1. Patchew
  2. linux
  3. View series

[PATCH] context_tracking: Have ct_state() use preempt_disable_notrace()

Steven Rostedt posted 1 patch 2 years, 10 months ago 
include/linux/context_tracking_state.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
[PATCH] context_tracking: Have ct_state() use preempt_disable_notrace()
Posted by Steven Rostedt 2 years, 10 months ago 
From: "Steven Rostedt (Google)" <rostedt@goodmis.org>

One of my tests triggered:

 ------------[ cut here ]------------
 WARNING: CPU: 0 PID: 1 at include/trace/events/preemptirq.h:51 trace_preempt_off+0x7c/0x80
 Modules linked in:
 CPU: 0 PID: 1 Comm: init Not tainted 6.3.0-rc3-test-00011-ge11b521a7b69-dirty #31
 Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.0-debian-1.16.0-5 04/01/2014
 RIP: 0010:trace_preempt_off+0x7c/0x80
 Code: 74 0f 48 8b 78 08 48 89 f2 48 89 de e8 5d f9 ff ff 65 ff 0d 6e d6 ce 68 75 a2 0f 1f 44 00 00 eb 9b e8 28 ea f3 ff 84 c0 75 a5 <0f> 0b eb a1 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 f3 0f
 RSP: 0000:ffffaf0880013f00 EFLAGS: 00010046
 RAX: 0000000000000000 RBX: ffffffff97fafe59 RCX: 0000000000000001
 RDX: 0000000000000000 RSI: ffffffff986f4b71 RDI: ffffffff986fa0dd
 RBP: ffffffff97fafe59 R08: 0000000000000000 R09: 0000000000000000
 R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000
 R13: 0000000000000014 R14: 0000000000000000 R15: 0000000000000000
 FS:  0000000000000000(0000) GS:ffff8adf7bc00000(0000) knlGS:0000000000000000
 CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
 CR2: 00007f125de099c0 CR3: 000000010498a001 CR4: 0000000000170ef0
 Call Trace:
  <TASK>
  irqentry_enter_from_user_mode+0x39/0x80
  irqentry_enter+0x51/0x60
  exc_page_fault+0x3d/0x200
  asm_exc_page_fault+0x26/0x30
 RIP: 0033:0x7f125de099c0
 Code: Unable to access opcode bytes at 0x7f125de09996.
 RSP: 002b:00007ffd1fdbf220 EFLAGS: 00010202
 RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000000
 RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000
 RBP: 0000000000000000 R08: 0000000000000000 R09: 0000000000000000
 R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000
 R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000
  </TASK>
 irq event stamp: 13182054
 hardirqs last  enabled at (13182053): [<ffffffff97002e3f>] ret_from_fork+0x1f/0x50
 hardirqs last disabled at (13182054): [<ffffffff97faff01>] irqentry_enter+0x51/0x60
 softirqs last  enabled at (13181978): [<ffffffff97fc97d2>] __do_softirq+0x322/0x429
 softirqs last disabled at (13181969): [<ffffffff971c2cce>] irq_exit_rcu+0xe/0x30
 ---[ end trace 0000000000000000 ]---
 ------------[ cut here ]------------

It appears that ct_state() can be called in noinst mode. Do not trace the
disabling of preemption.

Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
---
 include/linux/context_tracking_state.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/include/linux/context_tracking_state.h b/include/linux/context_tracking_state.h
index 4a4d56f77180..38e269174a7f 100644
--- a/include/linux/context_tracking_state.h
+++ b/include/linux/context_tracking_state.h
@@ -130,9 +130,9 @@ static __always_inline int ct_state(void)
 	if (!context_tracking_enabled())
 		return CONTEXT_DISABLED;
 
-	preempt_disable();
+	preempt_disable_notrace();
 	ret = __ct_state();
-	preempt_enable();
+	preempt_enable_notrace();
 
 	return ret;
 }
-- 
2.39.1
Re: [PATCH] context_tracking: Have ct_state() use preempt_disable_notrace()
Posted by Josh Poimboeuf 2 years, 10 months ago 
On Tue, Mar 21, 2023 at 08:45:24PM -0400, Steven Rostedt wrote:
> From: "Steven Rostedt (Google)" <rostedt@goodmis.org>
> 
> One of my tests triggered:
> 
>  ------------[ cut here ]------------
>  WARNING: CPU: 0 PID: 1 at include/trace/events/preemptirq.h:51 trace_preempt_off+0x7c/0x80
>  Modules linked in:
>  CPU: 0 PID: 1 Comm: init Not tainted 6.3.0-rc3-test-00011-ge11b521a7b69-dirty #31
>  Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.0-debian-1.16.0-5 04/01/2014
>  RIP: 0010:trace_preempt_off+0x7c/0x80
>  Code: 74 0f 48 8b 78 08 48 89 f2 48 89 de e8 5d f9 ff ff 65 ff 0d 6e d6 ce 68 75 a2 0f 1f 44 00 00 eb 9b e8 28 ea f3 ff 84 c0 75 a5 <0f> 0b eb a1 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 f3 0f
>  RSP: 0000:ffffaf0880013f00 EFLAGS: 00010046
>  RAX: 0000000000000000 RBX: ffffffff97fafe59 RCX: 0000000000000001
>  RDX: 0000000000000000 RSI: ffffffff986f4b71 RDI: ffffffff986fa0dd
>  RBP: ffffffff97fafe59 R08: 0000000000000000 R09: 0000000000000000
>  R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000
>  R13: 0000000000000014 R14: 0000000000000000 R15: 0000000000000000
>  FS:  0000000000000000(0000) GS:ffff8adf7bc00000(0000) knlGS:0000000000000000
>  CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>  CR2: 00007f125de099c0 CR3: 000000010498a001 CR4: 0000000000170ef0
>  Call Trace:
>   <TASK>
>   irqentry_enter_from_user_mode+0x39/0x80
>   irqentry_enter+0x51/0x60
>   exc_page_fault+0x3d/0x200
>   asm_exc_page_fault+0x26/0x30

I think this was just fixed in -tip today:

  https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git/commit/?id=f87d28673b71b35b248231a2086f9404afbb7f28

-- 
Josh
Re: [PATCH] context_tracking: Have ct_state() use preempt_disable_notrace()
Posted by Steven Rostedt 2 years, 10 months ago 
On Tue, 21 Mar 2023 17:58:21 -0700
Josh Poimboeuf <jpoimboe@kernel.org> wrote:

> On Tue, Mar 21, 2023 at 08:45:24PM -0400, Steven Rostedt wrote:
> > From: "Steven Rostedt (Google)" <rostedt@goodmis.org>
> > 
> > One of my tests triggered:
> > 
> >  ------------[ cut here ]------------
> >  WARNING: CPU: 0 PID: 1 at include/trace/events/preemptirq.h:51 trace_preempt_off+0x7c/0x80
> >  Modules linked in:
> >  CPU: 0 PID: 1 Comm: init Not tainted 6.3.0-rc3-test-00011-ge11b521a7b69-dirty #31
> >  Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.0-debian-1.16.0-5 04/01/2014
> >  RIP: 0010:trace_preempt_off+0x7c/0x80
> >  Code: 74 0f 48 8b 78 08 48 89 f2 48 89 de e8 5d f9 ff ff 65 ff 0d 6e d6 ce 68 75 a2 0f 1f 44 00 00 eb 9b e8 28 ea f3 ff 84 c0 75 a5 <0f> 0b eb a1 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 f3 0f
> >  RSP: 0000:ffffaf0880013f00 EFLAGS: 00010046
> >  RAX: 0000000000000000 RBX: ffffffff97fafe59 RCX: 0000000000000001
> >  RDX: 0000000000000000 RSI: ffffffff986f4b71 RDI: ffffffff986fa0dd
> >  RBP: ffffffff97fafe59 R08: 0000000000000000 R09: 0000000000000000
> >  R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000
> >  R13: 0000000000000014 R14: 0000000000000000 R15: 0000000000000000
> >  FS:  0000000000000000(0000) GS:ffff8adf7bc00000(0000) knlGS:0000000000000000
> >  CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> >  CR2: 00007f125de099c0 CR3: 000000010498a001 CR4: 0000000000170ef0
> >  Call Trace:
> >   <TASK>
> >   irqentry_enter_from_user_mode+0x39/0x80
> >   irqentry_enter+0x51/0x60
> >   exc_page_fault+0x3d/0x200
> >   asm_exc_page_fault+0x26/0x30  
> 
> I think this was just fixed in -tip today:
> 
>   https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git/commit/?id=f87d28673b71b35b248231a2086f9404afbb7f28
> 

Thanks. Either way, I need a fix to let me pass my tests on my own code ;-)

-- Steve

Patchew 2.1-devel-b67e4c2

© 2016 - 2026 Red Hat, Inc.