Re: [Mesa-dev] [PATCH 0/4] Update to GL_ARB_timer_query
Hi Marek, Please update src/gallium/docs for the gallium interface change, and implement the new function on galahad (I'm working towards getting galahad to support the full gallium interface and get it enabled by default on debug builds, as it is a shame not using it more). Otherwise looks good to me. Jose - Original Message - > Hi everyone, > > this series is a follow-up to the previous one sent on June 27. > Here's only what's changed. All commits are here: > http://cgit.freedesktop.org/~mareko/mesa/log/?h=timer-query > > Change log: > - reimplemented glGet(GL_TIMESTAMP) as per Eric Anholt's comment > - updated all gallium patches > - added softpipe support > > The craziness with the cpu-gpu difference is gone. Softpipe > (specifically the os_time_get function) uses gettimeofday instead of > the proposed xserver's GetTimeInMillis function, but I don't > consider that a big deal (especially if it's only softpipe using > it). > > Please review. > > Marek Olšák (4): > mesa: implement glGet(GL_TIMESTAMP) v2 > gallium: add QUERY_TIMESTAMP cap and get_timestamp screen > function > st/mesa: implement ARB_timer_query > softpipe: implement get_timestamp and expose ARB_timer_query > > src/gallium/drivers/i915/i915_screen.c |1 + > src/gallium/drivers/llvmpipe/lp_screen.c |1 + > src/gallium/drivers/nv30/nv30_screen.c |1 + > src/gallium/drivers/nv50/nv50_screen.c |1 + > src/gallium/drivers/nvc0/nvc0_screen.c |1 + > src/gallium/drivers/r300/r300_screen.c |1 + > src/gallium/drivers/r600/r600_pipe.c |1 + > src/gallium/drivers/softpipe/sp_screen.c | 10 ++ > src/gallium/drivers/svga/svga_screen.c |1 + > src/gallium/include/pipe/p_defines.h |3 ++- > src/gallium/include/pipe/p_screen.h |7 +++ > src/mesa/main/dd.h |6 ++ > src/mesa/main/get.c | 17 - > src/mesa/state_tracker/st_cb_queryobj.c | 14 ++ > src/mesa/state_tracker/st_extensions.c |5 + > 15 files changed, 68 insertions(+), 2 deletions(-) > > Marek > ___ > mesa-dev mailing list > mesa-dev@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/mesa-dev > ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH 04/22] gallium: add util_format_stencil helper function
"util_format_stencil" seems a too generic. Some sort of qualifier would be more
future proof, e.g., util_format_stencil_only.
- Original Message -
> used for stencil sampler views.
> ---
> src/gallium/auxiliary/util/u_format.h | 29
> +
> src/mesa/state_tracker/st_cb_drawpixels.c | 23
> ++-
> 2 files changed, 31 insertions(+), 21 deletions(-)
>
> diff --git a/src/gallium/auxiliary/util/u_format.h
> b/src/gallium/auxiliary/util/u_format.h
> index e35e164..7995b57 100644
> --- a/src/gallium/auxiliary/util/u_format.h
> +++ b/src/gallium/auxiliary/util/u_format.h
> @@ -882,6 +882,35 @@ util_format_linear(enum pipe_format format)
> }
>
> /**
> + * Given a depth-stencil format, return the corresponding
> stencil-only format.
> + * For stencil-only formats, return the format unchanged.
> + */
> +static INLINE enum pipe_format
> +util_format_stencil(enum pipe_format format)
> +{
> + switch (format) {
> + /* mask out the depth component */
> + case PIPE_FORMAT_Z24_UNORM_S8_UINT:
> + return PIPE_FORMAT_X24S8_UINT;
> + case PIPE_FORMAT_S8_UINT_Z24_UNORM:
> + return PIPE_FORMAT_S8X24_UINT;
> + case PIPE_FORMAT_Z32_FLOAT_S8X24_UINT:
> + return PIPE_FORMAT_X32_S8X24_UINT;
> +
> + /* stencil only formats */
> + case PIPE_FORMAT_X24S8_UINT:
> + case PIPE_FORMAT_S8X24_UINT:
> + case PIPE_FORMAT_X32_S8X24_UINT:
> + case PIPE_FORMAT_S8_UINT:
> + return format;
> +
> + default:
> + assert(0);
> + return PIPE_FORMAT_NONE;
> + }
> +}
> +
> +/**
> * Return the number of components stored.
> * Formats with block size != 1x1 will always have 1 component (the
> block).
> */
> diff --git a/src/mesa/state_tracker/st_cb_drawpixels.c
> b/src/mesa/state_tracker/st_cb_drawpixels.c
> index 2bcbada..727273b 100644
> --- a/src/mesa/state_tracker/st_cb_drawpixels.c
> +++ b/src/mesa/state_tracker/st_cb_drawpixels.c
> @@ -1165,27 +1165,8 @@ st_DrawPixels(struct gl_context *ctx, GLint x,
> GLint y,
> * The stencil is written using the shader stencil
> export
> * functionality. */
> if (write_stencil) {
> - enum pipe_format stencil_format = PIPE_FORMAT_NONE;
> -
> - switch (pt->format) {
> - case PIPE_FORMAT_Z24_UNORM_S8_UINT:
> - case PIPE_FORMAT_X24S8_UINT:
> - stencil_format = PIPE_FORMAT_X24S8_UINT;
> - break;
> - case PIPE_FORMAT_S8_UINT_Z24_UNORM:
> - case PIPE_FORMAT_S8X24_UINT:
> - stencil_format = PIPE_FORMAT_S8X24_UINT;
> - break;
> - case PIPE_FORMAT_Z32_FLOAT_S8X24_UINT:
> - case PIPE_FORMAT_X32_S8X24_UINT:
> - stencil_format = PIPE_FORMAT_X32_S8X24_UINT;
> - break;
> - case PIPE_FORMAT_S8_UINT:
> - stencil_format = PIPE_FORMAT_S8_UINT;
> - break;
> - default:
> - assert(0);
> - }
> + enum pipe_format stencil_format =
> + util_format_stencil(pt->format);
>
> sv[1] =
> st_create_texture_sampler_view_format(st->pipe, pt,
> stencil_format);
> --
> 1.7.9.5
>
> ___
> mesa-dev mailing list
> mesa-dev@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/mesa-dev
>
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH 22/22] r600g: fix uploading non-zero mipmap levels of depth textures
On Mon, 2012-07-09 at 21:15 +0200, Marek Olšák wrote: > This fixes piglit/depth-level-clamp. > --- > src/gallium/drivers/r600/r600_texture.c |5 +++-- > 1 file changed, 3 insertions(+), 2 deletions(-) > > diff --git a/src/gallium/drivers/r600/r600_texture.c > b/src/gallium/drivers/r600/r600_texture.c > index a6a83ca..d16c252 100644 > --- a/src/gallium/drivers/r600/r600_texture.c > +++ b/src/gallium/drivers/r600/r600_texture.c > @@ -926,8 +926,9 @@ void r600_texture_transfer_destroy(struct pipe_context > *ctx, > > u_box_origin_2d(texture->width0, texture->height0, > &sbox); > > - ctx->resource_copy_region(ctx, texture, 0, 0, 0, 0, > - &rtransfer->staging->b.b, 0, > + ctx->resource_copy_region(ctx, texture, transfer->level, > + 0, 0, transfer->box.z, > + &rtransfer->staging->b.b, > transfer->level, Does this really need to take into account transfer->box.z, but not .x and .y? -- Earthling Michel Dänzer | http://www.amd.com Libre software enthusiast | Debian, X and DRI developer ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH 4/9] automake: convert libmesa and libmesagallium
On 09/07/2012 17:31, Eric Anholt wrote: > Jon TURNEY writes: >> diff --git a/src/mesa/Makefile.am b/src/mesa/Makefile.am >> index bada760..08beaa6 100644 >> --- a/src/mesa/Makefile.am >> +++ b/src/mesa/Makefile.am > >> +libmesa_la_SOURCES = \ >> +$(MESA_FILES) \ >> +$(MESA_CXX_FILES) \ >> +$(MESA_ASM_FILES_FOR_ARCH) >> + >> +libmesa_la_LIBADD = $(top_builddir)/src/glsl/libglsl.la >> +libmesa_la_LDFLAGS = -static > > Doesn't this -static result in relocations for static libs getting > built, which would then be a problem for the linking into shared > libraries later? I don't really know this stuff well, but a -static in > the middle of a shared lib build seems strange. > > The helper lib for the asm code in libdricore built a .a file without > -static being specified. Yes, this is entirely wrong. I think I added the '-static' when reading 'mklib -static' and never came back to change it to a convenience library. > Also, with this commit, did you test that git_sha1 depencies were > working correctly and commit --amending would produce a new sha1 in your > drivers? This appears to work for me. I can see I have removed a dependency of depend: on git_sha1.h, but recomputing the dependencies every time that file changes seems a bit of overkill to me. The computed dependencies still include git_sha1.h and things which depend on it are rebuilt when it changes. ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH 9/9] Don't explicitly link libOsmesa with libmesa's dependency libglapi
On 09/07/2012 17:36, Eric Anholt wrote: >> We must either: >> (i) Not link libmesa with libglapi, and link anything that uses libmesa with >> libglapi as well, or >> (ii) Link libmesa with libglapi, and not link anything that uses libmesa with >> libglapi >> >> I choose (ii) just because it's least change, but I'm not sure it's right. >> >> Note that drivers/X11 makes libGL by linking with libmesa, but not libglapi, >> so it's a bit unclear to me how this was working correctly before in both >> cases >> (It might be some sort of accidental side-effect of what mklib's >> expand_archives() >> function does that this ever worked before?) >> --- >> src/mesa/drivers/osmesa/Makefile.am |3 +-- >> 1 files changed, 1 insertions(+), 2 deletions(-) >> >> diff --git a/src/mesa/drivers/osmesa/Makefile.am >> b/src/mesa/drivers/osmesa/Makefile.am >> index a91d9e2..cfba658 100644 >> --- a/src/mesa/drivers/osmesa/Makefile.am >> +++ b/src/mesa/drivers/osmesa/Makefile.am >> @@ -40,8 +40,7 @@ lib@OSMESA_LIB@_la_SOURCES = osmesa.c >> lib@OSMESA_LIB@_la_LDFLAGS = -module -version-number @OSMESA_VERSION@ >> -shared >> lib@OSMESA_LIB@_la_LIBADD = \ >> $(top_srcdir)/src/mesa/libmesa.la \ >> -$(top_srcdir)/src/mapi/glapi/libglapi.la \ >> -$(top_srcdir)/src/glsl/libglsl.la >> +$(top_srcdir)/src/mapi/glapi/libglapi.la > > This commit actually looks like it drops linking with libglsl. What's > up? Sorry, I had some kind of brain spasm when writing the comment for this commit, it should of course say libglsl everywhere it says libglapi. The libmesa convenience library is linked with the libglsl convenience library. libOsmesa is linked with libmesa, and also directly with libglsl. This gives rise to duplicate symbol errors. ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] Mesa (master): draw: fix flat shading and screen-space linear interpolation in clipper
Yep. The interfaces are busted. Without native integers we get DCL SV[0], INSTANCEID ... ARL ADDR[0].x, SV[0]. and with integers we get DCL SV[0], INSTANCEID ... UARL ADDR[0].x, SV[0]. Olivier's fix is incorrect. It works on the above cases by guessing the type, but if we have: DCL SV[0], INSTANCEID ... MOV TEMP[0].x, SV[0]. ARL ADDR[0].x, TEMP[0]. and DCL SV[0], INSTANCEID ... MOV TEMP[0].x, SV[0]. UARL ADDR[0].x, TEMP[0]. it is impossible to guess -- the change is merely replacing a bug with another. For the record, the problem also happens without LLVM: DRAW_USE_LLVM=0 draw-instanced -auto Anyway, AFAICT, all hardware out there that really supports INSTANCEID/VERTEXID also supports native integers, so this is a problem specific to the draw module. But given that draw module can support anything, this is actually self inflicted! In short, INSTANCEID/VERTEXID without integers is an historic artifact, that should not exist going forward. The right fix is merely making sure that PIPE_SHADER_CAP_INTEGERS is accurately advertised as 1 by the draw module (just like Stephane secretly did in 45fc069600ddbfe07a0a0cd5280161a8c7c55dd0 :) I'm working on a patch that does this, and tries to cleanup draw_get_shader_param() too. Jose - Original Message - > FWIW, I think Draw needs to know if a driver does or does not support > integers, so that it can return system values in the correct type > (not > sure if that includes gallivm too). The reason for that is that > glsl_to_tgsi generates *different* shaders depending on the native > integer cap. I think this is good; the drivers which don't have > integer system values and integer instructions just get shaders > prepared to be executed on floating-point units only. > > However the aforementioned modification of Draw would only be > required > if there were a driver using Draw for fallbacks, i.e. if a driver > doesn't know in advance whether a shader will be executed in software > or hardware (that's only svga, I guess?). > > The software drivers and the drivers using Draw on hardware without > vertex shaders can always use Draw with full integer support (and > glsl_to_tgsi should generate vertex shaders with integers*) and > nothing would need to be changed. Such drivers should return > draw_get_shader_param(..) in get_shader_param, so that the state > tracker knows exactly what the Draw module expects. > > * I just realized there is a bug in glsl_to_tgsi which doesn't follow > the integer shader cap correctly. I'll send patches. > > Marek > > On Tue, Jul 3, 2012 at 10:43 PM, Olivier Galibert > wrote: > > On Tue, Jul 03, 2012 at 12:39:47PM -0700, Jose Fonseca wrote: > >> Note that all registers are stored as floats (for convenience, and > >> because LLVM has no unions), so integers are bitcasted into floats > >> while storing/loading. And I'm not sure if your patch would break > >> that. > > > > I did test the patch with a llvmpipe in a glsl 120/no native > > integer > > setup. draw_instanced worked. I didn't try a full piglit though. > > > > > >> I still think that having draw/gallivm guessing whether native > >> integer support is intended or not is bad. Either: > >> 1) TGSI is extended (e.g., more type annotations) so that > >> native-integer support can inferred from it > >> 2) draw/gallivm need to now if the driver has native-integer or > >> not > >> > >> I'm inclined towards 1), as TGSI should be self-documented. That > >> is, > >> it should not be necessary to know if the driver has or not native > >> integer support to know whether system values should be assumed to > >> be integers or floats... > > > > It could be argued that dtype being TGSI_TYPE_FLOAT is the > > documentation on what is expected. But I'm quickly reaching the > > point > > where I don't really care, just tell me what you want. As long as > > textureFetch stays the only issue between llvmpipe and 1.30 I'm ok. > > > > Of course doing textureFetch right is going to require an > > interesting > > overhaul of the texture allocations... need to finish fixing the > > gm45 > > interpolation/clipping first. > > > > Best, > > > > OG. > > > > ___ > > mesa-dev mailing list > > mesa-dev@lists.freedesktop.org > > http://lists.freedesktop.org/mailman/listinfo/mesa-dev > ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] Mesa (master): draw: fix flat shading and screen-space linear interpolation in clipper
On Tue, Jul 10, 2012 at 3:30 PM, Jose Fonseca wrote: > Yep. The interfaces are busted. > > Without native integers we get > >DCL SV[0], INSTANCEID >... >ARL ADDR[0].x, SV[0]. > > and with integers we get > >DCL SV[0], INSTANCEID >... >UARL ADDR[0].x, SV[0]. > > Olivier's fix is incorrect. It works on the above cases by guessing the type, > but if we have: > >DCL SV[0], INSTANCEID >... >MOV TEMP[0].x, SV[0]. >ARL ADDR[0].x, TEMP[0]. > > and > >DCL SV[0], INSTANCEID >... >MOV TEMP[0].x, SV[0]. >UARL ADDR[0].x, TEMP[0]. > > it is impossible to guess -- the change is merely replacing a bug with > another. > > > For the record, the problem also happens without LLVM: > > DRAW_USE_LLVM=0 draw-instanced -auto > > > > Anyway, AFAICT, all hardware out there that really supports > INSTANCEID/VERTEXID also supports native integers, so this is a problem > specific to the draw module. > > But given that draw module can support anything, this is actually self > inflicted! In short, INSTANCEID/VERTEXID without integers is an historic > artifact, that should not exist going forward. > > The right fix is merely making sure that PIPE_SHADER_CAP_INTEGERS is > accurately advertised as 1 by the draw module (just like Stephane secretly > did in 45fc069600ddbfe07a0a0cd5280161a8c7c55dd0 :) I just wanted to tell you Stephane's change cannot work and it even has no effect at the moment. The native integer support is global in core Mesa. It's because integer uniforms are converted to floats based on the global NativeInteger flag for all shader stages and that can't be fixed easily, because uniforms can be shared between shaders. Basically, all drivers must advertise integer support either for all shader stages or none. Given that, I only see two possible outcomes: 1) Disable INSTANCEID support in DX9-level drivers using Draw (that's only i915g AFAIK) and only support INSTANCEID with integers. 2) Let drivers set the type of system values in Draw, so that Draw doesn't have to guess what it should be. I'm okay with either solution. Marek ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH 04/22] gallium: add util_format_stencil helper function
On Tue, Jul 10, 2012 at 8:29 AM, Jose Fonseca wrote:
> "util_format_stencil" seems a too generic. Some sort of qualifier would be
> more future proof, e.g., util_format_stencil_only.
Alright, consider it done. (without re-sending the patches)
Marek
>
> - Original Message -
>> used for stencil sampler views.
>> ---
>> src/gallium/auxiliary/util/u_format.h | 29
>> +
>> src/mesa/state_tracker/st_cb_drawpixels.c | 23
>> ++-
>> 2 files changed, 31 insertions(+), 21 deletions(-)
>>
>> diff --git a/src/gallium/auxiliary/util/u_format.h
>> b/src/gallium/auxiliary/util/u_format.h
>> index e35e164..7995b57 100644
>> --- a/src/gallium/auxiliary/util/u_format.h
>> +++ b/src/gallium/auxiliary/util/u_format.h
>> @@ -882,6 +882,35 @@ util_format_linear(enum pipe_format format)
>> }
>>
>> /**
>> + * Given a depth-stencil format, return the corresponding
>> stencil-only format.
>> + * For stencil-only formats, return the format unchanged.
>> + */
>> +static INLINE enum pipe_format
>> +util_format_stencil(enum pipe_format format)
>> +{
>> + switch (format) {
>> + /* mask out the depth component */
>> + case PIPE_FORMAT_Z24_UNORM_S8_UINT:
>> + return PIPE_FORMAT_X24S8_UINT;
>> + case PIPE_FORMAT_S8_UINT_Z24_UNORM:
>> + return PIPE_FORMAT_S8X24_UINT;
>> + case PIPE_FORMAT_Z32_FLOAT_S8X24_UINT:
>> + return PIPE_FORMAT_X32_S8X24_UINT;
>> +
>> + /* stencil only formats */
>> + case PIPE_FORMAT_X24S8_UINT:
>> + case PIPE_FORMAT_S8X24_UINT:
>> + case PIPE_FORMAT_X32_S8X24_UINT:
>> + case PIPE_FORMAT_S8_UINT:
>> + return format;
>> +
>> + default:
>> + assert(0);
>> + return PIPE_FORMAT_NONE;
>> + }
>> +}
>> +
>> +/**
>> * Return the number of components stored.
>> * Formats with block size != 1x1 will always have 1 component (the
>> block).
>> */
>> diff --git a/src/mesa/state_tracker/st_cb_drawpixels.c
>> b/src/mesa/state_tracker/st_cb_drawpixels.c
>> index 2bcbada..727273b 100644
>> --- a/src/mesa/state_tracker/st_cb_drawpixels.c
>> +++ b/src/mesa/state_tracker/st_cb_drawpixels.c
>> @@ -1165,27 +1165,8 @@ st_DrawPixels(struct gl_context *ctx, GLint x,
>> GLint y,
>> * The stencil is written using the shader stencil
>> export
>> * functionality. */
>> if (write_stencil) {
>> - enum pipe_format stencil_format = PIPE_FORMAT_NONE;
>> -
>> - switch (pt->format) {
>> - case PIPE_FORMAT_Z24_UNORM_S8_UINT:
>> - case PIPE_FORMAT_X24S8_UINT:
>> - stencil_format = PIPE_FORMAT_X24S8_UINT;
>> - break;
>> - case PIPE_FORMAT_S8_UINT_Z24_UNORM:
>> - case PIPE_FORMAT_S8X24_UINT:
>> - stencil_format = PIPE_FORMAT_S8X24_UINT;
>> - break;
>> - case PIPE_FORMAT_Z32_FLOAT_S8X24_UINT:
>> - case PIPE_FORMAT_X32_S8X24_UINT:
>> - stencil_format = PIPE_FORMAT_X32_S8X24_UINT;
>> - break;
>> - case PIPE_FORMAT_S8_UINT:
>> - stencil_format = PIPE_FORMAT_S8_UINT;
>> - break;
>> - default:
>> - assert(0);
>> - }
>> + enum pipe_format stencil_format =
>> + util_format_stencil(pt->format);
>>
>> sv[1] =
>> st_create_texture_sampler_view_format(st->pipe, pt,
>>
>> stencil_format);
>> --
>> 1.7.9.5
>>
>> ___
>> mesa-dev mailing list
>> mesa-dev@lists.freedesktop.org
>> http://lists.freedesktop.org/mailman/listinfo/mesa-dev
>>
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [PATCH 0/9] Automake continued (v2)
Updated to address various comments: [1/9] libOsmesa links with libaries in the builddir, not the srcdir [2/9] libglsl use AM_V_GEN to silence generated code rules and add BUILT_SOURCES to CLEANFILES [4/9] Remove stray -static from flags for libmesa and libmesagallium [6/9] Remove unndeded libglsl.a from GALLIUM_DRI_LIB_DEPS [9/9] Fix comment to talk about the right library and try to make sense If you should want to test this, these patches can be pulled from the branch automake-continued-v2 at git://people.freedesktop.org/~jturney/mesa Jon TURNEY (8): automake: convert libglsl Rename sparc/clip.S -> sparc/sparc_clip.S automake: convert libmesa and libmesagallium Remove unused MESA_MODULES autoconf variable Link dri drivers with mesa or dricore libtool library Rename X86-64_API -> X86_64_API automake: convert libglapi Don't explicitly link libOsmesa with libmesa's dependency libglsl Laurent Carlier (1): automake: convert libOSmesa building configs/current.in |3 +- configure.ac| 27 +++-- src/egl/main/Makefile.am|2 +- src/gallium/targets/Makefile.dri|4 +- src/glsl/.gitignore |1 + src/glsl/Makefile | 144 - src/glsl/Makefile.am| 112 + src/glsl/glcpp/.gitignore |1 + src/glsl/glcpp/Makefile.am | 33 + src/glsl/tests/Makefile.am |4 +- src/glx/Makefile.am |2 +- src/mapi/glapi/.gitignore |1 + src/mapi/glapi/Makefile | 77 src/mapi/glapi/Makefile.am | 69 ++ src/mapi/glapi/sources.mak |2 +- src/mapi/glapi/tests/Makefile.am|2 +- src/mesa/Makefile.am| 45 ++- src/mesa/Makefile.old | 57 +- src/mesa/SConscript |2 +- src/mesa/drivers/osmesa/.gitignore |1 + src/mesa/drivers/osmesa/Makefile| 51 src/mesa/drivers/osmesa/Makefile.am | 49 src/mesa/drivers/x11/Makefile.am|4 +- src/mesa/main/tests/Makefile.am |2 +- src/mesa/sources.mak|2 +- src/mesa/sparc/clip.S | 233 --- src/mesa/sparc/sparc_clip.S | 233 +++ 27 files changed, 572 insertions(+), 591 deletions(-) delete mode 100644 src/glsl/Makefile create mode 100644 src/glsl/Makefile.am create mode 100644 src/glsl/glcpp/Makefile.am delete mode 100644 src/mapi/glapi/Makefile create mode 100644 src/mapi/glapi/Makefile.am create mode 100644 src/mesa/drivers/osmesa/.gitignore delete mode 100644 src/mesa/drivers/osmesa/Makefile create mode 100644 src/mesa/drivers/osmesa/Makefile.am delete mode 100644 src/mesa/sparc/clip.S create mode 100644 src/mesa/sparc/sparc_clip.S -- 1.7.9 ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [PATCH 1/9] automake: convert libOSmesa building
From: Laurent Carlier This also currently fix the installation of libOSmesa. v2: Remove old Makefile, libOSmesa is now versioned, fix typos v3: Keep config substitution alphabetized v4: Update .gitignore v5: Libraries will be in the builddir, not the srcdir. Reviewed-by: Jon TURNEY --- configure.ac|5 +++ src/mesa/drivers/osmesa/.gitignore |1 + src/mesa/drivers/osmesa/Makefile| 51 --- src/mesa/drivers/osmesa/Makefile.am | 50 ++ 4 files changed, 56 insertions(+), 51 deletions(-) create mode 100644 src/mesa/drivers/osmesa/.gitignore delete mode 100644 src/mesa/drivers/osmesa/Makefile create mode 100644 src/mesa/drivers/osmesa/Makefile.am diff --git a/configure.ac b/configure.ac index 46265a2..e5ac791 100644 --- a/configure.ac +++ b/configure.ac @@ -1370,10 +1370,14 @@ if test "x$enable_osmesa" = xyes; then OSMESA_MESA_DEPS="" OSMESA_PC_LIB_PRIV="-lm -lpthread $SELINUX_LIBS $DLOPEN_LIBS" fi + +OSMESA_VERSION=`echo "$VERSION" | $SED 's/\./:/g'` + AC_SUBST([OSMESA_LIB_DEPS]) AC_SUBST([OSMESA_MESA_DEPS]) AC_SUBST([OSMESA_PC_REQ]) AC_SUBST([OSMESA_PC_LIB_PRIV]) +AC_SUBST([OSMESA_VERSION]) dnl dnl gbm configuration @@ -2188,6 +2192,7 @@ AC_CONFIG_FILES([configs/current src/mesa/drivers/dri/r200/Makefile src/mesa/drivers/dri/radeon/Makefile src/mesa/drivers/dri/swrast/Makefile + src/mesa/drivers/osmesa/Makefile src/mesa/drivers/x11/Makefile src/mesa/gl.pc src/mesa/osmesa.pc]) diff --git a/src/mesa/drivers/osmesa/.gitignore b/src/mesa/drivers/osmesa/.gitignore new file mode 100644 index 000..5fc607b --- /dev/null +++ b/src/mesa/drivers/osmesa/.gitignore @@ -0,0 +1 @@ +/Makefile diff --git a/src/mesa/drivers/osmesa/Makefile b/src/mesa/drivers/osmesa/Makefile deleted file mode 100644 index 39ab09a..000 --- a/src/mesa/drivers/osmesa/Makefile +++ /dev/null @@ -1,51 +0,0 @@ -# src/mesa/drivers/osmesa/Makefile for libOSMesa.so - -# Note that we may generate libOSMesa.so or libOSMesa16.so or libOSMesa32.so -# with this Makefile - - -TOP = ../../../.. - -include $(TOP)/configs/current - - - -SOURCES = osmesa.c - -OBJECTS = $(SOURCES:.c=.o) - -INCLUDE_DIRS = \ - -I$(TOP)/include \ - -I$(TOP)/src/mapi \ - -I$(TOP)/src/mesa \ - -I$(TOP)/src/mesa/main - -CORE_MESA = \ - $(TOP)/src/mesa/libmesa.a \ - $(TOP)/src/mapi/glapi/libglapi.a \ - $(TOP)/src/glsl/libglsl.a - -.c.o: - $(CC) -c $(INCLUDE_DIRS) $(CFLAGS) $< -o $@ - - -default: $(TOP)/$(LIB_DIR)/$(OSMESA_LIB_NAME) - - -# libOSMesa can be used in conjuction with libGL or with all other Mesa -# sources. We can also build libOSMesa16/libOSMesa32 by setting -# -DCHAN_BITS=16/32. -$(TOP)/$(LIB_DIR)/$(OSMESA_LIB_NAME): $(OBJECTS) $(CORE_MESA) - $(MKLIB) -o $(OSMESA_LIB) -linker '$(CXX)' -ldflags '$(LDFLAGS)' \ - -major $(MESA_MAJOR) -minor $(MESA_MINOR) -patch $(MESA_TINY) \ - -install $(TOP)/$(LIB_DIR) -cplusplus $(MKLIB_OPTIONS) \ - -id $(INSTALL_LIB_DIR)/lib$(OSMESA_LIB).$(MESA_MAJOR).dylib \ - $(OSMESA_LIB_DEPS) $(OBJECTS) $(CORE_MESA) - - - -clean: - -rm -f *.o *~ - - -# XXX todo install rule? diff --git a/src/mesa/drivers/osmesa/Makefile.am b/src/mesa/drivers/osmesa/Makefile.am new file mode 100644 index 000..3cf3005 --- /dev/null +++ b/src/mesa/drivers/osmesa/Makefile.am @@ -0,0 +1,50 @@ + + +# Copyright © 2012 Laurent Carlier +# +# Permission is hereby granted, free of charge, to any person obtaining a +# copy of this software and associated documentation files (the "Software"), +# to deal in the Software without restriction, including without limitation +# the rights to use, copy, modify, merge, publish, distribute, sublicense, +# and/or sell copies of the Software, and to permit persons to whom the +# Software is furnished to do so, subject to the following conditions: +# +# The above copyright notice and this permission notice (including the next +# paragraph) shall be included in all copies or substantial portions of the +# Software. +# +# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL +# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING +# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS +# IN THE SOFTWARE. + +# Hack to make some of the non-automake variables work. +TOP=$(top_builddir) + +AM_CFLAGS = \ + -I$(top_srcdir)/include \ + -I$(top_srcdir)/src/mapi \ + -I$(top_srcdir)/src/mesa/ \ + $(DEFINES) \ + $(API_DEFINES) + +if HAVE_OSMESA_DRIVER +lib_LTLIBRARIES = lib@O
[Mesa-dev] [PATCH 2/9] automake: convert libglsl
v2: Use AM_V_GEN to silence generated code rules. Add BUILT_SOURCES to CLEANFILES Signed-off-by: Jon TURNEY --- configure.ac|2 + src/glsl/.gitignore |1 + src/glsl/Makefile | 144 --- src/glsl/Makefile.am| 112 +++ src/glsl/glcpp/.gitignore |1 + src/glsl/glcpp/Makefile.am | 33 src/glsl/tests/Makefile.am |2 +- src/mesa/drivers/osmesa/Makefile.am |2 +- 8 files changed, 151 insertions(+), 146 deletions(-) delete mode 100644 src/glsl/Makefile create mode 100644 src/glsl/Makefile.am create mode 100644 src/glsl/glcpp/Makefile.am diff --git a/configure.ac b/configure.ac index e5ac791..93a8493 100644 --- a/configure.ac +++ b/configure.ac @@ -2170,6 +2170,8 @@ AC_CONFIG_FILES([configs/current src/egl/wayland/wayland-egl/Makefile src/egl/wayland/wayland-egl/wayland-egl.pc src/egl/wayland/wayland-drm/Makefile + src/glsl/Makefile + src/glsl/glcpp/Makefile src/glsl/tests/Makefile src/glx/Makefile src/glx/tests/Makefile diff --git a/src/glsl/.gitignore b/src/glsl/.gitignore index d26839a..e3531cd 100644 --- a/src/glsl/.gitignore +++ b/src/glsl/.gitignore @@ -6,3 +6,4 @@ glsl_parser.output builtin_function.cpp builtin_compiler glsl_test +/Makefile diff --git a/src/glsl/Makefile b/src/glsl/Makefile deleted file mode 100644 index 3cf9fc9..000 --- a/src/glsl/Makefile +++ /dev/null @@ -1,144 +0,0 @@ - -#src/glsl/pp/Makefile - -TOP = ../.. - -include $(TOP)/configs/current - -LIBNAME = glsl - -GLSL_SRCDIR=. -include Makefile.sources - -GLCPP_SOURCES = \ - $(LIBGLCPP_GENERATED_FILES) \ - $(LIBGLCPP_FILES) \ - ralloc.c \ - glcpp/glcpp.c - -C_SOURCES = \ - $(LIBGLCPP_GENERATED_FILES) \ - $(LIBGLCPP_FILES) \ - $(LIBGLSL_FILES) - -# common sources for builtin_compiler and libglsl -CXX_SOURCES = \ - $(BUILTIN_COMPILER_GENERATED_CXX_FILES) \ - $(LIBGLSL_CXX_FILES) - -LIBS = \ - $(TOP)/src/glsl/libglsl.a - -APPS = glsl_compiler glsl_test glcpp/glcpp - -GLSL2_C_SOURCES = \ - ../mesa/program/hash_table.c \ - ../mesa/program/symbol_table.c -GLSL2_CXX_SOURCES = \ - $(GLSL_COMPILER_CXX_FILES) - -GLSL2_OBJECTS = \ - $(GLSL2_C_SOURCES:.c=.o) \ - $(GLSL2_CXX_SOURCES:.cpp=.o) - -TEST_C_SOURCES = \ - ../mesa/program/hash_table.c \ - ../mesa/program/symbol_table.c - -TEST_CXX_SOURCES = \ - standalone_scaffolding.cpp \ - test.cpp \ - test_optpass.cpp - -TEST_OBJECTS = \ - $(TEST_C_SOURCES:.c=.o) \ - $(TEST_CXX_SOURCES:.cpp=.o) - -### Basic defines ### - -DEFINES += \ - $(LIBRARY_DEFINES) \ - $(API_DEFINES) - -GLCPP_OBJECTS = \ - $(GLCPP_SOURCES:.c=.o) \ - ../mesa/program/hash_table.o - -OBJECTS = \ - $(C_SOURCES:.c=.o) \ - $(CXX_SOURCES:.cpp=.o) - -INCLUDES = \ - -I. \ - -I../mesa \ - -I../mapi \ - -I../../include \ - $(LIBRARY_INCLUDES) - -ALL_SOURCES = \ - $(C_SOURCES) \ - $(CXX_SOURCES) \ - $(GLSL2_CXX_SOURCES) \ - $(GLSL2_C_SOURCES) \ - $(TEST_CXX_SOURCES) \ - $(TEST_C_SOURCES) - -# TARGETS # - -default: depend lib$(LIBNAME).a $(APPS) - -lib$(LIBNAME).a: $(OBJECTS) builtin_function.o Makefile $(TOP)/src/glsl/Makefile.template - $(MKLIB) -cplusplus -o $(LIBNAME) -static $(OBJECTS) builtin_function.o - -depend: $(ALL_SOURCES) Makefile - rm -f depend - touch depend - $(MKDEP) $(MKDEP_OPTIONS) $(INCLUDES) $(ALL_SOURCES) 2> /dev/null - -# Remove .o and backup files -clean: - rm -f $(GLCPP_OBJECTS) $(GLSL2_OBJECTS) $(TEST_OBJECTS) $(OBJECTS) lib$(LIBNAME).a depend depend.bak builtin_function.cpp builtin_function.o builtin_stubs.o builtin_compiler - -rm -f $(APPS) - -# Dummy target -install: - @echo -n "" - -# RULES # - -glsl_compiler: $(GLSL2_OBJECTS) libglsl.a - $(APP_CXX) $(INCLUDES) $(CFLAGS) $(LDFLAGS) $(GLSL2_OBJECTS) $(LIBS) -o $@ - -glsl_test: $(TEST_OBJECTS) libglsl.a - $(APP_CXX) $(INCLUDES) $(CFLAGS) $(LDFLAGS) $(TEST_OBJECTS) $(LIBS) -o $@ - -glcpp: glcpp/glcpp -glcpp/glcpp: $(GLCPP_OBJECTS) - $(APP_CC) $(INCLUDES) $(CFLAGS) $(LDFLAGS) $(GLCPP_OBJECTS) -o $@ - -.cpp.o: - $(CXX) -c $(INCLUDES) $(CXXFLAGS) $(DEFINES) $< -o $@ - -.c.o: - $(CC) -c $(INCLUDES) $(CFLAGS) $(DEFINES) $< -o $@ - -glsl_lexer.cpp: glsl_lexer.ll - $(FLEX) --nounistd -o$@ $< - -glsl_parser.cpp: glsl_parser.yy - $(BISON) -v -o "$@" -p "_mesa_glsl_" --defines=glsl_parser.h $< - -glcpp/glcpp-lex.c: glcpp/glcpp-lex.l - $(FLEX) --nounistd -o$@ $< - -glcpp/glcpp-parse.c: glcpp/glcpp-parse.y - $(BISON) -v -o "$@" --defines=glcpp/glcpp-parse.h $< - -builtin_compiler: $(GLSL2_OBJECTS) $(OBJECTS) buil
[Mesa-dev] [PATCH 3/9] Rename sparc/clip.S -> sparc/sparc_clip.S
Automake can't handle having both clip.S and clip.c, even though they have
different paths
"src/mesa/Makefile.am: object `clip.lo' created by `$(SRCDIR)/sparc/clip.S' and
`$(SRCDIR)/main/clip.c'"
Signed-off-by: Jon TURNEY
---
src/mesa/SConscript |2 +-
src/mesa/sources.mak|2 +-
src/mesa/sparc/clip.S | 233 ---
src/mesa/sparc/sparc_clip.S | 233 +++
4 files changed, 235 insertions(+), 235 deletions(-)
delete mode 100644 src/mesa/sparc/clip.S
create mode 100644 src/mesa/sparc/sparc_clip.S
diff --git a/src/mesa/SConscript b/src/mesa/SConscript
index 819a0fd..906c579 100644
--- a/src/mesa/SConscript
+++ b/src/mesa/SConscript
@@ -430,7 +430,7 @@ if env['gcc'] and env['platform'] not in ('darwin',
'windows'):
elif env['machine'] == 'sparc':
mesa_sources += [
'sparc/sparc.c',
-'sparc/clip.S',
+'sparc/sparc_clip.S',
'sparc/norm.S',
'sparc/xform.S',
]
diff --git a/src/mesa/sources.mak b/src/mesa/sources.mak
index 16b1c39..d22f059 100644
--- a/src/mesa/sources.mak
+++ b/src/mesa/sources.mak
@@ -307,7 +307,7 @@ X86_64_FILES = \
$(SRCDIR)/x86-64/xform4.S
SPARC_FILES = \
- $(SRCDIR)/sparc/clip.S \
+ $(SRCDIR)/sparc/sparc_clip.S\
$(SRCDIR)/sparc/norm.S \
$(SRCDIR)/sparc/xform.S
diff --git a/src/mesa/sparc/clip.S b/src/mesa/sparc/clip.S
deleted file mode 100644
index dc23917..000
--- a/src/mesa/sparc/clip.S
+++ /dev/null
@@ -1,233 +0,0 @@
-/*
- * Clip testing in SPARC assembly
- */
-
-#if __arch64__
-#define LDPTR ldx
-#define V4F_DATA 0x00
-#define V4F_START 0x08
-#define V4F_COUNT 0x10
-#define V4F_STRIDE 0x14
-#define V4F_SIZE 0x18
-#define V4F_FLAGS 0x1c
-#else
-#define LDPTR ld
-#define V4F_DATA 0x00
-#define V4F_START 0x04
-#define V4F_COUNT 0x08
-#define V4F_STRIDE 0x0c
-#define V4F_SIZE 0x10
-#define V4F_FLAGS 0x14
-#endif
-
-#define VEC_SIZE_1 1
-#define VEC_SIZE_2 3
-#define VEC_SIZE_3 7
-#define VEC_SIZE_4 15
-
-.register %g2, #scratch
-.register %g3, #scratch
-
- .text
- .align 64
-
-one_dot_zero:
- .word 0x3f80 /* 1.0f */
-
- /* This trick is shamelessly stolen from the x86
-* Mesa asm. Very clever, and we can do it too
-* since we have the necessary add with carry
-* instructions on Sparc.
-*/
-clip_table:
- .byte0, 1, 0, 2, 4, 5, 4, 6
- .byte0, 1, 0, 2, 8, 9, 8, 10
- .byte 32, 33, 32, 34, 36, 37, 36, 38
- .byte 32, 33, 32, 34, 40, 41, 40, 42
- .byte0, 1, 0, 2, 4, 5, 4, 6
- .byte0, 1, 0, 2, 8, 9, 8, 10
- .byte 16, 17, 16, 18, 20, 21, 20, 22
- .byte 16, 17, 16, 18, 24, 25, 24, 26
- .byte 63, 61, 63, 62, 55, 53, 55, 54
- .byte 63, 61, 63, 62, 59, 57, 59, 58
- .byte 47, 45, 47, 46, 39, 37, 39, 38
- .byte 47, 45, 47, 46, 43, 41, 43, 42
- .byte 63, 61, 63, 62, 55, 53, 55, 54
- .byte 63, 61, 63, 62, 59, 57, 59, 58
- .byte 31, 29, 31, 30, 23, 21, 23, 22
- .byte 31, 29, 31, 30, 27, 25, 27, 26
-
-/* GLvector4f *clip_vec, GLvector4f *proj_vec,
- GLubyte clipMask[], GLubyte *orMask, GLubyte *andMask,
- GLboolean viewport_z_enable */
-
- .align 64
-__pc_tramp:
- retl
-nop
-
- .globl _mesa_sparc_cliptest_points4
-_mesa_sparc_cliptest_points4:
- save%sp, -64, %sp
- call__pc_tramp
-sub%o7, (. - one_dot_zero - 4), %g1
- ld [%g1 + 0x0], %f4
- add %g1, 0x4, %g1
-
- ld [%i0 + V4F_STRIDE], %l1
- ld [%i0 + V4F_COUNT], %l3
- LDPTR [%i0 + V4F_START], %i0
- LDPTR [%i1 + V4F_START], %i5
- ldub[%i3], %g2
- ldub[%i4], %g3
- sll %g3, 8, %g3
- or %g2, %g3, %g2
-
- ld [%i1 + V4F_FLAGS], %g3
- or %g3, VEC_SIZE_4, %g3
- st %g3, [%i1 + V4F_FLAGS]
- mov 3, %g3
- st %g3, [%i1 + V4F_SIZE]
- st %l3, [%i1 + V4F_COUNT]
- clr %l2
- clr %l0
-
- /* l0: i
-* l3: count
-* l1: stride
-* l2: c
-* g2: (tmpAndMask << 8) | tmpOrMask
-* g1: clip_table
-* i0: from[stride][i]
-* i2: clipMask
-* i5: vProj[4][i]
-*/
-
-1: ld [%i0 + 0x0c], %f3 ! LSU Group
- ld [%i0 + 0x0c], %g5 ! LSU Group
- ld [%i0 + 0x08], %g4 ! LSU Gr
[Mesa-dev] [PATCH 5/9] Remove unused MESA_MODULES autoconf variable
--- configs/current.in |1 - configure.ac |3 --- 2 files changed, 0 insertions(+), 4 deletions(-) diff --git a/configs/current.in b/configs/current.in index f4858bd..ca817c4 100644 --- a/configs/current.in +++ b/configs/current.in @@ -147,7 +147,6 @@ VG_LIB_DEPS = $(EXTRA_LIB_PATH) @VG_LIB_DEPS@ GLAPI_LIB_DEPS = $(EXTRA_LIB_PATH) @GLAPI_LIB_DEPS@ # DRI dependencies -MESA_MODULES = @MESA_MODULES@ DRI_LIB_DEPS = $(EXTRA_LIB_PATH) @DRI_LIB_DEPS@ LIBDRM_CFLAGS = @LIBDRM_CFLAGS@ LIBDRM_LIB = @LIBDRM_LIBS@ diff --git a/configure.ac b/configure.ac index 93a8493..ae28007 100644 --- a/configure.ac +++ b/configure.ac @@ -1086,15 +1086,12 @@ AC_SUBST([GLAPI_LIB_DEPS]) dnl Setup default DRI CFLAGS DRI_LIB_DEPS='$(TOP)/src/mesa/libmesa.a' -MESA_MODULES='$(TOP)/src/mesa/libmesa.a' if test "x$enable_dri" = xyes && test "x$driglx_direct" = xyes ; then DRI_LIB_DEPS="-L\$(TOP)/\$(LIB_DIR) -ldricore$VERSION" -MESA_MODULES="\$(TOP)/\$(LIB_DIR)/libdricore$VERSION.so" HAVE_DRICORE=yes fi AM_CONDITIONAL(HAVE_DRICORE, test x$HAVE_DRICORE = xyes) -AC_SUBST([MESA_MODULES]) AC_SUBST([HAVE_XF86VIDMODE]) -- 1.7.9 ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [PATCH 4/9] automake: convert libmesa and libmesagallium
* "configure substitutions are not allowed in _SOURCES variables" in automake, so instead of MESA_ASM_FILES, use some AM_CONDITIONALS to choose which architecture's asm sources are used in libmesa_la_SOURCES. (Can't remove MESA_ASM_FILES autoconf variable as it's still used in sources.mak) * Update to link with the .la file in other Makefile.am files, and make a link to the .a file for the convenience of other Makefiles which have not yet been converted to automake v2: Remove stray -static from LDFLAGS Signed-off-by: Jon TURNEY --- src/glsl/tests/Makefile.am |2 +- src/mesa/Makefile.am| 45 +-- src/mesa/Makefile.old | 57 +-- src/mesa/drivers/osmesa/Makefile.am |2 +- src/mesa/drivers/x11/Makefile.am|2 +- src/mesa/main/tests/Makefile.am |2 +- 6 files changed, 46 insertions(+), 64 deletions(-) diff --git a/src/glsl/tests/Makefile.am b/src/glsl/tests/Makefile.am index 4829fb2..33e634d 100644 --- a/src/glsl/tests/Makefile.am +++ b/src/glsl/tests/Makefile.am @@ -28,7 +28,7 @@ uniform_initializer_test_SOURCES =\ uniform_initializer_test_LDADD = \ $(top_builddir)/src/gtest/libgtest.la \ $(top_builddir)/src/glsl/libglsl.la \ - $(top_builddir)/src/mesa/libmesa.a \ + $(top_builddir)/src/mesa/libmesa.la \ -lpthread ralloc_test_SOURCES = ralloc_test.cpp $(top_builddir)/src/glsl/ralloc.c diff --git a/src/mesa/Makefile.am b/src/mesa/Makefile.am index bada760..54ee253 100644 --- a/src/mesa/Makefile.am +++ b/src/mesa/Makefile.am @@ -93,14 +93,51 @@ program/lex.yy.c: program/program_lexer.l mkdir -p program $(LEX) --never-interactive --outfile=$@ $< -all-local: - $(MAKE) -f $(srcdir)/Makefile.old +noinst_LTLIBRARIES = libmesa.la libmesagallium.la + +SRCDIR = $(top_srcdir)/src/mesa +include sources.mak + +AM_CFLAGS = $(API_DEFINES) $(DEFINES) $(INCLUDE_DIRS) $(LLVM_CFLAGS) $(CFLAGS) +AM_CXXFLAGS = $(API_DEFINES) $(DEFINES) $(INCLUDE_DIRS) $(LLVM_CFLAGS) $(CXXFLAGS) + +# cannot just add $(MESA_ASM_FILES) to libmesa_la_SOURCES as it contains a configure substitution +MESA_ASM_FILES_FOR_ARCH = + +if HAVE_X86_ASM +MESA_ASM_FILES_FOR_ARCH += $(X86_FILES) +endif +if HAVE_X86_64_ASM +MESA_ASM_FILES_FOR_ARCH += $(X86_64_FILES) +endif +if HAVE_SPARC_ASM +MESA_ASM_FILES_FOR_ARCH += $(SPARC_FILES) +endif + +libmesa_la_SOURCES = \ + $(MESA_FILES) \ + $(MESA_CXX_FILES) \ +$(MESA_ASM_FILES_FOR_ARCH) + +libmesa_la_LIBADD = $(top_builddir)/src/glsl/libglsl.la +libmesa_la_LDFLAGS = + +libmesagallium_la_SOURCES = \ + $(MESA_GALLIUM_FILES) \ + $(MESA_GALLIUM_CXX_FILES) \ +$(MESA_ASM_FILES_FOR_ARCH) + +libmesagallium_la_LIBADD = $(top_builddir)/src/glsl/libglsl.la +libmesagallium_la_LDFLAGS = install-exec-local: $(MAKE) -f $(srcdir)/Makefile.old install -clean-local: - $(MAKE) -f $(srcdir)/Makefile.old clean +# Provide compatibility with scripts for the old Mesa build system for +# a while by putting a link to the library in the current directory. +all-local: libmesa.la libmesagallium.la + ln -f .libs/libmesa.a . + ln -f .libs/libmesagallium.a . pkgconfigdir = $(libdir)/pkgconfig diff --git a/src/mesa/Makefile.old b/src/mesa/Makefile.old index 4ea70d4..3266a5d 100644 --- a/src/mesa/Makefile.old +++ b/src/mesa/Makefile.old @@ -3,58 +3,10 @@ TOP = ../.. include $(TOP)/configs/current -MESA_LIBS := libmesa.a libmesagallium.a -DEPENDS := depend - SRCDIR = . include sources.mak -# define preprocessor flags -MESA_CPPFLAGS := $(API_DEFINES) $(DEFINES) - -# append include dirs -MESA_CPPFLAGS += $(INCLUDE_DIRS) - -# tidy compiler flags -CFLAGS := $(filter-out $(DEFINES), $(CFLAGS)) -CXXFLAGS := $(filter-out $(DEFINES), $(CXXFLAGS)) - -# LLVM is needed for the state tracker -MESA_CFLAGS := $(LLVM_CFLAGS) $(CFLAGS) -MESA_CXXFLAGS := $(LLVM_CFLAGS) $(CXXFLAGS) - -%.o: %.c - $(CC) -c -o $@ $< $(MESA_CPPFLAGS) $(MESA_CFLAGS) - -%.o: %.cpp - $(CXX) -c -o $@ $< $(MESA_CPPFLAGS) $(MESA_CXXFLAGS) - -%.o: %.S - $(CC) -c -o $@ $< $(MESA_CPPFLAGS) $(MESA_CFLAGS) - -# Default: build dependencies, then asm_subdirs, GLSL built-in lib, -# then convenience libs (.a) and finally the device drivers: -default: $(DEPENDS) $(MESA_LIBS) - -## -# Helper libraries used by many drivers: - -# Make archive of core mesa object files -libmesa.a: $(MESA_OBJECTS) $(GLSL_LIBS) - @ $(MKLIB) -o mesa -static $(MESA_OBJECTS) $(GLSL_LIBS) - -# Make archive of subset of core mesa object files for gallium -libmesagallium.a: $(MESA_GALLIUM_OBJECTS) $(GLSL_LIBS) - @ $(MKLIB) -o mesagallium -static $(MESA_GALLIUM_OBJECTS) $(GLSL_LIBS) - -## -# Dependency generation - -depend: $(ALL_FILES) main/git_sha1.h -
[Mesa-dev] [PATCH 6/9] Link dri drivers with mesa or dricore libtool library
Now mesa/drivers/dri is converted to automake, we want to update DRI_LIB_DEPS so that we link with the libmesa or libdricore libtool library, as appropriate. However, this is complicated by the fact that gallium/targets is not (yet) converted, so we can't share the DRI_LIB_DEPS autoconf variable with that anymore. Add an additional autoconf variable GALLIUM_DRI_LIB_DEPS, whic is now used in gallium/targets/Makefile.dri, to link with the libdircore or libmesa native library. Signed-off-by: Jon TURNEY --- configs/current.in |1 + configure.ac | 11 --- src/gallium/targets/Makefile.dri |4 ++-- 3 files changed, 11 insertions(+), 5 deletions(-) diff --git a/configs/current.in b/configs/current.in index ca817c4..e0c0130 100644 --- a/configs/current.in +++ b/configs/current.in @@ -148,6 +148,7 @@ GLAPI_LIB_DEPS = $(EXTRA_LIB_PATH) @GLAPI_LIB_DEPS@ # DRI dependencies DRI_LIB_DEPS = $(EXTRA_LIB_PATH) @DRI_LIB_DEPS@ +GALLIUM_DRI_LIB_DEPS = $(EXTRA_LIB_PATH) @GALLIUM_DRI_LIB_DEPS@ LIBDRM_CFLAGS = @LIBDRM_CFLAGS@ LIBDRM_LIB = @LIBDRM_LIBS@ DRI2PROTO_CFLAGS = @DRI2PROTO_CFLAGS@ diff --git a/configure.ac b/configure.ac index ae28007..5724f8a 100644 --- a/configure.ac +++ b/configure.ac @@ -1084,11 +1084,14 @@ GLAPI_LIB_DEPS="-lpthread $SELINUX_LIBS" AC_SUBST([GLAPI_LIB_DEPS]) -dnl Setup default DRI CFLAGS -DRI_LIB_DEPS='$(TOP)/src/mesa/libmesa.a' +dnl dri libraries are linking with mesa +DRI_LIB_DEPS='$(TOP)/src/mesa/libmesa.la' +GALLIUM_DRI_LIB_DEPS='$(TOP)/src/mesa/libmesa.a' +dnl ... or dricore? if test "x$enable_dri" = xyes && test "x$driglx_direct" = xyes ; then -DRI_LIB_DEPS="-L\$(TOP)/\$(LIB_DIR) -ldricore$VERSION" +DRI_LIB_DEPS='$(TOP)/src/mesa/libdricore/libdricore$VERSION.la' +GALLIUM_DRI_LIB_DEPS='$(TOP)/src/mesa/libdricore/libdricore$VERSION.a' HAVE_DRICORE=yes fi AM_CONDITIONAL(HAVE_DRICORE, test x$HAVE_DRICORE = xyes) @@ -1266,10 +1269,12 @@ if test "x$enable_dri" = xyes; then # put all the necessary libs together DRI_LIB_DEPS="$DRI_LIB_DEPS $SELINUX_LIBS $LIBDRM_LIBS $EXPAT_LIB -lm -lpthread $DLOPEN_LIBS" +GALLIUM_DRI_LIB_DEPS="$GALLIUM_DRI_LIB_DEPS $SELINUX_LIBS $LIBDRM_LIBS $EXPAT_LIB -lm -lpthread $DLOPEN_LIBS" fi AC_SUBST([DRI_DIRS]) AC_SUBST([EXPAT_INCLUDES]) AC_SUBST([DRI_LIB_DEPS]) +AC_SUBST([GALLIUM_DRI_LIB_DEPS]) case $DRI_DIRS in *i915*|*i965*) diff --git a/src/gallium/targets/Makefile.dri b/src/gallium/targets/Makefile.dri index 3abed82..5b6676d 100644 --- a/src/gallium/targets/Makefile.dri +++ b/src/gallium/targets/Makefile.dri @@ -73,8 +73,8 @@ $(LIBNAME): $(OBJECTS) $(MESA_MODULES) $(PIPE_DRIVERS) Makefile \ $(MKLIB) -o $@.tmp -noprefix -linker '$(CXX)' -ldflags '$(LDFLAGS)' \ $(OBJECTS) $(PIPE_DRIVERS) \ -Wl,--start-group $(MESA_MODULES) -Wl,--end-group \ - $(DRI_LIB_DEPS) $(DRIVER_EXTRAS) - $(CXX) $(CFLAGS) -o $@.test $(TOP)/src/mesa/drivers/dri/common/dri_test.o $@.tmp $(DRI_LIB_DEPS) $(LDFLAGS); + $(GALLIUM_DRI_LIB_DEPS) $(DRIVER_EXTRAS) + $(CXX) $(CFLAGS) -o $@.test $(TOP)/src/mesa/drivers/dri/common/dri_test.o $@.tmp $(GALLIUM_DRI_LIB_DEPS) $(LDFLAGS); @rm -f $@.test mv -f $@.tmp $@ -- 1.7.9 ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [PATCH 7/9] Rename X86-64_API -> X86_64_API
automake doesn't allow hyphens in variable names Signed-off-by: Jon TURNEY --- configure.ac |2 +- src/mapi/glapi/sources.mak |2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/configure.ac b/configure.ac index 5724f8a..c579fd3 100644 --- a/configure.ac +++ b/configure.ac @@ -509,7 +509,7 @@ if test "x$enable_asm" = xyes; then x86_64) DEFINES="$DEFINES -DUSE_X86_64_ASM" MESA_ASM_FILES='$(X86_64_FILES)' -GLAPI_ASM_SOURCES='$(X86-64_API)' +GLAPI_ASM_SOURCES='$(X86_64_API)' AC_MSG_RESULT([yes, x86_64]) ;; sparc) diff --git a/src/mapi/glapi/sources.mak b/src/mapi/glapi/sources.mak index dfc6cc6..aa8a4d4 100644 --- a/src/mapi/glapi/sources.mak +++ b/src/mapi/glapi/sources.mak @@ -12,7 +12,7 @@ GLAPI_SOURCES = \ X86_API = \ glapi_x86.S -X86-64_API = \ +X86_64_API = \ glapi_x86-64.S SPARC_API =\ -- 1.7.9 ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [PATCH 9/9] Don't explicitly link libOsmesa with libmesa's dependency libglsl
The libmesa convenience library is linked with the libglsl convenience library. libOsmesa is linked with libmesa, and also directly with libglsl. This gives rise to duplicate symbol errors. We must either: (i) Not link libmesa with libglsl, and link anything that uses libmesa with libglsl as well, or (ii) Link libmesa with libglsl, and not link anything that uses libmesa with libglsl I choose (ii) just because it's least change, but I'm not sure it's right. Note that drivers/X11 makes libGL by linking with libmesa, but not libglsl, so it's a bit unclear to me how this was working correctly before in both cases (It might be some sort of accidental side-effect of what mklib's expand_archives() function does that this ever worked before?) --- src/mesa/drivers/osmesa/Makefile.am |3 +-- 1 files changed, 1 insertions(+), 2 deletions(-) diff --git a/src/mesa/drivers/osmesa/Makefile.am b/src/mesa/drivers/osmesa/Makefile.am index 9989c92..7c151bf 100644 --- a/src/mesa/drivers/osmesa/Makefile.am +++ b/src/mesa/drivers/osmesa/Makefile.am @@ -40,8 +40,7 @@ lib@OSMESA_LIB@_la_SOURCES = osmesa.c lib@OSMESA_LIB@_la_LDFLAGS = -module -version-number @OSMESA_VERSION@ -shared lib@OSMESA_LIB@_la_LIBADD = \ $(top_builddir)/src/mesa/libmesa.la \ - $(top_builddir)/src/mapi/glapi/libglapi.la \ - $(top_builddir)/src/glsl/libglsl.la + $(top_builddir)/src/mapi/glapi/libglapi.la # Provide compatibility with scripts for the old Mesa build system for # a while by putting a link to the driver into /lib of the build tree. -- 1.7.9 ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [PATCH 8/9] automake: convert libglapi
* "configure substitutions are not allowed in _SOURCES variables" in automake, so remove the AC_SUBST'ed GLAPI_ASM_SOURCES and instead use some AM_CONDITIONALS to choose which asm sources are used * Change GLAPI_LIB to point to the .la file in other Makefile.am files, and make a link to the .a file for the convenience of other Makefiles which have not yet been converted to automake Signed-off-by: Jon TURNEY --- configs/current.in |1 - configure.ac|6 +-- src/egl/main/Makefile.am|2 +- src/glx/Makefile.am |2 +- src/mapi/glapi/.gitignore |1 + src/mapi/glapi/Makefile | 77 --- src/mapi/glapi/Makefile.am | 69 +++ src/mapi/glapi/tests/Makefile.am|2 +- src/mesa/drivers/osmesa/Makefile.am |2 +- src/mesa/drivers/x11/Makefile.am|2 +- 10 files changed, 76 insertions(+), 88 deletions(-) delete mode 100644 src/mapi/glapi/Makefile create mode 100644 src/mapi/glapi/Makefile.am diff --git a/configs/current.in b/configs/current.in index e0c0130..dc0dea8 100644 --- a/configs/current.in +++ b/configs/current.in @@ -47,7 +47,6 @@ DLOPEN_LIBS = @DLOPEN_LIBS@ # Source selection MESA_ASM_FILES = @MESA_ASM_FILES@ -GLAPI_ASM_SOURCES = @GLAPI_ASM_SOURCES@ # Misc tools and flags MAKE = @MAKE@ diff --git a/configure.ac b/configure.ac index c579fd3..7e55907 100644 --- a/configure.ac +++ b/configure.ac @@ -454,7 +454,6 @@ AC_ARG_ENABLE([asm], ) asm_arch="" MESA_ASM_FILES="" -GLAPI_ASM_SOURCES="" AC_MSG_CHECKING([whether to enable assembly]) test "x$enable_asm" = xno && AC_MSG_RESULT([no]) # disable if cross compiling on x86/x86_64 since we must run gen_matypes @@ -503,19 +502,16 @@ if test "x$enable_asm" = xyes; then x86) DEFINES="$DEFINES -DUSE_X86_ASM -DUSE_MMX_ASM -DUSE_3DNOW_ASM -DUSE_SSE_ASM" MESA_ASM_FILES='$(X86_FILES)' -GLAPI_ASM_SOURCES='$(X86_API)' AC_MSG_RESULT([yes, x86]) ;; x86_64) DEFINES="$DEFINES -DUSE_X86_64_ASM" MESA_ASM_FILES='$(X86_64_FILES)' -GLAPI_ASM_SOURCES='$(X86_64_API)' AC_MSG_RESULT([yes, x86_64]) ;; sparc) DEFINES="$DEFINES -DUSE_SPARC_ASM" MESA_ASM_FILES='$(SPARC_FILES)' -GLAPI_ASM_SOURCES='$(SPARC_API)' AC_MSG_RESULT([yes, sparc]) ;; *) @@ -524,7 +520,6 @@ if test "x$enable_asm" = xyes; then esac fi AC_SUBST([MESA_ASM_FILES]) -AC_SUBST([GLAPI_ASM_SOURCES]) dnl PIC code macro MESA_PIC_FLAGS @@ -2177,6 +2172,7 @@ AC_CONFIG_FILES([configs/current src/glsl/tests/Makefile src/glx/Makefile src/glx/tests/Makefile + src/mapi/glapi/Makefile src/mapi/glapi/gen/Makefile src/mapi/shared-glapi/Makefile src/mapi/glapi/tests/Makefile diff --git a/src/egl/main/Makefile.am b/src/egl/main/Makefile.am index 9c3935b..ca5257a 100644 --- a/src/egl/main/Makefile.am +++ b/src/egl/main/Makefile.am @@ -19,7 +19,7 @@ # FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS # IN THE SOFTWARE. -GLAPI_LIB = ../mapi/glapi/libglapi.a +GLAPI_LIB = ../mapi/glapi/libglapi.la if HAVE_XF86VIDMODE EXTRA_DEFINES_XF86VIDMODE = -DXF86VIDMODE diff --git a/src/glx/Makefile.am b/src/glx/Makefile.am index 37a938d..6ca85cd 100644 --- a/src/glx/Makefile.am +++ b/src/glx/Makefile.am @@ -26,7 +26,7 @@ endif SUBDIRS=tests -GLAPI_LIB = ../mapi/glapi/libglapi.a +GLAPI_LIB = ../mapi/glapi/libglapi.la if HAVE_XF86VIDMODE EXTRA_DEFINES_XF86VIDMODE = -DXF86VIDMODE diff --git a/src/mapi/glapi/.gitignore b/src/mapi/glapi/.gitignore index ecae56a..25d3dfd 100644 --- a/src/mapi/glapi/.gitignore +++ b/src/mapi/glapi/.gitignore @@ -7,3 +7,4 @@ glapi_x86.S glapitable.h glapitemp.h glprocs.h +\Makefile diff --git a/src/mapi/glapi/Makefile b/src/mapi/glapi/Makefile deleted file mode 100644 index 211f384..000 --- a/src/mapi/glapi/Makefile +++ /dev/null @@ -1,77 +0,0 @@ -# src/mapi/glapi/Makefile - -TOP = ../../.. -include $(TOP)/configs/current - -TARGET = glapi - -MAPI = $(TOP)/src/mapi/mapi - -include sources.mak -include $(MAPI)/sources.mak - -glapi_CPPFLAGS := \ - -I$(TOP)/include \ - -I$(TOP)/src/mapi \ - -I$(TOP)/src/mesa - -ifeq ($(SHARED_GLAPI),1) -glapi_CPPFLAGS += \ - -DMAPI_MODE_BRIDGE \ - -DMAPI_ABI_HEADER=\"glapi/glapi_mapi_tmp.h\" -glapi_SOURCES := $(MAPI_BRIDGE_FILES) - -glapi_GLAPI_OBJECTS := -glapi_ASM_OBJECTS := -glapi_MAPI_OBJECTS := $(notdir $(MAPI_BRIDGE_FILES:.c=.o)) -else -glapi_CPPFLAGS += -DMAPI_MODE_UTIL -glapi_SOURCES := $(GLAPI_SOURCES) $(MAPI_UTIL_FILES) - -glapi_GLAPI_OBJECTS := $(GLAPI_SOURCES:.c=.o) -glapi_ASM_OBJECTS := $(GLAPI_ASM_SOURCES:.S=.o) -glapi_MAPI_OBJECTS := $(notdir $(MAPI_UTIL_FILES:.c=.o)) -endif # SHARED_GLAPI - -glapi_OBJECTS := \ - $(glapi_G
Re: [Mesa-dev] [PATCH] configure.ac: Add --with-(gl|glu|osmesa)-lib-name options
On 07/09/2012 04:57 PM, Brad King wrote: > Running "git bisect" points to commit 2d4b77c7 (automake: > Convert src/mesa/drivers/x11/Makefile, 2012-06-12). The change > made by the commit does not obviously drop use of GL_LIB. Upon closer inspection it *does* obviously drop use of GL_LIB. Now "libGL" is hard-coded in "Makefile.am". Naive replacement like that below is not valid automake code AFAICT. I'm not familiar enough with automake to know how to make the library name configurable. Does anyone know? Thanks, -Brad diff --git a/src/mesa/drivers/x11/Makefile.am b/src/mesa/drivers/x11/Makefile.am index cced238..af3fa8f 100644 --- a/src/mesa/drivers/x11/Makefile.am +++ b/src/mesa/drivers/x11/Makefile.am @@ -33,10 +33,10 @@ AM_CPPFLAGS = \ $(DEFINES) if HAVE_X11_DRIVER -lib_LTLIBRARIES = libGL.la +lib_LTLIBRARIES = lib$(GL_LIB).la endif -libGL_la_SOURCES = \ +lib$(GL_LIB)_la_SOURCES = \ glxapi.h \ glxheader.h \ xfonts.h \ @@ -55,20 +55,20 @@ GL_MAJOR = 1 GL_MINOR = 6 GL_PATCH = 0 -libGL_la_LIBADD = \ +lib$(GL_LIB)_la_LIBADD = \ $(top_builddir)/src/mesa/libmesa.a \ $(top_builddir)/src/mapi/glapi/libglapi.a -libGL_la_LDFLAGS = \ +lib$(GL_LIB)_la_LDFLAGS = \ -version-number $(GL_MAJOR):$(GL_MINOR):$(GL_PATCH) \ $(GL_LIB_DEPS) if HAVE_SHARED_GLAPI -libGL_la_LDFLAGS += -L$(top_builddir)/$(LIB_DIR) -l$(GLAPI_LIB) +lib$(GL_LIB)_la_LDFLAGS += -L$(top_builddir)/$(LIB_DIR) -l$(GLAPI_LIB) endif # Provide compatibility with scripts for the old Mesa build system for # a while by putting a link to the driver into /lib of the build tree. -all-local: libGL.la +all-local: lib$(GL_LIB).la $(MKDIR_P) $(top_builddir)/$(LIB_DIR); - ln -f .libs/libGL.so.$(GL_MAJOR).$(GL_MINOR).$(GL_PATCH) $(top_builddir)/$(LIB_DIR)/libGL.so.1 - ln -sf libGL.so.1 $(top_builddir)/$(LIB_DIR)/libGL.so + ln -f .libs/lib$(GL_LIB).so.$(GL_MAJOR).$(GL_MINOR).$(GL_PATCH) $(top_builddir)/$(LIB_DIR)/lib$(GL_LIB).so.1 + ln -sf lib$(GL_LIB).so.1 $(top_builddir)/$(LIB_DIR)/lib$(GL_LIB).so ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH 00/10] ARB_transform_feedback3 & ARB_transform_feedback_instanced
Hi, I plan to push this series in two days if there's no other feedback. Marek On Thu, Jul 5, 2012 at 1:20 PM, Marek Olšák wrote: > Hi everyone, > > the title says it all. These are the remaining transform feedback extensions > from GL4. Nothing needs to be done for this in drivers, provided they fully > implement the internal interfaces. > > This doesn't add support for multiple vertex streams. It seems like > ARB_transform_feedback3 adds vertex streams, but it only adds the API. The > main part of the functionality is defined in ARB_gpu_shader5. > > There are new tests for this on the piglit mailing list. > > Please review. > > Marek Olšák (10): > glapi: add ARB_transform_feedback3 > glapi: add ARB_transform_feedback_instanced > glsl: implement ARB_transform_feedback3 in the linker > mesa: implement glGet queries and error handling for > ARB_transform_feedback3 > mesa: implement indexed query functions from ARB_transform_feedback3 > mesa: implement display list support for indexed query functions > mesa: implement display list support for new DrawTransformFeedback > functions > mesa: implement new DrawTransformFeedback functions > mesa: add ARB_transform_feedback_instanced extension enable flag > st/mesa: expose new transform feedback extensions > > src/glsl/linker.cpp| 119 ++ > src/mapi/glapi/gen/gl_API.xml | 50 - > src/mesa/main/api_validate.c | 26 +-- > src/mesa/main/api_validate.h |4 +- > src/mesa/main/context.c|1 + > src/mesa/main/dd.h |8 ++ > src/mesa/main/dlist.c | 125 > +++- > src/mesa/main/extensions.c |2 + > src/mesa/main/get.c|9 +++ > src/mesa/main/mtypes.h |3 + > src/mesa/main/queryobj.c | 81 + > src/mesa/main/transformfeedback.c | 32 > src/mesa/main/vtxfmt.c |5 ++ > src/mesa/state_tracker/st_extensions.c |6 ++ > src/mesa/vbo/vbo_exec_array.c | 62 ++-- > src/mesa/vbo/vbo_save_api.c| 45 +++- > 16 files changed, 533 insertions(+), 45 deletions(-) > > Marek ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH 00/22] Gallium and R600 improvements
- Original Message - > Hi everyone, > > this series has two parts. > > Patches 1-12 are Mesa/Gallium improvements. Most importantly, this > adds acceleration of stencil copying (including glBlitFramebuffer) > using the shader stencil export functionality. I am afraid r600g and > radeonsi are the only drivers which can make use of it at the > moment. There is also acceleration for depth texture mipmap > generation. I've skimmed through these, and didn't notice anything wrong. Jose ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH 00/22] Gallium and R600 improvements
On Mon, Jul 9, 2012 at 3:15 PM, Marek Olšák wrote: > Hi everyone, > > this series has two parts. > > Patches 1-12 are Mesa/Gallium improvements. Most importantly, this adds > acceleration of stencil copying (including glBlitFramebuffer) using the > shader stencil export functionality. I am afraid r600g and radeonsi are the > only drivers which can make use of it at the moment. There is also > acceleration for depth texture mipmap generation. > > Patches 13-22 are R600 improvements. There is a new derived state for the CB > registers which kept cumulating in the draw_vbo function, and some > improvements for depth-stencil flushing and uploading. My changes to the > CB_TARGET_MASK emission in the compute code (patch #13) might need some > review from somebody working on compute. > > The R600 changes depend on the not-yet-committed patch "r600g: improve > flushed depth texture handling v2" from Vadim Girlin. > > The series fixes 23 depth-stencil piglit tests on r600g. For the series: Reviewed-by: Alex Deucher > > Marek Olšák (22): > mesa: remove assertions that do not allow compressed 2D_ARRAY textures > gallium/u_gen_mipmap: accelerate depth texture mipmap generation > gallium/u_blitter: minify depth0 when initializing last_layer > gallium: add util_format_stencil helper function > gallium/u_blitter: accelerate depth-stencil copying using shader > stencil export > gallium/u_blitter: accelerate stencil-only copying > gallium/util: move pipe_tex_to_tgsi_tex helper function into u_inlines > gallium/u_blit: don't do two copies for non-2D textures > gallium/u_blit: drop not-very-useful wrapper around > util_blit_pixels_writemask > gallium/u_blit: remove useless memset calls > st/mesa: set colormask to zero when blitting depth > st/mesa: implement accelerated stencil blitting using shader stencil > export > r600g: move CB_TARGET_MASK setup into new cb_misc_state > r600g: move MULTIWRITE setup into cb_misc_state for r6xx-r7xx > r600g: move CB_SHADER_MASK setup into cb_misc_state > r600g: set DISABLE in CB_COLOR_CONTROL if colormask is 0 > r600g: remove is_flush from DSA state > r600g: do fine-grained depth texture flushing > r600g: flush depth textures bound to vertex shaders > r600g: don't set dirty_db_mask for a flushed depth texture > r600g: don't flush depth textures set as colorbuffers > r600g: fix uploading non-zero mipmap levels of depth textures > > src/gallium/auxiliary/postprocess/pp_mlaa.c|3 +- > src/gallium/auxiliary/postprocess/pp_run.c |3 +- > src/gallium/auxiliary/util/u_blit.c| 337 > +--- > src/gallium/auxiliary/util/u_blit.h| 20 +- > src/gallium/auxiliary/util/u_blitter.c | 153 ++--- > src/gallium/auxiliary/util/u_blitter.h |4 +- > src/gallium/auxiliary/util/u_format.h | 29 ++ > src/gallium/auxiliary/util/u_gen_mipmap.c | 84 +++-- > src/gallium/auxiliary/util/u_inlines.h | 26 ++ > src/gallium/auxiliary/util/u_simple_shaders.c | 100 ++ > src/gallium/auxiliary/util/u_simple_shaders.h | 12 + > src/gallium/drivers/r600/evergreen_compute.c |3 +- > .../drivers/r600/evergreen_compute_internal.c |2 +- > src/gallium/drivers/r600/evergreen_hw_context.c|4 - > src/gallium/drivers/r600/evergreen_state.c | 36 ++- > src/gallium/drivers/r600/evergreend.h |2 + > src/gallium/drivers/r600/r600_blit.c | 99 +++--- > src/gallium/drivers/r600/r600_hw_context.c |4 +- > src/gallium/drivers/r600/r600_pipe.h | 27 +- > src/gallium/drivers/r600/r600_resource.h |6 +- > src/gallium/drivers/r600/r600_state.c | 42 ++- > src/gallium/drivers/r600/r600_state_common.c | 88 ++--- > src/gallium/drivers/r600/r600_texture.c| 23 +- > src/gallium/drivers/r600/r600d.h |2 + > src/mesa/main/formats.c|6 +- > src/mesa/state_tracker/st_cb_blit.c| 37 ++- > src/mesa/state_tracker/st_cb_drawpixels.c | 23 +- > src/mesa/state_tracker/st_cb_texture.c | 32 +- > src/mesa/state_tracker/st_context.c|2 + > src/mesa/state_tracker/st_context.h|2 +- > 30 files changed, 838 insertions(+), 373 deletions(-) > > Marek > ___ > mesa-dev mailing list > mesa-dev@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/mesa-dev ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] Mesa (master): draw: fix flat shading and screen-space linear interpolation in clipper
- Original Message - > On Tue, Jul 10, 2012 at 3:30 PM, Jose Fonseca > wrote: > > Yep. The interfaces are busted. > > > > Without native integers we get > > > >DCL SV[0], INSTANCEID > >... > >ARL ADDR[0].x, SV[0]. > > > > and with integers we get > > > >DCL SV[0], INSTANCEID > >... > >UARL ADDR[0].x, SV[0]. > > > > Olivier's fix is incorrect. It works on the above cases by guessing > > the type, but if we have: > > > >DCL SV[0], INSTANCEID > >... > >MOV TEMP[0].x, SV[0]. > >ARL ADDR[0].x, TEMP[0]. > > > > and > > > >DCL SV[0], INSTANCEID > >... > >MOV TEMP[0].x, SV[0]. > >UARL ADDR[0].x, TEMP[0]. > > > > it is impossible to guess -- the change is merely replacing a bug > > with another. > > > > > > For the record, the problem also happens without LLVM: > > > > DRAW_USE_LLVM=0 draw-instanced -auto > > > > > > > > Anyway, AFAICT, all hardware out there that really supports > > INSTANCEID/VERTEXID also supports native integers, so this is a > > problem specific to the draw module. > > > > But given that draw module can support anything, this is actually > > self inflicted! In short, INSTANCEID/VERTEXID without integers is > > an historic artifact, that should not exist going forward. > > > > The right fix is merely making sure that PIPE_SHADER_CAP_INTEGERS > > is accurately advertised as 1 by the draw module (just like > > Stephane secretly did in 45fc069600ddbfe07a0a0cd5280161a8c7c55dd0 > > :) > > I just wanted to tell you Stephane's change cannot work and it even > has no effect at the moment. The native integer support is global in > core Mesa. It's because integer uniforms are converted to floats > based > on the global NativeInteger flag for all shader stages and that can't > be fixed easily, because uniforms can be shared between shaders. > Basically, all drivers must advertise integer support either for all > shader stages or none. I see.. > Given that, I only see two possible outcomes: > > 1) Disable INSTANCEID support in DX9-level drivers using Draw (that's > only i915g AFAIK) and only support INSTANCEID with integers. > 2) Let drivers set the type of system values in Draw, so that Draw > doesn't have to guess what it should be. My issue with 2) with this is that TGSI semantics depend on an external state. I think this is ugly and confusing. If Stephane's OK with 1), I'd prefer taking it. Jose ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] Mesa 8.0.4 release
Mesa 8.0.4 has been released. Mesa 8.0.4 is a bug fix release which fixes bugs found since the 8.0.3 release. The tag in the GIT repository for Mesa 8.0.4 is 'mesa-8.0.4'. Mesa 8.0.4 is available for download at ftp://freedesktop.org/pub/mesa/8.0.4/ I was originally planning for Mesa 8.0.4 to be released back on June 15th and Mesa 8.0.5 to be released on July 20th. Things happened (as they always do!), and 8.0.4 got delayed. I don't see a lot of value in doing another release in 10 days. Unless there are objections, I think we'll just skip the July release. The next releases will be mid-August. As far as I know, the plan is still to release Mesa 8.1 and Mesa 8.0.next simultaneously on (or around) August 17th. md5sums: d546f988adfdf986cff45b1efa2d8a46 MesaLib-8.0.4.tar.bz2 02b96082d2f1ad72e7385f4022afafb9 MesaLib-8.0.4.tar.gz 1f0fdabe6e8019d4de6c16e20e74d163 MesaLib-8.0.4.zip I have verified building from the .tar.bz2 file by doing: tar -xjf MesaLib-8.0.4.tar.bz2 cd Mesa-8.0.4 ./configure --enable-dri --with-dri-drivers=yes \ --enable-gles1 --enable-gles2 --enable-glx-tls \ --disable-glu --with-gallium-drivers \ --enable-gallium-llvm make -j6 I have also verified that I pushed all the cherry picks and the tag. ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] Mesa (master): draw: fix flat shading and screen-space linear interpolation in clipper
On Tue, Jul 10, 2012 at 9:04 AM, Jose Fonseca wrote: > > > - Original Message - >> On Tue, Jul 10, 2012 at 3:30 PM, Jose Fonseca >> wrote: >> > Yep. The interfaces are busted. >> > >> > Without native integers we get >> > >> >DCL SV[0], INSTANCEID >> >... >> >ARL ADDR[0].x, SV[0]. >> > >> > and with integers we get >> > >> >DCL SV[0], INSTANCEID >> >... >> >UARL ADDR[0].x, SV[0]. >> > >> > Olivier's fix is incorrect. It works on the above cases by guessing >> > the type, but if we have: >> > >> >DCL SV[0], INSTANCEID >> >... >> >MOV TEMP[0].x, SV[0]. >> >ARL ADDR[0].x, TEMP[0]. >> > >> > and >> > >> >DCL SV[0], INSTANCEID >> >... >> >MOV TEMP[0].x, SV[0]. >> >UARL ADDR[0].x, TEMP[0]. >> > >> > it is impossible to guess -- the change is merely replacing a bug >> > with another. >> > >> > >> > For the record, the problem also happens without LLVM: >> > >> > DRAW_USE_LLVM=0 draw-instanced -auto >> > >> > >> > >> > Anyway, AFAICT, all hardware out there that really supports >> > INSTANCEID/VERTEXID also supports native integers, so this is a >> > problem specific to the draw module. >> > >> > But given that draw module can support anything, this is actually >> > self inflicted! In short, INSTANCEID/VERTEXID without integers is >> > an historic artifact, that should not exist going forward. >> > >> > The right fix is merely making sure that PIPE_SHADER_CAP_INTEGERS >> > is accurately advertised as 1 by the draw module (just like >> > Stephane secretly did in 45fc069600ddbfe07a0a0cd5280161a8c7c55dd0 >> > :) >> >> I just wanted to tell you Stephane's change cannot work and it even >> has no effect at the moment. The native integer support is global in >> core Mesa. It's because integer uniforms are converted to floats >> based >> on the global NativeInteger flag for all shader stages and that can't >> be fixed easily, because uniforms can be shared between shaders. >> Basically, all drivers must advertise integer support either for all >> shader stages or none. > > I see.. > >> Given that, I only see two possible outcomes: >> >> 1) Disable INSTANCEID support in DX9-level drivers using Draw (that's >> only i915g AFAIK) and only support INSTANCEID with integers. > >> 2) Let drivers set the type of system values in Draw, so that Draw >> doesn't have to guess what it should be. > > My issue with 2) with this is that TGSI semantics depend on an external > state. I think this is ugly and confusing. > > If Stephane's OK with 1), I'd prefer taking it. > There is also option 3): revert the two patches causing the regression. Stéphane ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [PATCH] docs/relnotes-8.0.4: fix html markup
--- docs/relnotes-8.0.4.html | 85 ++ 1 files changed, 41 insertions(+), 44 deletions(-) diff --git a/docs/relnotes-8.0.4.html b/docs/relnotes-8.0.4.html index dd89b95..529140a 100644 --- a/docs/relnotes-8.0.4.html +++ b/docs/relnotes-8.0.4.html @@ -1,16 +1,13 @@ - - +http://www.w3.org/TR/html4/loose.dtd";> + -Mesa Release Notes - - + + Mesa Release Notes + + - - - - -Mesa 8.0.4 Release Notes / July 10, 2012 +Mesa 8.0.4 Release Notes / July 10, 2012 Mesa 8.0.4 is a bug fix release which fixes bugs found since the 8.0.2 release. @@ -70,18 +67,18 @@ d546f988adfdf986cff45b1efa2d8a46 MesaLib-8.0.4.tar.bz2 git log mesa-8.0.3..mesa-8.0.4 -Andreas Betz (1): +Andreas Betz (1): vega: fix 565 color unpacking bug - + -Antoine Labour (2): +Antoine Labour (2): meta: Cleanup the resources we allocate. mesa: Free uniforms correclty. - + -Brian Paul (22): +Brian Paul (22): docs: add link to 8.0.3 release notes mesa: fix Z32_FLOAT -> uint conversion functions @@ -105,39 +102,39 @@ d546f988adfdf986cff45b1efa2d8a46 MesaLib-8.0.4.tar.bz2 st/mesa: fix mipmap image size computation w.r.t. texture arrays draw: fix missing immediates bug in polygon stipple code st/mesa: fix max_offset computation for base vertex - + -Christoph Bumiller (1): +Christoph Bumiller (1): nv50: handle NEG,ABS modifiers for short RCP encoding - + -Dylan Noblesmith (1): +Dylan Noblesmith (1): mesa: require GL_MAX_SAMPLES >= 4 for GL 3.0 - + -Eric Anholt (1): +Eric Anholt (1): i965/vs: Fix texelFetchOffset() - + -Ian Romanick (5): +Ian Romanick (5): docs: Add 8.0.3 release md5sums glx/tests: Fix off-by-one error in allocating extension string buffer glsl: Remove spurious printf messages glsl: Fix pi/2 constant in acos built-in function mesa: Bump version number to 8.0.4 - + -José Fonseca (2): +José Fonseca (2): mesa: Avoid void acinclude.m4 Android.common.mk Android.mk autogen.sh bin common.py configs configure.ac docs doxygen include Makefile scons SConstruct src tests arithmetic. draw: Ensure that prepare is always run after LLVM garbagge collection. - + -Kenneth Graunke (15): +Kenneth Graunke (15): mesa: Check for a negative "size" parameter in glCopyBufferSubData(). i965: Fix brw_swap_cmod() for LE/GE comparisons. @@ -154,51 +151,51 @@ d546f988adfdf986cff45b1efa2d8a46 MesaLib-8.0.4.tar.bz2 i965/fs: Fix user-defined FS outputs with less than four components. glsl: Hook up loop_variable_state destructor to plug a memory leak. glsl: Don't trust loop analysis in the presence of function calls. - + -Kurt Roeckx (1): +Kurt Roeckx (1): i830: Fix crash for GL_STENCIL_TEST in i830Enable() - + -Lukas Rössler (1): +Lukas Rössler (1): glu: fix two Clang warnings - + -Marek Olšák (2): +Marek Olšák (2): mesa: allow exposing GL3 without EXT_texture_integer st/mesa: don't do srgb->linear conversion in decompress_with_blit - + -Paul Seidler (1): +Paul Seidler (1): tests: include mesa headers - + -Stéphane Marchesin (3): +Stéphane Marchesin (3): glx: Handle a null reply in QueryVersion. i915g: Don't invert signalled/unsignalled fences i915g: Don't avoid flushing when we have a pending fence. - + -Thomas Gstädtner (1): +Thomas Gstädtner (1): gallium/targets: pass ldflags parameter to MKLIB - + -Vadim Girlin (2): +Vadim Girlin (2): st/mesa: set stObj->lastLevel in guess_and_alloc_texture r600g: check gpr count limit - + -Vinson Lee (1): +Vinson Lee (1): st/mesa: Fix uninitialized members in glsl_to_tgsi_visitor constructor. - + -- 1.7.1 ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH] docs/relnotes-8.0.4: fix html markup
Right... I forgot that the HTML formatting got changed around. I just copied the 8.0.3 release notes in the 8.0 branch as the starting point. I think it should be fine apply this to master, and I can pick it over to 8.0 branch later... though I don't think it will matter there. I'll just have to remember to not make the same mistake again. :) Reviewed-by: Ian Romanick On 07/10/2012 10:09 AM, Andreas Boll wrote: --- docs/relnotes-8.0.4.html | 85 ++ 1 files changed, 41 insertions(+), 44 deletions(-) diff --git a/docs/relnotes-8.0.4.html b/docs/relnotes-8.0.4.html index dd89b95..529140a 100644 --- a/docs/relnotes-8.0.4.html +++ b/docs/relnotes-8.0.4.html @@ -1,16 +1,13 @@ - - +http://www.w3.org/TR/html4/loose.dtd";> + -Mesa Release Notes - - + + Mesa Release Notes + + - - - - -Mesa 8.0.4 Release Notes / July 10, 2012 +Mesa 8.0.4 Release Notes / July 10, 2012 Mesa 8.0.4 is a bug fix release which fixes bugs found since the 8.0.2 release. @@ -70,18 +67,18 @@ d546f988adfdf986cff45b1efa2d8a46 MesaLib-8.0.4.tar.bz2 git log mesa-8.0.3..mesa-8.0.4 -Andreas Betz (1): +Andreas Betz (1): vega: fix 565 color unpacking bug - + -Antoine Labour (2): +Antoine Labour (2): meta: Cleanup the resources we allocate. mesa: Free uniforms correclty. - + -Brian Paul (22): +Brian Paul (22): docs: add link to 8.0.3 release notes mesa: fix Z32_FLOAT -> uint conversion functions @@ -105,39 +102,39 @@ d546f988adfdf986cff45b1efa2d8a46 MesaLib-8.0.4.tar.bz2 st/mesa: fix mipmap image size computation w.r.t. texture arrays draw: fix missing immediates bug in polygon stipple code st/mesa: fix max_offset computation for base vertex - + -Christoph Bumiller (1): +Christoph Bumiller (1): nv50: handle NEG,ABS modifiers for short RCP encoding - + -Dylan Noblesmith (1): +Dylan Noblesmith (1): mesa: require GL_MAX_SAMPLES >= 4 for GL 3.0 - + -Eric Anholt (1): +Eric Anholt (1): i965/vs: Fix texelFetchOffset() - + -Ian Romanick (5): +Ian Romanick (5): docs: Add 8.0.3 release md5sums glx/tests: Fix off-by-one error in allocating extension string buffer glsl: Remove spurious printf messages glsl: Fix pi/2 constant in acos built-in function mesa: Bump version number to 8.0.4 - + -José Fonseca (2): +José Fonseca (2): mesa: Avoid void acinclude.m4 Android.common.mk Android.mk autogen.sh bin common.py configs configure.ac docs doxygen include Makefile scons SConstruct src tests arithmetic. draw: Ensure that prepare is always run after LLVM garbagge collection. - + -Kenneth Graunke (15): +Kenneth Graunke (15): mesa: Check for a negative "size" parameter in glCopyBufferSubData(). i965: Fix brw_swap_cmod() for LE/GE comparisons. @@ -154,51 +151,51 @@ d546f988adfdf986cff45b1efa2d8a46 MesaLib-8.0.4.tar.bz2 i965/fs: Fix user-defined FS outputs with less than four components. glsl: Hook up loop_variable_state destructor to plug a memory leak. glsl: Don't trust loop analysis in the presence of function calls. - + -Kurt Roeckx (1): +Kurt Roeckx (1): i830: Fix crash for GL_STENCIL_TEST in i830Enable() - + -Lukas Rössler (1): +Lukas Rössler (1): glu: fix two Clang warnings - + -Marek Olšák (2): +Marek Olšák (2): mesa: allow exposing GL3 without EXT_texture_integer st/mesa: don't do srgb->linear conversion in decompress_with_blit - + -Paul Seidler (1): +Paul Seidler (1): tests: include mesa headers - + -Stéphane Marchesin (3): +Stéphane Marchesin (3): glx: Handle a null reply in QueryVersion. i915g: Don't invert signalled/unsignalled fences i915g: Don't avoid flushing when we have a pending fence. - + -Thomas Gstädtner (1): +Thomas Gstädtner (1): gallium/targets: pass ldflags parameter to MKLIB - + -Vadim Girlin (2): +Vadim Girlin (2): st/mesa: set stObj->lastLevel in guess_and_alloc_texture r600g: check gpr count limit - + -Vinson Lee (1): +Vinson Lee (1): st/mesa: Fix uninitialized members in glsl_to_tgsi_visitor constructor. - + ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH 22/22] r600g: fix uploading non-zero mipmap levels of depth textures
On Tue, Jul 10, 2012 at 11:44 AM, Michel Dänzer wrote: > On Mon, 2012-07-09 at 21:15 +0200, Marek Olšák wrote: >> This fixes piglit/depth-level-clamp. >> --- >> src/gallium/drivers/r600/r600_texture.c |5 +++-- >> 1 file changed, 3 insertions(+), 2 deletions(-) >> >> diff --git a/src/gallium/drivers/r600/r600_texture.c >> b/src/gallium/drivers/r600/r600_texture.c >> index a6a83ca..d16c252 100644 >> --- a/src/gallium/drivers/r600/r600_texture.c >> +++ b/src/gallium/drivers/r600/r600_texture.c >> @@ -926,8 +926,9 @@ void r600_texture_transfer_destroy(struct pipe_context >> *ctx, >> >> u_box_origin_2d(texture->width0, texture->height0, >> &sbox); >> >> - ctx->resource_copy_region(ctx, texture, 0, 0, 0, 0, >> - &rtransfer->staging->b.b, 0, >> + ctx->resource_copy_region(ctx, texture, >> transfer->level, >> + 0, 0, transfer->box.z, >> + &rtransfer->staging->b.b, >> transfer->level, > > Does this really need to take into account transfer->box.z, but not .x > and .y? Yes, it could be optimized more. I'd rather leave the optimization to some other patch. The read transfers always copy the whole layer too. Marek ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH] r600g: improve flushed depth texture handling v2
On Tue, Jul 10, 2012 at 6:40 AM, Vadim Girlin wrote: > On Sat, 2012-07-07 at 01:48 +0200, Marek Olšák wrote: >> On Wed, Jun 27, 2012 at 1:34 AM, Vadim Girlin wrote: >> > Use r600_resource_texture::flished_depth_texture for GPU access, and >> > allocate it in the VRAM. For transfers we'll allocate untiled texture in >> > the >> > GTT and store it in the r600_transfer::staging. >> > >> > Improves performance when flushed depth texture is frequently used by the >> > GPU (about 30% for Lightsmark). >> > >> > Signed-off-by: Vadim Girlin >> > --- >> > >> > Fixes fbo-clear-formats, fbo-generatemipmap-formats, no regressions on >> > evergreen >> >> Hi, >> >> is there any reason this patch hasn't been committed yet? >> > > Hi, > > I have some doubts because it was benchmarked by phoronix and there were > regressions, though I suspect that something is wrong with the results: > > http://www.phoronix.com/scan.php?page=article&item=amd_r600g_texdepth&num=4 > > I was going to look into it but had no time yet. I'd like to be sure > that there are no regressions before committing. Well, there's nothing wrong with your patch. I wouldn't trust benchmarks run with the Unity desktop so much. I myself had to switch from Unity 2D to Xfce just to get consistent results when testing performance. Now that your patch separates flushing for texturing and transfers, I think we could make it a little bit faster by imlementing an in-place flush for texturing (that is without having to allocate another resource). Marek ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [PATCH 1/5] glsl: Initialize coordinate to NULL in ir_texture constructor.
I ran into this while trying to create a TXS query, which doesn't have a
coordinate. Since it didn't get initialized to NULL, a bunch of
visitors tried to access it and crashed.
Most of the time, this won't be a problem, but it's just a good idea.
Signed-off-by: Kenneth Graunke
---
src/glsl/ir.h |3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/src/glsl/ir.h b/src/glsl/ir.h
index b54e2f2..9bbf3b7 100644
--- a/src/glsl/ir.h
+++ b/src/glsl/ir.h
@@ -1341,7 +1341,8 @@ enum ir_texture_opcode {
class ir_texture : public ir_rvalue {
public:
ir_texture(enum ir_texture_opcode op)
- : op(op), projector(NULL), shadow_comparitor(NULL), offset(NULL)
+ : op(op), coordinate(NULL), projector(NULL), shadow_comparitor(NULL),
+offset(NULL)
{
this->ir_type = ir_type_texture;
}
--
1.7.10.4
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] Reworking textureGrad on shadow samplers
Intel hardware doesn't natively support textureGrad with shadow comparisons. This patch series implements a GLSL IR level lowering pass to compute the LOD level that would be selected based on the given gradient values, based on the equations on page 205 of the OpenGL 3.0 specification. It then simply uses textureLod(). The old method had many problems: - The shadow comparison was done post-filtering. - It required state-dependent recompiles whenever the comparison function changed. - It didn't even work: many cases hit assertion failures. - I never implemented it for the VS. The new method is not perfect; it only passes 36/45 of oglconform's shadow-grad tests. That said, it seems to work a lot better than the previous method, and the remaining cases are so obscure that I'd like to just push it as is and fix up the remaining cases sometime in the future (i.e. bug reports welcome). One of the main advantages is that this reduces state-dependent recompiles, and thus increases performance certain applications. (Said applications aren't actually using TXD - we were just being overzealous in our program key handling.) ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [PATCH 2/5] glsl/ir_builder: Add a generic constructor for unary expressions.
I needed to compute logs and square roots in a patch I was working on,
and wanted to use the convenient interface. We already have a similar
constructor for binops; adding one for unops seems reasonable.
Signed-off-by: Kenneth Graunke
---
src/glsl/ir_builder.cpp |8
src/glsl/ir_builder.h |1 +
2 files changed, 9 insertions(+)
diff --git a/src/glsl/ir_builder.cpp b/src/glsl/ir_builder.cpp
index 9a16c90..fc6ee96 100644
--- a/src/glsl/ir_builder.cpp
+++ b/src/glsl/ir_builder.cpp
@@ -143,6 +143,14 @@ swizzle_xyzw(operand a)
}
ir_expression *
+expr(ir_expression_operation op, operand a)
+{
+ void *mem_ctx = ralloc_parent(a.val);
+
+ return new(mem_ctx) ir_expression(op, a.val);
+}
+
+ir_expression *
expr(ir_expression_operation op, operand a, operand b)
{
void *mem_ctx = ralloc_parent(a.val);
diff --git a/src/glsl/ir_builder.h b/src/glsl/ir_builder.h
index 0ebcbab..410b08c 100644
--- a/src/glsl/ir_builder.h
+++ b/src/glsl/ir_builder.h
@@ -83,6 +83,7 @@ public:
ir_assignment *assign(deref lhs, operand rhs);
ir_assignment *assign(deref lhs, operand rhs, int writemask);
+ir_expression *expr(ir_expression_operation op, operand a);
ir_expression *expr(ir_expression_operation op, operand a, operand b);
ir_expression *add(operand a, operand b);
ir_expression *sub(operand a, operand b);
--
1.7.10.4
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [PATCH 3/5] glsl/ir_builder: Add a new swizzle_for_size() function.
This swizzles away unwanted components, while preserving the order of
the ones that remain.
Signed-off-by: Kenneth Graunke
---
src/glsl/ir_builder.cpp | 15 +++
src/glsl/ir_builder.h |5 +
2 files changed, 20 insertions(+)
diff --git a/src/glsl/ir_builder.cpp b/src/glsl/ir_builder.cpp
index fc6ee96..d96e25c 100644
--- a/src/glsl/ir_builder.cpp
+++ b/src/glsl/ir_builder.cpp
@@ -77,6 +77,21 @@ swizzle(operand a, int swizzle, int components)
}
ir_swizzle *
+swizzle_for_size(operand a, int components)
+{
+ void *mem_ctx = ralloc_parent(a.val);
+
+ if (a.val->type->vector_elements < components)
+ components = a.val->type->vector_elements;
+
+ unsigned s[4] = { 0, 1, 2, 3 };
+ for (int i = components; i < 4; i++)
+ s[i] = components - 1;
+
+ return new(mem_ctx) ir_swizzle(a.val, s, components);
+}
+
+ir_swizzle *
swizzle_(operand a)
{
return swizzle(a, SWIZZLE_, 4);
diff --git a/src/glsl/ir_builder.h b/src/glsl/ir_builder.h
index 410b08c..7a0a196 100644
--- a/src/glsl/ir_builder.h
+++ b/src/glsl/ir_builder.h
@@ -91,6 +91,11 @@ ir_expression *mul(operand a, operand b);
ir_expression *dot(operand a, operand b);
ir_expression *saturate(operand a);
+/**
+ * Swizzle away later components, but preserve the ordering.
+ */
+ir_swizzle *swizzle_for_size(operand a, int components);
+
ir_swizzle *swizzle_(operand a);
ir_swizzle *swizzle_(operand a);
ir_swizzle *swizzle_(operand a);
--
1.7.10.4
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [PATCH 4/5] i965: Add a lowering pass to convert TXD to TXL by computing the LOD.
Intel hardware doesn't natively support textureGrad with shadow
comparisons. So we need to generate code to handle it somehow.
Based on the equations of page 205 of the OpenGL 3.0 specification,
it's possible to compute the LOD value that would be selected given the
gradient values. Then, we can simply convert the TXD to a TXL.
Currently, this passes 38/46 of oglconform's shadow-grad subtests,
and selects the wrong level for the other 6 (but doesn't crash).
While we should investigate the remaining failures, this new method at
least works better than the old one. Texturing with explicit gradients,
shadow sampling, and either projection or texel offsets is pretty
obscure anyway.
Signed-off-by: Kenneth Graunke
---
src/mesa/drivers/dri/i965/Makefile.sources |1 +
src/mesa/drivers/dri/i965/brw_context.h|1 +
.../dri/i965/brw_lower_texture_gradients.cpp | 153
src/mesa/drivers/dri/i965/brw_shader.cpp |1 +
4 files changed, 156 insertions(+)
create mode 100644 src/mesa/drivers/dri/i965/brw_lower_texture_gradients.cpp
diff --git a/src/mesa/drivers/dri/i965/Makefile.sources
b/src/mesa/drivers/dri/i965/Makefile.sources
index 82143ac..334bfd9 100644
--- a/src/mesa/drivers/dri/i965/Makefile.sources
+++ b/src/mesa/drivers/dri/i965/Makefile.sources
@@ -125,6 +125,7 @@ i965_CXX_FILES = \
brw_fs_reg_allocate.cpp \
brw_fs_schedule_instructions.cpp \
brw_fs_vector_splitting.cpp \
+ brw_lower_texture_gradients.cpp \
brw_shader.cpp \
brw_vec4.cpp \
brw_vec4_emit.cpp \
diff --git a/src/mesa/drivers/dri/i965/brw_context.h
b/src/mesa/drivers/dri/i965/brw_context.h
index c1f2314..b4868fe 100644
--- a/src/mesa/drivers/dri/i965/brw_context.h
+++ b/src/mesa/drivers/dri/i965/brw_context.h
@@ -1276,6 +1276,7 @@ brw_program_reloc(struct brw_context *brw, uint32_t
state_offset,
}
bool brw_do_cubemap_normalize(struct exec_list *instructions);
+bool brw_lower_texture_gradients(struct exec_list *instructions);
#ifdef __cplusplus
}
diff --git a/src/mesa/drivers/dri/i965/brw_lower_texture_gradients.cpp
b/src/mesa/drivers/dri/i965/brw_lower_texture_gradients.cpp
new file mode 100644
index 000..2d4f2ab
--- /dev/null
+++ b/src/mesa/drivers/dri/i965/brw_lower_texture_gradients.cpp
@@ -0,0 +1,153 @@
+/*
+ * Copyright © 2012 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+/**
+ * \file brw_lower_texture_gradients.cpp
+ */
+
+#include "glsl/ir.h"
+#include "glsl/ir_builder.h"
+
+using namespace ir_builder;
+
+class lower_texture_grad_visitor : public ir_hierarchical_visitor {
+public:
+ lower_texture_grad_visitor()
+ {
+ progress = false;
+ }
+
+ ir_visitor_status visit_leave(ir_texture *ir);
+
+
+ bool progress;
+
+private:
+ void emit(ir_variable *, ir_rvalue *);
+};
+
+/**
+ * Emit a variable declaration and an assignment to initialize it.
+ */
+void
+lower_texture_grad_visitor::emit(ir_variable *var, ir_rvalue *value)
+{
+ base_ir->insert_before(var);
+ base_ir->insert_before(assign(var, value));
+}
+
+static const glsl_type *
+txs_type(const glsl_type *type)
+{
+ unsigned dims;
+ switch (type->sampler_dimensionality) {
+ case GLSL_SAMPLER_DIM_1D:
+ dims = 1;
+ break;
+ case GLSL_SAMPLER_DIM_2D:
+ case GLSL_SAMPLER_DIM_RECT:
+ case GLSL_SAMPLER_DIM_CUBE:
+ dims = 2;
+ break;
+ case GLSL_SAMPLER_DIM_3D:
+ dims = 3;
+ break;
+ default:
+ assert(!"Should not get here: invalid sampler dimensionality");
+ }
+
+ if (type->sampler_array)
+ dims++;
+
+ return glsl_type::get_instance(GLSL_TYPE_INT, dims, 1);
+}
+
+ir_visitor_status
+lower_texture_grad_visitor::visit_leave(ir_texture *ir)
+{
+ /* Only lower textureGrad with shadow samplers */
+ if (ir->op != ir_txd || !ir->shadow_comparitor)
+ return visit_continue;
+
+ /* Cube
[Mesa-dev] [PATCH 5/5] i965: Delete previous workaround for textureGrad with shadow samplers.
It had many problems:
- The shadow comparison was done post-filtering.
- It required state-dependent recompiles whenever the comparison
function changed.
- It didn't even work: many cases hit assertion failures.
- I never implemented it for the VS.
The new lowering pass which converts textureGrad to textureLod by
computing the LOD value works much better, and eliminates the
state-dependent recompiles that were contributing to a performance
problem in a closed-source application we're investigating.
Signed-off-by: Kenneth Graunke
---
src/mesa/drivers/dri/i965/brw_fs.cpp |3 --
src/mesa/drivers/dri/i965/brw_fs_visitor.cpp | 62 +++---
src/mesa/drivers/dri/i965/brw_program.h | 12 -
src/mesa/drivers/dri/i965/brw_wm.c |3 --
4 files changed, 5 insertions(+), 75 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/brw_fs.cpp
b/src/mesa/drivers/dri/i965/brw_fs.cpp
index 175e36e..b3b25cc 100644
--- a/src/mesa/drivers/dri/i965/brw_fs.cpp
+++ b/src/mesa/drivers/dri/i965/brw_fs.cpp
@@ -2163,9 +2163,6 @@ brw_fs_precompile(struct gl_context *ctx, struct
gl_shader_program *prog)
key.clamp_fragment_color = true;
for (int i = 0; i < BRW_MAX_TEX_UNIT; i++) {
- if (fp->Base.ShadowSamplers & (1 << i))
-key.tex.compare_funcs[i] = GL_LESS;
-
/* FINISHME: depth compares might use (0,0,0,W) for example */
key.tex.swizzles[i] = SWIZZLE_XYZW;
}
diff --git a/src/mesa/drivers/dri/i965/brw_fs_visitor.cpp
b/src/mesa/drivers/dri/i965/brw_fs_visitor.cpp
index 28adf33..7224cbe 100644
--- a/src/mesa/drivers/dri/i965/brw_fs_visitor.cpp
+++ b/src/mesa/drivers/dri/i965/brw_fs_visitor.cpp
@@ -712,7 +712,7 @@ fs_visitor::emit_texture_gen4(ir_texture *ir, fs_reg dst,
fs_reg coordinate,
/* g0 header. */
mlen = 1;
- if (ir->shadow_comparitor && ir->op != ir_txd) {
+ if (ir->shadow_comparitor) {
for (int i = 0; i < ir->coordinate->type->vector_elements; i++) {
emit(BRW_OPCODE_MOV, fs_reg(MRF, base_mrf + mlen + i), coordinate);
coordinate.reg_offset++;
@@ -931,7 +931,7 @@ fs_visitor::emit_texture_gen5(ir_texture *ir, fs_reg dst,
fs_reg coordinate,
}
mlen += vector_elements * reg_width;
- if (ir->shadow_comparitor && ir->op != ir_txd) {
+ if (ir->shadow_comparitor) {
mlen = MAX2(mlen, header_present + 4 * reg_width);
ir->shadow_comparitor->accept(this);
@@ -1038,7 +1038,7 @@ fs_visitor::emit_texture_gen7(ir_texture *ir, fs_reg dst,
fs_reg coordinate,
base_mrf--;
}
- if (ir->shadow_comparitor && ir->op != ir_txd) {
+ if (ir->shadow_comparitor) {
ir->shadow_comparitor->accept(this);
emit(BRW_OPCODE_MOV, fs_reg(MRF, base_mrf + mlen), this->result);
mlen += reg_width;
@@ -1163,19 +1163,6 @@ fs_visitor::visit(ir_texture *ir)
int sampler = _mesa_get_sampler_uniform_value(ir->sampler, prog, &fp->Base);
sampler = fp->Base.SamplerUnits[sampler];
- /* Our hardware doesn't have a sample_d_c message, so shadow compares
-* for textureGrad/TXD need to be emulated with instructions.
-*/
- bool hw_compare_supported = ir->op != ir_txd;
- if (ir->shadow_comparitor && !hw_compare_supported) {
- assert(c->key.tex.compare_funcs[sampler] != GL_NONE);
- /* No need to even sample for GL_ALWAYS or GL_NEVER...bail early */
- if (c->key.tex.compare_funcs[sampler] == GL_ALWAYS)
-return swizzle_result(ir, fs_reg(1.0f), sampler);
- else if (c->key.tex.compare_funcs[sampler] == GL_NEVER)
-return swizzle_result(ir, fs_reg(0.0f), sampler);
- }
-
if (ir->coordinate)
ir->coordinate->accept(this);
fs_reg coordinate = this->result;
@@ -1325,47 +1312,8 @@ fs_visitor::visit(ir_texture *ir)
inst->sampler = sampler;
- if (ir->shadow_comparitor) {
- if (hw_compare_supported) {
-inst->shadow_compare = true;
- } else {
-ir->shadow_comparitor->accept(this);
-fs_reg ref = this->result;
-
-fs_reg value = dst;
-dst = fs_reg(this, glsl_type::vec4_type);
-
-/* FINISHME: This needs to be done pre-filtering. */
-
-uint32_t conditional = 0;
-switch (c->key.tex.compare_funcs[sampler]) {
-/* GL_ALWAYS and GL_NEVER were handled at the top of the function */
-case GL_LESS: conditional = BRW_CONDITIONAL_L; break;
-case GL_GREATER: conditional = BRW_CONDITIONAL_G; break;
-case GL_LEQUAL: conditional = BRW_CONDITIONAL_LE; break;
-case GL_GEQUAL: conditional = BRW_CONDITIONAL_GE; break;
-case GL_EQUAL:conditional = BRW_CONDITIONAL_EQ; break;
-case GL_NOTEQUAL: conditional = BRW_CONDITIONAL_NEQ; break;
-default: assert(!"Should not get here: bad shadow compare function");
-}
-
-/* Use conditional moves to load 0 or 1 as the result */
-this->current_annotation = "manual shadow comparison";
-for (int i = 0; i < 4; i++) {
-
[Mesa-dev] [Bug 51956] New: EnumStrings.LookUpByNumber regression
https://bugs.freedesktop.org/show_bug.cgi?id=51956 Bug #: 51956 Summary: EnumStrings.LookUpByNumber regression Classification: Unclassified Product: Mesa Version: git Platform: x86-64 (AMD64) OS/Version: Linux (All) Status: NEW Severity: normal Priority: medium Component: Mesa core AssignedTo: mesa-dev@lists.freedesktop.org ReportedBy: v...@freedesktop.org CC: mar...@gmail.com mesa: 67a8ee891b2e119d826d8f830c1cbbe64ecb4f82 (master) $ make check [...] [==] Running 4 tests from 1 test case. [--] Global test environment set-up. [--] 4 tests from EnumStrings [ RUN ] EnumStrings.LookUpByNumber enum_strings.cpp:54: Failure Value of: _mesa_lookup_enum_by_nr(everything[i].value) Actual: "GL_TIME_ELAPSED" Expected: everything[i].name Which is: "GL_TIME_ELAPSED_EXT" [ FAILED ] EnumStrings.LookUpByNumber (1 ms) 4952caad2dbf8668d86c3043b5761c61382bbb4c is the first bad commit commit 4952caad2dbf8668d86c3043b5761c61382bbb4c Author: Marek Olšák Date: Tue Jun 26 01:36:44 2012 +0200 glapi: add ARB_timer_query Reviewed-by: Brian Paul :04 04 7e6016519ee95a7681a00bb7df928403f54bdbd2 e260426c8823c355cc6ca16eb789bed0beeeddaf M src bisect run success -- Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are the assignee for the bug. ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH 4/5] mesa: Add function for decoding ETC1 textures
On 07/09/2012 10:46 AM, Chad Versace wrote:
> Add function _mesa_etc1_image_decode_to_abgr(). It is intended to be
> used by glCompressedTexSubImage2D to translate ETC1 textures into ABGR.
>
> Signed-off-by: Chad Versace
> ---
> src/mesa/main/texcompress_etc.c | 82
> +
> src/mesa/main/texcompress_etc.h | 6 +++
> 2 files changed, 88 insertions(+)
>
> diff --git a/src/mesa/main/texcompress_etc.c b/src/mesa/main/texcompress_etc.c
> index e3b2e83..d84ee0e 100644
> --- a/src/mesa/main/texcompress_etc.c
> +++ b/src/mesa/main/texcompress_etc.c
> @@ -178,3 +178,85 @@ _mesa_fetch_texel_2d_f_etc1_rgb8(const struct
> swrast_texture_image *texImage,
> texel[BCOMP] = UBYTE_TO_FLOAT(bgr888[2]);
> texel[ACOMP] = 1.0f;
> }
> +
> +static int
> +align(int x, int m)
> +{
> + return (m * x + m - 1) / m;
> +}
Perhaps we should just move the ALIGN and ROUND_DOWN_TO macros from
intel_context.h into macros.h?
For reference, ALIGN() is:
#define ALIGN(value, alignment) (((value) + alignment - 1) &
~(alignment - 1))
Either way, this patch looks right to me.
Reviewed-by: Kenneth Graunke
> +
> +/**
> + * Decode texture data in format `MESA_FORMAT_ETC1_RGB8` to
> + * `MESA_FORMAT_ABGR`.
> + *
> + * The size of the source data must be a multiple of the ETC1 block size,
> + * which is 8, even if the texture image's dimensions are not aligned to 4.
> + * From the GL_OES_compressed_ETC1_RGB8_texture spec:
> + * The texture is described as a number of 4x4 pixel blocks. If the
> + * texture (or a particular mip-level) is smaller than 4 pixels in
> + * any dimension (such as a 2x2 or a 8x1 texture), the texture is
> + * found in the upper left part of the block(s), and the rest of the
> + * pixels are not used. For instance, a texture of size 4x2 will be
> + * placed in the upper half of a 4x4 block, and the lower half of the
> + * pixels in the block will not be accessed.
> + *
> + * \param src_width in pixels
> + * \param src_height in pixels
> + * \param dst_rowstride in bytes
> + */
> +void
> +_mesa_etc1_image_decode_to_abgr(const uint8_t *src,
> +int src_width,
> +int src_height,
> +uint8_t *dst,
> +int dst_rowstride)
> +{
> + /* An XRGB pixel is 4 bytes. */
> + const int pixel_bytes = 4;
> +
> + /* A block contains 4x4 pixels. */
> + const int block_width = 4;
> + const int block_height = 4;
> +
> + /* A block contains 8 bytes. */
> + const int block_bytes = 8;
> +
> + /* The stride in units of blocks. */
> + const int block_stride = align(src_width, block_width) / block_width;
> +
> + /* The block being decoded. */
> + int block_x;
> + int block_y;
> +
> + /* Iterate over each block. */
> + for (block_y = 0; block_y * block_height < src_height; ++block_y) {
> + for (block_x = 0; block_x * block_width < src_width; ++block_x) {
> + struct etc1_block block;
> +
> + /* The pixel's position relative to the block's base. */
> + int pixel_x;
> + int pixel_y;
> +
> + const uint8_t *src_block_base =
> + src + block_bytes * (block_y * block_stride + block_x);
> +
> + uint8_t *dst_block_base = dst
> + + block_y * block_height * dst_rowstride
> + + block_x * block_width * pixel_bytes;
> +
> + etc1_parse_block(&block, src_block_base);
> +
> + /* Iterate over each pixel in block. */
> + for (pixel_y = 0; pixel_y < block_height; ++pixel_y) {
> +for (pixel_x = 0; pixel_x < block_width; ++pixel_x) {
> + uint8_t *dst_texel = dst_block_base
> + + pixel_y * dst_rowstride
> + + pixel_x * pixel_bytes;
> +
> + etc1_fetch_texel(&block, pixel_x, pixel_y, dst_texel);
> + dst_texel[3] = 0xff;
> +}
> + }
> + }
> + }
> +}
> diff --git a/src/mesa/main/texcompress_etc.h b/src/mesa/main/texcompress_etc.h
> index 8e8427f..552a7bd 100644
> --- a/src/mesa/main/texcompress_etc.h
> +++ b/src/mesa/main/texcompress_etc.h
> @@ -36,5 +36,11 @@ _mesa_texstore_etc1_rgb8(TEXSTORE_PARAMS);
> void
> _mesa_fetch_texel_2d_f_etc1_rgb8(const struct swrast_texture_image *texImage,
> GLint i, GLint j, GLint k, GLfloat *texel);
> +void
> +_mesa_etc1_image_decode_to_abgr(const uint8_t *src,
> +int src_width,
> +int src_height,
> +uint8_t *dst,
> +int dst_rowstride);
>
> #endif
>
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH 1/5] mesa: Fold texcompress_etc1_tmp.h into texcompress_etc.c
On 07/09/2012 10:46 AM, Chad Versace wrote: > The header defined the functions in with C-style macro templates. Since > there was only one user of the header, the templates were unnecessary. > This patch redefines the functions as vanilla, non-templated functions. > > This patch also removes the header and folds its contents into > texcompress_etc.c. The header defined many inline helper functions for > texcompress_etc.c, but, since no other file included it, it never made > sense to keep the header separate from the .c file. > > Signed-off-by: Chad Versace > --- > src/mesa/main/texcompress_etc.c | 119 +-- > src/mesa/main/texcompress_etc_tmp.h | 136 > > 2 files changed, 112 insertions(+), 143 deletions(-) > delete mode 100644 src/mesa/main/texcompress_etc_tmp.h This is also used by src/gallium/auxiliary/util/u_format_etc.c. Otherwise, this series looks good to me! Reviewed-by: Kenneth Graunke ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH 3/5] build: fix location of generated files in src/mesa
On Sun, Jul 8, 2012 at 10:44 PM, Christopher James Halse Rogers wrote: > On Sun, 2012-07-08 at 19:39 -0700, Kenneth Graunke wrote: >> On 07/08/2012 04:14 PM, Christopher James Halse Rogers wrote: >> > Signed-off-by: Christopher James Halse Rogers >> > >> > --- >> > src/mesa/sources.mak |9 + >> > 1 file changed, 5 insertions(+), 4 deletions(-) >> > >> > diff --git a/src/mesa/sources.mak b/src/mesa/sources.mak >> > index 16b1c39..87191fd 100644 >> > --- a/src/mesa/sources.mak >> > +++ b/src/mesa/sources.mak >> > @@ -4,8 +4,8 @@ SRCDIR ?= . >> > >> > # this is part of MAIN_FILES >> > MAIN_ES_FILES = \ >> > - $(SRCDIR)/main/api_exec_es1.c \ >> > - $(SRCDIR)/main/api_exec_es2.c >> > + $(TOP)/src/mesa/main/api_exec_es1.c \ >> > + $(TOP)/src/mesa/main/api_exec_es2.c >> >> Could we use $(top_builddir) instead of $(TOP)? As your first patch >> demonstrates, there are really two top-level directories, so $(TOP) is >> ambiguous/confusing. I'd rather see us stop using it, for clarity. > > It's included in a number of non-automake Makefiles, where IIUC > $(top_builddir) is not normally defined. > > On the other hand, making these Makefiles define $(top_builddir) instead > of $(TOP) would be easy, and probably makes sense given the transition > to autotools. I'll do that. Ultimately, I think we should kill off $(TOP). It seems to often be the source of top_srcdir/top_builddir confusion. ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [Bug 51960] New: [bisected] Mesa master fails to build
https://bugs.freedesktop.org/show_bug.cgi?id=51960 Bug #: 51960 Summary: [bisected] Mesa master fails to build Classification: Unclassified Product: Mesa Version: git Platform: Other OS/Version: All Status: NEW Keywords: regression Severity: normal Priority: medium Component: Mesa core AssignedTo: mesa-dev@lists.freedesktop.org ReportedBy: pavel.ondra...@email.cz CC: bri...@vmware.com Created attachment 64082 --> https://bugs.freedesktop.org/attachment.cgi?id=64082 build log CXXLD libdricore8.1.0.la .libs/execmem.o: file not recognized: File truncated collect2: error: ld returned 1 exit status gmake[4]: *** [libdricore8.1.0.la] Error 1 Full build log attached. I'm building with "./autogen.sh --enable-debug --with-gallium-drivers=swrast,r300 --with-dri-drivers= --disable-egl --enable-texture-float --disable-glu". 51b0a0b33ca16599a5972c3e77607284a1b7fb9b is the first bad commit commit 51b0a0b33ca16599a5972c3e77607284a1b7fb9b Author: Brian Paul Date: Tue Jun 26 14:47:19 2012 -0600 mesa: update glext.h to version 81 I double checked the bisect, and also reverting this commit from master fixes it. This is up to date Fedora 17, gcc (GCC) 4.7.0 20120507 (Red Hat 4.7.0-5), -- Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are the assignee for the bug. ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH] docs/relnotes-8.0.4: fix html markup
2012/7/10 Ian Romanick : > Right... I forgot that the HTML formatting got changed around. I just > copied the 8.0.3 release notes in the 8.0 branch as the starting point. I > think it should be fine apply this to master, and I can pick it over to 8.0 > branch later... though I don't think it will matter there. I'll just have > to remember to not make the same mistake again. :) > > Reviewed-by: Ian Romanick > > No problem. Please push this patch, I don't have commit access. Andreas. ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH 1/5] mesa: Fold texcompress_etc1_tmp.h into texcompress_etc.c
On 07/10/2012 11:42 AM, Kenneth Graunke wrote: > On 07/09/2012 10:46 AM, Chad Versace wrote: >> The header defined the functions in with C-style macro templates. Since >> there was only one user of the header, the templates were unnecessary. >> This patch redefines the functions as vanilla, non-templated functions. >> >> This patch also removes the header and folds its contents into >> texcompress_etc.c. The header defined many inline helper functions for >> texcompress_etc.c, but, since no other file included it, it never made >> sense to keep the header separate from the .c file. >> >> Signed-off-by: Chad Versace >> --- >> src/mesa/main/texcompress_etc.c | 119 +-- >> src/mesa/main/texcompress_etc_tmp.h | 136 >> >> 2 files changed, 112 insertions(+), 143 deletions(-) >> delete mode 100644 src/mesa/main/texcompress_etc_tmp.h > > This is also used by src/gallium/auxiliary/util/u_format_etc.c. Argh, I fail at grep. I'll de-templatize the header but not delete it. > Otherwise, this series looks good to me! > Reviewed-by: Kenneth Graunke ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] Mesa master branch: forced update
I just fetched from the master branch of the fdo mesa repo and was greeted with a "forced update" message, and the gitweb interface shows several days of history are missing from the master branch. olv appears to be the last user to modify the master branch: tstellar@annarchy:~$ ls -l /git/mesa/mesa.git/refs/heads/master -rw-rw-r-- 1 olv mesa 41 Jul 10 11:41 /git/mesa/mesa.git/refs/heads/master Anyone know what happened? -Tom ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH] r600g: improve flushed depth texture handling v2
On Tue, Jul 10, 2012 at 2:10 PM, Marek Olšák wrote: > On Tue, Jul 10, 2012 at 6:40 AM, Vadim Girlin wrote: >> On Sat, 2012-07-07 at 01:48 +0200, Marek Olšák wrote: >>> On Wed, Jun 27, 2012 at 1:34 AM, Vadim Girlin wrote: >>> > Use r600_resource_texture::flished_depth_texture for GPU access, and >>> > allocate it in the VRAM. For transfers we'll allocate untiled texture in >>> > the >>> > GTT and store it in the r600_transfer::staging. >>> > >>> > Improves performance when flushed depth texture is frequently used by the >>> > GPU (about 30% for Lightsmark). >>> > >>> > Signed-off-by: Vadim Girlin >>> > --- >>> > >>> > Fixes fbo-clear-formats, fbo-generatemipmap-formats, no regressions on >>> > evergreen >>> >>> Hi, >>> >>> is there any reason this patch hasn't been committed yet? >>> >> >> Hi, >> >> I have some doubts because it was benchmarked by phoronix and there were >> regressions, though I suspect that something is wrong with the results: >> >> http://www.phoronix.com/scan.php?page=article&item=amd_r600g_texdepth&num=4 >> >> I was going to look into it but had no time yet. I'd like to be sure >> that there are no regressions before committing. > > Well, there's nothing wrong with your patch. I wouldn't trust > benchmarks run with the Unity desktop so much. I myself had to switch > from Unity 2D to Xfce just to get consistent results when testing > performance. > > Now that your patch separates flushing for texturing and transfers, I > think we could make it a little bit faster by imlementing an in-place > flush for texturing (that is without having to allocate another > resource). > > Marek In place flush are useful for the case where you know you wont reuse the depth buffer as a depth buffer, or if you know next operation will be a gClear on the depth buffer. What i am worried about is that recompression might not work in place, for it to work you need to have db decompressed into db tiling format and not cb tiling format. Cheers, Jerome ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] Mesa master branch: forced update
On 07/10/2012 12:50 PM, Tom Stellard wrote: > > I just fetched from the master branch of the fdo mesa repo and was > greeted with a "forced update" message, and the gitweb interface shows > several days of history are missing from the master branch. > > olv appears to be the last user to modify the master branch: > > tstellar@annarchy:~$ ls -l /git/mesa/mesa.git/refs/heads/master > -rw-rw-r-- 1 olv mesa 41 Jul 10 11:41 /git/mesa/mesa.git/refs/heads/master > > Anyone know what happened? > > -Tom I'm not sure what happened, but I just (non-force) pushed my copy of master to restore the lost history. Ian, Chad, Matt, Jordan, Anuj, Paul, and I looked at our copies and it appeared that I had the most recent history (I'd pulled literally a few hours ago). The new head is: commit 67a8ee891b2e119d826d8f830c1cbbe64ecb4f82 Author: Marek Olšák Date: Tue Jul 10 18:55:46 2012 +0200 gallium/docs: document interface changes for timestamp query Hopefully this doesn't make things worse... ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [Bug 51956] EnumStrings.LookUpByNumber regression
https://bugs.freedesktop.org/show_bug.cgi?id=51956 --- Comment #1 from Ian Romanick 2012-07-10 20:18:07 PDT --- Changing the name is fine. The test just needs to have the new name replace the old name too. -- Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are the assignee for the bug. ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] Mesa master branch: forced update
On 10-07-12 22:13, Kenneth Graunke wrote: On 07/10/2012 12:50 PM, Tom Stellard wrote: I just fetched from the master branch of the fdo mesa repo and was greeted with a "forced update" message, and the gitweb interface shows several days of history are missing from the master branch. olv appears to be the last user to modify the master branch: tstellar@annarchy:~$ ls -l /git/mesa/mesa.git/refs/heads/master -rw-rw-r-- 1 olv mesa 41 Jul 10 11:41 /git/mesa/mesa.git/refs/heads/master Anyone know what happened? Login on the server, and look at the git logs. The commits are not lost, just not visible. logs are in: /logs or do: cd git reflog -Tom I'm not sure what happened, but I just (non-force) pushed my copy of master to restore the lost history. Ian, Chad, Matt, Jordan, Anuj, Paul, and I looked at our copies and it appeared that I had the most recent history (I'd pulled literally a few hours ago). The new head is: commit 67a8ee891b2e119d826d8f830c1cbbe64ecb4f82 Author: Marek Olšák Date: Tue Jul 10 18:55:46 2012 +0200 gallium/docs: document interface changes for timestamp query Hopefully this doesn't make things worse... ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev -- Ferry Huberts ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] Mesa master branch: forced update
On Tue, Jul 10, 2012 at 4:24 PM, Ferry Huberts wrote: > > On 10-07-12 22:13, Kenneth Graunke wrote: >> >> On 07/10/2012 12:50 PM, Tom Stellard wrote: >>> >>> >>> I just fetched from the master branch of the fdo mesa repo and was >>> greeted with a "forced update" message, and the gitweb interface shows >>> several days of history are missing from the master branch. >>> >>> olv appears to be the last user to modify the master branch: >>> >>> tstellar@annarchy:~$ ls -l /git/mesa/mesa.git/refs/heads/master >>> -rw-rw-r-- 1 olv mesa 41 Jul 10 11:41 >>> /git/mesa/mesa.git/refs/heads/master >>> >>> Anyone know what happened? > > > Login on the server, and look at the git logs. > The commits are not lost, just not visible. > > logs are in: > /logs > > or do: > cd > git reflog I already did that, there are no reflogs in the mesa git repo. The repo is older than the reflog feature. The best we can do is to look at the master ref. It's possible that this was an attack to alter history (sneak in a backdoor, for example, the dri drivers run as root in aiglx in most distros). However, the commit that was pushed matches the older commit (which is why Kenneth was able to pull and fast-forward) and git fsck verifies that the history hasn't been tampered with. That is, it is possible to hand edit a commit object to include changes that wasn't originally there and then just force the SHA1 to match what is was before. git fsck will catch that, but only in a new clone, since when you pull from an existing repo, git won't fetch old objects. More unlikely, history was altered in a way such that code was inserted but the sha1 was preserved (ie sha1 was compromised). I'm on a bad connection right now, but I'll do a fresh clone of the mesa repo and do a git fsck there as well as comparing the contents of a recent commit with what I have locally to see if the contents has been changed while preserving the sha1 validity. Of course, if you were looking to compromise the repo, you wouldn't push back history and attract attention like this, but better safe than sorry. Kristian >>> -Tom >> >> >> I'm not sure what happened, but I just (non-force) pushed my copy of >> master to restore the lost history. Ian, Chad, Matt, Jordan, Anuj, >> Paul, and I looked at our copies and it appeared that I had the most >> recent history (I'd pulled literally a few hours ago). >> >> The new head is: >> >> commit 67a8ee891b2e119d826d8f830c1cbbe64ecb4f82 >> Author: Marek Olšák >> Date: Tue Jul 10 18:55:46 2012 +0200 >> >> gallium/docs: document interface changes for timestamp query >> >> Hopefully this doesn't make things worse... >> ___ >> mesa-dev mailing list >> mesa-dev@lists.freedesktop.org >> http://lists.freedesktop.org/mailman/listinfo/mesa-dev >> > > -- > Ferry Huberts > > > > ___ > mesa-dev mailing list > mesa-dev@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/mesa-dev ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH] r600g: improve flushed depth texture handling v2
On Tue, Jul 10, 2012 at 10:00 PM, Jerome Glisse wrote: > On Tue, Jul 10, 2012 at 2:10 PM, Marek Olšák wrote: >> On Tue, Jul 10, 2012 at 6:40 AM, Vadim Girlin wrote: >>> On Sat, 2012-07-07 at 01:48 +0200, Marek Olšák wrote: On Wed, Jun 27, 2012 at 1:34 AM, Vadim Girlin wrote: > Use r600_resource_texture::flished_depth_texture for GPU access, and > allocate it in the VRAM. For transfers we'll allocate untiled texture in > the > GTT and store it in the r600_transfer::staging. > > Improves performance when flushed depth texture is frequently used by the > GPU (about 30% for Lightsmark). > > Signed-off-by: Vadim Girlin > --- > > Fixes fbo-clear-formats, fbo-generatemipmap-formats, no regressions on > evergreen Hi, is there any reason this patch hasn't been committed yet? >>> >>> Hi, >>> >>> I have some doubts because it was benchmarked by phoronix and there were >>> regressions, though I suspect that something is wrong with the results: >>> >>> http://www.phoronix.com/scan.php?page=article&item=amd_r600g_texdepth&num=4 >>> >>> I was going to look into it but had no time yet. I'd like to be sure >>> that there are no regressions before committing. >> >> Well, there's nothing wrong with your patch. I wouldn't trust >> benchmarks run with the Unity desktop so much. I myself had to switch >> from Unity 2D to Xfce just to get consistent results when testing >> performance. >> >> Now that your patch separates flushing for texturing and transfers, I >> think we could make it a little bit faster by imlementing an in-place >> flush for texturing (that is without having to allocate another >> resource). >> >> Marek > > In place flush are useful for the case where you know you wont reuse > the depth buffer as a depth buffer, or if you know next operation will > be a gClear on the depth buffer. What i am worried about is that > recompression might not work in place, for it to work you need to have > db decompressed into db tiling format and not cb tiling format. The case where the depth is not reused is the most common one. It might even be the only one in practice. Depth textures are most commonly used for shadow mapping, which is the not-reusing case. They can also be used to implement deferred rendering (though that's not very common), which means the same as shadow mapping for us. Actually, no graphics algorithm comes to mind that would do write-texture-write with the same depth buffer. Marek ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [Bug 51956] EnumStrings.LookUpByNumber regression
https://bugs.freedesktop.org/show_bug.cgi?id=51956 Ian Romanick changed: What|Removed |Added Status|NEW |RESOLVED Resolution||FIXED --- Comment #2 from Ian Romanick 2012-07-10 21:48:11 PDT --- commit 4fae5e32d5272986e9e64303eadc974d57e7b3ed Author: Ian Romanick Date: Tue Jul 10 13:37:06 2012 -0700 mesa/test: Update name of GL_TIME_ELAPSED 4952caa caused the _EXT to fall off the name of this enum. This is fine. Update the unit test to expect the new value. Signed-off-by: Ian Romanick Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=51956 -- Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are the assignee for the bug. ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH 2/9] automake: convert libglsl
On Tue, Jul 10, 2012 at 10:08 AM, Jon TURNEY wrote: > v2: Use AM_V_GEN to silence generated code rules. Add BUILT_SOURCES to > CLEANFILES > > Signed-off-by: Jon TURNEY > --- > configure.ac|2 + > src/glsl/.gitignore |1 + > src/glsl/Makefile | 144 > --- > src/glsl/Makefile.am| 112 +++ > src/glsl/glcpp/.gitignore |1 + > src/glsl/glcpp/Makefile.am | 33 > src/glsl/tests/Makefile.am |2 +- > src/mesa/drivers/osmesa/Makefile.am |2 +- > 8 files changed, 151 insertions(+), 146 deletions(-) > delete mode 100644 src/glsl/Makefile > create mode 100644 src/glsl/Makefile.am > create mode 100644 src/glsl/glcpp/Makefile.am > > diff --git a/configure.ac b/configure.ac > index e5ac791..93a8493 100644 > --- a/configure.ac > +++ b/configure.ac > @@ -2170,6 +2170,8 @@ AC_CONFIG_FILES([configs/current > src/egl/wayland/wayland-egl/Makefile > src/egl/wayland/wayland-egl/wayland-egl.pc > src/egl/wayland/wayland-drm/Makefile > + src/glsl/Makefile > + src/glsl/glcpp/Makefile > src/glsl/tests/Makefile > src/glx/Makefile > src/glx/tests/Makefile > diff --git a/src/glsl/.gitignore b/src/glsl/.gitignore > index d26839a..e3531cd 100644 > --- a/src/glsl/.gitignore > +++ b/src/glsl/.gitignore > @@ -6,3 +6,4 @@ glsl_parser.output > builtin_function.cpp > builtin_compiler > glsl_test > +/Makefile > diff --git a/src/glsl/Makefile b/src/glsl/Makefile > deleted file mode 100644 > index 3cf9fc9..000 > --- a/src/glsl/Makefile > +++ /dev/null > @@ -1,144 +0,0 @@ > - > -#src/glsl/pp/Makefile > - > -TOP = ../.. > - > -include $(TOP)/configs/current > - > -LIBNAME = glsl > - > -GLSL_SRCDIR=. > -include Makefile.sources > - > -GLCPP_SOURCES = \ > - $(LIBGLCPP_GENERATED_FILES) \ > - $(LIBGLCPP_FILES) \ > - ralloc.c \ > - glcpp/glcpp.c > - > -C_SOURCES = \ > - $(LIBGLCPP_GENERATED_FILES) \ > - $(LIBGLCPP_FILES) \ > - $(LIBGLSL_FILES) > - > -# common sources for builtin_compiler and libglsl > -CXX_SOURCES = \ > - $(BUILTIN_COMPILER_GENERATED_CXX_FILES) \ > - $(LIBGLSL_CXX_FILES) > - > -LIBS = \ > - $(TOP)/src/glsl/libglsl.a > - > -APPS = glsl_compiler glsl_test glcpp/glcpp > - > -GLSL2_C_SOURCES = \ > - ../mesa/program/hash_table.c \ > - ../mesa/program/symbol_table.c > -GLSL2_CXX_SOURCES = \ > - $(GLSL_COMPILER_CXX_FILES) > - > -GLSL2_OBJECTS = \ > - $(GLSL2_C_SOURCES:.c=.o) \ > - $(GLSL2_CXX_SOURCES:.cpp=.o) > - > -TEST_C_SOURCES = \ > - ../mesa/program/hash_table.c \ > - ../mesa/program/symbol_table.c > - > -TEST_CXX_SOURCES = \ > - standalone_scaffolding.cpp \ > - test.cpp \ > - test_optpass.cpp > - > -TEST_OBJECTS = \ > - $(TEST_C_SOURCES:.c=.o) \ > - $(TEST_CXX_SOURCES:.cpp=.o) > - > -### Basic defines ### > - > -DEFINES += \ > - $(LIBRARY_DEFINES) \ > - $(API_DEFINES) > - > -GLCPP_OBJECTS = \ > - $(GLCPP_SOURCES:.c=.o) \ > - ../mesa/program/hash_table.o > - > -OBJECTS = \ > - $(C_SOURCES:.c=.o) \ > - $(CXX_SOURCES:.cpp=.o) > - > -INCLUDES = \ > - -I. \ > - -I../mesa \ > - -I../mapi \ > - -I../../include \ > - $(LIBRARY_INCLUDES) > - > -ALL_SOURCES = \ > - $(C_SOURCES) \ > - $(CXX_SOURCES) \ > - $(GLSL2_CXX_SOURCES) \ > - $(GLSL2_C_SOURCES) \ > - $(TEST_CXX_SOURCES) \ > - $(TEST_C_SOURCES) > - > -# TARGETS # > - > -default: depend lib$(LIBNAME).a $(APPS) > - > -lib$(LIBNAME).a: $(OBJECTS) builtin_function.o Makefile > $(TOP)/src/glsl/Makefile.template > - $(MKLIB) -cplusplus -o $(LIBNAME) -static $(OBJECTS) > builtin_function.o > - > -depend: $(ALL_SOURCES) Makefile > - rm -f depend > - touch depend > - $(MKDEP) $(MKDEP_OPTIONS) $(INCLUDES) $(ALL_SOURCES) 2> /dev/null > - > -# Remove .o and backup files > -clean: > - rm -f $(GLCPP_OBJECTS) $(GLSL2_OBJECTS) $(TEST_OBJECTS) $(OBJECTS) > lib$(LIBNAME).a depend depend.bak builtin_function.cpp builtin_function.o > builtin_stubs.o builtin_compiler > - -rm -f $(APPS) > - > -# Dummy target > -install: > - @echo -n "" > - > -# RULES # > - > -glsl_compiler: $(GLSL2_OBJECTS) libglsl.a > - $(APP_CXX) $(INCLUDES) $(CFLAGS) $(LDFLAGS) $(GLSL2_OBJECTS) $(LIBS) > -o $@ > - > -glsl_test: $(TEST_OBJECTS) libglsl.a > - $(APP_CXX) $(INCLUDES) $(CFLAGS) $(LDFLAGS) $(TEST_OBJECTS) $(LIBS) > -o $@ > - > -glcpp: glcpp/glcpp > -glcpp/glcpp: $(GLCPP_OBJECTS) > - $(APP_CC) $(INCLUDES) $(CFLAGS) $(LDFLAGS) $(GLCPP_OBJECTS) -o $@ > - > -.cpp.o: > - $(CXX) -c $(INCLUDES) $(CXXFLAGS) $(DEFINES) $< -o $@ > - > -.c.o: > - $(CC) -c $(INCLUDES) $(CFLAGS) $(DEFINES) $< -o $@ > - >
Re: [Mesa-dev] [PATCH 3/9] Rename sparc/clip.S -> sparc/sparc_clip.S
On Tue, Jul 10, 2012 at 10:08 AM, Jon TURNEY wrote: > Automake can't handle having both clip.S and clip.c, even though they have > different paths > > "src/mesa/Makefile.am: object `clip.lo' created by `$(SRCDIR)/sparc/clip.S' > and `$(SRCDIR)/main/clip.c'" > > Signed-off-by: Jon TURNEY > --- Yeah, ran into this too at some point. Reviewed-by: Matt Turner ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] Recent master build failure (upgrade gl2ext.h / glDrawBuffersNV fallout)
Mesa master stopped building successfully on my tinderbox on 7/6. The relevant changes from the previous successful build were 95ce454..c445b0f, and I bisected it to the upgrade of gl2ext.h to version 18099 (see below). ../../../src/mapi/glapi/glapitemp.h:4635:1: error: no previous prototype for 'glDrawBuffersNV' [-Werror=missing-prototypes] The commit mentions that this new failure is expected, so it would be nice if the fix for the original issue could now be included as well. --Jeremy 3ed8d4285360047d78cfd3218ac4f8f325ae5d6e is the first bad commit commit 3ed8d4285360047d78cfd3218ac4f8f325ae5d6e Author: Kristian Høgsberg Date: Tue Jul 3 20:47:04 2012 -0400 GLES2: upgrade gl2ext.h to version 18099 Redo this commit, and remove the inclusion of gl2ext.h from src/mapi/glapi/glapi_priv.h. The include was added in 8f3be339850ead96f9c6200db4e0db1f74e39d13 to fix a missing prototype for glDrawBuffersNV and others, but it's not possible to include both glext.h and gl2ext.h from the same file. I don't see the missing prototype here (with or without shared glapi) so I'm just removing the offending #include. Also, since we're redoing this, update to the most recent gl2ext.2. Signed-off-by: Kristian Høgsberg :04 04 3076bd4623a3727eb3690e39801c459ea2c08706 7754f21c4cee87eb86ccb5ac8b9cbf810d0835ff M include :04 04 0e602fff716a390251eacb932e7f406eb399e2e0 ed4ecdd9127f4541ed0f148f3b97e6eefb6572df M src ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH] r600g: improve flushed depth texture handling v2
On Tue, 2012-07-10 at 20:10 +0200, Marek Olšák wrote: > On Tue, Jul 10, 2012 at 6:40 AM, Vadim Girlin wrote: > > On Sat, 2012-07-07 at 01:48 +0200, Marek Olšák wrote: > >> On Wed, Jun 27, 2012 at 1:34 AM, Vadim Girlin > >> wrote: > >> > Use r600_resource_texture::flished_depth_texture for GPU access, and > >> > allocate it in the VRAM. For transfers we'll allocate untiled texture in > >> > the > >> > GTT and store it in the r600_transfer::staging. > >> > > >> > Improves performance when flushed depth texture is frequently used by the > >> > GPU (about 30% for Lightsmark). > >> > > >> > Signed-off-by: Vadim Girlin > >> > --- > >> > > >> > Fixes fbo-clear-formats, fbo-generatemipmap-formats, no regressions on > >> > evergreen > >> > >> Hi, > >> > >> is there any reason this patch hasn't been committed yet? > >> > > > > Hi, > > > > I have some doubts because it was benchmarked by phoronix and there were > > regressions, though I suspect that something is wrong with the results: > > > > http://www.phoronix.com/scan.php?page=article&item=amd_r600g_texdepth&num=4 > > > > I was going to look into it but had no time yet. I'd like to be sure > > that there are no regressions before committing. > > Well, there's nothing wrong with your patch. I wouldn't trust > benchmarks run with the Unity desktop so much. I myself had to switch > from Unity 2D to Xfce just to get consistent results when testing > performance. > OK, I've pushed it. Thanks for reviewing. Vadim > Now that your patch separates flushing for texturing and transfers, I > think we could make it a little bit faster by imlementing an in-place > flush for texturing (that is without having to allocate another > resource). > > Marek ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] Mesa master branch: forced update
On Tue, Jul 10, 2012 at 4:54 PM, Kristian Høgsberg wrote: > On Tue, Jul 10, 2012 at 4:24 PM, Ferry Huberts wrote: >> >> On 10-07-12 22:13, Kenneth Graunke wrote: >>> >>> On 07/10/2012 12:50 PM, Tom Stellard wrote: I just fetched from the master branch of the fdo mesa repo and was greeted with a "forced update" message, and the gitweb interface shows several days of history are missing from the master branch. olv appears to be the last user to modify the master branch: tstellar@annarchy:~$ ls -l /git/mesa/mesa.git/refs/heads/master -rw-rw-r-- 1 olv mesa 41 Jul 10 11:41 /git/mesa/mesa.git/refs/heads/master Anyone know what happened? >> >> >> Login on the server, and look at the git logs. >> The commits are not lost, just not visible. >> >> logs are in: >> /logs >> >> or do: >> cd >> git reflog > > I already did that, there are no reflogs in the mesa git repo. The > repo is older than the reflog feature. The best we can do is to look > at the master ref. > > It's possible that this was an attack to alter history (sneak in a > backdoor, for example, the dri drivers run as root in aiglx in most > distros). However, the commit that was pushed matches the older > commit (which is why Kenneth was able to pull and fast-forward) and > git fsck verifies that the history hasn't been tampered with. That > is, it is possible to hand edit a commit object to include changes > that wasn't originally there and then just force the SHA1 to match > what is was before. git fsck will catch that, but only in a new > clone, since when you pull from an existing repo, git won't fetch old > objects. More unlikely, history was altered in a way such that code > was inserted but the sha1 was preserved (ie sha1 was compromised). > I'm on a bad connection right now, but I'll do a fresh clone of the > mesa repo and do a git fsck there as well as comparing the contents of > a recent commit with what I have locally to see if the contents has > been changed while preserving the sha1 validity. And the results are in: freshly cloned mesa repo goes through git fsck without problems and just to be completely paranoid I checked against compromised sha1sums (that is, attack by inserting code without affecting the sha1sums) by comparing the output of git archive of 40742fa6864000d431b81c3769a3136b7ff4a0d1 in both my previous checkout and the fresh clone and they match. So while it's suspicious that Chia-I hasn't been active for a long time and the suddenly pushes a forced update of the repo, I don't think anything was compromised or any history lost. The freedesktop.org account has been disabled until we hear back from Chia-I. Kristian ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [PATCH 0/3 v2] intel: Enable GL_OES_compressed_ETC1_RGB8_texture
This series lives on my etc1-v9 branch. This series enables GL_OES_compressed_ETC1_RGB8_texture for all Intel hardware by simply decoding the ETC1 data into RGBX data at time of glTexImage. No current Intel hardware supports ETC1. v2: Don't break gallium. Chad Versace (3): gallium/util, mesa: Refactor etc1 unpack function mesa: Add function for decoding ETC1 textures intel: Enable GL_OES_compressed_ETC1_RGB8_texture src/gallium/auxiliary/util/u_format_etc.c| 25 + src/mesa/drivers/dri/i965/brw_wm_surface_state.c | 5 ++ src/mesa/drivers/dri/intel/intel_extensions.c| 1 + src/mesa/drivers/dri/intel/intel_mipmap_tree.c | 70 +++- src/mesa/drivers/dri/intel/intel_mipmap_tree.h | 15 + src/mesa/main/texcompress_etc.c | 32 +++ src/mesa/main/texcompress_etc.h | 8 +++ src/mesa/main/texcompress_etc_tmp.h | 34 8 files changed, 165 insertions(+), 25 deletions(-) -- 1.7.11.1 ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [PATCH 1/3] gallium/util, mesa: Refactor etc1 unpack function
Move the body of util_etc1_rgb8_unpack_rgba_unorm8 into a new function
that can be shared between gallium and dri drivers,
texcompress_etc_tmp.h:etc1_unpack_rgba.
CC: Chia-I
CC: Kenneth Graunke
Signed-off-by: Chad Versace
---
src/gallium/auxiliary/util/u_format_etc.c | 25 +--
src/mesa/main/texcompress_etc_tmp.h | 34 +++
2 files changed, 35 insertions(+), 24 deletions(-)
diff --git a/src/gallium/auxiliary/util/u_format_etc.c
b/src/gallium/auxiliary/util/u_format_etc.c
index 7500e1e..f909b16 100644
--- a/src/gallium/auxiliary/util/u_format_etc.c
+++ b/src/gallium/auxiliary/util/u_format_etc.c
@@ -13,30 +13,7 @@
void
util_format_etc1_rgb8_unpack_rgba_8unorm(uint8_t *dst_row, unsigned
dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width,
unsigned height)
{
- const unsigned bw = 4, bh = 4, bs = 8, comps = 4;
- struct etc1_block block;
- unsigned x, y, i, j;
-
- for (y = 0; y < height; y += bh) {
- const uint8_t *src = src_row;
-
- for (x = 0; x < width; x+= bw) {
- etc1_parse_block(&block, src);
-
- for (j = 0; j < bh; j++) {
-uint8_t *dst = dst_row + (y + j) * dst_stride + x * comps;
-for (i = 0; i < bw; i++) {
- etc1_fetch_texel(&block, i, j, dst);
- dst[3] = 255;
- dst += comps;
-}
- }
-
- src += bs;
- }
-
- src_row += src_stride;
- }
+ etc1_unpack_rgba(dst_row, dst_stride, src_row, src_stride, width,
height);
}
void
diff --git a/src/mesa/main/texcompress_etc_tmp.h
b/src/mesa/main/texcompress_etc_tmp.h
index 5c8c6de..8bbb2cd 100644
--- a/src/mesa/main/texcompress_etc_tmp.h
+++ b/src/mesa/main/texcompress_etc_tmp.h
@@ -134,3 +134,37 @@ TAG(etc1_fetch_texel)(const struct TAG(etc1_block) *block,
dst[1] = TAG(etc1_clamp)(base_color[1], modifier);
dst[2] = TAG(etc1_clamp)(base_color[2], modifier);
}
+
+static void
+etc1_unpack_rgba(uint8_t *dst_row,
+ unsigned dst_stride,
+ const uint8_t *src_row,
+ unsigned src_stride,
+ unsigned width,
+ unsigned height)
+{
+ const unsigned bw = 4, bh = 4, bs = 8, comps = 4;
+ struct etc1_block block;
+ unsigned x, y, i, j;
+
+ for (y = 0; y < height; y += bh) {
+ const uint8_t *src = src_row;
+
+ for (x = 0; x < width; x+= bw) {
+ etc1_parse_block(&block, src);
+
+ for (j = 0; j < bh; j++) {
+uint8_t *dst = dst_row + (y + j) * dst_stride + x * comps;
+for (i = 0; i < bw; i++) {
+ etc1_fetch_texel(&block, i, j, dst);
+ dst[3] = 255;
+ dst += comps;
+}
+ }
+
+ src += bs;
+ }
+
+ src_row += src_stride;
+ }
+}
--
1.7.11.1
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [PATCH 2/3] mesa: Add function for decoding ETC1 textures
Add function _mesa_etc1_unpack_rgba. It is intended to be used by
glCompressedTexSubImage2D to decode ETC1 textures into RGBA.
CC: Chia-I
CC: Kenneth Graunke
Signed-off-by: Chad Versace
---
src/mesa/main/texcompress_etc.c | 32
src/mesa/main/texcompress_etc.h | 8
2 files changed, 40 insertions(+)
diff --git a/src/mesa/main/texcompress_etc.c b/src/mesa/main/texcompress_etc.c
index 5b331a9..c645f52 100644
--- a/src/mesa/main/texcompress_etc.c
+++ b/src/mesa/main/texcompress_etc.c
@@ -69,3 +69,35 @@ _mesa_fetch_texel_2d_f_etc1_rgb8(const struct
swrast_texture_image *texImage,
texel[BCOMP] = UBYTE_TO_FLOAT(dst[2]);
texel[ACOMP] = 1.0f;
}
+
+/**
+ * Decode texture data in format `MESA_FORMAT_ETC1_RGB8` to
+ * `MESA_FORMAT_ABGR`.
+ *
+ * The size of the source data must be a multiple of the ETC1 block size,
+ * which is 8, even if the texture image's dimensions are not aligned to 4.
+ * From the GL_OES_compressed_ETC1_RGB8_texture spec:
+ * The texture is described as a number of 4x4 pixel blocks. If the
+ * texture (or a particular mip-level) is smaller than 4 pixels in
+ * any dimension (such as a 2x2 or a 8x1 texture), the texture is
+ * found in the upper left part of the block(s), and the rest of the
+ * pixels are not used. For instance, a texture of size 4x2 will be
+ * placed in the upper half of a 4x4 block, and the lower half of the
+ * pixels in the block will not be accessed.
+ *
+ * \param src_width in pixels
+ * \param src_height in pixels
+ * \param dst_stride in bytes
+ */
+void
+_mesa_etc1_unpack_rgba(uint8_t *dst_row,
+ unsigned dst_stride,
+ const uint8_t *src_row,
+ unsigned src_stride,
+ unsigned src_width,
+ unsigned src_height)
+{
+ etc1_unpack_rgba(dst_row, dst_stride,
+src_row, src_stride,
+src_width, src_height);
+}
diff --git a/src/mesa/main/texcompress_etc.h b/src/mesa/main/texcompress_etc.h
index 8e8427f..bfba4ff 100644
--- a/src/mesa/main/texcompress_etc.h
+++ b/src/mesa/main/texcompress_etc.h
@@ -37,4 +37,12 @@ void
_mesa_fetch_texel_2d_f_etc1_rgb8(const struct swrast_texture_image *texImage,
GLint i, GLint j, GLint k, GLfloat *texel);
+void
+_mesa_etc1_unpack_rgba(uint8_t *dst_row,
+ unsigned dst_stride,
+ const uint8_t *src_row,
+ unsigned src_stride,
+ unsigned src_width,
+ unsigned src_height);
+
#endif
--
1.7.11.1
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [PATCH 3/3] intel: Enable GL_OES_compressed_ETC1_RGB8_texture
Enable it for all hardware.
No current hardware supports ETC1, so this patch implements it by
translating the ETC1 data to RGBX data during the call to
glCompressedTexImage2D(). For details, see the doxygen for
intel_mipmap_tree::wraps_etc1.
Passes the Piglit test spec/OES_compressed_ETC1_RGB8_texture/miptree (test
is pending review) and the ETC1 test in the GLES2 conformance suite.
Reviewed-by: Kenneth Graunke
Signed-off-by: Chad Versace
---
src/mesa/drivers/dri/i965/brw_wm_surface_state.c | 5 ++
src/mesa/drivers/dri/intel/intel_extensions.c| 1 +
src/mesa/drivers/dri/intel/intel_mipmap_tree.c | 70 +++-
src/mesa/drivers/dri/intel/intel_mipmap_tree.h | 15 +
4 files changed, 90 insertions(+), 1 deletion(-)
diff --git a/src/mesa/drivers/dri/i965/brw_wm_surface_state.c
b/src/mesa/drivers/dri/i965/brw_wm_surface_state.c
index 26e65af..794aadb 100644
--- a/src/mesa/drivers/dri/i965/brw_wm_surface_state.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_surface_state.c
@@ -551,6 +551,11 @@ brw_init_surface_formats(struct brw_context *brw)
ctx->TextureFormatSupported[MESA_FORMAT_Z32_FLOAT] = true;
ctx->TextureFormatSupported[MESA_FORMAT_Z32_FLOAT_X24S8] = true;
ctx->TextureFormatSupported[MESA_FORMAT_Z16] = true;
+
+ /* On hardware that lacks support for ETC1, we map ETC1 to RGBX
+* during glCompressedTexImage2D(). See intel_mipmap_tree::wraps_etc1.
+*/
+ ctx->TextureFormatSupported[MESA_FORMAT_ETC1_RGB8] = true;
}
bool
diff --git a/src/mesa/drivers/dri/intel/intel_extensions.c
b/src/mesa/drivers/dri/intel/intel_extensions.c
index 953b983..844125d 100755
--- a/src/mesa/drivers/dri/intel/intel_extensions.c
+++ b/src/mesa/drivers/dri/intel/intel_extensions.c
@@ -91,6 +91,7 @@ intelInitExtensions(struct gl_context *ctx)
ctx->Extensions.OES_EGL_image = true;
#endif
ctx->Extensions.OES_draw_texture = true;
+ ctx->Extensions.OES_compressed_ETC1_RGB8_texture = true;
if (intel->gen >= 6)
ctx->Const.GLSLVersion = 130;
diff --git a/src/mesa/drivers/dri/intel/intel_mipmap_tree.c
b/src/mesa/drivers/dri/intel/intel_mipmap_tree.c
index 7018c37..74c9924 100644
--- a/src/mesa/drivers/dri/intel/intel_mipmap_tree.c
+++ b/src/mesa/drivers/dri/intel/intel_mipmap_tree.c
@@ -38,6 +38,7 @@
#include "main/enums.h"
#include "main/formats.h"
#include "main/image.h"
+#include "main/texcompress_etc.h"
#include "main/teximage.h"
#define FILE_DEBUG_FLAG DEBUG_MIPTREE
@@ -180,7 +181,15 @@ intel_miptree_create(struct intel_context *intel,
{
struct intel_mipmap_tree *mt;
uint32_t tiling = I915_TILING_NONE;
- GLenum base_format = _mesa_get_format_base_format(format);
+ GLenum base_format;
+ bool wraps_etc1 = false;
+
+ if (format == MESA_FORMAT_ETC1_RGB8) {
+ format = MESA_FORMAT_RGBX_REV;
+ wraps_etc1 = true;
+ }
+
+ base_format = _mesa_get_format_base_format(format);
if (intel->use_texture_tiling && !_mesa_is_format_compressed(format)) {
if (intel->gen >= 4 &&
@@ -227,6 +236,7 @@ intel_miptree_create(struct intel_context *intel,
return NULL;
}
+ mt->wraps_etc1 = wraps_etc1;
mt->region = intel_region_alloc(intel->intelScreen,
tiling,
mt->cpp,
@@ -978,6 +988,60 @@ intel_miptree_unmap_s8(struct intel_context *intel,
free(map->buffer);
}
+static void
+intel_miptree_map_etc1(struct intel_context *intel,
+ struct intel_mipmap_tree *mt,
+ struct intel_miptree_map *map,
+ unsigned int level,
+ unsigned int slice)
+{
+ /* For justification of these invariants,
+* see intel_mipmap_tree:wraps_etc1.
+*/
+ assert(mt->wraps_etc1);
+ assert(mt->format == MESA_FORMAT_RGBX_REV);
+
+ /* From the GL_OES_compressed_ETC1_RGB8_texture spec:
+* INVALID_OPERATION is generated by CompressedTexSubImage2D,
+* TexSubImage2D, or CopyTexSubImage2D if the texture image
+* bound to has internal format ETC1_RGB8_OES.
+*
+* This implies that intel_miptree_map_etc1() can only be called from
+* glCompressedTexImage2D, and hence the assertions below hold.
+*/
+ assert(map->mode & GL_MAP_WRITE_BIT);
+ assert(map->mode & GL_MAP_INVALIDATE_RANGE_BIT);
+ assert(map->x == 0);
+ assert(map->y == 0);
+
+ /* Each ETC1 block contains 4x4 pixels in 8 bytes. */
+ map->stride = 2 * map->w;
+ map->buffer = map->ptr = malloc(map->stride * map->h);
+}
+
+static void
+intel_miptree_unmap_etc1(struct intel_context *intel,
+ struct intel_mipmap_tree *mt,
+ struct intel_miptree_map *map,
+ unsigned int level,
+ unsigned int slice)
+{
+ uint32_t image_x;
+ uint32_t image_y;
+ intel_miptree_get_image_offset(mt, level, 0, slice, &image_x, &image_y);
+
+ uint8_t *xbgr = intel_region_map(intel, mt->region, map->
Re: [Mesa-dev] [PATCH 4/5] i965/fs.h: Refactor tests for instructions modifying a register.
On 07/09/2012 03:40 PM, Eric Anholt wrote:
> ---
> src/mesa/drivers/dri/i965/brw_fs.cpp | 25
> ++--
> src/mesa/drivers/dri/i965/brw_fs.h |1 +
> .../drivers/dri/i965/brw_fs_copy_propagation.cpp | 16 ++---
> src/mesa/drivers/dri/i965/brw_fs_cse.cpp |8 +--
> 4 files changed, 16 insertions(+), 34 deletions(-)
>
> diff --git a/src/mesa/drivers/dri/i965/brw_fs.cpp
> b/src/mesa/drivers/dri/i965/brw_fs.cpp
> index 374a059..354134c 100644
> --- a/src/mesa/drivers/dri/i965/brw_fs.cpp
> +++ b/src/mesa/drivers/dri/i965/brw_fs.cpp
> @@ -168,6 +168,15 @@ fs_inst::regs_written()
> }
>
> bool
> +fs_inst::overwrites_reg(const fs_reg ®)
> +{
> + return (reg.file == dst.file &&
> + reg.reg == dst.reg &&
> + reg.reg_offset >= dst.reg_offset &&
> + reg.reg_offset < dst.reg_offset + regs_written());
> +}
> +
> +bool
> fs_inst::is_tex()
> {
> return (opcode == SHADER_OPCODE_TEX ||
> @@ -1400,9 +1409,7 @@ fs_visitor::propagate_constants()
>}
>
>if (scan_inst->dst.file == GRF &&
> - scan_inst->dst.reg == inst->dst.reg &&
> - (scan_inst->dst.reg_offset == inst->dst.reg_offset ||
> - scan_inst->regs_written() > 1)) {
> + scan_inst->overwrites_reg(inst->dst)) {
Technically this is a change in behavior: instead of breaking whenever
regs_written() > 1, we actually check if it's in the range.
That's a good change to make, but it might be nice to note it in the
commit message.
Reviewed-by: Kenneth Graunke
> break;
>}
>}
> @@ -1602,16 +1609,8 @@ fs_visitor::register_coalesce()
> !scan_inst->is_tail_sentinel();
> scan_inst = (fs_inst *)scan_inst->next) {
>if (scan_inst->dst.file == GRF) {
> - if (scan_inst->dst.reg == inst->dst.reg &&
> - (scan_inst->dst.reg_offset == inst->dst.reg_offset ||
> - scan_inst->regs_written() > 1)) {
> -interfered = true;
> -break;
> - }
> - if (inst->src[0].file == GRF &&
> - scan_inst->dst.reg == inst->src[0].reg &&
> - (scan_inst->dst.reg_offset == inst->src[0].reg_offset ||
> - scan_inst->regs_written() > 1)) {
> + if (scan_inst->overwrites_reg(inst->dst) ||
> +scan_inst->overwrites_reg(inst->src[0])) {
> interfered = true;
> break;
> }
> diff --git a/src/mesa/drivers/dri/i965/brw_fs.h
> b/src/mesa/drivers/dri/i965/brw_fs.h
> index deb58d8..6e3c46a 100644
> --- a/src/mesa/drivers/dri/i965/brw_fs.h
> +++ b/src/mesa/drivers/dri/i965/brw_fs.h
> @@ -150,6 +150,7 @@ public:
>
> bool equals(fs_inst *inst);
> int regs_written();
> + bool overwrites_reg(const fs_reg ®);
> bool is_tex();
> bool is_math();
>
> diff --git a/src/mesa/drivers/dri/i965/brw_fs_copy_propagation.cpp
> b/src/mesa/drivers/dri/i965/brw_fs_copy_propagation.cpp
> index 106964d..1870f43 100644
> --- a/src/mesa/drivers/dri/i965/brw_fs_copy_propagation.cpp
> +++ b/src/mesa/drivers/dri/i965/brw_fs_copy_propagation.cpp
> @@ -98,23 +98,11 @@ fs_visitor::opt_copy_propagate_local(void *mem_ctx,
>
>/* kill the destination from the ACP */
>if (inst->dst.file == GRF) {
> - int start_offset = inst->dst.reg_offset;
> - int end_offset = start_offset + inst->regs_written();
> -
>foreach_list_safe(entry_node, acp) {
> acp_entry *entry = (acp_entry *)entry_node;
>
> - if (entry->dst.file == GRF &&
> - entry->dst.reg == inst->dst.reg &&
> - entry->dst.reg_offset >= start_offset &&
> - entry->dst.reg_offset < end_offset) {
> -entry->remove();
> -continue;
> - }
> - if (entry->src.file == GRF &&
> - entry->src.reg == inst->dst.reg &&
> - entry->src.reg_offset >= start_offset &&
> - entry->src.reg_offset < end_offset) {
> + if (inst->overwrites_reg(entry->dst) ||
> +inst->overwrites_reg(entry->src)) {
> entry->remove();
> }
>}
> diff --git a/src/mesa/drivers/dri/i965/brw_fs_cse.cpp
> b/src/mesa/drivers/dri/i965/brw_fs_cse.cpp
> index fd28e14..7bf6698 100644
> --- a/src/mesa/drivers/dri/i965/brw_fs_cse.cpp
> +++ b/src/mesa/drivers/dri/i965/brw_fs_cse.cpp
> @@ -150,17 +150,11 @@ fs_visitor::opt_cse_local(fs_bblock *block, exec_list
> *aeb)
>}
>
>/* Kill all AEB entries that use the destination. */
> - int start_offset = inst->dst.reg_offset;
> - int end_offset = start_offset + inst->regs_written();
> -
>foreach_list_safe(entry_node, aeb) {
>aeb_entry *entry = (aeb_entry *)entry_node;
>
>for (int i = 0; i < 3; i++) {
> - if (entry->generator->src[i].file == inst->dst.file &&
> - entry->generator->src[i].reg == inst->dst.reg &&
> -
Re: [Mesa-dev] [PATCH 1/5] i965/fs: Move a block out of a loop in live variables setup.
On 07/09/2012 03:40 PM, Eric Anholt wrote:
> This was accidentally copy-and-pasted inside.
> ---
> src/mesa/drivers/dri/i965/brw_fs_live_variables.cpp |9 +
> 1 file changed, 5 insertions(+), 4 deletions(-)
>
> diff --git a/src/mesa/drivers/dri/i965/brw_fs_live_variables.cpp
> b/src/mesa/drivers/dri/i965/brw_fs_live_variables.cpp
> index 46408da..40e7ae5 100644
> --- a/src/mesa/drivers/dri/i965/brw_fs_live_variables.cpp
> +++ b/src/mesa/drivers/dri/i965/brw_fs_live_variables.cpp
> @@ -192,11 +192,12 @@ fs_visitor::calculate_live_intervals()
>
> use[reg] = ip;
>}
> - if (inst->dst.file == GRF) {
> - int reg = inst->dst.reg;
> + }
>
> - def[reg] = MIN2(def[reg], ip);
> - }
> + if (inst->dst.file == GRF) {
> + int reg = inst->dst.reg;
> +
> + def[reg] = MIN2(def[reg], ip);
>}
>
>ip++;
Patches 1-4 are:
Reviewed-by: Kenneth Graunke
(though I had comments on 3 and 4)
I'm still trying to work through patch 5. It looks like some parts
aren't adding up, but I'm not sure whether that's a mistake in your code
or in my ability to read the patch.
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH] r600g: improve flushed depth texture handling v2
On Tue, Jul 10, 2012 at 5:16 PM, Marek Olšák wrote: > On Tue, Jul 10, 2012 at 10:00 PM, Jerome Glisse wrote: >> On Tue, Jul 10, 2012 at 2:10 PM, Marek Olšák wrote: >>> On Tue, Jul 10, 2012 at 6:40 AM, Vadim Girlin wrote: On Sat, 2012-07-07 at 01:48 +0200, Marek Olšák wrote: > On Wed, Jun 27, 2012 at 1:34 AM, Vadim Girlin > wrote: > > Use r600_resource_texture::flished_depth_texture for GPU access, and > > allocate it in the VRAM. For transfers we'll allocate untiled texture > > in the > > GTT and store it in the r600_transfer::staging. > > > > Improves performance when flushed depth texture is frequently used by > > the > > GPU (about 30% for Lightsmark). > > > > Signed-off-by: Vadim Girlin > > --- > > > > Fixes fbo-clear-formats, fbo-generatemipmap-formats, no regressions on > > evergreen > > Hi, > > is there any reason this patch hasn't been committed yet? > Hi, I have some doubts because it was benchmarked by phoronix and there were regressions, though I suspect that something is wrong with the results: http://www.phoronix.com/scan.php?page=article&item=amd_r600g_texdepth&num=4 I was going to look into it but had no time yet. I'd like to be sure that there are no regressions before committing. >>> >>> Well, there's nothing wrong with your patch. I wouldn't trust >>> benchmarks run with the Unity desktop so much. I myself had to switch >>> from Unity 2D to Xfce just to get consistent results when testing >>> performance. >>> >>> Now that your patch separates flushing for texturing and transfers, I >>> think we could make it a little bit faster by imlementing an in-place >>> flush for texturing (that is without having to allocate another >>> resource). >>> >>> Marek >> >> In place flush are useful for the case where you know you wont reuse >> the depth buffer as a depth buffer, or if you know next operation will >> be a gClear on the depth buffer. What i am worried about is that >> recompression might not work in place, for it to work you need to have >> db decompressed into db tiling format and not cb tiling format. > > The case where the depth is not reused is the most common one. It > might even be the only one in practice. Depth textures are most > commonly used for shadow mapping, which is the not-reusing case. They > can also be used to implement deferred rendering (though that's not > very common), which means the same as shadow mapping for us. Actually, > no graphics algorithm comes to mind that would do write-texture-write > with the same depth buffer. > > Marek I am not saying it's not the most common one, i am saying that recompressing might be more complex (recompress to different buffer then copy back to original buffer, or copy buffer and uncompress from copy). Cheers, Jerome ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] Recent master build failure (upgrade gl2ext.h / glDrawBuffersNV fallout)
On 07/10/2012 02:39 PM, Jeremy Huddleston Sequoia wrote: Mesa master stopped building successfully on my tinderbox on 7/6. The relevant changes from the previous successful build were 95ce454..c445b0f, and I bisected it to the upgrade of gl2ext.h to version 18099 (see below). ../../../src/mapi/glapi/glapitemp.h:4635:1: error: no previous prototype for 'glDrawBuffersNV' [-Werror=missing-prototypes] The commit mentions that this new failure is expected, so it would be nice if the fix for the original issue could now be included as well. This commit fixed another, even more catastrophic, problem. The fix is to get a new gl2ext.h, when it's available, that fixes the original problem. In the mean time, I think tinderbox will remain broken. :( See http://lists.freedesktop.org/archives/mesa-dev/2012-July/023670.html --Jeremy 3ed8d4285360047d78cfd3218ac4f8f325ae5d6e is the first bad commit commit 3ed8d4285360047d78cfd3218ac4f8f325ae5d6e Author: Kristian Høgsberg Date: Tue Jul 3 20:47:04 2012 -0400 GLES2: upgrade gl2ext.h to version 18099 Redo this commit, and remove the inclusion of gl2ext.h from src/mapi/glapi/glapi_priv.h. The include was added in 8f3be339850ead96f9c6200db4e0db1f74e39d13 to fix a missing prototype for glDrawBuffersNV and others, but it's not possible to include both glext.h and gl2ext.h from the same file. I don't see the missing prototype here (with or without shared glapi) so I'm just removing the offending #include. Also, since we're redoing this, update to the most recent gl2ext.2. Signed-off-by: Kristian Høgsberg :04 04 3076bd4623a3727eb3690e39801c459ea2c08706 7754f21c4cee87eb86ccb5ac8b9cbf810d0835ff M include :04 04 0e602fff716a390251eacb932e7f406eb399e2e0 ed4ecdd9127f4541ed0f148f3b97e6eefb6572df M src ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [Bug 51972] New: Compilation error on x86-64 with --enable-32-bit option
https://bugs.freedesktop.org/show_bug.cgi?id=51972 Bug #: 51972 Summary: Compilation error on x86-64 with --enable-32-bit option Classification: Unclassified Product: Mesa Version: git Platform: x86-64 (AMD64) OS/Version: Linux (All) Status: NEW Severity: normal Priority: medium Component: Mesa core AssignedTo: mesa-dev@lists.freedesktop.org ReportedBy: virgile.bes...@free.fr I'm on Debian Sid and I can't manage to compile mesa with the --enable-32-bit option whereas it works without this option. I typed this command in my shell : ./autogen.sh --with-gallium-drivers=r600 --with-dri-drivers= --enable-32-bit and the output was : [...] checking expat.h usability... no checking expat.h presence... yes configure: WARNING: expat.h: present but cannot be compiled configure: WARNING: expat.h: check for missing prerequisite headers? configure: WARNING: expat.h: see the Autoconf documentation configure: WARNING: expat.h: section "Present But Cannot Be Compiled" configure: WARNING: expat.h: proceeding with the compiler's result configure: WARNING: ## -- ## configure: WARNING: ## Report this to https://bugs.freedesktop.org/enter_bug.cgi?product=Mesa ## configure: WARNING: ## -- ## checking for expat.h... no configure: error: Expat required for DRI. I don't know if it's a mesa bug or an Autoconf bug... Thanks V.B. -- Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are the assignee for the bug. ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [PATCH] egl_dri2: NULL check for EGLNativeWindowType
Some application calls eglCreateWindowSurface with
EGLNativeWindowType parameter having zero value. It causes SEGV
and disturbs error handling like EGL_NO_SURFACE.
Signed-off-by: Elvis Lee
---
src/egl/drivers/dri2/platform_drm.c |2 ++
1 file changed, 2 insertions(+)
diff --git a/src/egl/drivers/dri2/platform_drm.c
b/src/egl/drivers/dri2/platform_drm.c
index 54067ff..9aafb52 100644
--- a/src/egl/drivers/dri2/platform_drm.c
+++ b/src/egl/drivers/dri2/platform_drm.c
@@ -108,6 +108,8 @@ dri2_create_surface(_EGLDriver *drv, _EGLDisplay *disp,
EGLint type,
switch (type) {
case EGL_WINDOW_BIT:
+ if (!window)
+ return NULL;
surf = gbm_dri_surface((struct gbm_surface *) window);
dri2_surf->gbm_surf = surf;
dri2_surf->base.Width = surf->base.width;
--
1.7.9.5
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] Recent master build failure (upgrade gl2ext.h / glDrawBuffersNV fallout)
Hi Jeremy, I wasn't able to reproduce this problem when I committed the patch to update gl2ext.h, but I now figured out what triggers it. It only happens when the asm stubs are disabled (either pass --disable-asm to configure or buid on an arch where we fallback to c stubs) and when shared glapi is disabled. Anyway, having reproduced it, I see that it's just glDrawBuffersNV that's the problem. I assumed that this was just the first missing prototype and you didn't list the rest. Digging further, it turns out that we don't generate entry points for GLES extension functions in general, which is why only glDrawBuffersNV is a problem. The GL_NV_draw_buffers extension defines the entry point is in ARB_draw_buffers.xml which holds GL extensions too. All other GLES extensions are listed in es_EXT.xml, which isn't used to generate the public entry points. So it looks like the real fix is the attached patch - move the GL_NV_draw_buffer definition to es_EXT.xml. It does fix the error here and doesn't affect the other configuration combinations. Kristian On Tue, Jul 10, 2012 at 9:00 PM, Ian Romanick wrote: > On 07/10/2012 02:39 PM, Jeremy Huddleston Sequoia wrote: >> >> Mesa master stopped building successfully on my tinderbox on 7/6. The >> relevant changes from the previous successful build were 95ce454..c445b0f, >> and I bisected it to the upgrade of gl2ext.h to version 18099 (see below). >> >> ../../../src/mapi/glapi/glapitemp.h:4635:1: error: no previous prototype >> for 'glDrawBuffersNV' [-Werror=missing-prototypes] >> >> The commit mentions that this new failure is expected, so it would be nice >> if the fix for the original issue could now be included as well. > > > This commit fixed another, even more catastrophic, problem. The fix is to > get a new gl2ext.h, when it's available, that fixes the original problem. > In the mean time, I think tinderbox will remain broken. :( > > See http://lists.freedesktop.org/archives/mesa-dev/2012-July/023670.html > >> --Jeremy >> >> 3ed8d4285360047d78cfd3218ac4f8f325ae5d6e is the first bad commit >> commit 3ed8d4285360047d78cfd3218ac4f8f325ae5d6e >> Author: Kristian Høgsberg >> Date: Tue Jul 3 20:47:04 2012 -0400 >> >> GLES2: upgrade gl2ext.h to version 18099 >> >> Redo this commit, and remove the inclusion of gl2ext.h >> from src/mapi/glapi/glapi_priv.h. The include was added in >> 8f3be339850ead96f9c6200db4e0db1f74e39d13 to fix a missing prototype >> for >> glDrawBuffersNV and others, but it's not possible to include both >> glext.h and gl2ext.h from the same file. >> >> I don't see the missing prototype here (with or without shared glapi) >> so I'm just removing the offending #include. >> >> Also, since we're redoing this, update to the most recent gl2ext.2. >> >> Signed-off-by: Kristian Høgsberg >> >> :04 04 3076bd4623a3727eb3690e39801c459ea2c08706 >> 7754f21c4cee87eb86ccb5ac8b9cbf810d0835ff M include >> :04 04 0e602fff716a390251eacb932e7f406eb399e2e0 >> ed4ecdd9127f4541ed0f148f3b97e6eefb6572df M src >> >> ___ >> mesa-dev mailing list >> mesa-dev@lists.freedesktop.org >> http://lists.freedesktop.org/mailman/listinfo/mesa-dev >> > > 0001-mapi-Move-GL_NV_draw_buffers-extension-to-es_EXT.xml.patch Description: Binary data ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] Mesa (master): docs: Update GL3.txt.
Kenneth Graunke writes: > inverse() has been done for a while. Does the inverse() builtin constant expression handling work for you? It doesn't here. > None of us know what "highp change" means; GLSL 1.40 spec: "Make the default precision qualification for fragment shader be high." -- it was also on our task list. pgpxO39mw0NHk.pgp Description: PGP signature ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] Mesa master branch: forced update
On Wed, Jul 11, 2012 at 7:15 AM, Kristian Høgsberg wrote: > On Tue, Jul 10, 2012 at 4:54 PM, Kristian Høgsberg wrote: >> On Tue, Jul 10, 2012 at 4:24 PM, Ferry Huberts wrote: >>> >>> On 10-07-12 22:13, Kenneth Graunke wrote: On 07/10/2012 12:50 PM, Tom Stellard wrote: > > > I just fetched from the master branch of the fdo mesa repo and was > greeted with a "forced update" message, and the gitweb interface shows > several days of history are missing from the master branch. > > olv appears to be the last user to modify the master branch: > > tstellar@annarchy:~$ ls -l /git/mesa/mesa.git/refs/heads/master > -rw-rw-r-- 1 olv mesa 41 Jul 10 11:41 > /git/mesa/mesa.git/refs/heads/master > > Anyone know what happened? >>> >>> >>> Login on the server, and look at the git logs. >>> The commits are not lost, just not visible. >>> >>> logs are in: >>> /logs >>> >>> or do: >>> cd >>> git reflog >> >> I already did that, there are no reflogs in the mesa git repo. The >> repo is older than the reflog feature. The best we can do is to look >> at the master ref. >> >> It's possible that this was an attack to alter history (sneak in a >> backdoor, for example, the dri drivers run as root in aiglx in most >> distros). However, the commit that was pushed matches the older >> commit (which is why Kenneth was able to pull and fast-forward) and >> git fsck verifies that the history hasn't been tampered with. That >> is, it is possible to hand edit a commit object to include changes >> that wasn't originally there and then just force the SHA1 to match >> what is was before. git fsck will catch that, but only in a new >> clone, since when you pull from an existing repo, git won't fetch old >> objects. More unlikely, history was altered in a way such that code >> was inserted but the sha1 was preserved (ie sha1 was compromised). >> I'm on a bad connection right now, but I'll do a fresh clone of the >> mesa repo and do a git fsck there as well as comparing the contents of >> a recent commit with what I have locally to see if the contents has >> been changed while preserving the sha1 validity. > > And the results are in: freshly cloned mesa repo goes through git fsck > without problems and just to be completely paranoid I checked against > compromised sha1sums (that is, attack by inserting code without > affecting the sha1sums) by comparing the output of git archive of > 40742fa6864000d431b81c3769a3136b7ff4a0d1 in both my previous checkout > and the fresh clone and they match. So while it's suspicious that > Chia-I hasn't been active for a long time and the suddenly pushes a > forced update of the repo, I don't think anything was compromised or > any history lost. The freedesktop.org account has been disabled until > we hear back from Chia-I. That was my fault. I intended to "git push -f" mesa to my other machine last night, but I forgot to specify the remote. The command was interrupted before completing, and I thought the changes were not pushed. I am sorry for the troubles... I wonder if it is feasible to set receive.denyNonFastForwards to true on the server to avoid such errors in the future? -- o...@lunarg.com ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
