From: "Christian König" <christian.koenig@amd.com>
To: Lang Yu <Lang.Yu@amd.com>, amd-gfx@lists.freedesktop.org
Cc: Alex Deucher <alexander.deucher@amd.com>, Huang Rui <ray.huang@amd.com>
Subject: Re: [PATCH] drm/amdgpu: add workarounds for Raven VCN TMZ issue
Date: Fri, 4 Mar 2022 10:25:35 +0100 [thread overview]
Message-ID: <6edf5bbd-e498-5979-9449-a824706532d3@amd.com> (raw)
In-Reply-To: <20220304091555.4094582-1-Lang.Yu@amd.com>
Am 04.03.22 um 10:15 schrieb Lang Yu:
> It is a hardware issue that VCN can't handle a GTT
> backing stored TMZ buffer on Raven.
>
> Move such a TMZ buffer to VRAM domain before command
> submission.
Please don't touch the CS code with that. What you can do is to
implement a Raven specific patch_cs_in_place callback in the vcn_v*.c file.
See amdgpu_uvd_force_into_uvd_segment() how we handle the 256MiB window
on really old UVD blocks. You just don't need to implement the full
parsing of the command stream (yet).
Regards,
Christian.
>
> Signed-off-by: Lang Yu <Lang.Yu@amd.com>
> ---
> drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 27 ++++++++++++++++++++++++++
> 1 file changed, 27 insertions(+)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> index af12256e1bd3..66345f2ce6ba 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> @@ -393,6 +393,24 @@ void amdgpu_cs_report_moved_bytes(struct amdgpu_device *adev, u64 num_bytes,
> spin_unlock(&adev->mm_stats.lock);
> }
>
> +static int raven_vcn_tmz_quirks(struct amdgpu_cs_parser *p, struct amdgpu_bo *bo, uint32_t *domain)
> +{
> + struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
> + struct amdgpu_ring *ring = to_amdgpu_ring(p->entity->rq->sched);
> +
> + if ((adev->asic_type == CHIP_RAVEN) &&
> + (bo->flags & AMDGPU_GEM_CREATE_ENCRYPTED) &&
> + (ring->funcs->type == AMDGPU_HW_IP_VCN_DEC ||
> + ring->funcs->type == AMDGPU_HW_IP_VCN_ENC ||
> + ring->funcs->type == AMDGPU_HW_IP_VCN_JPEG)) {
> + if (domain)
> + *domain = AMDGPU_GEM_DOMAIN_VRAM;
> + return 1;
> + }
> +
> + return 0;
> +}
> +
> static int amdgpu_cs_bo_validate(void *param, struct amdgpu_bo *bo)
> {
> struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
> @@ -403,6 +421,7 @@ static int amdgpu_cs_bo_validate(void *param, struct amdgpu_bo *bo)
> .resv = bo->tbo.base.resv
> };
> uint32_t domain;
> + bool need_retry = 1;
> int r;
>
> if (bo->tbo.pin_count)
> @@ -431,6 +450,8 @@ static int amdgpu_cs_bo_validate(void *param, struct amdgpu_bo *bo)
> domain = bo->allowed_domains;
> }
>
> + raven_vcn_tmz_quirks(p, bo, &domain);
> +
> retry:
> amdgpu_bo_placement_from_domain(bo, domain);
> r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
> @@ -442,6 +463,12 @@ static int amdgpu_cs_bo_validate(void *param, struct amdgpu_bo *bo)
>
> if (unlikely(r == -ENOMEM) && domain != bo->allowed_domains) {
> domain = bo->allowed_domains;
> + if (raven_vcn_tmz_quirks(p, bo, &domain)) {
> + if (need_retry)
> + need_retry = 0;
> + else
> + return r;
> + }
> goto retry;
> }
>
next prev parent reply other threads:[~2022-03-04 9:25 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-03-04 9:15 [PATCH] drm/amdgpu: add workarounds for Raven VCN TMZ issue Lang Yu
2022-03-04 9:25 ` Christian König [this message]
2022-03-04 9:38 ` Lang Yu
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=6edf5bbd-e498-5979-9449-a824706532d3@amd.com \
--to=christian.koenig@amd.com \
--cc=Lang.Yu@amd.com \
--cc=alexander.deucher@amd.com \
--cc=amd-gfx@lists.freedesktop.org \
--cc=ray.huang@amd.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox