Bug 212262 - x11/nvidia-driver: after update to 367.35, resume (after suspend-to-RAM) fails to re-initialize display
Summary: x11/nvidia-driver: after update to 367.35, resume (after suspend-to-RAM) fail...
Status: Closed FIXED
Alias: None
Product: Ports & Packages
Classification: Unclassified
Component: Individual Port(s) (show other bugs)
Version: Latest
Hardware: amd64 Any
: --- Affects Only Me
Assignee: Alexey Dokuchaev
Depends on:
Reported: 2016-08-30 16:14 UTC by david
Modified: 2019-10-21 20:04 UTC (History)
3 users (show)

See Also:
bugzilla: maintainer-feedback? (danfe)


Note You need to log in before you can comment on or make changes to this bug.
Description david 2016-08-30 16:14:55 UTC
Prior to 367.35, I had (most recently) been running 346.96 -- and resume worked.  This was under stable/10, stable/11, and head (all amd64 -- all the same physical hardware; different slices of boot drive).

I realize that not everyone needs suspend/resume to work, but the machine in question is my laptop, and I carry it in my backpack as I cycle between home and the shuttle stop (as part of my commute to & from work).  Having the laptop "on" and cooking itself in the backpack isn't a reasonable option, and powering it off is ... well, a regression, at best.

Might it be worthwhile to create an x11/nvidia-driver-346, pending a fix for the newer version of the driver?

(I have verbose dmesg.boot for the above 3 versions of the OS available, as well as a fair amount of other info, should that be wanted.  And I'm happy to test.)
Comment 1 david 2016-08-30 23:07:55 UTC
I just tried downgrading to x11/nvidia-driver-340; that seems to work (including suspend/resume), at least under stable/11.
Comment 2 Stefan 2016-08-31 17:42:18 UTC
Just FYI, removing driver "device VESA" (included in GENERIC) from the kernel config, switching to sc console and building + installing the new kernel solved the problem for me.
Comment 3 david 2016-08-31 20:04:21 UTC
A few weeks ago -- when I started running stable/11 for more than merely smoke-testing, and I thus needed to be able to suspend & resume -- I appended the line:


to /boot/loader.conf, and that got suspend/resume working for me (both for stable/11 and head; stable/10 didn't need it).  This was while running x11/nvidia-driver-346.96.

I have (of course) left that specification in while I attempted to use x11/nvidia-driver-367.35.

So while it's good to note that switching from vt to sc while running nvidia-driver-367.35 works for some, it does not work for all... even in configurations where nvidia-driver-346.96 worked (and -340.96 works).
Comment 4 Walter Schwarzenfeld freebsd_triage 2018-01-14 00:05:17 UTC
Maintainer feedback? (Resp. is this still relevant?),
Comment 5 Erik Kurzinger 2019-04-18 23:19:07 UTC
I think https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=237050 should resolve this, if anyone wants to double check.
Comment 6 david 2019-10-21 12:01:04 UTC
Thank you; I was able to go back to x11/nvidia-driver (nvidia-driver-390.87_3, as of this writing) yesterday, and suspend/resume works fine (in stable/11, stable/12, and head).  I did need to (also) load nvidia-modeset.ko (as expected).

I believe that this bug report may be closed.