[FFmpeg-devel] [PATCHv13] libavcodec: v4l2: add support for v4l2 mem2mem codecs

Jorge Ramirez-Ortiz jorge.ramirez-ortiz at linaro.org
Thu Sep 28 20:29:29 EEST 2017

On 09/28/2017 09:48 AM, Thilo Borgmann wrote:
> Am 23.09.17 um 16:48 schrieb Michael Niedermayer:
>> On Sat, Sep 23, 2017 at 12:28:01AM -0700, Jorge Ramirez-Ortiz wrote:
>>> On 09/22/2017 11:49 PM, wm4 wrote:
>>>> On Wed, 20 Sep 2017 18:55:40 -0700
>>>> Jorge Ramirez-Ortiz <jorge.ramirez-ortiz at linaro.org> wrote:
>>>>>      This patchset enhances Alexis Ballier's original patch and validates
>>>>>      it using Qualcomm's Venus hardware (driver recently landed upstream
>>>>>      [1]).
>>>> Pushed to master.
>>> thanks!
>>> will add DRM support over the coming weeks.
>>> btw would you know the process to add a new board to the fate farm?
>> If you wish to run the fate tests regularly and submit results to the
>> server see: https://ffmpeg.org/fate.html
> Someone donated a dragonboard-410c for that purpose during ELCE 2016.
> I plan to set it up as a FATE client during the next weeks.
> When creating test cases, please consider to reuse files already in the fate-suite.

sounds great.
I was just re-testing on 410c and there are a couple of issues in the kernel 
that will need address

I will post this one to ffmpeg once I know the fix has been merged in kernel.org:

For the time being, this kernel patch is required needed for ffmpeg to work with 
the latest 410c release:

diff --git a/drivers/media/platform/qcom/venus/vdec.c 
index 82b4d32..d4badf0 100644
--- a/drivers/media/platform/qcom/venus/vdec.c
+++ b/drivers/media/platform/qcom/venus/vdec.c
@@ -145,9 +145,11 @@ find_format_by_index(struct venus_inst *inst, unsigned int 
index, u32 type)
         if (i == size)
                 return NULL;

+#if 0
         if (type == V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE &&
             !venus_helper_check_codec(inst, fmt->pixfmt))
                 return NULL;

         return &fmt[i];
@@ -489,7 +491,6 @@ vdec_decoder_cmd(struct file *file, void *fh, struct 
v4l2_decoder_cmd *cmd)
         inst->cmd_stop = true;

-   hfi_session_flush(inst);

         return 0;

> -Thilo
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

More information about the ffmpeg-devel mailing list