Revert "bpf: remove unnecessary rcu_read_{lock,unlock}() in multi-uprobe attach logic"

[PATCH] Revert "bpf: remove unnecessary rcu_read_{lock,unlock}() in multi-uprobe attach logic"

Posted by Xuewen Yan 7 months ago

From: Di Shen <di.shen@unisoc.com>

This reverts commit 4a8f635a60540888dab3804992e86410360339c8.

Althought get_pid_task() internally already calls rcu_read_lock() and
rcu_read_unlock(), the find_vpid() was not.

The documentation for find_vpid() clearly states:

  "Must be called with the tasklist_lock or rcu_read_lock() held."

Add proper rcu_read_lock/unlock() to protect the find_vpid().

Reported-by: Xuewen Yan <xuewen.yan@unisoc.com>
Signed-off-by: Di Shen <di.shen@unisoc.com>
---
 kernel/trace/bpf_trace.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
index 187dc37d61d4..0c4b6af10601 100644
--- a/kernel/trace/bpf_trace.c
+++ b/kernel/trace/bpf_trace.c
@@ -3417,7 +3417,9 @@ int bpf_uprobe_multi_link_attach(const union bpf_attr *attr, struct bpf_prog *pr
 	}
 
 	if (pid) {
+		rcu_read_lock();
 		task = get_pid_task(find_vpid(pid), PIDTYPE_TGID);
+		rcu_read_unlock();
 		if (!task) {
 			err = -ESRCH;
 			goto error_path_put;
-- 
2.25.1

Re: [PATCH] Revert "bpf: remove unnecessary rcu_read_{lock,unlock}() in multi-uprobe attach logic"

Posted by Alexei Starovoitov 7 months ago

On Mon, May 19, 2025 at 10:51 PM Xuewen Yan <xuewen.yan@unisoc.com> wrote:
>
> From: Di Shen <di.shen@unisoc.com>
>
> This reverts commit 4a8f635a60540888dab3804992e86410360339c8.
>
> Althought get_pid_task() internally already calls rcu_read_lock() and
> rcu_read_unlock(), the find_vpid() was not.
>
> The documentation for find_vpid() clearly states:
>
>   "Must be called with the tasklist_lock or rcu_read_lock() held."
>
> Add proper rcu_read_lock/unlock() to protect the find_vpid().
>
> Reported-by: Xuewen Yan <xuewen.yan@unisoc.com>
> Signed-off-by: Di Shen <di.shen@unisoc.com>
> ---
>  kernel/trace/bpf_trace.c | 2 ++
>  1 file changed, 2 insertions(+)
>
> diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
> index 187dc37d61d4..0c4b6af10601 100644
> --- a/kernel/trace/bpf_trace.c
> +++ b/kernel/trace/bpf_trace.c
> @@ -3417,7 +3417,9 @@ int bpf_uprobe_multi_link_attach(const union bpf_attr *attr, struct bpf_prog *pr
>         }
>
>         if (pid) {
> +               rcu_read_lock();
>                 task = get_pid_task(find_vpid(pid), PIDTYPE_TGID);
> +               rcu_read_unlock();
>                 if (!task) {
>                         err = -ESRCH;
>                         goto error_path_put;

hmm. indeed.

Jiri ?

Re: [PATCH] Revert "bpf: remove unnecessary rcu_read_{lock,unlock}() in multi-uprobe attach logic"

Posted by Andrii Nakryiko 7 months ago

On Tue, May 20, 2025 at 1:08 PM Alexei Starovoitov
<alexei.starovoitov@gmail.com> wrote:
>
> On Mon, May 19, 2025 at 10:51 PM Xuewen Yan <xuewen.yan@unisoc.com> wrote:
> >
> > From: Di Shen <di.shen@unisoc.com>
> >
> > This reverts commit 4a8f635a60540888dab3804992e86410360339c8.
> >
> > Althought get_pid_task() internally already calls rcu_read_lock() and
> > rcu_read_unlock(), the find_vpid() was not.
> >
> > The documentation for find_vpid() clearly states:
> >
> >   "Must be called with the tasklist_lock or rcu_read_lock() held."
> >
> > Add proper rcu_read_lock/unlock() to protect the find_vpid().
> >
> > Reported-by: Xuewen Yan <xuewen.yan@unisoc.com>
> > Signed-off-by: Di Shen <di.shen@unisoc.com>
> > ---
> >  kernel/trace/bpf_trace.c | 2 ++
> >  1 file changed, 2 insertions(+)
> >
> > diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
> > index 187dc37d61d4..0c4b6af10601 100644
> > --- a/kernel/trace/bpf_trace.c
> > +++ b/kernel/trace/bpf_trace.c
> > @@ -3417,7 +3417,9 @@ int bpf_uprobe_multi_link_attach(const union bpf_attr *attr, struct bpf_prog *pr
> >         }
> >
> >         if (pid) {
> > +               rcu_read_lock();
> >                 task = get_pid_task(find_vpid(pid), PIDTYPE_TGID);
> > +               rcu_read_unlock();
> >                 if (!task) {
> >                         err = -ESRCH;
> >                         goto error_path_put;
>
> hmm. indeed.
>

yep, my bad, missed find_vpid() restrictions. revert LGTM

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

> Jiri ?