| Summary: | Application: KWin (kwin), signal: Segmentation fault | ||
|---|---|---|---|
| Product: | [Plasma] kwin | Reporter: | omega <biasquez> |
| Component: | general | Assignee: | KWin default assignee <kwin-bugs-null> |
| Status: | RESOLVED UPSTREAM | ||
| Severity: | crash | ||
| Priority: | NOR | ||
| Version: | unspecified | ||
| Target Milestone: | --- | ||
| Platform: | Ubuntu | ||
| OS: | Linux | ||
| Latest Commit: | Version Fixed In: | ||
| Sentry Crash Report: | |||
|
Description
omega
2011-03-20 23:06:31 UTC
Weird backtrace....
a) this seems to come from the logout effect, can you ensure you did not click some "logout" button?
b) the crash happens in an OpenGL call of the mipmap fallback to GL_LINEAR, which
b1) would be an OpenGL / Upstream bug
b2) makes no sense because every nvidia GPU supported by that driver should be fully capable of mipmapping -> no need for the fallback path
Do all backtraces look exactly like this?
The trace makes few sense to me. It crashes in an OpenGL call (what should not happen anyway) of the fallback* routine nof the mipmap generation**
So this could defer from some memory corruption.
-> do all backtraces look the same?
-> is the condition always the same ("clicked chromium icon")
-> Please don't feel offended, but do you actually have an nvidia GPU supported by that driver (>= 6xxx)? - i'm just asking because i recall someone who wanted to boost his system's OpenGL support by installing the nvidia driver on an intel chip ;-)
a) i didn't click on logout button but on chromium browser icon on task bar b) i have nvidia 9600GT card there's likely some overflow then. a) "do all backtraces look the same"?? b) "click chromium icon" mean "to launch it" or for a running application. In the latter case, what's the state of the browser (minimized, obscured by other windows, things like this) c) does this /only/ happen for chromium? a) i had this issue once b) chromium browser was minimized ( no launch it but only running application) c) this happened once with chromium browser Then we have a heisenbug :-( Could have been the un/minimization or the slideback effect - or somewhere else. Could even be corrupted HW. The provied trace makes little sense, sorry. as it crashes in the NVIDIA blob, I set to upstream |