* [PATCH V2] lib/smp_processor_id: Don't warn if there is only one CPU
@ 2026-11-15 9:55 Huacai Chen
2026-11-16 9:55 ` sashiko-bot
0 siblings, 1 reply; 3+ messages in thread
From: Huacai Chen @ 2026-11-15 9:55 UTC (permalink / raw)
To: Ingo Molnar, Peter Zijlstra, Juri Lelli, Vincent Guittot,
Dietmar Eggemann, Steven Rostedt, Ben Segall, Mel Gorman,
Andrew Morton, Huacai Chen
Cc: linux-mm, linux-kernel, Huacai Chen, stable, Tiezhu Yang
When CONFIG_DEBUG_PREEMPT is enabled on a Non-SMP (single-core) kernel,
calls to smp_processor_id() in preemptible contexts trigger some false
positive "BUG: using smp_processor_id() in preemptible" warnings, which
are harmless noise because the process always executes on CPU 0 and CPU
migration is impossible.
Here are some of the trimmed log messages:
Call Trace:
[<9000000000d33e28>] show_stack+0xf8/0x1c0
[<9000000000d2fdac>] dump_stack_lvl+0x4c/0x78
[<9000000002456fc0>] check_preemption_disabled.isra.0+0x100/0x110
[<9000000002456ff8>] debug_smp_processor_id+0x18/0x40
[<9000000000d31118>] init_cache_level+0x18/0x90
[<9000000001adbe2c>] detect_cache_attributes+0x4c/0x640
[<9000000001adc44c>] cacheinfo_cpu_online+0x1c/0x310
[<9000000000d5ae04>] cpuhp_issue_call.constprop.0+0x94/0xd0
[<9000000000d5b278>] __cpuhp_setup_state_cpuslocked+0x208/0x320
[<9000000000d30148>] do_one_initcall+0x68/0x200
[<90000000024711b8>] kernel_init_freeable+0x1fc/0x294
[<900000000245824c>] kernel_init+0x20/0x134
[<9000000002455b54>] ret_from_kernel_thread+0x24/0x1d0
[<9000000000d31650>] ret_from_kernel_thread_asm+0xc/0x9c
This warning only exists for the !CONFIG_SMP build. This is because for
CONFIG_SMP cpuhp_issue_call() invokes callbacks in cpuhp_thread_fun()
which is a percpu thread, but for !CONFIG_SMP build it invoke callbacks
in the init process which is not a percpu thread. And there is such a
skip condition in check_preemption_disabled():
if (is_percpu_thread())
goto out;
Since there is only one CPU for the !CONFIG_SMP, we can simply skip the
warning if (!IS_ENABLED(CONFIG_SMP)) in check_preemption_disabled(). The
effect is partially reverting 570a752b7a9bd03b50 ("lib/smp_processor_id:
Use is_percpu_thread() instead of nr_cpus_allowed"), because now in the
UP case nr_cpus_allowed is 1 but PF_NO_SETAFFINITY is unset.
Cc: stable@vger.kernel.org
Reported-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
---
V2: Keep the behaviors for CONFIG_SMP.
lib/smp_processor_id.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/lib/smp_processor_id.c b/lib/smp_processor_id.c
index 94b3f6b19538..be079b908504 100644
--- a/lib/smp_processor_id.c
+++ b/lib/smp_processor_id.c
@@ -22,6 +22,9 @@ unsigned int check_preemption_disabled(const char *what1, const char *what2)
if (is_percpu_thread())
goto out;
+ if (!IS_ENABLED(CONFIG_SMP))
+ goto out;
+
if (current->migration_disabled)
goto out;
--
2.52.0
^ permalink raw reply related [flat|nested] 3+ messages in thread* Re: [PATCH V2] lib/smp_processor_id: Don't warn if there is only one CPU 2026-11-15 9:55 [PATCH V2] lib/smp_processor_id: Don't warn if there is only one CPU Huacai Chen @ 2026-11-16 9:55 ` sashiko-bot 2026-11-18 7:40 ` Huacai Chen 0 siblings, 1 reply; 3+ messages in thread From: sashiko-bot @ 2026-11-16 9:55 UTC (permalink / raw) To: Huacai Chen, Ingo Molnar, Peter Zijlstra, Juri Lelli, Vincent Guittot, Dietmar Eggemann, Steven Rostedt, Ben Segall, Mel Gorman, Andrew Morton, Huacai Chen Cc: linux-mm, Tiezhu Yang Thank you for your contribution! Sashiko AI review found 2 potential issues to consider: - [Medium] lib/smp_processor_id: check_preemption_disabled() skips checks on !CONFIG_SMP - [Low] lib/smp_processor_id: missing Fixes trailer on stable bug-fix commit -- commit 62cd1e08e5d017184a1415887d93a1188fd3f001 Author: Huacai Chen <chenhuacai@kernel.org> > lib/smp_processor_id: Don't warn if there is only one CPU > > When CONFIG_DEBUG_PREEMPT is enabled on a Non-SMP (single-core) kernel, > calls to smp_processor_id() in preemptible contexts trigger some false > positive "BUG: using smp_processor_id() in preemptible" warnings, which > are harmless noise because the process always executes on CPU 0 and CPU > migration is impossible. > > Here are some of the trimmed log messages: > > Call Trace: > [<9000000000d33e28>] show_stack+0xf8/0x1c0 > [<9000000000d2fdac>] dump_stack_lvl+0x4c/0x78 > [<9000000002456fc0>] check_preemption_disabled.isra.0+0x100/0x110 > [<9000000002456ff8>] debug_smp_processor_id+0x18/0x40 > [<9000000000d31118>] init_cache_level+0x18/0x90 > [<9000000001adbe2c>] detect_cache_attributes+0x4c/0x640 > [<9000000001adc44c>] cacheinfo_cpu_online+0x1c/0x310 > [<9000000000d5ae04>] cpuhp_issue_call.constprop.0+0x94/0xd0 > [<9000000000d5b278>] __cpuhp_setup_state_cpuslocked+0x208/0x320 > [<9000000000d30148>] do_one_initcall+0x68/0x200 > [<90000000024711b8>] kernel_init_freeable+0x1fc/0x294 > [<900000000245824c>] kernel_init+0x20/0x134 > [<9000000002455b54>] ret_from_kernel_thread+0x24/0x1d0 > [<9000000000d31650>] ret_from_kernel_thread_asm+0xc/0x9c > > This warning only exists for the !CONFIG_SMP build. This is because for > CONFIG_SMP cpuhp_issue_call() invokes callbacks in cpuhp_thread_fun() > which is a percpu thread, but for !CONFIG_SMP build it invoke callbacks > in the init process which is not a percpu thread. And there is such a > skip condition in check_preemption_disabled(): > > if (is_percpu_thread()) > goto out; > > Since there is only one CPU for the !CONFIG_SMP, we can simply skip the > warning if (!IS_ENABLED(CONFIG_SMP)) in check_preemption_disabled(). The > effect is partially reverting 570a752b7a9bd03b50 ("lib/smp_processor_id: > Use is_percpu_thread() instead of nr_cpus_allowed"), because now in the > UP case nr_cpus_allowed is 1 but PF_NO_SETAFFINITY is unset. > > Cc: stable@vger.kernel.org > Reported-by: Tiezhu Yang <yangtiezhu@loongson.cn> > Signed-off-by: Huacai Chen <chenhuacai@loongson.cn> [Severity: Low] Since this commit is tagged for stable and addresses the !CONFIG_SMP behavior changed when is_percpu_thread() stopped unconditionally returning true on UP builds, should it include a Fixes: tag such as: Fixes: cac5cefbade9 ("sched/smp: Make SMP unconditional") Fixes: 06ddd17521bf ("sched/smp: Always define is_percpu_thread() and scheduler_ipi()") > diff --git a/lib/smp_processor_id.c b/lib/smp_processor_id.c > index 94b3f6b195388..be079b9085045 100644 > --- a/lib/smp_processor_id.c > +++ b/lib/smp_processor_id.c > @@ -22,6 +22,9 @@ unsigned int check_preemption_disabled(const char *what1, const char *what2) > if (is_percpu_thread()) > goto out; > > + if (!IS_ENABLED(CONFIG_SMP)) > + goto out; > + > if (current->migration_disabled) > goto out; [Severity: Medium] As reported by Peter Zijlstra on v1 [1], does returning early on !IS_ENABLED(CONFIG_SMP) disable all CONFIG_DEBUG_PREEMPT checks in check_preemption_disabled() on UP builds? check_preemption_disabled() also backs __this_cpu_preempt_check(), which guards non-atomic __this_cpu_*() read-modify-write operations that can still race across same-CPU preemption on a preemptible UP kernel, and suppresses warnings for smp_processor_id() callers that would be broken on SMP. Because !IS_ENABLED(CONFIG_SMP) is a compile-time constant, this also turns the remainder of check_preemption_disabled(), including the current->migration_disabled check made unconditional of CONFIG_SMP by 2885daf47081 ("lib/smp_processor_id: Make migration check unconditional of SMP"), into dead code on !CONFIG_SMP. In the reported trace, init_cache_level(unsigned int cpu) in arch/loongarch/kernel/cacheinfo.c already receives cpu as an argument but reads current_cpu_data via smp_processor_id() instead of cpu_data[cpu], while cpuhp_issue_call() invokes the callback directly in preemptible task context on !CONFIG_SMP. [1] https://lore.kernel.org/r/20261111170600.GS1779021@noisy.programming.kicks-ass.net -- Sashiko AI review · https://sashiko.dev/#/patchset/20261115095503.2922598-1-chenhuacai@loongson.cn?part=1 ^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH V2] lib/smp_processor_id: Don't warn if there is only one CPU 2026-11-16 9:55 ` sashiko-bot @ 2026-11-18 7:40 ` Huacai Chen 0 siblings, 0 replies; 3+ messages in thread From: Huacai Chen @ 2026-11-18 7:40 UTC (permalink / raw) To: sashiko-reviews Cc: Huacai Chen, Ingo Molnar, Peter Zijlstra, Juri Lelli, Vincent Guittot, Dietmar Eggemann, Steven Rostedt, Ben Segall, Mel Gorman, Andrew Morton, linux-mm, Tiezhu Yang Hi, Sashiko, On Mon, Nov 16, 2026 at 5:55 PM <sashiko-bot@kernel.org> wrote: > > Thank you for your contribution! Sashiko AI review found 2 potential > issues to consider: > - [Medium] lib/smp_processor_id: check_preemption_disabled() skips > checks on !CONFIG_SMP > - [Low] lib/smp_processor_id: missing Fixes trailer on stable bug-fix > commit > -- > > commit 62cd1e08e5d017184a1415887d93a1188fd3f001 > Author: Huacai Chen <chenhuacai@kernel.org> > > > lib/smp_processor_id: Don't warn if there is only one CPU > > > > When CONFIG_DEBUG_PREEMPT is enabled on a Non-SMP (single-core) kernel, > > calls to smp_processor_id() in preemptible contexts trigger some false > > positive "BUG: using smp_processor_id() in preemptible" warnings, which > > are harmless noise because the process always executes on CPU 0 and CPU > > migration is impossible. > > > > Here are some of the trimmed log messages: > > > > Call Trace: > > [<9000000000d33e28>] show_stack+0xf8/0x1c0 > > [<9000000000d2fdac>] dump_stack_lvl+0x4c/0x78 > > [<9000000002456fc0>] check_preemption_disabled.isra.0+0x100/0x110 > > [<9000000002456ff8>] debug_smp_processor_id+0x18/0x40 > > [<9000000000d31118>] init_cache_level+0x18/0x90 > > [<9000000001adbe2c>] detect_cache_attributes+0x4c/0x640 > > [<9000000001adc44c>] cacheinfo_cpu_online+0x1c/0x310 > > [<9000000000d5ae04>] cpuhp_issue_call.constprop.0+0x94/0xd0 > > [<9000000000d5b278>] __cpuhp_setup_state_cpuslocked+0x208/0x320 > > [<9000000000d30148>] do_one_initcall+0x68/0x200 > > [<90000000024711b8>] kernel_init_freeable+0x1fc/0x294 > > [<900000000245824c>] kernel_init+0x20/0x134 > > [<9000000002455b54>] ret_from_kernel_thread+0x24/0x1d0 > > [<9000000000d31650>] ret_from_kernel_thread_asm+0xc/0x9c > > > > This warning only exists for the !CONFIG_SMP build. This is because for > > CONFIG_SMP cpuhp_issue_call() invokes callbacks in cpuhp_thread_fun() > > which is a percpu thread, but for !CONFIG_SMP build it invoke callbacks > > in the init process which is not a percpu thread. And there is such a > > skip condition in check_preemption_disabled(): > > > > if (is_percpu_thread()) > > goto out; > > > > Since there is only one CPU for the !CONFIG_SMP, we can simply skip the > > warning if (!IS_ENABLED(CONFIG_SMP)) in check_preemption_disabled(). The > > effect is partially reverting 570a752b7a9bd03b50 ("lib/smp_processor_id: > > Use is_percpu_thread() instead of nr_cpus_allowed"), because now in the > > UP case nr_cpus_allowed is 1 but PF_NO_SETAFFINITY is unset. > > > > Cc: stable@vger.kernel.org > > Reported-by: Tiezhu Yang <yangtiezhu@loongson.cn> > > Signed-off-by: Huacai Chen <chenhuacai@loongson.cn> > > [Severity: Low] > Since this commit is tagged for stable and addresses the !CONFIG_SMP > behavior changed when is_percpu_thread() stopped unconditionally returning > true on UP builds, should it include a Fixes: tag such as: > > Fixes: cac5cefbade9 ("sched/smp: Make SMP unconditional") > Fixes: 06ddd17521bf ("sched/smp: Always define is_percpu_thread() and scheduler_ipi()") > > > diff --git a/lib/smp_processor_id.c b/lib/smp_processor_id.c > > index 94b3f6b195388..be079b9085045 100644 > > --- a/lib/smp_processor_id.c > > +++ b/lib/smp_processor_id.c > > @@ -22,6 +22,9 @@ unsigned int check_preemption_disabled(const char *what1, const char *what2) > > if (is_percpu_thread()) > > goto out; > > > > + if (!IS_ENABLED(CONFIG_SMP)) > > + goto out; > > + > > if (current->migration_disabled) > > goto out; > > [Severity: Medium] > As reported by Peter Zijlstra on v1 [1], does returning early on > !IS_ENABLED(CONFIG_SMP) disable all CONFIG_DEBUG_PREEMPT checks in > check_preemption_disabled() on UP builds? As I said, the effect of this patch is partially reverting 570a752b7a9bd03b50 ("lib/smp_processor_id: Use is_percpu_thread() instead of nr_cpus_allowed"). I think 570a752b7a9bd03b50 is mostly optimising the checking of CONFIG_SMP, so partially reverting is acceptable? > > check_preemption_disabled() also backs __this_cpu_preempt_check(), which > guards non-atomic __this_cpu_*() read-modify-write operations that can > still race across same-CPU preemption on a preemptible UP kernel, and > suppresses warnings for smp_processor_id() callers that would be broken on > SMP. Because !IS_ENABLED(CONFIG_SMP) is a compile-time constant, this > also turns the remainder of check_preemption_disabled(), including the > current->migration_disabled check made unconditional of CONFIG_SMP by > 2885daf47081 ("lib/smp_processor_id: Make migration check unconditional of > SMP"), into dead code on !CONFIG_SMP. > > In the reported trace, init_cache_level(unsigned int cpu) in > arch/loongarch/kernel/cacheinfo.c already receives cpu as an argument but > reads current_cpu_data via smp_processor_id() instead of cpu_data[cpu], > while cpuhp_issue_call() invokes the callback directly in preemptible task > context on !CONFIG_SMP. Yes, cacheinfo.c is just an example and can be fixed in other ways, but that doesn't mean this patch is unnecessary. Huacai > > [1] https://lore.kernel.org/r/20261111170600.GS1779021@noisy.programming.kicks-ass.net > > -- > Sashiko AI review · https://sashiko.dev/#/patchset/20261115095503.2922598-1-chenhuacai@loongson.cn?part=1 ^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2026-11-18 7:40 UTC | newest] Thread overview: 3+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2026-11-15 9:55 [PATCH V2] lib/smp_processor_id: Don't warn if there is only one CPU Huacai Chen 2026-11-16 9:55 ` sashiko-bot 2026-11-18 7:40 ` Huacai Chen
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.