Re: [PATCH] perf/core: Prevent WARN_ON(!ctx) in __free_event for partial init

From: Gabriel
Date: Sun Apr 06 2025 - 03:21:09 EST


On Sat, Apr 5, 2025 at 4:30 PM Gabriel Shahrouzi <gshahrouzi@xxxxxxxxx> wrote:
>
> Move the get_ctx(child_ctx) call and the child_event->ctx assignment to
> occur immediately after the child event is allocated. Ensure that
> child_event->ctx is non-NULL before any subsequent error path within
> inherit_event calls free_event(), satisfying the assumptions of the
> cleanup code.
>
> Reported-by: syzbot+ff3aa851d46ab82953a3@xxxxxxxxxxxxxxxxxxxxxxxxx
> Closes: https://syzkaller.appspot.com/bug?extid=ff3aa851d46ab82953a3
> Signed-off-by: Gabriel Shahrouzi <gshahrouzi@xxxxxxxxx>
> ---
> kernel/events/core.c | 6 +++---
> 1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/kernel/events/core.c b/kernel/events/core.c
> index 0bb21659e252..153ba622cfa0 100644
> --- a/kernel/events/core.c
> +++ b/kernel/events/core.c
> @@ -14016,6 +14016,9 @@ inherit_event(struct perf_event *parent_event,
> if (IS_ERR(child_event))
> return child_event;
>
> + get_ctx(child_ctx);
> + child_event->ctx = child_ctx;
> +
> pmu_ctx = find_get_pmu_context(child_event->pmu, child_ctx, child_event);
> if (IS_ERR(pmu_ctx)) {
> free_event(child_event);
> @@ -14037,8 +14040,6 @@ inherit_event(struct perf_event *parent_event,
> return NULL;
> }
>
> - get_ctx(child_ctx);
> -
> /*
> * Make the child state follow the state of the parent event,
> * not its attr.disabled bit. We hold the parent's mutex,
> @@ -14059,7 +14060,6 @@ inherit_event(struct perf_event *parent_event,
> local64_set(&hwc->period_left, sample_period);
> }
>
> - child_event->ctx = child_ctx;
> child_event->overflow_handler = parent_event->overflow_handler;
> child_event->overflow_handler_context
> = parent_event->overflow_handler_context;
> --
> 2.43.0
>
I want to clarify three things [1] why the Fixes: tag was not used;
[2] what the author’s original intent was and how to maintain it; [3]
how to improve my patch.

[1] This bug is a side-effect of multiple interacting commits over
time (up to 15 years old), not a single regression. Bisection yielded
a false positive
(https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next.git/commit/?id=00b35530811f2aa3d7ceec2dbada80861c7632a8),
and testing on an earlier branch (6.14-c8) hit unrelated issues,
confirming no single commit is responsible. A Fixes: tag is therefore
unsuitable.

[2] The code initially incremented refcount then assigned context
immediately after the child_event was created. Later, an early
validity check for child_event was added before the
refcount/assignment. Even later, a WARN_ON_ONCE cleanup check was
added, assuming event->ctx is valid if the pmu_ctx is valid.
The problem is that the WARN_ON_ONCE could trigger after the initial
check passed but before child_event->ctx was assigned, violating its
precondition. The solution is to assign child_event->ctx right after
its initial validation. This ensures the context exists for any
subsequent checks or cleanup routines, resolving the WARN_ON_ONCE.

[3]. Defer the refcount update and child_event->ctx assignment
directly after child_event->pmu_ctx is set but before checking if the
parent event is orphaned. The cleanup routine depends on
event->pmu_ctx being non-NULL before it verifies event->ctx is
non-NULL. This also maintains the author's original intent of passing
in child_ctx to find_get_pmu_context before its refcount/assignment.