[PATCH v3] arm64: Support arch_irq_work_raise() via self IPIs

From: Larry Bassel
Date: Mon May 12 2014 - 11:49:08 EST

Next message: Sasha Levin: "perf: NULL ptr deref in perf_event_mmap"
Previous message: Vlastimil Babka: "Re: [PATCH] mm, compaction: properly signal and act upon lock and need_sched() contention"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Support for arch_irq_work_raise() was missing from
arm64 (a prerequisite for FULL_NOHZ).

This patch is based on the arm32 patches ARM 7872/1
and 7887/1.

commit bf18525fd793101df42a1344ecc48b49b62e48c9
Author: Stephen Boyd <sboyd@xxxxxxxxxxxxxx>
Date: Tue Oct 29 20:32:56 2013 +0100

ARM: 7872/1: Support arch_irq_work_raise() via self IPIs

By default, IRQ work is run from the tick interrupt (see
irq_work_run() in update_process_times()). When we're in full
NOHZ mode, restarting the tick requires the use of IRQ work and
if the only place we run IRQ work is in the tick interrupt we
have an unbreakable cycle. Implement arch_irq_work_raise() via
self IPIs to break this cycle and get the tick started again.
Note that we implement this via IPIs which are only available on
SMP builds. This shouldn't be a problem because full NOHZ is only
supported on SMP builds anyway.

Signed-off-by: Stephen Boyd <sboyd@xxxxxxxxxxxxxx>
Reviewed-by: Kevin Hilman <khilman@xxxxxxxxxx>
Cc: Frederic Weisbecker <fweisbec@xxxxxxxxx>
Signed-off-by: Russell King <rmk+kernel@xxxxxxxxxxxxxxxx>

commit c682e51dbc9837f4aa61180775e9ca4f2a463adf
Author: Stephen Boyd <sboyd@xxxxxxxxxxxxxx>
Date: Sat Nov 9 00:38:24 2013 +0100

ARM: 7887/1: Don't smp_cross_call() on UP devices in arch_irq_work_raise()

If we're running a kernel compiled with SMP_ON_UP=y and the
hardware only supports UP operation there isn't any
smp_cross_call function assigned. Unfortunately, we call
smp_cross_call() unconditionally in arch_irq_work_raise() and
crash the kernel on UP devices. Check to make sure we're running
on an SMP device before calling smp_cross_call() here.

Unable to handle kernel NULL pointer dereference at virtual address 00000000
pgd = c0004000
[00000000] *pgd=00000000
Internal error: Oops: 80000005 [#1] SMP ARM
Modules linked in:
CPU: 0 PID: 1 Comm: swapper/0 Not tainted 3.12.0-rc6-00018-g8d45144-dirty #16
task: de05b440 ti: de05c000 task.ti: de05c000
PC is at 0x0
LR is at arch_irq_work_raise+0x3c/0x48
pc : [<00000000>] lr : [<c0019590>] psr: 60000193
sp : de05dd60 ip : 00000001 fp : 00000000
r10: c085e2f0 r9 : de05c000 r8 : c07be0a4
r7 : de05c000 r6 : de05c000 r5 : c07c5778 r4 : c0824554
r3 : 00000000 r2 : 00000000 r1 : 00000006 r0 : c0529a58
Flags: nZCv IRQs off FIQs on Mode SVC_32 ISA ARM Segment kernel
Control: 10c5387d Table: 80004019 DAC: 00000017
Process swapper/0 (pid: 1, stack limit = 0xde05c248)
Stack: (0xde05dd60 to 0xde05e000)
dd60: c07b9dbc c00cb2dc 00000001 c08242c0 c08242c0 60000113 c07be0a8 c00b0590
dd80: de05c000 c085e2f0 c08242c0 c08242c0 c1414c28 c00b07cc de05b440 c1414c28
dda0: c08242c0 c00b0af8 c0862bb0 c0862db0 c1414cd8 de05c028 c0824840 de05ddb8
ddc0: 00000000 00000009 00000001 00000024 c07be0a8 c07be0a4 de05c000 c085e2f0
dde0: 00000000 c004a4b0 00000010 de00d2dc 00000054 00000100 00000024 00000000
de00: de05c028 0000000a ffff8ae7 00200040 00000016 de05c000 60000193 de05c000
de20: 00000054 00000000 00000000 00000000 00000000 c004a704 00000000 de05c008
de40: c07ba254 c004aa1c c07c5778 c0014b70 fa200000 00000054 de05de80 c0861244
de60: 00000000 c0008634 de05b440 c051c778 20000113 ffffffff de05deb4 c051d0a4
de80: 00000001 00000001 00000000 de05b440 c082afac de057ac0 de057ac0 de0443c0
dea0: 00000000 00000000 00000000 00000000 c082afbc de05dec8 c009f2a0 c051c778
dec0: 20000113 ffffffff 00000000 c016edb0 00000000 000002b0 de057ac0 de057ac0
dee0: 00000000 c016ee40 c0875e50 de05df2e de057ac0 00000000 00000013 00000000
df00: 00000000 c016f054 de043600 de0443c0 c008eb38 de004ec0 c0875e50 c008eb44
df20: 00000012 00000000 00000000 3931f0f8 00000000 00000000 00000014 c0822e84
df40: 00000000 c008ed2c 00000000 00000000 00000000 c07b7490 c07b7490 c075ab3c
df60: 00000000 c00701ac 00000002 00000000 c0070160 dffadb73 7bf8edb4 00000000
df80: c051092c 00000000 00000000 00000000 00000000 00000000 00000000 c0510934
dfa0: de05aa40 00000000 c051092c c0013ce8 00000000 00000000 00000000 00000000
dfc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
dfe0: 00000000 00000000 00000000 00000000 00000013 00000000 07efffe5 4dfac6f5
[<c0019590>] (arch_irq_work_raise+0x3c/0x48) from [<c00cb2dc>] (irq_work_queue+0xe4/0xf8)
[<c00cb2dc>] (irq_work_queue+0xe4/0xf8) from [<c00b0590>] (rcu_accelerate_cbs+0x1d4/0x1d8)
[<c00b0590>] (rcu_accelerate_cbs+0x1d4/0x1d8) from [<c00b07cc>] (rcu_start_gp+0x34/0x48)
[<c00b07cc>] (rcu_start_gp+0x34/0x48) from [<c00b0af8>] (rcu_process_callbacks+0x318/0x608)
[<c00b0af8>] (rcu_process_callbacks+0x318/0x608) from [<c004a4b0>] (__do_softirq+0x114/0x2a0)
[<c004a4b0>] (__do_softirq+0x114/0x2a0) from [<c004a704>] (do_softirq+0x6c/0x74)
[<c004a704>] (do_softirq+0x6c/0x74) from [<c004aa1c>] (irq_exit+0xac/0x100)
[<c004aa1c>] (irq_exit+0xac/0x100) from [<c0014b70>] (handle_IRQ+0x54/0xb4)
[<c0014b70>] (handle_IRQ+0x54/0xb4) from [<c0008634>] (omap3_intc_handle_irq+0x60/0x74)
[<c0008634>] (omap3_intc_handle_irq+0x60/0x74) from [<c051d0a4>] (__irq_svc+0x44/0x5c)
Exception stack(0xde05de80 to 0xde05dec8)
de80: 00000001 00000001 00000000 de05b440 c082afac de057ac0 de057ac0 de0443c0
dea0: 00000000 00000000 00000000 00000000 c082afbc de05dec8 c009f2a0 c051c778
dec0: 20000113 ffffffff
[<c051d0a4>] (__irq_svc+0x44/0x5c) from [<c051c778>] (_raw_spin_unlock_irq+0x28/0x2c)
[<c051c778>] (_raw_spin_unlock_irq+0x28/0x2c) from [<c016edb0>] (proc_alloc_inum+0x30/0xa8)
[<c016edb0>] (proc_alloc_inum+0x30/0xa8) from [<c016ee40>] (proc_register+0x18/0x130)
[<c016ee40>] (proc_register+0x18/0x130) from [<c016f054>] (proc_mkdir_data+0x44/0x6c)
[<c016f054>] (proc_mkdir_data+0x44/0x6c) from [<c008eb44>] (register_irq_proc+0x6c/0x128)
[<c008eb44>] (register_irq_proc+0x6c/0x128) from [<c008ed2c>] (init_irq_proc+0x74/0xb0)
[<c008ed2c>] (init_irq_proc+0x74/0xb0) from [<c075ab3c>] (kernel_init_freeable+0x84/0x1c8)
[<c075ab3c>] (kernel_init_freeable+0x84/0x1c8) from [<c0510934>] (kernel_init+0x8/0x150)
[<c0510934>] (kernel_init+0x8/0x150) from [<c0013ce8>] (ret_from_fork+0x14/0x2c)
Code: bad PC value

Fixes: bf18525fd79 "ARM: 7872/1: Support arch_irq_work_raise() via self IPIs"

Reported-by: Olof Johansson <olof@xxxxxxxxx>
Signed-off-by: Stephen Boyd <sboyd@xxxxxxxxxxxxxx>
Tested-by: Olof Johansson <olof@xxxxxxxxx>
Signed-off-by: Russell King <rmk+kernel@xxxxxxxxxxxxxxxx>

Changes v2 to v3:

* Do not call is_smp() as this is only defined on arm32

Changes v1 to v2:

* Include ARM 7887/1 bugfix

Signed-off-by: Larry Bassel <larry.bassel@xxxxxxxxxx>
Reviewed-by: Kevin Hilman <khilman@xxxxxxxxxx>
---
arch/arm64/include/asm/hardirq.h | 2 +-
arch/arm64/kernel/smp.c | 19 +++++++++++++++++++
2 files changed, 20 insertions(+), 1 deletion(-)

diff --git a/arch/arm64/include/asm/hardirq.h b/arch/arm64/include/asm/hardirq.h
index ae4801d..0be6782 100644
--- a/arch/arm64/include/asm/hardirq.h
+++ b/arch/arm64/include/asm/hardirq.h
@@ -20,7 +20,7 @@
#include <linux/threads.h>
#include <asm/irq.h>

-#define NR_IPI 5
+#define NR_IPI 6

typedef struct {
unsigned int __softirq_pending;
diff --git a/arch/arm64/kernel/smp.c b/arch/arm64/kernel/smp.c
index f0a141d..049aa8d 100644
--- a/arch/arm64/kernel/smp.c
+++ b/arch/arm64/kernel/smp.c
@@ -35,6 +35,7 @@
#include <linux/clockchips.h>
#include <linux/completion.h>
#include <linux/of.h>
+#include <linux/irq_work.h>

#include <asm/atomic.h>
#include <asm/cacheflush.h>
@@ -62,6 +63,7 @@ enum ipi_msg_type {
IPI_CALL_FUNC_SINGLE,
IPI_CPU_STOP,
IPI_TIMER,
+ IPI_IRQ_WORK,
};

/*
@@ -455,6 +457,14 @@ void arch_send_call_function_single_ipi(int cpu)
smp_cross_call(cpumask_of(cpu), IPI_CALL_FUNC_SINGLE);
}

+#ifdef CONFIG_IRQ_WORK
+void arch_irq_work_raise(void)
+{
+ if (smp_cross_call)
+ smp_cross_call(cpumask_of(smp_processor_id()), IPI_IRQ_WORK);
+}
+#endif
+
static const char *ipi_types[NR_IPI] = {
#define S(x,s) [x - IPI_RESCHEDULE] = s
S(IPI_RESCHEDULE, "Rescheduling interrupts"),
@@ -462,6 +472,7 @@ static const char *ipi_types[NR_IPI] = {
S(IPI_CALL_FUNC_SINGLE, "Single function call interrupts"),
S(IPI_CPU_STOP, "CPU stop interrupts"),
S(IPI_TIMER, "Timer broadcast interrupts"),
+ S(IPI_IRQ_WORK, "IRQ work interrupts"),
};

void show_ipi_list(struct seq_file *p, int prec)
@@ -554,6 +565,14 @@ void handle_IPI(int ipinr, struct pt_regs *regs)
break;
#endif

+#ifdef CONFIG_IRQ_WORK
+ case IPI_IRQ_WORK:
+ irq_enter();
+ irq_work_run();
+ irq_exit();
+ break;
+#endif
+
default:
pr_crit("CPU%u: Unknown IPI message 0x%x\n", cpu, ipinr);
break;
--
1.8.3.2

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Next message: Sasha Levin: "perf: NULL ptr deref in perf_event_mmap"
Previous message: Vlastimil Babka: "Re: [PATCH] mm, compaction: properly signal and act upon lock and need_sched() contention"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]