Messages in this thread | | | Date | Wed, 9 Sep 2009 02:38:47 -0400 | Subject | Xorg server hanging/crashing kernel, esp. 2.6.31-rc9 (-rc8) | From | Robert Bradbury <> |
| |
I've got a situation [1] where I am trying to configure the Xorg server (xorg.conf) to control multiple graphics cards with multiple monitors. I've got a specific case where the Xorg server appears to be segfaulting when it is supposed to enter its "OPERATING" state and in the process seems to take down the consoles and perhaps even the kernel itself.
(I consider a kernel in a multi-user state with logged in telnet sessions to be "up" while a kernel with unresponsive telnet sessions, no response to ping's and no console VT switching capabilities to be "hung" (it is reasonably useless)).
Now I would normally consider that Xorg crashing should not be able to hang the kernel and it normally doesn't (returning to the console state, leaving a stack trace in the Xorg.#.log file, etc.). But I've got clear cases in 2.6.31 where that doesn't appear to be the case. The problem may extend back to ~2.6.28 or so when the KMS / UXA / EXA transition started, maybe even earlier.
The details of the bug and associated files are discussed in the Gentoo and Freedesktop bugzillas: http://bugs.gentoo.org/show_bug.cgi?id=284223 https://bugs.freedesktop.org/show_bug.cgi?id=23804
Any help, esp. why is there doesn't appear to be a Linux Kernel bugzilla and whether there is any documentation on how to debug a "hung" kernel when the consoles & remote terminal access doesn't work would be appreciated.
1. The situation involves trying to make a 2-PCI device (1 motherboard Intel 915 + 1 PCIe R3450 (Radeon 620) graphics card) drive 3 monitors.
| |