[viewer] [crash] Crash: Freeze, then crash #595
Labels
No labels
amd
avatars
bug
builder
cannot-reproduce
crash
epic
feature
from-viewer
gl-removal
known-limitation
monitoring
parity
perf
platform
priority
critical
priority
high
priority
low
priority
medium
provenance
rendering
rt
status
in-progress
status
testing
status
up-next
triage
ui
upstream
No milestone
No project
No assignees
2 participants
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
sky_line/Slipstream#595
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
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?
Source: in-viewer report (Help > Report Bug)
Reporter (avatar name as given; not verified):
Filed after a crash. Previous session ended with:
System
Description
Steps to reproduce
not given
Log excerpt
Full scrubbed excerpt: slipstream-log-excerpt.txt
Last 40 lines
Filed by the Slipstream report relay. The reporter's identity is not verified; system details are as sent by the reporter.
Thanks for the report and the log. Cause found: when video memory runs out (here when another app such as Discord takes some, or in busy places on 6-8 GB cards), a couple of buffer-resize and reflection-probe paths treated the failed allocation as fatal instead of backing off. Same root cause as #595, #597 and #599; a fix that makes these paths degrade quality instead of crashing is in progress.
Fix:
649ca0dfa8-- running out of video memory at runtime now degrades instead of hitting the fatal handler (relief + retry, soft refusal of large over-budget requests, buffer grows fall back to the exact size, geometry defers overflow, the VRAM ladder fast-steps), plus #600 (~320 MB back with ray-traced projector shadows) andd1ced71a7e(fast texture recovery). Verified with a 2000 MB fake budget plus injected allocation failures: full degrade, no crash, survives teleports, recovers. Needs confirmation on a real 8 GB card. Lands in 1.2.0 Beta 8.Fixed in 1.2.0: running out of video memory degrades quality instead of crashing. 1.3.0 Beta 7 goes further: a failed video or system memory allocation is retried later, then the asset is skipped and logged, never a crash. Closing. If you still see a freeze followed by a crash, please file a new report from the crash dialog.