Created attachment 93816 [details] boot log from 3.14-rc2 with nouveau.debug=PDISP=debug,VBIOS=trace Using dual monitors, one on DVI and one on Displayport, the displayport output is blank. The monitor will not come out of sleep mode. xrandr reports both displays and indicates both displays are active. Have tried disabling/enabling outputs, and changing orientation. Tested against kernels: 3.8.13 - works ok on both outputs 3.9.11 - no DP output 3.10.26 - no DP output 3.14-rc2 - no DP output Tested against following cards: Quadro 400 (GT216GL) - no DP output Quadro 600 (GF108FL) - no DP output Quadro K600 (GK107) - both outputs fine OS is stock Debian Wheezy 7.2 with updated kernel. Attached boot log from 3.14-rc2 with 'nouveau.debug=PDISP=debug,VBIOS=trace'
There's another bug about a DP regression... bug #67628, which bisected down to 5cc027f6b1e ("drm/nv50-/disp: handle supervisor tasks from workqueue"), first appearing in 3.9-rc1. Could you verify that you're having trouble with the same commit? If so, we can mark as a duplicate. Otherwise, which log is that log from? Could you upload your VBIOS?
(In reply to comment #1) > There's another bug about a DP regression... bug #67628, which bisected down > to 5cc027f6b1e ("drm/nv50-/disp: handle supervisor tasks from workqueue"), > first appearing in 3.9-rc1. Could you verify that you're having trouble with > the same commit? If so, we can mark as a duplicate. > > Otherwise, which log is that log from? Could you upload your VBIOS? Unfortunately that doesn't seem to be the commit that caused the problem (unless I did something wrong in testing). It does appear to be between tags 'v3.8' and 'v3.9-rc1'. I'm trying to do a git bisect but it looks like it will take a while (13 steps estimated, but the first kernel wouldn't boot...). Any tips for speeding up that process? Any other info I can provide?
After seeing that commit 5cc027f6b1e was still good, I bisected the rest of the commits in that merge, and came to the very next commit, 0a0afd282f ("drm/nv50-/disp: move DP link training to core and train from supervisor") that breaks the displayport output (which is not surprising given the commit message). What additional information can I provide to help find a resolution to this bug?
Oops, after looking at the referenced bug #67628, that it actually finds the same commit (0a0afd282f, not 5cc027f6b1e) as the first bad commit. So, I'll mark this as a duplicate of the that bug. Sorry for the confusion.
*** This bug has been marked as a duplicate of bug 67628 ***
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.