[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Xen-users] [resend] getting X to work, GPU lockups on dom0



(resending without attachments, in hope of getting past the spam filter)
Hi all,

Debian sid system.  NV34M graphics card (GeForce FX Go5200 64M).  I'm
trying to get started using Xen on my laptop.  CPU is a Pentium 4 with
no vmx, so this is all PV.  I install the kernel and hypervisor, boot
into the Xen dom0, and try to start X so I can launch a web browser
and see what to do next.

Unfortunately X fails to start.  The screen turns black and stays that
way, though I can blindly run a few commands.  Xorg.0.log has a few
hundred

        (EE) FBDEV(0): FBIOPUTCMAP: Invalid argument

lines.

Looking more closely, I see that the actual trouble happened a little
earlier: dmesg informs me:

| [    8.703713] [drm] nouveau 0000:01:00.0: PFIFO_DMA_PUSHER - Ch 0 Get 
0x10000000 Put 0x0000fdc0 State 0xc0000000 (err: MEM_FAULT) Push 0x00000000
| [    8.704276] Console: switching to colour frame buffer device 175x65
| [   10.553368] [drm] nouveau 0000:01:00.0: PFIFO_DMA_PUSHER - Ch 0 Get 
0x10000000 Put 0x00010048 State 0xc0000000 (err: MEM_FAULT) Push 0x00000000
| [   12.402789] [drm] nouveau 0000:01:00.0: PFIFO_DMA_PUSHER - Ch 0 Get 
0x10000000 Put 0x000101c8 State 0xc0000000 (err: MEM_FAULT) Push 0x00000000
| [   14.252200] [drm] nouveau 0000:01:00.0: PFIFO_DMA_PUSHER - Ch 0 Get 
0x10000000 Put 0x000103cc State 0xc0000000 (err: MEM_FAULT) Push 0x00000000
| [   14.260058] [drm] nouveau 0000:01:00.0: GPU lockup - switching to software 
fbcon

By contrast, when I boot the same kernel on bare hardware I don't get
the PFIFO_DMA_PUSHER or GPU lockup messages and X works fine.

Hypervisor is xen-hypervisor-4.1-i386 from Debian, version 4.1.2-2.
Reproducible with linux-image-3.2.0-1-686-pae version 3.2.4-1 and with
linus's master (8597559a78e1, 2012-02-06) and linux-next
(e9e792b45a9b, 2012-02-06).

dmesg from running the native kernel (good) and dom0 (bad) attached.

I also tried blacklisting the nouveau kernel module and starting X
with the vesa driver.  That also didn't work.

Is this behavior expected?  Does starting X in the dom0 work on other
cards?

Curious,
Jonathan

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.