From: Jordan Crouse <jordan@cosmicpenguin.net>
To: Rob Clark <robdclark@gmail.com>
Cc: Rob Clark <robdclark@chromium.org>,
"open list:DRM DRIVER FOR MSM ADRENO GPU"
<freedreno@lists.freedesktop.org>,
Sai Prakash Ranjan <saiprakash.ranjan@codeaurora.org>,
Jonathan Marek <jonathan@marek.ca>,
David Airlie <airlied@linux.ie>,
"open list:DRM DRIVER FOR MSM ADRENO GPU"
<linux-arm-msm@vger.kernel.org>,
Sharat Masetty <smasetty@codeaurora.org>,
Konrad Dybcio <konrad.dybcio@somainline.org>,
Akhil P Oommen <akhilpo@codeaurora.org>,
dri-devel@lists.freedesktop.org,
Douglas Anderson <dianders@chromium.org>,
"Kristian H. Kristensen" <hoegsberg@google.com>,
AngeloGioacchino Del Regno
<angelogioacchino.delregno@somainline.org>,
Marijn Suijten <marijn.suijten@somainline.org>,
Sean Paul <sean@poorly.run>,
open list <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 2/2] drm/msm: Periodically update RPTR shadow
Date: Sun, 2 May 2021 13:47:00 -0600 [thread overview]
Message-ID: <20210502194700.si5jysge2gftjdpw@cosmicpenguin.net> (raw)
In-Reply-To: <20210428193654.1498482-3-robdclark@gmail.com>
On Wed, Apr 28, 2021 at 12:36:49PM -0700, Rob Clark wrote:
> From: Rob Clark <robdclark@chromium.org>
>
> On a5xx and a6xx devices that are using CP_WHERE_AM_I to update a
> ringbuffer read-ptr shadow value, periodically emit a CP_WHERE_AM_I
> every 32 commands, so that a later submit waiting for ringbuffer
> space to become available sees partial progress, rather than not
> seeing rptr advance at all until the GPU gets to the end of the
> submit that it is currently chewing on.
Acked-by: Jordan Crouse <jordan@cosmicpenguin.net>
> Signed-off-by: Rob Clark <robdclark@chromium.org>
> ---
> drivers/gpu/drm/msm/adreno/a5xx_gpu.c | 29 ++++++++++++++++++++++-----
> drivers/gpu/drm/msm/adreno/a6xx_gpu.c | 27 +++++++++++++++++++------
> 2 files changed, 45 insertions(+), 11 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c
> index 0c8faad3b328..5202f1498a48 100644
> --- a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c
> +++ b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c
> @@ -18,6 +18,18 @@ static void a5xx_dump(struct msm_gpu *gpu);
>
> #define GPU_PAS_ID 13
>
> +static void update_shadow_rptr(struct msm_gpu *gpu, struct msm_ringbuffer *ring)
> +{
> + struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
> + struct a5xx_gpu *a5xx_gpu = to_a5xx_gpu(adreno_gpu);
> +
> + if (a5xx_gpu->has_whereami) {
> + OUT_PKT7(ring, CP_WHERE_AM_I, 2);
> + OUT_RING(ring, lower_32_bits(shadowptr(a5xx_gpu, ring)));
> + OUT_RING(ring, upper_32_bits(shadowptr(a5xx_gpu, ring)));
> + }
> +}
> +
> void a5xx_flush(struct msm_gpu *gpu, struct msm_ringbuffer *ring,
> bool sync)
> {
> @@ -30,11 +42,8 @@ void a5xx_flush(struct msm_gpu *gpu, struct msm_ringbuffer *ring,
> * Most flush operations need to issue a WHERE_AM_I opcode to sync up
> * the rptr shadow
> */
> - if (a5xx_gpu->has_whereami && sync) {
> - OUT_PKT7(ring, CP_WHERE_AM_I, 2);
> - OUT_RING(ring, lower_32_bits(shadowptr(a5xx_gpu, ring)));
> - OUT_RING(ring, upper_32_bits(shadowptr(a5xx_gpu, ring)));
> - }
> + if (sync)
> + update_shadow_rptr(gpu, ring);
>
> if (unlikely(ring->overflow))
> return;
> @@ -171,6 +180,16 @@ static void a5xx_submit(struct msm_gpu *gpu, struct msm_gem_submit *submit)
> ibs++;
> break;
> }
> +
> + /*
> + * Periodically update shadow-wptr if needed, so that we
> + * can see partial progress of submits with large # of
> + * cmds.. otherwise we could needlessly stall waiting for
> + * ringbuffer state, simply due to looking at a shadow
> + * rptr value that has not been updated
> + */
> + if ((ibs % 32) == 0)
> + update_shadow_rptr(gpu, ring);
> }
>
> /*
> diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
> index 4a4728a774c0..2986e36ffd8d 100644
> --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
> +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
> @@ -52,21 +52,25 @@ static bool a6xx_idle(struct msm_gpu *gpu, struct msm_ringbuffer *ring)
> return true;
> }
>
> -static void a6xx_flush(struct msm_gpu *gpu, struct msm_ringbuffer *ring)
> +static void update_shadow_rptr(struct msm_gpu *gpu, struct msm_ringbuffer *ring)
> {
> struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
> struct a6xx_gpu *a6xx_gpu = to_a6xx_gpu(adreno_gpu);
> - uint32_t wptr;
> - unsigned long flags;
>
> /* Expanded APRIV doesn't need to issue the WHERE_AM_I opcode */
> if (a6xx_gpu->has_whereami && !adreno_gpu->base.hw_apriv) {
> - struct a6xx_gpu *a6xx_gpu = to_a6xx_gpu(adreno_gpu);
> -
> OUT_PKT7(ring, CP_WHERE_AM_I, 2);
> OUT_RING(ring, lower_32_bits(shadowptr(a6xx_gpu, ring)));
> OUT_RING(ring, upper_32_bits(shadowptr(a6xx_gpu, ring)));
> }
> +}
> +
> +static void a6xx_flush(struct msm_gpu *gpu, struct msm_ringbuffer *ring)
> +{
> + uint32_t wptr;
> + unsigned long flags;
> +
> + update_shadow_rptr(gpu, ring);
>
> if (unlikely(ring->overflow))
> return;
> @@ -148,7 +152,7 @@ static void a6xx_submit(struct msm_gpu *gpu, struct msm_gem_submit *submit)
> struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
> struct a6xx_gpu *a6xx_gpu = to_a6xx_gpu(adreno_gpu);
> struct msm_ringbuffer *ring = submit->ring;
> - unsigned int i;
> + unsigned int i, ibs = 0;
>
> a6xx_set_pagetable(a6xx_gpu, ring, submit->queue->ctx);
>
> @@ -184,8 +188,19 @@ static void a6xx_submit(struct msm_gpu *gpu, struct msm_gem_submit *submit)
> OUT_RING(ring, lower_32_bits(submit->cmd[i].iova));
> OUT_RING(ring, upper_32_bits(submit->cmd[i].iova));
> OUT_RING(ring, submit->cmd[i].size);
> + ibs++;
> break;
> }
> +
> + /*
> + * Periodically update shadow-wptr if needed, so that we
> + * can see partial progress of submits with large # of
> + * cmds.. otherwise we could needlessly stall waiting for
> + * ringbuffer state, simply due to looking at a shadow
> + * rptr value that has not been updated
> + */
> + if ((ibs % 32) == 0)
> + update_shadow_rptr(gpu, ring);
> }
>
> get_stats_counter(ring, REG_A6XX_RBBM_PERFCTR_CP_0_LO,
> --
> 2.30.2
>
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
prev parent reply other threads:[~2021-05-02 19:47 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-04-28 19:36 [PATCH 0/2] drm/msm: Smooth out ringbuffer-full handling Rob Clark
2021-04-28 19:36 ` [PATCH 1/2] drm/msm: Handle ringbuffer overflow Rob Clark
2021-11-25 7:36 ` Dmitry Baryshkov
2021-11-25 17:20 ` Rob Clark
2021-04-28 19:36 ` [PATCH 2/2] drm/msm: Periodically update RPTR shadow Rob Clark
2021-05-02 19:47 ` Jordan Crouse [this message]
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=20210502194700.si5jysge2gftjdpw@cosmicpenguin.net \
--to=jordan@cosmicpenguin.net \
--cc=airlied@linux.ie \
--cc=akhilpo@codeaurora.org \
--cc=angelogioacchino.delregno@somainline.org \
--cc=dianders@chromium.org \
--cc=dri-devel@lists.freedesktop.org \
--cc=freedreno@lists.freedesktop.org \
--cc=hoegsberg@google.com \
--cc=jonathan@marek.ca \
--cc=konrad.dybcio@somainline.org \
--cc=linux-arm-msm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=marijn.suijten@somainline.org \
--cc=robdclark@chromium.org \
--cc=robdclark@gmail.com \
--cc=saiprakash.ranjan@codeaurora.org \
--cc=sean@poorly.run \
--cc=smasetty@codeaurora.org \
/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;
as well as URLs for NNTP newsgroup(s).