Re: [Libva] [PATCH][libva-intel-driver] i965_validate_config: return unsupported profile
MJPEG Encoding will fail on more than just SNB - IVB, Haswell, Broadwell, etc. Make sure the tests identify the requisite platforms. On Wed, Oct 19, 2016 at 6:25 AM, Xiang, Haihao wrote: > > Neither JPEG decoding nor JPEG encoding is supported on SNB, JPEG cases > will fail on SNB if don't change JPEG cases accordingly. On the old g4x, > AVC cases will fail too. > > Thanks > Haihao > > > >-Original Message- > >From: Libva [mailto:libva-boun...@lists.freedesktop.org] On Behalf Of > Eoff, > >Ullysses A > >Sent: Wednesday, October 19, 2016 2:26 AM > >To: Charles, Daniel > >Cc: libva@lists.freedesktop.org > >Subject: Re: [Libva] [PATCH][libva-intel-driver] i965_validate_config: > return > >unsupported profile > > > > > >> -Original Message- > >> From: Charles, Daniel [mailto:daniel.char...@intel.com] > >> Sent: Tuesday, October 18, 2016 11:09 AM > >> To: Eoff, Ullysses A > >> Cc: libva@lists.freedesktop.org > >> Subject: Re: [Libva] [PATCH][libva-intel-driver] i965_validate_config: > return > >unsupported profile > >> > >> On Tue, Oct 18, 2016 at 10:48 AM, Eoff, Ullysses A > >> wrote: > >> >> -Original Message----- > >> >> From: Charles, Daniel [mailto:daniel.char...@intel.com] > >> >> Sent: Tuesday, October 18, 2016 10:20 AM > >> >> To: Eoff, Ullysses A > >> >> Cc: libva@lists.freedesktop.org > >> >> Subject: Re: [Libva] [PATCH][libva-intel-driver] > i965_validate_config: > >return unsupported profile > >> >> > >> >> On Tue, Oct 18, 2016 at 8:25 AM, Eoff, Ullysses A > >> >> wrote: > >> >> > Please run the tests... I suspect this patch will cause some to > fail. Those > >tests will need updated too. > >> >> > > >> >> > >> >> Ran the test suite on ToT, no regression found with this patch. > >> >> > >> > > >> > Ah yes, I see why now. The AVC and JPEG config tests still pass on > most > >HW because UNSUPPORTED_PROFILE is only returned if > >> *both* encode *and* decode are unsupported. Thus, if there is such HW > >that doesn't support one or the other then those tests will > >> be regressed. > >> > >> I am not sure I follow you here. > >> > >> Unsupported Profile is sent when neither encoder and decoder are > >> supported (VC1 is only decoder, VPP is a special case and VP9 decoder > >> can also be hybrid on some h/w). When h/w only supports one of them > >> (regardless which it is), it should return UNSUPPORTED_ENTRYPOINT as > >> profile is supported at least once. This patch is not changing this > >> last scenario. My testing was done on a h/w that has Jpeg Decoder but > >> no Jpeg Encoder and everything is okay. > >> > > > >Right now, there are tests for AVC and JPEG that create a config for > encode > >and decode entrypoints. Each test defines its own expected result of > >CreateConfig and compares with the actual result. If there is a platform > that > >doesn't support any of the entrypoints for AVC or JPEG, then the expected > >result will be calculated wrong (i.e. the test expectation is not taking > into > >account your new changes). (See i965_avce_config_test.cpp, > >i965_avcd_config_test.cpp, i965_jpege_config_test.cpp and > >i965_jpegd_config_test.cpp). > > > >U. Artie > > > >> -- > >> Daniel. > >> > > >> >> -- > >> >> Daniel. > >> >> > U. Artie > >> >> > > >> >> >> -Original Message- > >> >> >> From: Libva [mailto:libva-boun...@lists.freedesktop.org] On > Behalf > >Of Daniel Charles > >> >> >> Sent: Monday, October 17, 2016 5:45 PM > >> >> >> To: libva@lists.freedesktop.org > >> >> >> Subject: [Libva] [PATCH][libva-intel-driver] i965_validate_config: > >return unsupported profile > >> >> >> > >> >> >> When all the profiles are not supported return > >> >> >> VA_STATUS_ERROR_UNSUPPORTED_PROFILE instead of > >> >> >> VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT. > >> >> >> > >> >> >> Also change the style on the code modified to be common > >> >> >> on all cases > >> >> >> > >> >> >> Signed-off-by: Dani
Re: [Libva] [PATCH][libva-intel-driver] i965_validate_config: return unsupported profile
I'll add the patch to the test for jpeg and avc to adapt to this change as the tests currently only account for UNSUPPORTED_ENTRYPOINT Thanks for the feedback Haihao and Artie. -- Daniel. On Wed, Oct 19, 2016 at 6:25 AM, Xiang, Haihao wrote: > > Neither JPEG decoding nor JPEG encoding is supported on SNB, JPEG cases will > fail on SNB if don't change JPEG cases accordingly. On the old g4x, AVC cases > will fail too. > > Thanks > Haihao > > >>-Original Message- >>From: Libva [mailto:libva-boun...@lists.freedesktop.org] On Behalf Of Eoff, >>Ullysses A >>Sent: Wednesday, October 19, 2016 2:26 AM >>To: Charles, Daniel >>Cc: libva@lists.freedesktop.org >>Subject: Re: [Libva] [PATCH][libva-intel-driver] i965_validate_config: return >>unsupported profile >> >> >>> -Original Message- >>> From: Charles, Daniel [mailto:daniel.char...@intel.com] >>> Sent: Tuesday, October 18, 2016 11:09 AM >>> To: Eoff, Ullysses A >>> Cc: libva@lists.freedesktop.org >>> Subject: Re: [Libva] [PATCH][libva-intel-driver] i965_validate_config: >>> return >>unsupported profile >>> >>> On Tue, Oct 18, 2016 at 10:48 AM, Eoff, Ullysses A >>> wrote: >>> >> -Original Message----- >>> >> From: Charles, Daniel [mailto:daniel.char...@intel.com] >>> >> Sent: Tuesday, October 18, 2016 10:20 AM >>> >> To: Eoff, Ullysses A >>> >> Cc: libva@lists.freedesktop.org >>> >> Subject: Re: [Libva] [PATCH][libva-intel-driver] i965_validate_config: >>return unsupported profile >>> >> >>> >> On Tue, Oct 18, 2016 at 8:25 AM, Eoff, Ullysses A >>> >> wrote: >>> >> > Please run the tests... I suspect this patch will cause some to fail. >>> >> > Those >>tests will need updated too. >>> >> > >>> >> >>> >> Ran the test suite on ToT, no regression found with this patch. >>> >> >>> > >>> > Ah yes, I see why now. The AVC and JPEG config tests still pass on most >>HW because UNSUPPORTED_PROFILE is only returned if >>> *both* encode *and* decode are unsupported. Thus, if there is such HW >>that doesn't support one or the other then those tests will >>> be regressed. >>> >>> I am not sure I follow you here. >>> >>> Unsupported Profile is sent when neither encoder and decoder are >>> supported (VC1 is only decoder, VPP is a special case and VP9 decoder >>> can also be hybrid on some h/w). When h/w only supports one of them >>> (regardless which it is), it should return UNSUPPORTED_ENTRYPOINT as >>> profile is supported at least once. This patch is not changing this >>> last scenario. My testing was done on a h/w that has Jpeg Decoder but >>> no Jpeg Encoder and everything is okay. >>> >> >>Right now, there are tests for AVC and JPEG that create a config for encode >>and decode entrypoints. Each test defines its own expected result of >>CreateConfig and compares with the actual result. If there is a platform that >>doesn't support any of the entrypoints for AVC or JPEG, then the expected >>result will be calculated wrong (i.e. the test expectation is not taking into >>account your new changes). (See i965_avce_config_test.cpp, >>i965_avcd_config_test.cpp, i965_jpege_config_test.cpp and >>i965_jpegd_config_test.cpp). >> >>U. Artie >> >>> -- >>> Daniel. >>> > >>> >> -- >>> >> Daniel. >>> >> > U. Artie >>> >> > >>> >> >> -Original Message- >>> >> >> From: Libva [mailto:libva-boun...@lists.freedesktop.org] On Behalf >>Of Daniel Charles >>> >> >> Sent: Monday, October 17, 2016 5:45 PM >>> >> >> To: libva@lists.freedesktop.org >>> >> >> Subject: [Libva] [PATCH][libva-intel-driver] i965_validate_config: >>return unsupported profile >>> >> >> >>> >> >> When all the profiles are not supported return >>> >> >> VA_STATUS_ERROR_UNSUPPORTED_PROFILE instead of >>> >> >> VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT. >>> >> >> >>> >> >> Also change the style on the code modified to be common >>> >> >> on all cases >>> >> >> >>> >> >> Signed-off-by: Daniel Charles >>> >> >&
Re: [Libva] [PATCH][libva-intel-driver] i965_validate_config: return unsupported profile
Neither JPEG decoding nor JPEG encoding is supported on SNB, JPEG cases will fail on SNB if don't change JPEG cases accordingly. On the old g4x, AVC cases will fail too. Thanks Haihao >-Original Message- >From: Libva [mailto:libva-boun...@lists.freedesktop.org] On Behalf Of Eoff, >Ullysses A >Sent: Wednesday, October 19, 2016 2:26 AM >To: Charles, Daniel >Cc: libva@lists.freedesktop.org >Subject: Re: [Libva] [PATCH][libva-intel-driver] i965_validate_config: return >unsupported profile > > >> -Original Message- >> From: Charles, Daniel [mailto:daniel.char...@intel.com] >> Sent: Tuesday, October 18, 2016 11:09 AM >> To: Eoff, Ullysses A >> Cc: libva@lists.freedesktop.org >> Subject: Re: [Libva] [PATCH][libva-intel-driver] i965_validate_config: return >unsupported profile >> >> On Tue, Oct 18, 2016 at 10:48 AM, Eoff, Ullysses A >> wrote: >> >> -Original Message- >> >> From: Charles, Daniel [mailto:daniel.char...@intel.com] >> >> Sent: Tuesday, October 18, 2016 10:20 AM >> >> To: Eoff, Ullysses A >> >> Cc: libva@lists.freedesktop.org >> >> Subject: Re: [Libva] [PATCH][libva-intel-driver] i965_validate_config: >return unsupported profile >> >> >> >> On Tue, Oct 18, 2016 at 8:25 AM, Eoff, Ullysses A >> >> wrote: >> >> > Please run the tests... I suspect this patch will cause some to fail. >> >> > Those >tests will need updated too. >> >> > >> >> >> >> Ran the test suite on ToT, no regression found with this patch. >> >> >> > >> > Ah yes, I see why now. The AVC and JPEG config tests still pass on most >HW because UNSUPPORTED_PROFILE is only returned if >> *both* encode *and* decode are unsupported. Thus, if there is such HW >that doesn't support one or the other then those tests will >> be regressed. >> >> I am not sure I follow you here. >> >> Unsupported Profile is sent when neither encoder and decoder are >> supported (VC1 is only decoder, VPP is a special case and VP9 decoder >> can also be hybrid on some h/w). When h/w only supports one of them >> (regardless which it is), it should return UNSUPPORTED_ENTRYPOINT as >> profile is supported at least once. This patch is not changing this >> last scenario. My testing was done on a h/w that has Jpeg Decoder but >> no Jpeg Encoder and everything is okay. >> > >Right now, there are tests for AVC and JPEG that create a config for encode >and decode entrypoints. Each test defines its own expected result of >CreateConfig and compares with the actual result. If there is a platform that >doesn't support any of the entrypoints for AVC or JPEG, then the expected >result will be calculated wrong (i.e. the test expectation is not taking into >account your new changes). (See i965_avce_config_test.cpp, >i965_avcd_config_test.cpp, i965_jpege_config_test.cpp and >i965_jpegd_config_test.cpp). > >U. Artie > >> -- >> Daniel. >> > >> >> -- >> >> Daniel. >> >> > U. Artie >> >> > >> >> >> -Original Message- >> >> >> From: Libva [mailto:libva-boun...@lists.freedesktop.org] On Behalf >Of Daniel Charles >> >> >> Sent: Monday, October 17, 2016 5:45 PM >> >> >> To: libva@lists.freedesktop.org >> >> >> Subject: [Libva] [PATCH][libva-intel-driver] i965_validate_config: >return unsupported profile >> >> >> >> >> >> When all the profiles are not supported return >> >> >> VA_STATUS_ERROR_UNSUPPORTED_PROFILE instead of >> >> >> VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT. >> >> >> >> >> >> Also change the style on the code modified to be common >> >> >> on all cases >> >> >> >> >> >> Signed-off-by: Daniel Charles >> >> >> --- >> >> >> src/i965_drv_video.c | 49 >- >> >> >> 1 file changed, 40 insertions(+), 9 deletions(-) >> >> >> >> >> >> diff --git a/src/i965_drv_video.c b/src/i965_drv_video.c >> >> >> index fbf2cda..0830ae0 100644 >> >> >> --- a/src/i965_drv_video.c >> >> >> +++ b/src/i965_drv_video.c >> >> >> @@ -711,6 +711,8 @@ i965_validate_config(VADriverContextP ctx, >VAProfile profile, >> >> >> if ((HAS_MPEG2_DECODING(i965) && en
Re: [Libva] [PATCH][libva-intel-driver] i965_validate_config: return unsupported profile
> -Original Message-
> From: Charles, Daniel [mailto:daniel.char...@intel.com]
> Sent: Tuesday, October 18, 2016 11:09 AM
> To: Eoff, Ullysses A
> Cc: libva@lists.freedesktop.org
> Subject: Re: [Libva] [PATCH][libva-intel-driver] i965_validate_config: return
> unsupported profile
>
> On Tue, Oct 18, 2016 at 10:48 AM, Eoff, Ullysses A
> wrote:
> >> -Original Message-
> >> From: Charles, Daniel [mailto:daniel.char...@intel.com]
> >> Sent: Tuesday, October 18, 2016 10:20 AM
> >> To: Eoff, Ullysses A
> >> Cc: libva@lists.freedesktop.org
> >> Subject: Re: [Libva] [PATCH][libva-intel-driver] i965_validate_config:
> >> return unsupported profile
> >>
> >> On Tue, Oct 18, 2016 at 8:25 AM, Eoff, Ullysses A
> >> wrote:
> >> > Please run the tests... I suspect this patch will cause some to fail.
> >> > Those tests will need updated too.
> >> >
> >>
> >> Ran the test suite on ToT, no regression found with this patch.
> >>
> >
> > Ah yes, I see why now. The AVC and JPEG config tests still pass on most
> > HW because UNSUPPORTED_PROFILE is only returned if
> *both* encode *and* decode are unsupported. Thus, if there is such HW that
> doesn't support one or the other then those tests will
> be regressed.
>
> I am not sure I follow you here.
>
> Unsupported Profile is sent when neither encoder and decoder are
> supported (VC1 is only decoder, VPP is a special case and VP9 decoder
> can also be hybrid on some h/w). When h/w only supports one of them
> (regardless which it is), it should return UNSUPPORTED_ENTRYPOINT as
> profile is supported at least once. This patch is not changing this
> last scenario. My testing was done on a h/w that has Jpeg Decoder but
> no Jpeg Encoder and everything is okay.
>
Right now, there are tests for AVC and JPEG that create a config for encode and
decode entrypoints. Each test defines its own expected result of CreateConfig
and compares with the actual result. If there is a platform that doesn't
support any of the entrypoints for AVC or JPEG, then the expected result will
be calculated wrong (i.e. the test expectation is not taking into account your
new changes). (See i965_avce_config_test.cpp, i965_avcd_config_test.cpp,
i965_jpege_config_test.cpp and i965_jpegd_config_test.cpp).
U. Artie
> --
> Daniel.
> >
> >> --
> >> Daniel.
> >> > U. Artie
> >> >
> >> >> -Original Message-
> >> >> From: Libva [mailto:libva-boun...@lists.freedesktop.org] On Behalf Of
> >> >> Daniel Charles
> >> >> Sent: Monday, October 17, 2016 5:45 PM
> >> >> To: libva@lists.freedesktop.org
> >> >> Subject: [Libva] [PATCH][libva-intel-driver] i965_validate_config:
> >> >> return unsupported profile
> >> >>
> >> >> When all the profiles are not supported return
> >> >> VA_STATUS_ERROR_UNSUPPORTED_PROFILE instead of
> >> >> VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT.
> >> >>
> >> >> Also change the style on the code modified to be common
> >> >> on all cases
> >> >>
> >> >> Signed-off-by: Daniel Charles
> >> >> ---
> >> >> src/i965_drv_video.c | 49
> >> >> -
> >> >> 1 file changed, 40 insertions(+), 9 deletions(-)
> >> >>
> >> >> diff --git a/src/i965_drv_video.c b/src/i965_drv_video.c
> >> >> index fbf2cda..0830ae0 100644
> >> >> --- a/src/i965_drv_video.c
> >> >> +++ b/src/i965_drv_video.c
> >> >> @@ -711,6 +711,8 @@ i965_validate_config(VADriverContextP ctx,
> >> >> VAProfile profile,
> >> >> if ((HAS_MPEG2_DECODING(i965) && entrypoint ==
> >> >> VAEntrypointVLD) ||
> >> >> (HAS_MPEG2_ENCODING(i965) && entrypoint ==
> >> >> VAEntrypointEncSlice)) {
> >> >> va_status = VA_STATUS_SUCCESS;
> >> >> +} else if (!HAS_MPEG2_DECODING(i965) &&
> >> >> !HAS_MPEG2_ENCODING(i965)){
> >> >> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
> >> >> } else {
> >> >> va_status = VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT;
> >> >> }
> >> >> @@ -723,6 +725,9 @@ i965_validate_config(VADriverContextP ctx,
> >> >>
Re: [Libva] [PATCH][libva-intel-driver] i965_validate_config: return unsupported profile
On Tue, Oct 18, 2016 at 10:48 AM, Eoff, Ullysses A
wrote:
>> -Original Message-
>> From: Charles, Daniel [mailto:daniel.char...@intel.com]
>> Sent: Tuesday, October 18, 2016 10:20 AM
>> To: Eoff, Ullysses A
>> Cc: libva@lists.freedesktop.org
>> Subject: Re: [Libva] [PATCH][libva-intel-driver] i965_validate_config:
>> return unsupported profile
>>
>> On Tue, Oct 18, 2016 at 8:25 AM, Eoff, Ullysses A
>> wrote:
>> > Please run the tests... I suspect this patch will cause some to fail.
>> > Those tests will need updated too.
>> >
>>
>> Ran the test suite on ToT, no regression found with this patch.
>>
>
> Ah yes, I see why now. The AVC and JPEG config tests still pass on most HW
> because UNSUPPORTED_PROFILE is only returned if *both* encode *and* decode
> are unsupported. Thus, if there is such HW that doesn't support one or the
> other then those tests will be regressed.
I am not sure I follow you here.
Unsupported Profile is sent when neither encoder and decoder are
supported (VC1 is only decoder, VPP is a special case and VP9 decoder
can also be hybrid on some h/w). When h/w only supports one of them
(regardless which it is), it should return UNSUPPORTED_ENTRYPOINT as
profile is supported at least once. This patch is not changing this
last scenario. My testing was done on a h/w that has Jpeg Decoder but
no Jpeg Encoder and everything is okay.
--
Daniel.
>
>> --
>> Daniel.
>> > U. Artie
>> >
>> >> -Original Message-
>> >> From: Libva [mailto:libva-boun...@lists.freedesktop.org] On Behalf Of
>> >> Daniel Charles
>> >> Sent: Monday, October 17, 2016 5:45 PM
>> >> To: libva@lists.freedesktop.org
>> >> Subject: [Libva] [PATCH][libva-intel-driver] i965_validate_config: return
>> >> unsupported profile
>> >>
>> >> When all the profiles are not supported return
>> >> VA_STATUS_ERROR_UNSUPPORTED_PROFILE instead of
>> >> VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT.
>> >>
>> >> Also change the style on the code modified to be common
>> >> on all cases
>> >>
>> >> Signed-off-by: Daniel Charles
>> >> ---
>> >> src/i965_drv_video.c | 49
>> >> -
>> >> 1 file changed, 40 insertions(+), 9 deletions(-)
>> >>
>> >> diff --git a/src/i965_drv_video.c b/src/i965_drv_video.c
>> >> index fbf2cda..0830ae0 100644
>> >> --- a/src/i965_drv_video.c
>> >> +++ b/src/i965_drv_video.c
>> >> @@ -711,6 +711,8 @@ i965_validate_config(VADriverContextP ctx, VAProfile
>> >> profile,
>> >> if ((HAS_MPEG2_DECODING(i965) && entrypoint == VAEntrypointVLD)
>> >> ||
>> >> (HAS_MPEG2_ENCODING(i965) && entrypoint ==
>> >> VAEntrypointEncSlice)) {
>> >> va_status = VA_STATUS_SUCCESS;
>> >> +} else if (!HAS_MPEG2_DECODING(i965) &&
>> >> !HAS_MPEG2_ENCODING(i965)){
>> >> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
>> >> } else {
>> >> va_status = VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT;
>> >> }
>> >> @@ -723,6 +725,9 @@ i965_validate_config(VADriverContextP ctx, VAProfile
>> >> profile,
>> >> (HAS_H264_ENCODING(i965) && entrypoint ==
>> >> VAEntrypointEncSlice) ||
>> >> (HAS_LP_H264_ENCODING(i965) && entrypoint ==
>> >> VAEntrypointEncSliceLP)) {
>> >> va_status = VA_STATUS_SUCCESS;
>> >> +} else if (!HAS_H264_DECODING(i965) && !HAS_H264_ENCODING(i965)
>> >> &&
>> >> + !HAS_LP_H264_ENCODING(i965)){
>> >> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
>> >> } else {
>> >> va_status = VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT;
>> >> }
>> >> @@ -733,6 +738,8 @@ i965_validate_config(VADriverContextP ctx, VAProfile
>> >> profile,
>> >> case VAProfileVC1Advanced:
>> >> if (HAS_VC1_DECODING(i965) && entrypoint == VAEntrypointVLD) {
>> >> va_status = VA_STATUS_SUCCESS;
>> >> +} else if (!HAS_VC1_DECODING(i965)) {
>> >> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
>> >> } else {
>> >
Re: [Libva] [PATCH][libva-intel-driver] i965_validate_config: return unsupported profile
> -Original Message-
> From: Charles, Daniel [mailto:daniel.char...@intel.com]
> Sent: Tuesday, October 18, 2016 10:20 AM
> To: Eoff, Ullysses A
> Cc: libva@lists.freedesktop.org
> Subject: Re: [Libva] [PATCH][libva-intel-driver] i965_validate_config: return
> unsupported profile
>
> On Tue, Oct 18, 2016 at 8:25 AM, Eoff, Ullysses A
> wrote:
> > Please run the tests... I suspect this patch will cause some to fail.
> > Those tests will need updated too.
> >
>
> Ran the test suite on ToT, no regression found with this patch.
>
Ah yes, I see why now. The AVC and JPEG config tests still pass on most HW
because UNSUPPORTED_PROFILE is only returned if *both* encode *and* decode are
unsupported. Thus, if there is such HW that doesn't support one or the other
then those tests will be regressed.
> --
> Daniel.
> > U. Artie
> >
> >> -Original Message-
> >> From: Libva [mailto:libva-boun...@lists.freedesktop.org] On Behalf Of
> >> Daniel Charles
> >> Sent: Monday, October 17, 2016 5:45 PM
> >> To: libva@lists.freedesktop.org
> >> Subject: [Libva] [PATCH][libva-intel-driver] i965_validate_config: return
> >> unsupported profile
> >>
> >> When all the profiles are not supported return
> >> VA_STATUS_ERROR_UNSUPPORTED_PROFILE instead of
> >> VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT.
> >>
> >> Also change the style on the code modified to be common
> >> on all cases
> >>
> >> Signed-off-by: Daniel Charles
> >> ---
> >> src/i965_drv_video.c | 49
> >> -
> >> 1 file changed, 40 insertions(+), 9 deletions(-)
> >>
> >> diff --git a/src/i965_drv_video.c b/src/i965_drv_video.c
> >> index fbf2cda..0830ae0 100644
> >> --- a/src/i965_drv_video.c
> >> +++ b/src/i965_drv_video.c
> >> @@ -711,6 +711,8 @@ i965_validate_config(VADriverContextP ctx, VAProfile
> >> profile,
> >> if ((HAS_MPEG2_DECODING(i965) && entrypoint == VAEntrypointVLD) ||
> >> (HAS_MPEG2_ENCODING(i965) && entrypoint ==
> >> VAEntrypointEncSlice)) {
> >> va_status = VA_STATUS_SUCCESS;
> >> +} else if (!HAS_MPEG2_DECODING(i965) &&
> >> !HAS_MPEG2_ENCODING(i965)){
> >> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
> >> } else {
> >> va_status = VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT;
> >> }
> >> @@ -723,6 +725,9 @@ i965_validate_config(VADriverContextP ctx, VAProfile
> >> profile,
> >> (HAS_H264_ENCODING(i965) && entrypoint ==
> >> VAEntrypointEncSlice) ||
> >> (HAS_LP_H264_ENCODING(i965) && entrypoint ==
> >> VAEntrypointEncSliceLP)) {
> >> va_status = VA_STATUS_SUCCESS;
> >> +} else if (!HAS_H264_DECODING(i965) && !HAS_H264_ENCODING(i965) &&
> >> + !HAS_LP_H264_ENCODING(i965)){
> >> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
> >> } else {
> >> va_status = VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT;
> >> }
> >> @@ -733,6 +738,8 @@ i965_validate_config(VADriverContextP ctx, VAProfile
> >> profile,
> >> case VAProfileVC1Advanced:
> >> if (HAS_VC1_DECODING(i965) && entrypoint == VAEntrypointVLD) {
> >> va_status = VA_STATUS_SUCCESS;
> >> +} else if (!HAS_VC1_DECODING(i965)) {
> >> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
> >> } else {
> >> va_status = VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT;
> >> }
> >> @@ -741,6 +748,8 @@ i965_validate_config(VADriverContextP ctx, VAProfile
> >> profile,
> >> case VAProfileNone:
> >> if (HAS_VPP(i965) && VAEntrypointVideoProc == entrypoint) {
> >> va_status = VA_STATUS_SUCCESS;
> >> +} else if (!HAS_VPP(i965)){
> >> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
> >> } else {
> >> va_status = VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT;
> >> }
> >> @@ -750,6 +759,8 @@ i965_validate_config(VADriverContextP ctx, VAProfile
> >> profile,
> >> if ((HAS_JPEG_DECODING(i965) && entrypoint == VAEntrypointVLD) ||
> >> (HAS_JPEG_ENCODING(i965) &
Re: [Libva] [PATCH][libva-intel-driver] i965_validate_config: return unsupported profile
On Tue, Oct 18, 2016 at 8:25 AM, Eoff, Ullysses A
wrote:
> Please run the tests... I suspect this patch will cause some to fail. Those
> tests will need updated too.
>
Ran the test suite on ToT, no regression found with this patch.
--
Daniel.
> U. Artie
>
>> -Original Message-
>> From: Libva [mailto:libva-boun...@lists.freedesktop.org] On Behalf Of Daniel
>> Charles
>> Sent: Monday, October 17, 2016 5:45 PM
>> To: libva@lists.freedesktop.org
>> Subject: [Libva] [PATCH][libva-intel-driver] i965_validate_config: return
>> unsupported profile
>>
>> When all the profiles are not supported return
>> VA_STATUS_ERROR_UNSUPPORTED_PROFILE instead of
>> VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT.
>>
>> Also change the style on the code modified to be common
>> on all cases
>>
>> Signed-off-by: Daniel Charles
>> ---
>> src/i965_drv_video.c | 49 -
>> 1 file changed, 40 insertions(+), 9 deletions(-)
>>
>> diff --git a/src/i965_drv_video.c b/src/i965_drv_video.c
>> index fbf2cda..0830ae0 100644
>> --- a/src/i965_drv_video.c
>> +++ b/src/i965_drv_video.c
>> @@ -711,6 +711,8 @@ i965_validate_config(VADriverContextP ctx, VAProfile
>> profile,
>> if ((HAS_MPEG2_DECODING(i965) && entrypoint == VAEntrypointVLD) ||
>> (HAS_MPEG2_ENCODING(i965) && entrypoint ==
>> VAEntrypointEncSlice)) {
>> va_status = VA_STATUS_SUCCESS;
>> +} else if (!HAS_MPEG2_DECODING(i965) && !HAS_MPEG2_ENCODING(i965)){
>> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
>> } else {
>> va_status = VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT;
>> }
>> @@ -723,6 +725,9 @@ i965_validate_config(VADriverContextP ctx, VAProfile
>> profile,
>> (HAS_H264_ENCODING(i965) && entrypoint == VAEntrypointEncSlice)
>> ||
>> (HAS_LP_H264_ENCODING(i965) && entrypoint ==
>> VAEntrypointEncSliceLP)) {
>> va_status = VA_STATUS_SUCCESS;
>> +} else if (!HAS_H264_DECODING(i965) && !HAS_H264_ENCODING(i965) &&
>> + !HAS_LP_H264_ENCODING(i965)){
>> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
>> } else {
>> va_status = VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT;
>> }
>> @@ -733,6 +738,8 @@ i965_validate_config(VADriverContextP ctx, VAProfile
>> profile,
>> case VAProfileVC1Advanced:
>> if (HAS_VC1_DECODING(i965) && entrypoint == VAEntrypointVLD) {
>> va_status = VA_STATUS_SUCCESS;
>> +} else if (!HAS_VC1_DECODING(i965)) {
>> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
>> } else {
>> va_status = VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT;
>> }
>> @@ -741,6 +748,8 @@ i965_validate_config(VADriverContextP ctx, VAProfile
>> profile,
>> case VAProfileNone:
>> if (HAS_VPP(i965) && VAEntrypointVideoProc == entrypoint) {
>> va_status = VA_STATUS_SUCCESS;
>> +} else if (!HAS_VPP(i965)){
>> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
>> } else {
>> va_status = VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT;
>> }
>> @@ -750,6 +759,8 @@ i965_validate_config(VADriverContextP ctx, VAProfile
>> profile,
>> if ((HAS_JPEG_DECODING(i965) && entrypoint == VAEntrypointVLD) ||
>> (HAS_JPEG_ENCODING(i965) && entrypoint ==
>> VAEntrypointEncPicture)) {
>> va_status = VA_STATUS_SUCCESS;
>> +} else if (!HAS_JPEG_DECODING(i965) && !HAS_JPEG_ENCODING(i965)){
>> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
>> } else {
>> va_status = VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT;
>> }
>> @@ -759,6 +770,8 @@ i965_validate_config(VADriverContextP ctx, VAProfile
>> profile,
>> if ((HAS_VP8_DECODING(i965) && entrypoint == VAEntrypointVLD) ||
>> (HAS_VP8_ENCODING(i965) && entrypoint == VAEntrypointEncSlice))
>> {
>> va_status = VA_STATUS_SUCCESS;
>> +} else if (!HAS_VP8_DECODING(i965) && !HAS_VP8_ENCODING(i965)){
>> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
>> } else {
>> va_status = VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT;
>> }
>> @@ -768,8 +781,12 @@ i965_validate_config(VADriverContextP ctx, VAProfile
>> profile,
>> case VAProfileH264StereoHigh:
>> if ((HAS_H264_MVC_DECODING_PROFILE(i965, profile) &&
>> entrypoint == VAEntrypointVLD) ||
>> -(HAS_H264_MVC_ENCODING(i965) && entrypoint ==
>> VAEntrypointEncSlice)) {
>> +(HAS_H264_MVC_ENCODING(i965) &&
>> + entrypoint == VAEntrypointEncSlice)) {
>> va_status = VA_STATUS_SUCCESS;
>> +} else if(!HAS_H264_MVC_DECODING_PROFILE(i965, profile) &&
>> + !HAS_H264_MVC_ENCODING(i965)) {
>> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
>> } else {
>> v
Re: [Libva] [PATCH][libva-intel-driver] i965_validate_config: return unsupported profile
Please run the tests... I suspect this patch will cause some to fail. Those
tests will need updated too.
U. Artie
> -Original Message-
> From: Libva [mailto:libva-boun...@lists.freedesktop.org] On Behalf Of Daniel
> Charles
> Sent: Monday, October 17, 2016 5:45 PM
> To: libva@lists.freedesktop.org
> Subject: [Libva] [PATCH][libva-intel-driver] i965_validate_config: return
> unsupported profile
>
> When all the profiles are not supported return
> VA_STATUS_ERROR_UNSUPPORTED_PROFILE instead of
> VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT.
>
> Also change the style on the code modified to be common
> on all cases
>
> Signed-off-by: Daniel Charles
> ---
> src/i965_drv_video.c | 49 -
> 1 file changed, 40 insertions(+), 9 deletions(-)
>
> diff --git a/src/i965_drv_video.c b/src/i965_drv_video.c
> index fbf2cda..0830ae0 100644
> --- a/src/i965_drv_video.c
> +++ b/src/i965_drv_video.c
> @@ -711,6 +711,8 @@ i965_validate_config(VADriverContextP ctx, VAProfile
> profile,
> if ((HAS_MPEG2_DECODING(i965) && entrypoint == VAEntrypointVLD) ||
> (HAS_MPEG2_ENCODING(i965) && entrypoint ==
> VAEntrypointEncSlice)) {
> va_status = VA_STATUS_SUCCESS;
> +} else if (!HAS_MPEG2_DECODING(i965) && !HAS_MPEG2_ENCODING(i965)){
> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
> } else {
> va_status = VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT;
> }
> @@ -723,6 +725,9 @@ i965_validate_config(VADriverContextP ctx, VAProfile
> profile,
> (HAS_H264_ENCODING(i965) && entrypoint == VAEntrypointEncSlice)
> ||
> (HAS_LP_H264_ENCODING(i965) && entrypoint ==
> VAEntrypointEncSliceLP)) {
> va_status = VA_STATUS_SUCCESS;
> +} else if (!HAS_H264_DECODING(i965) && !HAS_H264_ENCODING(i965) &&
> + !HAS_LP_H264_ENCODING(i965)){
> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
> } else {
> va_status = VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT;
> }
> @@ -733,6 +738,8 @@ i965_validate_config(VADriverContextP ctx, VAProfile
> profile,
> case VAProfileVC1Advanced:
> if (HAS_VC1_DECODING(i965) && entrypoint == VAEntrypointVLD) {
> va_status = VA_STATUS_SUCCESS;
> +} else if (!HAS_VC1_DECODING(i965)) {
> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
> } else {
> va_status = VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT;
> }
> @@ -741,6 +748,8 @@ i965_validate_config(VADriverContextP ctx, VAProfile
> profile,
> case VAProfileNone:
> if (HAS_VPP(i965) && VAEntrypointVideoProc == entrypoint) {
> va_status = VA_STATUS_SUCCESS;
> +} else if (!HAS_VPP(i965)){
> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
> } else {
> va_status = VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT;
> }
> @@ -750,6 +759,8 @@ i965_validate_config(VADriverContextP ctx, VAProfile
> profile,
> if ((HAS_JPEG_DECODING(i965) && entrypoint == VAEntrypointVLD) ||
> (HAS_JPEG_ENCODING(i965) && entrypoint ==
> VAEntrypointEncPicture)) {
> va_status = VA_STATUS_SUCCESS;
> +} else if (!HAS_JPEG_DECODING(i965) && !HAS_JPEG_ENCODING(i965)){
> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
> } else {
> va_status = VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT;
> }
> @@ -759,6 +770,8 @@ i965_validate_config(VADriverContextP ctx, VAProfile
> profile,
> if ((HAS_VP8_DECODING(i965) && entrypoint == VAEntrypointVLD) ||
> (HAS_VP8_ENCODING(i965) && entrypoint == VAEntrypointEncSlice)) {
> va_status = VA_STATUS_SUCCESS;
> +} else if (!HAS_VP8_DECODING(i965) && !HAS_VP8_ENCODING(i965)){
> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
> } else {
> va_status = VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT;
> }
> @@ -768,8 +781,12 @@ i965_validate_config(VADriverContextP ctx, VAProfile
> profile,
> case VAProfileH264StereoHigh:
> if ((HAS_H264_MVC_DECODING_PROFILE(i965, profile) &&
> entrypoint == VAEntrypointVLD) ||
> -(HAS_H264_MVC_ENCODING(i965) && entrypoint ==
> VAEntrypointEncSlice)) {
> +(HAS_H264_MVC_ENCODING(i965) &&
> + entrypoint == VAEntrypointEncSlice)) {
> va_status = VA_STATUS_SUCCESS;
> +} else if(!HAS_H264_MVC_DECODING_PROFILE(i965, profile) &&
> + !HAS_H264_MVC_ENCODING(i965)) {
> +va_status = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
> } else {
> va_status = VA_STATUS_ERROR_UNSUPPORTED_ENTRYPOINT;
> }
> @@ -778,32 +795,46 @@ i965_validate_config(VADriverContextP ctx, VAProfile
> profile,
>
> case VAProfileHEVCMain:
> if ((HAS_HEVC_DECODING(i965) && (entrypoint == VAE
