Re: [Mesa-dev] [PATCH 0/9] nir: Add better helpers for handling constant sources
I landed the first four but it'd be nice if I could get some review on the intel bits. --Jason On Sat, Oct 20, 2018 at 12:55 PM Jason Ekstrand wrote: > Previously, the only thing we had was nir_src_as_const_value which returns > a pointer to a nir_const_value which was NULL if the source wasn't actually > a constant. This was great except that almost none of the users considered > anything other than 32-bit values. With 8, 16, and 64-bit values floating > around, we really shouldn't be hand-rolling it everywhere. Most of the > code is currently safe if you operate under the assumption that things like > array indices are always 32 bits. The glaring exception was some of the > helpers in nir_search_helpers.h where we definitely cannot be making that > assumption but were anyway. > > I've converted core NIR and i965 but have not written patches for vc4, ir3, > or radeon. > > Cc: Connor Abbott > Cc: Timothy Arceri > Cc: Eric Anholt > Cc: Rob Clark > Cc: Karol Herbst > Cc: Bas Nieuwenhuizen > > Jason Ekstrand (9): > nir: Add some new helpers for working with const sources > nir/search: Use nir_src_is_const and friends > nir/search_helpers: Use nir_src_is_const and friends > nir: Use nir_src_is_const and nir_src_as_* in core code > intel/fs,vec4: Clean up a repeated pattern with SSBOs > intel/fs: Use the new nir_src_is_const and friends > intel/vec4: Use the new nir_src_is_const and friends > intel/analyze_ubo_ranges: Use nir_src_is_const and friends > anv: Use nir_src_is_const and friends in lowering code > > src/compiler/glsl/gl_nir_lower_samplers.c | 7 +- > src/compiler/nir/nir.c| 92 + > src/compiler/nir/nir.h| 16 + > src/compiler/nir/nir_deref.c | 14 +- > src/compiler/nir/nir_gather_info.c| 6 +- > src/compiler/nir/nir_gs_count_vertices.c | 7 +- > src/compiler/nir/nir_lower_clip.c | 2 +- > src/compiler/nir/nir_lower_indirect_derefs.c | 4 +- > src/compiler/nir/nir_lower_io.c | 6 +- > .../nir/nir_lower_io_arrays_to_elements.c | 11 +- > src/compiler/nir/nir_lower_locals_to_regs.c | 6 +- > src/compiler/nir/nir_lower_two_sided_color.c | 2 +- > src/compiler/nir/nir_lower_vars_to_ssa.c | 14 +- > src/compiler/nir/nir_opt_dead_cf.c| 7 +- > src/compiler/nir/nir_opt_find_array_copies.c | 13 +- > src/compiler/nir/nir_opt_intrinsics.c | 4 +- > src/compiler/nir/nir_opt_large_constants.c| 2 +- > src/compiler/nir/nir_search.c | 70 +--- > src/compiler/nir/nir_search_helpers.h | 49 ++- > src/compiler/nir/nir_split_vars.c | 17 +- > src/compiler/nir/tests/vars_tests.cpp | 10 +- > src/intel/compiler/brw_fs.h | 2 + > src/intel/compiler/brw_fs_nir.cpp | 315 +++--- > .../compiler/brw_nir_analyze_ubo_ranges.c | 15 +- > src/intel/compiler/brw_vec4.h | 2 + > src/intel/compiler/brw_vec4_gs_nir.cpp| 12 +- > src/intel/compiler/brw_vec4_nir.cpp | 190 --- > src/intel/compiler/brw_vec4_tcs.cpp | 6 +- > .../vulkan/anv_nir_apply_pipeline_layout.c| 15 +- > .../vulkan/anv_nir_lower_ycbcr_textures.c | 6 +- > 30 files changed, 423 insertions(+), 499 deletions(-) > > -- > 2.19.1 > > ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH 0/9] nir: Add better helpers for handling constant sources
Re: The list. It doesn't always happen every month, but I've seen it multiple times now, as far back as January 2018, though possibly earlier. The gzip mail archive however has all the messages that are missing from the web archive. https://lists.freedesktop.org/archives/mesa-dev/2018-October.txt.gz There are patches in that series missing however, as you pointed out. That said, looking at that file I see that the last message that didn't get processed had the following Message ID header: Message-ID: <20181020175547.9159-6-jason.ekstr...@intel.com> The messages are a bit out of order there, which may or may not be related. I can't see anything else that seems out of the ordinary. I also get DMARC issues for certain emails (they fail DMARC, meaning that they've not been handled properly when passing through the list). I notice this particularly for emails from Liviu Prodea liviupro...@yahoo.com but sometimes for other ppl too. Perhaps a later version of Mailman (eg: 2.1.29?) might resolve some/both of these issues? On Mon, 22 Oct 2018 at 22:31, Samuel Iglesias Gonsálvez < sigles...@igalia.com> wrote: > Patch 7/9 has never arrived my inbox and checking the archives [0], looks > like > the archive is messed up... it has only a few emails. Does somebody know > what > happened with the ML archive? > > Sam > > [0] https://lists.freedesktop.org/archives/mesa-dev/2018-October/date.html > > On Saturday, 20 October 2018 19:55:38 (CEST) Jason Ekstrand wrote: > > Previously, the only thing we had was nir_src_as_const_value which > returns > > a pointer to a nir_const_value which was NULL if the source wasn't > actually > > a constant. This was great except that almost none of the users > considered > > anything other than 32-bit values. With 8, 16, and 64-bit values > floating > > around, we really shouldn't be hand-rolling it everywhere. Most of the > > code is currently safe if you operate under the assumption that things > like > > array indices are always 32 bits. The glaring exception was some of the > > helpers in nir_search_helpers.h where we definitely cannot be making that > > assumption but were anyway. > > > > I've converted core NIR and i965 but have not written patches for vc4, > ir3, > > or radeon. > > > > Cc: Connor Abbott > > Cc: Timothy Arceri > > Cc: Eric Anholt > > Cc: Rob Clark > > Cc: Karol Herbst > > Cc: Bas Nieuwenhuizen > > > > Jason Ekstrand (9): > > nir: Add some new helpers for working with const sources > > nir/search: Use nir_src_is_const and friends > > nir/search_helpers: Use nir_src_is_const and friends > > nir: Use nir_src_is_const and nir_src_as_* in core code > > intel/fs,vec4: Clean up a repeated pattern with SSBOs > > intel/fs: Use the new nir_src_is_const and friends > > intel/vec4: Use the new nir_src_is_const and friends > > intel/analyze_ubo_ranges: Use nir_src_is_const and friends > > anv: Use nir_src_is_const and friends in lowering code > > > > src/compiler/glsl/gl_nir_lower_samplers.c | 7 +- > > src/compiler/nir/nir.c| 92 + > > src/compiler/nir/nir.h| 16 + > > src/compiler/nir/nir_deref.c | 14 +- > > src/compiler/nir/nir_gather_info.c| 6 +- > > src/compiler/nir/nir_gs_count_vertices.c | 7 +- > > src/compiler/nir/nir_lower_clip.c | 2 +- > > src/compiler/nir/nir_lower_indirect_derefs.c | 4 +- > > src/compiler/nir/nir_lower_io.c | 6 +- > > .../nir/nir_lower_io_arrays_to_elements.c | 11 +- > > src/compiler/nir/nir_lower_locals_to_regs.c | 6 +- > > src/compiler/nir/nir_lower_two_sided_color.c | 2 +- > > src/compiler/nir/nir_lower_vars_to_ssa.c | 14 +- > > src/compiler/nir/nir_opt_dead_cf.c| 7 +- > > src/compiler/nir/nir_opt_find_array_copies.c | 13 +- > > src/compiler/nir/nir_opt_intrinsics.c | 4 +- > > src/compiler/nir/nir_opt_large_constants.c| 2 +- > > src/compiler/nir/nir_search.c | 70 +--- > > src/compiler/nir/nir_search_helpers.h | 49 ++- > > src/compiler/nir/nir_split_vars.c | 17 +- > > src/compiler/nir/tests/vars_tests.cpp | 10 +- > > src/intel/compiler/brw_fs.h | 2 + > > src/intel/compiler/brw_fs_nir.cpp | 315 +++--- > > .../compiler/brw_nir_analyze_ubo_ranges.c | 15 +- > > src/intel/compiler/brw_vec4.h | 2 + > > src/intel/compiler/brw_vec4_gs_nir.cpp| 12 +- > > src/intel/compiler/brw_vec4_nir.cpp | 190 --- > > src/intel/compiler/brw_vec4_tcs.cpp | 6 +- > > .../vulkan/anv_nir_apply_pipeline_layout.c| 15 +- > > .../vulkan/anv_nir_lower_ycbcr_textures.c | 6 +- > > 30 files changed, 423 insertions(+), 499 deletions(-) > > ___ > mesa-dev mailing list > mesa-dev@lists.freedesktop.org > https://lists.freedesktop.org/mailman
Re: [Mesa-dev] [PATCH 0/9] nir: Add better helpers for handling constant sources
Yeah... Bas said he had the same problem. The series can be found on my wip/nir-const-src branch of you want to see it that way. I can also resend if it's helpful. --Jason On October 22, 2018 06:31:01 Samuel Iglesias Gonsálvez wrote: Patch 7/9 has never arrived my inbox and checking the archives [0], looks like the archive is messed up... it has only a few emails. Does somebody know what happened with the ML archive? Sam [0] https://lists.freedesktop.org/archives/mesa-dev/2018-October/date.html On Saturday, 20 October 2018 19:55:38 (CEST) Jason Ekstrand wrote: Previously, the only thing we had was nir_src_as_const_value which returns a pointer to a nir_const_value which was NULL if the source wasn't actually a constant. This was great except that almost none of the users considered anything other than 32-bit values. With 8, 16, and 64-bit values floating around, we really shouldn't be hand-rolling it everywhere. Most of the code is currently safe if you operate under the assumption that things like array indices are always 32 bits. The glaring exception was some of the helpers in nir_search_helpers.h where we definitely cannot be making that assumption but were anyway. I've converted core NIR and i965 but have not written patches for vc4, ir3, or radeon. Cc: Connor Abbott Cc: Timothy Arceri Cc: Eric Anholt Cc: Rob Clark Cc: Karol Herbst Cc: Bas Nieuwenhuizen Jason Ekstrand (9): nir: Add some new helpers for working with const sources nir/search: Use nir_src_is_const and friends nir/search_helpers: Use nir_src_is_const and friends nir: Use nir_src_is_const and nir_src_as_* in core code intel/fs,vec4: Clean up a repeated pattern with SSBOs intel/fs: Use the new nir_src_is_const and friends intel/vec4: Use the new nir_src_is_const and friends intel/analyze_ubo_ranges: Use nir_src_is_const and friends anv: Use nir_src_is_const and friends in lowering code src/compiler/glsl/gl_nir_lower_samplers.c | 7 +- src/compiler/nir/nir.c| 92 + src/compiler/nir/nir.h| 16 + src/compiler/nir/nir_deref.c | 14 +- src/compiler/nir/nir_gather_info.c| 6 +- src/compiler/nir/nir_gs_count_vertices.c | 7 +- src/compiler/nir/nir_lower_clip.c | 2 +- src/compiler/nir/nir_lower_indirect_derefs.c | 4 +- src/compiler/nir/nir_lower_io.c | 6 +- .../nir/nir_lower_io_arrays_to_elements.c | 11 +- src/compiler/nir/nir_lower_locals_to_regs.c | 6 +- src/compiler/nir/nir_lower_two_sided_color.c | 2 +- src/compiler/nir/nir_lower_vars_to_ssa.c | 14 +- src/compiler/nir/nir_opt_dead_cf.c| 7 +- src/compiler/nir/nir_opt_find_array_copies.c | 13 +- src/compiler/nir/nir_opt_intrinsics.c | 4 +- src/compiler/nir/nir_opt_large_constants.c| 2 +- src/compiler/nir/nir_search.c | 70 +--- src/compiler/nir/nir_search_helpers.h | 49 ++- src/compiler/nir/nir_split_vars.c | 17 +- src/compiler/nir/tests/vars_tests.cpp | 10 +- src/intel/compiler/brw_fs.h | 2 + src/intel/compiler/brw_fs_nir.cpp | 315 +++--- .../compiler/brw_nir_analyze_ubo_ranges.c | 15 +- src/intel/compiler/brw_vec4.h | 2 + src/intel/compiler/brw_vec4_gs_nir.cpp| 12 +- src/intel/compiler/brw_vec4_nir.cpp | 190 --- src/intel/compiler/brw_vec4_tcs.cpp | 6 +- .../vulkan/anv_nir_apply_pipeline_layout.c| 15 +- .../vulkan/anv_nir_lower_ycbcr_textures.c | 6 +- 30 files changed, 423 insertions(+), 499 deletions(-) ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH 0/9] nir: Add better helpers for handling constant sources
Patch 7/9 has never arrived my inbox and checking the archives [0], looks like the archive is messed up... it has only a few emails. Does somebody know what happened with the ML archive? Sam [0] https://lists.freedesktop.org/archives/mesa-dev/2018-October/date.html On Saturday, 20 October 2018 19:55:38 (CEST) Jason Ekstrand wrote: > Previously, the only thing we had was nir_src_as_const_value which returns > a pointer to a nir_const_value which was NULL if the source wasn't actually > a constant. This was great except that almost none of the users considered > anything other than 32-bit values. With 8, 16, and 64-bit values floating > around, we really shouldn't be hand-rolling it everywhere. Most of the > code is currently safe if you operate under the assumption that things like > array indices are always 32 bits. The glaring exception was some of the > helpers in nir_search_helpers.h where we definitely cannot be making that > assumption but were anyway. > > I've converted core NIR and i965 but have not written patches for vc4, ir3, > or radeon. > > Cc: Connor Abbott > Cc: Timothy Arceri > Cc: Eric Anholt > Cc: Rob Clark > Cc: Karol Herbst > Cc: Bas Nieuwenhuizen > > Jason Ekstrand (9): > nir: Add some new helpers for working with const sources > nir/search: Use nir_src_is_const and friends > nir/search_helpers: Use nir_src_is_const and friends > nir: Use nir_src_is_const and nir_src_as_* in core code > intel/fs,vec4: Clean up a repeated pattern with SSBOs > intel/fs: Use the new nir_src_is_const and friends > intel/vec4: Use the new nir_src_is_const and friends > intel/analyze_ubo_ranges: Use nir_src_is_const and friends > anv: Use nir_src_is_const and friends in lowering code > > src/compiler/glsl/gl_nir_lower_samplers.c | 7 +- > src/compiler/nir/nir.c| 92 + > src/compiler/nir/nir.h| 16 + > src/compiler/nir/nir_deref.c | 14 +- > src/compiler/nir/nir_gather_info.c| 6 +- > src/compiler/nir/nir_gs_count_vertices.c | 7 +- > src/compiler/nir/nir_lower_clip.c | 2 +- > src/compiler/nir/nir_lower_indirect_derefs.c | 4 +- > src/compiler/nir/nir_lower_io.c | 6 +- > .../nir/nir_lower_io_arrays_to_elements.c | 11 +- > src/compiler/nir/nir_lower_locals_to_regs.c | 6 +- > src/compiler/nir/nir_lower_two_sided_color.c | 2 +- > src/compiler/nir/nir_lower_vars_to_ssa.c | 14 +- > src/compiler/nir/nir_opt_dead_cf.c| 7 +- > src/compiler/nir/nir_opt_find_array_copies.c | 13 +- > src/compiler/nir/nir_opt_intrinsics.c | 4 +- > src/compiler/nir/nir_opt_large_constants.c| 2 +- > src/compiler/nir/nir_search.c | 70 +--- > src/compiler/nir/nir_search_helpers.h | 49 ++- > src/compiler/nir/nir_split_vars.c | 17 +- > src/compiler/nir/tests/vars_tests.cpp | 10 +- > src/intel/compiler/brw_fs.h | 2 + > src/intel/compiler/brw_fs_nir.cpp | 315 +++--- > .../compiler/brw_nir_analyze_ubo_ranges.c | 15 +- > src/intel/compiler/brw_vec4.h | 2 + > src/intel/compiler/brw_vec4_gs_nir.cpp| 12 +- > src/intel/compiler/brw_vec4_nir.cpp | 190 --- > src/intel/compiler/brw_vec4_tcs.cpp | 6 +- > .../vulkan/anv_nir_apply_pipeline_layout.c| 15 +- > .../vulkan/anv_nir_lower_ycbcr_textures.c | 6 +- > 30 files changed, 423 insertions(+), 499 deletions(-) signature.asc Description: This is a digitally signed message part. ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [PATCH 0/9] nir: Add better helpers for handling constant sources
Previously, the only thing we had was nir_src_as_const_value which returns a pointer to a nir_const_value which was NULL if the source wasn't actually a constant. This was great except that almost none of the users considered anything other than 32-bit values. With 8, 16, and 64-bit values floating around, we really shouldn't be hand-rolling it everywhere. Most of the code is currently safe if you operate under the assumption that things like array indices are always 32 bits. The glaring exception was some of the helpers in nir_search_helpers.h where we definitely cannot be making that assumption but were anyway. I've converted core NIR and i965 but have not written patches for vc4, ir3, or radeon. Cc: Connor Abbott Cc: Timothy Arceri Cc: Eric Anholt Cc: Rob Clark Cc: Karol Herbst Cc: Bas Nieuwenhuizen Jason Ekstrand (9): nir: Add some new helpers for working with const sources nir/search: Use nir_src_is_const and friends nir/search_helpers: Use nir_src_is_const and friends nir: Use nir_src_is_const and nir_src_as_* in core code intel/fs,vec4: Clean up a repeated pattern with SSBOs intel/fs: Use the new nir_src_is_const and friends intel/vec4: Use the new nir_src_is_const and friends intel/analyze_ubo_ranges: Use nir_src_is_const and friends anv: Use nir_src_is_const and friends in lowering code src/compiler/glsl/gl_nir_lower_samplers.c | 7 +- src/compiler/nir/nir.c| 92 + src/compiler/nir/nir.h| 16 + src/compiler/nir/nir_deref.c | 14 +- src/compiler/nir/nir_gather_info.c| 6 +- src/compiler/nir/nir_gs_count_vertices.c | 7 +- src/compiler/nir/nir_lower_clip.c | 2 +- src/compiler/nir/nir_lower_indirect_derefs.c | 4 +- src/compiler/nir/nir_lower_io.c | 6 +- .../nir/nir_lower_io_arrays_to_elements.c | 11 +- src/compiler/nir/nir_lower_locals_to_regs.c | 6 +- src/compiler/nir/nir_lower_two_sided_color.c | 2 +- src/compiler/nir/nir_lower_vars_to_ssa.c | 14 +- src/compiler/nir/nir_opt_dead_cf.c| 7 +- src/compiler/nir/nir_opt_find_array_copies.c | 13 +- src/compiler/nir/nir_opt_intrinsics.c | 4 +- src/compiler/nir/nir_opt_large_constants.c| 2 +- src/compiler/nir/nir_search.c | 70 +--- src/compiler/nir/nir_search_helpers.h | 49 ++- src/compiler/nir/nir_split_vars.c | 17 +- src/compiler/nir/tests/vars_tests.cpp | 10 +- src/intel/compiler/brw_fs.h | 2 + src/intel/compiler/brw_fs_nir.cpp | 315 +++--- .../compiler/brw_nir_analyze_ubo_ranges.c | 15 +- src/intel/compiler/brw_vec4.h | 2 + src/intel/compiler/brw_vec4_gs_nir.cpp| 12 +- src/intel/compiler/brw_vec4_nir.cpp | 190 --- src/intel/compiler/brw_vec4_tcs.cpp | 6 +- .../vulkan/anv_nir_apply_pipeline_layout.c| 15 +- .../vulkan/anv_nir_lower_ycbcr_textures.c | 6 +- 30 files changed, 423 insertions(+), 499 deletions(-) -- 2.19.1 ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev