Philippe Gerum.
+ *
+ * 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, Inc., 675 Mass Ave, Cambridge MA 02139,
+ * USA; 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.
+ */
+
+#ifndef __MIPS_IPIPE_BASE_H
+#define __MIPS_IPIPE_BASE_H
+
+#include <linux/threads.h>
+#include <asm/irq.h>
+
+#define IPIPE_NR_XIRQS NR_IRQS
+#define IPIPE_IRQ_ISHIFT 5
+
+/* MIPS Exceptions */
+#define IPIPE_TRAP_INT 0 /* Interrupt */
+#define IPIPE_TRAP_MOD 1 /* TLB Modification Exception */
+#define IPIPE_TRAP_TLBL 2 /* TLB Exception (load or instruction fetch) */
+#define IPIPE_TRAP_TLBS 3 /* TLB Exception (store) */
+#define IPIPE_TRAP_ADEL 4 /* Address error exception (load or instruction fetch) */
+#define IPIPE_TRAP_ADES 5 /* Address error exception (store) */
+#define IPIPE_TRAP_IBE 6 /* Bus error exception (instruction fetch) */
+#define IPIPE_TRAP_DBE 7 /* Bus error exception (data reference: load or store) */
+#define IPIPE_TRAP_SYS 8 /* Syscall exception */
+#define IPIPE_TRAP_BP 9 /* Breakpoint exception */
+#define IPIPE_TRAP_RI 10 /* Reserved instruction exception */
+#define IPIPE_TRAP_CPU 11 /* Coprocessor Unusable exception */
+#define IPIPE_TRAP_OV 12 /* Arithmetic Overflow exception */
+#define IPIPE_TRAP_TRAP 13 /* Trap exception */
+#define IPIPE_TRAP_VCEI 14 /* Virtual Coherency Exception instruction */
+#define IPIPE_TRAP_FPE 15 /* Floating-Point exception */
+
+#define IPIPE_NR_FAULTS 16
+
+
+/* Pseudo-vectors used for kernel events */
+#define IPIPE_FIRST_EVENT (IPIPE_NR_FAULTS)
+#define IPIPE_EVENT_SYSCALL (IPIPE_FIRST_EVENT)
+#define IPIPE_EVENT_SCHEDULE (IPIPE_FIRST_EVENT + 1)
+#define IPIPE_EVENT_SIGWAKE (IPIPE_FIRST_EVENT + 2)
+#define IPIPE_EVENT_SETSCHED (IPIPE_FIRST_EVENT + 3)
+#define IPIPE_EVENT_INIT (IPIPE_FIRST_EVENT + 4)
+#define IPIPE_EVENT_EXIT (IPIPE_FIRST_EVENT + 5)
+#define IPIPE_EVENT_CLEANUP (IPIPE_FIRST_EVENT + 6)
+#define IPIPE_LAST_EVENT (IPIPE_EVENT_CLEANUP)
+#define IPIPE_NR_EVENTS (IPIPE_LAST_EVENT + 1)
+
+#ifndef __ASSEMBLY__
+
+#include <asm/irqflags.h>
+
+#ifdef CONFIG_SMP
+
+#error "SMP not implemented"
+#define __ipipe_root_status ipipe_root_cpudom_var(status)
+
+#else /* !CONFIG_SMP */
+
+#if __GNUC__ >= 4
+
+/* Alias to ipipe_root_cpudom_var(status) */
+extern unsigned long __ipipe_root_status;
+
+#else
+
+extern unsigned long *const __ipipe_root_status_addr;
+#define __ipipe_root_status (*__ipipe_root_status_addr)
+
+#endif
+
+static inline void __ipipe_stall_root(void)
+{
+ unsigned long flags;
+
+ local_irq_save_hw(flags);
+ __ipipe_root_status |= 1;
+ local_irq_restore_hw(flags);
+}
+
+static inline unsigned long __ipipe_test_root(void)
+{
+ return __ipipe_root_status & 1;
+}
+
+static inline unsigned long __ipipe_test_and_stall_root(void)
+{
+ unsigned long flags, res;
+
+ local_irq_save_hw(flags);
+ res = __ipipe_root_status;
+ __ipipe_root_status = res | 1;
+ local_irq_restore_hw(flags);
+
+ return res & 1;
+}
+
+#endif /* CONFIG_SMP */
+
+#endif /* !__ASSEMBLY__ */
+
+#endif /* __MIPS_IPIPE_BASE_H */
diff --git a/include/asm-mips/irq.h b/include/asm-mips/irq.h
index a58f0ee..e095195 100644
--- a/include/asm-mips/irq.h
+++ b/include/asm-mips/irq.h
@@ -119,6 +119,12 @@ do { \
* Ideally there should be away to get this into kernel/irq/handle.c to
* avoid the overhead of a call for just a tiny function ...
*/
+#ifdef CONFIG_IPIPE
+#define do_IRQ(irq) \
+do { \
+ __ipipe_handle_irq(irq, get_irq_regs()); \
+} while (0)
+#else /* !CONFIG_IPIPE */
#define do_IRQ(irq) \
do { \
irq_enter(); \
@@ -126,6 +132,7 @@ do { \
generic_handle_irq(irq); \
irq_exit(); \
} while (0)
+#endif /* CONFIG_IPIPE */
#ifdef CONFIG_MIPS_MT_SMTC_IRQAFF
/*
diff --git a/include/asm-mips/irqflags.h b/include/asm-mips/irqflags.h
index 881e886..f04cf32 100644
--- a/include/asm-mips/irqflags.h
+++ b/include/asm-mips/irqflags.h
@@ -16,8 +16,10 @@
#include <linux/compiler.h>
#include <asm/hazards.h>
+#include <linux/ipipe_trace.h>
+
__asm__(
- " .macro raw_local_irq_enable \n"
+ " .macro local_irq_enable_hw_notrace \n"
" .set push \n"
" .set reorder \n"
" .set noat \n"
@@ -38,10 +40,10 @@ __asm__(
" .set pop \n"
" .endm");
-static inline void raw_local_irq_enable(void)
+static inline void local_irq_enable_hw_notrace(void)
{
__asm__ __volatile__(
- "raw_local_irq_enable"
+ "local_irq_enable_hw_notrace"
: /* no outputs */
: /* no inputs */
: "memory");
@@ -66,7 +68,7 @@ static inline void raw_local_irq_enable(void)
* Workaround: mask EXL bit of the result or place a nop before mfc0.
*/
__asm__(
- " .macro raw_local_irq_disable\n"
+ " .macro local_irq_disable_hw_notrace\n"
" .set push \n"
" .set noat \n"
#ifdef CONFIG_MIPS_MT_SMTC
@@ -87,17 +89,17 @@ __asm__(
" .set pop \n"
" .endm \n");
-static inline void raw_local_irq_disable(void)
+static inline void local_irq_disable_hw_notrace(void)
{
__asm__ __volatile__(
- "raw_local_irq_disable"
+ "local_irq_disable_hw_notrace"
: /* no outputs */
: /* no inputs */
: "memory");
}
__asm__(
- " .macro raw_local_save_flags flags \n"
+ " .macro local_save_flags_hw flags \n"
" .set push \n"
" .set reorder \n"
#ifdef CONFIG_MIPS_MT_SMTC
@@ -108,13 +110,13 @@ __asm__(
" .set pop \n"
" .endm \n");
-#define raw_local_save_flags(x) \
+#define local_save_flags_hw(x) \
__asm__ __volatile__( \
- "raw_local_save_flags %0" \
+ "local_save_flags_hw %0" \
: "=r" (x))
__asm__(
- " .macro raw_local_irq_save result \n"
+ " .macro local_irq_save_hw_notrace result \n"
" .set push \n"
" .set reorder \n"
" .set noat \n"
@@ -138,15 +140,15 @@ __asm__(
" .set pop \n"
" .endm \n");
-#define raw_local_irq_save(x) \
+#define local_irq_save_hw_notrace(x) \
__asm__ __volatile__( \
- "raw_local_irq_save\t%0" \
+ "local_irq_save_hw_notrace\t%0" \
: "=r" (x) \
: /* no inputs */ \
: "memory")
__asm__(
- " .macro raw_local_irq_restore flags \n"
+ " .macro local_irq_restore_hw_notrace flags \n"
" .set push \n"
" .set noreorder \n"
" .set noat \n"
@@ -187,7 +189,7 @@ __asm__(
extern void smtc_ipi_replay(void);
-static inline void raw_local_irq_restore(unsigned long flags)
+static inline void local_irq_restore_hw_notrace(unsigned long flags)
{
unsigned long __tmp1;
@@ -202,7 +204,7 @@ static inline void raw_local_irq_restore(unsigned long flags)
#endif
__asm__ __volatile__(
- "raw_local_irq_restore\t%0"
+ "local_irq_restore_hw_notrace\t%0"
: "=r" (__tmp1)
: "0" (flags)
: "memory");
@@ -220,6 +222,21 @@ static inline int raw_irqs_disabled_flags(unsigned long flags)
#endif
}
+// FIXME: Is this valid for MIPS ?
+static inline unsigned long raw_mangle_irq_bits(int virt, unsigned long real)
+{
+ /* Merge virtual and real interrupt mask bits into a single
+ 32bit word. */
+ return (real & ~(1L << 31)) | ((virt != 0) << 31);
+}
+
+static inline int raw_demangle_irq_bits(unsigned long *x)
+{
+ int virt = (*x & (1 << 31)) != 0;
+ *x &= ~(1L << 31);
+ return virt;
+}
+
#endif
/*
@@ -260,4 +277,35 @@ static inline int raw_irqs_disabled_flags(unsigned long flags)
# define TRACE_IRQS_OFF
#endif
+#ifdef CONFIG_IPIPE
+
+# define raw_local_irq_save(flags) ((flags) = !__ipipe_test_and_stall_root())
+# define raw_local_irq_enable() __ipipe_unstall_root()
+# define raw_local_irq_disable() __ipipe_stall_root()
+# define raw_local_save_flags(flags) ((flags) = !__ipipe_test_root())
+# define raw_local_irq_restore(flags) __ipipe_restore_root(!flags)
+
+# ifdef CONFIG_IPIPE_TRACE_IRQSOFF
+# error "CONFIG_IPIPE_TRACE_IRQSOFF not implemented yet."
+# else /* !CONFIG_IPIPE_TRACE_IRQSOFF */
+# define local_irq_save_hw(flags) local_irq_save_hw_notrace(flags)
+# define local_irq_enable_hw() local_irq_enable_hw_notrace()
+# define local_irq_disable_hw() local_irq_disable_hw_notrace()
+# define local_irq_restore_hw(flags) local_irq_restore_hw_notrace(flags)
+# endif /* CONFIG_IPIPE_TRACE_IRQSOFF */
+
+#else /* !CONFIG_IPIPE */
+
+# define raw_local_irq_save(flags) local_irq_save_hw_notrace(flags)
+# define raw_local_irq_enable() local_irq_enable_hw_notrace()
+# define raw_local_irq_disable() local_irq_disable_hw_notrace()
+# define raw_local_save_flags(flags) local_save_flags_hw(flags)
+# define raw_local_irq_restore(flags) local_irq_restore_hw_notrace(flags)
+
+# define local_irq_save_hw(flags) local_irq_save_hw_notrace(flags)
+# define local_irq_enable_hw() local_irq_enable_hw_notrace()
+# define local_irq_disable_hw() local_irq_disable_hw_notrace()
+# define local_irq_restore_hw(flags) local_irq_restore_hw_notrace(flags)
+#endif /* CONFIG_IPIPE */
+
#endif /* _ASM_IRQFLAGS_H */
diff --git a/include/asm-mips/system.h b/include/asm-mips/system.h
index a944eda..c4a273a 100644
--- a/include/asm-mips/system.h
+++ b/include/asm-mips/system.h
@@ -118,10 +118,10 @@ static inline unsigned long __xchg_u32(volatile int * m, unsigned int val)
} else {
unsigned long flags;
- raw_local_irq_save(flags);
+ local_irq_save_hw(flags);
retval = *m;
*m = val;
- raw_local_irq_restore(flags); /* implies memory barrier */
+ local_irq_restore_hw(flags); /* implies memory barrier */
}
smp_llsc_mb();
@@ -166,10 +166,10 @@ static inline __u64 __xchg_u64(volatile __u64 * m, __u64 val)
} else {
unsigned long flags;
- raw_local_irq_save(flags);
+ local_irq_save_hw(flags);
retval = *m;
*m = val;
- raw_local_irq_restore(flags); /* implies memory barrier */
+ local_irq_restore_hw(flags); /* implies memory barrier */
}
smp_llsc_mb();
diff --git a/include/linux/hardirq.h b/include/linux/hardirq.h
index 8d30229..74b355d 100644
--- a/include/linux/hardirq.h
+++ b/include/linux/hardirq.h
@@ -146,7 +146,7 @@ extern void irq_enter(void);
*/
extern void irq_exit(void);
-#define nmi_enter() do { lockdep_off(); __irq_enter(); } while (0)
-#define nmi_exit() do { __irq_exit(); lockdep_on(); } while (0)
+#define nmi_enter() do { if (ipipe_root_domain_p) { lockdep_off(); __irq_enter(); } } while (0)
+#define nmi_exit() do { if (ipipe_root_domain_p) { __irq_exit(); lockdep_on(); } } while (0)
#endif /* LINUX_HARDIRQ_H */
diff --git a/include/linux/ipipe.h b/include/linux/ipipe.h
new file mode 100644
index 0000000..d77f173
--- /dev/null
+++ b/include/linux/ipipe.h
@@ -0,0 +1,582 @@
+/* -*- linux-c -*-
+ * include/linux/ipipe.h
+ *
+ * Copyright (C) 2002-2007 Philippe Gerum.
+ *
+ * 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, Inc., 675 Mass Ave, Cambridge MA 02139,
+ * USA; 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.
+ */
+
+#ifndef __LINUX_IPIPE_H
+#define __LINUX_IPIPE_H
+
+#include <linux/spinlock.h>
+#include <linux/cache.h>
+#include <linux/percpu.h>
+#include <linux/mutex.h>
+#include <linux/linkage.h>
+#include <linux/ipipe_base.h>
+#include <linux/ipipe_compat.h>
+#include <asm/ipipe.h>
+
+#ifdef CONFIG_IPIPE
+
+/*
+ * Sanity check: IPIPE_VIRQ_BASE depends on CONFIG_NR_CPUS, and if the
+ * latter gets too large, we fail to map the virtual interrupts.
+ */
+#if IPIPE_VIRQ_BASE / BITS_PER_LONG > BITS_PER_LONG
+#error "CONFIG_NR_CPUS is too large, please lower it."
+#endif
+
+#define IPIPE_VERSION_STRING IPIPE_ARCH_STRING
+#define IPIPE_RELEASE_NUMBER ((IPIPE_MAJOR_NUMBER << 16) | \
+ (IPIPE_MINOR_NUMBER << 8) | \
+ (IPIPE_PATCH_NUMBER))
+
+#ifndef BROKEN_BUILTIN_RETURN_ADDRESS
+#define __BUILTIN_RETURN_ADDRESS0 ((unsigned long)__builtin_return_address(0))
+#define __BUILTIN_RETURN_ADDRESS1 ((unsigned long)__builtin_return_address(1))
+#endif /* !BUILTIN_RETURN_ADDRESS */
+
+#define IPIPE_ROOT_PRIO 100
+#define IPIPE_ROOT_ID 0
+#define IPIPE_ROOT_NPTDKEYS 4 /* Must be <= BITS_PER_LONG */
+
+#define IPIPE_RESET_TIMER 0x1
+#define IPIPE_GRAB_TIMER 0x2
+
+/* Global domain flags */
+#define IPIPE_SPRINTK_FLAG 0 /* Synchronous printk() allowed */
+#define IPIPE_AHEAD_FLAG 1 /* Domain always heads the pipeline */
+
+/* Interrupt control bits */
+#define IPIPE_HANDLE_FLAG 0
+#define IPIPE_PASS_FLAG 1
+#define IPIPE_ENABLE_FLAG 2
+#define IPIPE_DYNAMIC_FLAG IPIPE_HANDLE_FLAG
+#define IPIPE_STICKY_FLAG 3
+#define IPIPE_SYSTEM_FLAG 4
+#define IPIPE_LOCK_FLAG 5
+#define IPIPE_WIRED_FLAG 6
+#define IPIPE_EXCLUSIVE_FLAG 7
+
+#define IPIPE_HANDLE_MASK (1 << IPIPE_HANDLE_FLAG)
+#define IPIPE_PASS_MASK (1 << IPIPE_PASS_FLAG)
+#define IPIPE_ENABLE_MASK (1 << IPIPE_ENABLE_FLAG)
+#define IPIPE_DYNAMIC_MASK IPIPE_HANDLE_MASK
+#define IPIPE_STICKY_MASK (1 << IPIPE_STICKY_FLAG)
+#define IPIPE_SYSTEM_MASK (1 << IPIPE_SYSTEM_FLAG)
+#define IPIPE_LOCK_MASK (1 << IPIPE_LOCK_FLAG)
+#define IPIPE_WIRED_MASK (1 << IPIPE_WIRED_FLAG)
+#define IPIPE_EXCLUSIVE_MASK (1 << IPIPE_EXCLUSIVE_FLAG)
+
+#define IPIPE_DEFAULT_MASK (IPIPE_HANDLE_MASK|IPIPE_PASS_MASK)
+#define IPIPE_STDROOT_MASK (IPIPE_HANDLE_MASK|IPIPE_PASS_MASK|IPIPE_SYSTEM_MASK)
+
+#define IPIPE_EVENT_SELF 0x80000000
+
+#define IPIPE_NR_CPUS NR_CPUS
+
+#define ipipe_current_domain ipipe_cpu_var(ipipe_percpu_domain)
+
+#define ipipe_virtual_irq_p(irq) ((irq) >= IPIPE_VIRQ_BASE && \
+ (irq) < IPIPE_NR_IRQS)
+
+#define IPIPE_SAME_HANDLER ((ipipe_irq_handler_t)(-1))
+
+typedef int (*ipipe_irq_ackfn_t)(unsigned irq);
+
+typedef int (*ipipe_event_handler_t)(unsigned event,
+ struct ipipe_domain *from,
+ void *data);
+struct ipipe_domain {
+
+ int slot; /* Slot number in percpu domain data array. */
+ struct list_head p_link; /* Link in pipeline */
+ ipipe_event_handler_t evhand[IPIPE_NR_EVENTS]; /* Event handlers. */
+ unsigned long long evself; /* Self-monitored event bits. */
+
+ struct {
+ unsigned long control;
+ ipipe_irq_ackfn_t acknowledge;
+ ipipe_irq_handler_t handler;
+ void *cookie;
+ } ____cacheline_aligned irqs[IPIPE_NR_IRQS];
+
+ int priority;
+ void *pdd;
+ unsigned long flags;
+ unsigned domid;
+ const char *name;
+ struct mutex mutex;
+};
+
+#define IPIPE_HEAD_PRIORITY (-1) /* For domains always heading the pipeline */
+
+struct ipipe_domain_attr {
+
+ unsigned domid; /* Domain identifier -- Magic value set by caller */
+ const char *name; /* Domain name -- Warning: won't be dup'ed! */
+ int priority; /* Priority in interrupt pipeline */
+ void (*entry) (void); /* Domain entry point */
+ void *pdd; /* Per-domain (opaque) data pointer */
+};
+
+#ifdef CONFIG_SMP
+/* These ops must start and complete on the same CPU: care for
+ * migration. */
+#define set_bit_safe(b, a) \
+ ({ unsigned long __flags; \
+ local_irq_save_hw_notrace(__flags); \
+ __set_bit(b, a); \
+ local_irq_restore_hw_notrace(__flags); })
+#define test_and_set_bit_safe(b, a) \
+ ({ unsigned long __flags, __x; \
+ local_irq_save_hw_notrace(__flags); \
+ __x = __test_and_set_bit(b, a); \
+ local_irq_restore_hw_notrace(__flags); __x; })
+#define clear_bit_safe(b, a) \
+ ({ unsigned long __flags; \
+ local_irq_save_hw_notrace(__flags); \
+ __clear_bit(b, a); \
+ local_irq_restore_hw_notrace(__flags); })
+#else
+#define set_bit_safe(b, a) set_bit(b, a)
+#define test_and_set_bit_safe(b, a) test_and_set_bit(b, a)
+#define clear_bit_safe(b, a) clear_bit(b, a)
+#endif
+
+#define __ipipe_irq_cookie(ipd, irq) (ipd)->irqs[irq].cookie
+#define __ipipe_irq_handler(ipd, irq) (ipd)->irqs[irq].handler
+#define __ipipe_cpudata_irq_hits(ipd, cpu, irq) ipipe_percpudom(ipd, irqall, cpu)[irq]
+
+extern unsigned __ipipe_printk_virq;
+
+extern unsigned long __ipipe_virtual_irq_map;
+
+extern struct list_head __ipipe_pipeline;
+
+extern int __ipipe_event_monitors[];
+
+/* Private interface */
+
+void ipipe_init(void);
+
+#ifdef CONFIG_PROC_FS
+void ipipe_init_proc(void);
+
+#ifdef CONFIG_IPIPE_TRACE
+void __ipipe_init_tracer(void);
+#else /* !CONFIG_IPIPE_TRACE */
+#define __ipipe_init_tracer() do { } while(0)
+#endif /* CONFIG_IPIPE_TRACE */
+
+#else /* !CONFIG_PROC_FS */
+#define ipipe_init_proc() do { } while(0)
+#endif /* CONFIG_PROC_FS */
+
+void __ipipe_init_stage(struct ipipe_domain *ipd);
+
+void __ipipe_cleanup_domain(struct ipipe_domain *ipd);
+
+void __ipipe_add_domain_proc(struct ipipe_domain *ipd);
+
+void __ipipe_remove_domain_proc(struct ipipe_domain *ipd);
+
+void __ipipe_flush_printk(unsigned irq, void *cookie);
+
+void __ipipe_walk_pipeline(struct list_head *pos);
+
+int __ipipe_schedule_irq(unsigned irq, struct list_head *head);
+
+int __ipipe_dispatch_event(unsigned event, void *data);
+
+int __ipipe_dispatch_wired(struct ipipe_domain *head_domain, unsigned irq);
+
+void __ipipe_sync_stage(unsigned long syncmask);
+
+void __ipipe_set_irq_pending(struct ipipe_domain *ipd, unsigned irq);
+
+void __ipipe_lock_irq(struct ipipe_domain *ipd, int cpu, unsigned irq);
+
+void __ipipe_unlock_irq(struct ipipe_domain *ipd, unsigned irq);
+
+#define __ipipe_pin_range_globally(start, end) do { } while(0)
+
+/* Must be called hw IRQs off. */
+static inline void ipipe_irq_lock(unsigned irq)
+{
+ __ipipe_lock_irq(ipipe_current_domain, ipipe_processor_id(), irq);
+}
+
+/* Must be called hw IRQs off. */
+static inline void ipipe_irq_unlock(unsigned irq)
+{
+ __ipipe_unlock_irq(ipipe_current_domain, irq);
+}
+
+#ifndef __ipipe_sync_pipeline
+#define __ipipe_sync_pipeline(syncmask) __ipipe_sync_stage(syncmask)
+#endif
+
+#ifndef __ipipe_run_irqtail
+#define __ipipe_run_irqtail() do { } while(0)
+#endif
+
+#define __ipipe_pipeline_head_p(ipd) (&(ipd)->p_link == __ipipe_pipeline.next)
+
+/*
+ * Keep the following as a macro, so that client code could check for
+ * the support of the invariant pipeline head optimization.
+ */
+#define __ipipe_pipeline_head() list_entry(__ipipe_pipeline.next,struct ipipe_domain,p_link)
+
+#define __ipipe_event_monitored_p(ev) \
+ (__ipipe_event_monitors[ev] > 0 || (ipipe_current_domain->evself & (1LL << ev)))
+
+#ifdef CONFIG_SMP
+
+cpumask_t __ipipe_set_irq_affinity(unsigned irq,
+ cpumask_t cpumask);
+
+int __ipipe_send_ipi(unsigned ipi,
+ cpumask_t cpumask);
+
+#endif /* CONFIG_SMP */
+
+#define ipipe_sigwake_notify(p) \
+do { \
+ if (((p)->flags & PF_EVNOTIFY) && __ipipe_event_monitored_p(IPIPE_EVENT_SIGWAKE)) \
+ __ipipe_dispatch_event(IPIPE_EVENT_SIGWAKE,p); \
+} while(0)
+
+#define ipipe_exit_notify(p) \
+do { \
+ if (((p)->flags & PF_EVNOTIFY) && __ipipe_event_monitored_p(IPIPE_EVENT_EXIT)) \
+ __ipipe_dispatch_event(IPIPE_EVENT_EXIT,p); \
+} while(0)
+
+#define ipipe_setsched_notify(p) \
+do { \
+ if (((p)->flags & PF_EVNOTIFY) && __ipipe_event_monitored_p(IPIPE_EVENT_SETSCHED)) \
+ __ipipe_dispatch_event(IPIPE_EVENT_SETSCHED,p); \
+} while(0)
+
+#define ipipe_schedule_notify(prev, next) \
+do { \
+ if ((((prev)->flags|(next)->flags) & PF_EVNOTIFY) && \
+ __ipipe_event_monitored_p(IPIPE_EVENT_SCHEDULE)) \
+ __ipipe_dispatch_event(IPIPE_EVENT_SCHEDULE,next); \
+} while(0)
+
+#define ipipe_trap_notify(ex, regs) \
+({ \
+ int ret = 0; \
+ if ((test_bit(IPIPE_NOSTACK_FLAG, &ipipe_this_cpudom_var(status)) || \
+ ((current)->flags & PF_EVNOTIFY)) && \
+ __ipipe_event_monitored_p(ex)) \
+ ret = __ipipe_dispatch_event(ex, regs); \
+ ret; \
+})
+
+static inline void ipipe_init_notify(struct task_struct *p)
+{
+ if (__ipipe_event_monitored_p(IPIPE_EVENT_INIT))
+ __ipipe_dispatch_event(IPIPE_EVENT_INIT,p);
+}
+
+struct mm_struct;
+
+static inline void ipipe_cleanup_notify(struct mm_struct *mm)
+{
+ if (__ipipe_event_monitored_p(IPIPE_EVENT_CLEANUP))
+ __ipipe_dispatch_event(IPIPE_EVENT_CLEANUP,mm);
+}
+
+/* Public interface */
+
+int ipipe_register_domain(struct ipipe_domain *ipd,
+ struct ipipe_domain_attr *attr);
+
+int ipipe_unregister_domain(struct ipipe_domain *ipd);
+
+void ipipe_suspend_domain(void);
+
+int ipipe_virtualize_irq(struct ipipe_domain *ipd,
+ unsigned irq,
+ ipipe_irq_handler_t handler,
+ void *cookie,
+ ipipe_irq_ackfn_t acknowledge,
+ unsigned modemask);
+
+int ipipe_control_irq(unsigned irq,
+ unsigned clrmask,
+ unsigned setmask);
+
+unsigned ipipe_alloc_virq(void);
+
+int ipipe_free_virq(unsigned virq);
+
+int ipipe_trigger_irq(unsigned irq);
+
+static inline int ipipe_propagate_irq(unsigned irq)
+{
+ return __ipipe_schedule_irq(irq, ipipe_current_domain->p_link.next);
+}
+
+static inline int ipipe_schedule_irq(unsigned irq)
+{
+ return __ipipe_schedule_irq(irq, &ipipe_current_domain->p_link);
+}
+
+void ipipe_stall_pipeline_from(struct ipipe_domain *ipd);
+
+unsigned long ipipe_test_and_stall_pipeline_from(struct ipipe_domain *ipd);
+
+void ipipe_unstall_pipeline_from(struct ipipe_domain *ipd);
+
+unsigned long ipipe_test_and_unstall_pipeline_from(struct ipipe_domain *ipd);
+
+void ipipe_restore_pipeline_from(struct ipipe_domain *ipd,
+ unsigned long x);
+
+static inline unsigned long ipipe_test_pipeline_from(struct ipipe_domain *ipd)
+{
+ return test_bit(IPIPE_STALL_FLAG, &ipipe_cpudom_var(ipd, status));
+}
+
+static inline void ipipe_stall_pipeline_head(void)
+{
+ local_irq_disable_hw();
+ __set_bit(IPIPE_STALL_FLAG, &ipipe_head_cpudom_var(status));
+}
+
+static inline unsigned long ipipe_test_and_stall_pipeline_head(void)
+{
+ local_irq_disable_hw();
+ return __test_and_set_bit(IPIPE_STALL_FLAG, &ipipe_head_cpudom_var(status));
+}
+
+void ipipe_unstall_pipeline_head(void);
+
+void __ipipe_restore_pipeline_head(unsigned long x);
+
+static inline void ipipe_restore_pipeline_head(unsigned long x)
+{
+ /* On some archs, __test_and_set_bit() might return different
+ * truth value than test_bit(), so we test the exclusive OR of
+ * both statuses, assuming that the lowest bit is always set in
+ * the truth value (if this is wrong, the failed optimization will
+ * be caught in __ipipe_restore_pipeline_head() if
+ * CONFIG_DEBUG_KERNEL is set). */
+ if ((x ^ test_bit(IPIPE_STALL_FLAG, &ipipe_head_cpudom_var(status))) & 1)
+ __ipipe_restore_pipeline_head(x);
+}
+
+#define ipipe_unstall_pipeline() \
+ ipipe_unstall_pipeline_from(ipipe_current_domain)
+
+#define ipipe_test_and_unstall_pipeline() \
+ ipipe_test_and_unstall_pipeline_from(ipipe_current_domain)
+
+#define ipipe_test_pipeline() \
+ ipipe_test_pipeline_from(ipipe_current_domain)
+
+#define ipipe_test_and_stall_pipeline() \
+ ipipe_test_and_stall_pipeline_from(ipipe_current_domain)
+
+#define ipipe_stall_pipeline() \
+ ipipe_stall_pipeline_from(ipipe_current_domain)
+
+#define ipipe_restore_pipeline(x) \
+ ipipe_restore_pipeline_from(ipipe_current_domain, (x))
+
+void ipipe_init_attr(struct ipipe_domain_attr *attr);
+
+int ipipe_get_sysinfo(struct ipipe_sysinfo *sysinfo);
+
+unsigned long ipipe_critical_enter(void (*syncfn) (void));
+
+void ipipe_critical_exit(unsigned long flags);
+
+static inline void ipipe_set_printk_sync(struct ipipe_domain *ipd)
+{
+ set_bit(IPIPE_SPRINTK_FLAG, &ipd->flags);
+}
+
+static inline void ipipe_set_printk_async(struct ipipe_domain *ipd)
+{
+ clear_bit(IPIPE_SPRINTK_FLAG, &ipd->flags);
+}
+
+static inline void ipipe_set_foreign_stack(struct ipipe_domain *ipd)
+{
+ /* Must be called hw interrupts off. */
+ __set_bit(IPIPE_NOSTACK_FLAG, &ipipe_cpudom_var(ipd, status));
+}
+
+static inline void ipipe_clear_foreign_stack(struct ipipe_domain *ipd)
+{
+ /* Must be called hw interrupts off. */
+ __clear_bit(IPIPE_NOSTACK_FLAG, &ipipe_cpudom_var(ipd, status));
+}
+
+#ifndef ipipe_safe_current
+#define ipipe_safe_current() \
+({ \
+ struct task_struct *p; \
+ p = test_bit(IPIPE_NOSTACK_FLAG, \
+ &ipipe_this_cpudom_var(status)) ? &init_task : current; \
+ p; \
+})
+#endif
+
+ipipe_event_handler_t ipipe_catch_event(struct ipipe_domain *ipd,
+ unsigned event,
+ ipipe_event_handler_t handler);
+
+cpumask_t ipipe_set_irq_affinity(unsigned irq,
+ cpumask_t cpumask);
+
+int ipipe_send_ipi(unsigned ipi,
+ cpumask_t cpumask);
+
+int ipipe_setscheduler_root(struct task_struct *p,
+ int policy,
+ int prio);
+
+int ipipe_reenter_root(struct task_struct *prev,
+ int policy,
+ int prio);
+
+int ipipe_alloc_ptdkey(void);
+
+int ipipe_free_ptdkey(int key);
+
+int ipipe_set_ptd(int key,
+ void *value);
+
+void *ipipe_get_ptd(int key);
+
+int ipipe_disable_ondemand_mappings(struct task_struct *tsk);
+
+#define local_irq_enable_hw_cond() local_irq_enable_hw()
+#define local_irq_disable_hw_cond() local_irq_disable_hw()
+#define local_irq_save_hw_cond(flags) local_irq_save_hw(flags)
+#define local_irq_restore_hw_cond(flags) local_irq_restore_hw(flags)
+
+#define local_irq_enable_nohead(ipd) \
+ do { \
+ if (!__ipipe_pipeline_head_p(ipd)) \
+ local_irq_enable_hw(); \
+ } while(0)
+
+#define local_irq_disable_nohead(ipd) \
+ do { \
+ if (!__ipipe_pipeline_head_p(ipd)) \
+ local_irq_disable_hw(); \
+ } while(0)
+
+#define local_irq_save_full(vflags, rflags) \
+ do { \
+ local_irq_save(vflags); \
+ local_irq_save_hw(rflags); \
+ } while(0)
+
+#define local_irq_restore_full(vflags, rflags) \
+ do { \
+ local_irq_restore_hw(rflags); \
+ local_irq_restore(vflags); \
+ } while(0)
+
+static inline void local_irq_restore_nosync(unsigned long x)
+{
+ if (x)
+ set_bit(IPIPE_STALL_FLAG, &ipipe_cpudom_var(ipipe_root_domain, status));
+ else
+ clear_bit(IPIPE_STALL_FLAG, &ipipe_cpudom_var(ipipe_root_domain, status));
+}
+
+#define ipipe_root_domain_p (ipipe_current_domain == ipipe_root_domain)
+
+#else /* !CONFIG_IPIPE */
+
+#define ipipe_init() do { } while(0)
+#define ipipe_suspend_domain() do { } while(0)
+#define ipipe_sigwake_notify(p) do { } while(0)
+#define ipipe_setsched_notify(p) do { } while(0)
+#define ipipe_init_notify(p) do { } while(0)
+#define ipipe_exit_notify(p) do { } while(0)
+#define ipipe_cleanup_notify(mm) do { } while(0)
+#define ipipe_trap_notify(t,r) 0
+#define ipipe_init_proc() do { } while(0)
+#define __ipipe_pin_range_globally(start, end) do { } while(0)
+
+#define local_irq_enable_hw_cond() do { } while(0)
+#define local_irq_disable_hw_cond() do { } while(0)
+#define local_irq_save_hw_cond(flags) do { (void)(flags); } while(0)
+#define local_irq_restore_hw_cond(flags) do { } while(0)
+
+#define ipipe_irq_lock(irq) do { } while(0)
+#define ipipe_irq_unlock(irq) do { } while(0)
+
+#define ipipe_root_domain_p 1
+#define ipipe_safe_current current
+
+#define local_irq_disable_head() local_irq_disable()
+
+#define local_irq_save_full(vflags, rflags) do { (void)(vflags); local_irq_save(rflags); } while(0)
+#define local_irq_restore_full(vflags, rflags) do { (void)(vflags); local_irq_restore(rflags); } while(0)
+#define local_irq_restore_nosync(vflags) local_irq_restore(vflags)
+
+#endif /* CONFIG_IPIPE */
+
+#ifdef CONFIG_IPIPE_DEBUG_CONTEXT
+
+#include <linux/cpumask.h>
+#include <asm/system.h>
+
+static inline int ipipe_disable_context_check(int cpu)
+{
+ return xchg(&per_cpu(ipipe_percpu_context_check, cpu), 0);
+}
+
+static inline void ipipe_restore_context_check(int cpu, int old_state)
+{
+ per_cpu(ipipe_percpu_context_check, cpu) = old_state;
+}
+
+static inline void ipipe_context_check_off(void)
+{
+ int cpu;
+ for_each_online_cpu(cpu)
+ per_cpu(ipipe_percpu_context_check, cpu) = 0;
+}
+
+#else /* !CONFIG_IPIPE_DEBUG_CONTEXT */
+
+static inline int ipipe_disable_context_check(int cpu)
+{
+ return 0;
+}
+
+static inline void ipipe_restore_context_check(int cpu, int old_state) { }
+
+static inline void ipipe_context_check_off(void) { }
+
+#endif /* !CONFIG_IPIPE_DEBUG_CONTEXT */
+
+#endif /* !__LINUX_IPIPE_H */
diff --git a/include/linux/ipipe_base.h b/include/linux/ipipe_base.h
new file mode 100644
index 0000000..b31c8a2
--- /dev/null
+++ b/include/linux/ipipe_base.h
@@ -0,0 +1,86 @@
+/* -*- linux-c -*-
+ * include/linux/ipipe_base.h
+ *
+ * Copyright (C) 2002-2007 Philippe Gerum.
+ * 2007 Jan Kiszka.
+ *
+ * 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, Inc., 675 Mass Ave, Cambridge MA 02139,
+ * USA; 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.
+ */
+
+#ifndef __LINUX_IPIPE_BASE_H
+#define __LINUX_IPIPE_BASE_H
+
+#ifdef CONFIG_IPIPE
+
+#include <linux/bitops.h>
+#include <asm/ipipe_base.h>
+
+/* Number of virtual IRQs */
+#define IPIPE_NR_VIRQS BITS_PER_LONG
+/* First virtual IRQ # */
+#define IPIPE_VIRQ_BASE (((IPIPE_NR_XIRQS + BITS_PER_LONG - 1) / BITS_PER_LONG) * BITS_PER_LONG)
+/* Total number of IRQ slots */
+#define IPIPE_NR_IRQS (IPIPE_VIRQ_BASE + IPIPE_NR_VIRQS)
+/* Number of indirect words needed to map the whole IRQ space. */
+#define IPIPE_IRQ_IWORDS ((IPIPE_NR_IRQS + BITS_PER_LONG - 1) / BITS_PER_LONG)
+#define IPIPE_IRQ_IMASK (BITS_PER_LONG - 1)
+#define IPIPE_IRQMASK_ANY (~0L)
+#define IPIPE_IRQMASK_VIRT (IPIPE_IRQMASK_ANY << (IPIPE_VIRQ_BASE / BITS_PER_LONG))
+
+/* Per-cpu pipeline status */
+#define IPIPE_STALL_FLAG 0 /* Stalls a pipeline stage -- guaranteed at bit #0 */
+#define IPIPE_SYNC_FLAG 1 /* The interrupt syncer is running for the domain */
+#define IPIPE_NOSTACK_FLAG 2 /* Domain currently runs on a foreign stack */
+
+#define IPIPE_STALL_MASK (1L << IPIPE_STALL_FLAG)
+#define IPIPE_SYNC_MASK (1L << IPIPE_SYNC_FLAG)
+
+typedef void (*ipipe_irq_handler_t)(unsigned irq,
+ void *cookie);
+
+extern struct ipipe_domain ipipe_root;
+
+#define ipipe_root_domain (&ipipe_root)
+
+void __ipipe_unstall_root(void);
+
+void __ipipe_restore_root(unsigned long x);
+
+#define ipipe_preempt_disable(flags) local_irq_save_hw(flags)
+#define ipipe_preempt_enable(flags) local_irq_restore_hw(flags)
+
+#ifdef CONFIG_IPIPE_DEBUG_CONTEXT
+void ipipe_check_context(struct ipipe_domain *border_ipd);
+#else /* !CONFIG_IPIPE_DEBUG_CONTEXT */
+static inline void ipipe_check_context(struct ipipe_domain *border_ipd) { }
+#endif /* !CONFIG_IPIPE_DEBUG_CONTEXT */
+
+/* Generic features */
+
+#ifdef CONFIG_GENERIC_CLOCKEVENTS
+#define __IPIPE_FEATURE_REQUEST_TICKDEV 1
+#endif
+
+#else /* !CONFIG_IPIPE */
+#define ipipe_preempt_disable(flags) do { \
+ preempt_disable(); \
+ (void)(flags); \
+ } while (0)
+#define ipipe_preempt_enable(flags) preempt_enable()
+#define ipipe_check_context(ipd) do { } while(0)
+#endif /* CONFIG_IPIPE */
+
+#endif /* !__LINUX_IPIPE_BASE_H */
diff --git a/include/linux/ipipe_compat.h b/include/linux/ipipe_compat.h
new file mode 100644
index 0000000..50a245c
--- /dev/null
+++ b/include/linux/ipipe_compat.h
@@ -0,0 +1,54 @@
+/* -*- linux-c -*-
+ * include/linux/ipipe_compat.h
+ *
+ * Copyright (C) 2007 Philippe Gerum.
+ *
+ * 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, Inc., 675 Mass Ave, Cambridge MA 02139,
+ * USA; 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.
+ */
+
+#ifndef __LINUX_IPIPE_COMPAT_H
+#define __LINUX_IPIPE_COMPAT_H
+
+#ifdef CONFIG_IPIPE_COMPAT
+/*
+ * OBSOLETE: defined only for backward compatibility. Will be removed
+ * in future releases, please update client code accordingly.
+ */
+
+#ifdef CONFIG_SMP
+#define ipipe_declare_cpuid int cpuid
+#define ipipe_load_cpuid() do { \
+ cpuid = ipipe_processor_id(); \
+ } while(0)
+#define ipipe_lock_cpu(flags) do { \
+ local_irq_save_hw(flags); \
+ cpuid = ipipe_processor_id(); \
+ } while(0)
+#define ipipe_unlock_cpu(flags) local_irq_restore_hw(flags)
+#define ipipe_get_cpu(flags) ipipe_lock_cpu(flags)
+#define ipipe_put_cpu(flags) ipipe_unlock_cpu(flags)
+#else /* !CONFIG_SMP */
+#define ipipe_declare_cpuid const int cpuid = 0
+#define ipipe_load_cpuid() do { } while(0)
+#define ipipe_lock_cpu(flags) local_irq_save_hw(flags)
+#define ipipe_unlock_cpu(flags) local_irq_restore_hw(flags)
+#define ipipe_get_cpu(flags) do { (void)(flags); } while(0)
+#define ipipe_put_cpu(flags) do { } while(0)
+#endif /* CONFIG_SMP */
+
+#endif /* CONFIG_IPIPE_COMPAT */
+
+#endif /* !__LINUX_IPIPE_COMPAT_H */
diff --git a/include/linux/ipipe_percpu.h b/include/linux/ipipe_percpu.h
new file mode 100644
index 0000000..170acac
--- /dev/null
+++ b/include/linux/ipipe_percpu.h
@@ -0,0 +1,75 @@
+/* -*- linux-c -*-
+ * include/linux/ipipe_percpu.h
+ *
+ * Copyright (C) 2007 Philippe Gerum.
+ *
+ * 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, Inc., 675 Mass Ave, Cambridge MA 02139,
+ * USA; 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.
+ */
+
+#ifndef __LINUX_IPIPE_PERCPU_H
+#define __LINUX_IPIPE_PERCPU_H
+
+#include <asm/percpu.h>
+#include <asm/ptrace.h>
+
+struct ipipe_domain;
+
+struct ipipe_percpu_domain_data {
+ unsigned long status; /* <= Must be first in struct. */
+ unsigned long irqpend_himask;
+ unsigned long irqpend_lomask[IPIPE_IRQ_IWORDS];
+ unsigned long irqheld_mask[IPIPE_IRQ_IWORDS];
+ unsigned long irqall[IPIPE_NR_IRQS];
+ u64 evsync;
+};
+
+#ifdef CONFIG_SMP
+#define ipipe_percpudom(ipd, var, cpu) \
+ (per_cpu(ipipe_percpu_darray, cpu)[(ipd)->slot].var)
+#define ipipe_cpudom_var(ipd, var) \
+ (__raw_get_cpu_var(ipipe_percpu_darray)[(ipd)->slot].var)
+#else
+DECLARE_PER_CPU(struct ipipe_percpu_domain_data *, ipipe_percpu_daddr[CONFIG_IPIPE_DOMAINS]);
+#define ipipe_percpudom(ipd, var, cpu) \
+ (per_cpu(ipipe_percpu_daddr, cpu)[(ipd)->slot]->var)
+#define ipipe_cpudom_var(ipd, var) \
+ (__raw_get_cpu_var(ipipe_percpu_daddr)[(ipd)->slot]->var)
+#endif
+
+#define IPIPE_ROOT_SLOT 0
+#define IPIPE_HEAD_SLOT (CONFIG_IPIPE_DOMAINS - 1)
+
+DECLARE_PER_CPU(struct ipipe_percpu_domain_data, ipipe_percpu_darray[CONFIG_IPIPE_DOMAINS]);
+
+DECLARE_PER_CPU(struct ipipe_domain *, ipipe_percpu_domain);
+
+#ifdef CONFIG_IPIPE_DEBUG_CONTEXT
+DECLARE_PER_CPU(int, ipipe_percpu_context_check);
+#endif
+
+#define ipipe_percpu(var, cpu) per_cpu(var, cpu)
+#define ipipe_cpu_var(var) __raw_get_cpu_var(var)
+
+#define ipipe_root_cpudom_var(var) \
+ __raw_get_cpu_var(ipipe_percpu_darray)[IPIPE_ROOT_SLOT].var
+
+#define ipipe_this_cpudom_var(var) \
+ ipipe_cpudom_var(ipipe_current_domain, var)
+
+#define ipipe_head_cpudom_var(var) \
+ __raw_get_cpu_var(ipipe_percpu_darray)[IPIPE_HEAD_SLOT].var
+
+#endif /* !__LINUX_IPIPE_PERCPU_H */
diff --git a/include/linux/ipipe_tickdev.h b/include/linux/ipipe_tickdev.h
new file mode 100644
index 0000000..4a1cb1b
--- /dev/null
+++ b/include/linux/ipipe_tickdev.h
@@ -0,0 +1,58 @@
+/* -*- linux-c -*-
+ * include/linux/ipipe_tickdev.h
+ *
+ * Copyright (C) 2007 Philippe Gerum.
+ *
+ * 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, Inc., 675 Mass Ave, Cambridge MA 02139,
+ * USA; 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.
+ */
+
+#ifndef __LINUX_IPIPE_TICKDEV_H
+#define __LINUX_IPIPE_TICKDEV_H
+
+#if defined(CONFIG_IPIPE) && defined(CONFIG_GENERIC_CLOCKEVENTS)
+
+#include <linux/clockchips.h>
+
+struct tick_device;
+
+struct ipipe_tick_device {
+
+ void (*emul_set_mode)(enum clock_event_mode,
+ struct clock_event_device *cdev);
+ int (*emul_set_tick)(unsigned long delta,
+ struct clock_event_device *cdev);
+ void (*real_set_mode)(enum clock_event_mode mode,
+ struct clock_event_device *cdev);
+ int (*real_set_tick)(unsigned long delta,
+ struct clock_event_device *cdev);
+ struct tick_device *slave;
+ unsigned long real_max_delta_ns;
+ unsigned long real_mult;
+ int real_shift;
+};
+
+int ipipe_request_tickdev(const char *devname,
+ void (*emumode)(enum clock_event_mode mode,
+ struct clock_event_device *cdev),
+ int (*emutick)(unsigned long evt,
+ struct clock_event_device *cdev),
+ int cpu, unsigned long *tmfreq);
+
+void ipipe_release_tickdev(int cpu);
+
+#endif /* CONFIG_IPIPE && CONFIG_GENERIC_CLOCKEVENTS */
+
+#endif /* !__LINUX_IPIPE_TICKDEV_H */
diff --git a/include/linux/ipipe_trace.h b/include/linux/ipipe_trace.h
new file mode 100644
index 0000000..815b5ad
--- /dev/null
+++ b/include/linux/ipipe_trace.h
@@ -0,0 +1,70 @@
+/* -*- linux-c -*-
+ * include/linux/ipipe_trace.h
+ *
+ * Copyright (C) 2005 Luotao Fu.
+ * 2005-2007 Jan Kiszka.
+ *
+ * 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, Inc., 675 Mass Ave, Cambridge MA 02139,
+ * USA; 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.
+ */
+
+#ifndef _LINUX_IPIPE_TRACE_H
+#define _LINUX_IPIPE_TRACE_H
+
+#ifdef CONFIG_IPIPE_TRACE
+
+#include <linux/types.h>
+
+void ipipe_trace_begin(unsigned long v);
+void ipipe_trace_end(unsigned long v);
+void ipipe_trace_freeze(unsigned long v);
+void ipipe_trace_special(unsigned char special_id, unsigned long v);
+void ipipe_trace_pid(pid_t pid, short prio);
+int ipipe_trace_max_reset(void);
+int ipipe_trace_frozen_reset(void);
+
+#else /* !CONFIG_IPIPE_TRACE */
+
+#define ipipe_trace_begin(v) do { (void)(v); } while(0)
+#define ipipe_trace_end(v) do { (void)(v); } while(0)
+#define ipipe_trace_freeze(v) do { (void)(v); } while(0)
+#define ipipe_trace_special(id, v) do { (void)(id); (void)(v); } while(0)
+#define ipipe_trace_pid(pid, prio) do { (void)(pid); (void)(prio); } while(0)
+#define ipipe_trace_max_reset() do { } while(0)
+#define ipipe_trace_froze_reset() do { } while(0)
+
+#endif /* !CONFIG_IPIPE_TRACE */
+
+#ifdef CONFIG_IPIPE_TRACE_PANIC
+void ipipe_trace_panic_freeze(void);
+void ipipe_trace_panic_dump(void);
+#else
+static inline void ipipe_trace_panic_freeze(void) { }
+static inline void ipipe_trace_panic_dump(void) { }
+#endif
+
+#ifdef CONFIG_IPIPE_TRACE_IRQSOFF
+#define ipipe_trace_irq_entry(irq) ipipe_trace_begin(irq)
+#define ipipe_trace_irq_exit(irq) ipipe_trace_end(irq)
+#define ipipe_trace_irqsoff() ipipe_trace_begin(0x80000000UL)
+#define ipipe_trace_irqson() ipipe_trace_end(0x80000000UL)
+#else
+#define ipipe_trace_irq_entry(irq) do { (void)(irq);} while(0)
+#define ipipe_trace_irq_exit(irq) do { (void)(irq);} while(0)
+#define ipipe_trace_irqsoff() do { } while(0)
+#define ipipe_trace_irqson() do { } while(0)
+#endif
+
+#endif /* !__LINUX_IPIPE_TRACE_H */
diff --git a/include/linux/irq.h b/include/linux/irq.h
index 1fc1cb8..f8aa83a 100644
--- a/include/linux/irq.h
+++ b/include/linux/irq.h
@@ -150,6 +150,14 @@ struct irq_chip {
* @name: flow handler name for /proc/interrupts output
*/
struct irq_desc {
+#ifdef CONFIG_IPIPE
+ void fastcall (*ipipe_ack)(unsigned int irq,
+ struct irq_desc *desc);
+ void fastcall (*ipipe_demux)(unsigned int irq,
+ struct irq_desc *desc);
+ void fastcall (*ipipe_end)(unsigned int irq,
+ struct irq_desc *desc);
+#endif /* CONFIG_IPIPE */
irq_flow_handler_t handle_irq;
struct irq_chip *chip;
struct msi_desc *msi_desc;
@@ -370,6 +378,14 @@ set_irq_chained_handler(unsigned int irq,
extern void set_irq_noprobe(unsigned int irq);
extern void set_irq_probe(unsigned int irq);
+#ifdef CONFIG_IPIPE
+extern void
+__set_irq_demux_handler(unsigned int irq,
+ void fastcall (*decode)(unsigned int, struct irq_desc *),
+ int is_chained,
+ const char *name);
+#endif /* CONFIG_IPIPE */
+
/* Handle dynamic irq creation and destruction */
extern int create_irq(void);
extern void destroy_irq(unsigned int irq);
diff --git a/include/linux/kernel.h b/include/linux/kernel.h
index 94bc996..0510140 100644
--- a/include/linux/kernel.h
+++ b/include/linux/kernel.h
@@ -14,6 +14,7 @@
#include <linux/compiler.h>
#include <linux/bitops.h>
#include <linux/log2.h>
+#include <linux/ipipe_base.h>
#include <asm/byteorder.h>
#include <asm/bug.h>
@@ -106,9 +107,12 @@ struct user;
*/
#ifdef CONFIG_PREEMPT_VOLUNTARY
extern int cond_resched(void);
-# define might_resched() cond_resched()
+# define might_resched() do { \
+ ipipe_check_context(ipipe_root_domain); \
+ cond_resched(); \
+ } while (0)
#else
-# define might_resched() do { } while (0)
+# define might_resched() ipipe_check_context(ipipe_root_domain)
#endif
#ifdef CONFIG_DEBUG_SPINLOCK_SLEEP
diff --git a/include/linux/linkage.h b/include/linux/linkage.h
index ff203dd..3791605 100644
--- a/include/linux/linkage.h
+++ b/include/linux/linkage.h
@@ -70,4 +70,8 @@
#define fastcall
#endif
+#ifndef notrace
+#define notrace __attribute__((no_instrument_function))
+#endif
+
#endif
diff --git a/include/linux/mm.h b/include/linux/mm.h
index 1b7b95c..7f27db6 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -104,6 +104,7 @@ extern unsigned int kobjsize(const void *objp);
#define VM_MAPPED_COPY 0x01000000 /* T if mapped copy of data (nommu mmap) */
#define VM_INSERTPAGE 0x02000000 /* The vma has had "vm_insert_page()" done on it */
#define VM_ALWAYSDUMP 0x04000000 /* Always include in core dumps */
+#define VM_PINNED 0x08000000 /* Disable faults for the vma */
#define VM_CAN_NONLINEAR 0x08000000 /* Has ->fault & does nonlinear pages */
diff --git a/include/linux/preempt.h b/include/linux/preempt.h
index 484988e..dc27067 100644
--- a/include/linux/preempt.h
+++ b/include/linux/preempt.h
@@ -9,13 +9,20 @@
#include <linux/thread_info.h>
#include <linux/linkage.h>
#include <linux/list.h>
+#include <linux/ipipe_base.h>
#ifdef CONFIG_DEBUG_PREEMPT
extern void fastcall add_preempt_count(int val);
extern void fastcall sub_preempt_count(int val);
#else
-# define add_preempt_count(val) do { preempt_count() += (val); } while (0)
-# define sub_preempt_count(val) do { preempt_count() -= (val); } while (0)
+# define add_preempt_count(val) do { \
+ ipipe_check_context(ipipe_root_domain); \
+ preempt_count() += (val); \
+ } while (0)
+# define sub_preempt_count(val) do { \
+ ipipe_check_context(ipipe_root_domain); \
+ preempt_count() -= (val); \
+ } while (0)
#endif
#define inc_preempt_count() add_preempt_count(1)
diff --git a/include/linux/sched.h b/include/linux/sched.h
index cc14656..0cdea5a 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -58,6 +58,7 @@ struct sched_param {
#include <linux/errno.h>
#include <linux/nodemask.h>
#include <linux/mm_types.h>
+#include <linux/ipipe.h>
#include <asm/system.h>
#include <asm/semaphore.h>
@@ -177,6 +178,13 @@ print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
#define EXIT_DEAD 32
/* in tsk->state again */
#define TASK_DEAD 64
+#ifdef CONFIG_IPIPE
+#define TASK_ATOMICSWITCH 512
+#define TASK_NOWAKEUP 1024
+#else /* !CONFIG_IPIPE */
+#define TASK_ATOMICSWITCH 0
+#define TASK_NOWAKEUP 0
+#endif /* CONFIG_IPIPE */
#define __set_task_state(tsk, state_value) \
do { (tsk)->state = (state_value); } while (0)
@@ -1166,6 +1174,9 @@ struct task_struct {
#endif
atomic_t fs_excl; /* holding fs exclusive resources */
struct rcu_head rcu;
+#ifdef CONFIG_IPIPE
+ void *ptd[IPIPE_ROOT_NPTDKEYS];
+#endif
/*
* cache last used pipe for splice
@@ -1383,6 +1394,11 @@ static inline void put_task_struct(struct task_struct *t)
#define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
#define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
#define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */
+#ifdef CONFIG_IPIPE
+#define PF_EVNOTIFY 0x80000000 /* Notify other domains about internal events */
+#else
+#define PF_EVNOTIFY 0
+#endif /* CONFIG_IPIPE */
/*
* Only the _current_ task can read/write to tsk->flags, but other
diff --git a/include/linux/spinlock.h b/include/linux/spinlock.h
index c376f3b..2d9fbbc 100644
--- a/include/linux/spinlock.h
+++ b/include/linux/spinlock.h
@@ -89,10 +89,14 @@ extern int __lockfunc generic__raw_read_trylock(raw_rwlock_t *lock);
# include <linux/spinlock_up.h>
#endif
+#undef TYPE_EQUAL
+#define TYPE_EQUAL(lock, type) \
+ __builtin_types_compatible_p(typeof(lock), type *)
+
#ifdef CONFIG_DEBUG_SPINLOCK
extern void __spin_lock_init(spinlock_t *lock, const char *name,
struct lock_class_key *key);
-# define spin_lock_init(lock) \
+# define _spin_lock_init(lock) \
do { \
static struct lock_class_key __key; \
\
@@ -100,10 +104,21 @@ do { \
} while (0)
#else
-# define spin_lock_init(lock) \
+# define _spin_lock_init(lock) \
do { *(lock) = SPIN_LOCK_UNLOCKED; } while (0)
#endif
+# define spin_lock_init(lock) \
+ do { \
+ if (TYPE_EQUAL((lock), __ipipe_spinlock_t)) \
+ do { \
+ IPIPE_DEFINE_SPINLOCK(__lock__); \
+ *((ipipe_spinlock_t *)lock) = __lock__; \
+ } while(0); \
+ else \
+ _spin_lock_init((spinlock_t *)lock); \
+ } while(0)
+
#ifdef CONFIG_DEBUG_SPINLOCK
extern void __rwlock_init(rwlock_t *lock, const char *name,
struct lock_class_key *key);
@@ -172,7 +187,86 @@ do { \
#define read_trylock(lock) __cond_lock(lock, _read_trylock(lock))
#define write_trylock(lock) __cond_lock(lock, _write_trylock(lock))
-#define spin_lock(lock) _spin_lock(lock)
+#define PICK_SPINOP(op, lock) \
+do { \
+ if (TYPE_EQUAL((lock), __ipipe_spinlock_t)) \
+ __raw_spin##op(&((__ipipe_spinlock_t *)(lock))->__raw_lock); \
+ else if (TYPE_EQUAL(lock, spinlock_t)) \
+ _spin##op((spinlock_t *)(lock)); \
+} while (0)
+
+#define PICK_SPINOP_RAW(op, lock) \
+do { \
+ if (TYPE_EQUAL((lock), __ipipe_spinlock_t)) \
+ __raw_spin##op(&((__ipipe_spinlock_t *)(lock))->__raw_lock); \
+ else if (TYPE_EQUAL(lock, spinlock_t)) \
+ __raw_spin##op(&((spinlock_t *)(lock))->raw_lock); \
+} while (0)
+
+#define PICK_SPINLOCK_IRQ(lock) \
+do { \
+ if (TYPE_EQUAL((lock), __ipipe_spinlock_t)) { \
+ __ipipe_spin_lock_irq(&((__ipipe_spinlock_t *)(lock))->__raw_lock); \
+ } else if (TYPE_EQUAL(lock, spinlock_t)) \
+ _spin_lock_irq((spinlock_t *)(lock)); \
+} while (0)
+
+#define PICK_SPINUNLOCK_IRQ(lock) \
+do { \
+ if (TYPE_EQUAL((lock), __ipipe_spinlock_t)) { \
+ __ipipe_spin_unlock_irq(&((__ipipe_spinlock_t *)(lock))->__raw_lock); \
+ } else if (TYPE_EQUAL(lock, spinlock_t)) \
+ _spin_unlock_irq((spinlock_t *)(lock)); \
+} while (0)
+
+#define PICK_SPINLOCK_IRQ_RAW(lock) \
+do { \
+ if (TYPE_EQUAL((lock), __ipipe_spinlock_t)) { \
+ __ipipe_spin_lock_irq(&((__ipipe_spinlock_t *)(lock))->__raw_lock); \
+ } else if (TYPE_EQUAL(lock, spinlock_t)) \
+ local_irq_disable(); \
+ __raw_spin_lock(&((spinlock_t *)(lock))->raw_lock); \
+} while (0)
+
+#define PICK_SPINUNLOCK_IRQ_RAW(lock) \
+do { \
+ if (TYPE_EQUAL((lock), __ipipe_spinlock_t)) { \
+ __ipipe_spin_unlock_irq(&((__ipipe_spinlock_t *)(lock))->__raw_lock); \
+ } else if (TYPE_EQUAL(lock, spinlock_t)) \
+ __raw_spin_unlock(&((spinlock_t *)(lock))->raw_lock); \
+ local_irq_enable(); \
+} while (0)
+
+#if defined(CONFIG_SMP) || defined(CONFIG_DEBUG_SPINLOCK)
+extern int __bad_spinlock_type(void);
+
+#define PICK_SPINLOCK_IRQSAVE(lock, flags) \
+do { \
+ if (TYPE_EQUAL((lock), __ipipe_spinlock_t)) { \
+ (flags) = __ipipe_spin_lock_irqsave(&((__ipipe_spinlock_t *)(lock))->__raw_lock); \
+ } else if (TYPE_EQUAL(lock, spinlock_t)) \
+ flags = _spin_lock_irqsave((spinlock_t *)(lock)); \
+ else __bad_spinlock_type(); \
+} while (0)
+#else
+#define PICK_SPINLOCK_IRQSAVE(lock, flags) \
+do { \
+ if (TYPE_EQUAL((lock), __ipipe_spinlock_t)) { \
+ (flags) = __ipipe_spin_lock_irqsave(&((__ipipe_spinlock_t *)(lock))->__raw_lock); \
+ } else if (TYPE_EQUAL(lock, spinlock_t)) \
+ _spin_lock_irqsave((spinlock_t *)(lock), flags); \
+} while (0)
+#endif
+
+#define PICK_SPINUNLOCK_IRQRESTORE(lock, flags) \
+ do { \
+ if (TYPE_EQUAL((lock), __ipipe_spinlock_t)) { \
+ __ipipe_spin_unlock_irqrestore(&((__ipipe_spinlock_t *)(lock))->__raw_lock, flags); \
+ } else if (TYPE_EQUAL(lock, spinlock_t)) \
+ _spin_unlock_irqrestore((spinlock_t *)(lock), flags); \
+} while (0)
+
+#define spin_lock(lock) PICK_SPINOP(_lock, lock)
#ifdef CONFIG_DEBUG_LOCK_ALLOC
# define spin_lock_nested(lock, subclass) _spin_lock_nested(lock, subclass)
@@ -185,7 +279,7 @@ do { \
#if defined(CONFIG_SMP) || defined(CONFIG_DEBUG_SPINLOCK)
-#define spin_lock_irqsave(lock, flags) flags = _spin_lock_irqsave(lock)
+#define spin_lock_irqsave(lock, flags) PICK_SPINLOCK_IRQSAVE(lock, flags)
#define read_lock_irqsave(lock, flags) flags = _read_lock_irqsave(lock)
#define write_lock_irqsave(lock, flags) flags = _write_lock_irqsave(lock)
@@ -199,7 +293,7 @@ do { \
#else
-#define spin_lock_irqsave(lock, flags) _spin_lock_irqsave(lock, flags)
+#define spin_lock_irqsave(lock, flags) PICK_SPINLOCK_IRQSAVE(lock, flags)
#define read_lock_irqsave(lock, flags) _read_lock_irqsave(lock, flags)
#define write_lock_irqsave(lock, flags) _write_lock_irqsave(lock, flags)
#define spin_lock_irqsave_nested(lock, flags, subclass) \
@@ -207,7 +301,7 @@ do { \
#endif
-#define spin_lock_irq(lock) _spin_lock_irq(lock)
+#define spin_lock_irq(lock) PICK_SPINLOCK_IRQ(lock)
#define spin_lock_bh(lock) _spin_lock_bh(lock)
#define read_lock_irq(lock) _read_lock_irq(lock)
@@ -221,32 +315,40 @@ do { \
*/
#if defined(CONFIG_DEBUG_SPINLOCK) || defined(CONFIG_PREEMPT) || \
!defined(CONFIG_SMP)
-# define spin_unlock(lock) _spin_unlock(lock)
+#define spin_unlock(lock) PICK_SPINOP(_unlock, lock)
# define read_unlock(lock) _read_unlock(lock)
# define write_unlock(lock) _write_unlock(lock)
-# define spin_unlock_irq(lock) _spin_unlock_irq(lock)
-# define read_unlock_irq(lock) _read_unlock_irq(lock)
-# define write_unlock_irq(lock) _write_unlock_irq(lock)
+# define spin_unlock_irq(lock) PICK_SPINUNLOCK_IRQ(lock)
+# define read_unlock_irq(lock) _read_unlock_irq(lock)
+# define write_unlock_irq(lock) _write_unlock_irq(lock)
#else
-# define spin_unlock(lock) \
- do {__raw_spin_unlock(&(lock)->raw_lock); __release(lock); } while (0)
-# define read_unlock(lock) \
- do {__raw_read_unlock(&(lock)->raw_lock); __release(lock); } while (0)
-# define write_unlock(lock) \
- do {__raw_write_unlock(&(lock)->raw_lock); __release(lock); } while (0)
-# define spin_unlock_irq(lock) \
+# define spin_unlock(lock) \
do { \
- __raw_spin_unlock(&(lock)->raw_lock); \
+ PICK_SPINOP_RAW(_unlock, lock); \
+ __release(lock); \
+} while(0)
+# define read_unlock(lock) \
+do { \
+ __raw_read_unlock(&(lock)->raw_lock); \
+ __release(lock); \
+} while (0)
+# define write_unlock(lock) \
+do { \
+ __raw_write_unlock(&(lock)->raw_lock); \
__release(lock); \
- local_irq_enable(); \
} while (0)
-# define read_unlock_irq(lock) \
+# define spin_unlock_irq(lock) \
+do { \
+ PICK_SPINUNLOCK_IRQ_RAW(lock); \
+ __release(lock); \
+} while(0)
+# define read_unlock_irq(lock) \
do { \
__raw_read_unlock(&(lock)->raw_lock); \
__release(lock); \
local_irq_enable(); \
} while (0)
-# define write_unlock_irq(lock) \
+# define write_unlock_irq(lock) \
do { \
__raw_write_unlock(&(lock)->raw_lock); \
__release(lock); \
@@ -254,8 +356,8 @@ do { \
} while (0)
#endif
-#define spin_unlock_irqrestore(lock, flags) \
- _spin_unlock_irqrestore(lock, flags)
+#define spin_unlock_irqrestore(lock, flags) \
+ PICK_SPINUNLOCK_IRQRESTORE(lock, flags)
#define spin_unlock_bh(lock) _spin_unlock_bh(lock)
#define read_unlock_irqrestore(lock, flags) \
@@ -346,4 +448,29 @@ extern int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock);
*/
#define spin_can_lock(lock) (!spin_is_locked(lock))
+#ifdef CONFIG_IPIPE
+void fastcall __ipipe_spin_lock_irq(raw_spinlock_t *lock);
+void fastcall __ipipe_spin_unlock_irq(raw_spinlock_t *lock);
+unsigned long fastcall __ipipe_spin_lock_irqsave(raw_spinlock_t *lock);
+void fastcall __ipipe_spin_unlock_irqrestore(raw_spinlock_t *lock,
+ unsigned long x);
+void fastcall __ipipe_spin_unlock_irqbegin(ipipe_spinlock_t *lock);
+void fastcall __ipipe_spin_unlock_irqcomplete(unsigned long x);
+#define spin_lock_irqsave_cond(lock, flags) \
+ spin_lock_irqsave(lock, flags)
+#define spin_unlock_irqrestore_cond(lock, flags) \
+ spin_unlock_irqrestore(lock, flags)
+#else
+#define spin_lock_irqsave_cond(lock, flags) \
+ do { (void)(flags); spin_lock(lock); } while(0)
+#define spin_unlock_irqrestore_cond(lock, flags) \
+ spin_unlock(lock)
+#define __ipipe_spin_lock_irq(lock) do { } while(0)
+#define __ipipe_spin_unlock_irq(lock) do { } while(0)
+#define __ipipe_spin_lock_irqsave(lock) 0
+#define __ipipe_spin_unlock_irqrestore(lock, x) do { (void)(x); } while(0)
+#define __ipipe_spin_unlock_irqbegin(lock) do { } while(0)
+#define __ipipe_spin_unlock_irqcomplete(x) do { (void)(x); } while(0)
+#endif
+
#endif /* __LINUX_SPINLOCK_H */
diff --git a/include/linux/spinlock_types.h b/include/linux/spinlock_types.h
index f6a3a95..4ca73f5 100644
--- a/include/linux/spinlock_types.h
+++ b/include/linux/spinlock_types.h
@@ -31,6 +31,10 @@ typedef struct {
#endif
} spinlock_t;
+typedef struct {
+ raw_spinlock_t __raw_lock;
+} __ipipe_spinlock_t;
+
#define SPINLOCK_MAGIC 0xdead4ead
typedef struct {
@@ -92,9 +96,21 @@ typedef struct {
* __SPIN_LOCK_UNLOCKED()/__RW_LOCK_UNLOCKED() as appropriate.
*/
#define SPIN_LOCK_UNLOCKED __SPIN_LOCK_UNLOCKED(old_style_spin_init)
+#define IPIPE_SPIN_LOCK_UNLOCKED \
+ (__ipipe_spinlock_t) { .__raw_lock = __RAW_SPIN_LOCK_UNLOCKED }
#define RW_LOCK_UNLOCKED __RW_LOCK_UNLOCKED(old_style_rw_init)
#define DEFINE_SPINLOCK(x) spinlock_t x = __SPIN_LOCK_UNLOCKED(x)
#define DEFINE_RWLOCK(x) rwlock_t x = __RW_LOCK_UNLOCKED(x)
+#ifdef CONFIG_IPIPE
+# define ipipe_spinlock_t __ipipe_spinlock_t
+# define IPIPE_DEFINE_SPINLOCK(x) ipipe_spinlock_t x = IPIPE_SPIN_LOCK_UNLOCKED
+# define IPIPE_DECLARE_SPINLOCK(x) extern ipipe_spinlock_t x
+#else
+# define ipipe_spinlock_t spinlock_t
+# define IPIPE_DEFINE_SPINLOCK(x) DEFINE_SPINLOCK(x)
+# define IPIPE_DECLARE_SPINLOCK(x) extern spinlock_t x
+#endif
+
#endif /* __LINUX_SPINLOCK_TYPES_H */
diff --git a/init/Kconfig b/init/Kconfig
index b9d11a8..b40b352 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -64,6 +64,7 @@ config INIT_ENV_ARG_LIMIT
config LOCALVERSION
string "Local version - append to kernel release"
+ default "-ipipe"
help
Append an extra string to the end of your kernel version.
This will show up when you type uname, for example.
diff --git a/init/main.c b/init/main.c
index 80b04b6..a34d817 100644
--- a/init/main.c
+++ b/init/main.c
@@ -519,7 +519,7 @@ asmlinkage void __init start_kernel(void)
lockdep_init();
cgroup_init_early();
- local_irq_disable();
+ local_irq_disable_hw();
early_boot_irqs_off();
early_init_irq_lock_class();
@@ -572,10 +572,16 @@ asmlinkage void __init start_kernel(void)
softirq_init();
timekeeping_init();
time_init();
+ /*
+ * We need to wait for the interrupt and time subsystems to be
+ * initialized before enabling the pipeline.
+ */
+ ipipe_init();
profile_init();
if (!irqs_disabled())
printk("start_kernel(): bug: interrupts were enabled early\n");
early_boot_irqs_on();
+
local_irq_enable();
/*
@@ -663,7 +669,7 @@ static void __init do_initcalls(void)
{
initcall_t *call;
int count = preempt_count();
-
+
for (call = __initcall_start; call < __initcall_end; call++) {
ktime_t t0, t1, delta;
char *msg = NULL;
@@ -679,7 +685,7 @@ static void __init do_initcalls(void)
}
result = (*call)();
-
+
if (initcall_debug) {
t1 = ktime_get();
delta = ktime_sub(t1, t0);
@@ -733,6 +739,7 @@ static void __init do_basic_setup(void)
usermodehelper_init();
driver_init();
init_irq_proc();
+ ipipe_init_proc();
do_initcalls();
}
diff --git a/kernel/Makefile b/kernel/Makefile
index dfa9695..c78c5cd 100644
--- a/kernel/Makefile
+++ b/kernel/Makefile
@@ -53,6 +53,7 @@ obj-$(CONFIG_GENERIC_HARDIRQS) += irq/
obj-$(CONFIG_SECCOMP) += seccomp.o
obj-$(CONFIG_RCU_TORTURE_TEST) += rcutorture.o
obj-$(CONFIG_RELAY) += relay.o
+obj-$(CONFIG_IPIPE) += ipipe/
obj-$(CONFIG_SYSCTL) += utsname_sysctl.o
obj-$(CONFIG_TASK_DELAY_ACCT) += delayacct.o
obj-$(CONFIG_TASKSTATS) += taskstats.o tsacct.o
@@ -80,3 +81,5 @@ quiet_cmd_ikconfiggz = IKCFG $@
targets += config_data.h
$(obj)/config_data.h: $(obj)/config_data.gz FORCE
$(call if_changed,ikconfiggz)
+
+# obj-$(CONFIG_XENOMAI) += xenomai/
diff --git a/kernel/exit.c b/kernel/exit.c
index 549c055..41e975a 100644
--- a/kernel/exit.c
+++ b/kernel/exit.c
@@ -994,6 +994,7 @@ fastcall NORET_TYPE void do_exit(long code)
if (group_dead)
acct_process();
+ ipipe_exit_notify(tsk);
exit_sem(tsk);
__exit_files(tsk);
__exit_fs(tsk);
diff --git a/kernel/fork.c b/kernel/fork.c
index 8dd8ff2..0e97457 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -403,6 +403,7 @@ void mmput(struct mm_struct *mm)
if (atomic_dec_and_test(&mm->mm_users)) {
exit_aio(mm);
exit_mmap(mm);
+ ipipe_cleanup_notify(mm);
if (!list_empty(&mm->mmlist)) {
spin_lock(&mmlist_lock);
list_del(&mm->mmlist);
@@ -938,7 +939,7 @@ static void copy_flags(unsigned long clone_flags, struct task_struct *p)
{
unsigned long new_flags = p->flags;
- new_flags &= ~PF_SUPERPRIV;
+ new_flags &= ~(PF_SUPERPRIV | PF_EVNOTIFY);
new_flags |= PF_FORKNOEXEC;
if (!(clone_flags & CLONE_PTRACE))
p->ptrace = 0;
@@ -1312,6 +1313,9 @@ static struct task_struct *copy_process(unsigned long clone_flags,
write_unlock_irq(&tasklist_lock);
proc_fork_connector(p);
cgroup_post_fork(p);
+#ifdef CONFIG_IPIPE
+ memset(p->ptd, 0, sizeof(p->ptd));
+#endif /* CONFIG_IPIPE */
return p;
bad_fork_free_pid:
diff --git a/kernel/ipipe/Kconfig b/kernel/ipipe/Kconfig
new file mode 100644
index 0000000..d523b18
--- /dev/null
+++ b/kernel/ipipe/Kconfig
@@ -0,0 +1,25 @@
+config IPIPE
+ bool "Interrupt pipeline"
+ default y
+ ---help---
+ Activate this option if you want the interrupt pipeline to be
+ compiled in.
+
+config IPIPE_DOMAINS
+ int "Max domains"
+ depends on IPIPE
+ default 4
+ ---help---
+ The maximum number of I-pipe domains to run concurrently.
+
+config IPIPE_COMPAT
+ bool "Maintain code compatibility with older releases"
+ depends on IPIPE
+ default y
+ ---help---
+ Activate this option if you want the compatibility code to be
+ defined, so that older I-pipe clients may use obsolete
+ constructs. WARNING: obsolete code will be eventually
+ deprecated in future I-pipe releases, and removed from the
+ compatibility support as time passes. Please fix I-pipe
+ clients to get rid of such uses as soon as possible.
diff --git a/kernel/ipipe/Kconfig.debug b/kernel/ipipe/Kconfig.debug
new file mode 100644
index 0000000..64a692c
--- /dev/null
+++ b/kernel/ipipe/Kconfig.debug
@@ -0,0 +1,88 @@
+config IPIPE_DEBUG
+ bool "I-pipe debugging"
+ depends on IPIPE
+
+config IPIPE_DEBUG_CONTEXT
+ bool "Check for illicit cross-domain calls"
+ depends on IPIPE_DEBUG
+ default y
+ ---help---
+ Enable this feature to arm checkpoints in the kernel that
+ verify the correct invocation context. On entry of critical
+ Linux services a warning is issued if the caller is not
+ running over the root domain.
+
+config IPIPE_TRACE
+ bool "Latency tracing"
+ depends on IPIPE_DEBUG
+ select FRAME_POINTER
+ select KALLSYMS
+ select PROC_FS
+ ---help---
+ Activate this option if you want to use per-function tracing of
+ the kernel. The tracer will collect data via instrumentation
+ features like the one below or with the help of explicite calls
+ of ipipe_trace_xxx(). See include/linux/ipipe_trace.h for the
+ in-kernel tracing API. The collected data and runtime control
+ is available via /proc/ipipe/trace/*.
+
+if IPIPE_TRACE
+
+config IPIPE_TRACE_ENABLE
+ bool "Enable tracing on boot"
+ default y
+ ---help---
+ Disable this option if you want to arm the tracer after booting
+ manually ("echo 1 > /proc/ipipe/tracer/enable"). This can reduce
+ boot time on slow embedded devices due to the tracer overhead.
+
+config IPIPE_TRACE_MCOUNT
+ bool "Instrument function entries"
+ default y
+ ---help---
+ When enabled, records every kernel function entry in the tracer
+ log. While this slows down the system noticeably, it provides
+ the highest level of information about the flow of events.
+ However, it can be switch off in order to record only explicit
+ I-pipe trace points.
+
+config IPIPE_TRACE_IRQSOFF
+ bool "Trace IRQs-off times"
+ default y
+ ---help---
+ Activate this option if I-pipe shall trace the longest path
+ with hard-IRQs switched off.
+
+config IPIPE_TRACE_SHIFT
+ int "Depth of trace log (14 => 16Kpoints, 15 => 32Kpoints)"
+ range 10 18
+ default 14
+ ---help---
+ The number of trace points to hold tracing data for each
+ trace path, as a power of 2.
+
+config IPIPE_TRACE_VMALLOC
+ bool "Use vmalloc'ed trace buffer"
+ default y if EMBEDDED
+ ---help---
+ Instead of reserving static kernel data, the required buffer
+ is allocated via vmalloc during boot-up when this option is
+ enabled. This can help to start systems that are low on memory,
+ but it slightly degrades overall performance. Try this option
+ when a traced kernel hangs unexpectedly at boot time.
+
+config IPIPE_TRACE_PANIC
+ bool "Enable panic back traces"
+ default y
+ ---help---
+ Provides services to freeze and dump a back trace on panic
+ situations. This is used on IPIPE_DEBUG_CONTEXT exceptions
+ as well as ordinary kernel oopses. You can control the number
+ of printed back trace points via /proc/ipipe/trace.
+
+config IPIPE_TRACE_ENABLE_VALUE
+ int
+ default 0 if !IPIPE_TRACE_ENABLE
+ default 1 if IPIPE_TRACE_ENABLE
+
+endif
diff --git a/kernel/ipipe/Makefile b/kernel/ipipe/Makefile
new file mode 100644
index 0000000..6257dfa
--- /dev/null
+++ b/kernel/ipipe/Makefile
@@ -0,0 +1,3 @@
+
+obj-$(CONFIG_IPIPE) += core.o
+obj-$(CONFIG_IPIPE_TRACE) += tracer.o
diff --git a/kernel/ipipe/core.c b/kernel/ipipe/core.c
new file mode 100644
index 0000000..c59c4d4
--- /dev/null
+++ b/kernel/ipipe/core.c
@@ -0,0 +1,1631 @@
+/* -*- linux-c -*-
+ * linux/kernel/ipipe/core.c
+ *
+ * Copyright (C) 2002-2005 Philippe Gerum.
+ *
+ * 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, Inc., 675 Mass Ave, Cambridge MA 02139,
+ * USA; 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.
+ *
+ * Architecture-independent I-PIPE core support.
+ */
+
+#include <linux/version.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/sched.h>
+#include <linux/sched.h>
+#include <linux/kallsyms.h>
+#include <linux/interrupt.h>
+#include <linux/bitops.h>
+#include <linux/tick.h>
+#ifdef CONFIG_PROC_FS
+#include <linux/proc_fs.h>
+#include <linux/seq_file.h>
+#endif /* CONFIG_PROC_FS */
+#include <linux/ipipe_trace.h>
+#include <linux/ipipe_tickdev.h>
+#include <linux/irq.h>
+
+static int __ipipe_ptd_key_count;
+
+static unsigned long __ipipe_ptd_key_map;
+
+static unsigned long __ipipe_domain_slot_map;
+
+struct ipipe_domain ipipe_root;
+
+#ifndef CONFIG_SMP
+/*
+ * Create an alias to the unique root status, so that arch-dep code
+ * may get simple and easy access to this percpu variable. We also
+ * create an array of pointers to the percpu domain data; this tends
+ * to produce a better code when reaching non-root domains. We make
+ * sure that the early boot code would be able to dereference the
+ * pointer to the root domain data safely by statically initializing
+ * its value (local_irq*() routines depend on this).
+ */
+#if __GNUC__ >= 4
+extern unsigned long __ipipe_root_status
+__attribute__((alias(__stringify(__raw_get_cpu_var(ipipe_percpu_darray)))));
+EXPORT_SYMBOL(__ipipe_root_status);
+#else /* __GNUC__ < 4 */
+/*
+ * Work around a GCC 3.x issue making alias symbols unusable as
+ * constant initializers.
+ */
+unsigned long *const __ipipe_root_status_addr =
+ &__raw_get_cpu_var(ipipe_percpu_darray)[IPIPE_ROOT_SLOT].status;
+EXPORT_SYMBOL(__ipipe_root_status_addr);
+#endif /* __GNUC__ < 4 */
+
+DEFINE_PER_CPU(struct ipipe_percpu_domain_data *, ipipe_percpu_daddr[CONFIG_IPIPE_DOMAINS]) =
+{ [IPIPE_ROOT_SLOT] = (struct ipipe_percpu_domain_data *)&__raw_get_cpu_var(ipipe_percpu_darray) };
+EXPORT_PER_CPU_SYMBOL(ipipe_percpu_daddr);
+#endif /* !CONFIG_SMP */
+
+DEFINE_PER_CPU(struct ipipe_percpu_domain_data, ipipe_percpu_darray[CONFIG_IPIPE_DOMAINS]) =
+{ [IPIPE_ROOT_SLOT] = { .status = IPIPE_STALL_MASK } }; /* Root domain stalled on each CPU at startup. */
+
+DEFINE_PER_CPU(struct ipipe_domain *, ipipe_percpu_domain) = { &ipipe_root };
+
+static IPIPE_DEFINE_SPINLOCK(__ipipe_pipelock);
+
+LIST_HEAD(__ipipe_pipeline);
+
+unsigned long __ipipe_virtual_irq_map;
+
+#ifdef CONFIG_PRINTK
+unsigned __ipipe_printk_virq;
+#endif /* CONFIG_PRINTK */
+
+int __ipipe_event_monitors[IPIPE_NR_EVENTS];
+
+#ifdef CONFIG_GENERIC_CLOCKEVENTS
+
+DECLARE_PER_CPU(struct tick_device, tick_cpu_device);
+
+static DEFINE_PER_CPU(struct ipipe_tick_device, ipipe_tick_cpu_device);
+
+int ipipe_request_tickdev(const char *devname,
+ void (*emumode)(enum clock_event_mode mode,
+ struct clock_event_device *cdev),
+ int (*emutick)(unsigned long delta,
+ struct clock_event_device *cdev),
+ int cpu, unsigned long *tmfreq)
+{
+ struct ipipe_tick_device *itd;
+ struct tick_device *slave;
+ struct clock_event_device *evtdev;
+ unsigned long long freq;
+ unsigned long flags;
+ int status;
+
+ flags = ipipe_critical_enter(NULL);
+
+ itd = &per_cpu(ipipe_tick_cpu_device, cpu);
+
+ if (itd->slave != NULL) {
+ status = -EBUSY;
+ goto out;
+ }
+
+ slave = &per_cpu(tick_cpu_device, cpu);
+
+ if (strcmp(slave->evtdev->name, devname)) {
+ /*
+ * No conflict so far with the current tick device,
+ * check whether the requested device is sane and has
+ * been blessed by the kernel.
+ */
+ status = __ipipe_check_tickdev(devname) ?
+ CLOCK_EVT_MODE_UNUSED : CLOCK_EVT_MODE_SHUTDOWN;
+ goto out;
+ }
+
+ /*
+ * Our caller asks for using the same clock event device for
+ * ticking than we do, let's create a tick emulation device to
+ * interpose on the set_next_event() method, so that we may
+ * both manage the device in oneshot mode. Only the tick
+ * emulation code will actually program the clockchip hardware
+ * for the next shot, though.
+ *
+ * CAUTION: we still have to grab the tick device even when it
+ * current runs in periodic mode, since the kernel may switch
+ * to oneshot dynamically (highres/no_hz tick mode).
+ */
+
+ evtdev = slave->evtdev;
+ status = evtdev->mode;
+
+ if (status == CLOCK_EVT_MODE_SHUTDOWN)
+ goto out;
+
+ itd->slave = slave;
+ itd->emul_set_mode = emumode;
+ itd->emul_set_tick = emutick;
+ itd->real_set_mode = evtdev->set_mode;
+ itd->real_set_tick = evtdev->set_next_event;
+ itd->real_max_delta_ns = evtdev->max_delta_ns;
+ itd->real_mult = evtdev->mult;
+ itd->real_shift = evtdev->shift;
+ freq = (1000000000ULL * evtdev->mult) >> evtdev->shift;
+ *tmfreq = (unsigned long)freq;
+ evtdev->set_mode = emumode;
+ evtdev->set_next_event = emutick;
+ evtdev->max_delta_ns = ULONG_MAX;
+ evtdev->mult = 1;
+ evtdev->shift = 0;
+out:
+ ipipe_critical_exit(flags);
+
+ return status;
+}
+
+void ipipe_release_tickdev(int cpu)
+{
+ struct ipipe_tick_device *itd;
+ struct tick_device *slave;
+ struct clock_event_device *evtdev;
+ unsigned long flags;
+
+ flags = ipipe_critical_enter(NULL);
+
+ itd = &per_cpu(ipipe_tick_cpu_device, cpu);
+
+ if (itd->slave != NULL) {
+ slave = &per_cpu(tick_cpu_device, cpu);
+ evtdev = slave->evtdev;
+ evtdev->set_mode = itd->real_set_mode;
+ evtdev->set_next_event = itd->real_set_tick;
+ evtdev->max_delta_ns = itd->real_max_delta_ns;
+ evtdev->mult = itd->real_mult;
+ evtdev->shift = itd->real_shift;
+ itd->slave = NULL;
+ }
+
+ ipipe_critical_exit(flags);
+}
+
+#endif /* CONFIG_GENERIC_CLOCKEVENTS */
+
+/*
+ * ipipe_init() -- Initialization routine of the IPIPE layer. Called
+ * by the host kernel early during the boot procedure.
+ */
+void __init ipipe_init(void)
+{
+ struct ipipe_domain *ipd = &ipipe_root;
+
+ __ipipe_check_platform(); /* Do platform dependent checks first. */
+
+ /*
+ * A lightweight registration code for the root domain. We are
+ * running on the boot CPU, hw interrupts are off, and
+ * secondary CPUs are still lost in space.
+ */
+
+ /* Reserve percpu data slot #0 for the root domain. */
+ ipd->slot = 0;
+ set_bit(0, &__ipipe_domain_slot_map);
+
+ ipd->name = "Linux";
+ ipd->domid = IPIPE_ROOT_ID;
+ ipd->priority = IPIPE_ROOT_PRIO;
+
+ __ipipe_init_stage(ipd);
+
+ INIT_LIST_HEAD(&ipd->p_link);
+ list_add_tail(&ipd->p_link, &__ipipe_pipeline);
+
+ __ipipe_init_platform();
+
+#ifdef CONFIG_PRINTK
+ __ipipe_printk_virq = ipipe_alloc_virq(); /* Cannot fail here. */
+ ipd->irqs[__ipipe_printk_virq].handler = &__ipipe_flush_printk;
+ ipd->irqs[__ipipe_printk_virq].cookie = NULL;
+ ipd->irqs[__ipipe_printk_virq].acknowledge = NULL;
+ ipd->irqs[__ipipe_printk_virq].control = IPIPE_HANDLE_MASK;
+#endif /* CONFIG_PRINTK */
+
+ __ipipe_enable_pipeline();
+
+ printk(KERN_INFO "I-pipe %s: pipeline enabled.\n",
+ IPIPE_VERSION_STRING);
+}
+
+void __ipipe_init_stage(struct ipipe_domain *ipd)
+{
+ int cpu, n;
+
+ for_each_online_cpu(cpu) {
+
+ ipipe_percpudom(ipd, irqpend_himask, cpu) = 0;
+
+ for (n = 0; n < IPIPE_IRQ_IWORDS; n++) {
+ ipipe_percpudom(ipd, irqpend_lomask, cpu)[n] = 0;
+ ipipe_percpudom(ipd, irqheld_mask, cpu)[n] = 0;
+ }
+
+ for (n = 0; n < IPIPE_NR_IRQS; n++)
+ ipipe_percpudom(ipd, irqall, cpu)[n] = 0;
+
+ ipipe_percpudom(ipd, evsync, cpu) = 0;
+ }
+
+ for (n = 0; n < IPIPE_NR_IRQS; n++) {
+ ipd->irqs[n].acknowledge = NULL;
+ ipd->irqs[n].handler = NULL;
+ ipd->irqs[n].control = IPIPE_PASS_MASK; /* Pass but don't handle */
+ }
+
+ for (n = 0; n < IPIPE_NR_EVENTS; n++)
+ ipd->evhand[n] = NULL;
+
+ ipd->evself = 0LL;
+ mutex_init(&ipd->mutex);
+
+ __ipipe_hook_critical_ipi(ipd);
+}
+
+void __ipipe_cleanup_domain(struct ipipe_domain *ipd)
+{
+ ipipe_unstall_pipeline_from(ipd);
+
+#ifdef CONFIG_SMP
+ {
+ int cpu;
+
+ for_each_online_cpu(cpu) {
+ while (ipipe_percpudom(ipd, irqpend_himask, cpu) != 0)
+ cpu_relax();
+ }
+ }
+#else
+ __raw_get_cpu_var(ipipe_percpu_daddr)[ipd->slot] = NULL;
+#endif
+
+ clear_bit(ipd->slot, &__ipipe_domain_slot_map);
+}
+
+void __ipipe_unstall_root(void)
+{
+#ifndef CONFIG_IPIPE_DEBUG_CONTEXT
+ BUG_ON(!ipipe_root_domain_p);
+#endif
+ local_irq_disable_hw();
+
+ __clear_bit(IPIPE_STALL_FLAG, &ipipe_root_cpudom_var(status));
+
+ if (unlikely(ipipe_root_cpudom_var(irqpend_himask) != 0))
+ __ipipe_sync_pipeline(IPIPE_IRQMASK_ANY);
+
+ local_irq_enable_hw();
+}
+
+void __ipipe_restore_root(unsigned long x)
+{
+#ifndef CONFIG_IPIPE_DEBUG_CONTEXT
+ BUG_ON(!ipipe_root_domain_p);
+#endif
+
+ if (x)
+ __ipipe_stall_root();
+ else
+ __ipipe_unstall_root();
+}
+
+void ipipe_stall_pipeline_from(struct ipipe_domain *ipd)
+{
+ set_bit_safe(IPIPE_STALL_FLAG, &ipipe_cpudom_var(ipd, status));
+
+ if (__ipipe_pipeline_head_p(ipd))
+ local_irq_disable_hw();
+}
+
+unsigned long ipipe_test_and_stall_pipeline_from(struct ipipe_domain *ipd)
+{
+ unsigned long x;
+
+ x = test_and_set_bit_safe(IPIPE_STALL_FLAG, &ipipe_cpudom_var(ipd, status));
+
+ if (__ipipe_pipeline_head_p(ipd))
+ local_irq_disable_hw();
+
+ return x;
+}
+
+/*
+ * ipipe_unstall_pipeline_from() -- Unstall the pipeline and
+ * synchronize pending interrupts for a given domain. See
+ * __ipipe_walk_pipeline() for more information.
+ */
+void ipipe_unstall_pipeline_from(struct ipipe_domain *ipd)
+{
+ struct list_head *pos;
+ unsigned long flags;
+
+ local_irq_save_hw(flags);
+
+ __clear_bit(IPIPE_STALL_FLAG, &ipipe_cpudom_var(ipd, status));
+
+ if (ipd == ipipe_current_domain)
+ pos = &ipd->p_link;
+ else
+ pos = __ipipe_pipeline.next;
+
+ __ipipe_walk_pipeline(pos);
+
+ if (likely(__ipipe_pipeline_head_p(ipd)))
+ local_irq_enable_hw();
+ else
+ local_irq_restore_hw(flags);
+}
+
+unsigned long ipipe_test_and_unstall_pipeline_from(struct ipipe_domain *ipd)
+{
+ unsigned long x;
+
+ x = test_bit(IPIPE_STALL_FLAG, &ipipe_cpudom_var(ipd, status));
+ ipipe_unstall_pipeline_from(ipd);
+
+ return x;
+}
+
+void ipipe_restore_pipeline_from(struct ipipe_domain *ipd,
+ unsigned long x)
+{
+ if (x)
+ ipipe_stall_pipeline_from(ipd);
+ else
+ ipipe_unstall_pipeline_from(ipd);
+}
+
+void ipipe_unstall_pipeline_head(void)
+{
+ local_irq_disable_hw();
+
+ __clear_bit(IPIPE_STALL_FLAG, &ipipe_head_cpudom_var(status));
+
+ if (unlikely(ipipe_head_cpudom_var(irqpend_himask) != 0)) {
+ struct ipipe_domain *head_domain = __ipipe_pipeline_head();
+ if (likely(head_domain == ipipe_current_domain))
+ __ipipe_sync_pipeline(IPIPE_IRQMASK_ANY);
+ else
+ __ipipe_walk_pipeline(&head_domain->p_link);
+ }
+
+ local_irq_enable_hw();
+}
+
+void __ipipe_restore_pipeline_head(unsigned long x)
+{
+ local_irq_disable_hw();
+
+ if (x) {
+#ifdef CONFIG_DEBUG_KERNEL
+ static int warned;
+ if (!warned && test_and_set_bit(IPIPE_STALL_FLAG, &ipipe_head_cpudom_var(status))) {
+ /*
+ * Already stalled albeit ipipe_restore_pipeline_head()
+ * should have detected it? Send a warning once.
+ */
+ warned = 1;
+ printk(KERN_WARNING
+ "I-pipe: ipipe_restore_pipeline_head() optimization failed.\n");
+ dump_stack();
+ }
+#else /* !CONFIG_DEBUG_KERNEL */
+ set_bit(IPIPE_STALL_FLAG, &ipipe_head_cpudom_var(status));
+#endif /* CONFIG_DEBUG_KERNEL */
+ }
+ else {
+ __clear_bit(IPIPE_STALL_FLAG, &ipipe_head_cpudom_var(status));
+ if (unlikely(ipipe_head_cpudom_var(irqpend_himask) != 0)) {
+ struct ipipe_domain *head_domain = __ipipe_pipeline_head();
+ if (likely(head_domain == ipipe_current_domain))
+ __ipipe_sync_pipeline(IPIPE_IRQMASK_ANY);
+ else
+ __ipipe_walk_pipeline(&head_domain->p_link);
+ }
+ local_irq_enable_hw();
+ }
+}
+
+void __ipipe_spin_lock_irq(raw_spinlock_t *lock)
+{
+ local_irq_disable_hw();
+ __raw_spin_lock(lock);
+ __set_bit(IPIPE_STALL_FLAG, &ipipe_this_cpudom_var(status));
+}
+
+void __ipipe_spin_unlock_irq(raw_spinlock_t *lock)
+{
+ __raw_spin_unlock(lock);
+ __clear_bit(IPIPE_STALL_FLAG, &ipipe_this_cpudom_var(status));
+ local_irq_enable_hw();
+}
+
+unsigned long __ipipe_spin_lock_irqsave(raw_spinlock_t *lock)
+{
+ unsigned long flags;
+ int s;
+
+ local_irq_save_hw(flags);
+ __raw_spin_lock(lock);
+ s = __test_and_set_bit(IPIPE_STALL_FLAG, &ipipe_this_cpudom_var(status));
+
+ return raw_mangle_irq_bits(s, flags);
+}
+
+void __ipipe_spin_unlock_irqrestore(raw_spinlock_t *lock, unsigned long x)
+{
+ __raw_spin_unlock(lock);
+ if (!raw_demangle_irq_bits(&x))
+ __clear_bit(IPIPE_STALL_FLAG, &ipipe_this_cpudom_var(status));
+ local_irq_restore_hw(x);
+}
+
+void __ipipe_spin_unlock_irqbegin(ipipe_spinlock_t *lock)
+{
+ __raw_spin_unlock(&lock->__raw_lock);
+}
+
+void __ipipe_spin_unlock_irqcomplete(unsigned long x)
+{
+ if (!raw_demangle_irq_bits(&x))
+ __clear_bit(IPIPE_STALL_FLAG, &ipipe_this_cpudom_var(status));
+ local_irq_restore_hw(x);
+}
+
+/* Must be called hw IRQs off. */
+void __ipipe_set_irq_pending(struct ipipe_domain *ipd, unsigned irq)
+{
+ int level = irq >> IPIPE_IRQ_ISHIFT, rank = irq & IPIPE_IRQ_IMASK;
+
+ if (likely(!test_bit(IPIPE_LOCK_FLAG, &ipd->irqs[irq].control))) {
+ __set_bit(rank, &ipipe_cpudom_var(ipd, irqpend_lomask)[level]);
+ __set_bit(level,&ipipe_cpudom_var(ipd, irqpend_himask));
+ } else
+ __set_bit(rank, &ipipe_cpudom_var(ipd, irqheld_mask)[level]);
+
+ ipipe_cpudom_var(ipd, irqall)[irq]++;
+}
+
+/* Must be called hw IRQs off. */
+void __ipipe_lock_irq(struct ipipe_domain *ipd, int cpu, unsigned irq)
+{
+ if (likely(!test_and_set_bit(IPIPE_LOCK_FLAG, &ipd->irqs[irq].control))) {
+ int level = irq >> IPIPE_IRQ_ISHIFT, rank = irq & IPIPE_IRQ_IMASK;
+ if (__test_and_clear_bit(rank, &ipipe_percpudom(ipd, irqpend_lomask, cpu)[level]))
+ __set_bit(rank, &ipipe_cpudom_var(ipd, irqheld_mask)[level]);
+ if (ipipe_percpudom(ipd, irqpend_lomask, cpu)[level] == 0)
+ __clear_bit(level, &ipipe_percpudom(ipd, irqpend_himask, cpu));
+ }
+}
+
+/* Must be called hw IRQs off. */
+void __ipipe_unlock_irq(struct ipipe_domain *ipd, unsigned irq)
+{
+ int cpu;
+
+ if (likely(test_and_clear_bit(IPIPE_LOCK_FLAG, &ipd->irqs[irq].control))) {
+ int level = irq >> IPIPE_IRQ_ISHIFT, rank = irq & IPIPE_IRQ_IMASK;
+ for_each_online_cpu(cpu) {
+ if (test_and_clear_bit(rank, &ipipe_percpudom(ipd, irqheld_mask, cpu)[level])) {
+ /* We need atomic ops here: */
+ set_bit(rank, &ipipe_percpudom(ipd, irqpend_lomask, cpu)[level]);
+ set_bit(level, &ipipe_percpudom(ipd, irqpend_himask, cpu));
+ }
+ }
+ }
+}
+
+/* __ipipe_walk_pipeline(): Plays interrupts pending in the log. Must
+ be called with local hw interrupts disabled. */
+
+void __ipipe_walk_pipeline(struct list_head *pos)
+{
+ struct ipipe_domain *this_domain = ipipe_current_domain, *next_domain;
+
+ while (pos != &__ipipe_pipeline) {
+
+ next_domain = list_entry(pos, struct ipipe_domain, p_link);
+
+ if (test_bit(IPIPE_STALL_FLAG, &ipipe_cpudom_var(next_domain, status)))
+ break; /* Stalled stage -- do not go further. */
+
+ if (ipipe_cpudom_var(next_domain, irqpend_himask) != 0) {
+
+ if (next_domain == this_domain)
+ __ipipe_sync_pipeline(IPIPE_IRQMASK_ANY);
+ else {
+
+ ipipe_cpudom_var(this_domain, evsync) = 0;
+ ipipe_current_domain = next_domain;
+ ipipe_suspend_domain(); /* Sync stage and propagate interrupts. */
+
+ if (ipipe_current_domain == next_domain)
+ ipipe_current_domain = this_domain;
+ /*
+ * Otherwise, something changed the current domain under our
+ * feet recycling the register set; do not override the new
+ * domain.
+ */
+
+ if (ipipe_cpudom_var(this_domain, irqpend_himask) != 0 &&
+ !test_bit(IPIPE_STALL_FLAG,
+ &ipipe_cpudom_var(this_domain, status)))
+ __ipipe_sync_pipeline(IPIPE_IRQMASK_ANY);
+ }
+
+ break;
+ } else if (next_domain == this_domain)
+ break;
+
+ pos = next_domain->p_link.next;
+ }
+}
+
+/*
+ * ipipe_suspend_domain() -- Suspend the current domain, switching to
+ * the next one which has pending work down the pipeline.
+ */
+void ipipe_suspend_domain(void)
+{
+ struct ipipe_domain *this_domain, *next_domain;
+ struct list_head *ln;
+ unsigned long flags;
+
+ local_irq_save_hw(flags);
+
+ this_domain = next_domain = ipipe_current_domain;
+
+ __clear_bit(IPIPE_STALL_FLAG, &ipipe_cpudom_var(this_domain, status));
+
+ if (ipipe_cpudom_var(this_domain, irqpend_himask) != 0)
+ goto sync_stage;
+
+ for (;;) {
+ ln = next_domain->p_link.next;
+
+ if (ln == &__ipipe_pipeline)
+ break;
+
+ next_domain = list_entry(ln, struct ipipe_domain, p_link);
+
+ if (test_bit(IPIPE_STALL_FLAG,
+ &ipipe_cpudom_var(next_domain, status)) != 0)
+ break;
+
+ if (ipipe_cpudom_var(next_domain, irqpend_himask) == 0)
+ continue;
+
+ ipipe_current_domain = next_domain;
+
+sync_stage:
+ __ipipe_sync_pipeline(IPIPE_IRQMASK_ANY);
+
+ if (ipipe_current_domain != next_domain)
+ /*
+ * Something has changed the current domain under our
+ * feet, recycling the register set; take note.
+ */
+ this_domain = ipipe_current_domain;
+ }
+
+ ipipe_current_domain = this_domain;
+
+ local_irq_restore_hw(flags);
+}
+
+/* ipipe_alloc_virq() -- Allocate a pipelined virtual/soft interrupt.
+ * Virtual interrupts are handled in exactly the same way than their
+ * hw-generated counterparts wrt pipelining.
+ */
+unsigned ipipe_alloc_virq(void)
+{
+ unsigned long flags, irq = 0;
+ int ipos;
+
+ spin_lock_irqsave(&__ipipe_pipelock, flags);
+
+ if (__ipipe_virtual_irq_map != ~0) {
+ ipos = ffz(__ipipe_virtual_irq_map);
+ set_bit(ipos, &__ipipe_virtual_irq_map);
+ irq = ipos + IPIPE_VIRQ_BASE;
+ }
+
+ spin_unlock_irqrestore(&__ipipe_pipelock, flags);
+
+ return irq;
+}
+
+/* ipipe_virtualize_irq() -- Attach a handler (and optionally a hw
+ acknowledge routine) to an interrupt for a given domain. */
+
+int ipipe_virtualize_irq(struct ipipe_domain *ipd,
+ unsigned irq,
+ ipipe_irq_handler_t handler,
+ void *cookie,
+ ipipe_irq_ackfn_t acknowledge,
+ unsigned modemask)
+{
+ unsigned long flags;
+ int err;
+
+ if (irq >= IPIPE_NR_IRQS)
+ return -EINVAL;
+
+ if (ipd->irqs[irq].control & IPIPE_SYSTEM_MASK)
+ return -EPERM;
+
+ if (!test_bit(IPIPE_AHEAD_FLAG, &ipd->flags))
+ /* Silently unwire interrupts for non-heading domains. */
+ modemask &= ~IPIPE_WIRED_MASK;
+
+ spin_lock_irqsave(&__ipipe_pipelock, flags);
+
+ if (handler != NULL) {
+ if (handler == IPIPE_SAME_HANDLER) {
+ handler = ipd->irqs[irq].handler;
+ cookie = ipd->irqs[irq].cookie;
+
+ if (handler == NULL) {
+ err = -EINVAL;
+ goto unlock_and_exit;
+ }
+ } else if ((modemask & IPIPE_EXCLUSIVE_MASK) != 0 &&
+ ipd->irqs[irq].handler != NULL) {
+ err = -EBUSY;
+ goto unlock_and_exit;
+ }
+
+ /* Wired interrupts can only be delivered to domains
+ * always heading the pipeline, and using dynamic
+ * propagation. */
+
+ if ((modemask & IPIPE_WIRED_MASK) != 0) {
+ if ((modemask & (IPIPE_PASS_MASK | IPIPE_STICKY_MASK)) != 0) {
+ err = -EINVAL;
+ goto unlock_and_exit;
+ }
+ modemask |= (IPIPE_HANDLE_MASK);
+ }
+
+ if ((modemask & IPIPE_STICKY_MASK) != 0)
+ modemask |= IPIPE_HANDLE_MASK;
+ } else
+ modemask &=
+ ~(IPIPE_HANDLE_MASK | IPIPE_STICKY_MASK |
+ IPIPE_EXCLUSIVE_MASK | IPIPE_WIRED_MASK);
+
+ if (acknowledge == NULL && !ipipe_virtual_irq_p(irq))
+ /* Acknowledge handler unspecified for a hw interrupt:
+ use the Linux-defined handler instead. */
+ acknowledge = ipipe_root_domain->irqs[irq].acknowledge;
+
+ ipd->irqs[irq].handler = handler;
+ ipd->irqs[irq].cookie = cookie;
+ ipd->irqs[irq].acknowledge = acknowledge;
+ ipd->irqs[irq].control = modemask;
+
+ if (irq < NR_IRQS && handler != NULL && !ipipe_virtual_irq_p(irq)) {
+ __ipipe_enable_irqdesc(ipd, irq);
+
+ if ((modemask & IPIPE_ENABLE_MASK) != 0) {
+ if (ipd != ipipe_current_domain) {
+ /* IRQ enable/disable state is domain-sensitive, so we may
+ not change it for another domain. What is allowed
+ however is forcing some domain to handle an interrupt
+ source, by passing the proper 'ipd' descriptor which
+ thus may be different from ipipe_current_domain. */
+ err = -EPERM;
+ goto unlock_and_exit;
+ }
+ __ipipe_enable_irq(irq);
+ }
+ }
+
+ err = 0;
+
+ unlock_and_exit:
+
+ spin_unlock_irqrestore(&__ipipe_pipelock, flags);
+
+ return err;
+}
+
+/* ipipe_control_irq() -- Change modes of a pipelined interrupt for
+ * the current domain. */
+
+int ipipe_control_irq(unsigned irq, unsigned clrmask, unsigned setmask)
+{
+ struct ipipe_domain *ipd;
+ unsigned long flags;
+
+ if (irq >= IPIPE_NR_IRQS)
+ return -EINVAL;
+
+ ipd = ipipe_current_domain;
+
+ if (ipd->irqs[irq].control & IPIPE_SYSTEM_MASK)
+ return -EPERM;
+
+ if (ipd->irqs[irq].handler == NULL)
+ setmask &= ~(IPIPE_HANDLE_MASK | IPIPE_STICKY_MASK);
+
+ if ((setmask & IPIPE_STICKY_MASK) != 0)
+ setmask |= IPIPE_HANDLE_MASK;
+
+ if ((clrmask & (IPIPE_HANDLE_MASK | IPIPE_STICKY_MASK)) != 0) /* If one goes, both go. */
+ clrmask |= (IPIPE_HANDLE_MASK | IPIPE_STICKY_MASK);
+
+ spin_lock_irqsave(&__ipipe_pipelock, flags);
+
+ ipd->irqs[irq].control &= ~clrmask;
+ ipd->irqs[irq].control |= setmask;
+
+ if ((setmask & IPIPE_ENABLE_MASK) != 0)
+ __ipipe_enable_irq(irq);
+ else if ((clrmask & IPIPE_ENABLE_MASK) != 0)
+ __ipipe_disable_irq(irq);
+
+ spin_unlock_irqrestore(&__ipipe_pipelock, flags);
+
+ return 0;
+}
+
+/* __ipipe_dispatch_event() -- Low-level event dispatcher. */
+
+int __ipipe_dispatch_event (unsigned event, void *data)
+{
+ struct ipipe_domain *start_domain, *this_domain, *next_domain;
+ ipipe_event_handler_t evhand;
+ struct list_head *pos, *npos;
+ unsigned long flags;
+ int propagate = 1;
+
+ local_irq_save_hw(flags);
+
+ start_domain = this_domain = ipipe_current_domain;
+
+ list_for_each_safe(pos, npos, &__ipipe_pipeline) {
+ /*
+ * Note: Domain migration may occur while running
+ * event or interrupt handlers, in which case the
+ * current register set is going to be recycled for a
+ * different domain than the initiating one. We do
+ * care for that, always tracking the current domain
+ * descriptor upon return from those handlers.
+ */
+ next_domain = list_entry(pos, struct ipipe_domain, p_link);
+
+ /*
+ * Keep a cached copy of the handler's address since
+ * ipipe_catch_event() may clear it under our feet.
+ */
+ evhand = next_domain->evhand[event];
+
+ if (evhand != NULL) {
+ ipipe_current_domain = next_domain;
+ ipipe_cpudom_var(next_domain, evsync) |= (1LL << event);
+ local_irq_restore_hw(flags);
+ propagate = !evhand(event, start_domain, data);
+ local_irq_save_hw(flags);
+ ipipe_cpudom_var(next_domain, evsync) &= ~(1LL << event);
+ if (ipipe_current_domain != next_domain)
+ this_domain = ipipe_current_domain;
+ }
+
+ if (next_domain != ipipe_root_domain && /* NEVER sync the root stage here. */
+ ipipe_cpudom_var(next_domain, irqpend_himask) != 0 &&
+ !test_bit(IPIPE_STALL_FLAG, &ipipe_cpudom_var(next_domain, status))) {
+ ipipe_current_domain = next_domain;
+ __ipipe_sync_pipeline(IPIPE_IRQMASK_ANY);
+ if (ipipe_current_domain != next_domain)
+ this_domain = ipipe_current_domain;
+ }
+
+ ipipe_current_domain = this_domain;
+
+ if (next_domain == this_domain || !propagate)
+ break;
+ }
+
+ local_irq_restore_hw(flags);
+
+ return !propagate;
+}
+
+/*
+ * __ipipe_dispatch_wired -- Wired interrupt dispatcher. Wired
+ * interrupts are immediately and unconditionally delivered to the
+ * domain heading the pipeline upon receipt, and such domain must have
+ * been registered as an invariant head for the system (priority ==
+ * IPIPE_HEAD_PRIORITY). The motivation for using wired interrupts is
+ * to get an extra-fast dispatching path for those IRQs, by relying on
+ * a straightforward logic based on assumptions that must always be
+ * true for invariant head domains. The following assumptions are
+ * made when dealing with such interrupts:
+ *
+ * 1- Wired interrupts are purely dynamic, i.e. the decision to
+ * propagate them down the pipeline must be done from the head domain
+ * ISR.
+ * 2- Wired interrupts cannot be shared or sticky.
+ * 3- The root domain cannot be an invariant pipeline head, in
+ * consequence of what the root domain cannot handle wired
+ * interrupts.
+ * 4- Wired interrupts must have a valid acknowledge handler for the
+ * head domain (if needed), and in any case, must not rely on handlers
+ * provided by lower priority domains during the acknowledge cycle
+ * (see __ipipe_handle_irq).
+ *
+ * Called with hw interrupts off.
+ */
+
+int __ipipe_dispatch_wired(struct ipipe_domain *head_domain, unsigned irq)
+{
+ struct ipipe_domain *old;
+
+ if (test_bit(IPIPE_LOCK_FLAG, &head_domain->irqs[irq].control)) {
+ /* If we can't process this IRQ right now, we must
+ * mark it as held, so that it will get played during
+ * normal log sync when the corresponding interrupt
+ * source is eventually unlocked. */
+ ipipe_cpudom_var(head_domain, irqall)[irq]++;
+ __set_bit(irq & IPIPE_IRQ_IMASK, &ipipe_cpudom_var(head_domain, irqheld_mask)[irq >> IPIPE_IRQ_ISHIFT]);
+ return 0;
+ }
+
+ if (test_bit(IPIPE_STALL_FLAG, &ipipe_cpudom_var(head_domain, status))) {
+ __ipipe_set_irq_pending(head_domain, irq);
+ return 0;
+ }
+
+ old = ipipe_current_domain;
+ ipipe_current_domain = head_domain; /* Switch to the head domain. */
+
+ ipipe_cpudom_var(head_domain, irqall)[irq]++;
+ __set_bit(IPIPE_STALL_FLAG, &ipipe_cpudom_var(head_domain, status));
+ head_domain->irqs[irq].handler(irq, head_domain->irqs[irq].cookie); /* Call the ISR. */
+ __ipipe_run_irqtail();
+ __clear_bit(IPIPE_STALL_FLAG, &ipipe_cpudom_var(head_domain, status));
+
+ /* We expect the caller to start a complete pipeline walk upon
+ * return, so that propagated interrupts will get played. */
+
+ if (ipipe_current_domain == head_domain)
+ ipipe_current_domain = old; /* Back to the preempted domain. */
+
+ return 1;
+}
+
+/*
+ * __ipipe_sync_stage() -- Flush the pending IRQs for the current
+ * domain (and processor). This routine flushes the interrupt log
+ * (see "Optimistic interrupt protection" from D. Stodolsky et al. for
+ * more on the deferred interrupt scheme). Every interrupt that
+ * occurred while the pipeline was stalled gets played. WARNING:
+ * callers on SMP boxen should always check for CPU migration on
+ * return of this routine. One can control the kind of interrupts
+ * which are going to be sync'ed using the syncmask
+ * parameter. IPIPE_IRQMASK_ANY plays them all, IPIPE_IRQMASK_VIRT
+ * plays virtual interrupts only.
+ *
+ * This routine must be called with hw interrupts off.
+ */
+void __ipipe_sync_stage(unsigned long syncmask)
+{
+ unsigned long mask, submask;
+ struct ipipe_domain *ipd;
+ int level, rank, cpu;
+ unsigned irq;
+
+ if (__test_and_set_bit(IPIPE_SYNC_FLAG, &ipipe_this_cpudom_var(status)))
+ return;
+
+ ipd = ipipe_current_domain;
+ cpu = ipipe_processor_id();
+
+ /*
+ * The policy here is to keep the dispatching code interrupt-free
+ * by stalling the current stage. If the upper domain handler
+ * (which we call) wants to re-enable interrupts while in a safe
+ * portion of the code (e.g. SA_INTERRUPT flag unset for Linux's
+ * sigaction()), it will have to unstall (then stall again before
+ * returning to us!) the stage when it sees fit.
+ */
+ while ((mask = (ipipe_this_cpudom_var(irqpend_himask) & syncmask)) != 0) {
+ level = __ipipe_ffnz(mask);
+
+ while ((submask = ipipe_this_cpudom_var(irqpend_lomask)[level]) != 0) {
+ rank = __ipipe_ffnz(submask);
+ irq = (level << IPIPE_IRQ_ISHIFT) + rank;
+
+ __clear_bit(rank, &ipipe_this_cpudom_var(irqpend_lomask)[level]);
+
+ if (ipipe_this_cpudom_var(irqpend_lomask)[level] == 0)
+ __clear_bit(level, &ipipe_this_cpudom_var(irqpend_himask));
+
+ if (test_bit(IPIPE_LOCK_FLAG, &ipd->irqs[irq].control))
+ continue;
+
+ __set_bit(IPIPE_STALL_FLAG, &ipipe_this_cpudom_var(status));
+
+ if (ipd == ipipe_root_domain)
+ trace_hardirqs_off();
+
+ __ipipe_run_isr(ipd, irq);
+#ifdef CONFIG_SMP
+ {
+ int newcpu = ipipe_processor_id();
+
+ if (newcpu != cpu) { /* Handle CPU migration. */
+ /*
+ * We expect any domain to clear the SYNC bit each
+ * time it switches in a new task, so that preemptions
+ * and/or CPU migrations (in the SMP case) over the
+ * ISR do not lock out the log syncer for some
+ * indefinite amount of time. In the Linux case,
+ * schedule() handles this (see kernel/sched.c). For
+ * this reason, we don't bother clearing it here for
+ * the source CPU in the migration handling case,
+ * since it must have scheduled another task in by
+ * now.
+ */
+ __set_bit(IPIPE_SYNC_FLAG, &ipipe_this_cpudom_var(status));
+ cpu = newcpu;
+ }
+ }
+#endif /* CONFIG_SMP */
+ if (ipd == ipipe_root_domain &&
+ test_bit(IPIPE_STALL_FLAG, &ipipe_this_cpudom_var(status)))
+ trace_hardirqs_on();
+
+ __clear_bit(IPIPE_STALL_FLAG, &ipipe_this_cpudom_var(status));
+ }
+ }
+
+ __clear_bit(IPIPE_SYNC_FLAG, &ipipe_this_cpudom_var(status));
+}
+
+/* ipipe_register_domain() -- Link a new domain to the pipeline. */
+
+int ipipe_register_domain(struct ipipe_domain *ipd,
+ struct ipipe_domain_attr *attr)
+{
+ struct ipipe_domain *_ipd;
+ struct list_head *pos = NULL;
+ unsigned long flags;
+
+ if (!ipipe_root_domain_p) {
+ printk(KERN_WARNING
+ "I-pipe: Only the root domain may register a new domain.\n");
+ return -EPERM;
+ }
+
+ flags = ipipe_critical_enter(NULL);
+
+ if (attr->priority == IPIPE_HEAD_PRIORITY) {
+ if (test_bit(IPIPE_HEAD_SLOT, &__ipipe_domain_slot_map)) {
+ ipipe_critical_exit(flags);
+ return -EAGAIN; /* Cannot override current head. */
+ }
+ ipd->slot = IPIPE_HEAD_SLOT;
+ } else
+ ipd->slot = ffz(__ipipe_domain_slot_map);
+
+ if (ipd->slot < CONFIG_IPIPE_DOMAINS) {
+ set_bit(ipd->slot, &__ipipe_domain_slot_map);
+ list_for_each(pos, &__ipipe_pipeline) {
+ _ipd = list_entry(pos, struct ipipe_domain, p_link);
+ if (_ipd->domid == attr->domid)
+ break;
+ }
+ }
+
+ ipipe_critical_exit(flags);
+
+ if (pos != &__ipipe_pipeline) {
+ if (ipd->slot < CONFIG_IPIPE_DOMAINS)
+ clear_bit(ipd->slot, &__ipipe_domain_slot_map);
+ return -EBUSY;
+ }
+
+#ifndef CONFIG_SMP
+ /*
+ * Set up the perdomain pointers for direct access to the
+ * percpu domain data. This saves a costly multiply each time
+ * we need to refer to the contents of the percpu domain data
+ * array.
+ */
+ __raw_get_cpu_var(ipipe_percpu_daddr)[ipd->slot] = &__raw_get_cpu_var(ipipe_percpu_darray)[ipd->slot];
+#endif
+
+ ipd->name = attr->name;
+ ipd->domid = attr->domid;
+ ipd->pdd = attr->pdd;
+ ipd->flags = 0;
+
+ if (attr->priority == IPIPE_HEAD_PRIORITY) {
+ ipd->priority = INT_MAX;
+ __set_bit(IPIPE_AHEAD_FLAG,&ipd->flags);
+ }
+ else
+ ipd->priority = attr->priority;
+
+ __ipipe_init_stage(ipd);
+
+ INIT_LIST_HEAD(&ipd->p_link);
+
+#ifdef CONFIG_PROC_FS
+ __ipipe_add_domain_proc(ipd);
+#endif /* CONFIG_PROC_FS */
+
+ flags = ipipe_critical_enter(NULL);
+
+ list_for_each(pos, &__ipipe_pipeline) {
+ _ipd = list_entry(pos, struct ipipe_domain, p_link);
+ if (ipd->priority > _ipd->priority)
+ break;
+ }
+
+ list_add_tail(&ipd->p_link, pos);
+
+ ipipe_critical_exit(flags);
+
+ printk(KERN_INFO "I-pipe: Domain %s registered.\n", ipd->name);
+
+ /*
+ * Finally, allow the new domain to perform its initialization
+ * chores.
+ */
+
+ if (attr->entry != NULL) {
+ ipipe_current_domain = ipd;
+ attr->entry();
+ ipipe_current_domain = ipipe_root_domain;
+
+ local_irq_save_hw(flags);
+
+ if (ipipe_root_cpudom_var(irqpend_himask) != 0 &&
+ !test_bit(IPIPE_STALL_FLAG, &ipipe_root_cpudom_var(status)))
+ __ipipe_sync_pipeline(IPIPE_IRQMASK_ANY);
+
+ local_irq_restore_hw(flags);
+ }
+
+ return 0;
+}
+
+/* ipipe_unregister_domain() -- Remove a domain from the pipeline. */
+
+int ipipe_unregister_domain(struct ipipe_domain *ipd)
+{
+ unsigned long flags;
+
+ if (!ipipe_root_domain_p) {
+ printk(KERN_WARNING
+ "I-pipe: Only the root domain may unregister a domain.\n");
+ return -EPERM;
+ }
+
+ if (ipd == ipipe_root_domain) {
+ printk(KERN_WARNING
+ "I-pipe: Cannot unregister the root domain.\n");
+ return -EPERM;
+ }
+#ifdef CONFIG_SMP
+ {
+ unsigned irq;
+ int cpu;
+
+ /*
+ * In the SMP case, wait for the logged events to drain on
+ * other processors before eventually removing the domain
+ * from the pipeline.
+ */
+
+ ipipe_unstall_pipeline_from(ipd);
+
+ flags = ipipe_critical_enter(NULL);
+
+ for (irq = 0; irq < IPIPE_NR_IRQS; irq++) {
+ clear_bit(IPIPE_HANDLE_FLAG, &ipd->irqs[irq].control);
+ clear_bit(IPIPE_STICKY_FLAG, &ipd->irqs[irq].control);
+ set_bit(IPIPE_PASS_FLAG, &ipd->irqs[irq].control);
+ }
+
+ ipipe_critical_exit(flags);
+
+ for_each_online_cpu(cpu) {
+ while (ipipe_percpudom(ipd, irqpend_himask, cpu) > 0)
+ cpu_relax();
+ }
+ }
+#endif /* CONFIG_SMP */
+
+ mutex_lock(&ipd->mutex);
+
+#ifdef CONFIG_PROC_FS
+ __ipipe_remove_domain_proc(ipd);
+#endif /* CONFIG_PROC_FS */
+
+ /*
+ * Simply remove the domain from the pipeline and we are almost done.
+ */
+
+ flags = ipipe_critical_enter(NULL);
+ list_del_init(&ipd->p_link);
+ ipipe_critical_exit(flags);
+
+ __ipipe_cleanup_domain(ipd);
+
+ mutex_unlock(&ipd->mutex);
+
+ printk(KERN_INFO "I-pipe: Domain %s unregistered.\n", ipd->name);
+
+ return 0;
+}
+
+/*
+ * ipipe_propagate_irq() -- Force a given IRQ propagation on behalf of
+ * a running interrupt handler to the next domain down the pipeline.
+ * ipipe_schedule_irq() -- Does almost the same as above, but attempts
+ * to pend the interrupt for the current domain first.
+ */
+int __ipipe_schedule_irq(unsigned irq, struct list_head *head)
+{
+ struct ipipe_domain *ipd;
+ struct list_head *ln;
+ unsigned long flags;
+
+ if (irq >= IPIPE_NR_IRQS ||
+ (ipipe_virtual_irq_p(irq)
+ && !test_bit(irq - IPIPE_VIRQ_BASE, &__ipipe_virtual_irq_map)))
+ return -EINVAL;
+
+ local_irq_save_hw(flags);
+
+ ln = head;
+
+ while (ln != &__ipipe_pipeline) {
+
+ ipd = list_entry(ln, struct ipipe_domain, p_link);
+
+ if (test_bit(IPIPE_HANDLE_FLAG, &ipd->irqs[irq].control)) {
+ __ipipe_set_irq_pending(ipd, irq);
+ local_irq_restore_hw(flags);
+ return 1;
+ }
+
+ ln = ipd->p_link.next;
+ }
+
+ local_irq_restore_hw(flags);
+
+ return 0;
+}
+
+/* ipipe_free_virq() -- Release a virtual/soft interrupt. */
+
+int ipipe_free_virq(unsigned virq)
+{
+ if (!ipipe_virtual_irq_p(virq))
+ return -EINVAL;
+
+ clear_bit(virq - IPIPE_VIRQ_BASE, &__ipipe_virtual_irq_map);
+
+ return 0;
+}
+
+void ipipe_init_attr(struct ipipe_domain_attr *attr)
+{
+ attr->name = "anon";
+ attr->domid = 1;
+ attr->entry = NULL;
+ attr->priority = IPIPE_ROOT_PRIO;
+ attr->pdd = NULL;
+}
+
+/*
+ * ipipe_catch_event() -- Interpose or remove an event handler for a
+ * given domain.
+ */
+ipipe_event_handler_t ipipe_catch_event(struct ipipe_domain *ipd,
+ unsigned event,
+ ipipe_event_handler_t handler)
+{
+ ipipe_event_handler_t old_handler;
+ unsigned long flags;
+ int self = 0, cpu;
+
+ if (event & IPIPE_EVENT_SELF) {
+ event &= ~IPIPE_EVENT_SELF;
+ self = 1;
+ }
+
+ if (event >= IPIPE_NR_EVENTS)
+ return NULL;
+
+ flags = ipipe_critical_enter(NULL);
+
+ if (!(old_handler = xchg(&ipd->evhand[event],handler))) {
+ if (handler) {
+ if (self)
+ ipd->evself |= (1LL << event);
+ else
+ __ipipe_event_monitors[event]++;
+ }
+ }
+ else if (!handler) {
+ if (ipd->evself & (1LL << event))
+ ipd->evself &= ~(1LL << event);
+ else
+ __ipipe_event_monitors[event]--;
+ } else if ((ipd->evself & (1LL << event)) && !self) {
+ __ipipe_event_monitors[event]++;
+ ipd->evself &= ~(1LL << event);
+ } else if (!(ipd->evself & (1LL << event)) && self) {
+ __ipipe_event_monitors[event]--;
+ ipd->evself |= (1LL << event);
+ }
+
+ ipipe_critical_exit(flags);
+
+ if (!handler && ipipe_root_domain_p) {
+ /*
+ * If we cleared a handler on behalf of the root
+ * domain, we have to wait for any current invocation
+ * to drain, since our caller might subsequently unmap
+ * the target domain. To this aim, this code
+ * synchronizes with __ipipe_dispatch_event(),
+ * guaranteeing that either the dispatcher sees a null
+ * handler in which case it discards the invocation
+ * (which also prevents from entering a livelock), or
+ * finds a valid handler and calls it. Symmetrically,
+ * ipipe_catch_event() ensures that the called code
+ * won't be unmapped under our feet until the event
+ * synchronization flag is cleared for the given event
+ * on all CPUs.
+ */
+
+ for_each_online_cpu(cpu) {
+ while (ipipe_percpudom(ipd, evsync, cpu) & (1LL << event))
+ schedule_timeout_interruptible(HZ / 50);
+ }
+ }
+
+ return old_handler;
+}
+
+cpumask_t ipipe_set_irq_affinity (unsigned irq, cpumask_t cpumask)
+{
+#ifdef CONFIG_SMP
+ if (irq >= IPIPE_NR_XIRQS)
+ /* Allow changing affinity of external IRQs only. */
+ return CPU_MASK_NONE;
+
+ if (num_online_cpus() > 1)
+ return __ipipe_set_irq_affinity(irq,cpumask);
+#endif /* CONFIG_SMP */
+
+ return CPU_MASK_NONE;
+}
+
+int ipipe_send_ipi (unsigned ipi, cpumask_t cpumask)
+
+{
+#ifdef CONFIG_SMP
+ return __ipipe_send_ipi(ipi,cpumask);
+#else /* !CONFIG_SMP */
+ return -EINVAL;
+#endif /* CONFIG_SMP */
+}
+
+int ipipe_alloc_ptdkey (void)
+{
+ unsigned long flags;
+ int key = -1;
+
+ spin_lock_irqsave(&__ipipe_pipelock,flags);
+
+ if (__ipipe_ptd_key_count < IPIPE_ROOT_NPTDKEYS) {
+ key = ffz(__ipipe_ptd_key_map);
+ set_bit(key,&__ipipe_ptd_key_map);
+ __ipipe_ptd_key_count++;
+ }
+
+ spin_unlock_irqrestore(&__ipipe_pipelock,flags);
+
+ return key;
+}
+
+int ipipe_free_ptdkey (int key)
+{
+ unsigned long flags;
+
+ if (key < 0 || key >= IPIPE_ROOT_NPTDKEYS)
+ return -EINVAL;
+
+ spin_lock_irqsave(&__ipipe_pipelock,flags);
+
+ if (test_and_clear_bit(key,&__ipipe_ptd_key_map))
+ __ipipe_ptd_key_count--;
+
+ spin_unlock_irqrestore(&__ipipe_pipelock,flags);
+
+ return 0;
+}
+
+int ipipe_set_ptd (int key, void *value)
+
+{
+ if (key < 0 || key >= IPIPE_ROOT_NPTDKEYS)
+ return -EINVAL;
+
+ current->ptd[key] = value;
+
+ return 0;
+}
+
+void *ipipe_get_ptd (int key)
+
+{
+ if (key < 0 || key >= IPIPE_ROOT_NPTDKEYS)
+ return NULL;
+
+ return current->ptd[key];
+}
+
+#ifdef CONFIG_PROC_FS
+
+struct proc_dir_entry *ipipe_proc_root;
+
+static int __ipipe_version_info_proc(char *page,
+ char **start,
+ off_t off, int count, int *eof, void *data)
+{
+ int len = sprintf(page, "%s\n", IPIPE_VERSION_STRING);
+
+ len -= off;
+
+ if (len <= off + count)
+ *eof = 1;
+
+ *start = page + off;
+
+ if(len > count)
+ len = count;
+
+ if(len < 0)
+ len = 0;
+
+ return len;
+}
+
+static int __ipipe_common_info_show(struct seq_file *p, void *data)
+{
+ struct ipipe_domain *ipd = (struct ipipe_domain *)p->private;
+ char handling, stickiness, lockbit, exclusive, virtuality;
+
+ unsigned long ctlbits;
+ unsigned irq;
+
+ seq_printf(p, " +----- Handling ([A]ccepted, [G]rabbed, [W]ired, [D]iscarded)\n");
+ seq_printf(p, " |+---- Sticky\n");
+ seq_printf(p, " ||+--- Locked\n");
+ seq_printf(p, " |||+-- Exclusive\n");
+ seq_printf(p, " ||||+- Virtual\n");
+ seq_printf(p, "[IRQ] |||||\n");
+
+ mutex_lock(&ipd->mutex);
+
+ for (irq = 0; irq < IPIPE_NR_IRQS; irq++) {
+ /* Remember to protect against
+ * ipipe_virtual_irq/ipipe_control_irq if more fields
+ * get involved. */
+ ctlbits = ipd->irqs[irq].control;
+
+ if (irq >= IPIPE_NR_XIRQS && !ipipe_virtual_irq_p(irq))
+ /*
+ * There might be a hole between the last external
+ * IRQ and the first virtual one; skip it.
+ */
+ continue;
+
+ if (ipipe_virtual_irq_p(irq)
+ && !test_bit(irq - IPIPE_VIRQ_BASE, &__ipipe_virtual_irq_map))
+ /* Non-allocated virtual IRQ; skip it. */
+ continue;
+
+ /*
+ * Statuses are as follows:
+ * o "accepted" means handled _and_ passed down the pipeline.
+ * o "grabbed" means handled, but the interrupt might be
+ * terminated _or_ passed down the pipeline depending on
+ * what the domain handler asks for to the I-pipe.
+ * o "wired" is basically the same as "grabbed", except that
+ * the interrupt is unconditionally delivered to an invariant
+ * pipeline head domain.
+ * o "passed" means unhandled by the domain but passed
+ * down the pipeline.
+ * o "discarded" means unhandled and _not_ passed down the
+ * pipeline. The interrupt merely disappears from the
+ * current domain down to the end of the pipeline.
+ */
+ if (ctlbits & IPIPE_HANDLE_MASK) {
+ if (ctlbits & IPIPE_PASS_MASK)
+ handling = 'A';
+ else if (ctlbits & IPIPE_WIRED_MASK)
+ handling = 'W';
+ else
+ handling = 'G';
+ } else if (ctlbits & IPIPE_PASS_MASK)
+ /* Do not output if no major action is taken. */
+ continue;
+ else
+ handling = 'D';
+
+ if (ctlbits & IPIPE_STICKY_MASK)
+ stickiness = 'S';
+ else
+ stickiness = '.';
+
+ if (ctlbits & IPIPE_LOCK_MASK)
+ lockbit = 'L';
+ else
+ lockbit = '.';
+
+ if (ctlbits & IPIPE_EXCLUSIVE_MASK)
+ exclusive = 'X';
+ else
+ exclusive = '.';
+
+ if (ipipe_virtual_irq_p(irq))
+ virtuality = 'V';
+ else
+ virtuality = '.';
+
+ seq_printf(p, " %3u: %c%c%c%c%c\n",
+ irq, handling, stickiness, lockbit, exclusive, virtuality);
+ }
+
+ seq_printf(p, "[Domain info]\n");
+
+ seq_printf(p, "id=0x%.8x\n", ipd->domid);
+
+ if (test_bit(IPIPE_AHEAD_FLAG,&ipd->flags))
+ seq_printf(p, "priority=topmost\n");
+ else
+ seq_printf(p, "priority=%d\n", ipd->priority);
+
+ mutex_unlock(&ipd->mutex);
+
+ return 0;
+}
+
+static int __ipipe_common_info_open(struct inode *inode, struct file *file)
+{
+ return single_open(file, __ipipe_common_info_show, PROC_I(inode)->pde->data);
+}
+
+static struct file_operations __ipipe_info_proc_ops = {
+ .owner = THIS_MODULE,
+ .open = __ipipe_common_info_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+ .release = single_release,
+};
+
+void __ipipe_add_domain_proc(struct ipipe_domain *ipd)
+{
+ struct proc_dir_entry *e = create_proc_entry(ipd->name, 0444, ipipe_proc_root);
+ if (e) {
+ e->proc_fops = &__ipipe_info_proc_ops;
+ e->data = (void*) ipd;
+ }
+}
+
+void __ipipe_remove_domain_proc(struct ipipe_domain *ipd)
+{
+ remove_proc_entry(ipd->name,ipipe_proc_root);
+}
+
+void __init ipipe_init_proc(void)
+{
+ ipipe_proc_root = create_proc_entry("ipipe",S_IFDIR, 0);
+ create_proc_read_entry("version",0444,ipipe_proc_root,&__ipipe_version_info_proc,NULL);
+ __ipipe_add_domain_proc(ipipe_root_domain);
+
+ __ipipe_init_tracer();
+}
+
+#endif /* CONFIG_PROC_FS */
+
+#ifdef CONFIG_IPIPE_DEBUG_CONTEXT
+
+DEFINE_PER_CPU(int, ipipe_percpu_context_check) = { 1 };
+
+void ipipe_check_context(struct ipipe_domain *border_ipd)
+{
+ /* Note: We don't make the per_cpu access atomic. We assume that code
+ which temporarily disables the check does this in atomic context
+ only. */
+ if (likely(ipipe_current_domain->priority <= border_ipd->priority) ||
+ !per_cpu(ipipe_percpu_context_check, ipipe_processor_id()))
+ return;
+
+ ipipe_context_check_off();
+
+ ipipe_trace_panic_freeze();
+ ipipe_set_printk_sync(ipipe_current_domain);
+ printk(KERN_ERR "I-pipe: Detected illicit call from domain '%s'\n"
+ KERN_ERR " into a service reserved for domain '%s' and "
+ "below.\n",
+ ipipe_current_domain->name, border_ipd->name);
+ dump_stack();
+ ipipe_trace_panic_dump();
+}
+
+EXPORT_SYMBOL(ipipe_check_context);
+#endif /* CONFIG_IPIPE_DEBUG_CONTEXT */
+
+EXPORT_SYMBOL(ipipe_virtualize_irq);
+EXPORT_SYMBOL(ipipe_control_irq);
+EXPORT_SYMBOL(ipipe_suspend_domain);
+EXPORT_SYMBOL(ipipe_alloc_virq);
+EXPORT_PER_CPU_SYMBOL(ipipe_percpu_domain);
+EXPORT_PER_CPU_SYMBOL(ipipe_percpu_darray);
+EXPORT_SYMBOL(ipipe_root);
+EXPORT_SYMBOL(ipipe_stall_pipeline_from);
+EXPORT_SYMBOL(ipipe_test_and_stall_pipeline_from);
+EXPORT_SYMBOL(ipipe_unstall_pipeline_from);
+EXPORT_SYMBOL(ipipe_restore_pipeline_from);
+EXPORT_SYMBOL(ipipe_test_and_unstall_pipeline_from);
+EXPORT_SYMBOL(ipipe_unstall_pipeline_head);
+EXPORT_SYMBOL(__ipipe_restore_pipeline_head);
+EXPORT_SYMBOL(__ipipe_unstall_root);
+EXPORT_SYMBOL(__ipipe_restore_root);
+EXPORT_SYMBOL(__ipipe_spin_lock_irq);
+EXPORT_SYMBOL(__ipipe_spin_unlock_irq);
+EXPORT_SYMBOL(__ipipe_spin_lock_irqsave);
+EXPORT_SYMBOL(__ipipe_spin_unlock_irqrestore);
+EXPORT_SYMBOL(__ipipe_pipeline);
+EXPORT_SYMBOL(__ipipe_lock_irq);
+EXPORT_SYMBOL(__ipipe_unlock_irq);
+EXPORT_SYMBOL(ipipe_register_domain);
+EXPORT_SYMBOL(ipipe_unregister_domain);
+EXPORT_SYMBOL(ipipe_free_virq);
+EXPORT_SYMBOL(ipipe_init_attr);
+EXPORT_SYMBOL(ipipe_catch_event);
+EXPORT_SYMBOL(ipipe_alloc_ptdkey);
+EXPORT_SYMBOL(ipipe_free_ptdkey);
+EXPORT_SYMBOL(ipipe_set_ptd);
+EXPORT_SYMBOL(ipipe_get_ptd);
+EXPORT_SYMBOL(ipipe_set_irq_affinity);
+EXPORT_SYMBOL(ipipe_send_ipi);
+EXPORT_SYMBOL(__ipipe_schedule_irq);
+#ifdef CONFIG_GENERIC_CLOCKEVENTS
+EXPORT_SYMBOL(ipipe_request_tickdev);
+EXPORT_SYMBOL(ipipe_release_tickdev);
+#endif
+
+EXPORT_SYMBOL(ipipe_critical_enter);
+EXPORT_SYMBOL(ipipe_critical_exit);
+EXPORT_SYMBOL(ipipe_trigger_irq);
+EXPORT_SYMBOL(ipipe_get_sysinfo);
diff --git a/kernel/ipipe/tracer.c b/kernel/ipipe/tracer.c
new file mode 100644
index 0000000..fc7fe99
--- /dev/null
+++ b/kernel/ipipe/tracer.c
@@ -0,0 +1,1336 @@
+/* -*- linux-c -*-
+ * kernel/ipipe/tracer.c
+ *
+ * Copyright (C) 2005 Luotao Fu.
+ * 2005-2008 Jan Kiszka.
+ *
+ * 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, Inc., 675 Mass Ave, Cambridge MA 02139,
+ * USA; 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/kernel.h>
+#include <linux/module.h>
+#include <linux/version.h>
+#include <linux/kallsyms.h>
+#include <linux/seq_file.h>
+#include <linux/proc_fs.h>
+#include <linux/ctype.h>
+#include <linux/vmalloc.h>
+#include <linux/utsrelease.h>
+#include <linux/sched.h>
+#include <linux/ipipe.h>
+#include <asm/uaccess.h>
+
+#define IPIPE_TRACE_PATHS 4 /* <!> Do not lower below 3 */
+#define IPIPE_DEFAULT_ACTIVE 0
+#define IPIPE_DEFAULT_MAX 1
+#define IPIPE_DEFAULT_FROZEN 2
+
+#define IPIPE_TRACE_POINTS (1 << CONFIG_IPIPE_TRACE_SHIFT)
+#define WRAP_POINT_NO(point) ((point) & (IPIPE_TRACE_POINTS-1))
+
+#define IPIPE_DEFAULT_PRE_TRACE 10
+#define IPIPE_DEFAULT_POST_TRACE 10
+#define IPIPE_DEFAULT_BACK_TRACE 100
+
+#define IPIPE_DELAY_NOTE 1000 /* in nanoseconds */
+#define IPIPE_DELAY_WARN 10000 /* in nanoseconds */
+
+#define IPIPE_TFLG_NMI_LOCK 0x0001
+#define IPIPE_TFLG_NMI_HIT 0x0002
+#define IPIPE_TFLG_NMI_FREEZE_REQ 0x0004
+
+#define IPIPE_TFLG_HWIRQ_OFF 0x0100
+#define IPIPE_TFLG_FREEZING 0x0200
+#define IPIPE_TFLG_CURRDOM_SHIFT 10 /* bits 10..11: current domain */
+#define IPIPE_TFLG_CURRDOM_MASK 0x0C00
+#define IPIPE_TFLG_DOMSTATE_SHIFT 12 /* bits 12..15: domain stalled? */
+#define IPIPE_TFLG_DOMSTATE_BITS 3
+
+#define IPIPE_TFLG_DOMAIN_STALLED(point, n) \
+ (point->flags & (1 << (n + IPIPE_TFLG_DOMSTATE_SHIFT)))
+#define IPIPE_TFLG_CURRENT_DOMAIN(point) \
+ ((point->flags & IPIPE_TFLG_CURRDOM_MASK) >> IPIPE_TFLG_CURRDOM_SHIFT)
+
+struct ipipe_trace_point {
+ short type;
+ short flags;
+ unsigned long eip;
+ unsigned long parent_eip;
+ unsigned long v;
+ unsigned long long timestamp;
+};
+
+struct ipipe_trace_path {
+ volatile int flags;
+ int dump_lock; /* separated from flags due to cross-cpu access */
+ int trace_pos; /* next point to fill */
+ int begin, end; /* finalised path begin and end */
+ int post_trace; /* non-zero when in post-trace phase */
+ unsigned long long length; /* max path length in cycles */
+ unsigned long nmi_saved_eip; /* for deferred requests from NMIs */
+ unsigned long nmi_saved_parent_eip;
+ unsigned long nmi_saved_v;
+ struct ipipe_trace_point point[IPIPE_TRACE_POINTS];
+} ____cacheline_aligned_in_smp;
+
+enum ipipe_trace_type
+{
+ IPIPE_TRACE_FUNC = 0,
+ IPIPE_TRACE_BEGIN,
+ IPIPE_TRACE_END,
+ IPIPE_TRACE_FREEZE,
+ IPIPE_TRACE_SPECIAL,
+ IPIPE_TRACE_PID,
+};
+
+#define IPIPE_TYPE_MASK 0x0007
+#define IPIPE_TYPE_BITS 3
+
+#ifdef CONFIG_IPIPE_TRACE_VMALLOC
+static DEFINE_PER_CPU(struct ipipe_trace_path *, trace_path);
+#else /* !CONFIG_IPIPE_TRACE_VMALLOC */
+static DEFINE_PER_CPU(struct ipipe_trace_path, trace_path[IPIPE_TRACE_PATHS]) =
+ { [0 ... IPIPE_TRACE_PATHS-1] = { .begin = -1, .end = -1 } };
+#endif /* CONFIG_IPIPE_TRACE_VMALLOC */
+
+int ipipe_trace_enable = 0;
+
+static DEFINE_PER_CPU(int, active_path) = { IPIPE_DEFAULT_ACTIVE };
+static DEFINE_PER_CPU(int, max_path) = { IPIPE_DEFAULT_MAX };
+static DEFINE_PER_CPU(int, frozen_path) = { IPIPE_DEFAULT_FROZEN };
+static IPIPE_DEFINE_SPINLOCK(global_path_lock);
+static int pre_trace = IPIPE_DEFAULT_PRE_TRACE;
+static int post_trace = IPIPE_DEFAULT_POST_TRACE;
+static int back_trace = IPIPE_DEFAULT_BACK_TRACE;
+static int verbose_trace = 1;
+static unsigned long trace_overhead;
+
+static unsigned long trigger_begin;
+static unsigned long trigger_end;
+
+static DEFINE_MUTEX(out_mutex);
+static struct ipipe_trace_path *print_path;
+#ifdef CONFIG_IPIPE_TRACE_PANIC
+static struct ipipe_trace_path *panic_path;
+#endif /* CONFIG_IPIPE_TRACE_PANIC */
+static int print_pre_trace;
+static int print_post_trace;
+
+
+static long __ipipe_signed_tsc2us(long long tsc);
+static void
+__ipipe_trace_point_type(char *buf, struct ipipe_trace_point *point);
+static void __ipipe_print_symname(struct seq_file *m, unsigned long eip);
+
+
+static notrace void
+__ipipe_store_domain_states(struct ipipe_trace_point *point)
+{
+ struct ipipe_domain *ipd;
+ struct list_head *pos;
+ int i = 0;
+
+ list_for_each_prev(pos, &__ipipe_pipeline) {
+ ipd = list_entry(pos, struct ipipe_domain, p_link);
+
+ if (test_bit(IPIPE_STALL_FLAG, &ipipe_cpudom_var(ipd, status)))
+ point->flags |= 1 << (i + IPIPE_TFLG_DOMSTATE_SHIFT);
+
+ if (ipd == ipipe_current_domain)
+ point->flags |= i << IPIPE_TFLG_CURRDOM_SHIFT;
+
+ if (++i > IPIPE_TFLG_DOMSTATE_BITS)
+ break;
+ }
+}
+
+static notrace int __ipipe_get_free_trace_path(int old, int cpu)
+{
+ int new_active = old;
+ struct ipipe_trace_path *tp;
+
+ do {
+ if (++new_active == IPIPE_TRACE_PATHS)
+ new_active = 0;
+ tp = &per_cpu(trace_path, cpu)[new_active];
+ } while (new_active == per_cpu(max_path, cpu) ||
+ new_active == per_cpu(frozen_path, cpu) ||
+ tp->dump_lock);
+
+ return new_active;
+}
+
+static notrace void
+__ipipe_migrate_pre_trace(struct ipipe_trace_path *new_tp,
+ struct ipipe_trace_path *old_tp, int old_pos)
+{
+ int i;
+
+ new_tp->trace_pos = pre_trace+1;
+
+ for (i = new_tp->trace_pos; i > 0; i--)
+ memcpy(&new_tp->point[WRAP_POINT_NO(new_tp->trace_pos-i)],
+ &old_tp->point[WRAP_POINT_NO(old_pos-i)],
+ sizeof(struct ipipe_trace_point));
+
+ /* mark the end (i.e. the point before point[0]) invalid */
+ new_tp->point[IPIPE_TRACE_POINTS-1].eip = 0;
+}
+
+static notrace struct ipipe_trace_path *
+__ipipe_trace_end(int cpu, struct ipipe_trace_path *tp, int pos)
+{
+ struct ipipe_trace_path *old_tp = tp;
+ long active = per_cpu(active_path, cpu);
+ unsigned long long length;
+
+ /* do we have a new worst case? */
+ length = tp->point[tp->end].timestamp -
+ tp->point[tp->begin].timestamp;
+ if (length > per_cpu(trace_path, cpu)[per_cpu(max_path, cpu)].length) {
+ /* we need protection here against other cpus trying
+ to start a proc dump */
+ spin_lock(&global_path_lock);
+
+ /* active path holds new worst case */
+ tp->length = length;
+ per_cpu(max_path, cpu) = active;
+
+ /* find next unused trace path */
+ active = __ipipe_get_free_trace_path(active, cpu);
+
+ spin_unlock(&global_path_lock);
+
+ tp = &per_cpu(trace_path, cpu)[active];
+
+ /* migrate last entries for pre-tracing */
+ __ipipe_migrate_pre_trace(tp, old_tp, pos);
+ }
+
+ return tp;
+}
+
+static notrace struct ipipe_trace_path *
+__ipipe_trace_freeze(int cpu, struct ipipe_trace_path *tp, int pos)
+{
+ struct ipipe_trace_path *old_tp = tp;
+ long active = per_cpu(active_path, cpu);
+ int n;
+
+ /* frozen paths have no core (begin=end) */
+ tp->begin = tp->end;
+
+ /* we need protection here against other cpus trying
+ * to set their frozen path or to start a proc dump */
+ spin_lock(&global_path_lock);
+
+ per_cpu(frozen_path, cpu) = active;
+
+ /* find next unused trace path */
+ active = __ipipe_get_free_trace_path(active, cpu);
+
+ /* check if this is the first frozen path */
+ for_each_possible_cpu(n) {
+ if (n != cpu &&
+ per_cpu(trace_path, n)[per_cpu(frozen_path, n)].end >= 0)
+ tp->end = -1;
+ }
+
+ spin_unlock(&global_path_lock);
+
+ tp = &per_cpu(trace_path, cpu)[active];
+
+ /* migrate last entries for pre-tracing */
+ __ipipe_migrate_pre_trace(tp, old_tp, pos);
+
+ return tp;
+}
+
+void notrace
+__ipipe_trace(enum ipipe_trace_type type, unsigned long eip,
+ unsigned long parent_eip, unsigned long v)
+{
+ struct ipipe_trace_path *tp, *old_tp;
+ int pos, next_pos, begin;
+ struct ipipe_trace_point *point;
+ unsigned long flags;
+ int cpu;
+
+ local_irq_save_hw_notrace(flags);
+
+ cpu = ipipe_processor_id();
+ restart:
+ tp = old_tp = &per_cpu(trace_path, cpu)[per_cpu(active_path, cpu)];
+
+ /* here starts a race window with NMIs - catched below */
+
+ /* check for NMI recursion */
+ if (unlikely(tp->flags & IPIPE_TFLG_NMI_LOCK)) {
+ tp->flags |= IPIPE_TFLG_NMI_HIT;
+
+ /* first freeze request from NMI context? */
+ if ((type == IPIPE_TRACE_FREEZE) &&
+ !(tp->flags & IPIPE_TFLG_NMI_FREEZE_REQ)) {
+ /* save arguments and mark deferred freezing */
+ tp->flags |= IPIPE_TFLG_NMI_FREEZE_REQ;
+ tp->nmi_saved_eip = eip;
+ tp->nmi_saved_parent_eip = parent_eip;
+ tp->nmi_saved_v = v;
+ }
+ return; /* no need for restoring flags inside IRQ */
+ }
+
+ /* clear NMI events and set lock (atomically per cpu) */
+ tp->flags = (tp->flags & ~(IPIPE_TFLG_NMI_HIT |
+ IPIPE_TFLG_NMI_FREEZE_REQ))
+ | IPIPE_TFLG_NMI_LOCK;
+
+ /* check active_path again - some nasty NMI may have switched
+ * it meanwhile */
+ if (unlikely(tp !=
+ &per_cpu(trace_path, cpu)[per_cpu(active_path, cpu)])) {
+ /* release lock on wrong path and restart */
+ tp->flags &= ~IPIPE_TFLG_NMI_LOCK;
+
+ /* there is no chance that the NMI got deferred
+ * => no need to check for pending freeze requests */
+ goto restart;
+ }
+
+ /* get the point buffer */
+ pos = tp->trace_pos;
+ point = &tp->point[pos];
+
+ /* store all trace point data */
+ point->type = type;
+ point->flags = raw_irqs_disabled_flags(flags) ? IPIPE_TFLG_HWIRQ_OFF : 0;
+ point->eip = eip;
+ point->parent_eip = parent_eip;
+ point->v = v;
+ ipipe_read_tsc(point->timestamp);
+
+ __ipipe_store_domain_states(point);
+
+ /* forward to next point buffer */
+ next_pos = WRAP_POINT_NO(pos+1);
+ tp->trace_pos = next_pos;
+
+ /* only mark beginning if we haven't started yet */
+ begin = tp->begin;
+ if (unlikely(type == IPIPE_TRACE_BEGIN) && (begin < 0))
+ tp->begin = pos;
+
+ /* end of critical path, start post-trace if not already started */
+ if (unlikely(type == IPIPE_TRACE_END) &&
+ (begin >= 0) && !tp->post_trace)
+ tp->post_trace = post_trace + 1;
+
+ /* freeze only if the slot is free and we are not already freezing */
+ if ((unlikely(type == IPIPE_TRACE_FREEZE) ||
+ (unlikely(eip >= trigger_begin && eip <= trigger_end) &&
+ type == IPIPE_TRACE_FUNC)) &&
+ per_cpu(trace_path, cpu)[per_cpu(frozen_path, cpu)].begin < 0 &&
+ !(tp->flags & IPIPE_TFLG_FREEZING)) {
+ tp->post_trace = post_trace + 1;
+ tp->flags |= IPIPE_TFLG_FREEZING;
+ }
+
+ /* enforce end of trace in case of overflow */
+ if (unlikely(WRAP_POINT_NO(next_pos + 1) == begin)) {
+ tp->end = pos;
+ goto enforce_end;
+ }
+
+ /* stop tracing this path if we are in post-trace and
+ * a) that phase is over now or
+ * b) a new TRACE_BEGIN came in but we are not freezing this path */
+ if (unlikely((tp->post_trace > 0) && ((--tp->post_trace == 0) ||
+ ((type == IPIPE_TRACE_BEGIN) &&
+ !(tp->flags & IPIPE_TFLG_FREEZING))))) {
+ /* store the path's end (i.e. excluding post-trace) */
+ tp->end = WRAP_POINT_NO(pos - post_trace + tp->post_trace);
+
+ enforce_end:
+ if (tp->flags & IPIPE_TFLG_FREEZING)
+ tp = __ipipe_trace_freeze(cpu, tp, pos);
+ else
+ tp = __ipipe_trace_end(cpu, tp, pos);
+
+ /* reset the active path, maybe already start a new one */
+ tp->begin = (type == IPIPE_TRACE_BEGIN) ?
+ WRAP_POINT_NO(tp->trace_pos - 1) : -1;
+ tp->end = -1;
+ tp->post_trace = 0;
+ tp->flags = 0;
+
+ /* update active_path not earlier to avoid races with NMIs */
+ per_cpu(active_path, cpu) = tp - per_cpu(trace_path, cpu);
+ }
+
+ /* we still have old_tp and point,
+ * let's reset NMI lock and check for catches */
+ old_tp->flags &= ~IPIPE_TFLG_NMI_LOCK;
+ if (unlikely(old_tp->flags & IPIPE_TFLG_NMI_HIT)) {
+ /* well, this late tagging may not immediately be visible for
+ * other cpus already dumping this path - a minor issue */
+ point->flags |= IPIPE_TFLG_NMI_HIT;
+
+ /* handle deferred freezing from NMI context */
+ if (old_tp->flags & IPIPE_TFLG_NMI_FREEZE_REQ)
+ __ipipe_trace(IPIPE_TRACE_FREEZE, old_tp->nmi_saved_eip,
+ old_tp->nmi_saved_parent_eip,
+ old_tp->nmi_saved_v);
+ }
+
+ local_irq_restore_hw_notrace(flags);
+}
+
+static unsigned long __ipipe_global_path_lock(void)
+{
+ unsigned long flags;
+ int cpu;
+ struct ipipe_trace_path *tp;
+
+ spin_lock_irqsave(&global_path_lock, flags);
+
+ cpu = ipipe_processor_id();
+ restart:
+ tp = &per_cpu(trace_path, cpu)[per_cpu(active_path, cpu)];
+
+ /* here is small race window with NMIs - catched below */
+
+ /* clear NMI events and set lock (atomically per cpu) */
+ tp->flags = (tp->flags & ~(IPIPE_TFLG_NMI_HIT |
+ IPIPE_TFLG_NMI_FREEZE_REQ))
+ | IPIPE_TFLG_NMI_LOCK;
+
+ /* check active_path again - some nasty NMI may have switched
+ * it meanwhile */
+ if (tp != &per_cpu(trace_path, cpu)[per_cpu(active_path, cpu)]) {
+ /* release lock on wrong path and restart */
+ tp->flags &= ~IPIPE_TFLG_NMI_LOCK;
+
+ /* there is no chance that the NMI got deferred
+ * => no need to check for pending freeze requests */
+ goto restart;
+ }
+
+ return flags;
+}
+
+static void __ipipe_global_path_unlock(unsigned long flags)
+{
+ int cpu;
+ struct ipipe_trace_path *tp;
+
+ /* release spinlock first - it's not involved in the NMI issue */
+ __ipipe_spin_unlock_irqbegin(&global_path_lock);
+
+ cpu = ipipe_processor_id();
+ tp = &per_cpu(trace_path, cpu)[per_cpu(active_path, cpu)];
+
+ tp->flags &= ~IPIPE_TFLG_NMI_LOCK;
+
+ /* handle deferred freezing from NMI context */
+ if (tp->flags & IPIPE_TFLG_NMI_FREEZE_REQ)
+ __ipipe_trace(IPIPE_TRACE_FREEZE, tp->nmi_saved_eip,
+ tp->nmi_saved_parent_eip, tp->nmi_saved_v);
+
+ /* See __ipipe_spin_lock_irqsave() and friends. */
+ __ipipe_spin_unlock_irqcomplete(flags);
+}
+
+void notrace ipipe_trace_begin(unsigned long v)
+{
+ if (!ipipe_trace_enable)
+ return;
+ __ipipe_trace(IPIPE_TRACE_BEGIN, __BUILTIN_RETURN_ADDRESS0,
+ __BUILTIN_RETURN_ADDRESS1, v);
+}
+EXPORT_SYMBOL(ipipe_trace_begin);
+
+void notrace ipipe_trace_end(unsigned long v)
+{
+ if (!ipipe_trace_enable)
+ return;
+ __ipipe_trace(IPIPE_TRACE_END, __BUILTIN_RETURN_ADDRESS0,
+ __BUILTIN_RETURN_ADDRESS1, v);
+}
+EXPORT_SYMBOL(ipipe_trace_end);
+
+void notrace ipipe_trace_freeze(unsigned long v)
+{
+ if (!ipipe_trace_enable)
+ return;
+ __ipipe_trace(IPIPE_TRACE_FREEZE, __BUILTIN_RETURN_ADDRESS0,
+ __BUILTIN_RETURN_ADDRESS1, v);
+}
+EXPORT_SYMBOL(ipipe_trace_freeze);
+
+void notrace ipipe_trace_special(unsigned char id, unsigned long v)
+{
+ if (!ipipe_trace_enable)
+ return;
+ __ipipe_trace(IPIPE_TRACE_SPECIAL | (id << IPIPE_TYPE_BITS),
+ __BUILTIN_RETURN_ADDRESS0,
+ __BUILTIN_RETURN_ADDRESS1, v);
+}
+EXPORT_SYMBOL(ipipe_trace_special);
+
+void notrace ipipe_trace_pid(pid_t pid, short prio)
+{
+ if (!ipipe_trace_enable)
+ return;
+ __ipipe_trace(IPIPE_TRACE_PID | (prio << IPIPE_TYPE_BITS),
+ __BUILTIN_RETURN_ADDRESS0,
+ __BUILTIN_RETURN_ADDRESS1, pid);
+}
+EXPORT_SYMBOL(ipipe_trace_pid);
+
+int ipipe_trace_max_reset(void)
+{
+ int cpu;
+ unsigned long flags;
+ struct ipipe_trace_path *path;
+ int ret = 0;
+
+ flags = __ipipe_global_path_lock();
+
+ for_each_possible_cpu(cpu) {
+ path = &per_cpu(trace_path, cpu)[per_cpu(max_path, cpu)];
+
+ if (path->dump_lock) {
+ ret = -EBUSY;
+ break;
+ }
+
+ path->begin = -1;
+ path->end = -1;
+ path->trace_pos = 0;
+ path->length = 0;
+ }
+
+ __ipipe_global_path_unlock(flags);
+
+ return ret;
+}
+EXPORT_SYMBOL(ipipe_trace_max_reset);
+
+int ipipe_trace_frozen_reset(void)
+{
+ int cpu;
+ unsigned long flags;
+ struct ipipe_trace_path *path;
+ int ret = 0;
+
+ flags = __ipipe_global_path_lock();
+
+ for_each_online_cpu(cpu) {
+ path = &per_cpu(trace_path, cpu)[per_cpu(frozen_path, cpu)];
+
+ if (path->dump_lock) {
+ ret = -EBUSY;
+ break;
+ }
+
+ path->begin = -1;
+ path->end = -1;
+ path->trace_pos = 0;
+ path->length = 0;
+ }
+
+ __ipipe_global_path_unlock(flags);
+
+ return ret;
+}
+EXPORT_SYMBOL(ipipe_trace_frozen_reset);
+
+static void
+__ipipe_get_task_info(char *task_info, struct ipipe_trace_point *point,
+ int trylock)
+{
+ struct task_struct *task = NULL;
+ char buf[8];
+ int i;
+ int locked = 1;
+
+ if (trylock) {
+ if (!read_trylock(&tasklist_lock))
+ locked = 0;
+ } else
+ read_lock(&tasklist_lock);
+
+ if (locked)
+ task = find_task_by_pid((pid_t)point->v);
+
+ if (task)
+ strncpy(task_info, task->comm, 11);
+ else
+ strcpy(task_info, "-<?>-");
+
+ if (locked)
+ read_unlock(&tasklist_lock);
+
+ for (i = strlen(task_info); i < 11; i++)
+ task_info[i] = ' ';
+
+ sprintf(buf, " %d ", point->type >> IPIPE_TYPE_BITS);
+ strcpy(task_info + (11 - strlen(buf)), buf);
+}
+
+#ifdef CONFIG_IPIPE_TRACE_PANIC
+void ipipe_trace_panic_freeze(void)
+{
+ unsigned long flags;
+ int cpu;
+
+ if (!ipipe_trace_enable)
+ return;
+
+ ipipe_trace_enable = 0;
+ local_irq_save_hw_notrace(flags);
+
+ cpu = ipipe_processor_id();
+
+ panic_path = &per_cpu(trace_path, cpu)[per_cpu(active_path, cpu)];
+
+ local_irq_restore_hw(flags);
+}
+EXPORT_SYMBOL(ipipe_trace_panic_freeze);
+
+void ipipe_trace_panic_dump(void)
+{
+ int cnt = back_trace;
+ int start, pos;
+ char task_info[12];
+
+ if (!panic_path)
+ return;
+
+ ipipe_context_check_off();
+
+ printk("I-pipe tracer log (%d points):\n", cnt);
+
+ start = pos = WRAP_POINT_NO(panic_path->trace_pos-1);
+
+ while (cnt-- > 0) {
+ struct ipipe_trace_point *point = &panic_path->point[pos];
+ long time;
+ char buf[16];
+ int i;
+
+ printk(" %c",
+ (point->flags & IPIPE_TFLG_HWIRQ_OFF) ? '|' : ' ');
+
+ for (i = IPIPE_TFLG_DOMSTATE_BITS; i >= 0; i--)
+ printk("%c",
+ (IPIPE_TFLG_CURRENT_DOMAIN(point) == i) ?
+ (IPIPE_TFLG_DOMAIN_STALLED(point, i) ?
+ '#' : '+') :
+ (IPIPE_TFLG_DOMAIN_STALLED(point, i) ?
+ '*' : ' '));
+
+ if (!point->eip)
+ printk("-<invalid>-\n");
+ else {
+ __ipipe_trace_point_type(buf, point);
+ printk(buf);
+
+ switch (point->type & IPIPE_TYPE_MASK) {
+ case IPIPE_TRACE_FUNC:
+ printk(" ");
+ break;
+
+ case IPIPE_TRACE_PID:
+ __ipipe_get_task_info(task_info,
+ point, 1);
+ printk(task_info);
+ break;
+
+ default:
+ printk("0x%08lx ", point->v);
+ }
+
+ time = __ipipe_signed_tsc2us(point->timestamp -
+ panic_path->point[start].timestamp);
+ printk(" %5ld ", time);
+
+ __ipipe_print_symname(NULL, point->eip);
+ printk(" (");
+ __ipipe_print_symname(NULL, point->parent_eip);
+ printk(")\n");
+ }
+ pos = WRAP_POINT_NO(pos - 1);
+ }
+
+ panic_path = NULL;
+}
+EXPORT_SYMBOL(ipipe_trace_panic_dump);
+#endif /* CONFIG_IPIPE_TRACE_PANIC */
+
+
+/* --- /proc output --- */
+
+static notrace int __ipipe_in_critical_trpath(long point_no)
+{
+ return ((WRAP_POINT_NO(point_no-print_path->begin) <
+ WRAP_POINT_NO(print_path->end-print_path->begin)) ||
+ ((print_path->end == print_path->begin) &&
+ (WRAP_POINT_NO(point_no-print_path->end) >
+ print_post_trace)));
+}
+
+static long __ipipe_signed_tsc2us(long long tsc)
+{
+ unsigned long long abs_tsc;
+ long us;
+
+ /* ipipe_tsc2us works on unsigned => handle sign separately */
+ abs_tsc = (tsc >= 0) ? tsc : -tsc;
+ us = ipipe_tsc2us(abs_tsc);
+ if (tsc < 0)
+ return -us;
+ else
+ return us;
+}
+
+static void
+__ipipe_trace_point_type(char *buf, struct ipipe_trace_point *point)
+{
+ switch (point->type & IPIPE_TYPE_MASK) {
+ case IPIPE_TRACE_FUNC:
+ strcpy(buf, "func ");
+ break;
+
+ case IPIPE_TRACE_BEGIN:
+ strcpy(buf, "begin ");
+ break;
+
+ case IPIPE_TRACE_END:
+ strcpy(buf, "end ");
+ break;
+
+ case IPIPE_TRACE_FREEZE:
+ strcpy(buf, "freeze ");
+ break;
+
+ case IPIPE_TRACE_SPECIAL:
+ sprintf(buf, "(0x%02x) ",
+ point->type >> IPIPE_TYPE_BITS);
+ break;
+
+ case IPIPE_TRACE_PID:
+ sprintf(buf, "[%5d] ", (pid_t)point->v);
+ break;
+ }
+}
+
+static void
+__ipipe_print_pathmark(struct seq_file *m, struct ipipe_trace_point *point)
+{
+ char mark = ' ';
+ int point_no = point - print_path->point;
+ int i;
+
+ if (print_path->end == point_no)
+ mark = '<';
+ else if (print_path->begin == point_no)
+ mark = '>';
+ else if (__ipipe_in_critical_trpath(point_no))
+ mark = ':';
+ seq_printf(m, "%c%c", mark,
+ (point->flags & IPIPE_TFLG_HWIRQ_OFF) ? '|' : ' ');
+
+ if (!verbose_trace)
+ return;
+
+ for (i = IPIPE_TFLG_DOMSTATE_BITS; i >= 0; i--)
+ seq_printf(m, "%c",
+ (IPIPE_TFLG_CURRENT_DOMAIN(point) == i) ?
+ (IPIPE_TFLG_DOMAIN_STALLED(point, i) ?
+ '#' : '+') :
+ (IPIPE_TFLG_DOMAIN_STALLED(point, i) ? '*' : ' '));
+}
+
+static void
+__ipipe_print_delay(struct seq_file *m, struct ipipe_trace_point *point)
+{
+ unsigned long delay = 0;
+ int next;
+ char *mark = " ";
+
+ next = WRAP_POINT_NO(point+1 - print_path->point);
+
+ if (next != print_path->trace_pos)
+ delay = ipipe_tsc2ns(print_path->point[next].timestamp -
+ point->timestamp);
+
+ if (__ipipe_in_critical_trpath(point - print_path->point)) {
+ if (delay > IPIPE_DELAY_WARN)
+ mark = "! ";
+ else if (delay > IPIPE_DELAY_NOTE)
+ mark = "+ ";
+ }
+ seq_puts(m, mark);
+
+ if (verbose_trace)
+ seq_printf(m, "%3lu.%03lu%c ", delay/1000, delay%1000,
+ (point->flags & IPIPE_TFLG_NMI_HIT) ? 'N' : ' ');
+ else
+ seq_puts(m, " ");
+}
+
+static void __ipipe_print_symname(struct seq_file *m, unsigned long eip)
+{
+ char namebuf[KSYM_NAME_LEN+1];
+ unsigned long size, offset;
+ const char *sym_name;
+ char *modname;
+
+ sym_name = kallsyms_lookup(eip, &size, &offset, &modname, namebuf);
+
+#ifdef CONFIG_IPIPE_TRACE_PANIC
+ if (!m) {
+ /* panic dump */
+ if (sym_name) {
+ printk("%s+0x%lx", sym_name, offset);
+ if (modname)
+ printk(" [%s]", modname);
+ }
+ } else
+#endif /* CONFIG_IPIPE_TRACE_PANIC */
+ {
+ if (sym_name) {
+ if (verbose_trace) {
+ seq_printf(m, "%s+0x%lx", sym_name, offset);
+ if (modname)
+ seq_printf(m, " [%s]", modname);
+ } else
+ seq_puts(m, sym_name);
+ } else
+ seq_printf(m, "<%08lx>", eip);
+ }
+}
+
+static void __ipipe_print_headline(struct seq_file *m)
+{
+ seq_printf(m, "Calibrated minimum trace-point overhead: %lu.%03lu "
+ "us\n\n", trace_overhead/1000, trace_overhead%1000);
+
+ if (verbose_trace) {
+ const char *name[4] = { [0 ... 3] = "<unused>" };
+ struct list_head *pos;
+ int i = 0;
+
+ list_for_each_prev(pos, &__ipipe_pipeline) {
+ struct ipipe_domain *ipd =
+ list_entry(pos, struct ipipe_domain, p_link);
+
+ name[i] = ipd->name;
+ if (++i > 3)
+ break;
+ }
+
+ seq_printf(m,
+ " +----- Hard IRQs ('|': locked)\n"
+ " |+---- %s\n"
+ " ||+--- %s\n"
+ " |||+-- %s\n"
+ " ||||+- %s%s\n"
+ " ||||| +---------- "
+ "Delay flag ('+': > %d us, '!': > %d us)\n"
+ " ||||| | +- "
+ "NMI noise ('N')\n"
+ " ||||| | |\n"
+ " Type User Val. Time Delay Function "
+ "(Parent)\n",
+ name[3], name[2], name[1], name[0],
+ name[0] ? " ('*': domain stalled, '+': current, "
+ "'#': current+stalled)" : "",
+ IPIPE_DELAY_NOTE/1000, IPIPE_DELAY_WARN/1000);
+ } else
+ seq_printf(m,
+ " +--------------- Hard IRQs ('|': locked)\n"
+ " | +- Delay flag "
+ "('+': > %d us, '!': > %d us)\n"
+ " | |\n"
+ " Type Time Function (Parent)\n",
+ IPIPE_DELAY_NOTE/1000, IPIPE_DELAY_WARN/1000);
+}
+
+static void *__ipipe_max_prtrace_start(struct seq_file *m, loff_t *pos)
+{
+ loff_t n = *pos;
+
+ mutex_lock(&out_mutex);
+
+ if (!n) {
+ struct ipipe_trace_path *tp;
+ unsigned long length_usecs;
+ int points, cpu;
+ unsigned long flags;
+
+ /* protect against max_path/frozen_path updates while we
+ * haven't locked our target path, also avoid recursively
+ * taking global_path_lock from NMI context */
+ flags = __ipipe_global_path_lock();
+
+ /* find the longest of all per-cpu paths */
+ print_path = NULL;
+ for_each_online_cpu(cpu) {
+ tp = &per_cpu(trace_path, cpu)[per_cpu(max_path, cpu)];
+ if ((print_path == NULL) ||
+ (tp->length > print_path->length)) {
+ print_path = tp;
+ break;
+ }
+ }
+ print_path->dump_lock = 1;
+
+ __ipipe_global_path_unlock(flags);
+
+ /* does this path actually contain data? */
+ if (print_path->end == print_path->begin)
+ return NULL;
+
+ /* number of points inside the critical path */
+ points = WRAP_POINT_NO(print_path->end-print_path->begin+1);
+
+ /* pre- and post-tracing length, post-trace length was frozen
+ in __ipipe_trace, pre-trace may have to be reduced due to
+ buffer overrun */
+ print_pre_trace = pre_trace;
+ print_post_trace = WRAP_POINT_NO(print_path->trace_pos -
+ print_path->end - 1);
+ if (points+pre_trace+print_post_trace > IPIPE_TRACE_POINTS - 1)
+ print_pre_trace = IPIPE_TRACE_POINTS - 1 - points -
+ print_post_trace;
+
+ length_usecs = ipipe_tsc2us(print_path->length);
+ seq_printf(m, "I-pipe worst-case tracing service on %s/ipipe-%s\n"
+ "------------------------------------------------------------\n",
+ UTS_RELEASE, IPIPE_ARCH_STRING);
+ seq_printf(m, "CPU: %d, Begin: %lld cycles, Trace Points: "
+ "%d (-%d/+%d), Length: %lu us\n",
+ cpu, print_path->point[print_path->begin].timestamp,
+ points, print_pre_trace, print_post_trace, length_usecs);
+ __ipipe_print_headline(m);
+ }
+
+ /* check if we are inside the trace range */
+ if (n >= WRAP_POINT_NO(print_path->end - print_path->begin + 1 +
+ print_pre_trace + print_post_trace))
+ return NULL;
+
+ /* return the next point to be shown */
+ return &print_path->point[WRAP_POINT_NO(print_path->begin -
+ print_pre_trace + n)];
+}
+
+static void *__ipipe_prtrace_next(struct seq_file *m, void *p, loff_t *pos)
+{
+ loff_t n = ++*pos;
+
+ /* check if we are inside the trace range with the next entry */
+ if (n >= WRAP_POINT_NO(print_path->end - print_path->begin + 1 +
+ print_pre_trace + print_post_trace))
+ return NULL;
+
+ /* return the next point to be shown */
+ return &print_path->point[WRAP_POINT_NO(print_path->begin -
+ print_pre_trace + *pos)];
+}
+
+static void __ipipe_prtrace_stop(struct seq_file *m, void *p)
+{
+ if (print_path)
+ print_path->dump_lock = 0;
+ mutex_unlock(&out_mutex);
+}
+
+static int __ipipe_prtrace_show(struct seq_file *m, void *p)
+{
+ long time;
+ struct ipipe_trace_point *point = p;
+ char buf[16];
+
+ if (!point->eip) {
+ seq_puts(m, "-<invalid>-\n");
+ return 0;
+ }
+
+ __ipipe_print_pathmark(m, point);
+ __ipipe_trace_point_type(buf, point);
+ seq_puts(m, buf);
+ if (verbose_trace)
+ switch (point->type & IPIPE_TYPE_MASK) {
+ case IPIPE_TRACE_FUNC:
+ seq_puts(m, " ");
+ break;
+
+ case IPIPE_TRACE_PID:
+ __ipipe_get_task_info(buf, point, 0);
+ seq_puts(m, buf);
+ break;
+
+ default:
+ seq_printf(m, "0x%08lx ", point->v);
+ }
+
+ time = __ipipe_signed_tsc2us(point->timestamp -
+ print_path->point[print_path->begin].timestamp);
+ seq_printf(m, "%5ld", time);
+
+ __ipipe_print_delay(m, point);
+ __ipipe_print_symname(m, point->eip);
+ seq_puts(m, " (");
+ __ipipe_print_symname(m, point->parent_eip);
+ seq_puts(m, ")\n");
+
+ return 0;
+}
+
+static struct seq_operations __ipipe_max_ptrace_ops = {
+ .start = __ipipe_max_prtrace_start,
+ .next = __ipipe_prtrace_next,
+ .stop = __ipipe_prtrace_stop,
+ .show = __ipipe_prtrace_show
+};
+
+static int __ipipe_max_prtrace_open(struct inode *inode, struct file *file)
+{
+ return seq_open(file, &__ipipe_max_ptrace_ops);
+}
+
+static ssize_t
+__ipipe_max_reset(struct file *file, const char __user *pbuffer,
+ size_t count, loff_t *data)
+{
+ mutex_lock(&out_mutex);
+ ipipe_trace_max_reset();
+ mutex_unlock(&out_mutex);
+
+ return count;
+}
+
+struct file_operations __ipipe_max_prtrace_fops = {
+ .open = __ipipe_max_prtrace_open,
+ .read = seq_read,
+ .write = __ipipe_max_reset,
+ .llseek = seq_lseek,
+ .release = seq_release,
+};
+
+static void *__ipipe_frozen_prtrace_start(struct seq_file *m, loff_t *pos)
+{
+ loff_t n = *pos;
+
+ mutex_lock(&out_mutex);
+
+ if (!n) {
+ struct ipipe_trace_path *tp;
+ int cpu;
+ unsigned long flags;
+
+ /* protect against max_path/frozen_path updates while we
+ * haven't locked our target path, also avoid recursively
+ * taking global_path_lock from NMI context */
+ flags = __ipipe_global_path_lock();
+
+ /* find the first of all per-cpu frozen paths */
+ print_path = NULL;
+ for_each_online_cpu(cpu) {
+ tp = &per_cpu(trace_path, cpu)[per_cpu(frozen_path, cpu)];
+ if (tp->end >= 0) {
+ print_path = tp;
+ break;
+ }
+ }
+ if (print_path)
+ print_path->dump_lock = 1;
+
+ __ipipe_global_path_unlock(flags);
+
+ if (!print_path)
+ return NULL;
+
+ /* back- and post-tracing length, post-trace length was frozen
+ in __ipipe_trace, back-trace may have to be reduced due to
+ buffer overrun */
+ print_pre_trace = back_trace-1; /* substract freeze point */
+ print_post_trace = WRAP_POINT_NO(print_path->trace_pos -
+ print_path->end - 1);
+ if (1+pre_trace+print_post_trace > IPIPE_TRACE_POINTS - 1)
+ print_pre_trace = IPIPE_TRACE_POINTS - 2 -
+ print_post_trace;
+
+ seq_printf(m, "I-pipe frozen back-tracing service on %s/ipipe-%s\n"
+ "------------------------------------------------------"
+ "------\n",
+ UTS_RELEASE, IPIPE_ARCH_STRING);
+ seq_printf(m, "CPU: %d, Freeze: %lld cycles, Trace Points: %d (+%d)\n",
+ cpu, print_path->point[print_path->begin].timestamp,
+ print_pre_trace+1, print_post_trace);
+ __ipipe_print_headline(m);
+ }
+
+ /* check if we are inside the trace range */
+ if (n >= print_pre_trace + 1 + print_post_trace)
+ return NULL;
+
+ /* return the next point to be shown */
+ return &print_path->point[WRAP_POINT_NO(print_path->begin-
+ print_pre_trace+n)];
+}
+
+static struct seq_operations __ipipe_frozen_ptrace_ops = {
+ .start = __ipipe_frozen_prtrace_start,
+ .next = __ipipe_prtrace_next,
+ .stop = __ipipe_prtrace_stop,
+ .show = __ipipe_prtrace_show
+};
+
+static int __ipipe_frozen_prtrace_open(struct inode *inode, struct file *file)
+{
+ return seq_open(file, &__ipipe_frozen_ptrace_ops);
+}
+
+static ssize_t
+__ipipe_frozen_ctrl(struct file *file, const char __user *pbuffer,
+ size_t count, loff_t *data)
+{
+ char *end, buf[16];
+ int val;
+ int n;
+
+ n = (count > sizeof(buf) - 1) ? sizeof(buf) - 1 : count;
+
+ if (copy_from_user(buf, pbuffer, n))
+ return -EFAULT;
+
+ buf[n] = '\0';
+ val = simple_strtol(buf, &end, 0);
+
+ if (((*end != '\0') && !isspace(*end)) || (val < 0))
+ return -EINVAL;
+
+ mutex_lock(&out_mutex);
+ ipipe_trace_frozen_reset();
+ if (val > 0)
+ ipipe_trace_freeze(-1);
+ mutex_unlock(&out_mutex);
+
+ return count;
+}
+
+struct file_operations __ipipe_frozen_prtrace_fops = {
+ .open = __ipipe_frozen_prtrace_open,
+ .read = seq_read,
+ .write = __ipipe_frozen_ctrl,
+ .llseek = seq_lseek,
+ .release = seq_release,
+};
+
+static int __ipipe_rd_proc_val(char *page, char **start, off_t off,
+ int count, int *eof, void *data)
+{
+ int len;
+
+ len = sprintf(page, "%u\n", *(int *)data);
+ len -= off;
+ if (len <= off + count)
+ *eof = 1;
+ *start = page + off;
+ if (len > count)
+ len = count;
+ if (len < 0)
+ len = 0;
+
+ return len;
+}
+
+static int __ipipe_wr_proc_val(struct file *file, const char __user *buffer,
+ unsigned long count, void *data)
+{
+ char *end, buf[16];
+ int val;
+ int n;
+
+ n = (count > sizeof(buf) - 1) ? sizeof(buf) - 1 : count;
+
+ if (copy_from_user(buf, buffer, n))
+ return -EFAULT;
+
+ buf[n] = '\0';
+ val = simple_strtol(buf, &end, 0);
+
+ if (((*end != '\0') && !isspace(*end)) || (val < 0))
+ return -EINVAL;
+
+ mutex_lock(&out_mutex);
+ *(int *)data = val;
+ mutex_unlock(&out_mutex);
+
+ return count;
+}
+
+static int __ipipe_rd_trigger(char *page, char **start, off_t off, int count,
+ int *eof, void *data)
+{
+ int len;
+
+ if (!trigger_begin)
+ return 0;
+
+ len = sprint_symbol(page, trigger_begin);
+ page[len++] = '\n';
+
+ len -= off;
+ if (len <= off + count)
+ *eof = 1;
+ *start = page + off;
+ if (len > count)
+ len = count;
+ if (len < 0)
+ len = 0;
+
+ return len;
+}
+
+static int __ipipe_wr_trigger(struct file *file, const char __user *buffer,
+ unsigned long count, void *data)
+{
+ char buf[KSYM_SYMBOL_LEN];
+ unsigned long begin, end;
+
+ if (count > sizeof(buf) - 1)
+ count = sizeof(buf) - 1;
+ if (copy_from_user(buf, buffer, count))
+ return -EFAULT;
+ buf[count] = 0;
+ if (buf[count-1] == '\n')
+ buf[count-1] = 0;
+
+ begin = kallsyms_lookup_name(buf);
+ if (!begin || !kallsyms_lookup_size_offset(begin, &end, NULL))
+ return -ENOENT;
+ end += begin - 1;
+
+ mutex_lock(&out_mutex);
+ /* invalidate the current range before setting a new one */
+ trigger_end = 0;
+ wmb();
+ ipipe_trace_frozen_reset();
+
+ /* set new range */
+ trigger_begin = begin;
+ wmb();
+ trigger_end = end;
+ mutex_unlock(&out_mutex);
+
+ return count;
+}
+
+extern struct proc_dir_entry *ipipe_proc_root;
+
+static void __init
+__ipipe_create_trace_proc_val(struct proc_dir_entry *trace_dir,
+ const char *name, int *value_ptr)
+{
+ struct proc_dir_entry *entry;
+
+ entry = create_proc_entry(name, 0644, trace_dir);
+ if (entry) {
+ entry->data = value_ptr;
+ entry->read_proc = __ipipe_rd_proc_val;
+ entry->write_proc = __ipipe_wr_proc_val;
+ entry->owner = THIS_MODULE;
+ }
+}
+
+void __init __ipipe_init_tracer(void)
+{
+ struct proc_dir_entry *trace_dir;
+ struct proc_dir_entry *entry;
+ unsigned long long start, end, min = ULLONG_MAX;
+ int i;
+#ifdef CONFIG_IPIPE_TRACE_VMALLOC
+ int cpu, path;
+
+ for_each_possible_cpu(cpu) {
+ struct ipipe_trace_path *tp_buf;
+
+ tp_buf = vmalloc_node(sizeof(struct ipipe_trace_path) *
+ IPIPE_TRACE_PATHS, cpu_to_node(cpu));
+ if (!tp_buf) {
+ printk(KERN_ERR "I-pipe: "
+ "insufficient memory for trace buffer.\n");
+ return;
+ }
+ memset(tp_buf, 0,
+ sizeof(struct ipipe_trace_path) * IPIPE_TRACE_PATHS);
+ for (path = 0; path < IPIPE_TRACE_PATHS; path++) {
+ tp_buf[path].begin = -1;
+ tp_buf[path].end = -1;
+ }
+ per_cpu(trace_path, cpu) = tp_buf;
+ }
+#endif /* CONFIG_IPIPE_TRACE_VMALLOC */
+ ipipe_trace_enable = CONFIG_IPIPE_TRACE_ENABLE_VALUE;
+
+ /* Calculate minimum overhead of __ipipe_trace() */
+ local_irq_disable_hw();
+ for (i = 0; i < 100; i++) {
+ ipipe_read_tsc(start);
+ __ipipe_trace(IPIPE_TRACE_FUNC, __BUILTIN_RETURN_ADDRESS0,
+ __BUILTIN_RETURN_ADDRESS1, 0);
+ ipipe_read_tsc(end);
+
+ end -= start;
+ if (end < min)
+ min = end;
+ }
+ local_irq_enable_hw();
+ trace_overhead = ipipe_tsc2ns(min);
+
+ trace_dir = create_proc_entry("trace", S_IFDIR, ipipe_proc_root);
+
+ entry = create_proc_entry("max", 0644, trace_dir);
+ if (entry)
+ entry->proc_fops = &__ipipe_max_prtrace_fops;
+
+ entry = create_proc_entry("frozen", 0644, trace_dir);
+ if (entry)
+ entry->proc_fops = &__ipipe_frozen_prtrace_fops;
+
+ entry = create_proc_entry("trigger", 0644, trace_dir);
+ if (entry) {
+ entry->read_proc = __ipipe_rd_trigger;
+ entry->write_proc = __ipipe_wr_trigger;
+ entry->owner = THIS_MODULE;
+ }
+
+ __ipipe_create_trace_proc_val(trace_dir, "pre_trace_points",
+ &pre_trace);
+ __ipipe_create_trace_proc_val(trace_dir, "post_trace_points",
+ &post_trace);
+ __ipipe_create_trace_proc_val(trace_dir, "back_trace_points",
+ &back_trace);
+ __ipipe_create_trace_proc_val(trace_dir, "verbose",
+ &verbose_trace);
+ __ipipe_create_trace_proc_val(trace_dir, "enable",
+ &ipipe_trace_enable);
+}
diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c
index e4e1c99..4e01167 100644
--- a/kernel/irq/chip.c
+++ b/kernel/irq/chip.c
@@ -353,7 +353,9 @@ handle_level_irq(unsigned int irq, struct irq_desc *desc)
irqreturn_t action_ret;
spin_lock(&desc->lock);
+#ifndef CONFIG_IPIPE
mask_ack_irq(desc, irq);
+#endif /* CONFIG_IPIPE */
if (unlikely(desc->status & IRQ_INPROGRESS))
goto out_unlock;
@@ -430,8 +432,13 @@ handle_fasteoi_irq(unsigned int irq, struct irq_desc *desc)
spin_lock(&desc->lock);
desc->status &= ~IRQ_INPROGRESS;
+#ifdef CONFIG_IPIPE
+ desc->chip->unmask(irq);
+out:
+#else
out:
desc->chip->eoi(irq);
+#endif
spin_unlock(&desc->lock);
}
@@ -475,8 +482,10 @@ handle_edge_irq(unsigned int irq, struct irq_desc *desc)
kstat_cpu(cpu).irqs[irq]++;
+#ifndef CONFIG_IPIPE
/* Start handling the irq */
desc->chip->ack(irq);
+#endif /* CONFIG_IPIPE */
/* Mark the IRQ currently in progress.*/
desc->status |= IRQ_INPROGRESS;
@@ -516,6 +525,120 @@ out_unlock:
spin_unlock(&desc->lock);
}
+#ifdef CONFIG_IPIPE
+
+void fastcall __ipipe_ack_simple_irq(unsigned irq, struct irq_desc *desc)
+{
+}
+
+void fastcall __ipipe_end_simple_irq(unsigned irq, struct irq_desc *desc)
+{
+}
+
+void fastcall __ipipe_ack_level_irq(unsigned irq, struct irq_desc *desc)
+{
+ mask_ack_irq(desc, irq);
+}
+
+void fastcall __ipipe_end_level_irq(unsigned irq, struct irq_desc *desc)
+{
+ if (desc->chip->unmask)
+ desc->chip->unmask(irq);
+}
+
+void fastcall __ipipe_ack_fasteoi_irq(unsigned irq, struct irq_desc *desc)
+{
+ desc->chip->eoi(irq);
+}
+
+void fastcall __ipipe_end_fasteoi_irq(unsigned irq, struct irq_desc *desc)
+{
+ desc->chip->unmask(irq);
+}
+
+void fastcall __ipipe_ack_edge_irq(unsigned irq, struct irq_desc *desc)
+{
+ desc->chip->ack(irq);
+}
+
+void fastcall __ipipe_ack_percpu_irq(unsigned irq, struct irq_desc *desc)
+{
+ if (desc->chip->ack)
+ desc->chip->ack(irq);
+}
+
+void fastcall __ipipe_end_percpu_irq(unsigned irq, struct irq_desc *desc)
+{
+ if (desc->chip->eoi)
+ desc->chip->eoi(irq);
+}
+
+void fastcall __ipipe_end_edge_irq(unsigned irq, struct irq_desc *desc)
+{
+}
+
+void fastcall __ipipe_ack_demux_irq(unsigned irq, struct irq_desc *desc)
+{
+ /*
+ * Handling is delegated to some demultiplexer routine,
+ * e.g. GPIO. We mask_ack it, then call back into the demux
+ * handler, which should decode the interrupt and feed the
+ * pipeline as needed.
+ */
+ if (desc->chip->mask)
+ desc->chip->mask(irq);
+ desc->ipipe_demux(irq, desc);
+}
+
+void fastcall __ipipe_end_demux_irq(unsigned irq, struct irq_desc *desc)
+{
+ if (desc->chip->unmask)
+ desc->chip->unmask(irq);
+}
+
+void fastcall
+handle_demux_irq(unsigned int irq, struct irq_desc *desc)
+{
+ /*
+ * The regular IRQ handler will run last of all GPIO handlers,
+ * to unmask the demux IRQ.
+ */
+ __ipipe_end_demux_irq(irq, desc);
+}
+
+void __set_irq_demux_handler(unsigned int irq,
+ void fastcall (*decode)(unsigned int, struct irq_desc *),
+ int is_chained,
+ const char *name)
+{
+ struct irq_desc *desc = irq_desc + irq;
+ __set_irq_handler(irq, &handle_demux_irq, is_chained, name);
+ desc->ipipe_demux = decode;
+}
+
+void fastcall __ipipe_ack_bad_irq(unsigned irq, struct irq_desc *desc)
+{
+ static int done;
+
+ handle_bad_irq(irq, desc);
+
+ if (!done) {
+ printk(KERN_WARNING "%s: unknown flow handler for IRQ %d\n",
+ __FUNCTION__, irq);
+ done = 1;
+ }
+}
+
+void fastcall __ipipe_noack_irq(unsigned irq, struct irq_desc *desc)
+{
+}
+
+void fastcall __ipipe_noend_irq(unsigned irq, struct irq_desc *desc)
+{
+}
+
+#endif /* CONFIG_IPIPE */
+
/**
* handle_percpu_IRQ - Per CPU local irq handler
* @irq: the interrupt number
@@ -530,8 +653,10 @@ handle_percpu_irq(unsigned int irq, struct irq_desc *desc)
kstat_this_cpu.irqs[irq]++;
+#ifndef CONFIG_IPIPE
if (desc->chip->ack)
desc->chip->ack(irq);
+#endif /* CONFIG_IPIPE */
action_ret = handle_IRQ_event(irq, desc->action);
if (!noirqdebug)
@@ -558,6 +683,34 @@ __set_irq_handler(unsigned int irq, irq_flow_handler_t handle, int is_chained,
if (!handle)
handle = handle_bad_irq;
+#ifdef CONFIG_IPIPE
+ else if (handle == &handle_simple_irq) {
+ desc->ipipe_ack = &__ipipe_ack_simple_irq;
+ desc->ipipe_end = &__ipipe_end_simple_irq;
+ }
+ else if (handle == &handle_level_irq) {
+ desc->ipipe_ack = &__ipipe_ack_level_irq;
+ desc->ipipe_end = &__ipipe_end_level_irq;
+ }
+ else if (handle == &handle_edge_irq) {
+ desc->ipipe_ack = &__ipipe_ack_edge_irq;
+ desc->ipipe_end = &__ipipe_end_edge_irq;
+ }
+ else if (handle == &handle_fasteoi_irq) {
+ desc->ipipe_ack = &__ipipe_ack_fasteoi_irq;
+ desc->ipipe_end = &__ipipe_end_fasteoi_irq;
+ }
+#ifdef CONFIG_SMP
+ else if (handle == &handle_percpu_irq) {
+ desc->ipipe_ack = &__ipipe_ack_percpu_irq;
+ desc->ipipe_end = &__ipipe_end_percpu_irq;
+ }
+#endif /* CONFIG_SMP */
+ else if (handle == &handle_demux_irq) {
+ desc->ipipe_ack = &__ipipe_ack_demux_irq;
+ desc->ipipe_end = &__ipipe_end_demux_irq;
+ }
+#endif /* CONFIG_IPIPE */
else if (desc->chip == &no_irq_chip) {
printk(KERN_WARNING "Trying to install %sinterrupt handler "
"for IRQ%d\n", is_chained ? "chained " : "", irq);
@@ -569,7 +722,17 @@ __set_irq_handler(unsigned int irq, irq_flow_handler_t handle, int is_chained,
* dummy_irq_chip for easy transition.
*/
desc->chip = &dummy_irq_chip;
+#ifdef CONFIG_IPIPE
+ desc->ipipe_ack = &__ipipe_noack_irq;
+ desc->ipipe_end = &__ipipe_noend_irq;
+#endif /* CONFIG_IPIPE */
}
+#ifdef CONFIG_IPIPE
+ else {
+ desc->ipipe_ack = &__ipipe_ack_bad_irq;
+ desc->ipipe_end = &__ipipe_noend_irq;
+ }
+#endif /* CONFIG_IPIPE */
spin_lock_irqsave(&desc->lock, flags);
@@ -579,6 +742,10 @@ __set_irq_handler(unsigned int irq, irq_flow_handler_t handle, int is_chained,
mask_ack_irq(desc, irq);
desc->status |= IRQ_DISABLED;
desc->depth = 1;
+#ifdef CONFIG_IPIPE
+ desc->ipipe_ack = &__ipipe_ack_bad_irq;
+ desc->ipipe_end = &__ipipe_noend_irq;
+#endif /* CONFIG_IPIPE */
}
desc->handle_irq = handle;
desc->name = name;
diff --git a/kernel/power/disk.c b/kernel/power/disk.c
index 05b6479..8446b38 100644
--- a/kernel/power/disk.c
+++ b/kernel/power/disk.c
@@ -154,6 +154,7 @@ int create_image(int platform_mode)
return error;
local_irq_disable();
+ local_irq_disable_hw_cond();
/* At this point, device_suspend() has been called, but *not*
* device_power_down(). We *must* call device_power_down() now.
* Otherwise, drivers for some devices (e.g. interrupt controllers)
@@ -180,6 +181,7 @@ int create_image(int platform_mode)
*/
device_power_up();
Enable_irqs:
+ local_irq_enable_hw_cond();
local_irq_enable();
return error;
}
diff --git a/kernel/power/swsusp.c b/kernel/power/swsusp.c
index e1722d3..d7eeb4b 100644
--- a/kernel/power/swsusp.c
+++ b/kernel/power/swsusp.c
@@ -275,6 +275,7 @@ int swsusp_resume(void)
int error;
local_irq_disable();
+ local_irq_disable_hw_cond();
/* NOTE: device_power_down() is just a suspend() with irqs off;
* it has no special "power things down" semantics
*/
@@ -301,6 +302,7 @@ int swsusp_resume(void)
restore_processor_state();
touch_softlockup_watchdog();
device_power_up();
+ local_irq_enable_hw_cond();
local_irq_enable();
return error;
}
diff --git a/kernel/printk.c b/kernel/printk.c
index 89011bf..c217844 100644
--- a/kernel/printk.c
+++ b/kernel/printk.c
@@ -590,6 +590,41 @@ static int have_callable_console(void)
return 0;
}
+#ifdef CONFIG_IPIPE
+
+static ipipe_spinlock_t __ipipe_printk_lock = IPIPE_SPIN_LOCK_UNLOCKED;
+
+static int __ipipe_printk_fill;
+
+static char __ipipe_printk_buf[__LOG_BUF_LEN];
+
+void __ipipe_flush_printk (unsigned virq, void *cookie)
+{
+ char *p = __ipipe_printk_buf;
+ int len, lmax, out = 0;
+ unsigned long flags;
+
+ goto start;
+
+ do {
+ spin_unlock_irqrestore(&__ipipe_printk_lock, flags);
+ start:
+ lmax = __ipipe_printk_fill;
+ while (out < lmax) {
+ len = strlen(p) + 1;
+ printk("%s",p);
+ p += len;
+ out += len;
+ }
+ spin_lock_irqsave(&__ipipe_printk_lock, flags);
+ }
+ while (__ipipe_printk_fill != lmax);
+
+ __ipipe_printk_fill = 0;
+
+ spin_unlock_irqrestore(&__ipipe_printk_lock, flags);
+}
+
/**
* printk - print a kernel message
* @fmt: format string
@@ -615,6 +650,47 @@ static int have_callable_console(void)
asmlinkage int printk(const char *fmt, ...)
{
+ int r, fbytes, oldcount, cs = -1;
+ unsigned long flags;
+ va_list args;
+
+ va_start(args, fmt);
+
+ if (test_bit(IPIPE_SPRINTK_FLAG,&ipipe_current_domain->flags) ||
+ oops_in_progress)
+ cs = ipipe_disable_context_check(ipipe_processor_id());
+
+ if (ipipe_current_domain == ipipe_root_domain || cs != -1) {
+ r = vprintk(fmt, args);
+ if (cs != -1)
+ ipipe_restore_context_check(ipipe_processor_id(), cs);
+ goto out;
+ }
+
+ spin_lock_irqsave(&__ipipe_printk_lock, flags);
+
+ oldcount = __ipipe_printk_fill;
+ fbytes = __LOG_BUF_LEN - oldcount;
+
+ if (fbytes > 1) {
+ r = vscnprintf(__ipipe_printk_buf + __ipipe_printk_fill,
+ fbytes, fmt, args) + 1; /* account for the null byte */
+ __ipipe_printk_fill += r;
+ } else
+ r = 0;
+
+ spin_unlock_irqrestore(&__ipipe_printk_lock, flags);
+
+ if (oldcount == 0)
+ ipipe_trigger_irq(__ipipe_printk_virq);
+out:
+ va_end(args);
+
+ return r;
+}
+#else /* !CONFIG_IPIPE */
+asmlinkage int printk(const char *fmt, ...)
+{
va_list args;
int r;
@@ -624,6 +700,7 @@ asmlinkage int printk(const char *fmt, ...)
return r;
}
+#endif /* CONFIG_IPIPE */
/* cpu currently holding logbuf_lock */
static volatile unsigned int printk_cpu = UINT_MAX;
diff --git a/kernel/profile.c b/kernel/profile.c
index 5e95330..bc99d53 100644
--- a/kernel/profile.c
+++ b/kernel/profile.c
@@ -418,6 +418,7 @@ void profile_tick(int type)
if (type == CPU_PROFILING && timer_hook)
timer_hook(regs);
+
if (!user_mode(regs) && cpu_isset(smp_processor_id(), prof_cpu_mask))
profile_hit(type, (void *)profile_pc(regs));
}
diff --git a/kernel/sched.c b/kernel/sched.c
index 5ba5db9..e8f30b5 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -1518,7 +1518,7 @@ static int try_to_wake_up(struct task_struct *p, unsigned int state, int sync)
rq = task_rq_lock(p, &flags);
old_state = p->state;
- if (!(old_state & state))
+ if (!(old_state & state) || (old_state & TASK_NOWAKEUP))
goto out;
if (p->se.on_rq)
@@ -1920,13 +1920,15 @@ asmlinkage void schedule_tail(struct task_struct *prev)
#endif
if (current->set_child_tid)
put_user(task_pid_vnr(current), current->set_child_tid);
+
+ ipipe_init_notify(current);
}
/*
* context_switch - switch to the new MM and the new
* thread's register state.
*/
-static inline void
+static inline int
context_switch(struct rq *rq, struct task_struct *prev,
struct task_struct *next)
{
@@ -1967,12 +1969,17 @@ context_switch(struct rq *rq, struct task_struct *prev,
switch_to(prev, next, prev);
barrier();
+
+ if (task_hijacked(prev))
+ return 1;
/*
* this_rq must be evaluated again because prev may have moved
* CPUs since it called schedule(), thus the 'rq' on its stack
* frame will be invalid.
*/
finish_task_switch(this_rq(), prev);
+
+ return 0;
}
/*
@@ -3503,6 +3510,7 @@ void scheduler_tick(void)
void fastcall add_preempt_count(int val)
{
+ ipipe_check_context(ipipe_root_domain);
/*
* Underflow?
*/
@@ -3519,6 +3527,7 @@ EXPORT_SYMBOL(add_preempt_count);
void fastcall sub_preempt_count(int val)
{
+ ipipe_check_context(ipipe_root_domain);
/*
* Underflow?
*/
@@ -3630,6 +3639,11 @@ need_resched:
rcu_qsctr_inc(cpu);
prev = rq->curr;
switch_count = &prev->nivcsw;
+ if (unlikely(prev->state & TASK_ATOMICSWITCH)) {
+ prev->state &= ~TASK_ATOMICSWITCH;
+ /* Pop one disable level -- one still remains. */
+ preempt_enable();
+ }
release_kernel_lock(prev);
need_resched_nonpreemptible:
@@ -3667,7 +3681,8 @@ need_resched_nonpreemptible:
rq->curr = next;
++*switch_count;
- context_switch(rq, prev, next); /* unlocks the rq */
+ if (context_switch(rq, prev, next)) /* unlocks the rq unless hijacked */
+ return;
} else
spin_unlock_irq(&rq->lock);
@@ -3695,6 +3710,7 @@ asmlinkage void __sched preempt_schedule(void)
struct task_struct *task = current;
int saved_lock_depth;
#endif
+ ipipe_check_context(ipipe_root_domain);
/*
* If there is a non-zero preempt_count or interrupts are disabled,
* we do not want to preempt the current task. Just return..
@@ -4345,6 +4361,7 @@ recheck:
oldprio = p->prio;
__setscheduler(rq, p, policy, param->sched_priority);
+ ipipe_setsched_notify(p);
if (running)
p->sched_class->set_curr_task(rq);
@@ -7398,3 +7415,66 @@ struct cgroup_subsys cpuacct_subsys = {
.subsys_id = cpuacct_subsys_id,
};
#endif /* CONFIG_CGROUP_CPUACCT */
+
+#ifdef CONFIG_IPIPE
+
+int ipipe_setscheduler_root (struct task_struct *p, int policy, int prio)
+{
+ int oldprio, on_rq, running;
+ unsigned long flags;
+ struct rq *rq;
+
+ spin_lock_irqsave(&p->pi_lock, flags);
+ rq = __task_rq_lock(p);
+ update_rq_clock(rq);
+ on_rq = p->se.on_rq;
+ running = task_running(rq, p);
+
+ if (on_rq) {
+ deactivate_task(rq, p, 0);
+ if (running)
+ p->sched_class->put_prev_task(rq, p);
+ }
+
+ oldprio = p->prio;
+ __setscheduler(rq, p, policy, prio);
+ ipipe_setsched_notify(p);
+
+ if (on_rq) {
+ if (running)
+ p->sched_class->set_curr_task(rq);
+ activate_task(rq, p, 0);
+
+ if (running) {
+ if (p->prio > oldprio)
+ resched_task(rq->curr);
+ } else {
+ check_preempt_curr(rq, p);
+ }
+ }
+ __task_rq_unlock(rq);
+ spin_unlock_irqrestore(&p->pi_lock, flags);
+
+ rt_mutex_adjust_pi(p);
+
+ return 0;
+}
+
+EXPORT_SYMBOL(ipipe_setscheduler_root);
+
+int ipipe_reenter_root (struct task_struct *prev, int policy, int prio)
+{
+ finish_task_switch(this_rq(), prev);
+
+ (void)reacquire_kernel_lock(current);
+ preempt_enable_no_resched();
+
+ if (current->policy != policy || current->rt_priority != prio)
+ return ipipe_setscheduler_root(current, policy, prio);
+
+ return 0;
+}
+
+EXPORT_SYMBOL(ipipe_reenter_root);
+
+#endif /* CONFIG_IPIPE */
diff --git a/kernel/signal.c b/kernel/signal.c
index afa4f78..e1eb7ae 100644
--- a/kernel/signal.c
+++ b/kernel/signal.c
@@ -454,6 +454,7 @@ void signal_wake_up(struct task_struct *t, int resume)
unsigned int mask;
set_tsk_thread_flag(t, TIF_SIGPENDING);
+ ipipe_sigwake_notify(t); /* TIF_SIGPENDING must be set first. */
/*
* For SIGKILL, we want to wake it up in the stopped/traced case.
diff --git a/kernel/spinlock.c b/kernel/spinlock.c
index cd72424..3e74c9b 100644
--- a/kernel/spinlock.c
+++ b/kernel/spinlock.c
@@ -88,7 +88,7 @@ unsigned long __lockfunc _spin_lock_irqsave(spinlock_t *lock)
* _raw_spin_lock_flags() code, because lockdep assumes
* that interrupts are not re-enabled during lock-acquire:
*/
-#ifdef CONFIG_LOCKDEP
+#if defined(CONFIG_LOCKDEP) || defined(CONFIG_IPIPE)
LOCK_CONTENDED(lock, _raw_spin_trylock, _raw_spin_lock);
#else
_raw_spin_lock_flags(lock, &flags);
@@ -305,7 +305,7 @@ unsigned long __lockfunc _spin_lock_irqsave_nested(spinlock_t *lock, int subclas
* _raw_spin_lock_flags() code, because lockdep assumes
* that interrupts are not re-enabled during lock-acquire:
*/
-#ifdef CONFIG_LOCKDEP
+#if defined(CONFIG_LOCKDEP) || defined(CONFIG_IPIPE)
LOCK_CONTENDED(lock, _raw_spin_trylock, _raw_spin_lock);
#else
_raw_spin_lock_flags(lock, &flags);
diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
index a601093..6372a6f 100644
--- a/lib/Kconfig.debug
+++ b/lib/Kconfig.debug
@@ -79,6 +79,8 @@ config HEADERS_CHECK
exported to $(INSTALL_HDR_PATH) (usually 'usr/include' in
your build tree), to make sure they're suitable.
+source "kernel/ipipe/Kconfig.debug"
+
config DEBUG_KERNEL
bool "Kernel debugging"
help
diff --git a/lib/bust_spinlocks.c b/lib/bust_spinlocks.c
index 486da62..0d33232 100644
--- a/lib/bust_spinlocks.c
+++ b/lib/bust_spinlocks.c
@@ -12,16 +12,19 @@
#include <linux/tty.h>
#include <linux/wait.h>
#include <linux/vt_kern.h>
+#include <linux/ipipe_trace.h>
void __attribute__((weak)) bust_spinlocks(int yes)
{
if (yes) {
+ ipipe_trace_panic_freeze();
++oops_in_progress;
} else {
#ifdef CONFIG_VT
unblank_screen();
#endif
+ ipipe_trace_panic_dump();
if (--oops_in_progress == 0)
wake_up_klogd();
}
diff --git a/lib/ioremap.c b/lib/ioremap.c
index 14c6078..a275469 100644
--- a/lib/ioremap.c
+++ b/lib/ioremap.c
@@ -85,8 +85,8 @@ int ioremap_page_range(unsigned long addr,
if (err)
break;
} while (pgd++, addr = next, addr != end);
-
- flush_cache_vmap(start, end);
+ __ipipe_pin_range_globally(start, end);
+ flush_cache_vmap(start, end);
return err;
}
diff --git a/lib/smp_processor_id.c b/lib/smp_processor_id.c
index eddc9b3..66302ec 100644
--- a/lib/smp_processor_id.c
+++ b/lib/smp_processor_id.c
@@ -13,10 +13,13 @@ unsigned int debug_smp_processor_id(void)
int this_cpu = raw_smp_processor_id();
cpumask_t this_mask;
+ if (!ipipe_root_domain_p)
+ goto out;
+
if (likely(preempt_count))
goto out;
- if (irqs_disabled())
+ if (irqs_disabled() || irqs_disabled_hw())
goto out;
/*
diff --git a/lib/spinlock_debug.c b/lib/spinlock_debug.c
index 9c4b025..08f096b 100644
--- a/lib/spinlock_debug.c
+++ b/lib/spinlock_debug.c
@@ -133,6 +133,8 @@ void _raw_spin_lock(spinlock_t *lock)
debug_spin_lock_after(lock);
}
+EXPORT_SYMBOL(_raw_spin_lock);
+
int _raw_spin_trylock(spinlock_t *lock)
{
int ret = __raw_spin_trylock(&lock->raw_lock);
@@ -148,12 +150,16 @@ int _raw_spin_trylock(spinlock_t *lock)
return ret;
}
+EXPORT_SYMBOL(_raw_spin_trylock);
+
void _raw_spin_unlock(spinlock_t *lock)
{
debug_spin_unlock(lock);
__raw_spin_unlock(&lock->raw_lock);
}
+EXPORT_SYMBOL(_raw_spin_unlock);
+
static void rwlock_bug(rwlock_t *lock, const char *msg)
{
if (!debug_locks_off())
@@ -199,6 +205,8 @@ void _raw_read_lock(rwlock_t *lock)
__raw_read_lock(&lock->raw_lock);
}
+EXPORT_SYMBOL(_raw_read_lock);
+
int _raw_read_trylock(rwlock_t *lock)
{
int ret = __raw_read_trylock(&lock->raw_lock);
@@ -212,12 +220,16 @@ int _raw_read_trylock(rwlock_t *lock)
return ret;
}
+EXPORT_SYMBOL(_raw_read_trylock);
+
void _raw_read_unlock(rwlock_t *lock)
{
RWLOCK_BUG_ON(lock->magic != RWLOCK_MAGIC, lock, "bad magic");
__raw_read_unlock(&lock->raw_lock);
}
+EXPORT_SYMBOL(_raw_read_unlock);
+
static inline void debug_write_lock_before(rwlock_t *lock)
{
RWLOCK_BUG_ON(lock->magic != RWLOCK_MAGIC, lock, "bad magic");
@@ -275,6 +287,8 @@ void _raw_write_lock(rwlock_t *lock)
debug_write_lock_after(lock);
}
+EXPORT_SYMBOL(_raw_write_lock);
+
int _raw_write_trylock(rwlock_t *lock)
{
int ret = __raw_write_trylock(&lock->raw_lock);
@@ -290,8 +304,12 @@ int _raw_write_trylock(rwlock_t *lock)
return ret;
}
+EXPORT_SYMBOL(_raw_write_trylock);
+
void _raw_write_unlock(rwlock_t *lock)
{
debug_write_unlock(lock);
__raw_write_unlock(&lock->raw_lock);
}
+
+EXPORT_SYMBOL(_raw_write_unlock);
diff --git a/mm/memory.c b/mm/memory.c
index da8b74b..87ea304 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -50,6 +50,7 @@
#include <linux/delayacct.h>
#include <linux/init.h>
#include <linux/writeback.h>
+#include <linux/vmalloc.h>
#include <asm/pgalloc.h>
#include <asm/uaccess.h>
@@ -415,6 +416,34 @@ struct page *vm_normal_page(struct vm_area_struct *vma, unsigned long addr, pte_
return pfn_to_page(pfn);
}
+static inline void cow_user_page(struct page *dst, struct page *src, unsigned long va, struct vm_area_struct *vma)
+{
+ /*
+ * If the source page was a PFN mapping, we don't have
+ * a "struct page" for it. We do a best-effort copy by
+ * just copying from the original user address. If that
+ * fails, we just zero-fill it. Live with it.
+ */
+ if (unlikely(!src)) {
+ void *kaddr = kmap_atomic(dst, KM_USER0);
+ void __user *uaddr = (void __user *)(va & PAGE_MASK);
+
+ /*
+ * This really shouldn't fail, because the page is there
+ * in the page tables. But it might just be unreadable,
+ * in which case we just give up and fill the result with
+ * zeroes.
+ */
+ if (__copy_from_user_inatomic(kaddr, uaddr, PAGE_SIZE))
+ memset(kaddr, 0, PAGE_SIZE);
+ kunmap_atomic(kaddr, KM_USER0);
+ flush_dcache_page(dst);
+ return;
+
+ }
+ copy_user_highpage(dst, src, va, vma);
+}
+
/*
* copy one vm_area from one task to the other. Assumes the page tables
* already present in the new task to be cleared in the whole range
@@ -423,8 +452,8 @@ struct page *vm_normal_page(struct vm_area_struct *vma, unsigned long addr, pte_
static inline void
copy_one_pte(struct mm_struct *dst_mm, struct mm_struct *src_mm,
- pte_t *dst_pte, pte_t *src_pte, struct vm_area_struct *vma,
- unsigned long addr, int *rss)
+ pte_t *dst_pte, pte_t *src_pte, struct vm_area_struct *vma,
+ unsigned long addr, int *rss, struct page *uncow_page)
{
unsigned long vm_flags = vma->vm_flags;
pte_t pte = *src_pte;
@@ -463,6 +492,21 @@ copy_one_pte(struct mm_struct *dst_mm, struct mm_struct *src_mm,
* in the parent and the child
*/
if (is_cow_mapping(vm_flags)) {
+#ifdef CONFIG_IPIPE
+ if (uncow_page) {
+ struct page *old_page = vm_normal_page(vma, addr, pte);
+ cow_user_page(uncow_page, old_page, addr, vma);
+ pte = mk_pte(uncow_page, vma->vm_page_prot);
+
+ if (vm_flags & VM_SHARED)
+ pte = pte_mkclean(pte);
+ pte = pte_mkold(pte);
+
+ page_dup_rmap(uncow_page, vma, addr);
+ rss[!!PageAnon(uncow_page)]++;
+ goto out_set_pte;
+ }
+#endif /* CONFIG_IPIPE */
ptep_set_wrprotect(src_mm, addr, src_pte);
pte = pte_wrprotect(pte);
}
@@ -493,13 +537,27 @@ static int copy_pte_range(struct mm_struct *dst_mm, struct mm_struct *src_mm,
pte_t *src_pte, *dst_pte;
spinlock_t *src_ptl, *dst_ptl;
int progress = 0;
+ struct page *uncow_page = NULL;
int rss[2];
-
+#ifdef CONFIG_IPIPE
+ int do_cow_break = 0;
+again:
+ if (do_cow_break) {
+ uncow_page = alloc_page_vma(GFP_HIGHUSER, vma, addr);
+ if (!uncow_page)
+ return -ENOMEM;
+ do_cow_break = 0;
+ }
+#else
again:
+#endif
rss[1] = rss[0] = 0;
dst_pte = pte_alloc_map_lock(dst_mm, dst_pmd, addr, &dst_ptl);
- if (!dst_pte)
+ if (!dst_pte) {
+ if (uncow_page)
+ page_cache_release(uncow_page);
return -ENOMEM;
+ }
src_pte = pte_offset_map_nested(src_pmd, addr);
src_ptl = pte_lockptr(src_mm, src_pmd);
spin_lock_nested(src_ptl, SINGLE_DEPTH_NESTING);
@@ -521,7 +579,20 @@ again:
progress++;
continue;
}
- copy_one_pte(dst_mm, src_mm, dst_pte, src_pte, vma, addr, rss);
+#ifdef CONFIG_IPIPE
+ if (likely(uncow_page == NULL) && likely(pte_present(*src_pte))) {
+ if (is_cow_mapping(vma->vm_flags)) {
+ if (((vma->vm_flags|src_mm->def_flags) & (VM_LOCKED|VM_PINNED))
+ == (VM_LOCKED|VM_PINNED)) {
+ do_cow_break = 1;
+ break;
+ }
+ }
+ }
+#endif
+ copy_one_pte(dst_mm, src_mm, dst_pte,
+ src_pte, vma, addr, rss, uncow_page);
+ uncow_page = NULL;
progress += 8;
} while (dst_pte++, src_pte++, addr += PAGE_SIZE, addr != end);
@@ -1498,34 +1569,6 @@ static inline pte_t maybe_mkwrite(pte_t pte, struct vm_area_struct *vma)
return pte;
}
-static inline void cow_user_page(struct page *dst, struct page *src, unsigned long va, struct vm_area_struct *vma)
-{
- /*
- * If the source page was a PFN mapping, we don't have
- * a "struct page" for it. We do a best-effort copy by
- * just copying from the original user address. If that
- * fails, we just zero-fill it. Live with it.
- */
- if (unlikely(!src)) {
- void *kaddr = kmap_atomic(dst, KM_USER0);
- void __user *uaddr = (void __user *)(va & PAGE_MASK);
-
- /*
- * This really shouldn't fail, because the page is there
- * in the page tables. But it might just be unreadable,
- * in which case we just give up and fill the result with
- * zeroes.
- */
- if (__copy_from_user_inatomic(kaddr, uaddr, PAGE_SIZE))
- memset(kaddr, 0, PAGE_SIZE);
- kunmap_atomic(kaddr, KM_USER0);
- flush_dcache_page(dst);
- return;
-
- }
- copy_user_highpage(dst, src, va, vma);
-}
-
/*
* This routine handles present pages, when users try to write
* to a shared page. It is done by copying the page to a new address
@@ -2758,3 +2801,110 @@ int access_process_vm(struct task_struct *tsk, unsigned long addr, void *buf, in
return buf - old_buf;
}
+
+#ifdef CONFIG_IPIPE
+
+static inline int ipipe_pin_pte_range(struct mm_struct *mm, pmd_t *pmd,
+ struct vm_area_struct *vma,
+ unsigned long addr, unsigned long end)
+{
+ spinlock_t *ptl;
+ pte_t *pte;
+
+ do {
+ pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
+ if (!pte)
+ continue;
+
+ if (!pte_present(*pte)) {
+ pte_unmap_unlock(pte, ptl);
+ continue;
+ }
+
+ if (do_wp_page(mm, vma, addr, pte, pmd, ptl, *pte) == VM_FAULT_OOM)
+ return -ENOMEM;
+ } while (addr += PAGE_SIZE, addr != end);
+ return 0;
+}
+
+static inline int ipipe_pin_pmd_range(struct mm_struct *mm, pud_t *pud,
+ struct vm_area_struct *vma,
+ unsigned long addr, unsigned long end)
+{
+ unsigned long next;
+ pmd_t *pmd;
+
+ pmd = pmd_offset(pud, addr);
+ do {
+ next = pmd_addr_end(addr, end);
+ if (pmd_none_or_clear_bad(pmd))
+ continue;
+ if (ipipe_pin_pte_range(mm, pmd, vma, addr, next))
+ return -ENOMEM;
+ } while (pmd++, addr = next, addr != end);
+ return 0;
+}
+
+static inline int ipipe_pin_pud_range(struct mm_struct *mm, pgd_t *pgd,
+ struct vm_area_struct *vma,
+ unsigned long addr, unsigned long end)
+{
+ unsigned long next;
+ pud_t *pud;
+
+ pud = pud_offset(pgd, addr);
+ do {
+ next = pud_addr_end(addr, end);
+ if (pud_none_or_clear_bad(pud))
+ continue;
+ if (ipipe_pin_pmd_range(mm, pud, vma, addr, next))
+ return -ENOMEM;
+ } while (pud++, addr = next, addr != end);
+ return 0;
+}
+
+int ipipe_disable_ondemand_mappings(struct task_struct *tsk)
+{
+ unsigned long addr, next, end;
+ struct vm_area_struct *vma;
+ struct mm_struct *mm;
+ int result = 0;
+ pgd_t *pgd;
+
+ mm = get_task_mm(tsk);
+ if (!mm)
+ return -EPERM;
+
+ down_write(&mm->mmap_sem);
+ if (mm->def_flags & VM_PINNED)
+ goto done_mm;
+
+ for (vma = mm->mmap; vma; vma = vma->vm_next) {
+ if (!is_cow_mapping(vma->vm_flags))
+ continue;
+
+ addr = vma->vm_start;
+ end = vma->vm_end;
+
+ pgd = pgd_offset(mm, addr);
+ do {
+ next = pgd_addr_end(addr, end);
+ if (pgd_none_or_clear_bad(pgd))
+ continue;
+ if (ipipe_pin_pud_range(mm, pgd, vma, addr, next)) {
+ result = -ENOMEM;
+ goto done_mm;
+ }
+ } while (pgd++, addr = next, addr != end);
+ }
+ mm->def_flags |= VM_PINNED;
+
+ done_mm:
+ up_write(&mm->mmap_sem);
+ mmput(mm);
+ return result;
+}
+
+EXPORT_SYMBOL(ipipe_disable_ondemand_mappings);
+
+#endif
diff --git a/mm/mlock.c b/mm/mlock.c
index 7b26560..d2a318a 100644
--- a/mm/mlock.c
+++ b/mm/mlock.c
@@ -173,10 +173,10 @@ asmlinkage long sys_munlock(unsigned long start, size_t len)
static int do_mlockall(int flags)
{
struct vm_area_struct * vma, * prev = NULL;
- unsigned int def_flags = 0;
+ unsigned int def_flags = current->mm->def_flags & VM_PINNED;
if (flags & MCL_FUTURE)
- def_flags = VM_LOCKED;
+ def_flags |= VM_LOCKED;
current->mm->def_flags = def_flags;
if (flags == MCL_FUTURE)
goto out;
diff --git a/mm/vmalloc.c b/mm/vmalloc.c
index af77e17..5bfff7a 100644
--- a/mm/vmalloc.c
+++ b/mm/vmalloc.c
@@ -161,6 +161,7 @@ int map_vm_area(struct vm_struct *area, pgprot_t prot, struct page ***pages)
if (err)
break;
} while (pgd++, addr = next, addr != end);
+ __ipipe_pin_range_globally((unsigned long) area->addr, end);
flush_cache_vmap((unsigned long) area->addr, end);
return err;
}
[-- Attachment #3: config-2.6.24-malta-network-adeos --]
[-- Type: application/octet-stream, Size: 22540 bytes --]
#
# Automatically generated make config: don't edit
# Linux kernel version: 2.6.24.4
# Thu Jul 10 10:51:27 2008
#
CONFIG_MIPS=y
#
# Machine selection
#
CONFIG_ZONE_DMA=y
# CONFIG_MACH_ALCHEMY is not set
# CONFIG_BASLER_EXCITE is not set
# CONFIG_BCM47XX is not set
# CONFIG_MIPS_COBALT is not set
# CONFIG_MACH_DECSTATION is not set
# CONFIG_MACH_JAZZ is not set
# CONFIG_LASAT is not set
# CONFIG_LEMOTE_FULONG is not set
# CONFIG_MIPS_ATLAS is not set
CONFIG_MIPS_MALTA=y
# CONFIG_MIPS_SEAD is not set
# CONFIG_MIPS_SIM is not set
# CONFIG_MARKEINS is not set
# CONFIG_MACH_VR41XX is not set
# CONFIG_PNX8550_JBS is not set
# CONFIG_PNX8550_STB810 is not set
# CONFIG_PMC_MSP is not set
# CONFIG_PMC_YOSEMITE is not set
# CONFIG_QEMU is not set
# CONFIG_SGI_IP22 is not set
# CONFIG_SGI_IP27 is not set
# CONFIG_SGI_IP32 is not set
# CONFIG_SIBYTE_CRHINE is not set
# CONFIG_SIBYTE_CARMEL is not set
# CONFIG_SIBYTE_CRHONE is not set
# CONFIG_SIBYTE_RHONE is not set
# CONFIG_SIBYTE_SWARM is not set
# CONFIG_SIBYTE_LITTLESUR is not set
# CONFIG_SIBYTE_SENTOSA is not set
# CONFIG_SIBYTE_PTSWARM is not set
# CONFIG_SIBYTE_BIGSUR is not set
# CONFIG_SNI_RM is not set
# CONFIG_TOSHIBA_JMR3927 is not set
# CONFIG_TOSHIBA_RBTX4927 is not set
# CONFIG_TOSHIBA_RBTX4938 is not set
# CONFIG_WR_PPMC is not set
CONFIG_RWSEM_GENERIC_SPINLOCK=y
# CONFIG_ARCH_HAS_ILOG2_U32 is not set
# CONFIG_ARCH_HAS_ILOG2_U64 is not set
CONFIG_ARCH_SUPPORTS_OPROFILE=y
CONFIG_GENERIC_FIND_NEXT_BIT=y
CONFIG_GENERIC_HWEIGHT=y
CONFIG_GENERIC_CALIBRATE_DELAY=y
CONFIG_GENERIC_CLOCKEVENTS=y
CONFIG_GENERIC_TIME=y
CONFIG_GENERIC_CMOS_UPDATE=y
CONFIG_SCHED_NO_NO_OMIT_FRAME_POINTER=y
# CONFIG_GENERIC_HARDIRQS_NO__DO_IRQ is not set
CONFIG_ARCH_MAY_HAVE_PC_FDC=y
CONFIG_CEVT_R4K=y
CONFIG_CSRC_R4K=y
CONFIG_DMA_NONCOHERENT=y
CONFIG_DMA_NEED_PCI_MAP_STATE=y
CONFIG_EARLY_PRINTK=y
CONFIG_SYS_HAS_EARLY_PRINTK=y
# CONFIG_HOTPLUG_CPU is not set
CONFIG_I8259=y
CONFIG_MIPS_BONITO64=y
CONFIG_MIPS_MSC=y
# CONFIG_NO_IOPORT is not set
CONFIG_GENERIC_ISA_DMA=y
CONFIG_CPU_BIG_ENDIAN=y
# CONFIG_CPU_LITTLE_ENDIAN is not set
CONFIG_SYS_SUPPORTS_BIG_ENDIAN=y
CONFIG_SYS_SUPPORTS_LITTLE_ENDIAN=y
CONFIG_IRQ_CPU=y
CONFIG_MIPS_BOARDS_GEN=y
CONFIG_PCI_GT64XXX_PCI0=y
CONFIG_SWAP_IO_SPACE=y
CONFIG_BOOT_ELF32=y
CONFIG_MIPS_L1_CACHE_SHIFT=5
#
# CPU selection
#
# CONFIG_CPU_LOONGSON2 is not set
CONFIG_CPU_MIPS32_R1=y
# CONFIG_CPU_MIPS32_R2 is not set
# CONFIG_CPU_MIPS64_R1 is not set
# CONFIG_CPU_MIPS64_R2 is not set
# CONFIG_CPU_R3000 is not set
# CONFIG_CPU_TX39XX is not set
# CONFIG_CPU_VR41XX is not set
# CONFIG_CPU_R4300 is not set
# CONFIG_CPU_R4X00 is not set
# CONFIG_CPU_TX49XX is not set
# CONFIG_CPU_R5000 is not set
# CONFIG_CPU_R5432 is not set
# CONFIG_CPU_R6000 is not set
# CONFIG_CPU_NEVADA is not set
# CONFIG_CPU_R8000 is not set
# CONFIG_CPU_R10000 is not set
# CONFIG_CPU_RM7000 is not set
# CONFIG_CPU_RM9000 is not set
# CONFIG_CPU_SB1 is not set
CONFIG_SYS_HAS_CPU_MIPS32_R1=y
CONFIG_SYS_HAS_CPU_MIPS32_R2=y
CONFIG_SYS_HAS_CPU_MIPS64_R1=y
CONFIG_SYS_HAS_CPU_NEVADA=y
CONFIG_SYS_HAS_CPU_RM7000=y
CONFIG_CPU_MIPS32=y
CONFIG_CPU_MIPSR1=y
CONFIG_SYS_SUPPORTS_32BIT_KERNEL=y
CONFIG_SYS_SUPPORTS_64BIT_KERNEL=y
CONFIG_CPU_SUPPORTS_32BIT_KERNEL=y
#
# Kernel type
#
CONFIG_32BIT=y
# CONFIG_64BIT is not set
CONFIG_PAGE_SIZE_4KB=y
# CONFIG_PAGE_SIZE_8KB is not set
# CONFIG_PAGE_SIZE_16KB is not set
# CONFIG_PAGE_SIZE_64KB is not set
CONFIG_BOARD_SCACHE=y
CONFIG_MIPS_CPU_SCACHE=y
CONFIG_CPU_HAS_PREFETCH=y
CONFIG_MIPS_MT_DISABLED=y
# CONFIG_MIPS_MT_SMP is not set
# CONFIG_MIPS_MT_SMTC is not set
CONFIG_SYS_SUPPORTS_MULTITHREADING=y
# CONFIG_MIPS_VPE_LOADER is not set
CONFIG_CPU_HAS_LLSC=y
# CONFIG_CPU_HAS_SMARTMIPS is not set
CONFIG_CPU_HAS_SYNC=y
CONFIG_GENERIC_HARDIRQS=y
CONFIG_GENERIC_IRQ_PROBE=y
CONFIG_CPU_SUPPORTS_HIGHMEM=y
CONFIG_SYS_SUPPORTS_SMARTMIPS=y
CONFIG_ARCH_FLATMEM_ENABLE=y
CONFIG_ARCH_POPULATES_NODE_MAP=y
CONFIG_SELECT_MEMORY_MODEL=y
CONFIG_FLATMEM_MANUAL=y
# CONFIG_DISCONTIGMEM_MANUAL is not set
# CONFIG_SPARSEMEM_MANUAL is not set
CONFIG_FLATMEM=y
CONFIG_FLAT_NODE_MEM_MAP=y
# CONFIG_SPARSEMEM_STATIC is not set
# CONFIG_SPARSEMEM_VMEMMAP_ENABLE is not set
CONFIG_SPLIT_PTLOCK_CPUS=4
# CONFIG_RESOURCES_64BIT is not set
CONFIG_ZONE_DMA_FLAG=1
CONFIG_BOUNCE=y
CONFIG_VIRT_TO_BUS=y
# CONFIG_TICK_ONESHOT is not set
# CONFIG_NO_HZ is not set
# CONFIG_HIGH_RES_TIMERS is not set
CONFIG_GENERIC_CLOCKEVENTS_BUILD=y
# CONFIG_HZ_48 is not set
# CONFIG_HZ_100 is not set
# CONFIG_HZ_128 is not set
CONFIG_HZ_250=y
# CONFIG_HZ_256 is not set
# CONFIG_HZ_1000 is not set
# CONFIG_HZ_1024 is not set
CONFIG_SYS_SUPPORTS_ARBIT_HZ=y
CONFIG_HZ=250
CONFIG_PREEMPT_NONE=y
# CONFIG_PREEMPT_VOLUNTARY is not set
# CONFIG_PREEMPT is not set
CONFIG_IPIPE=y
CONFIG_IPIPE_DOMAINS=4
CONFIG_IPIPE_COMPAT=y
# CONFIG_KEXEC is not set
CONFIG_SECCOMP=y
CONFIG_LOCKDEP_SUPPORT=y
CONFIG_STACKTRACE_SUPPORT=y
CONFIG_DEFCONFIG_LIST="/lib/modules/$UNAME_RELEASE/.config"
#
# General setup
#
CONFIG_EXPERIMENTAL=y
CONFIG_BROKEN_ON_SMP=y
CONFIG_INIT_ENV_ARG_LIMIT=32
CONFIG_LOCALVERSION=""
CONFIG_LOCALVERSION_AUTO=y
CONFIG_SWAP=y
CONFIG_SYSVIPC=y
CONFIG_SYSVIPC_SYSCTL=y
# CONFIG_POSIX_MQUEUE is not set
# CONFIG_BSD_PROCESS_ACCT is not set
# CONFIG_TASKSTATS is not set
# CONFIG_USER_NS is not set
# CONFIG_PID_NS is not set
# CONFIG_AUDIT is not set
CONFIG_IKCONFIG=y
CONFIG_IKCONFIG_PROC=y
CONFIG_LOG_BUF_SHIFT=14
# CONFIG_CGROUPS is not set
CONFIG_FAIR_GROUP_SCHED=y
CONFIG_FAIR_USER_SCHED=y
# CONFIG_FAIR_CGROUP_SCHED is not set
CONFIG_SYSFS_DEPRECATED=y
# CONFIG_RELAY is not set
# CONFIG_BLK_DEV_INITRD is not set
# CONFIG_CC_OPTIMIZE_FOR_SIZE is not set
CONFIG_SYSCTL=y
CONFIG_EMBEDDED=y
CONFIG_SYSCTL_SYSCALL=y
CONFIG_KALLSYMS=y
# CONFIG_KALLSYMS_ALL is not set
CONFIG_KALLSYMS_EXTRA_PASS=y
CONFIG_HOTPLUG=y
CONFIG_PRINTK=y
CONFIG_BUG=y
CONFIG_ELF_CORE=y
CONFIG_BASE_FULL=y
CONFIG_FUTEX=y
CONFIG_ANON_INODES=y
CONFIG_EPOLL=y
CONFIG_SIGNALFD=y
CONFIG_EVENTFD=y
CONFIG_SHMEM=y
CONFIG_VM_EVENT_COUNTERS=y
# CONFIG_SLAB is not set
# CONFIG_SLUB is not set
CONFIG_SLOB=y
CONFIG_RT_MUTEXES=y
# CONFIG_TINY_SHMEM is not set
CONFIG_BASE_SMALL=0
CONFIG_MODULES=y
CONFIG_MODULE_UNLOAD=y
CONFIG_MODULE_FORCE_UNLOAD=y
CONFIG_MODVERSIONS=y
# CONFIG_MODULE_SRCVERSION_ALL is not set
# CONFIG_KMOD is not set
CONFIG_BLOCK=y
# CONFIG_LBD is not set
# CONFIG_BLK_DEV_IO_TRACE is not set
# CONFIG_LSF is not set
# CONFIG_BLK_DEV_BSG is not set
#
# IO Schedulers
#
CONFIG_IOSCHED_NOOP=y
CONFIG_IOSCHED_AS=y
CONFIG_IOSCHED_DEADLINE=y
CONFIG_IOSCHED_CFQ=y
# CONFIG_DEFAULT_AS is not set
# CONFIG_DEFAULT_DEADLINE is not set
CONFIG_DEFAULT_CFQ=y
# CONFIG_DEFAULT_NOOP is not set
CONFIG_DEFAULT_IOSCHED="cfq"
#
# Bus options (PCI, PCMCIA, EISA, ISA, TC)
#
CONFIG_HW_HAS_PCI=y
CONFIG_PCI=y
CONFIG_PCI_DOMAINS=y
# CONFIG_ARCH_SUPPORTS_MSI is not set
CONFIG_PCI_LEGACY=y
# CONFIG_PCI_DEBUG is not set
CONFIG_MMU=y
CONFIG_I8253=y
# CONFIG_PCCARD is not set
# CONFIG_HOTPLUG_PCI is not set
#
# Executable file formats
#
CONFIG_BINFMT_ELF=y
CONFIG_BINFMT_MISC=y
CONFIG_TRAD_SIGNALS=y
#
# Power management options
#
# CONFIG_PM is not set
CONFIG_SUSPEND_UP_POSSIBLE=y
#
# Networking
#
CONFIG_NET=y
#
# Networking options
#
CONFIG_PACKET=y
CONFIG_PACKET_MMAP=y
CONFIG_UNIX=y
# CONFIG_NET_KEY is not set
CONFIG_INET=y
# CONFIG_IP_MULTICAST is not set
# CONFIG_IP_ADVANCED_ROUTER is not set
CONFIG_IP_FIB_HASH=y
# CONFIG_IP_PNP is not set
# CONFIG_NET_IPIP is not set
# CONFIG_NET_IPGRE is not set
# CONFIG_ARPD is not set
# CONFIG_SYN_COOKIES is not set
# CONFIG_INET_AH is not set
# CONFIG_INET_ESP is not set
# CONFIG_INET_IPCOMP is not set
# CONFIG_INET_XFRM_TUNNEL is not set
# CONFIG_INET_TUNNEL is not set
# CONFIG_INET_XFRM_MODE_TRANSPORT is not set
# CONFIG_INET_XFRM_MODE_TUNNEL is not set
# CONFIG_INET_XFRM_MODE_BEET is not set
# CONFIG_INET_LRO is not set
# CONFIG_INET_DIAG is not set
# CONFIG_TCP_CONG_ADVANCED is not set
CONFIG_TCP_CONG_CUBIC=y
CONFIG_DEFAULT_TCP_CONG="cubic"
# CONFIG_TCP_MD5SIG is not set
# CONFIG_IPV6 is not set
# CONFIG_INET6_XFRM_TUNNEL is not set
# CONFIG_INET6_TUNNEL is not set
# CONFIG_NETLABEL is not set
# CONFIG_NETWORK_SECMARK is not set
# CONFIG_NETFILTER is not set
# CONFIG_IP_DCCP is not set
# CONFIG_IP_SCTP is not set
# CONFIG_TIPC is not set
# CONFIG_ATM is not set
# CONFIG_BRIDGE is not set
# CONFIG_VLAN_8021Q is not set
# CONFIG_DECNET is not set
# CONFIG_LLC2 is not set
# CONFIG_IPX is not set
# CONFIG_ATALK is not set
# CONFIG_X25 is not set
# CONFIG_LAPB is not set
# CONFIG_ECONET is not set
# CONFIG_WAN_ROUTER is not set
# CONFIG_NET_SCHED is not set
#
# Network testing
#
# CONFIG_NET_PKTGEN is not set
# CONFIG_HAMRADIO is not set
# CONFIG_IRDA is not set
# CONFIG_BT is not set
# CONFIG_AF_RXRPC is not set
#
# Wireless
#
# CONFIG_CFG80211 is not set
# CONFIG_WIRELESS_EXT is not set
# CONFIG_MAC80211 is not set
# CONFIG_IEEE80211 is not set
# CONFIG_RFKILL is not set
# CONFIG_NET_9P is not set
#
# Device Drivers
#
#
# Generic Driver Options
#
CONFIG_UEVENT_HELPER_PATH="/sbin/hotplug"
# CONFIG_STANDALONE is not set
# CONFIG_PREVENT_FIRMWARE_BUILD is not set
CONFIG_FW_LOADER=y
# CONFIG_DEBUG_DRIVER is not set
# CONFIG_DEBUG_DEVRES is not set
# CONFIG_SYS_HYPERVISOR is not set
# CONFIG_CONNECTOR is not set
# CONFIG_MTD is not set
# CONFIG_PARPORT is not set
CONFIG_BLK_DEV=y
# CONFIG_BLK_DEV_FD is not set
# CONFIG_BLK_CPQ_DA is not set
# CONFIG_BLK_CPQ_CISS_DA is not set
# CONFIG_BLK_DEV_DAC960 is not set
# CONFIG_BLK_DEV_UMEM is not set
# CONFIG_BLK_DEV_COW_COMMON is not set
CONFIG_BLK_DEV_LOOP=y
# CONFIG_BLK_DEV_CRYPTOLOOP is not set
# CONFIG_BLK_DEV_NBD is not set
# CONFIG_BLK_DEV_SX8 is not set
# CONFIG_BLK_DEV_RAM is not set
# CONFIG_CDROM_PKTCDVD is not set
# CONFIG_ATA_OVER_ETH is not set
CONFIG_MISC_DEVICES=y
# CONFIG_PHANTOM is not set
# CONFIG_EEPROM_93CX6 is not set
# CONFIG_SGI_IOC4 is not set
# CONFIG_TIFM_CORE is not set
CONFIG_IDE=y
CONFIG_IDE_MAX_HWIFS=4
CONFIG_BLK_DEV_IDE=y
#
# Please see Documentation/ide.txt for help/info on IDE drives
#
# CONFIG_BLK_DEV_IDE_SATA is not set
CONFIG_BLK_DEV_IDEDISK=y
CONFIG_IDEDISK_MULTI_MODE=y
CONFIG_BLK_DEV_IDECD=y
# CONFIG_BLK_DEV_IDETAPE is not set
# CONFIG_BLK_DEV_IDEFLOPPY is not set
# CONFIG_IDE_TASK_IOCTL is not set
CONFIG_IDE_PROC_FS=y
#
# IDE chipset support/bugfixes
#
CONFIG_IDE_GENERIC=y
# CONFIG_BLK_DEV_PLATFORM is not set
#
# PCI IDE chipsets support
#
# CONFIG_IDEPCI_PCIBUS_ORDER is not set
# CONFIG_BLK_DEV_GENERIC is not set
# CONFIG_BLK_DEV_OPTI621 is not set
# CONFIG_BLK_DEV_AEC62XX is not set
# CONFIG_BLK_DEV_ALI15X3 is not set
# CONFIG_BLK_DEV_AMD74XX is not set
# CONFIG_BLK_DEV_CMD64X is not set
# CONFIG_BLK_DEV_TRIFLEX is not set
# CONFIG_BLK_DEV_CY82C693 is not set
# CONFIG_BLK_DEV_CS5520 is not set
# CONFIG_BLK_DEV_CS5530 is not set
# CONFIG_BLK_DEV_HPT34X is not set
# CONFIG_BLK_DEV_HPT366 is not set
# CONFIG_BLK_DEV_JMICRON is not set
# CONFIG_BLK_DEV_SC1200 is not set
# CONFIG_BLK_DEV_PIIX is not set
# CONFIG_BLK_DEV_IT8213 is not set
# CONFIG_BLK_DEV_IT821X is not set
# CONFIG_BLK_DEV_NS87415 is not set
# CONFIG_BLK_DEV_PDC202XX_OLD is not set
# CONFIG_BLK_DEV_PDC202XX_NEW is not set
# CONFIG_BLK_DEV_SVWKS is not set
# CONFIG_BLK_DEV_SIIMAGE is not set
# CONFIG_BLK_DEV_SLC90E66 is not set
# CONFIG_BLK_DEV_TRM290 is not set
# CONFIG_BLK_DEV_VIA82CXXX is not set
# CONFIG_BLK_DEV_TC86C001 is not set
# CONFIG_IDE_ARM is not set
# CONFIG_BLK_DEV_IDEDMA is not set
CONFIG_IDE_ARCH_OBSOLETE_INIT=y
# CONFIG_BLK_DEV_HD is not set
#
# SCSI device support
#
# CONFIG_RAID_ATTRS is not set
# CONFIG_SCSI is not set
# CONFIG_SCSI_DMA is not set
# CONFIG_SCSI_NETLINK is not set
# CONFIG_ATA is not set
# CONFIG_MD is not set
# CONFIG_FUSION is not set
#
# IEEE 1394 (FireWire) support
#
# CONFIG_FIREWIRE is not set
# CONFIG_IEEE1394 is not set
# CONFIG_I2O is not set
CONFIG_NETDEVICES=y
# CONFIG_NETDEVICES_MULTIQUEUE is not set
# CONFIG_DUMMY is not set
# CONFIG_BONDING is not set
# CONFIG_MACVLAN is not set
# CONFIG_EQUALIZER is not set
# CONFIG_TUN is not set
# CONFIG_VETH is not set
# CONFIG_ARCNET is not set
# CONFIG_PHYLIB is not set
CONFIG_NET_ETHERNET=y
CONFIG_MII=y
# CONFIG_AX88796 is not set
# CONFIG_HAPPYMEAL is not set
# CONFIG_SUNGEM is not set
# CONFIG_CASSINI is not set
# CONFIG_NET_VENDOR_3COM is not set
# CONFIG_DM9000 is not set
# CONFIG_NET_TULIP is not set
# CONFIG_HP100 is not set
# CONFIG_IBM_NEW_EMAC_ZMII is not set
# CONFIG_IBM_NEW_EMAC_RGMII is not set
# CONFIG_IBM_NEW_EMAC_TAH is not set
# CONFIG_IBM_NEW_EMAC_EMAC4 is not set
CONFIG_NET_PCI=y
CONFIG_PCNET32=y
# CONFIG_PCNET32_NAPI is not set
# CONFIG_AMD8111_ETH is not set
# CONFIG_ADAPTEC_STARFIRE is not set
# CONFIG_B44 is not set
# CONFIG_FORCEDETH is not set
# CONFIG_TC35815 is not set
# CONFIG_EEPRO100 is not set
# CONFIG_E100 is not set
# CONFIG_FEALNX is not set
# CONFIG_NATSEMI is not set
# CONFIG_NE2K_PCI is not set
# CONFIG_8139CP is not set
# CONFIG_8139TOO is not set
# CONFIG_SIS900 is not set
# CONFIG_EPIC100 is not set
# CONFIG_SUNDANCE is not set
# CONFIG_TLAN is not set
# CONFIG_VIA_RHINE is not set
# CONFIG_SC92031 is not set
# CONFIG_NETDEV_1000 is not set
# CONFIG_NETDEV_10000 is not set
# CONFIG_TR is not set
#
# Wireless LAN
#
# CONFIG_WLAN_PRE80211 is not set
# CONFIG_WLAN_80211 is not set
# CONFIG_WAN is not set
# CONFIG_FDDI is not set
# CONFIG_HIPPI is not set
# CONFIG_PPP is not set
# CONFIG_SLIP is not set
# CONFIG_SHAPER is not set
# CONFIG_NETCONSOLE is not set
# CONFIG_NETPOLL is not set
# CONFIG_NET_POLL_CONTROLLER is not set
# CONFIG_ISDN is not set
# CONFIG_PHONE is not set
#
# Input device support
#
CONFIG_INPUT=y
# CONFIG_INPUT_FF_MEMLESS is not set
# CONFIG_INPUT_POLLDEV is not set
#
# Userland interfaces
#
CONFIG_INPUT_MOUSEDEV=y
CONFIG_INPUT_MOUSEDEV_PSAUX=y
CONFIG_INPUT_MOUSEDEV_SCREEN_X=1024
CONFIG_INPUT_MOUSEDEV_SCREEN_Y=768
# CONFIG_INPUT_JOYDEV is not set
CONFIG_INPUT_EVDEV=y
# CONFIG_INPUT_EVBUG is not set
#
# Input Device Drivers
#
CONFIG_INPUT_KEYBOARD=y
CONFIG_KEYBOARD_ATKBD=y
# CONFIG_KEYBOARD_SUNKBD is not set
# CONFIG_KEYBOARD_LKKBD is not set
# CONFIG_KEYBOARD_XTKBD is not set
# CONFIG_KEYBOARD_NEWTON is not set
# CONFIG_KEYBOARD_STOWAWAY is not set
CONFIG_INPUT_MOUSE=y
CONFIG_MOUSE_PS2=y
CONFIG_MOUSE_PS2_ALPS=y
CONFIG_MOUSE_PS2_LOGIPS2PP=y
CONFIG_MOUSE_PS2_SYNAPTICS=y
CONFIG_MOUSE_PS2_LIFEBOOK=y
CONFIG_MOUSE_PS2_TRACKPOINT=y
# CONFIG_MOUSE_PS2_TOUCHKIT is not set
# CONFIG_MOUSE_SERIAL is not set
# CONFIG_MOUSE_VSXXXAA is not set
# CONFIG_INPUT_JOYSTICK is not set
# CONFIG_INPUT_TABLET is not set
# CONFIG_INPUT_TOUCHSCREEN is not set
CONFIG_INPUT_MISC=y
# CONFIG_INPUT_PCSPKR is not set
CONFIG_INPUT_UINPUT=y
#
# Hardware I/O ports
#
CONFIG_SERIO=y
CONFIG_SERIO_I8042=y
CONFIG_SERIO_SERPORT=y
# CONFIG_SERIO_PCIPS2 is not set
CONFIG_SERIO_LIBPS2=y
# CONFIG_SERIO_RAW is not set
# CONFIG_GAMEPORT is not set
#
# Character devices
#
CONFIG_VT=y
CONFIG_VT_CONSOLE=y
CONFIG_HW_CONSOLE=y
# CONFIG_VT_HW_CONSOLE_BINDING is not set
# CONFIG_SERIAL_NONSTANDARD is not set
#
# Serial drivers
#
CONFIG_SERIAL_8250=y
CONFIG_SERIAL_8250_CONSOLE=y
CONFIG_SERIAL_8250_PCI=y
CONFIG_SERIAL_8250_NR_UARTS=4
CONFIG_SERIAL_8250_RUNTIME_UARTS=4
# CONFIG_SERIAL_8250_EXTENDED is not set
#
# Non-8250 serial port support
#
CONFIG_SERIAL_CORE=y
CONFIG_SERIAL_CORE_CONSOLE=y
# CONFIG_SERIAL_JSM is not set
CONFIG_UNIX98_PTYS=y
# CONFIG_LEGACY_PTYS is not set
# CONFIG_IPMI_HANDLER is not set
# CONFIG_HW_RANDOM is not set
CONFIG_RTC=y
# CONFIG_R3964 is not set
# CONFIG_APPLICOM is not set
# CONFIG_RAW_DRIVER is not set
# CONFIG_TCG_TPM is not set
CONFIG_DEVPORT=y
# CONFIG_I2C is not set
#
# SPI support
#
# CONFIG_SPI is not set
# CONFIG_SPI_MASTER is not set
# CONFIG_W1 is not set
# CONFIG_POWER_SUPPLY is not set
# CONFIG_HWMON is not set
# CONFIG_WATCHDOG is not set
#
# Sonics Silicon Backplane
#
CONFIG_SSB_POSSIBLE=y
# CONFIG_SSB is not set
#
# Multifunction device drivers
#
# CONFIG_MFD_SM501 is not set
#
# Multimedia devices
#
# CONFIG_VIDEO_DEV is not set
# CONFIG_DVB_CORE is not set
# CONFIG_DAB is not set
#
# Graphics support
#
# CONFIG_DRM is not set
# CONFIG_VGASTATE is not set
# CONFIG_VIDEO_OUTPUT_CONTROL is not set
# CONFIG_FB is not set
# CONFIG_BACKLIGHT_LCD_SUPPORT is not set
#
# Display device support
#
# CONFIG_DISPLAY_SUPPORT is not set
#
# Console display driver support
#
CONFIG_VGA_CONSOLE=y
# CONFIG_VGACON_SOFT_SCROLLBACK is not set
CONFIG_DUMMY_CONSOLE=y
#
# Sound
#
# CONFIG_SOUND is not set
CONFIG_HID_SUPPORT=y
CONFIG_HID=y
# CONFIG_HID_DEBUG is not set
# CONFIG_HIDRAW is not set
# CONFIG_USB_SUPPORT is not set
# CONFIG_MMC is not set
# CONFIG_NEW_LEDS is not set
# CONFIG_INFINIBAND is not set
CONFIG_RTC_LIB=y
CONFIG_RTC_CLASS=y
CONFIG_RTC_HCTOSYS=y
CONFIG_RTC_HCTOSYS_DEVICE="rtc0"
# CONFIG_RTC_DEBUG is not set
#
# RTC interfaces
#
CONFIG_RTC_INTF_SYSFS=y
CONFIG_RTC_INTF_PROC=y
CONFIG_RTC_INTF_DEV=y
# CONFIG_RTC_INTF_DEV_UIE_EMUL is not set
# CONFIG_RTC_DRV_TEST is not set
#
# SPI RTC drivers
#
#
# Platform RTC drivers
#
# CONFIG_RTC_DRV_CMOS is not set
# CONFIG_RTC_DRV_DS1553 is not set
# CONFIG_RTC_DRV_STK17TA8 is not set
# CONFIG_RTC_DRV_DS1742 is not set
# CONFIG_RTC_DRV_M48T86 is not set
# CONFIG_RTC_DRV_M48T59 is not set
# CONFIG_RTC_DRV_V3020 is not set
#
# on-CPU RTC drivers
#
#
# Userspace I/O
#
# CONFIG_UIO is not set
#
# File systems
#
CONFIG_EXT2_FS=y
CONFIG_EXT2_FS_XATTR=y
CONFIG_EXT2_FS_POSIX_ACL=y
CONFIG_EXT2_FS_SECURITY=y
# CONFIG_EXT2_FS_XIP is not set
CONFIG_EXT3_FS=y
CONFIG_EXT3_FS_XATTR=y
CONFIG_EXT3_FS_POSIX_ACL=y
CONFIG_EXT3_FS_SECURITY=y
# CONFIG_EXT4DEV_FS is not set
CONFIG_JBD=y
CONFIG_FS_MBCACHE=y
# CONFIG_REISERFS_FS is not set
# CONFIG_JFS_FS is not set
CONFIG_FS_POSIX_ACL=y
# CONFIG_XFS_FS is not set
# CONFIG_GFS2_FS is not set
# CONFIG_OCFS2_FS is not set
# CONFIG_MINIX_FS is not set
# CONFIG_ROMFS_FS is not set
CONFIG_INOTIFY=y
CONFIG_INOTIFY_USER=y
CONFIG_QUOTA=y
# CONFIG_QUOTA_NETLINK_INTERFACE is not set
CONFIG_PRINT_QUOTA_WARNING=y
# CONFIG_QFMT_V1 is not set
CONFIG_QFMT_V2=y
CONFIG_QUOTACTL=y
CONFIG_DNOTIFY=y
CONFIG_AUTOFS_FS=y
CONFIG_AUTOFS4_FS=y
# CONFIG_FUSE_FS is not set
#
# CD-ROM/DVD Filesystems
#
CONFIG_ISO9660_FS=y
CONFIG_JOLIET=y
CONFIG_ZISOFS=y
CONFIG_UDF_FS=y
CONFIG_UDF_NLS=y
#
# DOS/FAT/NT Filesystems
#
CONFIG_FAT_FS=y
CONFIG_MSDOS_FS=y
CONFIG_VFAT_FS=y
CONFIG_FAT_DEFAULT_CODEPAGE=437
CONFIG_FAT_DEFAULT_IOCHARSET="iso8859-1"
CONFIG_NTFS_FS=y
# CONFIG_NTFS_DEBUG is not set
CONFIG_NTFS_RW=y
#
# Pseudo filesystems
#
CONFIG_PROC_FS=y
CONFIG_PROC_KCORE=y
CONFIG_PROC_SYSCTL=y
CONFIG_SYSFS=y
CONFIG_TMPFS=y
# CONFIG_TMPFS_POSIX_ACL is not set
# CONFIG_HUGETLB_PAGE is not set
# CONFIG_CONFIGFS_FS is not set
#
# Miscellaneous filesystems
#
# CONFIG_ADFS_FS is not set
# CONFIG_AFFS_FS is not set
# CONFIG_HFS_FS is not set
# CONFIG_HFSPLUS_FS is not set
# CONFIG_BEFS_FS is not set
# CONFIG_BFS_FS is not set
# CONFIG_EFS_FS is not set
CONFIG_CRAMFS=y
# CONFIG_VXFS_FS is not set
# CONFIG_HPFS_FS is not set
# CONFIG_QNX4FS_FS is not set
# CONFIG_SYSV_FS is not set
# CONFIG_UFS_FS is not set
CONFIG_NETWORK_FILESYSTEMS=y
# CONFIG_NFS_FS is not set
# CONFIG_NFSD is not set
# CONFIG_SMB_FS is not set
# CONFIG_CIFS is not set
# CONFIG_NCP_FS is not set
# CONFIG_CODA_FS is not set
# CONFIG_AFS_FS is not set
#
# Partition Types
#
# CONFIG_PARTITION_ADVANCED is not set
CONFIG_MSDOS_PARTITION=y
CONFIG_NLS=y
CONFIG_NLS_DEFAULT="iso8859-1"
CONFIG_NLS_CODEPAGE_437=y
# CONFIG_NLS_CODEPAGE_737 is not set
# CONFIG_NLS_CODEPAGE_775 is not set
CONFIG_NLS_CODEPAGE_850=y
# CONFIG_NLS_CODEPAGE_852 is not set
# CONFIG_NLS_CODEPAGE_855 is not set
# CONFIG_NLS_CODEPAGE_857 is not set
# CONFIG_NLS_CODEPAGE_860 is not set
# CONFIG_NLS_CODEPAGE_861 is not set
# CONFIG_NLS_CODEPAGE_862 is not set
# CONFIG_NLS_CODEPAGE_863 is not set
# CONFIG_NLS_CODEPAGE_864 is not set
# CONFIG_NLS_CODEPAGE_865 is not set
# CONFIG_NLS_CODEPAGE_866 is not set
# CONFIG_NLS_CODEPAGE_869 is not set
# CONFIG_NLS_CODEPAGE_936 is not set
# CONFIG_NLS_CODEPAGE_950 is not set
# CONFIG_NLS_CODEPAGE_932 is not set
# CONFIG_NLS_CODEPAGE_949 is not set
# CONFIG_NLS_CODEPAGE_874 is not set
# CONFIG_NLS_ISO8859_8 is not set
# CONFIG_NLS_CODEPAGE_1250 is not set
# CONFIG_NLS_CODEPAGE_1251 is not set
CONFIG_NLS_ASCII=y
CONFIG_NLS_ISO8859_1=y
# CONFIG_NLS_ISO8859_2 is not set
# CONFIG_NLS_ISO8859_3 is not set
# CONFIG_NLS_ISO8859_4 is not set
# CONFIG_NLS_ISO8859_5 is not set
# CONFIG_NLS_ISO8859_6 is not set
# CONFIG_NLS_ISO8859_7 is not set
# CONFIG_NLS_ISO8859_9 is not set
# CONFIG_NLS_ISO8859_13 is not set
# CONFIG_NLS_ISO8859_14 is not set
# CONFIG_NLS_ISO8859_15 is not set
# CONFIG_NLS_KOI8_R is not set
# CONFIG_NLS_KOI8_U is not set
CONFIG_NLS_UTF8=y
# CONFIG_DLM is not set
# CONFIG_INSTRUMENTATION is not set
#
# Kernel hacking
#
CONFIG_TRACE_IRQFLAGS_SUPPORT=y
# CONFIG_PRINTK_TIME is not set
CONFIG_ENABLE_WARN_DEPRECATED=y
CONFIG_ENABLE_MUST_CHECK=y
CONFIG_MAGIC_SYSRQ=y
# CONFIG_UNUSED_SYMBOLS is not set
# CONFIG_DEBUG_FS is not set
# CONFIG_HEADERS_CHECK is not set
CONFIG_IPIPE_DEBUG=y
# CONFIG_IPIPE_DEBUG_CONTEXT is not set
# CONFIG_IPIPE_TRACE is not set
CONFIG_DEBUG_KERNEL=y
# CONFIG_DEBUG_SHIRQ is not set
CONFIG_DETECT_SOFTLOCKUP=y
CONFIG_SCHED_DEBUG=y
# CONFIG_SCHEDSTATS is not set
# CONFIG_TIMER_STATS is not set
# CONFIG_DEBUG_RT_MUTEXES is not set
# CONFIG_RT_MUTEX_TESTER is not set
CONFIG_DEBUG_SPINLOCK=y
# CONFIG_DEBUG_MUTEXES is not set
# CONFIG_DEBUG_LOCK_ALLOC is not set
# CONFIG_PROVE_LOCKING is not set
# CONFIG_LOCK_STAT is not set
# CONFIG_DEBUG_SPINLOCK_SLEEP is not set
# CONFIG_DEBUG_LOCKING_API_SELFTESTS is not set
# CONFIG_DEBUG_KOBJECT is not set
CONFIG_DEBUG_INFO=y
# CONFIG_DEBUG_VM is not set
# CONFIG_DEBUG_LIST is not set
# CONFIG_DEBUG_SG is not set
CONFIG_FORCED_INLINING=y
# CONFIG_BOOT_PRINTK_DELAY is not set
# CONFIG_RCU_TORTURE_TEST is not set
# CONFIG_FAULT_INJECTION is not set
# CONFIG_SAMPLES is not set
CONFIG_CMDLINE=""
# CONFIG_DEBUG_STACK_USAGE is not set
# CONFIG_RUNTIME_DEBUG is not set
# CONFIG_MIPS_UNCACHED is not set
#
# Security options
#
CONFIG_KEYS=y
# CONFIG_KEYS_DEBUG_PROC_KEYS is not set
CONFIG_SECURITY=y
CONFIG_SECURITY_NETWORK=y
CONFIG_SECURITY_CAPABILITIES=y
# CONFIG_SECURITY_FILE_CAPABILITIES is not set
# CONFIG_CRYPTO is not set
#
# Library routines
#
CONFIG_BITREVERSE=y
# CONFIG_CRC_CCITT is not set
# CONFIG_CRC16 is not set
# CONFIG_CRC_ITU_T is not set
CONFIG_CRC32=y
# CONFIG_CRC7 is not set
# CONFIG_LIBCRC32C is not set
CONFIG_ZLIB_INFLATE=y
CONFIG_PLIST=y
CONFIG_HAS_IOMEM=y
CONFIG_HAS_IOPORT=y
CONFIG_HAS_DMA=y
^ permalink raw reply related [flat|nested] 6+ messages in thread
end of thread, other threads:[~2008-07-23 12:55 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2008-05-05 16:43 [Adeos-main] Advice regarding MIPS port Florent Audebert
2008-05-06 19:54 ` Jan Kiszka
2008-06-19 16:00 ` Florent Audebert
2008-06-20 10:15 ` Gilles Chanteperdrix
2008-07-17 15:12 ` Florent Audebert
2008-07-23 12:55 ` Florent Audebert
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.