In <[🔎] 201105081801.00768.bss@iguanasuicide.net>, Boyd Stephen Smith Jr. wrote: >I'm off to experiment some more and see if I can get accelerated nouveau >working on something. I guess even the testing/Sid kernels are worth >another attempt. I won't have the nvidia kernel module tainting the >kernel so which will be an advantage for debugging even if the crash still >occurs. 2.6.32 still fails when using the mesa version that has nouveau acceleration. Still something DRM/DRI related. Quite unfortunate. 2.6.38 (Sid) kernel panics due to an MCE thrown from somewhere in the nouveau kernel module after switching to the framebuffer. This is before I can get a console, so it's pretty hard to capture the associated trace; I may be able to get a USB-to-Serial and a Null-Modem cable this week in order to capture the requisite information for a bug report. 2.6.38 (Backports), 2.6.39-rc5 (experimental), and 2.6.39-rc6 (experimental) also exhibit the same behavior as 2.6.38 (Sid), although the backtrace is slightly different as far as I can tell. So, at least on my hardware, if you want to use nouveau, you have to forgo any type of acceleration AND you must use the kernel from stable. For me, this likely means switching back to the NVidia driver eventually. First, I'd like to get these bugs filed. Even after that, I will likely hold out until there's a new linux-kbuild package in Sid; at that point I will want to test out btrfs on the new kernel. -- Boyd Stephen Smith Jr. ,= ,-_-. =. bss@iguanasuicide.net ((_/)o o(\_)) ICQ: 514984 YM/AIM: DaTwinkDaddy `-'(. .)`-' http://iguanasuicide.net/ \_/
Attachment:
signature.asc
Description: This is a digitally signed message part.