[PATCH v2] drm/amdgpu: add a missing lock for AMDGPU_SCHED

[PATCH v2] drm/amdgpu: add a missing lock for AMDGPU_SCHED

[Chia-I Wu]

mgr->ctx_handles should be protected by mgr->lock.

v2: improve commit message

Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Cc: stable@vger.kernel.org
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c
index e9b45089a28a6..863b2a34b2d64 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c
@@ -38,6 +38,7 @@ static int amdgpu_sched_process_priority_override(struct amdgpu_device *adev,
 {
 	struct fd f = fdget(fd);
 	struct amdgpu_fpriv *fpriv;
+	struct amdgpu_ctx_mgr *mgr;
 	struct amdgpu_ctx *ctx;
 	uint32_t id;
 	int r;
@@ -51,8 +52,11 @@ static int amdgpu_sched_process_priority_override(struct amdgpu_device *adev,
 		return r;
 	}
 
-	idr_for_each_entry(&fpriv->ctx_mgr.ctx_handles, ctx, id)
+	mgr = &fpriv->ctx_mgr;
+	mutex_lock(&mgr->lock);
+	idr_for_each_entry(&mgr->ctx_handles, ctx, id)
 		amdgpu_ctx_priority_override(ctx, priority);
+	mutex_unlock(&mgr->lock);
 
 	fdput(f);
 	return 0;
-- 
2.40.1.495.gc816e09b53d-goog

Re: [PATCH v2] drm/amdgpu: add a missing lock for AMDGPU_SCHED

[Greg KH]

On Tue, Apr 25, 2023 at 11:17:14PM -0700, Chia-I Wu wrote:
> mgr->ctx_handles should be protected by mgr->lock.
> 
> v2: improve commit message
> 
> Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
> Cc: stable@vger.kernel.org

What commit id does this fix?  How far back in stable kernels should
this go?

thanks,

greg k-h

Re: [PATCH v2] drm/amdgpu: add a missing lock for AMDGPU_SCHED

[Christian König]

Am 26.04.23 um 08:17 schrieb Chia-I Wu:
> mgr->ctx_handles should be protected by mgr->lock.
>
> v2: improve commit message
>
> Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
> Cc: stable@vger.kernel.org

Please don't manually CC stable@vger.kernel.org while sending patches 
out, let us maintainers push that upstream with the appropriate tag and 
Greg picking it up from there.

A Fixes tag and figuring out to which stable versions this needs to be 
backported are nice to have as well, but Alex and I can take care of 
that as well.

Apart from that the technical side of the patch is Reviewed-by: 
Christian König <christian.koenig@amd.com>.

Regards,
Christian.

> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c | 6 +++++-
>   1 file changed, 5 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c
> index e9b45089a28a6..863b2a34b2d64 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c
> @@ -38,6 +38,7 @@ static int amdgpu_sched_process_priority_override(struct amdgpu_device *adev,
>   {
>   	struct fd f = fdget(fd);
>   	struct amdgpu_fpriv *fpriv;
> +	struct amdgpu_ctx_mgr *mgr;
>   	struct amdgpu_ctx *ctx;
>   	uint32_t id;
>   	int r;
> @@ -51,8 +52,11 @@ static int amdgpu_sched_process_priority_override(struct amdgpu_device *adev,
>   		return r;
>   	}
>   
> -	idr_for_each_entry(&fpriv->ctx_mgr.ctx_handles, ctx, id)
> +	mgr = &fpriv->ctx_mgr;
> +	mutex_lock(&mgr->lock);
> +	idr_for_each_entry(&mgr->ctx_handles, ctx, id)
>   		amdgpu_ctx_priority_override(ctx, priority);
> +	mutex_unlock(&mgr->lock);
>   
>   	fdput(f);
>   	return 0;

Re: [PATCH v2] drm/amdgpu: add a missing lock for AMDGPU_SCHED

[Chia-I Wu]

On Wed, Apr 26, 2023 at 4:05 AM Christian König
<christian.koenig@amd.com> wrote:
>
> Am 26.04.23 um 08:17 schrieb Chia-I Wu:
> > mgr->ctx_handles should be protected by mgr->lock.
> >
> > v2: improve commit message
> >
> > Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
> > Cc: stable@vger.kernel.org
>
> Please don't manually CC stable@vger.kernel.org while sending patches
> out, let us maintainers push that upstream with the appropriate tag and
> Greg picking it up from there.
>
> A Fixes tag and figuring out to which stable versions this needs to be
> backported are nice to have as well, but Alex and I can take care of
> that as well.
>
> Apart from that the technical side of the patch is Reviewed-by:
> Christian König <christian.koenig@amd.com>.
All done.  Thanks for clarifying the process and sorry for getting it
wrong in the first place :(
>
> Regards,
> Christian.
>
> > ---
> >   drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c | 6 +++++-
> >   1 file changed, 5 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c
> > index e9b45089a28a6..863b2a34b2d64 100644
> > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c
> > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c
> > @@ -38,6 +38,7 @@ static int amdgpu_sched_process_priority_override(struct amdgpu_device *adev,
> >   {
> >       struct fd f = fdget(fd);
> >       struct amdgpu_fpriv *fpriv;
> > +     struct amdgpu_ctx_mgr *mgr;
> >       struct amdgpu_ctx *ctx;
> >       uint32_t id;
> >       int r;
> > @@ -51,8 +52,11 @@ static int amdgpu_sched_process_priority_override(struct amdgpu_device *adev,
> >               return r;
> >       }
> >
> > -     idr_for_each_entry(&fpriv->ctx_mgr.ctx_handles, ctx, id)
> > +     mgr = &fpriv->ctx_mgr;
> > +     mutex_lock(&mgr->lock);
> > +     idr_for_each_entry(&mgr->ctx_handles, ctx, id)
> >               amdgpu_ctx_priority_override(ctx, priority);
> > +     mutex_unlock(&mgr->lock);
> >
> >       fdput(f);
> >       return 0;
>