From: Josh Poimboeuf <jpoimboe@kernel.org>

If the task doesn't have any memory, there's no stack to unwind.

Signed-off-by: Josh Poimboeuf <jpoimboe@kernel.org>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
---
Note from v4: https://lore.kernel.org/all/24cde1381495cd637f0d5663ab7629572bc0eca6.1737511963.git.jpoimboe@kernel.org/

- From the previous version, Peter Zijlstra brought up concern about
  io_uring helpers, but this was never resolved. No update was made.

 kernel/events/core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/events/core.c b/kernel/events/core.c
index 009f34bd9e63..40b98b05fb7c 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -8101,7 +8101,7 @@ struct perf_callchain_entry *
 perf_callchain(struct perf_event *event, struct pt_regs *regs)
 {
 	bool kernel = !event->attr.exclude_callchain_kernel;
-	bool user   = !event->attr.exclude_callchain_user;
+	bool user   = !event->attr.exclude_callchain_user && current->mm;
 	/* Disallow cross-task user callchains. */
 	bool crosstask = event->ctx->task && event->ctx->task != current;
 	const u32 max_stack = event->attr.sample_max_stack;
-- 
2.47.2

On Thu, Apr 24, 2025 at 12:25:41PM -0400, Steven Rostedt wrote:
> From: Josh Poimboeuf <jpoimboe@kernel.org>
> 
> If the task doesn't have any memory, there's no stack to unwind.
> 
> Signed-off-by: Josh Poimboeuf <jpoimboe@kernel.org>
> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
> ---
> Note from v4: https://lore.kernel.org/all/24cde1381495cd637f0d5663ab7629572bc0eca6.1737511963.git.jpoimboe@kernel.org/
> 
> - From the previous version, Peter Zijlstra brought up concern about
>   io_uring helpers, but this was never resolved. No update was made.

What is there to resolve? If you want to check for kthreads, don't check
for ->mm

> 
>  kernel/events/core.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/kernel/events/core.c b/kernel/events/core.c
> index 009f34bd9e63..40b98b05fb7c 100644
> --- a/kernel/events/core.c
> +++ b/kernel/events/core.c
> @@ -8101,7 +8101,7 @@ struct perf_callchain_entry *
>  perf_callchain(struct perf_event *event, struct pt_regs *regs)
>  {
>  	bool kernel = !event->attr.exclude_callchain_kernel;
> -	bool user   = !event->attr.exclude_callchain_user;
> +	bool user   = !event->attr.exclude_callchain_user && current->mm;
>  	/* Disallow cross-task user callchains. */
>  	bool crosstask = event->ctx->task && event->ctx->task != current;
>  	const u32 max_stack = event->attr.sample_max_stack;
> -- 
> 2.47.2
> 
>

On Thu, 24 Apr 2025 18:37:18 +0200
Peter Zijlstra <peterz@infradead.org> wrote:

> > Note from v4: https://lore.kernel.org/all/24cde1381495cd637f0d5663ab7629572bc0eca6.1737511963.git.jpoimboe@kernel.org/
> > 
> > - From the previous version, Peter Zijlstra brought up concern about
> >   io_uring helpers, but this was never resolved. No update was made.  
> 
> What is there to resolve? If you want to check for kthreads, don't check
> for ->mm

I misunderstood what the issue was then ;-)

OK, so the check should be task->flags & PF_KTHREAD ?

-- Steve