<html><head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
  </head>
  <body>
    <p>I didn't check the patch details, if it is for existing implicit
      sync of shared buffer, feel free go ahead.</p>
    <p>But if you add some description for its usage, that will be more
      clear to others.</p>
    <p>-David<br>
    </p>
    <div class="moz-cite-prefix">在 2020/6/11 15:19, Marek Olšák 写道:<br>
    </div>
    <blockquote type="cite" cite="mid:CAAxE2A4JsmQL8kUA1Z0aq3sK86D3oJ54YkRe=yEo1AXJ1MT-yg@mail.gmail.com">
      
      <div dir="auto">Hi David,
        <div dir="auto"><br>
        </div>
        <div dir="auto">Explicit sync has nothing to do with this. This
          is for implicit sync, which is required by DRI3. This fix
          allows removing existing inefficiencies from drivers, so it's
          a good thing.</div>
        <div dir="auto"><br>
        </div>
        <div dir="auto">Marek</div>
      </div>
      <br>
      <div class="gmail_quote">
        <div dir="ltr" class="gmail_attr">On Wed., Jun. 10, 2020, 03:56
          Chunming Zhou, <<a href="mailto:zhoucm1@amd.com" moz-do-not-send="true">zhoucm1@amd.com</a>> wrote:<br>
        </div>
        <blockquote class="gmail_quote" style="margin:0 0 0
          .8ex;border-left:1px #ccc solid;padding-left:1ex">
          <div>
            <p><br>
            </p>
            <div>在 2020/6/10 15:41, Christian König 写道:<br>
            </div>
            <blockquote type="cite">
              <div>That's true, but for now we are stuck with the
                implicit sync for quite a number of use cases.<br>
                <br>
                My problem is rather that we already tried this and it
                backfired immediately.<br>
                <br>
                I do remember that it was your patch who introduced the
                pipeline sync flag handling and I warned that this could
                be problematic. You then came back with a QA result
                saying that this is indeed causing a huge performance
                drop in one test case and we need to do something else.
                Together we then came up with the different handling
                between implicit and explicit sync.<br>
              </div>
            </blockquote>
            <p>Isn't pipeline sync flag to fix some issue because of
              parralel execution between jobs in one pipeline?  I really
              don't have this memory in mind why that's realted to this,
              Or do you mean extra sync hides many other potential
              issues?</p>
            <p>Anyway, when I go through Vulkan WSI code, the
              synchronization isn't so smooth between OS window system.
              And when I saw Jason drives explicit sync through the
              whole Linux ecosystem like Android window system does, I
              feel that's really a good direction.</p>
            <p>-David<br>
            </p>
            <blockquote type="cite">
              <div> <br>
                But I can't find that stupid mail thread any more. I
                knew that it was a couple of years ago when we started
                with the explicit sync for Vulkan.<br>
                <br>
                Christian.<br>
                <br>
                Am 10.06.20 um 08:29 schrieb Zhou, David(ChunMing):<br>
              </div>
              <blockquote type="cite">
                <div>
                  <p style="margin:0in;margin-bottom:.0001pt"><span style="font-size:10.0pt;font-family:"Arial",sans-serif;color:#0078d7">[AMD
                      Official Use Only - Internal Distribution Only]</span></p>
                  <p class="MsoNormal"> </p>
                  <p class="MsoNormal">Not sue if this is right
                    direction, I think usermode wants all
                    synchronizations to be explicit. Implicit sync often
                    confuses people who don’t know its history. I
                    remember Jason from Intel  is driving explicit
                    synchronization through the Linux ecosystem, which
                    even removes implicit sync of shared buffer.</p>
                  <p class="MsoNormal"> </p>
                  <p class="MsoNormal">-David</p>
                  <p class="MsoNormal"> </p>
                  <div style="border:none;border-top:solid #e1e1e1
                    1.0pt;padding:3.0pt 0in 0in 0in">
                    <p class="MsoNormal"><b>From:</b> amd-gfx <a href="mailto:amd-gfx-bounces@lists.freedesktop.org" target="_blank" rel="noreferrer" moz-do-not-send="true"><amd-gfx-bounces@lists.freedesktop.org></a>
                      <b>On Behalf Of </b>Marek Olšák<br>
                      <b>Sent:</b> Tuesday, June 9, 2020 6:58 PM<br>
                      <b>To:</b> amd-gfx mailing list <a href="mailto:amd-gfx@lists.freedesktop.org" target="_blank" rel="noreferrer" moz-do-not-send="true"><amd-gfx@lists.freedesktop.org></a><br>
                      <b>Subject:</b> [PATCH] drm/amdgpu: remove
                      distinction between explicit and implicit sync
                      (v2)</p>
                  </div>
                  <p class="MsoNormal"> </p>
                  <div>
                    <div>
                      <p class="MsoNormal">Hi,</p>
                    </div>
                    <div>
                      <p class="MsoNormal"> </p>
                    </div>
                    <div>
                      <p class="MsoNormal">This enables a full pipeline
                        sync for implicit sync. It's Christian's patch
                        with the driver version bumped. With this, user
                        mode drivers don't have to wait for idle at the
                        end of gfx IBs.</p>
                    </div>
                    <div>
                      <p class="MsoNormal"> </p>
                    </div>
                    <div>
                      <p class="MsoNormal">Any concerns?</p>
                    </div>
                    <div>
                      <p class="MsoNormal"> </p>
                    </div>
                    <div>
                      <p class="MsoNormal">Thanks,</p>
                    </div>
                    <div>
                      <p class="MsoNormal">Marek</p>
                    </div>
                  </div>
                </div>
                <br>
                <fieldset></fieldset>
                <pre>_______________________________________________
amd-gfx mailing list
<a href="mailto:amd-gfx@lists.freedesktop.org" target="_blank" rel="noreferrer" moz-do-not-send="true">amd-gfx@lists.freedesktop.org</a>
<a href="https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.freedesktop.org%2Fmailman%2Flistinfo%2Famd-gfx&data=02%7C01%7CDavid1.Zhou%40amd.com%7C0d3096fc043f4443f14e08d80dd7c674%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637274567683552668&sdata=xIHDswGRsdCP%2BE7MRI4nKXdoMgV2LBzFPP46zGpQusk%3D&reserved=0" originalsrc="https://lists.freedesktop.org/mailman/listinfo/amd-gfx" shash="hAGAyoS6auIG5z/ymuBRZ8NQ1FyVyzxjLTP8d56vUWHt3KOGxnZA9JR3arWRHVNGShrJPcF8kl5FVDL8/CaUnIELQRxp+YA+LqJpUCIUiFkJQUnkS8xLZ4pFCx17bcOL632sj4r1sETdA7zJMsD8dR7TUlzPxzmmzFhKnYvV+/w=" target="_blank" rel="noreferrer" moz-do-not-send="true">https://lists.freedesktop.org/mailman/listinfo/amd-gfx</a>
</pre>
              </blockquote>
              <br>
            </blockquote>
          </div>
        </blockquote>
      </div>
    </blockquote>
  </body>
</html>