Created attachment 71822 [details] i915_error_state after the hang Bug description: System hang when starting OpenGL java applications. System environment: -- chipset: Integrated Graphics Chipset: Intel(R) Ivybridge Mobile (GT2) -- system architecture: 32-bit -- xf86-video-intel: 2.20.15 -- xserver: 1.12.3 -- mesa: 9.0.1 -- libdrm: 2.4.40 -- kernel: 3.6.6-porteus -- Linux distribution: Porteus 2.0 (pre-RC1) -- Machine or mobo model: Intel ChiefRiver -- Display connector: HDMI1 Reproducing steps: I can reproduce this 100% of the time, on both the above-described hardware, and an Ivybridge desktop system. I am running a proprietary JOGL-based application. I do not get the hang with trivial apps like glxgears. Additional info: The current revision of Porteus is based on a graphic stack that is over a year old. That is completely stable on this hardware, but anti-aliasing does not work, so for the release, they built an updated version for me to try. That version hangs the GPU.
Created attachment 71823 [details] dmesg after the hang
Created attachment 71824 [details] xorg log with logverbose 6
Created attachment 71825 [details] lspci showing specific ivybridge hardware config
Created attachment 71870 [details] apitrace of the GL application up to and including the hang
Does replaying the trace reproduce the hang for you? It doesn't hang for me. I'm using an Ivybridge GT2 mobile (lspci -nn shows 8086:0166) and I tested with both Mesa master and 9.0.1. No hangs. I do have a newer kernel, however...I'm using f01af9f from Linus' tree (newer than 3.7). Would it be hard for you to try a newer kernel?
I'm doing this on top of Porteus, and I've just re-run my test using Porteus 2.0RC1, which is based on Linux 3.7.1. I'm still getting the hang on both my Desktop and Mobile Ivybridge systems.
After extensive trial and error, I have determined that this hang only happens when using 4x FSAA. If I setNumSamples(0) in the capabilities when creating the pbuffer, it works; but if I setNumSamples(4) in the capabilities, the GPU hangs. So I have a work-around for now (not using FSAA). I'm going to investigate MSAA, as that could be a suitable alternate workaround.
Dear Reporter, This Mesa bug has been in the "NEEDINFO" status for over 60 days. I am closing this bug based on lack of response but feel free to reopen if resolution is still needed. Please ensure you're supplying the correct information as requested. Thank you.
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.