Created attachment 101915 [details] linux 3.15.1 On kernels after ~3.14, some time after nouveau fails to initialize (<60 seconds) ~all of userland becomes mostly unreponsive (cpu time preemption?) [dmesg1]. The 'unresponsive' behavior starts at around the same time 56.740245 happens. I bisected, and found the first commit where this happens is 1f1ac3bf2a804911ad46076b10ce15364237a698. Removing this commit causes the previous behavior: initialization fails as usual but nothing else interesting happens.
Created attachment 101916 [details] linux-3.16-rc2+ and darktama/nouveau at 6dd425510 I observed behavior similar/identical to the former on linux 3.16-rc + darktama/nouveau at 6dd425510 [dmesg2].
I submitted an mmiotrace from nvidia-340.17 to mmio.dumps (module load, X startup, glxgears, X shutdown) with subject "bug 80627".
The ioread32 craziness vanished on or before darktama/nouveau @ 19ae6962 and kernel 3.17-mw.
The remaining issue is already discussed in bug 70354. *** This bug has been marked as a duplicate of bug 70354 ***
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.