I'm sending this one more time -- I didn't get a response last time. If I can
provide any more information, please let me know.
Hey folks,
I've been dealing with a strange X-locking problem for quite some time. I've
been trying to gather information, and I think I've gathered about as much as
I can. The problem only occurs when using Xinerama.
SYMPTOM:
If I go to certain web pages in firefox or mozilla, and then click on a link,
it will cause X to go unresponsive. The x server goes into a loop of:
select(256, [1 3 4 5 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
27 28 29 30 31], NULL, NULL, {103, 371000}) =3D 1 (in [26], left {103,
356000})
read(26, "\220\3\n\0\3\0@\2\5\0@\2\0\6@\0\0\2\0\0008\0008\0\4\0\4"..., 4096)
= 80
read(26, 0x856ecc0, 4096) =3D -1 EAGAIN (Resource temporarily unavailable)
gettimeofday({1094757921, 253529}, NULL) = 0
select(256, [1 3 4 5 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
27 28 29 30 31], NULL, NULL, {103, 356000}) = ? ERESTARTNOHAND (To be
restarted)
Firefox or mozilla are just sitting in a poll.
At this point my mouse can move, but no windows accept input. Focus does not
change, typing doesn't work, numlock/capslock won't work, ctrl-alt-backspace
won't work, etc.
SPECIFIC STEPS TO REPRODUCE (on my machine):
1. go to www.nmi-edit.org
2. Under "Proceedings Available for CAMP and Advanced CAMP" click "CAMP"
3. Click "register"
At this point X is now in the state described above.
DEBUG INFORMATION:
1. This ONLY happens if Xinerama is on. I have 3 monitors powered by 2 cards,
and if I turn Xinerama off, it doesn't happen. I've tried leaving Xinerama on
and using a few other combinations, and it's consistently reproducible.
2. I have an onboard AGP RV100 QY [Radeon 7000/VE] dual-headed card as well as
a PCI RV100 QY [Radeon 7000/VE] card.
3. I'm running a version of RHE3 built from source RPMs.
4. Problem happens in 2.4 or 2.6 kernels. Happens with noacpi or with acpi.
5. I have attached strace's of both X and firefox here:
http://sol.usc.edu/~phil/X.strace.gz
http://sol.usc.edu/~phil/firefox.strace.gz
6. Here's detailed lspci for both cards:
01:00.0 VGA compatible controller: ATI Technologies Inc Radeon RV100 QY
[Radeon 7000/VE] (prog-if 00 [VGA])
Subsystem: ATI Technologies Inc: Unknown device 1b8a
Flags: bus master, stepping, 66Mhz, medium devsel, latency 64, IRQ = 11
Memory at f0000000 (32-bit, prefetchable) [size=3D128M]
I/O ports at ec00 [size=3D256]
Memory at ff8f0000 (32-bit, non-prefetchable) [size=3D64K]
Expansion ROM at <unassigned> [disabled] [size=3D128K]
Capabilities: [58] AGP version 2.0
Capabilities: [50] Power Management version 2
02:08.0 VGA compatible controller: ATI Technologies Inc Radeon RV100 QY
[Radeon 7000/VE] (prog-if 00 [VGA])
Subsystem: PC Partner Limited: Unknown device 7c02
Flags: bus master, stepping, medium devsel, latency 64, IRQ 11
Memory at e8000000 (32-bit, prefetchable) [size=3D128M]
I/O ports at d800 [size=3D256]
Memory at ff6f0000 (32-bit, non-prefetchable) [size=3D64K]
Expansion ROM at 80000000 [disabled] [size=3D128K]
Capabilities: [50] Power Management version 2
7. X Version: 4.3.0 (actually XFree86-Xvfb-4.3.0-68.EL from RH)
--
Phil Dibowitz
Systems Architect and Administrator
Enterprise Infrastructure / ISD / USC
UCC 174 - 213-821-5427
Attachment:
pgp1z3Vub3VoY.pgp
Description: PGP signature
