Commit e7eb8964 authored by Martin Storsjö's avatar Martin Storsjö Committed by Rafaël Carré

omxil: Update the explanation of the OMX.SEC quirk/workaround

Signed-off-by: default avatarRafaël Carré <funman@videolan.org>
parent 5290741d
......@@ -512,14 +512,14 @@ static OMX_ERRORTYPE GetPortDefinition(decoder_t *p_dec, OmxPort *p_port,
strlen("OMX.qcom.video.decoder")))
def->format.video.eColorFormat = OMX_QCOM_COLOR_FormatYVU420SemiPlanar;
/* Hack: Galaxy S II (stock firmware) gives a slice height larger
* than the video height, but this doesn't imply padding between
* the video planes. Nexus S also has a slice height larger than
* the video height, but there it actually is real padding, thus
* Galaxy S II is the buggy one. The Galaxy S II decoder is
* named OMX.SEC.avcdec while the one on Nexus S is
* OMX.SEC.AVC.Decoder. Thus do this for any OMX.SEC. that don't
* contain the string ".Decoder". */
/* Hack: Some Samsung devices (e.g. Galaxy S III) have multiple
* H264 decoders with different quirks (OMX.SEC.avc.dec, OMX.SEC.avcdec
* and OMX.SEC.AVC.Decoder) - the latter is well-behaved while the
* former ones signal padding but in practice doesn't have any padding.
* We can't simply ignore the buggy ones, because some devices only
* have that one (Galaxy S II has only got one named OMX.SEC.avcdec,
* at least in the pre-4.0 firmwares). Thus, we enable this quirk on
* any OMX.SEC. decoder that doesn't contain the string ".Decoder". */
if(!strncmp(p_sys->psz_component, "OMX.SEC.", strlen("OMX.SEC.")) &&
!strstr(p_sys->psz_component, ".Decoder"))
def->format.video.nSliceHeight = 0;
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment