Summary: | Modesetting on Intel by default crashes on headless servers | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Audiodef <online> |
Component: | [OLD] Core system | Assignee: | Gentoo X packagers <x11> |
Status: | RESOLVED NEEDINFO | ||
Severity: | critical | CC: | galtgendo |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | x86 | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- |
Description
Audiodef
2010-03-02 16:31:28 UTC
IIRC, in a near future, intel driver is planned to become KMS-only. Any chance that your problem is something like the one xf86-video-ati people warn about, that is that you shouldn't use any other framebuffer driver, besides fbcon (including uvesafb, vga bot parameters, etc) See http://wiki.x.org/wiki/radeonBuildHowTo Probably there's a similar site for intel, but I don't have such card. Based on how many problems I and others have had with KMS, that would be a very bad idea. I know I've done everything "right" with KMS, including the things you've mentioned, and still have problems with it. I don't know who's idea it is to make KMS non-optional, but it would scare me off using Intel video to the point of getting ANY video card to replace any built-in Intel chip I have. So if this is Intel's idea, they're scaring off customers. If it's someone else's idea, they're scaring off Intel customers. Just. Not. A. Good. Idea. Better to leave it as an option. Gentoo is all about options and forcing users to accept a certain flavour is very un-Gentoo. My bad, you're out of luck - it has already happened: read http://www.mail-archive.com/xorg@lists.freedesktop.org/msg09895.html No problem - I just won't upgrade to that kernel on this machine! ;-) I think you've misunderstood something - 2.6.33 kernel seems to be required for the driver to function correctly, not the other way around. Also, to remind you about the fun ati-driver (the proprietary blob) had with upgrade to 1.7 xorg-server - eventually it was either new xorg-server or the old driver. Many have switched to xf86-video-ati then (especially people with older cards). Eventually, you'd be forced to upgrade too, so it'd better to figure out, what the real problem is. If you don't use X on that server, don't enable DRM in your kernel. That's possibly the easiest work-around. If you do want to use X, then you'll *have* to enable KMS. I think the KMS kernel option will be removed at some point, since enabling KMS is now mandatory with X drivers 2.10 and up. In any case, I'd recommend filing a bug upstream [1] to really fix the bug, it's in everyone's best interest. Please paste the URL here when you do. Thanks [1] http://intellinuxgraphics.org/how_to_report_bug.html Thanks for the tip on DRM - I appreciate it! I'll try to get around to an upstream bug report. |