Re: [Libva] [Libva-intel-driver][PATCH] Code cleanup for vme/mfc initializing on SKL

2016-09-07 Thread Sean V Kelley
On Wed, 2016-09-07 at 13:49 +0800, Zhao Yakui wrote: > On 09/07/2016 11:01 AM, Xiang, Haihao wrote: > > > > > > i965_encoder.c is a general file, it would be better not to include > > more HW/implementation related code in this file. > > > > Actually it is more clear if you look into the new > >

Re: [Libva] [Libva-intel-driver][PATCH] Code cleanup for vme/mfc initializing on SKL

2016-09-06 Thread Zhao Yakui
On 09/07/2016 11:01 AM, Xiang, Haihao wrote: i965_encoder.c is a general file, it would be better not to include more HW/implementation related code in this file. Actually it is more clear if you look into the new gen9_vme_context_init() and gen9_mfc_context_init(). Previous it selects differen

Re: [Libva] [Libva-intel-driver][PATCH] Code cleanup for vme/mfc initializing on SKL

2016-09-06 Thread Xiang, Haihao
i965_encoder.c is a general file, it would be better not to include more HW/implementation related code in this file.  Actually it is more clear if you look into the new gen9_vme_context_init() and gen9_mfc_context_init(). Previous it selects different path both in gen9_enc_hw_context_init(), gen

Re: [Libva] [Libva-intel-driver][PATCH] Code cleanup for vme/mfc initializing on SKL

2016-09-06 Thread Zhao Yakui
On 09/06/2016 11:39 PM, Xiang, Haihao wrote: It keeps i965_encoder.c simple Thanks for the patch. But I don't think that this patch is necessary. The code looks more clear if it can select the different initialization callback function earlier based on the corresponding profile/entrypoint .

[Libva] [Libva-intel-driver][PATCH] Code cleanup for vme/mfc initializing on SKL

2016-09-06 Thread Xiang, Haihao
It keeps i965_encoder.c simple Signed-off-by: Xiang, Haihao --- src/gen6_mfc.h | 6 ++ src/gen6_vme.h | 2 +- src/gen9_mfc.c | 25 +++-- src/gen9_vme.c | 15 ++- src/i965_encoder.c | 37 ++--- 5 files changed,