lkml.org 
[lkml]   [2008]   [Nov]   [16]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
SubjectRe: [BUG] using smp_processor_id() in preemptible

* Luis Henriques <henrix@sapo.pt> wrote:

> Hi Ingo,
>
> while running latencytop, got the following:
>
> [ 775.663239] BUG: using smp_processor_id() in preemptible [00000000] code: latencytop/6585
> [ 775.663303] caller is native_sched_clock+0x3a/0x80
> [ 775.663314] Pid: 6585, comm: latencytop Tainted: G W 2.6.28-rc4-00355-g9c7c354 #1
> [ 775.663322] Call Trace:
> [ 775.663343] [<ffffffff803a94e4>] debug_smp_processor_id+0xe4/0xf0
> [ 775.663356] [<ffffffff80213f7a>] native_sched_clock+0x3a/0x80
> [ 775.663368] [<ffffffff80213e19>] sched_clock+0x9/0x10
> [ 775.663381] [<ffffffff8024550d>] proc_sched_show_task+0x8bd/0x10e0
> [ 775.663395] [<ffffffff8034466e>] sched_show+0x3e/0x80
> [ 775.663408] [<ffffffff8031039b>] seq_read+0xdb/0x350
> [ 775.663421] [<ffffffff80368776>] ? security_file_permission+0x16/0x20
> [ 775.663435] [<ffffffff802f4198>] vfs_read+0xc8/0x170
> [ 775.663447] [<ffffffff802f4335>] sys_read+0x55/0x90
> [ 775.663460] [<ffffffff8020c67a>] system_call_fastpath+0x16/0x1b
> ...
>
> I am running 2.6.28-rc4-00355-g9c7c354 and, after taking a look at
> the code, it looks like the problem is in commit
> 7cbaef9c83e58bbd4bdd534b09052b6c5ec457d5, where native_sched_clock
> is modified to use __cycles_2_ns instead of cycles_2_ns.

indeed. Does the fix below work for you?

Thanks,

Ingo

-------------->
From 29d7b90c15035741d15421b36000509212b3e135 Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@elte.hu>
Date: Sun, 16 Nov 2008 08:07:15 +0100
Subject: [PATCH] sched: fix kernel warning on /proc/sched_debug access

Luis Henriques reported that with CONFIG_PREEMPT=y + CONFIG_PREEMPT_DEBUG=y +
CONFIG_SCHED_DEBUG=y + CONFIG_LATENCYTOP=y enabled, the following warning
triggers when using latencytop:

> [ 775.663239] BUG: using smp_processor_id() in preemptible [00000000] code: latencytop/6585
> [ 775.663303] caller is native_sched_clock+0x3a/0x80
> [ 775.663314] Pid: 6585, comm: latencytop Tainted: G W 2.6.28-rc4-00355-g9c7c354 #1
> [ 775.663322] Call Trace:
> [ 775.663343] [<ffffffff803a94e4>] debug_smp_processor_id+0xe4/0xf0
> [ 775.663356] [<ffffffff80213f7a>] native_sched_clock+0x3a/0x80
> [ 775.663368] [<ffffffff80213e19>] sched_clock+0x9/0x10
> [ 775.663381] [<ffffffff8024550d>] proc_sched_show_task+0x8bd/0x10e0
> [ 775.663395] [<ffffffff8034466e>] sched_show+0x3e/0x80
> [ 775.663408] [<ffffffff8031039b>] seq_read+0xdb/0x350
> [ 775.663421] [<ffffffff80368776>] ? security_file_permission+0x16/0x20
> [ 775.663435] [<ffffffff802f4198>] vfs_read+0xc8/0x170
> [ 775.663447] [<ffffffff802f4335>] sys_read+0x55/0x90
> [ 775.663460] [<ffffffff8020c67a>] system_call_fastpath+0x16/0x1b
> ...

This breakage was caused by me via:

7cbaef9: sched: optimize sched_clock() a bit

Change the calls to cpu_clock().

Reported-by: Luis Henriques <henrix@sapo.pt>
---
kernel/sched_debug.c | 5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/kernel/sched_debug.c b/kernel/sched_debug.c
index 48ecc51..26ed8e3 100644
--- a/kernel/sched_debug.c
+++ b/kernel/sched_debug.c
@@ -423,10 +423,11 @@ void proc_sched_show_task(struct task_struct *p, struct seq_file *m)
#undef __P

{
+ unsigned int this_cpu = raw_smp_processor_id();
u64 t0, t1;

- t0 = sched_clock();
- t1 = sched_clock();
+ t0 = cpu_clock(this_cpu);
+ t1 = cpu_clock(this_cpu);
SEQ_printf(m, "%-35s:%21Ld\n",
"clock-delta", (long long)(t1-t0));
}


\
 
 \ /
  Last update: 2008-11-16 08:17    [W:0.077 / U:0.288 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site