When we specify a divider we are looking for slightly different races that are in theory not so timing dependent (since we are using vrefresh windows). However, since they each take some multiple of vblanks to complete, they are much slower. We should be able to reduce the number of iterations without losing detection capability. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> |
||
|---|---|---|
| .. | ||
| .gitignore | ||
| DrawSegments.c | ||
| Makefile.am | ||
| README | ||
| basic-copyarea-size.c | ||
| basic-copyarea.c | ||
| basic-copyplane.c | ||
| basic-fillrect.c | ||
| basic-lines.c | ||
| basic-putimage.c | ||
| basic-rectangle.c | ||
| basic-stippledrect.c | ||
| basic-stress.c | ||
| basic-string.c | ||
| basic-tiledrect.c | ||
| cursor-test.c | ||
| dri2-race.c | ||
| dri2-speed.c | ||
| dri2-swap.c | ||
| dri2-test.c | ||
| dri2.c | ||
| dri2.h | ||
| dri3-test.c | ||
| dri3.c | ||
| dri3.h | ||
| lowlevel-blt-bench.c | ||
| mixed-stress.c | ||
| mkvsync.sh | ||
| present-race.c | ||
| present-speed.c | ||
| present-test.c | ||
| render-composite-solid-mask.c | ||
| render-composite-solid.c | ||
| render-copy-alphaless.c | ||
| render-copyarea-mask.c | ||
| render-copyarea-size.c | ||
| render-copyarea.c | ||
| render-fill-copy.c | ||
| render-fill.c | ||
| render-glyphs.c | ||
| render-trapezoid-image.c | ||
| render-trapezoid.c | ||
| render-triangle.c | ||
| shm-test.c | ||
| tearing.mp4 | ||
| test.h | ||
| test_display.c | ||
| test_image.c | ||
| test_log.c | ||
| test_render.c | ||
| virtual.conf | ||
README
These are no substitute for xts, rendercheck and cairo-test-suite. They are intended to exercise corner cases in the batch management of long drawing commands and more explicit checking of the acceleration paths. Useful tools: # Packed YUV Xv tester gst-launch-1.0 videotestsrc pattern=snow ! 'video/x-raw,format=UYVY,width=640,height=360' ! xvimagesink # Planar YUV Xv tester gst-launch-1.0 videotestsrc pattern=snow ! 'video/x-raw,format=I420,width=640,height=360' ! xvimagesink