mirror of
https://gitlab.freedesktop.org/mesa/mesa.git
synced 2026-05-22 23:58:10 +02:00
Found this while investigating iova faults, with impossibly low iova (ie. fault addresses like 0x0000000000181800). Looks like fetching past the end of memory mapped GMEM when doing a CP_BLIT resolve blit. But because the fault GPU is waiting in a CP_WAIT_FOR_ME in IB1 after the resolve blit, crashdec wasn't correctly finding the location. We should just skip all the complicated bits working backwards in IB1 to find the jump into IB2 if we see that the crash is actually in IB1. Signed-off-by: Rob Clark <robdclark@chromium.org> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27508> |
||
|---|---|---|
| .. | ||
| scripts | ||
| buffers.c | ||
| buffers.h | ||
| cffdec.c | ||
| cffdec.h | ||
| cffdump.c | ||
| crashdec-hfi.c | ||
| crashdec-mempool.c | ||
| crashdec-prefetch.c | ||
| crashdec.c | ||
| crashdec.h | ||
| io.c | ||
| io.h | ||
| meson.build | ||
| pager.c | ||
| pager.h | ||
| pgmdump.c | ||
| pgmdump2.c | ||
| rdcompiler-meson.build | ||
| rdcompiler-utils.h | ||
| rddecompiler.c | ||
| rdutil.h | ||
| replay.c | ||
| rnnutil.c | ||
| rnnutil.h | ||
| script.c | ||
| script.h | ||
| util.h | ||