1. 23 Sep, 2019 3 commits
  2. 10 Sep, 2019 1 commit
    • Doug Nazar's avatar
      alpha: Fix one_over_kc calculation · 1c6e3daa
      Doug Nazar authored
      On arm/aarch64, converting from float directly to unsigned int uses
      a different opcode and negative numbers result in 0. Cast to
      signed int first.
      1c6e3daa
  3. 24 Aug, 2019 2 commits
  4. 21 Aug, 2019 1 commit
  5. 20 Aug, 2019 3 commits
  6. 10 Aug, 2019 1 commit
  7. 08 Aug, 2019 6 commits
  8. 07 Aug, 2019 9 commits
    • Damian Hobson-Garcia's avatar
      v4l2bufferpool: return TRUE when buffer pool orphaning succeeds · 608a9dc1
      Damian Hobson-Garcia authored
      When trying to orphan a buffer pool, successfully return and unref
      the pool when the pool is either successfully stopped or orphaned.
      Indicate failure and leave the pool untouched otherwise.
      608a9dc1
    • Damian Hobson-Garcia's avatar
      v4l2bufferpool: Free orphaned allocator resources when buffers are released · e49cbdb3
      Damian Hobson-Garcia authored
      Allocator resources cannot be freed when a buffer pool is orphaned
      while its buffers are in use. They should, however, be freed once those
      buffers are no longer needed. This patch disposes of any buffers
      belonging to an orphaned pool as they are released, and makes sure
      that the allocator is cleaned up when the last buffer is returned.
      e49cbdb3
    • Damian Hobson-Garcia's avatar
      v4l2object: Orphan buffer pool on object_stop if supported · d20a82aa
      Damian Hobson-Garcia authored
      Use V4L2 buffer orphaning, on recent kernels so that
      the device can be restarted immediately with
      a new buffer pool during renogatiation.
      d20a82aa
    • Sebastian Dröge's avatar
      splitmuxsink: Only set running time on finalizing sink element when in async-finalize mode · 0f5dc763
      Sebastian Dröge authored
      There is only a single sink element in async-finalize mode, and we would
      keep the running time from previous fragments set in that case. As we
      don't ever set the running time for the very last fragment on EOS, this
      would mean that the closing time reported for the very last fragment is
      the same as the closing time of the previous fragment.
      0f5dc763
    • Nicolas Dufresne's avatar
      rtpsession: Always keep at least one NACK on early RTCP · 0d77e972
      Nicolas Dufresne authored
      We recently added code to remove outdate NACK to avoid using bandwidth
      for packet that have no chance of arriving on time. Though, this had a
      side effect, which is that it was to get an early RTCP packet with no
      feedback into it. This was pretty useless but also had a side effect,
      which is that the RTX RTT value would never be updated. So we we stared
      having late RTX request due to high RTT, we'd never manage to recover.
      
      This fixes the regression by making sure we keep at least one NACK in
      this situation. This is really light on the bandwidth and allow for
      quick recover after the RTT have spiked higher then the jitterbuffer
      capacity.
      0d77e972
    • Nicolas Dufresne's avatar
      rtpsession: Call on-new-ssrc earlier · 67b5edfe
      Nicolas Dufresne authored
      Right now, we may call on-new-ssrc after we have processed the first
      RTP packet. This prevents properly configuring the source as some
      property like "probation" are copied internally for use as a
      decreasing counter. For this specific property, it prevents the
      application from disabling probation on auxiliary sparse stream.
      
      Probation is harmful on sparse streams since the probation algorithm
      assume frequent and contiguous RTP packets.
      67b5edfe
    • Nicolas Dufresne's avatar
      0de3ebc5
    • Thiago Sousa Santos's avatar
      rtspsrc: do not try to send EOS with invalid seqnum · 9ee320f6
      Thiago Sousa Santos authored
      The second udpsrc (rtcp) might not have seen the segment event if it was
      not enabled or if rtcp is not available on the server. So if the
      application tries to send an EOS event it will try to set an invalid
      seqnum to the event.
      9ee320f6
    • Sebastian Dröge's avatar
      rtprawdepay: Don't get rid of the buffer pool on FLUSH_STOP · 49e87840
      Sebastian Dröge authored
      We expect there to be a pool as long as the caps are known and
      FLUSH_STOP is not resetting the caps. Getting rid of the pool would
      cause assertions.
      
      Fixes #584
      49e87840
  9. 06 Aug, 2019 2 commits
  10. 02 May, 2019 1 commit
  11. 18 Apr, 2019 4 commits
  12. 17 Apr, 2019 1 commit
  13. 16 Apr, 2019 1 commit
  14. 10 Apr, 2019 2 commits
  15. 09 Apr, 2019 3 commits
    • Tim-Philipp Müller's avatar
      rtpulpfecdec,enc: unbreak plugin gtk-doc build in autotools · c6c3bed0
      Tim-Philipp Müller authored
      Fix doc chunks to not use that syntax for links that have the
      url as description, it will be put verbatim into the xml/*.xml
      file and then the expat parser will throw a syntax error like:
      
        File "../../common/mangle-db.py", line 71, in <module>
          main()
        File "../../common/mangle-db.py", line 69, in main
          patch (details.replace("-details", ""), os.path.basename(details))
        File "../../common/mangle-db.py", line 20, in patch
          doc = xml.dom.minidom.parse(related)
        File "/usr/lib/python2.7/xml/dom/minidom.py", line 1918, in parse
          return expatbuilder.parse(file)
        File "/usr/lib/python2.7/xml/dom/expatbuilder.py", line 924, in parse
          result = builder.parseFile(fp)
        File "/usr/lib/python2.7/xml/dom/expatbuilder.py", line 207, in parseFile
          parser.Parse(buffer, 0)
      xml.parsers.expat.ExpatError: not well-formed (invalid token): line 84, column 7
      c6c3bed0
    • Antonio Ospite's avatar
      rtpvrawpay: preserve GST_BUFFER_FLAG_DISCONT on the first outputted buffer · 61c1385c
      Antonio Ospite authored
      If the incoming frame buffer has GST_BUFFER_FLAG_DISCONT set this should
      be preserved and set for the first output buffer too, like other
      payloaders do.
      
      Spotted with gst-validate-1.0 when adding integration tests for
      rtpsession, a minimal test to reproduce the issue is:
      
      $ gst-validate-1.0 videotestsrc num-buffers=1 ! rtpvrawpay ! identity ! fakesink
      Starting pipeline
      Pipeline started
         warning : Buffer didn't have expected DISCONT flag333 speed: 1.000000 />
                   Detected on <identity0:sink>
                   Detected on <identity0:src>
                   Detected on <fakesink0:sink>
                   Description : Buffers after SEGMENT and FLUSH must have a DISCONT flag
      
      Issues found: 1
      
      =======> Test PASSED (Return value: 0)
      61c1385c
    • Olivier Crête's avatar
      92138dc3