Fix GPU synchronization errors after frame dumps #162
No reviewers
Labels
No labels
accessibility
autorelease: pending
autorelease: tagged
bug
dependencies
documentation
duplicate
enhancement
github_actions
good first issue
help wanted
invalid
question
testing wanted
wontfix
No milestone
No assignees
1 participant
Due date
No due date set.
Dependencies
No dependencies set
Reference
MagicBeanDev/Hookshot!162
Loading…
Reference in a new issue
No description provided.
Delete branch "codex/dump-timeline-values"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
What this fixes
Returning from a frame dump to normal rendering can signal an already reached timeline value and report
VUID-VkSubmitInfo2-semaphore-03882. This can stall rendering or break synchronization after capture.Why merge
Change
gpu/vk/vk_replay.cpp: share timeline allocation across normal/dump modes and protect synchronous allocation/submission ordering.nx2/test_timeline_modes.py: check mode switches, queued/legacy paths and reserved present values.Tested
python nx2/test_timeline_modes.py --cxx clang++checks 100 transitions and allocation controls. Fails on the old code, passes now.03882after the frame dump. With this fix, a 258-second run completes the dump and continues rendering with zero validation errors. The same #217/#225 shader fixes are included in both runs.Not covered