Summary: | [radeonsi] outlast segfault on start | ||
---|---|---|---|
Product: | Mesa | Reporter: | Laurent carlier <lordheavym> |
Component: | Drivers/Gallium/radeonsi | Assignee: | Default DRI bug account <dri-devel> |
Status: | RESOLVED FIXED | QA Contact: | Default DRI bug account <dri-devel> |
Severity: | normal | ||
Priority: | medium | CC: | EoD, shawn.starr |
Version: | git | ||
Hardware: | x86-64 (AMD64) | ||
OS: | Linux (All) | ||
Whiteboard: | |||
i915 platform: | i915 features: | ||
Attachments: | Invalid reads flagged by valgrind |
Description
Laurent carlier
2016-02-24 22:49:00 UTC
Bisect gives me: 2b938a390c15a06be8cf706083890c822979508f is the first bad commit commit 2b938a390c15a06be8cf706083890c822979508f Author: Ilia Mirkin <imirkin@alum.mit.edu> Date: Thu Feb 18 01:04:13 2016 -0500 st/mesa: fix pbo uploads - LOD must be provided in .w for TXF (even for buffer textures) - User buffer must be valid at draw time - Must have a sampler associated with the sampler view This makes PBO uploads work again on nouveau. Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu> Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com> :040000 040000 ee1148bebe6a6658a6c4e26b9841e7eaf22805a3 3f90d6d20a63250fd784635976063f2da23217e2 M src I can reproduce the segfault with this apitrace: https://drive.google.com/file/d/0B1WCo3k21FK3RTVReVl3TkRZY3M/view?usp=sharing Created attachment 121980 [details] Invalid reads flagged by valgrind I got these when replaying the apitrace from bug 94242. Looks like there may be some sampler view / resource referencing imbalance somewhere, resulting in the radeonsi driver accessing memory of a resource which was already destroyed. *** Bug 94326 has been marked as a duplicate of this bug. *** Fixed by commit 28d2a7e67 "radeonsi: avoid crash when a sampler state is bound for a buffer texture" (In reply to Nicolai Hähnle from comment #5) > Fixed by commit 28d2a7e67 "radeonsi: avoid crash when a sampler state is > bound for a buffer texture" The commit fixes my apitrace ARK-replay of bug 94326, though ARK still crashes now with: Mesa: User error: GL_OUT_OF_MEMORY in glMapBufferRange(map failed) Assertion failed: Data != NULL [File:Runtime/OpenGLDrv/Public/OpenGLResources.h] [Line: 231] Signal 11 caught. Gtk-Message: GtkDialog mapped without a transient parent. This is discouraged. LowLevelFatalError [File:C:\SVN_Ark\Engine\Source\Runtime\Core\Private\GenericPlatform\GenericPlatformMemory.cpp] [Line: 51] Ran out of memory allocating 65536 bytes with alignment 0 Using binned. 4.5.1-0+UE4 7038 3077 402 5 Is this related? Almost certainly not, please open a separate bug report. |
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.