summaryrefslogtreecommitdiffstats
path: root/src
Commit message (Expand)AuthorAgeFilesLines
* gallium: Drop software-only primitive restart support.Eric Anholt2014-09-301-3/+2
* i965/fs: Properly calculate the number of instructions in calculate_register_...Jason Ekstrand2014-09-301-1/+3
* i965/fs: Use the GRF for FB writes on gen >= 7Jason Ekstrand2014-09-306-71/+142
* i965/fs: Handle COMPR4 in LOAD_PAYLOADJason Ekstrand2014-09-302-1/+36
* i965/fs: Constant propagate into LOAD_PAYLOADJason Ekstrand2014-09-301-0/+1
* i965/fs: Add split_virtual_grfs and compute_to_mrf after lower_load_payloadJason Ekstrand2014-09-301-0/+2
* i965/fs: Add a an optional source to the FS_OPCODE_FB_WRITE instructionJason Ekstrand2014-09-304-29/+28
* i965/fs: Use the GRF for UNTYPED_SURFACE_READ instructionsJason Ekstrand2014-09-304-16/+24
* i965/fs: Use the GRF for UNTYPED_ATOMIC instructionsJason Ekstrand2014-09-306-25/+36
* i965/fs: Add a function for getting a component of a 8 or 16-wide registerJason Ekstrand2014-09-301-0/+10
* i965/fs: Use the instruction execution size directly for texture generationJason Ekstrand2014-09-301-3/+10
* i965/fs: Use exec_size instead of force_uncompressed in dump_instructionJason Ekstrand2014-09-301-6/+7
* i965/fs: Use instruction execution sizes instead of heuristicsJason Ekstrand2014-09-303-23/+10
* i965/fs: Use instruction execution sizes to set compression stateJason Ekstrand2014-09-301-6/+19
* i965/fs: Remove unneeded uses of force_uncompressedJason Ekstrand2014-09-303-25/+9
* i965/fs: Derive force_uncompressed from instruction exec_sizeJason Ekstrand2014-09-301-0/+3
* i965/fs: Make fs_reg::effective_width take fs_inst* instead of fs_visitor*Jason Ekstrand2014-09-303-37/+43
* i965/fs: Better guess the width of LOAD_PAYLOADJason Ekstrand2014-09-301-2/+9
* i965/fs: Add an exec_size field to fs_instJason Ekstrand2014-09-305-32/+126
* i965/fs: Determine partial writes based on the destination widthJason Ekstrand2014-09-302-5/+3
* i965/fs: Fix a bug in register coalesceJason Ekstrand2014-09-301-0/+17
* i965/fs: Rework GEN5 texturing code to use fs_reg and offset()Jason Ekstrand2014-09-301-39/+38
* i965/fs_reg: Allocate double the number of vgrfs in SIMD16 modeJason Ekstrand2014-09-309-157/+371
* i965/fs: Handle printing of registers better.Jason Ekstrand2014-09-301-2/+6
* i965: Explicitly set widths on gen5 math instruction destinations.Jason Ekstrand2014-09-301-1/+1
* i965/fs: Make half() divide the register width by 2 and use it moreJason Ekstrand2014-09-302-5/+13
* i965/fs: Add a concept of a width to fs_regJason Ekstrand2014-09-302-4/+78
* i965/fs: A little harmless refactoring of register_coalesceJason Ekstrand2014-09-301-7/+7
* i965/brw_reg: Add a firsthalf function and use it in the generatorJason Ekstrand2014-09-302-29/+44
* i965/fs: Copy propagate partial reads.Jason Ekstrand2014-09-302-20/+64
* i965/fs: Refactor fs_inst::is_send_from_grf()Jason Ekstrand2014-09-301-9/+16
* i965/fs: Clean up emit_fb_writesJason Ekstrand2014-09-302-112/+85
* i965/fs: Print BAD_FILE registers in dump_instructionJason Ekstrand2014-09-301-1/+1
* i965/fs: Make compact_virtual_grfs an optimization passJason Ekstrand2014-09-302-8/+13
* i964/fs: Make immediate fs_reg constructors explicitJason Ekstrand2014-09-304-10/+11
* i965/fs: Make null_reg_* const members of fs_visitor instead of globalsJason Ekstrand2014-09-303-3/+12
* i965/fs: Use the var_from_vgrf helper function instead of doing it manuallyJason Ekstrand2014-09-301-4/+4
* i965/fs: Fix a bug with dead_code_eliminate on large writesJason Ekstrand2014-09-301-1/+1
* i965/fs: Use the UW type for the destination of VARYING_PULL_CONSTANT_LOAD in...Jason Ekstrand2014-09-301-2/+2
* i965/fs: Use offset a lot more placesJason Ekstrand2014-09-304-93/+78
* i965/fs: fix a comment in compact_virtual_grfsJason Ekstrand2014-09-301-1/+1
* i965/fs: Rewrite fs_visitor::split_virtual_grfsJason Ekstrand2014-09-301-47/+86
* i965/fs_live_variables: Use var_from_vgrf insead of repeating the calculationJason Ekstrand2014-09-301-2/+2
* i965/fs: Manually generate the meta fast-clear shaderJason Ekstrand2014-09-302-90/+35
* radeonsi: Pass the slice size to si_dma_copy_bufferMichel Dänzer2014-09-301-4/+4
* radeonsi: Catch more cases that can't be handled by si_dma_copy_buffer/tileMichel Dänzer2014-09-301-3/+11
* radeonsi: Fix si_dma_copy(_tile) for compressed formatsMichel Dänzer2014-09-301-2/+2
* radeonsi: Fix tiling mode index for stencil resourcesMichel Dänzer2014-09-301-2/+3
* ilo: fix format of edge flag pointerChia-I Wu2014-09-301-3/+5
* ilo: add a pass to finalize ilo_ve_stateChia-I Wu2014-09-308-158/+190