<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<style type="text/css" style="display:none;"> P {margin-top:0;margin-bottom:0;} </style>
</head>
<body dir="ltr">
<p style="font-family:Arial;font-size:10pt;color:#0000FF;margin:5pt;" align="Left">
[AMD Official Use Only]<br>
</p>
<br>
<div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Hi Christian,</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Since we have strict check on queue status, I don't think original design can cause issue here.<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
But this change should help improve below case:</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<ol>
<li><span>both enc thread and dec thread try to start begin_use.</span></li><li><span>dec thread gets the chance to finish begin_use process first.<br>
</span></li><li><span>before dec thread enters end_use, enc thread gets time slot to run through begin_use(No delay work scheduled at that time)</span></li><li><span><span>dec thread enters end_use</span>, scheduled a delay work<br>
</span></li><li><span><span> enc thread enters end_use</span>, modify this delay work.</span></li></ol>
<div>It will help reduce one delay work call at least. <span style="font-family:Calibri, Arial, Helvetica, sans-serif;font-size:12pt;color:rgb(0, 0, 0)">
<br>
</span></div>
<div><span style="font-family:Calibri, Arial, Helvetica, sans-serif;font-size:12pt;color:rgb(0, 0, 0)"><br>
</span></div>
</div>
<div>
<div id="Signature">
<div>
<div id="divtagdefaultwrapper" dir="ltr" style="font-size:12pt; color:#000000; font-family:Calibri,Arial,Helvetica,sans-serif">
<p style="margin-top: 0px; margin-bottom: 0px;">Thanks & Best Regards!</p>
<p style="margin-top: 0px; margin-bottom: 0px;"><br>
</p>
<p style="margin-top: 0px; margin-bottom: 0px;">James Zhu<br>
</p>
</div>
</div>
</div>
</div>
<div id="appendonsend"></div>
<hr style="display:inline-block;width:98%" tabindex="-1">
<div id="divRplyFwdMsg" dir="ltr"><font face="Calibri, sans-serif" style="font-size:11pt" color="#000000"><b>From:</b> Koenig, Christian <Christian.Koenig@amd.com><br>
<b>Sent:</b> Wednesday, August 11, 2021 5:34 PM<br>
<b>To:</b> Michel Dänzer <michel@daenzer.net>; Deucher, Alexander <Alexander.Deucher@amd.com><br>
<b>Cc:</b> Liu, Leo <Leo.Liu@amd.com>; Zhu, James <James.Zhu@amd.com>; amd-gfx@lists.freedesktop.org <amd-gfx@lists.freedesktop.org>; dri-devel@lists.freedesktop.org <dri-devel@lists.freedesktop.org><br>
<b>Subject:</b> AW: [PATCH 2/2] drm/amdgpu: Use mod_delayed_work in JPEG/UVD/VCE/VCN ring_end_use hooks</font>
<div> </div>
</div>
<div>
<div>NAK to at least this patch.</div>
<div><br>
</div>
<div>Since activating power management while submitting work is problematic cancel_delayed_work() must have been called during begin use or otherwise we have a serious coding problem in the first place.</div>
<div><br>
</div>
<div>So this change shouldn't make a difference and I suggest to really stick with schedule_delayed_work().</div>
<div><br>
</div>
<div>Maybe add a comment how this works?</div>
<div><br>
</div>
<div>Need to take a closer look at the first patch when I'm back from vacation, but it could be that this applies there as well.</div>
<div><br>
</div>
<div>Regards,</div>
<div>Christian.</div>
<div><br>
</div>
<hr tabindex="-1" style="display:inline-block; width:98%">
<div id="x_divRplyFwdMsg" dir="ltr"><font face="Calibri, sans-serif" color="#000000" style="font-size:11pt"><b>Von:</b> Michel Dänzer <michel@daenzer.net><br>
<b>Gesendet:</b> Mittwoch, 11. August 2021 18:52<br>
<b>An:</b> Deucher, Alexander <Alexander.Deucher@amd.com>; Koenig, Christian <Christian.Koenig@amd.com><br>
<b>Cc:</b> Liu, Leo <Leo.Liu@amd.com>; Zhu, James <James.Zhu@amd.com>; amd-gfx@lists.freedesktop.org <amd-gfx@lists.freedesktop.org>; dri-devel@lists.freedesktop.org <dri-devel@lists.freedesktop.org><br>
<b>Betreff:</b> [PATCH 2/2] drm/amdgpu: Use mod_delayed_work in JPEG/UVD/VCE/VCN ring_end_use hooks</font>
<div> </div>
</div>
<div class="x_BodyFragment"><font size="2"><span style="font-size:11pt">
<div class="x_PlainText">From: Michel Dänzer <mdaenzer@redhat.com><br>
<br>
In contrast to schedule_delayed_work, this pushes back the work if it<br>
was already scheduled before. Specific behaviour change:<br>
<br>
Before:<br>
<br>
The scheduled work ran ~1 second after the first time ring_end_use was<br>
called, even if the ring was used again during that second.<br>
<br>
After:<br>
<br>
The scheduled work runs ~1 second after the last time ring_end_use is<br>
called.<br>
<br>
Inspired by the corresponding change in amdgpu_gfx_off_ctrl. While I<br>
haven't run into specific issues in this case, the new behaviour makes<br>
more sense to me.<br>
<br>
Signed-off-by: Michel Dänzer <mdaenzer@redhat.com><br>
---<br>
drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c | 2 +-<br>
drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c | 2 +-<br>
drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c | 2 +-<br>
drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c | 2 +-<br>
4 files changed, 4 insertions(+), 4 deletions(-)<br>
<br>
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c<br>
index 8996cb4ed57a..2c0040153f6c 100644<br>
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c<br>
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c<br>
@@ -110,7 +110,7 @@ void amdgpu_jpeg_ring_begin_use(struct amdgpu_ring *ring)<br>
void amdgpu_jpeg_ring_end_use(struct amdgpu_ring *ring)<br>
{<br>
atomic_dec(&ring->adev->jpeg.total_submission_cnt);<br>
- schedule_delayed_work(&ring->adev->jpeg.idle_work, JPEG_IDLE_TIMEOUT);<br>
+ mod_delayed_work(system_wq, &ring->adev->jpeg.idle_work, JPEG_IDLE_TIMEOUT);<br>
}<br>
<br>
int amdgpu_jpeg_dec_ring_test_ring(struct amdgpu_ring *ring)<br>
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c<br>
index 0f576f294d8a..b6b1d7eeb8e5 100644<br>
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c<br>
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c<br>
@@ -1283,7 +1283,7 @@ void amdgpu_uvd_ring_begin_use(struct amdgpu_ring *ring)<br>
void amdgpu_uvd_ring_end_use(struct amdgpu_ring *ring)<br>
{<br>
if (!amdgpu_sriov_vf(ring->adev))<br>
- schedule_delayed_work(&ring->adev->uvd.idle_work, UVD_IDLE_TIMEOUT);<br>
+ mod_delayed_work(system_wq, &ring->adev->uvd.idle_work, UVD_IDLE_TIMEOUT);<br>
}<br>
<br>
/**<br>
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c<br>
index 1ae7f824adc7..2253c18a6688 100644<br>
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c<br>
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c<br>
@@ -401,7 +401,7 @@ void amdgpu_vce_ring_begin_use(struct amdgpu_ring *ring)<br>
void amdgpu_vce_ring_end_use(struct amdgpu_ring *ring)<br>
{<br>
if (!amdgpu_sriov_vf(ring->adev))<br>
- schedule_delayed_work(&ring->adev->vce.idle_work, VCE_IDLE_TIMEOUT);<br>
+ mod_delayed_work(system_wq, &ring->adev->vce.idle_work, VCE_IDLE_TIMEOUT);<br>
}<br>
<br>
/**<br>
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c<br>
index 284bb42d6c86..d5937ab5ac80 100644<br>
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c<br>
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c<br>
@@ -1874,7 +1874,7 @@ void vcn_v1_0_set_pg_for_begin_use(struct amdgpu_ring *ring, bool set_clocks)<br>
<br>
void vcn_v1_0_ring_end_use(struct amdgpu_ring *ring)<br>
{<br>
- schedule_delayed_work(&ring->adev->vcn.idle_work, VCN_IDLE_TIMEOUT);<br>
+ mod_delayed_work(system_wq, &ring->adev->vcn.idle_work, VCN_IDLE_TIMEOUT);<br>
mutex_unlock(&ring->adev->vcn.vcn1_jpeg1_workaround);<br>
}<br>
<br>
-- <br>
2.32.0<br>
<br>
</div>
</span></font></div>
</div>
</div>
</body>
</html>