[PATCH v2] perf: get_perf_callchain return NULL for crosstask

[PATCH v2] perf: get_perf_callchain return NULL for crosstask

[Jordan Rome]

Return NULL instead of returning 1 incorrect frame, which
currently happens when trying to walk the user stack for
any task that isn't current. Returning NULL is a better
indicator that this behavior is not supported.

This issue was found using bpf_get_task_stack inside a BPF
iterator ("iter/task"), which iterates over all tasks. The
single address/frame in the buffer when getting user stacks
for tasks that aren't current could not be symbolized (testing
multiple symbolizers).

Signed-off-by: Jordan Rome <linux@jordanrome.com>
---

Changes in v2:
* move user and crosstask check before get_callchain_entry

v1:
https://lore.kernel.org/linux-perf-users/CAEf4BzaWtOeTBb_+b7Td3NHaKjZU+OohuBJje_nvw9kd6xPA3g@mail.gmail.com/T/#t

 kernel/events/callchain.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/kernel/events/callchain.c b/kernel/events/callchain.c
index 1273be84392c..104ea2975a57 100644
--- a/kernel/events/callchain.c
+++ b/kernel/events/callchain.c
@@ -184,6 +184,9 @@ get_perf_callchain(struct pt_regs *regs, u32 init_nr, bool kernel, bool user,
 	struct perf_callchain_entry_ctx ctx;
 	int rctx;
 
+	if (user && crosstask)
+		return NULL;
+
 	entry = get_callchain_entry(&rctx);
 	if (!entry)
 		return NULL;
@@ -209,9 +212,6 @@ get_perf_callchain(struct pt_regs *regs, u32 init_nr, bool kernel, bool user,
 		}
 
 		if (regs) {
-			if (crosstask)
-				goto exit_put;
-
 			if (add_mark)
 				perf_callchain_store_context(&ctx, PERF_CONTEXT_USER);
 
@@ -219,7 +219,6 @@ get_perf_callchain(struct pt_regs *regs, u32 init_nr, bool kernel, bool user,
 		}
 	}
 
-exit_put:
 	put_callchain_entry(rctx);
 
 	return entry;
-- 
2.39.3

Re: [PATCH v2] perf: get_perf_callchain return NULL for crosstask

[Andrii Nakryiko]

On Sat, Nov 11, 2023 at 9:20 AM Jordan Rome <linux@jordanrome.com> wrote:
>
> Return NULL instead of returning 1 incorrect frame, which
> currently happens when trying to walk the user stack for
> any task that isn't current. Returning NULL is a better
> indicator that this behavior is not supported.
>
> This issue was found using bpf_get_task_stack inside a BPF
> iterator ("iter/task"), which iterates over all tasks. The
> single address/frame in the buffer when getting user stacks
> for tasks that aren't current could not be symbolized (testing
> multiple symbolizers).
>
> Signed-off-by: Jordan Rome <linux@jordanrome.com>
> ---
>
> Changes in v2:
> * move user and crosstask check before get_callchain_entry
>

LGTM.

Acked-by: Andrii Nakryiko <andrii@kernel.org>

> v1:
> https://lore.kernel.org/linux-perf-users/CAEf4BzaWtOeTBb_+b7Td3NHaKjZU+OohuBJje_nvw9kd6xPA3g@mail.gmail.com/T/#t
>
>  kernel/events/callchain.c | 7 +++----
>  1 file changed, 3 insertions(+), 4 deletions(-)
>
> diff --git a/kernel/events/callchain.c b/kernel/events/callchain.c
> index 1273be84392c..104ea2975a57 100644
> --- a/kernel/events/callchain.c
> +++ b/kernel/events/callchain.c
> @@ -184,6 +184,9 @@ get_perf_callchain(struct pt_regs *regs, u32 init_nr, bool kernel, bool user,
>         struct perf_callchain_entry_ctx ctx;
>         int rctx;
>
> +       if (user && crosstask)
> +               return NULL;
> +
>         entry = get_callchain_entry(&rctx);
>         if (!entry)
>                 return NULL;
> @@ -209,9 +212,6 @@ get_perf_callchain(struct pt_regs *regs, u32 init_nr, bool kernel, bool user,
>                 }
>
>                 if (regs) {
> -                       if (crosstask)
> -                               goto exit_put;
> -
>                         if (add_mark)
>                                 perf_callchain_store_context(&ctx, PERF_CONTEXT_USER);
>
> @@ -219,7 +219,6 @@ get_perf_callchain(struct pt_regs *regs, u32 init_nr, bool kernel, bool user,
>                 }
>         }
>
> -exit_put:
>         put_callchain_entry(rctx);
>
>         return entry;
> --
> 2.39.3
>

Re: [PATCH v2] perf: get_perf_callchain return NULL for crosstask

[Jiri Olsa]

On Sat, Nov 11, 2023 at 09:20:01AM -0800, Jordan Rome wrote:
> Return NULL instead of returning 1 incorrect frame, which
> currently happens when trying to walk the user stack for
> any task that isn't current. Returning NULL is a better
> indicator that this behavior is not supported.
> 
> This issue was found using bpf_get_task_stack inside a BPF
> iterator ("iter/task"), which iterates over all tasks. The
> single address/frame in the buffer when getting user stacks
> for tasks that aren't current could not be symbolized (testing
> multiple symbolizers).
> 
> Signed-off-by: Jordan Rome <linux@jordanrome.com>

Acked-by: Jiri Olsa <jolsa@kernel.org>

jirka

> ---
> 
> Changes in v2:
> * move user and crosstask check before get_callchain_entry
> 
> v1:
> https://lore.kernel.org/linux-perf-users/CAEf4BzaWtOeTBb_+b7Td3NHaKjZU+OohuBJje_nvw9kd6xPA3g@mail.gmail.com/T/#t
> 
>  kernel/events/callchain.c | 7 +++----
>  1 file changed, 3 insertions(+), 4 deletions(-)
> 
> diff --git a/kernel/events/callchain.c b/kernel/events/callchain.c
> index 1273be84392c..104ea2975a57 100644
> --- a/kernel/events/callchain.c
> +++ b/kernel/events/callchain.c
> @@ -184,6 +184,9 @@ get_perf_callchain(struct pt_regs *regs, u32 init_nr, bool kernel, bool user,
>  	struct perf_callchain_entry_ctx ctx;
>  	int rctx;
>  
> +	if (user && crosstask)
> +		return NULL;
> +
>  	entry = get_callchain_entry(&rctx);
>  	if (!entry)
>  		return NULL;
> @@ -209,9 +212,6 @@ get_perf_callchain(struct pt_regs *regs, u32 init_nr, bool kernel, bool user,
>  		}
>  
>  		if (regs) {
> -			if (crosstask)
> -				goto exit_put;
> -
>  			if (add_mark)
>  				perf_callchain_store_context(&ctx, PERF_CONTEXT_USER);
>  
> @@ -219,7 +219,6 @@ get_perf_callchain(struct pt_regs *regs, u32 init_nr, bool kernel, bool user,
>  		}
>  	}
>  
> -exit_put:
>  	put_callchain_entry(rctx);
>  
>  	return entry;
> -- 
> 2.39.3
> 
>