<div dir="ltr"><div>Hi Peter,<br><br>It looks as though this patch fixes your error and the memory leak as well.  It is against gst-plugins-bad-1.1.2.<br><br>--- gst-plugins-bad-1.1.2/sys/uvch264/gstuvch264_mjpgdemux.c    2013-04-23 08:38:28.000000000 -0400<br>
+++ gst-plugins-bad-1.1.2.new/sys/uvch264/gstuvch264_mjpgdemux.c    2013-07-26 14:22:51.177762515 -0400<br>@@ -711,6 +711,8 @@<br>   if (jpeg_buf)<br>     gst_buffer_unref (jpeg_buf);<br> <br>+  gst_buffer_unmap(buf, &info);<br>
+<br>   /* We must always unref the input buffer since we never push it out */<br>   gst_buffer_unref (buf);<br><br></div>Best Regards,<br><br>Rob Krakora<br><br></div><div class="gmail_extra"><br><br><div class="gmail_quote">
On Fri, Jul 26, 2013 at 12:53 PM, Robert Krakora <span dir="ltr"><<a href="mailto:rob.krakora@messagenetsystems.com" target="_blank">rob.krakora@messagenetsystems.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div dir="ltr">Making the change I just posted and no other change I can run indefinitely (no error parsing JPEG container) although there still seems to be a very, very slow memory leak in uvch264src...NO CHANGES were made to v4l2src at all to add back "always-copy"...<br>

<br>--- gst-plugins-bad-1.1.2/sys/<div class="im"><div>uvch264/gstuvch264_mjpgdemux.c    2013-04-23 08:38:28.000000000 -0400<br></div>+++ gst-plugins-bad-1.1.2.new/sys/<div>uvch264/gstuvch264_mjpgdemux.c    2013-07-26 12:20:52.625791724 -0400<br>


@@ -711,6 +711,8 @@<br>   if (jpeg_buf)<br>     gst_buffer_unref (jpeg_buf);<br> <br>+  gst_buffer_unmap(buf, &info);<div><br>+<br>   /* We must always unref the input buffer since we never push it out */<br>
   gst_buffer_unref (buf);</div></div><br><br></div>...<br></div><div class="HOEnZb"><div class="h5"><div class="gmail_extra"><br><br><div class="gmail_quote">On Fri, Jul 26, 2013 at 12:44 PM, Robert Krakora <span dir="ltr"><<a href="mailto:rob.krakora@messagenetsystems.com" target="_blank">rob.krakora@messagenetsystems.com</a>></span> wrote:<br>

<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">Hi Peter,<br><div><br>OK, I found a leak in uvch264src...<div><br><br>--- gst-plugins-bad-1.1.2/sys/uvch264/gstuvch264_mjpgdemux.c    2013-04-23 08:38:28.000000000 -0400<br>

</div>+++ gst-plugins-bad-1.1.2.new/sys/uvch264/gstuvch264_mjpgdemux.c    2013-07-26 12:20:52.625791724 -0400<br>
@@ -711,6 +711,8 @@<br>   if (jpeg_buf)<br>     gst_buffer_unref (jpeg_buf);<br> <br>+  gst_buffer_unmap(buf, &info);<div><br>+<br>   /* We must always unref the input buffer since we never push it out */<br>
   gst_buffer_unref (buf);<br>
<br></div></div><div>There is still a very slow leak may remain, however...<br></div><div><br><br></div></div><div><div><div class="gmail_extra"><br><br><div class="gmail_quote">On Fri, Jul 26, 2013 at 10:32 AM, Robert Krakora <span dir="ltr"><<a href="mailto:rob.krakora@messagenetsystems.com" target="_blank">rob.krakora@messagenetsystems.com</a>></span> wrote:<br>


<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">Sorry, I attached the patch files and they were scrubbed off by the forum...<br><br>--- gst-plugins-good-1.1.2/sys/v4l2/gstv4l2bufferpool.c    2013-07-08 10:27:16.000000000 -0400<br>


+++ gst-plugins-good-1.1.2.new/sys/v4l2/gstv4l2bufferpool.c    2013-07-26 09:47:22.584092634 -0400<br>
@@ -375,6 +375,7 @@<div><br>     {<br>       /* request a reasonable number of buffers when no max specified. We will<br>        * copy when we run out of buffers */<br></div>+      //max_buffers = 100;<br>       if (max_buffers == 0)<br>



         num_buffers = 4;<br>       else<br>@@ -411,11 +412,11 @@<div><br>       if (max_buffers == 0 || num_buffers < max_buffers) {<br>         /* if we are asked to provide more buffers than we have allocated, start<br>


          * copying buffers when we only have 2 buffers left in the pool */<br>
-        copy_threshold = 2;<br></div>+        copy_threshold = num_buffers + 1;<div><br>       } else {<br>         /* we are certain that we have enough buffers so we don't need to<br>          * copy */<br>
</div>-        copy_threshold = 0;<br>
+        copy_threshold = num_buffers + 1;<br>       }<br>       break;<br>     }<br><br>--- gst-plugins-bad-1.1.2/sys/uvch264/gstuvch264_mjpgdemux.c    2013-04-23 08:38:28.000000000 -0400<br>+++ gst-plugins-bad-1.1.2.new/sys/uvch264/gstuvch264_mjpgdemux.c    2013-07-26 10:07:18.232833437 -0400<br>



@@ -462,10 +462,10 @@<br> {<br>   GstUvcH264MjpgDemux *self;<br>   GstFlowReturn ret = GST_FLOW_OK;<br>-  GstBuffer *jpeg_buf = gst_buffer_copy_region (buf, GST_BUFFER_COPY_METADATA,<br>-      0, 0);<br>-  GstBuffer *aux_buf = NULL;<br>



+  GstBufferList *jpeg_buf = gst_buffer_list_new ();<br>+  GstBufferList *aux_buf = NULL;<br>   AuxiliaryStreamHeader aux_header = { 0 };<br>+  GstBuffer *sub_buffer = NULL;<br>   guint32 aux_size = 0;<br>   GstPad *aux_pad = NULL;<br>



   GstCaps **aux_caps = NULL;<br>@@ -514,9 +514,8 @@<br> <br>       /* Add JPEG data between the last offset and this market */<br>       if (i - last_offset > 0) {<br>-        GstMemory *m = gst_memory_copy (info.memory, last_offset,<br>



-            i - last_offset);<br>-        gst_buffer_append_memory (jpeg_buf, m);<br>+        sub_buffer = gst_buffer_copy_region (buf, GST_BUFFER_COPY_ALL, last_offset, i - last_offset);<br>+        gst_buffer_copy_into (sub_buffer, buf, GST_BUFFER_COPY_METADATA, 0, -1);<br>



       }<br>       last_offset = i + 2 + segment_size;<br> <br>@@ -624,7 +623,7 @@<br>           }<br> <br>           /* Create new auxiliary buffer list and adjust i/segment size */<br>-          aux_buf = gst_buffer_new ();<br>



+          aux_buf = gst_buffer_list_new ();<br>         }<br> <br>         i += sizeof (aux_header) + sizeof (aux_size);<br>@@ -640,15 +639,12 @@<br>       }<br> <br>       if (segment_size > 0) {<br>-        GstMemory *m;<br>



-        m = gst_memory_copy (info.memory, i, segment_size);<br>-<br>-        GST_BUFFER_DURATION (aux_buf) =<br>+        sub_buffer = gst_buffer_copy_region (buf, GST_BUFFER_COPY_ALL, i, segment_size);<br>+        GST_BUFFER_DURATION (sub_buffer) =<br>



             aux_header.frame_interval * 100 * GST_NSECOND;<br>+        gst_buffer_copy_into (sub_buffer, buf, GST_BUFFER_COPY_TIMESTAMPS, 0, -1);<br> <br>-        _pts_to_timestamp (self, aux_buf, aux_header.pts);<br>-<br>



-        gst_buffer_append_memory (aux_buf, m);<br>+        _pts_to_timestamp (self, sub_buffer, aux_header.pts);<br> <br>         aux_size -= segment_size;<br> <br>@@ -657,7 +653,7 @@<br>           GST_DEBUG_OBJECT (self, "Pushing %" GST_FOURCC_FORMAT<br>



               " auxiliary buffer %" GST_PTR_FORMAT,<br>               GST_FOURCC_ARGS (aux_header.type), *aux_caps);<br>-          ret = gst_pad_push (aux_pad, aux_buf);<br>+          ret = gst_pad_push_list (aux_pad, aux_buf);<br>



           aux_buf = NULL;<br>           if (ret != GST_FLOW_OK) {<br>             GST_WARNING_OBJECT (self, "Error pushing %" GST_FOURCC_FORMAT<br>@@ -669,13 +665,12 @@<br> <br>       i += segment_size - 1;<br>



     } else if (data[i] == 0xff && data[i + 1] == 0xda) {<br>-      GstMemory *m;<br> <br>       /* The APP4 markers must be before the SOS marker, so this is the end */<br>       GST_DEBUG_OBJECT (self, "Found SOS marker.");<br>



 <br>-      m = gst_memory_copy (info.memory, last_offset, size - last_offset);<br>-      gst_buffer_append_memory (jpeg_buf, m);<br>+      sub_buffer = gst_buffer_copy_region (buf, GST_BUFFER_COPY_ALL, last_offset, size - last_offset);<br>



+      gst_buffer_copy_into (sub_buffer, buf, GST_BUFFER_COPY_METADATA, 0, -1);<br>       last_offset = size;<br>       break;<br>     }<br>@@ -692,10 +687,10 @@<br>     /* this means there was no SOS marker in the jpg, so we assume the JPG was<br>



        just a container */<br>     GST_DEBUG_OBJECT (self, "SOS marker wasn't found. MJPG is container only");<br>-    gst_buffer_unref (jpeg_buf);<br>+    gst_buffer_list_unref (jpeg_buf);<br>     jpeg_buf = NULL;<br>



   } else {<br>-    ret = gst_pad_push (self->priv->jpeg_pad, jpeg_buf);<br>+    ret = gst_pad_push_list (self->priv->jpeg_pad, jpeg_buf);<br>     jpeg_buf = NULL;<br>   }<br> <br>@@ -707,9 +702,9 @@<br> done:<br>



   /* In case of error, unref whatever was left */<br>   if (aux_buf)<br>-    gst_buffer_unref (aux_buf);<br>+    gst_buffer_list_unref (aux_buf);<br>   if (jpeg_buf)<br>-    gst_buffer_unref (jpeg_buf);<br>+    gst_buffer_list_unref (jpeg_buf);<br>



 <br>   /* We must always unref the input buffer since we never push it out */<br>   gst_buffer_unref (buf);<br><br><br></div><div><div><div class="gmail_extra"><br><br><div class="gmail_quote">
On Fri, Jul 26, 2013 at 10:22 AM, Robert Krakora <span dir="ltr"><<a href="mailto:rob.krakora@messagenetsystems.com" target="_blank">rob.krakora@messagenetsystems.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div>Hi Peter,<br><br></div>Here is what I am running with currently...still has leak, but no error on JPEG parsing...<br>



<br>Best Regards,<br><br>Rob<br><br></div><div><div><div class="gmail_extra"><br><br><div class="gmail_quote">
On Fri, Jul 26, 2013 at 9:45 AM, Peter Rennert <span dir="ltr"><<a href="mailto:p.rennert@cs.ucl.ac.uk" target="_blank">p.rennert@cs.ucl.ac.uk</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">





  
    
  
  <div bgcolor="#FFFFFF" text="#000000">
    <div>After playing a bit around it with it,
      I think the memory leak arises if you set copy_threshold too high
      (25 seems to be still ok). In my patch in the previous email I had
      it set to 15 only, and capped the max_buffers to 100 in the
      gstv4l2src.c<br>
      <br>
      Alternatively, I can also set <br>
      <br>
      diff --git a/sys/v4l2/gstv4l2bufferpool.c
      b/sys/v4l2/gstv4l2bufferpool.c                                                                    
      <br>
      index 1e74fc7..90e8470
      100644                                                                                                               
      <br>
      --- a/sys/v4l2/gstv4l2bufferpool.c<br>
      +++ b/sys/v4l2/gstv4l2bufferpool.c<br>
      @@ -376,7 +376,7 @@ gst_v4l2_buffer_pool_set_config (GstBufferPool
      * bpool, GstStructure * config)<br>
             /* request a reasonable number of buffers when no max
      specified. We will<br>
              * copy when we run out of buffers */<br>
             if (max_buffers == 0)<br>
      -        num_buffers = 4;<br>
      +        num_buffers = 100;<br>
             else<br>
               num_buffers = max_buffers;<br>
      <br>
      (using the original gstv4l2src.c)<br>
      <br>
      leaving<br>
      <br>
      copy_threshold = 2;<br>
      <br>
      Intact. Now the pipeline runs without memory leak for 4.35min and
      crashes then. <br>
      <br>
      I get exactly the same behaviour if I set num_buffers = 100;
      copy_threshold = 25; or num_buffers = 200; copy_threshold = 25;<div><div><br>
      <br>
      On 07/26/2013 02:26 PM, Robert Krakora wrote:<br>
    </div></div></div><div><div>
    <blockquote type="cite">
      <div dir="ltr">Basically, you want to set up v4l2src so that it
        ALWAYS copies it's buffer pool buffers to freshly allocated
        buffers...this was it's default behaviour in 0.10.<br>
      </div>
      <div class="gmail_extra"><br>
        <br>
        <div class="gmail_quote">On Fri, Jul 26, 2013 at 9:24 AM, Robert
          Krakora <span dir="ltr"><<a href="mailto:rob.krakora@messagenetsystems.com" target="_blank">rob.krakora@messagenetsystems.com</a>></span>
          wrote:<br>
          <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
            <div dir="ltr">If the only thing that I do is set the copy
              threshold to 100 I can run until the memory leak that I
              mentioned before invokes the OS to kill gst-launch...I
              don't see the error reported...<br>
            </div>
            <div>
              <div>
                <div class="gmail_extra">
                  <br>
                  <br>
                  <div class="gmail_quote">On Fri, Jul 26, 2013 at 9:03
                    AM, Peter Rennert <span dir="ltr"><<a href="mailto:p.rennert@cs.ucl.ac.uk" target="_blank">p.rennert@cs.ucl.ac.uk</a>></span>
                    wrote:<br>
                    <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
                      <div bgcolor="#FFFFFF" text="#000000">
                        <div>Hmm that is not the solution. It still
                          fails after some minutes, reporting the same
                          error as before. I am trying to find the right
                          way to manipulate the buffer numbers now.
                          <div>
                            <div><br>
                              <br>
                              <br>
                              On 07/26/2013 01:13 PM, Peter Rennert
                              wrote:<br>
                            </div>
                          </div>
                        </div>
                        <div>
                          <div>
                            <blockquote type="cite">
                              <div>Rob,<br>
                                <br>
                                Well spotted!<br>
                                <br>
                                It works for me. I am monitoring the
                                memory only with htop at the moment, but
                                it seems to be stable. I think the
                                memory leak stems from the fact that
                                max_buffers is set to 0, so it will
                                always make the buffer bigger and
                                bigger. I think that is generally a bad
                                idea. This might affect also plain
                                v4l2src applications.<br>
                                <br>
                                Try this patch to get rid of the memory
                                leak:<br>
                                <br>
                                diff --git
                                a/sys/v4l2/gstv4l2bufferpool.c
                                b/sys/v4l2/gstv4l2bufferpool.c                                                                    


                                <br>
                                index 1e74fc7..282ac2b
                                100644                                                                                                               


                                <br>
                                --- a/sys/v4l2/gstv4l2bufferpool.c<br>
                                +++ b/sys/v4l2/gstv4l2bufferpool.c<br>
                                @@ -411,7 +411,7 @@
                                gst_v4l2_buffer_pool_set_config
                                (GstBufferPool * bpool, GstStructure *
                                config)<br>
                                       if (max_buffers == 0 ||
                                num_buffers < max_buffers) {<br>
                                         /* if we are asked to provide
                                more buffers than we have allocated,
                                start<br>
                                          * copying buffers when we only
                                have 2 buffers left in the pool */<br>
                                -        copy_threshold = 2;<br>
                                +        copy_threshold = 15;// 2;<br>
                                       } else {<br>
                                         /* we are certain that we have
                                enough buffers so we don't need to<br>
                                          * copy */<br>
                                diff --git a/sys/v4l2/gstv4l2src.c
                                b/sys/v4l2/gstv4l2src.c<br>
                                index 107ea21..0c5d91a 100644<br>
                                --- a/sys/v4l2/gstv4l2src.c<br>
                                +++ b/sys/v4l2/gstv4l2src.c<br>
                                @@ -529,7 +529,8 @@
                                gst_v4l2src_decide_allocation
                                (GstBaseSrc * bsrc, GstQuery * query)<br>
                                     update = TRUE;<br>
                                   } else {<br>
                                     pool = NULL;<br>
                                -    min = max = 0;<br>
                                +    min = 0;<br>
                                +    max = 100;<br>
                                     size = 0;<br>
                                     update = FALSE;<br>
                                   }<br>
                                <br>
                                <br>
                                On 07/25/2013 09:52 PM, Robert Krakora
                                wrote:<br>
                              </div>
                              <blockquote type="cite">
                                <div dir="ltr">
                                  <div>Hi Peter,<br>
                                    <br>
                                    I also wanted to note that when you
                                    apply the aforementioned "hack" to
                                    emulate the default operation of
                                    v4l2src in version 0.10
                                    ("always-copy=true") with the
                                    pipeline below, your system will run
                                    out of memory due to a memory leak. 
                                    It will then kill off the gst-launch
                                    process instantiated to run your
                                    pipeline.  However, your pipeline
                                    will run for quite a bit (a lot
                                    longer than the previous 36 seconds
                                    noted by Yusuf a couple of months
                                    ago).<br>
                                    <br>
                                    gst-launch-1.0 uvch264src
                                    device=/dev/video0 name=src
                                    auto-start=true src.vfsrc ! queue !
                                    fakesink src.vidsrc ! queue !
                                    video/x-h264 ! fakesink<br>
                                    <br>
                                  </div>
                                  Best Regards,<br>
                                  <br>
                                  Rob Krakora<br>
                                  <br>
                                </div>
                                <div class="gmail_extra"> <br>
                                  <br>
                                  <div class="gmail_quote">On Thu, Jul
                                    25, 2013 at 4:27 PM, Robert Krakora
                                    <span dir="ltr"><<a href="mailto:rob.krakora@messagenetsystems.com" target="_blank">rob.krakora@messagenetsystems.com</a>></span>
                                    wrote:<br>
                                    <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
                                      <div dir="ltr">
                                        <div>Hi Peter,<br>
                                          <br>
                                        </div>
                                        I forgot to mention that the
                                        file that needs modification is
                                        named gstv4l2bufferpool.c and is
                                        under sys/v4l2 under
                                        plugins-good.
                                        <div> <br>
                                          <br>
                                                if (max_buffers == 0 ||
                                          num_buffers < max_buffers)
                                          {<br>
                                                  /* if we are asked to
                                          provide more buffers than we
                                          have allocated, start<br>
                                                   * copying buffers
                                          when we only have 2 buffers
                                          left in the pool */<br>
                                                  copy_threshold = 100;
                                          //2;<br>
                                                } else {<br>
                                                  /* we are certain that
                                          we have enough buffers so we
                                          don't need to<br>
                                                   * copy */<br>
                                                  copy_threshold = 0;<br>
                                                }<br>
                                        </div>
                                      </div>
                                      <div>
                                        <div>
                                          <div class="gmail_extra"><br>
                                            <br>
                                            <div class="gmail_quote">On
                                              Thu, Jul 25, 2013 at 4:21
                                              PM, Robert Krakora <span dir="ltr"><<a href="mailto:rob.krakora@messagenetsystems.com" target="_blank">rob.krakora@messagenetsystems.com</a>></span>
                                              wrote:<br>
                                              <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
                                                <div dir="ltr">
                                                  <div>
                                                    <div>Hi Peter,<br>
                                                      <br>
                                                    </div>
                                                    In version 0.10
                                                    v4l2src used to have
                                                    a property to force
                                                    buffers from it's
                                                    pool to always be
                                                    copied prior to
                                                    being pushed out. 
                                                    This property was
                                                    called "always-copy"
                                                    and defaulted to
                                                    "true".  It seems
                                                    that this was
                                                    removed in version
                                                    1.x.  If you go to
                                                    version 1.x good
                                                    plugins and change
                                                    "copy_threadhold"
                                                    from 2 to 100 you
                                                    effectively get the
                                                    same behaviour with
                                                    1.x in this regard
                                                    (same as
                                                    "always-copy=true"
                                                    default in 0.10
                                                    v4l2src) and there
                                                    is no error after 30
                                                    some odd seconds
                                                    that causes the
                                                    stream to abort.<br>
                                                    <br>
                                                          if
                                                    (max_buffers == 0 ||
                                                    num_buffers <
                                                    max_buffers) {<br>
                                                            /* if we are
                                                    asked to provide
                                                    more buffers than we
                                                    have allocated,
                                                    start<br>
                                                             * copying
                                                    buffers when we only
                                                    have 2 buffers left
                                                    in the pool */<br>
                                                           
                                                    copy_threshold =
                                                    100; //2;<br>
                                                          } else {<br>
                                                            /* we are
                                                    certain that we have
                                                    enough buffers so we
                                                    don't need to<br>
                                                             * copy */<br>
                                                           
                                                    copy_threshold = 0;<br>
                                                          }<br>
                                                    <br>
                                                  </div>
                                                  Best Regards,<br>
                                                  <br>
                                                  Rob Krakora<br>
                                                  <br>
                                                </div>
                                                <div>
                                                  <div>
                                                    <div class="gmail_extra"><br>
                                                      <br>
                                                      <div class="gmail_quote">On
                                                        Thu, Jul 25,
                                                        2013 at 1:06 PM,
                                                        Robert Krakora <span dir="ltr"><<a href="mailto:rob.krakora@messagenetsystems.com" target="_blank">rob.krakora@messagenetsystems.com</a>></span>
                                                        wrote:<br>
                                                        <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
                                                          <div dir="ltr">
                                                          <div>
                                                          <div>
                                                          <div>Hi Peter,<br>
                                                          <br>
                                                          </div>
                                                          I did some
                                                          work yesterday
                                                          on this and
                                                          enabled
                                                          logging on
                                                          uvcvideo.ko
                                                          and was able
                                                          to correlate
                                                          the buffer
                                                          sizes reported
                                                          by it and by
                                                          the uvch264src
                                                          plugin.  <span style="background-color:rgb(255,255,0)">Below</span> is the frame that
                                                          failed...the
                                                          size in the
                                                          kernel module
                                                          was the same
                                                          as the size of
                                                          the buffer
                                                          once it got
                                                          back up to the
                                                          application to
                                                          v4l2src and
                                                          uvch264src.<br>
                                                          <br>
                                                          uvcvideo:
                                                          Frame complete
                                                          (EOF found).<br>
                                                          uvcvideo: EOF
                                                          in empty
                                                          payload.<br>
                                                          uvcvideo:
                                                          uvc_v4l2_poll<br>
                                                          uvcvideo:
                                                          uvc_v4l2_ioctl(VIDIOC_DQBUF)<br>
                                                          uvcvideo: HD
                                                          Pro Webcam
                                                          C920: PTS
                                                          1029592232 y
                                                          2948.098587
                                                          SOF
                                                          2948.098587
                                                          (x1 <a href="tel:2149480048" value="+12149480048" target="_blank">2149480048</a>
                                                          x2 <a href="tel:2179588048" value="+12179588048" target="_blank">2179588048</a>
                                                          y1 193593344
                                                          y2 199426048
                                                          SOF offset 39)<br>
                                                          uvcvideo: HD
                                                          Pro Webcam
                                                          C920: SOF
                                                          2948.098587 y
                                                          927116325 ts
                                                          589.071670 buf
                                                          ts 589.232519
                                                          (x1
                                                          197984256/205/906
                                                          x2
                                                          204537856/49/995
                                                          y1 1000000000
                                                          y2 1099975668)<br>
                                                          <span style="background-color:rgb(255,255,0)"><span>uvcvideo:


                                                          uvc_dequeue_buffer
                                                          -
                                                          buf->bytesused
                                                          = 220410</span></span><br>
                                                          uvcvideo:
                                                          uvc_v4l2_ioctl(VIDIOC_QBUF)<br>
                                                          <br>
                                                          </div>
                                                          Best Regards,<br>
                                                          <br>
                                                          </div>
                                                          Rob<br>
                                                          </div>
                                                          <div class="gmail_extra">
                                                          <div>
                                                          <div><br>
                                                          <br>
                                                          <div class="gmail_quote">On
                                                          Thu, Jul 25,
                                                          2013 at 12:22
                                                          PM, Peter
                                                          Rennert <span dir="ltr"><<a href="mailto:p.rennert@cs.ucl.ac.uk" target="_blank">p.rennert@cs.ucl.ac.uk</a>></span>
                                                          wrote:<br>
                                                          <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">A
                                                          bit of an
                                                          update today.<br>
                                                          <br>
                                                          To figure out
                                                          what is going
                                                          on normally
                                                          and what is
                                                          different in
                                                          the diseased
                                                          frame I
                                                          printed out
                                                          the debug line
                                                          [line 508]
                                                          that was
                                                          already in the
                                                          code,
                                                          augmented with
                                                          the total size
                                                          of the buffer.
                                                          The original
                                                          line was:<br>
                                                          <br>
                                                          GST_DEBUG_OBJECT
                                                          (self,<br>
                                                                   
                                                          "Found APP4
                                                          marker (%d).
                                                          JPG: %d-%d -
                                                          APP4: %d -
                                                          %d",
                                                          segment_size,<br>
                                                                   
                                                          last_offset,
                                                          i, i, i + 2 +
                                                          segment_size);<br>
                                                          <br>
                                                          I print for
                                                          each time in
                                                          the loop
                                                          (before the
                                                          sanity test
                                                          that will fail
                                                          finally):<br>
                                                          <br>
                                                          printf("Found
                                                          APP4 marker
                                                          (%d). JPG:
                                                          %d-%d - APP4:
                                                          %d - %d -
                                                          size: %d\n",
                                                          segment_size,<br>
                                                                   
                                                          last_offset,
                                                          i, i, i + 2 +
                                                          segment_size,
                                                          (int)size);<br>
                                                          <br>
                                                          What I get for
                                                          a normal frame
                                                          is about this.
                                                          The total
                                                          buffer size
                                                          changes and
                                                          also the size
                                                          of the first
                                                          APP4 chunk,
                                                          however, I
                                                          always seem to
                                                          get 4 blocks
                                                          for each
                                                          frame. The
                                                          segment size
                                                          of the first
                                                          marker differs
                                                          between the
                                                          individual
                                                          frames, while
                                                          the 2nd, 3rd
                                                          and 4th
                                                          markers are
                                                          the same size
                                                          all the time
                                                          as far as I
                                                          can tell:<br>
                                                          <br>
                                                          Found APP4
                                                          marker
                                                          (13010). JPG:
                                                          0-8 - APP4: 8
                                                          - 13020 -
                                                          size: 166660<br>
                                                          <br>
                                                          Found APP4
                                                          marker
                                                          (65533). JPG:
                                                          13020-13020 -
                                                          APP4: 13020 -
                                                          78555 - size:
                                                          166660<br>
                                                          <br>
                                                          Found APP4
                                                          marker
                                                          (65533). JPG:
                                                          <a href="tel:78555-78555" value="+17855578555" target="_blank">78555-78555</a>
                                                          - APP4: 78555
                                                          - 144090 -
                                                          size: 166660<br>
                                                          <br>
                                                          Found APP4
                                                          marker
                                                          (22566). JPG:
                                                          144090-144090
                                                          - APP4: 144090
                                                          - 166658 -
                                                          size: 166660<br>
                                                          <br>
                                                          <br>
                                                          The frame that
                                                          causes the
                                                          failure of the
                                                          system has the
                                                          following
                                                          output:<br>
                                                          <br>
                                                          Found APP4
                                                          marker
                                                          (12084). JPG:
                                                          0-8 - APP4: 8
                                                          - 12094 -
                                                          size: 165485<br>
                                                          <br>
                                                          Found APP4
                                                          marker
                                                          (65533). JPG:
                                                          12094-12094 -
                                                          APP4: 12094 -
                                                          77629 - size:
                                                          165485<br>
                                                          <br>
                                                          Found APP4
                                                          marker
                                                          (65533). JPG:
                                                          77629-77629 -
                                                          APP4: 77629 -
                                                          143164 - size:
                                                          165485<br>
                                                          <br>
                                                          Found APP4
                                                          marker
                                                          (22566). JPG:
                                                          143164-143164
                                                          - APP4: 143164
                                                          - 165732 -
                                                          size: 165485<br>
                                                          <br>
                                                          <br>
                                                          As you can see
                                                          it seems as
                                                          the last
                                                          segment
                                                          reaches out of
                                                          the total size
                                                          of the buffer.
                                                          As the last
                                                          segment seems
                                                          to be always
                                                          of a length
                                                          22566, I think
                                                          for some
                                                          reason (that
                                                          does not
                                                          happen in
                                                          gstreamer
                                                          0.10) a piece
                                                          of the h264
                                                          stream is
                                                          missing in
                                                          this
                                                          particular
                                                          buffer.<br>
                                                          <br>
                                                          The number of
                                                          bytes getting
                                                          lost is not
                                                          the same for
                                                          different
                                                          trials,
                                                          neither is the
                                                          cut-off. For
                                                          my second
                                                          trial I got:<br>
                                                          <br>
                                                          Found APP4
                                                          marker
                                                          (22566). JPG:
                                                          144796-144796
                                                          - APP4: 144796
                                                          - 167364 -
                                                          size: 165815<br>
                                                          <br>
                                                          So where are
                                                          these bytes?
                                                          And why are
                                                          they missing
                                                          so regularly?
                                                          <div>
                                                          <div><br>
_______________________________________________<br>
                                                          gstreamer-devel
                                                          mailing list<br>
                                                          <a href="mailto:gstreamer-devel@lists.freedesktop.org" target="_blank">gstreamer-devel@lists.freedesktop.org</a><br>
                                                          <a href="http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel" target="_blank">http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel</a><br>
                                                          </div>
                                                          </div>
                                                          </blockquote>
                                                          </div>
                                                          <br>
                                                          <br clear="all">
                                                          <br>
                                                          </div>
                                                          </div>
                                                          <div>-- <br>
                                                          <font color="#888888"><font color="blue">Rob
                                                          Krakora<br>
                                                          MessageNet
                                                          Systems <br>
                                                          101 East
                                                          Carmel Dr.
                                                          Suite 105 <br>
                                                          Carmel, IN
                                                          46032 <br>
                                                          <a value="+13175661677">(317)566-1677</a></font></font><font color="#888888"><font color="blue">
                                                          Ext 212<br>
                                                          <a value="+13176630808">(317)663-0808</a> Fax </font></font> </div>
                                                          </div>
                                                        </blockquote>
                                                      </div>
                                                      <br>
                                                      <br clear="all">
                                                      <br>
                                                      -- <br>
                                                      <font color="#888888"><font color="blue">Rob


                                                          Krakora<br>
                                                          MessageNet
                                                          Systems <br>
                                                          101 East
                                                          Carmel Dr.
                                                          Suite 105 <br>
                                                          Carmel, IN
                                                          46032 <br>
                                                          <a value="+13175661677">(317)566-1677</a></font></font><font color="#888888"><font color="blue">
                                                          Ext 212<br>
                                                          <a value="+13176630808">(317)663-0808</a> Fax </font></font> </div>
                                                  </div>
                                                </div>
                                              </blockquote>
                                            </div>
                                            <br>
                                            <br clear="all">
                                            <br>
                                            -- <br>
                                            <font color="#888888"><font color="blue">Rob Krakora<br>
                                                MessageNet Systems <br>
                                                101 East Carmel Dr.
                                                Suite 105 <br>
                                                Carmel, IN 46032 <br>
                                                <a value="+13175661677">(317)566-1677</a></font></font><font color="#888888"><font color="blue"> Ext 212<br>
                                                <a value="+13176630808">(317)663-0808</a>
                                                Fax </font></font> </div>
                                        </div>
                                      </div>
                                    </blockquote>
                                  </div>
                                  <br>
                                  <br clear="all">
                                  <br>
                                  -- <br>
                                  <font color="#888888"><font color="blue">Rob Krakora<br>
                                      MessageNet Systems <br>
                                      101 East Carmel Dr. Suite 105 <br>
                                      Carmel, IN 46032 <br>
                                      <a value="+13175661677">(317)566-1677</a></font></font><font color="#888888"><font color="blue">
                                      Ext 212<br>
                                      <a value="+13176630808">(317)663-0808</a>
                                      Fax </font></font> </div>
                                <br>
                                <fieldset></fieldset>
                                <br>
                                <pre>_______________________________________________
gstreamer-devel mailing list
<a href="mailto:gstreamer-devel@lists.freedesktop.org" target="_blank">gstreamer-devel@lists.freedesktop.org</a>
<a href="http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel" target="_blank">http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel</a>
</pre>
                              </blockquote>
                              <br>
                              <br>
                              <fieldset></fieldset>
                              <br>
                              <pre>_______________________________________________
gstreamer-devel mailing list
<a href="mailto:gstreamer-devel@lists.freedesktop.org" target="_blank">gstreamer-devel@lists.freedesktop.org</a>
<a href="http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel" target="_blank">http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel</a>
</pre>
                            </blockquote>
                            <br>
                          </div>
                        </div>
                      </div>
                      <br>
                      _______________________________________________<br>
                      gstreamer-devel mailing list<br>
                      <a href="mailto:gstreamer-devel@lists.freedesktop.org" target="_blank">gstreamer-devel@lists.freedesktop.org</a><br>
                      <a href="http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel" target="_blank">http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel</a><br>
                      <br>
                    </blockquote>
                  </div>
                  <br>
                  <br clear="all">
                  <br>
                  -- <br>
                  <font color="#888888"><font color="blue">Rob Krakora<br>
                      MessageNet Systems
                      <br>
                      101 East Carmel Dr. Suite 105
                      <br>
                      Carmel, IN 46032
                      <br>
                      <a value="+13175661677">(317)566-1677</a></font></font><font color="#888888"><font color="blue"> Ext 212<br>
                      <a value="+13176630808">(317)663-0808</a>
                      Fax
                    </font></font>
                </div>
              </div>
            </div>
          </blockquote>
        </div>
        <br>
        <br clear="all">
        <br>
        -- <br>
        <font color="#888888"><font color="blue">Rob Krakora<br>
            MessageNet Systems
            <br>
            101 East Carmel Dr. Suite 105
            <br>
            Carmel, IN 46032
            <br>
            <a value="+13175661677">(317)566-1677</a></font></font><font color="#888888"><font color="blue"> Ext 212<br>
            <a value="+13176630808">(317)663-0808</a>
            Fax
          </font></font>
      </div>
      <br>
      <fieldset></fieldset>
      <br>
      <pre>_______________________________________________
gstreamer-devel mailing list
<a href="mailto:gstreamer-devel@lists.freedesktop.org" target="_blank">gstreamer-devel@lists.freedesktop.org</a>
<a href="http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel" target="_blank">http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel</a>
</pre>
    </blockquote>
    <br>
  </div></div></div>

<br>_______________________________________________<br>
gstreamer-devel mailing list<br>
<a href="mailto:gstreamer-devel@lists.freedesktop.org" target="_blank">gstreamer-devel@lists.freedesktop.org</a><br>
<a href="http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel" target="_blank">http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel</a><br>
<br></blockquote></div><br><br clear="all"><br>-- <br><font color="#888888"><font color="blue">Rob Krakora<br>
MessageNet Systems
<br>
101 East Carmel Dr. Suite 105
<br>
Carmel, IN 46032
<br>
<a value="+13175661677">(317)566-1677</a></font></font><font color="#888888"><font color="blue"> Ext 212<br>
<a value="+13176630808">(317)663-0808</a> Fax
</font></font>
</div>
</div></div></blockquote></div><br><br clear="all"><br>-- <br><font color="#888888"><font color="blue">Rob Krakora<br>
MessageNet Systems
<br>
101 East Carmel Dr. Suite 105
<br>
Carmel, IN 46032
<br>
<a value="+13175661677">(317)566-1677</a></font></font><font color="#888888"><font color="blue"> Ext 212<br>
<a value="+13176630808">(317)663-0808</a> Fax
</font></font>
</div>
</div></div></blockquote></div><br><br clear="all"><br>-- <br><font color="#888888"><font color="blue">Rob Krakora<br>
MessageNet Systems
<br>
101 East Carmel Dr. Suite 105
<br>
Carmel, IN 46032
<br>
<a value="+13175661677">(317)566-1677</a></font></font><font color="#888888"><font color="blue"> Ext 212<br>
<a value="+13176630808">(317)663-0808</a> Fax
</font></font>
</div>
</div></div></blockquote></div><br><br clear="all"><br>-- <br><font color="#888888"><font color="blue">Rob Krakora<br>
MessageNet Systems
<br>
101 East Carmel Dr. Suite 105
<br>
Carmel, IN 46032
<br>
<a value="+13175661677">(317)566-1677</a></font></font><font color="#888888"><font color="blue"> Ext 212<br>
<a value="+13176630808">(317)663-0808</a> Fax
</font></font>
</div>
</div></div></blockquote></div><br><br clear="all"><br>-- <br><font color="#888888"><font color="blue">Rob Krakora<br>
MessageNet Systems
<br>
101 East Carmel Dr. Suite 105
<br>
Carmel, IN 46032
<br>
<a value="+13175661677">(317)566-1677</a></font></font><font color="#888888"><font color="blue"> Ext 212<br>
<a value="+13176630808">(317)663-0808</a> Fax
</font></font>
</div>