Closed Bug 622836 Opened 14 years ago Closed 12 years ago

[OSX] ray tracing demo causes OS level GL corruption, hangs / locks machine

Categories

(Core :: Graphics: CanvasWebGL, defect)

x86
macOS
defect
Not set
normal

Tracking

()

RESOLVED WORKSFORME
Tracking Status
blocking2.0 --- -

People

(Reporter: beltzner, Unassigned)

References

()

Details

(Whiteboard: rdar://8830307)

Attachments

(3 files)

The demo at http://www.lewpen.com/articles/3d-graphics/webgl/realtime-raytracer/raytracer-008-bumpy-textures/ reliably causes my 3-year old 13" MacBook to lock up and hang with corruption at what seems to be the OS level. I made a video of the symptom: http://www.youtube.com/watch?v=tj-Qep9COoQ Here's my graphics driver information from about:support Adapter Description 0x22600,0x20400 Vendor ID 0000 Device ID 0000 Adapter RAM Adapter Drivers Driver Version Driver Date Direct2D Enabled false DirectWrite Enabled false WebGL Renderer NVIDIA Corporation -- NVIDIA GeForce 9400M OpenGL Engine -- 2.1 NVIDIA-1.6.24GPU Accelerated Windows 1/1 OpenGL
blocking2.0: --- → ?
The bug does not show up on my new MacBook Air. Here's my graphics card info: WebGL Renderer NVIDIA Corporation -- NVIDIA GeForce 320M OpenGL Engine -- 2.1 NVIDIA-1.6.24 (the rest of the graphics section on about:support is identical to Beltzner's)
WFM with: Adapter Description 0x22600,0x20400 Vendor ID 0000 Device ID 0000 Adapter RAM Adapter Drivers Driver Version Driver Date Direct2D Enabled false DirectWrite Enabled false WebGL Renderer NVIDIA Corporation -- NVIDIA GeForce 9400M OpenGL Engine -- 2.1 NVIDIA-1.6.24GPU Accelerated Windows 1/1 OpenGL
Note that I have an ~1 year old 13" MBP
Cc Brian, his ~1yo MBP does fail.
Fails for me. My about:support doesn't have a WebGL Renderer line and the rest is identical to Beltzner's. Here's the information from System Profiler: Chipset Model: NVIDIA GeForce 9400M Type: GPU Bus: PCI VRAM (Total): 256 MB Vendor: NVIDIA (0x10de) Device ID: 0x0863 Revision ID: 0x00b1 ROM Revision: 3448
just some instrumentation to help understand this bug
This is all the shaders that are passed to compileShader() in this demo. Next step, in a rational world, would be to bisect them.
Note that you're looking at the shader sources pre-ANGLE translation; I think you want to look at both pre and post. There shouldn't be much difference for these shaders, but you never know.
(In reply to comment #8) > Note that you're looking at the shader sources pre-ANGLE translation; I think > you want to look at both pre and post. There shouldn't be much difference for > these shaders, but you never know. Mike confirmed that the problem persisted with webgl.shader_validator=false. But yes, someone bisecting this should keep that pref set to false.
Can someone tell me whether Chrome suffers the same issues?
(In reply to comment #10) > Can someone tell me whether Chrome suffers the same issues? I just tested on Chrome 10.0.612.3 dev, it suffers the same issue.
Thanks, Brian! I'm going to report this bug to Apple. This can't block, because there doesn't seem to be much we can do about it (it's an OS bug).
blocking2.0: ? → -
Attached file System Profiler dump
Here's a dump from my system profiler.
Whiteboard: rdar://8830307
What Mac OS versions are we talking about here? Can anyone reproduce this issue on 10.6 ?
This happens on 10.6, yeah. It's a known bug in OS X.
Blocks: 639671
WFM, but can I get confirmation if this is still a problem on the affected systems?
After a year and half of no comments, I'm marking this WFM. Please reopen if this reproduces for you.
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: