[PATCH] smpboot: Add NULL check before kthread_unpark()

[PATCH] smpboot: Add NULL check before kthread_unpark()

[Kaushlendra Kumar]

Add NULL check for tsk before calling kthread_unpark() in
smpboot_unpark_thread() to prevent potential null pointer
dereference.

Signed-off-by: Kaushlendra Kumar <kaushlendra.kumar@intel.com>
---
 kernel/smpboot.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/smpboot.c b/kernel/smpboot.c
index 4503b60ce9bd..a37d25b23f58 100644
--- a/kernel/smpboot.c
+++ b/kernel/smpboot.c
@@ -225,7 +225,7 @@ static void smpboot_unpark_thread(struct smp_hotplug_thread *ht, unsigned int cp
 {
 	struct task_struct *tsk = *per_cpu_ptr(ht->store, cpu);
 
-	if (!ht->selfparking)
+	if (tsk && !ht->selfparking)
 		kthread_unpark(tsk);
 }
 
-- 
2.34.1

Re: [PATCH] smpboot: Add NULL check before kthread_unpark()

[Thomas Gleixner]

On Mon, Nov 24 2025 at 10:41, Kaushlendra Kumar wrote:
> Add NULL check for tsk before calling kthread_unpark() in
> smpboot_unpark_thread() to prevent potential null pointer
> dereference.

How is that ever going to be NULL?

Thanks,

        tglx

RE: [PATCH] smpboot: Add NULL check before kthread_unpark()

[Kumar, Kaushlendra]

On Mon, Nov 24 2025, Thomas Gleixner wrote:
> On Mon, Nov 24 2025 at 10:41, Kaushlendra Kumar wrote:
>> Add NULL check for tsk before calling kthread_unpark() in
>> smpboot_unpark_thread() to prevent potential null pointer dereference.
>
> How is that ever going to be NULL?
Thanks for the review.
Here I was only concerned about edge cases during CPU hotplug state transitions.

Could you help clarify: are there any scenarios during the CPU hotplug state 
transition where this function might be called without a valid task, or is the 
current flow guaranteed to always have tsk initialized at this point?

If the existing code ensure tsk is always valid here, will drop this patch.



-Kaushlendra

RE: [PATCH] smpboot: Add NULL check before kthread_unpark()

[Thomas Gleixner]

On Wed, Dec 10 2025 at 08:21, Kaushlendra Kumar wrote:
> On Mon, Nov 24 2025, Thomas Gleixner wrote:
>> On Mon, Nov 24 2025 at 10:41, Kaushlendra Kumar wrote:
>>> Add NULL check for tsk before calling kthread_unpark() in
>>> smpboot_unpark_thread() to prevent potential null pointer dereference.
>>
>> How is that ever going to be NULL?
> Thanks for the review.
> Here I was only concerned about edge cases during CPU hotplug state transitions.

Then describe your concern and the edge cases in detail.

Thanks,

        tglx