I can't think of any cause for this. I've seen another report of
such a thing under Xinerama (or maybe it was the same report), but was
unable to reproduce. I spend hours in front of my Xinerama system
every day and it runs for weeks straight, being shut down only when
I upgrade XFree86. Unless I can repro, there's nothing for me to
go on, and I know of no outstanding Xinerama bugs.
Generally, when someone has a server problem when web browsing,
I tell them to comment out the "freetype" module in the XF86Config.
Historically, there have been problems with that module messing
up the font metrics. But that would happen with or without Xinerama,
though maybe it does but you only notice it in Xinerama because
the memory that gets clobbered only gets used in Xinerama.
Mark.
On Mon, 20 Sep 2004, Phil Dibowitz wrote:
> 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
>
>
_______________________________________________
XFree86 mailing list
XFree86@xxxxxxxxxxx
http://XFree86.Org/mailman/listinfo/xfree86
