MacBook Pro (Mid 2010 model) with Nvidia NV50 using nouveau driver on Arch Linux.
Using both the packaged xf86-video-nouveau (0.0.16_git20100819) and nouveau from git, the display fails to resume after sleep. Happens when closing the lid, and using `pm-suspend`, even after removing pm scripts 98video-quirks and 99-video as suggested in other bug trackers.
After sleeping, the display never comes on, nor does the display backlight. External displays connected to the Mini-DisplayPort->DVI stay asleep (amber LED). Sleeping from and/or waking to a non-X virtual terminal (eg, Ctrl-Alt-F1) does not restore the display, nor does switching after waking. The laptop is responsive, keyboard backlight comes on, and I can type `sudo reboot` into a terminal that I left focused before sleeping, and the laptop shuts down and reboots, with the display working after reboot.
Attached is messages.log of the sleep (first line), and the resume (line 35 or so). Please let me know what additonal information is needed.
fwiw, I see same issue on a MBP-5,4 (ubuntu 11.04)
fwiw, this issue is still present in kernel 3.3.1 (Arch Linux) with xf86-video-nouveau 0.0.16_git20120210.
Is this still an issue with the latest software?
I have the same issue on MacBook Pro (mid 2009 - NV96 + NVAC), running on 22.214.171.124-git5.1. However laptop is not responsive after wakening: fans don't go on, no keyboard backlight... It seems the computer is still sleeping; I'll try to ssh it and post the results here.
Created attachment 89730 [details]
126.96.36.199-git10.1 kernel log / suspend + resume
Adding 188.8.131.52-git10.1 kernel log, with suspend / resume / suspend /resume as a sequence.
Kernel options were: nouveau.noaccel=1 nouveau.modeset=0 3
Results: computer is up and running, interactive but screen does not light up
I tried with only nouveau.noaccel=1 but boot hangs up after initialising the nv96 card; it was working on kernel 3.11.x (though the screen was scrambled), so I'll fill a bug report for this regression and see if I can bisect it. Adding nouveau.config=NvMSI=0 changes nothing.
*** Bug 38350 has been marked as a duplicate of this bug. ***
I've been playing a bit with the apple-gmux driver to check if it was the one responsible but without success; the only difference between before suspending and after resuming being GMUX_PORT_SWITCH_GET_DISPLAY which used to be 24 before suspending, and equals to 0 when resuming (and the discrete card coming back online, but that was already dealt with). So, without Nouveau loaded, the screen will stay black.
However, if using patches from https://bugs.freedesktop.org/show_bug.cgi?id=27501#c29, the screen resumes correctly.
A friend tested on his MacBook Pro (Mid 2010) with an MCP89, running Linux 4.1.6 and it worked. Closing as fixed.