Bug 279542 – x11-drivers/nvidia-drivers-185.18.29 version bump. official release, not beta.

• Status: RESOLVED FIXED
• Product: Gentoo Linux

[Description Zeev Tarantov 2009-07-29 00:57:38 UTC]

From http://www.nvnews.net/vbulletin/showthread.php?t=122606:
Current official release: 185.18.29
Changelog indicates most changes are to cuda and vdpau bugfixes.
Not tested it yet myself.

Reproducible: Always

[Comment 1 James Earl Spahlinger 2009-07-29 01:20:20 UTC]

re-assigned to first developer in list of 4. Others are CC'd, feel free to modify further :).

[Comment 2 Zeev Tarantov 2009-07-29 01:34:42 UTC]

Just copied 185.18.14 ebuild from portage, it compiled, rebooted and played openarena fine on ~amd64 system with vanilla-sources-2.6.31-rc4.

[Comment 3 Michal Januszewski (RETIRED) 2009-07-29 05:12:14 UTC]

Added to CVS, thanks for reporting the new version.

[Comment 4 leeight 2009-07-30 01:54:52 UTC]

after upgrade to 185.18.29, i can't start my x server, so i mask it, and downgrade to 185.29.14, everything works find.

01:00.0 VGA compatible controller: nVidia Corporation Quadro NVS 140M (rev a1)
x11-base/xorg-server-1.6.2.901

[Comment 5 Zeev Tarantov 2009-07-30 05:58:29 UTC]

http://www.nvnews.net/vbulletin/showthread.php?t=136564

Says the newest version is broken on an Quadro NVS 140M.

That forum is the place to watch for official response from nVidia.

Can the ebuild be conditional on PCI id of the card? Or at least explicitly code an error on detecting the card? If not, at least add text that displays for everyone that says it doesn't work with those GPUs.

[Comment 6 Emond Papegaaij 2009-07-31 06:56:11 UTC]

I can confirm the problems with 185.18.29 on a Quadro NVS 140M. Starting X causes a kernel OOPS in the nvidia driver. Reverting to .14 fixed it.

[Comment 7 Michal Pytasz 2009-07-31 09:04:03 UTC]

Confirm problem starting X on HP Pavilion tx2010eo (GF-6150Go). Xorg.log ends with (no further lines):

(II) NVIDIA(0): ACPI display change hotkey events enabled: the X server is new
(II) NVIDIA(0):     enough to receive ACPI display change hotkey events.

blank screen, unable to switch terminals. Reproducible for xorg-server 1.6.2.901 and 1.6.2-r1, vanilla-sources 2.6.30.3, amd64 and x86.

Revert to 185.18.14 fixes the problem

[Comment 8 Dirk Heinrichs 2009-07-31 09:59:21 UTC]

(In reply to comment #6)
> I can confirm the problems with 185.18.29 on a Quadro NVS 140M. Starting X
> causes a kernel OOPS in the nvidia driver. Reverting to .14 fixed it.

/me too. So can .29 please be masked.

[Comment 9 Juergen Rose 2009-07-31 10:01:09 UTC]

Same problem here with Quadro NVS 140M (rev a1), see Bug  279588

[Comment 10 Dirk Heinrichs 2009-07-31 11:06:25 UTC]

BTW: This thread indicates that the problem may also occur with 190.16: http://www.nvnews.net/vbulletin/showthread.php?t=136143

[Comment 11 michel 2009-07-31 16:26:48 UTC]

oficial 185.18.31 is out now

[Comment 12 leeight 2009-08-03 01:49:18 UTC]

same problem with Quadro NVS 140M :(
(In reply to comment #11)
> oficial 185.18.31 is out now
> 

[Comment 13 Michal Januszewski (RETIRED) 2009-08-03 03:59:41 UTC]

I just added a device compatibility check for Quadro NVS 140M to .29 and .31.  If you try to emerge the ebuilds, you will be warned about the incompatibility.  Hope that helps.

[Comment 14 Zeev Tarantov 2009-08-03 04:02:27 UTC]

re: comment #12
Did you post an nvidia-bug-report.log on their forums like they ask in:
http://www.nvnews.net/vbulletin/showthread.php?t=46678 ?

This bad QA on the stable branch is surprising. They should at least be able to not cause regressions that are easy to reproduce.

re: comment #13
Thanks a lot!

[Comment 15 Doug Goldstein (RETIRED) 2009-08-03 04:19:19 UTC]

the Quadro NVS 140 issue is being tracked on bug #279588

When I get further info from NVIDIA I'll attach it there.