diff mbox series

drm/i915: Fixed NULL pointer dereference in capture_engine

Message ID 5a42bhkoaoed2mky6343qz35pdyzfz6pqgyczrywqgwb4ezipv@be2aiql3bmih (mailing list archive)
State New, archived
Headers show
Series drm/i915: Fixed NULL pointer dereference in capture_engine | expand

Commit Message

Eugene Kobyak Nov. 19, 2024, 5:16 p.m. UTC
When the intel_context structure contains NULL,
it raises a NULL pointer dereference error in drm_info().

This patch aims to resolve issue:
https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/12309

Signed-off-by: Eugene Kobyak <eugene.kobyak@intel.com>
---
 drivers/gpu/drm/i915/i915_gpu_error.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

Comments

Andi Shyti Nov. 20, 2024, 10:21 a.m. UTC | #1
Hi Eugene,

On Tue, Nov 19, 2024 at 05:16:44PM +0000, Eugene Kobyak wrote:
> When the intel_context structure contains NULL,
> it raises a NULL pointer dereference error in drm_info().
> 
> This patch aims to resolve issue:
> https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/12309
> 
> Signed-off-by: Eugene Kobyak <eugene.kobyak@intel.com>

Fixes: e8a3319c31a1 ("drm/i915: Allow error capture without a request")
Cc: John Harrison <John.C.Harrison@Intel.com>
Cc: <stable@vger.kernel.org> # v6.3+

> ---
>  drivers/gpu/drm/i915/i915_gpu_error.c | 10 ++++++----
>  1 file changed, 6 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
> index 135ded17334e..b00651ad8515 100644
> --- a/drivers/gpu/drm/i915/i915_gpu_error.c
> +++ b/drivers/gpu/drm/i915/i915_gpu_error.c
> @@ -1643,11 +1643,13 @@ capture_engine(struct intel_engine_cs *engine,
>  		return NULL;
>  
>  	intel_engine_get_hung_entity(engine, &ce, &rq);
> -	if (rq && !i915_request_started(rq))
> -		drm_info(&engine->gt->i915->drm, "Got hung context on %s with active request %lld:%lld [0x%04X] not yet started\n",
> -			 engine->name, rq->fence.context, rq->fence.seqno, ce->guc_id.id);
> -
>  	if (rq) {
> +		if (!i915_request_started(rq)) {

why are you breaking the if here?

> +			u16 guc_id = ce ? ce->guc_id.id : 0;

good catch!

Andi

> +			drm_info(&engine->gt->i915->drm, "Got hung context on %s with active request %lld:%lld [0x%04X] not yet started\n",
> +				 engine->name, rq->fence.context, rq->fence.seqno, guc_id);
> +		}
>  		capture = intel_engine_coredump_add_request(ee, rq, ATOMIC_MAYFAIL);
>  		i915_request_put(rq);
>  	} else if (ce) {
> -- 
> 2.34.1
Andi Shyti Nov. 20, 2024, 11:43 a.m. UTC | #2
Hi Eugene,

> >  	intel_engine_get_hung_entity(engine, &ce, &rq);
> > -	if (rq && !i915_request_started(rq))
> > -		drm_info(&engine->gt->i915->drm, "Got hung context on %s with active request %lld:%lld [0x%04X] not yet started\n",
> > -			 engine->name, rq->fence.context, rq->fence.seqno, ce->guc_id.id);
> > -
> >  	if (rq) {
> > +		if (!i915_request_started(rq)) {
> 
> why are you breaking the if here?

Just to be clear, this is not a binding comment as you are
merging to "if (rq)". But I prefer the previous style as the line
of this drm_info() is already too long and with one more level of
indentation is even longer.

In any case:

Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com> 

and now I'm really cc'eing John.

Andi

> > +			u16 guc_id = ce ? ce->guc_id.id : 0;
> 
> good catch!
> 
> Andi
> 
> > +			drm_info(&engine->gt->i915->drm, "Got hung context on %s with active request %lld:%lld [0x%04X] not yet started\n",
> > +				 engine->name, rq->fence.context, rq->fence.seqno, guc_id);
> > +		}
> >  		capture = intel_engine_coredump_add_request(ee, rq, ATOMIC_MAYFAIL);
> >  		i915_request_put(rq);
> >  	} else if (ce) {
> > -- 
> > 2.34.1
diff mbox series

Patch

diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
index 135ded17334e..b00651ad8515 100644
--- a/drivers/gpu/drm/i915/i915_gpu_error.c
+++ b/drivers/gpu/drm/i915/i915_gpu_error.c
@@ -1643,11 +1643,13 @@  capture_engine(struct intel_engine_cs *engine,
 		return NULL;
 
 	intel_engine_get_hung_entity(engine, &ce, &rq);
-	if (rq && !i915_request_started(rq))
-		drm_info(&engine->gt->i915->drm, "Got hung context on %s with active request %lld:%lld [0x%04X] not yet started\n",
-			 engine->name, rq->fence.context, rq->fence.seqno, ce->guc_id.id);
-
 	if (rq) {
+		if (!i915_request_started(rq)) {
+			u16 guc_id = ce ? ce->guc_id.id : 0;
+
+			drm_info(&engine->gt->i915->drm, "Got hung context on %s with active request %lld:%lld [0x%04X] not yet started\n",
+				 engine->name, rq->fence.context, rq->fence.seqno, guc_id);
+		}
 		capture = intel_engine_coredump_add_request(ee, rq, ATOMIC_MAYFAIL);
 		i915_request_put(rq);
 	} else if (ce) {