Patchwork 3.7-rc7: BUG: MAX_STACK_TRACE_ENTRIES too low!

login
register
mail settings
Submitter Li Zhong
Date Nov. 28, 2012, 8:41 a.m.
Message ID <1354092082.3054.19.camel@ThinkPad-T5421.cn.ibm.com>
Download mbox | patch
Permalink /patch/202385/
State Accepted
Commit e253ebab935693cc2eafbc93e94b68b47dc779db
Headers show

Comments

Li Zhong - Nov. 28, 2012, 8:41 a.m.
On Tue, 2012-11-27 at 19:22 -0800, Christian Kujau wrote:
> On Tue, 27 Nov 2012 at 19:06, Christian Kujau wrote:
> > the same thing[0] happened again in 3.7-rc7, after ~20h uptime:
> 
> I found the following on patchwork, but this seems to deal with powerpc64 
> only, while this PowerBook G4 of mine is powerpc32:
> 
>   http://patchwork.ozlabs.org/patch/193414/
> 
> It looks related, but then again, I fail to parse assember...

Hi Christian, 

Would you please help to try the following fix? I don't have a powerpc32
machine for test...

Thanks, Zhong

==============================================

> 
> Christian.
> 
> > [40007.339487] [sched_delayed] sched: RT throttling activated
> > [69731.388717] BUG: MAX_STACK_TRACE_ENTRIES too low!
> > [69731.390371] turning off the locking correctness validator.
> > [69731.391942] Call Trace:
> > [69731.393525] [c9a61c10] [c0009064] show_stack+0x70/0x1bc (unreliable)
> > [69731.395152] [c9a61c50] [c0077460] save_trace+0xfc/0x114
> > [69731.396735] [c9a61c60] [c007be20] __lock_acquire+0x1568/0x19b8
> > [69731.398296] [c9a61d00] [c007c2c0] lock_acquire+0x50/0x70
> > [69731.399857] [c9a61d20] [c0550e28] _raw_spin_lock_irq+0x5c/0x78
> > [69731.401419] [c9a61d40] [c054fb58] __schedule+0xd8/0x534
> > [69731.402972] [c9a61da0] [c0550094] _cond_resched+0x50/0x68
> > [69731.404527] [c9a61db0] [c0479908] dst_gc_task+0xbc/0x258
> > [69731.406070] [c9a61e40] [c004eeb8] process_one_work+0x1f4/0x49c
> > [69731.407585] [c9a61e80] [c004f644] worker_thread+0x14c/0x400
> > [69731.409075] [c9a61eb0] [c0057634] kthread+0xbc/0xc0
> > [69731.410521] [c9a61f40] [c0011ad4] ret_from_kernel_thread+0x5c/0x64
> > [...repeated 54 times...]
> > 
> > Anyone knows what this is about?
> > 
> > Thanks,
> > Christian.
> > 
> > [0] http://lkml.indiana.edu/hypermail/linux/kernel/1211.0/03025.html
Christian Kujau - Nov. 28, 2012, 6:59 p.m.
On Wed, 28 Nov 2012 at 16:41, Li Zhong wrote:
> Would you please help to try the following fix? I don't have a powerpc32
> machine for test...

I've just applied this to 3.7-rc7 and booted the machine. I don't know how 
to trigger this bug, so it might take a while until it happens again - or 
not, now with your patch applied.

It happened only 2 times so far, after ~8h and after ~20h:

Nov  5 13:28:20 alice kernel: [    0.000000] Linux version 3.7.0-rc4 
Nov  5 21:00:26 alice kernel: [27148.965634] BUG: MAX_STACK_TRACE_ENTRIES too low!
Nov 26 21:53:43 alice kernel: [    0.000000] Linux version 3.7.0-rc7 
Nov 27 17:15:29 alice kernel: [69731.388717] BUG: MAX_STACK_TRACE_ENTRIES too low!

Thanks,
Christian.
Christian Kujau - Dec. 1, 2012, 8:32 p.m.
On Wed, 28 Nov 2012 at 16:41, Li Zhong wrote:
> On Tue, 2012-11-27 at 19:22 -0800, Christian Kujau wrote:
> > On Tue, 27 Nov 2012 at 19:06, Christian Kujau wrote:
> > > the same thing[0] happened again in 3.7-rc7, after ~20h uptime:
> > 
> > I found the following on patchwork, but this seems to deal with powerpc64 
> > only, while this PowerBook G4 of mine is powerpc32:
> > 
> >   http://patchwork.ozlabs.org/patch/193414/
> > 
> > It looks related, but then again, I fail to parse assember...
> 
> Hi Christian, 
> 
> Would you please help to try the following fix? I don't have a powerpc32
> machine for test...

After 3 days of uptime with your patch applied (and "normal" usage, just 
as before) no such BUG message occured. So, from my point of view, feel
free to add:

   Tested-by: Christian Kujau <lists@nerdbynature.de>

Thanks!
Christian.

> ==============================================
> diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
> index 9499385..cadebfd 100644
> --- a/arch/powerpc/kernel/entry_32.S
> +++ b/arch/powerpc/kernel/entry_32.S
> @@ -439,6 +439,8 @@ ret_from_fork:
>  ret_from_kernel_thread:
>  	REST_NVGPRS(r1)
>  	bl	schedule_tail
> +	li	r3,0
> +	stw	r3,0(r1)
>  	mtlr	r14
>  	mr	r3,r15
>  	PPC440EP_ERR42
> ==============================================
> 
> > 
> > Christian.
> > 
> > > [40007.339487] [sched_delayed] sched: RT throttling activated
> > > [69731.388717] BUG: MAX_STACK_TRACE_ENTRIES too low!
> > > [69731.390371] turning off the locking correctness validator.
> > > [69731.391942] Call Trace:
> > > [69731.393525] [c9a61c10] [c0009064] show_stack+0x70/0x1bc (unreliable)
> > > [69731.395152] [c9a61c50] [c0077460] save_trace+0xfc/0x114
> > > [69731.396735] [c9a61c60] [c007be20] __lock_acquire+0x1568/0x19b8
> > > [69731.398296] [c9a61d00] [c007c2c0] lock_acquire+0x50/0x70
> > > [69731.399857] [c9a61d20] [c0550e28] _raw_spin_lock_irq+0x5c/0x78
> > > [69731.401419] [c9a61d40] [c054fb58] __schedule+0xd8/0x534
> > > [69731.402972] [c9a61da0] [c0550094] _cond_resched+0x50/0x68
> > > [69731.404527] [c9a61db0] [c0479908] dst_gc_task+0xbc/0x258
> > > [69731.406070] [c9a61e40] [c004eeb8] process_one_work+0x1f4/0x49c
> > > [69731.407585] [c9a61e80] [c004f644] worker_thread+0x14c/0x400
> > > [69731.409075] [c9a61eb0] [c0057634] kthread+0xbc/0xc0
> > > [69731.410521] [c9a61f40] [c0011ad4] ret_from_kernel_thread+0x5c/0x64
> > > [...repeated 54 times...]
> > > 
> > > Anyone knows what this is about?
> > > 
> > > Thanks,
> > > Christian.
> > > 
> > > [0] http://lkml.indiana.edu/hypermail/linux/kernel/1211.0/03025.html
Li Zhong - Dec. 3, 2012, 2:09 a.m.
On Sat, 2012-12-01 at 12:32 -0800, Christian Kujau wrote:
> On Wed, 28 Nov 2012 at 16:41, Li Zhong wrote:
> > On Tue, 2012-11-27 at 19:22 -0800, Christian Kujau wrote:
> > > On Tue, 27 Nov 2012 at 19:06, Christian Kujau wrote:
> > > > the same thing[0] happened again in 3.7-rc7, after ~20h uptime:
> > > 
> > > I found the following on patchwork, but this seems to deal with powerpc64 
> > > only, while this PowerBook G4 of mine is powerpc32:
> > > 
> > >   http://patchwork.ozlabs.org/patch/193414/
> > > 
> > > It looks related, but then again, I fail to parse assember...
> > 
> > Hi Christian, 
> > 
> > Would you please help to try the following fix? I don't have a powerpc32
> > machine for test...
> 
> After 3 days of uptime with your patch applied (and "normal" usage, just 
> as before) no such BUG message occured. So, from my point of view, feel
> free to add:
> 
>    Tested-by: Christian Kujau <lists@nerdbynature.de>

Thank you, Christian. 

I'll try to submit a patch for it.

Thanks, Zhong

> 
> Thanks!
> Christian.
> 
> > ==============================================
> > diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
> > index 9499385..cadebfd 100644
> > --- a/arch/powerpc/kernel/entry_32.S
> > +++ b/arch/powerpc/kernel/entry_32.S
> > @@ -439,6 +439,8 @@ ret_from_fork:
> >  ret_from_kernel_thread:
> >  	REST_NVGPRS(r1)
> >  	bl	schedule_tail
> > +	li	r3,0
> > +	stw	r3,0(r1)
> >  	mtlr	r14
> >  	mr	r3,r15
> >  	PPC440EP_ERR42
> > ==============================================
> > 
> > > 
> > > Christian.
> > > 
> > > > [40007.339487] [sched_delayed] sched: RT throttling activated
> > > > [69731.388717] BUG: MAX_STACK_TRACE_ENTRIES too low!
> > > > [69731.390371] turning off the locking correctness validator.
> > > > [69731.391942] Call Trace:
> > > > [69731.393525] [c9a61c10] [c0009064] show_stack+0x70/0x1bc (unreliable)
> > > > [69731.395152] [c9a61c50] [c0077460] save_trace+0xfc/0x114
> > > > [69731.396735] [c9a61c60] [c007be20] __lock_acquire+0x1568/0x19b8
> > > > [69731.398296] [c9a61d00] [c007c2c0] lock_acquire+0x50/0x70
> > > > [69731.399857] [c9a61d20] [c0550e28] _raw_spin_lock_irq+0x5c/0x78
> > > > [69731.401419] [c9a61d40] [c054fb58] __schedule+0xd8/0x534
> > > > [69731.402972] [c9a61da0] [c0550094] _cond_resched+0x50/0x68
> > > > [69731.404527] [c9a61db0] [c0479908] dst_gc_task+0xbc/0x258
> > > > [69731.406070] [c9a61e40] [c004eeb8] process_one_work+0x1f4/0x49c
> > > > [69731.407585] [c9a61e80] [c004f644] worker_thread+0x14c/0x400
> > > > [69731.409075] [c9a61eb0] [c0057634] kthread+0xbc/0xc0
> > > > [69731.410521] [c9a61f40] [c0011ad4] ret_from_kernel_thread+0x5c/0x64
> > > > [...repeated 54 times...]
> > > > 
> > > > Anyone knows what this is about?
> > > > 
> > > > Thanks,
> > > > Christian.
> > > > 
> > > > [0] http://lkml.indiana.edu/hypermail/linux/kernel/1211.0/03025.html
>
Christian Kujau - Jan. 12, 2013, 8:43 p.m.
On Wed, 28 Nov 2012 at 16:41, Li Zhong wrote:
> On Tue, 2012-11-27 at 19:22 -0800, Christian Kujau wrote:
> > On Tue, 27 Nov 2012 at 19:06, Christian Kujau wrote:
> > > the same thing[0] happened again in 3.7-rc7, after ~20h uptime:
> > 
> > I found the following on patchwork, but this seems to deal with powerpc64 
> > only, while this PowerBook G4 of mine is powerpc32:
> > 
> >   http://patchwork.ozlabs.org/patch/193414/
> > 
> > It looks related, but then again, I fail to parse assember...
> 
> Hi Christian, 
> 
> Would you please help to try the following fix? I don't have a powerpc32
> machine for test...

The patch hasn't made it into mainline yet. Any chance that this will be 
included in 3.8?

Thanks,
Christian.

> ==============================================
> diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
> index 9499385..cadebfd 100644
> --- a/arch/powerpc/kernel/entry_32.S
> +++ b/arch/powerpc/kernel/entry_32.S
> @@ -439,6 +439,8 @@ ret_from_fork:
>  ret_from_kernel_thread:
>  	REST_NVGPRS(r1)
>  	bl	schedule_tail
> +	li	r3,0
> +	stw	r3,0(r1)
>  	mtlr	r14
>  	mr	r3,r15
>  	PPC440EP_ERR42
> ==============================================
> 
> > 
> > Christian.
> > 
> > > [40007.339487] [sched_delayed] sched: RT throttling activated
> > > [69731.388717] BUG: MAX_STACK_TRACE_ENTRIES too low!
> > > [69731.390371] turning off the locking correctness validator.
> > > [69731.391942] Call Trace:
> > > [69731.393525] [c9a61c10] [c0009064] show_stack+0x70/0x1bc (unreliable)
> > > [69731.395152] [c9a61c50] [c0077460] save_trace+0xfc/0x114
> > > [69731.396735] [c9a61c60] [c007be20] __lock_acquire+0x1568/0x19b8
> > > [69731.398296] [c9a61d00] [c007c2c0] lock_acquire+0x50/0x70
> > > [69731.399857] [c9a61d20] [c0550e28] _raw_spin_lock_irq+0x5c/0x78
> > > [69731.401419] [c9a61d40] [c054fb58] __schedule+0xd8/0x534
> > > [69731.402972] [c9a61da0] [c0550094] _cond_resched+0x50/0x68
> > > [69731.404527] [c9a61db0] [c0479908] dst_gc_task+0xbc/0x258
> > > [69731.406070] [c9a61e40] [c004eeb8] process_one_work+0x1f4/0x49c
> > > [69731.407585] [c9a61e80] [c004f644] worker_thread+0x14c/0x400
> > > [69731.409075] [c9a61eb0] [c0057634] kthread+0xbc/0xc0
> > > [69731.410521] [c9a61f40] [c0011ad4] ret_from_kernel_thread+0x5c/0x64
> > > [...repeated 54 times...]
> > > 
> > > Anyone knows what this is about?
> > > 
> > > Thanks,
> > > Christian.
> > > 
> > > [0] http://lkml.indiana.edu/hypermail/linux/kernel/1211.0/03025.html
> 
> 
>
Li Zhong - Jan. 15, 2013, 6:59 a.m.
On Sat, 2013-01-12 at 12:43 -0800, Christian Kujau wrote:
> On Wed, 28 Nov 2012 at 16:41, Li Zhong wrote:
> > On Tue, 2012-11-27 at 19:22 -0800, Christian Kujau wrote:
> > > On Tue, 27 Nov 2012 at 19:06, Christian Kujau wrote:
> > > > the same thing[0] happened again in 3.7-rc7, after ~20h uptime:
> > > 
> > > I found the following on patchwork, but this seems to deal with powerpc64 
> > > only, while this PowerBook G4 of mine is powerpc32:
> > > 
> > >   http://patchwork.ozlabs.org/patch/193414/
> > > 
> > > It looks related, but then again, I fail to parse assember...
> > 
> > Hi Christian, 
> > 
> > Would you please help to try the following fix? I don't have a powerpc32
> > machine for test...
> 
> The patch hasn't made it into mainline yet. Any chance that this will be 
> included in 3.8?

I don't know...

FYI, it is already in the next of ppc tree 
http://git.kernel.org/?p=linux/kernel/git/benh/powerpc.git;a=shortlog;h=refs/heads/next

I guess it would get into 3.9, at least. 

Thanks, Zhong

> 
> Thanks,
> Christian.
> 
> > ==============================================
> > diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
> > index 9499385..cadebfd 100644
> > --- a/arch/powerpc/kernel/entry_32.S
> > +++ b/arch/powerpc/kernel/entry_32.S
> > @@ -439,6 +439,8 @@ ret_from_fork:
> >  ret_from_kernel_thread:
> >  	REST_NVGPRS(r1)
> >  	bl	schedule_tail
> > +	li	r3,0
> > +	stw	r3,0(r1)
> >  	mtlr	r14
> >  	mr	r3,r15
> >  	PPC440EP_ERR42
> > ==============================================
> > 
> > > 
> > > Christian.
> > > 
> > > > [40007.339487] [sched_delayed] sched: RT throttling activated
> > > > [69731.388717] BUG: MAX_STACK_TRACE_ENTRIES too low!
> > > > [69731.390371] turning off the locking correctness validator.
> > > > [69731.391942] Call Trace:
> > > > [69731.393525] [c9a61c10] [c0009064] show_stack+0x70/0x1bc (unreliable)
> > > > [69731.395152] [c9a61c50] [c0077460] save_trace+0xfc/0x114
> > > > [69731.396735] [c9a61c60] [c007be20] __lock_acquire+0x1568/0x19b8
> > > > [69731.398296] [c9a61d00] [c007c2c0] lock_acquire+0x50/0x70
> > > > [69731.399857] [c9a61d20] [c0550e28] _raw_spin_lock_irq+0x5c/0x78
> > > > [69731.401419] [c9a61d40] [c054fb58] __schedule+0xd8/0x534
> > > > [69731.402972] [c9a61da0] [c0550094] _cond_resched+0x50/0x68
> > > > [69731.404527] [c9a61db0] [c0479908] dst_gc_task+0xbc/0x258
> > > > [69731.406070] [c9a61e40] [c004eeb8] process_one_work+0x1f4/0x49c
> > > > [69731.407585] [c9a61e80] [c004f644] worker_thread+0x14c/0x400
> > > > [69731.409075] [c9a61eb0] [c0057634] kthread+0xbc/0xc0
> > > > [69731.410521] [c9a61f40] [c0011ad4] ret_from_kernel_thread+0x5c/0x64
> > > > [...repeated 54 times...]
> > > > 
> > > > Anyone knows what this is about?
> > > > 
> > > > Thanks,
> > > > Christian.
> > > > 
> > > > [0] http://lkml.indiana.edu/hypermail/linux/kernel/1211.0/03025.html
> > 
> > 
> > 
>
Christian Kujau - Jan. 27, 2013, 10:56 p.m.
On Tue, 15 Jan 2013 at 14:59, Li Zhong wrote:
> FYI, it is already in the next of ppc tree 
> http://git.kernel.org/?p=linux/kernel/git/benh/powerpc.git;a=shortlog;h=refs/heads/next
> 
> I guess it would get into 3.9, at least. 

Hm, is there no chance to get this into 3.8? I've been running with this 
patch applied since 3.7-rc7 and it got rid of this 
"MAX_STACK_TRACE_ENTRIES too low" message. I've just upgraded to 3.8-rc5 
and it's still not in mainline :-\

Christian.

[0] http://lkml.indiana.edu/hypermail/linux/kernel/1211.3/01836.html

> > > ==============================================
> > > diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
> > > index 9499385..cadebfd 100644
> > > --- a/arch/powerpc/kernel/entry_32.S
> > > +++ b/arch/powerpc/kernel/entry_32.S
> > > @@ -439,6 +439,8 @@ ret_from_fork:
> > >  ret_from_kernel_thread:
> > >  	REST_NVGPRS(r1)
> > >  	bl	schedule_tail
> > > +	li	r3,0
> > > +	stw	r3,0(r1)
> > >  	mtlr	r14
> > >  	mr	r3,r15
> > >  	PPC440EP_ERR42
> > > ==============================================
> > > 
> > > > 
> > > > Christian.
> > > > 
> > > > > [40007.339487] [sched_delayed] sched: RT throttling activated
> > > > > [69731.388717] BUG: MAX_STACK_TRACE_ENTRIES too low!
> > > > > [69731.390371] turning off the locking correctness validator.
> > > > > [69731.391942] Call Trace:
> > > > > [69731.393525] [c9a61c10] [c0009064] show_stack+0x70/0x1bc (unreliable)
> > > > > [69731.395152] [c9a61c50] [c0077460] save_trace+0xfc/0x114
> > > > > [69731.396735] [c9a61c60] [c007be20] __lock_acquire+0x1568/0x19b8
> > > > > [69731.398296] [c9a61d00] [c007c2c0] lock_acquire+0x50/0x70
> > > > > [69731.399857] [c9a61d20] [c0550e28] _raw_spin_lock_irq+0x5c/0x78
> > > > > [69731.401419] [c9a61d40] [c054fb58] __schedule+0xd8/0x534
> > > > > [69731.402972] [c9a61da0] [c0550094] _cond_resched+0x50/0x68
> > > > > [69731.404527] [c9a61db0] [c0479908] dst_gc_task+0xbc/0x258
> > > > > [69731.406070] [c9a61e40] [c004eeb8] process_one_work+0x1f4/0x49c
> > > > > [69731.407585] [c9a61e80] [c004f644] worker_thread+0x14c/0x400
> > > > > [69731.409075] [c9a61eb0] [c0057634] kthread+0xbc/0xc0
> > > > > [69731.410521] [c9a61f40] [c0011ad4] ret_from_kernel_thread+0x5c/0x64
> > > > > [...repeated 54 times...]
> > > > > 
> > > > > Anyone knows what this is about?
> > > > > 
> > > > > Thanks,
> > > > > Christian.
> > > > > 
> > > > > [0] http://lkml.indiana.edu/hypermail/linux/kernel/1211.0/03025.html
Christian Kujau - Feb. 4, 2013, 12:38 a.m.
On Sun, 27 Jan 2013 at 14:56, Christian Kujau wrote:
> Hm, is there no chance to get this into 3.8? I've been running with this 
> patch applied since 3.7-rc7 and it got rid of this 
> "MAX_STACK_TRACE_ENTRIES too low" message. I've just upgraded to 3.8-rc5 
> and it's still not in mainline :-\

Hah! I just noticed that it got merged the next day - Thanks!

Christian.

Patch

==============================================
diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
index 9499385..cadebfd 100644
--- a/arch/powerpc/kernel/entry_32.S
+++ b/arch/powerpc/kernel/entry_32.S
@@ -439,6 +439,8 @@  ret_from_fork:
 ret_from_kernel_thread:
 	REST_NVGPRS(r1)
 	bl	schedule_tail
+	li	r3,0
+	stw	r3,0(r1)
 	mtlr	r14
 	mr	r3,r15
 	PPC440EP_ERR42