summaryrefslogtreecommitdiff
AgeCommit message (Expand)AuthorFilesLines
2015-10-15armv7: Add in_reverse_8888_8888 fast pathBen Avison2-0/+17
2015-10-15armv7: Add IN, IN_REVERSE and OUT combiners, and rework OUT_REVERSEBen Avison2-106/+311
2015-10-15armv7: Stop over_8888_8888 relying on out_reverse_8888_8888 macrosBen Avison1-2/+14
2015-10-15armv7: Split macros for over_reverse_8888_n_8888 & out_reverse_8888_8888_8888Ben Avison1-2/+39
2015-10-15armv7: Add SRC combinerBen Avison2-0/+65
2015-10-15armv6: Add SRC combinerBen Avison2-0/+107
2015-10-15armv6: Add OVER combinerBen Avison2-0/+142
2015-10-15armv6: Add IN, IN_REVERSE, OUT and OUT_REVERSE combinersBen Avison2-0/+384
2015-10-15armv6: Add OVER_REVERSE combinerBen Avison2-0/+257
2015-10-15armv6: Add ADD combinerBen Avison2-0/+99
2015-10-15arm: Move BIND_COMBINE_U from NEON code to a generic ARM headerBen Avison2-30/+31
2015-10-15armv6: Add ability to generate single-scanline fast pathsBen Avison1-15/+89
2015-10-15armv6: Improved over_8888_8888 fast pathBen Avison1-69/+78
2015-10-15armv7: Move common bilinear macro definitions to a new header fileBen Avison4-281/+169
2015-10-15armv7: Use aligned memory writes in both copies of bilinear codeBen Avison1-8/+8
2015-10-15armv7: Use prefetch for small-width images tooBen Avison1-0/+9
2015-10-15armv7: Simplify constant loadBen Avison1-2/+1
2015-10-15armv7: Use VLD-to-all-lanesBen Avison1-84/+58
2015-10-15armv7: Faster fill operationsBen Avison1-0/+7
2015-10-15armv7: Coalesce scalar accesses where possibleBen Avison1-0/+4
2015-10-15Require destination images to be bitmapsBen Avison1-1/+2
2015-10-15composite flags: Early detection of fully opaque 1x1 repeating source imagesBen Avison1-3/+25
2015-10-15pixman-fast-path: Make bilinear cover fetcher use COVER_CLIP_TIGHT flagBen Avison1-12/+25
2015-10-15armv7/mips/sse2: Make bilinear cover fast paths use COVER_CLIP_TIGHT flagBen Avison1-4/+65
2015-10-15Introduce FAST_PATH_SAMPLES_COVER_CLIP_TIGHT_BILINEAR flagBen Avison2-0/+9
2015-10-13mmx: Use MMX2 intrinsics from xmmintrin.h directly.Matt Turner2-71/+8
2015-09-29vmx: implement fast path vmx_composite_over_n_8888Siarhei Siamashka1-0/+54
2015-09-25affine-bench: remove 8e margin from COVER areaBen Avison1-6/+18
2015-09-25Remove the 8e extra safety margin in COVER_CLIP analysisBen Avison1-13/+4
2015-09-25pixman-general: Tighten up calculation of temporary buffer sizesBen Avison1-2/+2
2015-09-22pixman-general: Fix stack related pointer arithmetic overflowSiarhei Siamashka1-9/+7
2015-09-20test: add a check for FE_DIVBYZEROThomas Petazzoni2-0/+7
2015-09-18vmx: Remove unused expensive functionsOded Gabbay1-196/+0
2015-09-18vmx: implement fast path vmx_composite_over_n_8_8888Oded Gabbay1-0/+111
2015-09-18vmx: optimize vmx_composite_over_n_8888_8888_caOded Gabbay1-31/+21
2015-09-18vmx: optimize scaled_nearest_scanline_vmx_8888_8888_OVEROded Gabbay1-62/+17
2015-09-17armv6: enable over_n_8888Pekka Paalanen1-5/+4
2015-09-17armv6: Add over_n_8888 fast path (disabled)Ben Avison2-0/+48
2015-09-16test: Add cover-test v5Ben Avison2-0/+450
2015-09-09implementation: add PIXMAN_DISABLE=wholeopsPekka Paalanen1-0/+16
2015-09-09utils.[ch]: add fence_get_page_size()Pekka Paalanen2-3/+22
2015-09-09utils.c: fix fallback code for fence_image_create_bits()Pekka Paalanen1-1/+1
2015-09-03test: add fence-image-self-testPekka Paalanen2-0/+238
2015-09-01utils.[ch]: add fence_image_create_bits ()Pekka Paalanen2-0/+112
2015-09-01utils.[ch]: add FENCE_MALLOC_ACTIVEPekka Paalanen2-3/+14
2015-08-28scaling-test: list more details when verboseBen Avison1-22/+44
2015-08-18lowlevel-blt-bench: make extra arguments an errorPekka Paalanen1-0/+6
2015-08-01Post-release version bump to 0.33.3Oded Gabbay1-1/+1
2015-08-01Pre-release version bump to 0.33.2Oded Gabbay1-1/+1
2015-07-16vmx: implement fast path iterator vmx_fetch_a8Oded Gabbay1-0/+46
generated by cgit v1.2.3 (git 2.39.1) at 2024-05-23 21:33:32 +0000