[PATCH] drm/amdgpu: Replace HQD terminology with slots naming
Alex Deucher
alexdeucher at gmail.com
Mon Jul 7 13:22:40 UTC 2025
On Mon, Jul 7, 2025 at 5:48 AM Christian König <christian.koenig at amd.com> wrote:
>
> On 04.07.25 09:26, Jesse Zhang wrote:
> > The term "HQD" is CP-specific and doesn't
> > accurately describe the queue resources for other IP blocks like SDMA,
> > VCN, or VPE. This change:
> >
> > 1. Renames `num_hqds` to `num_slots` in amdgpu_kms.c to better reflect
> > the generic nature of the resource counting
> > 2. Updates the UAPI struct member from `userq_num_hqds` to `userq_num_slots`
> > 3. Maintains the same functionality while using more appropriate terminology
> >
> > Signed-off-by: Jesse Zhang <Jesse.Zhang at amd.com>
>
> Acked-by: Christian König <christian.koenig at amd.com>
>
> BTW: Why userspace needs to know that in the first place?
So IGT can determine how many queues to create to hit over subscription.
Alex
>
> > ---
> > drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c | 10 +++++-----
> > include/uapi/drm/amdgpu_drm.h | 4 ++--
> > 2 files changed, 7 insertions(+), 7 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
> > index 4aab5e394ce2..e2aa2264fa0b 100644
> > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
> > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
> > @@ -399,7 +399,7 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev,
> > uint32_t ib_size_alignment = 0;
> > enum amd_ip_block_type type;
> > unsigned int num_rings = 0;
> > - uint32_t num_hqds = 0;
> > + uint32_t num_slots = 0;
> > unsigned int i, j;
> >
> > if (info->query_hw_ip.ip_instance >= AMDGPU_HW_IP_INSTANCE_MAX_COUNT)
> > @@ -415,7 +415,7 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev,
> >
> > if (!adev->gfx.disable_uq) {
> > for (i = 0; i < AMDGPU_MES_MAX_GFX_PIPES; i++)
> > - num_hqds += hweight32(adev->mes.gfx_hqd_mask[i]);
> > + num_slots += hweight32(adev->mes.gfx_hqd_mask[i]);
> > }
> >
> > ib_start_alignment = 32;
> > @@ -430,7 +430,7 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev,
> >
> > if (!adev->sdma.disable_uq) {
> > for (i = 0; i < AMDGPU_MES_MAX_COMPUTE_PIPES; i++)
> > - num_hqds += hweight32(adev->mes.compute_hqd_mask[i]);
> > + num_slots += hweight32(adev->mes.compute_hqd_mask[i]);
> > }
> >
> > ib_start_alignment = 32;
> > @@ -445,7 +445,7 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev,
> >
> > if (!adev->gfx.disable_uq) {
> > for (i = 0; i < AMDGPU_MES_MAX_SDMA_PIPES; i++)
> > - num_hqds += hweight32(adev->mes.sdma_hqd_mask[i]);
> > + num_slots += hweight32(adev->mes.sdma_hqd_mask[i]);
> > }
> >
> > ib_start_alignment = 256;
> > @@ -589,7 +589,7 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev,
> > }
> > result->capabilities_flags = 0;
> > result->available_rings = (1 << num_rings) - 1;
> > - result->userq_num_hqds = num_hqds;
> > + result->userq_num_slots = num_slots;
> > result->ib_start_alignment = ib_start_alignment;
> > result->ib_size_alignment = ib_size_alignment;
> > return 0;
> > diff --git a/include/uapi/drm/amdgpu_drm.h b/include/uapi/drm/amdgpu_drm.h
> > index 66c4a03ac9f9..bdedbaccf776 100644
> > --- a/include/uapi/drm/amdgpu_drm.h
> > +++ b/include/uapi/drm/amdgpu_drm.h
> > @@ -1493,8 +1493,8 @@ struct drm_amdgpu_info_hw_ip {
> > __u32 available_rings;
> > /** version info: bits 23:16 major, 15:8 minor, 7:0 revision */
> > __u32 ip_discovery_version;
> > - /* Userq available hqds */
> > - __u32 userq_num_hqds;
> > + /* Userq available slots */
> > + __u32 userq_num_slots;
> > };
> >
> > /* GFX metadata BO sizes and alignment info (in bytes) */
>
More information about the amd-gfx
mailing list