<html>
  <head>
    <meta content="text/html; charset=utf-8" http-equiv="Content-Type">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    <div class="moz-cite-prefix">Am 15.09.2016 um 23:07 schrieb Julien
      Isorce:<br>
    </div>
    <blockquote
cite="mid:CAHWPjbVxGe6dthMx2Kx27j0ke2zLQhz+i-vRew_WUgRXemN3gg@mail.gmail.com"
      type="cite">
      <div dir="ltr"><br>
        <div class="gmail_extra"><br>
          <div class="gmail_quote">On 15 September 2016 at 16:02, Leo
            Liu <span dir="ltr"><<a moz-do-not-send="true"
                href="mailto:leo.liu@amd.com" target="_blank">leo.liu@amd.com</a>></span>
            wrote:<br>
            <blockquote class="gmail_quote" style="margin:0px 0px 0px
              0.8ex;border-left:1px solid
              rgb(204,204,204);padding-left:1ex"><span class="gmail-"><br>
                <br>
                On 09/15/2016 10:43 AM, Andy Furniss wrote:<br>
                <blockquote class="gmail_quote" style="margin:0px 0px
                  0px 0.8ex;border-left:1px solid
                  rgb(204,204,204);padding-left:1ex">
                  Since bf901a2<br>
                  st/va: also honors interlaced preference when
                  providing a video format<br>
                  existing scripts and most use cases will need true.<br>
                  <br>
                  Signed-off-by: Andy Furniss <<a
                    moz-do-not-send="true"
                    href="mailto:adf.lists@gmail.com" target="_blank">adf.lists@gmail.com</a>><br>
                  ---<br>
                   src/gallium/state_trackers/<wbr>va/surface.c | 2 +-<br>
                   1 file changed, 1 insertion(+), 1 deletion(-)<br>
                  <br>
                  diff --git a/src/gallium/state_trackers/v<wbr>a/surface.c
                  b/src/gallium/state_trackers/v<wbr>a/surface.c<br>
                  index 00df69d..e73e17e 100644<br>
                  --- a/src/gallium/state_trackers/v<wbr>a/surface.c<br>
                  +++ b/src/gallium/state_trackers/v<wbr>a/surface.c<br>
                  @@ -43,7 +43,7 @@<br>
                  <br>
                   #include "va_private.h"<br>
                  <br>
                  -DEBUG_GET_ONCE_BOOL_OPTION(no<wbr>interlace,
                  "VAAPI_DISABLE_INTERLACE", FALSE);<br>
                  +DEBUG_GET_ONCE_BOOL_OPTION(no<wbr>interlace,
                  "VAAPI_DISABLE_INTERLACE", TRUE);<br>
                </blockquote>
                <br>
              </span>
              Like being mentioned,  It'll still override the preferred
              interlaced format when this env is not explicitly used.<br>
              <br>
              Not sure this will be okay with other case. @Julien?<br>
            </blockquote>
            <div><br>
            </div>
            <div>Hi,<br>
              <br>
            </div>
            <div>So the problem is that with radeon,
              PIPE_VIDEO_CAP_SUPPORTS_INTERLACED always returns false
              for encoding<br>
            </div>
            <div>but can return true for decoding (depending on the
              chipset and the codec). Then when doing transcoding you
              need all to be non interlaced to avoid extra
              copy/conversion and even a clash. Is it correct ?<br>
            </div>
          </div>
        </div>
      </div>
    </blockquote>
    <br>
    Yes correct. The decoder supports both interlaced as well as
    progressive memory layout, but the encoder can only handle
    progressive input.<br>
    <br>
    Interlaced memory layout is needed for things like adaptive
    deinterlacing as well as VDPAU OpenGL interop.<br>
    <br>
    <blockquote
cite="mid:CAHWPjbVxGe6dthMx2Kx27j0ke2zLQhz+i-vRew_WUgRXemN3gg@mail.gmail.com"
      type="cite">
      <div dir="ltr">
        <div class="gmail_extra">
          <div class="gmail_quote">
            <div><br>
            </div>
            <div>Should debug_get_option_nointerlace() be moved to
              radeon_video.c::rvid_get_video_param ?<br>
            </div>
          </div>
        </div>
      </div>
    </blockquote>
    <br>
    For the short term that sounds like a good idea to me.<br>
    <br>
    In the mid term we need to better handle this case. E.g. allocate
    the video buffer with the layout the driver reports as preferred, if
    that doesn't match the use case (transcoding, deinterlacing or
    interop) convert as needed.<br>
    <br>
    <blockquote
cite="mid:CAHWPjbVxGe6dthMx2Kx27j0ke2zLQhz+i-vRew_WUgRXemN3gg@mail.gmail.com"
      type="cite">
      <div dir="ltr">
        <div class="gmail_extra">
          <div class="gmail_quote">
            <div><br>
            </div>
            <div>Other question:<br>
              <br>
                  case PIPE_VIDEO_CAP_SUPPORTS_INTERLACED:<br>
                      if (rscreen->family < CHIP_PALM) {<br>
                          /* MPEG2 only with shaders and no support for<br>
                             interlacing on R6xx style UVD */<br>
                          return codec != PIPE_VIDEO_FORMAT_MPEG12
              &&<br>
                                 rscreen->family > CHIP_RV770;<br>
                      } else {<br>
                          if (u_reduce_video_profile(profile) ==
              PIPE_VIDEO_FORMAT_HEVC)<br>
                              return false; //The firmware doesn't
              support interlaced HEVC.<br>
                          return true;<br>
                      }<br>
            </div>
            <div>So if instead it would always return false then it will
              really work on hardware for which above code says true ?<br>
            </div>
          </div>
        </div>
      </div>
    </blockquote>
    <br>
    It would work, but the deinterlacing and interop use case I noted
    above would break.<br>
    <br>
    <blockquote
cite="mid:CAHWPjbVxGe6dthMx2Kx27j0ke2zLQhz+i-vRew_WUgRXemN3gg@mail.gmail.com"
      type="cite">
      <div dir="ltr">
        <div class="gmail_extra">
          <div class="gmail_quote">
            <div>Because my understanding is that for nvidia hardware
              this is not a preference but rather a requirement but I
              might be wrong.<br>
            </div>
          </div>
        </div>
      </div>
    </blockquote>
    <br>
    Yes, as far as I know that is correct. AMD hardware can handle both
    for most hardware/profile combinations, even the HEVC limit only
    applies to a certain firmware version IIRC.<br>
    <br>
    <blockquote
cite="mid:CAHWPjbVxGe6dthMx2Kx27j0ke2zLQhz+i-vRew_WUgRXemN3gg@mail.gmail.com"
      type="cite">
      <div dir="ltr">
        <div class="gmail_extra">
          <div class="gmail_quote">
            <div><br>
            </div>
            <div>In any case, with current upstream code and <span
                class="gmail-">VAAPI_DISABLE_INTERLACE=1 it hits "</span>assert(templat->interlaced);"
              in nouveau_vp3_video_buffer_create. If I remove the asset
              it crashes and can even stall the driver (just wanted to
              check ).<br>
              <br>
            </div>
          </div>
        </div>
      </div>
    </blockquote>
    <br>
    Crap, feel free to revert the patch setting it to true by default.<br>
    <br>
    Regards,<br>
    Christian.<br>
    <br>
    <blockquote
cite="mid:CAHWPjbVxGe6dthMx2Kx27j0ke2zLQhz+i-vRew_WUgRXemN3gg@mail.gmail.com"
      type="cite">
      <div dir="ltr">
        <div class="gmail_extra">
          <div class="gmail_quote">
            <div>Cheers<br>
            </div>
            <div>Julien<br>
            </div>
            <div><br>
            </div>
            <blockquote class="gmail_quote" style="margin:0px 0px 0px
              0.8ex;border-left:1px solid
              rgb(204,204,204);padding-left:1ex">
              <br>
              Regards,<br>
              Leo<br>
              <br>
              <blockquote class="gmail_quote" style="margin:0px 0px 0px
                0.8ex;border-left:1px solid
                rgb(204,204,204);padding-left:1ex">
                <br>
                 #include <va/va_drmcommon.h><br>
                <br>
              </blockquote>
              <br>
            </blockquote>
          </div>
          <br>
        </div>
      </div>
      <br>
      <fieldset class="mimeAttachmentHeader"></fieldset>
      <br>
      <pre wrap="">_______________________________________________
mesa-dev mailing list
<a class="moz-txt-link-abbreviated" href="mailto:mesa-dev@lists.freedesktop.org">mesa-dev@lists.freedesktop.org</a>
<a class="moz-txt-link-freetext" href="https://lists.freedesktop.org/mailman/listinfo/mesa-dev">https://lists.freedesktop.org/mailman/listinfo/mesa-dev</a>
</pre>
    </blockquote>
    <p><br>
    </p>
  </body>
</html>