[PATCH 3/5] drm/amdkfd: use map_queues for hiq on arcturus as well
Huang Rui
ray.huang at amd.com
Mon Jan 13 02:38:20 UTC 2020
On Sat, Jan 11, 2020 at 07:05:40AM +0800, Kuehling, Felix wrote:
> What happens on Arcturus without this patch? Does it oops with a null
> pointer dereference? If yes, then you should squash this patch into
> patch 2 to avoid a broken intermediate state.
Yes, Arcturus will get a null pointer panic without this patch becasue
hiq_mqd_load is not inited in arcturus_kfd2kgd. I will squash this patch
into the patch 2.
Thanks,
Ray
>
> Regards,
> Felix
>
> On 2020-01-10 1:37 a.m., Huang Rui wrote:
> > Align with gfx v9, use the map_queues packet to load hiq MQD.
> >
> > Signed-off-by: Huang Rui <ray.huang at amd.com>
> > ---
> > drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c | 1 +
> > drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c | 6 +++---
> > drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h | 3 +++
> > 3 files changed, 7 insertions(+), 3 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c
> > index 3c11940..8baad42 100644
> > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c
> > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c
> > @@ -281,6 +281,7 @@ const struct kfd2kgd_calls arcturus_kfd2kgd = {
> > .set_pasid_vmid_mapping = kgd_gfx_v9_set_pasid_vmid_mapping,
> > .init_interrupts = kgd_gfx_v9_init_interrupts,
> > .hqd_load = kgd_gfx_v9_hqd_load,
> > + .hiq_mqd_load = kgd_gfx_v9_hiq_mqd_load,
> > .hqd_sdma_load = kgd_hqd_sdma_load,
> > .hqd_dump = kgd_gfx_v9_hqd_dump,
> > .hqd_sdma_dump = kgd_hqd_sdma_dump,
> > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
> > index ab8c23a..d2f9396 100644
> > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
> > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
> > @@ -324,9 +324,9 @@ int kgd_gfx_v9_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
> > return 0;
> > }
> >
> > -static int kgd_gfx_v9_hiq_mqd_load(struct kgd_dev *kgd, void *mqd,
> > - uint32_t pipe_id, uint32_t queue_id,
> > - uint32_t doorbell_off)
> > +int kgd_gfx_v9_hiq_mqd_load(struct kgd_dev *kgd, void *mqd,
> > + uint32_t pipe_id, uint32_t queue_id,
> > + uint32_t doorbell_off)
> > {
> > struct amdgpu_device *adev = get_amdgpu_device(kgd);
> > struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring;
> > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h
> > index 02b1426..32dd1a9 100644
> > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h
> > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h
> > @@ -33,6 +33,9 @@ int kgd_gfx_v9_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
> > uint32_t queue_id, uint32_t __user *wptr,
> > uint32_t wptr_shift, uint32_t wptr_mask,
> > struct mm_struct *mm);
> > +int kgd_gfx_v9_hiq_mqd_load(struct kgd_dev *kgd, void *mqd,
> > + uint32_t pipe_id, uint32_t queue_id,
> > + uint32_t doorbell_off);
> > int kgd_gfx_v9_hqd_dump(struct kgd_dev *kgd,
> > uint32_t pipe_id, uint32_t queue_id,
> > uint32_t (**dump)[2], uint32_t *n_regs);
More information about the amd-gfx
mailing list