2.6.8-rc4 on sparc64 oopses at boot and no working XFree86

2.6.8-rc4 on sparc64 oopses at boot and no working XFree86

[Admar Schoonen]

Hi

I'm running 2.6.8-rc4 on a sun blade 100 (ultra sparc IIe) with 2 PGX64
(Ati mach64) framebuffers, running Debian unstable. When the kernel boots,
it looks like it produces an oops in the very beginning. This is not
visible on the console, but dmesg shows some traces:

  0053c3d0] class_device_add+0x50/0x140
   [0000000000514ce8] pci_scan_bus_parented+0xe8/0x180
   (...)
  bad: scheduling while atomic!
  Call Trace:
    [000000000045bc4c] call_usermodehelper+0xac/0xc0
    (...)
    (this goes on for over 200 lines and continues with:)
  PCI-IRQ: Routing bus[ 0] slot[ 8] map[0] to INO[23]
    (...)
  PCI-IRQ: Routing bus[ 1] slot[ 1] map[0] to INO[0a]
  PCI0(PBMA): Bus running at 33MHz
  isa0: [dma -> (floppy) (parallel)] [power] [serial] [serial]
  ebus0: [flashprom] [eeprom] [idprom]
  
This also happens with 2.6.8-rc1, but it doesn't with 2.6.7 and older.
On 2.6.6, it looks like this:

  NET: Registered protocol family 16
  PCI: Probing for controllers.
  PCI: Found SABRE, main regs at 000001fe00000000, wsync at 000001fe00001c20
  SABRE: Shared PCI config space at 000001fe01000000
  SABRE: DVMA at c0000000 [20000000]
  PCI-IRQ: Routing bus[ 0] slot[ 8] map[0] to INO[23]
    (...)
  PCI-IRQ: Routing bus[ 1] slot[ 1] map[0] to INO[0a]
  PCI0(PBMA): Bus running at 33MHz
  isa0: [dma -> (floppy) (parallel)] [power] [serial] [serial]
  ebus0: [flashprom] [eeprom] [idprom]
  
so perhaps something is changed in the pci / sabre code?

Furthermore, with 2.4.x, I could use both framebuffers on the console
and in X. With 2.6.x (x <= 7), I could only use the primary framebuffer;
secondary was always black on console and X. But with 2.6.8-rc1 and up,
I can use the secondary framebuffer (con2fbmap works), but X (configured
for only my primary framebuffer) gets a signal 11 before it can display
it's famouse black & white dots. X -configure also gets a signal 11.

I've put up my .config, dmesg log's and XFree86.0.log at
http://people.spacelabs.nl/~admar/2.6.8-rc4-sparc64/

Could anybody pleasse help me out with these problems?

Pleasse CC any replies to me, as I'm not subscribed to this list.

Regards

Admar Schoonen

Re: 2.6.8-rc4 on sparc64 oopses at boot and no working XFree86

[David S. Miller]

Turn off CONFIG_PREEMPT if you want that OOPS message to go
away, it happens to be harmless though.

I have no idea wrt. your xfree86 problems.

Why not post this on the correct list, sparclinux@vger.kernel.org,
btw?

Re: 2.6.8-rc4 on sparc64 oopses at boot and no working XFree86

[Admar Schoonen]

(sorry for the cross posting; I had the wrong list first)

On Fri, Aug 13, 2004 at 09:03:09AM -0700, David S. Miller wrote:
> I have no idea wrt. your xfree86 problems.

Well, I guess it's related to the fact that I have 2 Ati mach64
framebuffers (2 x PGX64) in the machine. On that machine XFree86
segfaults with 2.6.8-rc4 segfaults, while it works fine with 2.6.7.

I've just tried 2.6.8-rc4 on an ultra 5 with just 1 Ati mach64
framebuffer (PGX24) and XFree86 works fine there too, so I guess it's
due to the secondary framebuffer.

I'll ask the linux-fbdev guys about this issue.
 
> Why not post this on the correct list, sparclinux@vger.kernel.org,
> btw?

Sorry, I didn't know it existed. I should have looked better of course.

Admar

(Please CC any replies to me, as I'm not subscribed to this list.)