| Message ID | 1411390322-25212-1-git-send-email-sjoerd.simons@collabora.co.uk (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Hi Sjoerd, > From: linux-media-owner@vger.kernel.org [mailto:linux-media- > owner@vger.kernel.org] On Behalf Of Sjoerd Simons > Sent: Monday, September 22, 2014 2:52 PM > To: Kyungmin Park; Kamil Debski; Arun Kumar K > Cc: Mauro Carvalho Chehab; linux-arm-kernel@lists.infradead.org; linux- > media@vger.kernel.org; linux-kernel@vger.kernel.org; linux-samsung- > soc@vger.kernel.org; Daniel Drake; Sjoerd Simons > Subject: [PATCH] [media] s5p-mfc: Use decode status instead of display > status on MFCv5 > > Commit 90c0ae50097 changed how the frame_type of a decoded frame > gets determined, by switching from the get_dec_frame_type to > get_disp_frame_type operation. Unfortunately it seems that on MFC v5 > the > result of get_disp_frame_type is always 0 (no display) when decoding > (tested with H264), resulting in no frame ever being output from the > decoder. Could you tell me which firmware version do you use (date)? Best wishes,
Hey Kamil, On Tue, 2014-09-23 at 15:58 +0200, Kamil Debski wrote: > > Commit 90c0ae50097 changed how the frame_type of a decoded frame > > gets determined, by switching from the get_dec_frame_type to > > get_disp_frame_type operation. Unfortunately it seems that on MFC v5 > > the > > result of get_disp_frame_type is always 0 (no display) when decoding > > (tested with H264), resulting in no frame ever being output from the > > decoder. > > Could you tell me which firmware version do you use (date)? I'm using the firmware version as included in http://git.kernel.org/cgit/linux/kernel/git/firmware/linux-firmware.git/ Unfortunately there is no specific version information included about the firmware. The commit that added it is fb5cda9c70277f6 dated Nov 28 17:43:06 2012 +0530, so that at least gives some information about the time-frame.
diff --git a/drivers/media/platform/s5p-mfc/s5p_mfc.c b/drivers/media/platform/s5p-mfc/s5p_mfc.c index d35b041..27ca9d0 100644 --- a/drivers/media/platform/s5p-mfc/s5p_mfc.c +++ b/drivers/media/platform/s5p-mfc/s5p_mfc.c @@ -264,7 +264,12 @@ static void s5p_mfc_handle_frame_new(struct s5p_mfc_ctx *ctx, unsigned int err) unsigned int frame_type; dspl_y_addr = s5p_mfc_hw_call(dev->mfc_ops, get_dspl_y_adr, dev); - frame_type = s5p_mfc_hw_call(dev->mfc_ops, get_disp_frame_type, ctx); + if (IS_MFCV6_PLUS(dev)) + frame_type = s5p_mfc_hw_call(dev->mfc_ops, + get_disp_frame_type, ctx); + else + frame_type = s5p_mfc_hw_call(dev->mfc_ops, + get_dec_frame_type, dev); /* If frame is same as previous then skip and do not dequeue */ if (frame_type == S5P_FIMV_DECODE_FRAME_SKIPPED) {
Commit 90c0ae50097 changed how the frame_type of a decoded frame gets determined, by switching from the get_dec_frame_type to get_disp_frame_type operation. Unfortunately it seems that on MFC v5 the result of get_disp_frame_type is always 0 (no display) when decoding (tested with H264), resulting in no frame ever being output from the decoder. This patch reverts MFC v5 to the previous behaviour while keeping the new behaviour for v6 and up. Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk> --- drivers/media/platform/s5p-mfc/s5p_mfc.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)