Skip to contentSkip to conversation

06ddd17521: BUG:using_smp_processor_id()in_preemptible

tip:sched/core Applied kernel test robot 5 participants 12 messages Jul 10, 2025 – Jul 11, 2025

Subsystems (3)
lore
Share & tools
lore

kernel test robot

author build bot
#0 lore raw

Hello,

kernel test robot noticed "BUG:using_smp_processor_id()in_preemptible" on:

commit: 06ddd17521bf11a3e7f59dafdf5c148f29467d2c ("sched/smp: Always define is_percpu_thread() and scheduler_ipi()")
https://git.kernel.org/cgit/linux/kernel/git/tip/tip.git sched/core

[test failed on linux-next/master 835244aba90de290b4b0b1fa92b6734f3ee7b3d9]

in testcase: boot

config: x86_64-randconfig-123-20250702
compiler: gcc-12
test machine: qemu-system-x86_64 -enable-kvm -cpu SandyBridge -smp 2 -m 16G

(please refer to attached dmesg/kmsg for entire log/backtrace)

+----------------------------------------------------------+------------+------------+
|                                                          | cac5cefbad | 06ddd17521 |
+----------------------------------------------------------+------------+------------+
| boot_successes                                           | 20         | 0          |
| boot_failures                                            | 0          | 18         |
| BUG:using_smp_processor_id()in_preemptible               | 0          | 18         |
| BUG:using__this_cpu_write()in_preemptible[#]code:swapper | 0          | 18         |
| BUG:using__this_cpu_read()in_preemptible                 | 0          | 18         |
+----------------------------------------------------------+------------+------------+

If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <[email protected]>
| Closes: https://lore.kernel.org/oe-lkp/[email protected]

[   25.232998][    T1] BUG: using smp_processor_id() in preemptible [00000000] code: swapper/0/1
[ 25.233645][ T1] caller is __kvm_is_vmx_supported (arch/x86/include/asm/cpuid/api.h:74 arch/x86/include/asm/cpuid/api.h:113 arch/x86/kvm/vmx/vmx.c:2789) 
[   25.234128][    T1] CPU: 0 UID: 0 PID: 1 Comm: swapper/0 Not tainted 6.16.0-rc1-00023-g06ddd17521bf #1 PREEMPT(lazy)
[   25.234931][    T1] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.2-debian-1.16.2-1 04/01/2014
[   25.235357][    T1] Call Trace:
[   25.235357][    T1]  <TASK>
[ 25.235357][ T1] dump_stack_lvl (lib/dump_stack.c:123 (discriminator 1)) 
[ 25.235357][ T1] check_preemption_disabled (arch/x86/include/asm/preempt.h:85 lib/smp_processor_id.c:53) 
[ 25.235357][ T1] __kvm_is_vmx_supported (arch/x86/include/asm/cpuid/api.h:74 arch/x86/include/asm/cpuid/api.h:113 arch/x86/kvm/vmx/vmx.c:2789) 
[ 25.235357][ T1] ? __cpuid (arch/x86/include/asm/paravirt.h:179) 
[ 25.235357][ T1] ? __sanitizer_cov_trace_pc (kernel/kcov.c:217) 
[ 25.235357][ T1] ? ftrace_likely_update (arch/x86/include/asm/smap.h:53 kernel/trace/trace_branch.c:223) 
[ 25.235357][ T1] ? __sanitizer_cov_trace_pc (kernel/kcov.c:217) 
[ 25.235357][ T1] ? ftrace_likely_update (arch/x86/include/asm/smap.h:53 kernel/trace/trace_branch.c:223) 
[ 25.235357][ T1] ? __sanitizer_cov_trace_pc (kernel/kcov.c:217) 
[ 25.235357][ T1] ? ftrace_likely_update (arch/x86/include/asm/smap.h:53 kernel/trace/trace_branch.c:223) 
[ 25.235357][ T1] vmx_init (arch/x86/kvm/vmx/vmx.c:2808 arch/x86/kvm/vmx/vmx.c:8653) 
[ 25.235357][ T1] vt_init (arch/x86/kvm/vmx/main.c:1072) 
[ 25.235357][ T1] ? pi_init_cpu (arch/x86/kvm/vmx/main.c:1067) 
[ 25.235357][ T1] do_one_initcall (init/main.c:1273) 
[ 25.235357][ T1] ? trace_initcall_level (init/main.c:1264) 
[ 25.235357][ T1] ? __sanitizer_cov_trace_pc (kernel/kcov.c:217) 
[ 25.235357][ T1] ? ftrace_likely_update (arch/x86/include/asm/smap.h:53 kernel/trace/trace_branch.c:223) 
[ 25.235357][ T1] ? __sanitizer_cov_trace_pc (kernel/kcov.c:217) 
[ 25.235357][ T1] ? ftrace_likely_update (arch/x86/include/asm/smap.h:53 kernel/trace/trace_branch.c:223) 
[ 25.235357][ T1] do_initcalls (init/main.c:1334 init/main.c:1351) 
[ 25.235357][ T1] kernel_init_freeable (init/main.c:1587) 
[ 25.235357][ T1] ? rest_init (init/main.c:1465) 
[ 25.235357][ T1] kernel_init (init/main.c:1475) 
[ 25.235357][ T1] ? rest_init (init/main.c:1465) 
[ 25.235357][ T1] ret_from_fork (arch/x86/kernel/process.c:154) 
[ 25.235357][ T1] ? rest_init (init/main.c:1465) 
[ 25.235357][ T1] ret_from_fork_asm (arch/x86/entry/entry_64.S:255) 
[   25.235357][    T1]  </TASK>

The kernel config and materials to reproduce are available at:
https://download.01.org/0day-ci/archive/20250710/[email protected]

Signature · 2 lines
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
#1 lore raw
Quoted 63 lines · kernel test robot

Hello,

kernel test robot noticed "BUG:using_smp_processor_id()in_preemptible" on:

commit: 06ddd17521bf11a3e7f59dafdf5c148f29467d2c ("sched/smp: Always define is_percpu_thread() and scheduler_ipi()")
https://git.kernel.org/cgit/linux/kernel/git/tip/tip.git sched/core

[test failed on linux-next/master 835244aba90de290b4b0b1fa92b6734f3ee7b3d9]

in testcase: boot

config: x86_64-randconfig-123-20250702
compiler: gcc-12
test machine: qemu-system-x86_64 -enable-kvm -cpu SandyBridge -smp 2 -m 16G

(please refer to attached dmesg/kmsg for entire log/backtrace)

+----------------------------------------------------------+------------+------------+
|                                                          | cac5cefbad | 06ddd17521 |
+----------------------------------------------------------+------------+------------+
| boot_successes                                           | 20         | 0          |
| boot_failures                                            | 0          | 18         |
| BUG:using_smp_processor_id()in_preemptible               | 0          | 18         |
| BUG:using__this_cpu_write()in_preemptible[#]code:swapper | 0          | 18         |
| BUG:using__this_cpu_read()in_preemptible                 | 0          | 18         |
+----------------------------------------------------------+------------+------------+

If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <[email protected]>
| Closes: https://lore.kernel.org/oe-lkp/[email protected]

[   25.232998][    T1] BUG: using smp_processor_id() in preemptible [00000000] code: swapper/0/1
[ 25.233645][ T1] caller is __kvm_is_vmx_supported (arch/x86/include/asm/cpuid/api.h:74 arch/x86/include/asm/cpuid/api.h:113 arch/x86/kvm/vmx/vmx.c:2789)
[   25.234128][    T1] CPU: 0 UID: 0 PID: 1 Comm: swapper/0 Not tainted 6.16.0-rc1-00023-g06ddd17521bf #1 PREEMPT(lazy)
[   25.234931][    T1] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.2-debian-1.16.2-1 04/01/2014
[   25.235357][    T1] Call Trace:
[   25.235357][    T1]  <TASK>
[ 25.235357][ T1] dump_stack_lvl (lib/dump_stack.c:123 (discriminator 1))
[ 25.235357][ T1] check_preemption_disabled (arch/x86/include/asm/preempt.h:85 lib/smp_processor_id.c:53)
[ 25.235357][ T1] __kvm_is_vmx_supported (arch/x86/include/asm/cpuid/api.h:74 arch/x86/include/asm/cpuid/api.h:113 arch/x86/kvm/vmx/vmx.c:2789)
[ 25.235357][ T1] ? __cpuid (arch/x86/include/asm/paravirt.h:179)
[ 25.235357][ T1] ? __sanitizer_cov_trace_pc (kernel/kcov.c:217)
[ 25.235357][ T1] ? ftrace_likely_update (arch/x86/include/asm/smap.h:53 kernel/trace/trace_branch.c:223)
[ 25.235357][ T1] ? __sanitizer_cov_trace_pc (kernel/kcov.c:217)
[ 25.235357][ T1] ? ftrace_likely_update (arch/x86/include/asm/smap.h:53 kernel/trace/trace_branch.c:223)
[ 25.235357][ T1] ? __sanitizer_cov_trace_pc (kernel/kcov.c:217)
[ 25.235357][ T1] ? ftrace_likely_update (arch/x86/include/asm/smap.h:53 kernel/trace/trace_branch.c:223)
[ 25.235357][ T1] vmx_init (arch/x86/kvm/vmx/vmx.c:2808 arch/x86/kvm/vmx/vmx.c:8653)
[ 25.235357][ T1] vt_init (arch/x86/kvm/vmx/main.c:1072)
[ 25.235357][ T1] ? pi_init_cpu (arch/x86/kvm/vmx/main.c:1067)
[ 25.235357][ T1] do_one_initcall (init/main.c:1273)
[ 25.235357][ T1] ? trace_initcall_level (init/main.c:1264)
[ 25.235357][ T1] ? __sanitizer_cov_trace_pc (kernel/kcov.c:217)
[ 25.235357][ T1] ? ftrace_likely_update (arch/x86/include/asm/smap.h:53 kernel/trace/trace_branch.c:223)
[ 25.235357][ T1] ? __sanitizer_cov_trace_pc (kernel/kcov.c:217)
[ 25.235357][ T1] ? ftrace_likely_update (arch/x86/include/asm/smap.h:53 kernel/trace/trace_branch.c:223)
[ 25.235357][ T1] do_initcalls (init/main.c:1334 init/main.c:1351)
Quoted 9 lines
[ 25.235357][ T1] kernel_init_freeable (init/main.c:1587)
[ 25.235357][ T1] ? rest_init (init/main.c:1465)
[ 25.235357][ T1] kernel_init (init/main.c:1475)
[ 25.235357][ T1] ? rest_init (init/main.c:1465)
[ 25.235357][ T1] ret_from_fork (arch/x86/kernel/process.c:154)
[ 25.235357][ T1] ? rest_init (init/main.c:1465)
[ 25.235357][ T1] ret_from_fork_asm (arch/x86/entry/entry_64.S:255)
[   25.235357][    T1]  </TASK>

I took a glance at the warning, before this patch, is_percpu_thread() always return true when CONFIG_SMP is not set. After this patch,
is_percpu_thread() checks the current task's CPU affinity. So debug_smp_processor_id() -> check_preemption_disabled() -> is_percpu_thread() might not always return true anymore, which caused the warning.

Actually the issue is in __kvm_is_vmx_supported(), should we use something like this below:

arch/x86/kvm/vmx/vmx.c +2 -1
index 5c5766467a61..9cba7b061892 100644
@@ -2787,19 +2787,22 @@ static int setup_vmcs_config(struct vmcs_config *vmcs_conf,
 
 static bool __kvm_is_vmx_supported(void)
 {
-       int cpu = smp_processor_id();
+       int cpu = get_cpu();
 
        if (!(cpuid_ecx(1) & feature_bit(VMX))) {
                pr_err("VMX not supported by CPU %d\n", cpu);
+               put_cpu();
                return false;
        }
 
        if (!this_cpu_has(X86_FEATURE_MSR_IA32_FEAT_CTL) ||
            !this_cpu_has(X86_FEATURE_VMX)) {
                pr_err("VMX not enabled (by BIOS) in MSR_IA32_FEAT_CTL 

on CPU %d\n", cpu);

+               put_cpu();
                return false;
        }
+       put_cpu();
        return true;

}

thanks,
Chenyu

#2 lore raw
Quoted 5 lines · Yu C
Quoted 5 lines
[ 25.235357][ T1] dump_stack_lvl (lib/dump_stack.c:123 (discriminator 1))
[ 25.235357][ T1] check_preemption_disabled (arch/x86/include/asm/preempt.h:85 lib/smp_processor_id.c:53)
[ 25.235357][ T1] __kvm_is_vmx_supported (arch/x86/include/asm/cpuid/api.h:74 arch/x86/include/asm/cpuid/api.h:113 arch/x86/kvm/vmx/vmx.c:2789)
[ 25.235357][ T1] vmx_init (arch/x86/kvm/vmx/vmx.c:2808 arch/x86/kvm/vmx/vmx.c:8653)
[ 25.235357][ T1] vt_init (arch/x86/kvm/vmx/main.c:1072)

…

Quoted 10 lines

I took a glance at the warning, before this patch, is_percpu_thread() always return true when CONFIG_SMP is not set. After this patch,
is_percpu_thread() checks the current task's CPU affinity. So debug_smp_processor_id() -> check_preemption_disabled() -> is_percpu_thread() might not always return true anymore, which caused the warning.

Actually the issue is in __kvm_is_vmx_supported(), should we use something like this below:

No, it should not. If you look closely you will see that the call chains is
vmx_init() -> kvm_is_vmx_supported() -> __kvm_is_vmx_supported()

There is a migrate_disable() around __kvm_is_vmx_supported(). So why does this warning trigger then?

Sebastian

#3 lore raw

Hi Sebastian,

Quoted 18 lines · Sebastian Andrzej Siewior
Quoted 5 lines · Yu C
Quoted 5 lines
[ 25.235357][ T1] dump_stack_lvl (lib/dump_stack.c:123 (discriminator 1))
[ 25.235357][ T1] check_preemption_disabled (arch/x86/include/asm/preempt.h:85 lib/smp_processor_id.c:53)
[ 25.235357][ T1] __kvm_is_vmx_supported (arch/x86/include/asm/cpuid/api.h:74 arch/x86/include/asm/cpuid/api.h:113 arch/x86/kvm/vmx/vmx.c:2789)
[ 25.235357][ T1] vmx_init (arch/x86/kvm/vmx/vmx.c:2808 arch/x86/kvm/vmx/vmx.c:8653)
[ 25.235357][ T1] vt_init (arch/x86/kvm/vmx/main.c:1072)

…

Quoted 10 lines

I took a glance at the warning, before this patch, is_percpu_thread() always return true when CONFIG_SMP is not set. After this patch,
is_percpu_thread() checks the current task's CPU affinity. So debug_smp_processor_id() -> check_preemption_disabled() -> is_percpu_thread() might not always return true anymore, which caused the warning.

Actually the issue is in __kvm_is_vmx_supported(), should we use something like this below:

Quoted 7 lines

No, it should not. If you look closely you will see that the call chains is

    vmx_init() -> kvm_is_vmx_supported() -> __kvm_is_vmx_supported()

There is a migrate_disable() around __kvm_is_vmx_supported(). So why does this warning trigger then?

migrate_disable() disables the task migration between CPUs by restricting the task's affinity, but it does not disable the preemption on single CPUs IMO. The scope of guard(preempt)() in migrate_disable() is just within the migrate_disable(). debug_smp_processor_id() warns when the preemption is enabled.

Thanks,
Chenyu

#4 lore raw
Quoted 5 lines · Yu C

migrate_disable() disables the task migration between CPUs by restricting the task's affinity, but it does not disable the preemption on single CPUs IMO. The scope of guard(preempt)() in migrate_disable() is just within the migrate_disable(). debug_smp_processor_id() warns when the preemption is enabled.

does migrate_disable() really have no effect on
debug_smp_processor_id()?

Quoted 2 lines

Thanks,
Chenyu

Sebastian

#5 lore raw

PATCH 1/1 lib/smp_processor_id: Make migration check unconditional of SMP

Commit cac5cefbade90 ("sched/smp: Make SMP unconditional") migrate_disable() even on UP builds.
Commit 06ddd17521bf1 ("sched/smp: Always define is_percpu_thread() and scheduler_ipi()") made is_percpu_thread() check the affinity mask instead replying always true for UP mask.

As a consequence smp_processor_id() now complains if invoked within a migrate_disable() section because is_percpu_thread() checks its mask and the migration check is left out.

Make migration check unconditional of SMP.

---
 lib/smp_processor_id.c | 2 --
 1 file changed, 2 deletions(-)
lib/smp_processor_id.c +0 -2
index a2bb7738c373c..94b3f6b195388 100644
@@ -22,10 +22,8 @@ unsigned int check_preemption_disabled(const char *what1, const char *what2)
 	if (is_percpu_thread())
 		goto out;
 
-#ifdef CONFIG_SMP
 	if (current->migration_disabled)
 		goto out;
-#endif
 
 	/*
 	 * It is valid to assume CPU-locality during early bootup:
Signature · 1 line
2.50.0
#9 lore raw
In reply to Sebastian Andrzej Siewior
Quoted 10 lines · Sebastian Andrzej Siewior
Quoted 5 lines · Yu C

migrate_disable() disables the task migration between CPUs by restricting the task's affinity, but it does not disable the preemption on single CPUs IMO. The scope of guard(preempt)() in migrate_disable() is just within the migrate_disable(). debug_smp_processor_id() warns when the preemption is enabled.

does migrate_disable() really have no effect on
debug_smp_processor_id()?

I see. So there is still a CONFIG_SMP to check migrate_disbled in debug_smp_processor_id(), and that was skipped if we do not have CONFIG_SMP set.

thanks,
Chenyu>> Thanks,

Quoted 2 lines
Quoted 1 line

Chenyu

Sebastian

Reviewed-by on v1 1/1
#6 lore raw

Re: [PATCH] lib/smp_processor_id: Make migration check unconditional of SMP

In reply to Sebastian Andrzej Siewior
Quoted 10 lines · Sebastian Andrzej Siewior

Commit cac5cefbade90 ("sched/smp: Make SMP unconditional") migrate_disable() even on UP builds.
Commit 06ddd17521bf1 ("sched/smp: Always define is_percpu_thread() and scheduler_ipi()") made is_percpu_thread() check the affinity mask instead replying always true for UP mask.

As a consequence smp_processor_id() now complains if invoked within a migrate_disable() section because is_percpu_thread() checks its mask and the migration check is left out.

Quoted 7 lines

Make migration check unconditional of SMP.

---

From my understanding,

thank you for fixing this.

best,
Chenyu

#7 lore raw

Re: [PATCH] lib/smp_processor_id: Make migration check unconditional of SMP

Quoted 19 lines · Yu C
Quoted 10 lines · Sebastian Andrzej Siewior

Commit cac5cefbade90 ("sched/smp: Make SMP unconditional") migrate_disable() even on UP builds.
Commit 06ddd17521bf1 ("sched/smp: Always define is_percpu_thread() and scheduler_ipi()") made is_percpu_thread() check the affinity mask instead replying always true for UP mask.

As a consequence smp_processor_id() now complains if invoked within a migrate_disable() section because is_percpu_thread() checks its mask and the migration check is left out.

Quoted 7 lines

Make migration check unconditional of SMP.

---

Quoted 3 lines

From my understanding,

Thanks; let me go stick this in a tree somewhere.

#10 lore raw
In reply to Chen, Yu C
Quoted 3 lines · Yu C

I see. So there is still a CONFIG_SMP to check migrate_disbled in debug_smp_processor_id(), and that was skipped if we do not have CONFIG_SMP set.

Exactly, sorry for that.
I planned to let you do the figuring out and sending the patch but got confused midway, had to look it up and then had the patch ready and decided to send it…

Quoted 2 lines

thanks,
Chenyu>> Thanks,

Sebastian

#11 lore raw
Quoted 10 lines · Sebastian Andrzej Siewior
Quoted 3 lines · Yu C

I see. So there is still a CONFIG_SMP to check migrate_disbled in debug_smp_processor_id(), and that was skipped if we do not have CONFIG_SMP set.

Exactly, sorry for that.
I planned to let you do the figuring out and sending the patch but got confused midway, had to look it up and then had the patch ready and decided to send it…

No worries! Thanks for the patch. Glad it’s sorted.

thanks,
Chenyu>> thanks,

Quoted 2 lines
Quoted 1 line

Chenyu>> Thanks,

Sebastian

#8 lore raw

[tip: sched/core] lib/smp_processor_id: Make migration check unconditional of SMP

In reply to Sebastian Andrzej Siewior
The following commit has been merged into the sched/core branch of tip:

The following commit has been merged into the sched/core branch of tip:

Commit-ID:     2885daf47081dd1aaf1a588e9d001eb343df1f90
Gitweb:        https://git.kernel.org/tip/2885daf47081dd1aaf1a588e9d001eb343df1f90
Author:        Sebastian Andrzej Siewior <[email protected]>
AuthorDate:    Thu, 10 Jul 2025 10:27:48 +02:00
Committer:     Peter Zijlstra <[email protected]>
CommitterDate: Thu, 10 Jul 2025 17:52:13 +02:00

lib/smp_processor_id: Make migration check unconditional of SMP

Commit cac5cefbade90 ("sched/smp: Make SMP unconditional") migrate_disable() even on UP builds.
Commit 06ddd17521bf1 ("sched/smp: Always define is_percpu_thread() and scheduler_ipi()") made is_percpu_thread() check the affinity mask instead replying always true for UP mask.

As a consequence smp_processor_id() now complains if invoked within a migrate_disable() section because is_percpu_thread() checks its mask and the migration check is left out.

Make migration check unconditional of SMP.

---
 lib/smp_processor_id.c | 2 --
 1 file changed, 2 deletions(-)
lib/smp_processor_id.c +0 -2
index a2bb773..94b3f6b 100644
@@ -22,10 +22,8 @@ unsigned int check_preemption_disabled(const char *what1, const char *what2)
 	if (is_percpu_thread())
 		goto out;
 
-#ifdef CONFIG_SMP
 	if (current->migration_disabled)
 		goto out;
-#endif
 
 	/*
 	 * It is valid to assume CPU-locality during early bootup: