I could reproduce this with 2.2.18pre17 + Andrea's VM-global patch (-7),
pre20 with and without the VM patch.
Mainboard is a FIC VA503+, video is a G200.
Better run this with everything mounted RO:
$ insmod agpgart
$ rmmod agpgart
$ insmod agpgart
$ insmod mga
And here we go...
I tested a combinations, and what appears trigger the bug is the unloading
of agpgart and the subsequent loading of mga.
[Data gathered with vanilla pre20 - config file attached]
While in single-user mode, all FS mounted RO.
2 OOPSes at same EIP (locks_remove_flock+0x0E) for init are on screen,
similar ones for bash occured just before, although the 1st one for bash is
partly offscreen and I can't scrollback.
OOPSes per process are:
* paging request at e850097c (EFLAGS=00010286)
* NULL pointer dereference at 00...0008 (EFLAGS=00010292 in this case)
Not sure I remember the correct order, though.
A couple of "hda lost interrupt" also appear from time to time. All of this
(OOPS+lost intr) do not appear at once, but usually all of them come after
waiting some time. At the very least I only had bash not giving me a prompt
any more after hitting RETURN.
Anyone seing this ?
--
Yann Dirson <[email protected]> | Why make M$-Bill richer & richer ?
debian-email: <[email protected]> | Support Debian GNU/Linux:
| Cheaper, more Powerful, more Stable !
http://ydirson.free.fr/ | Check <http://www.debian.org/>