Re: Kernel BUG at include/linux/gfp.h:80

Re: Kernel BUG at include/linux/gfp.h:80

[Pekka J Enberg]

Hi Carsten,

On 1/19/06, Carsten Otto <c-otto@gmx.de> wrote:
> As indicated the sounddriver seems to have some sort of problem. I
> disabled it and now I can boot (without problems).

Does the following patch fix your problem?

			Pekka

Index: 2.6/lib/swiotlb.c
===================================================================
--- 2.6.orig/lib/swiotlb.c
+++ 2.6/lib/swiotlb.c
@@ -444,7 +444,8 @@ swiotlb_alloc_coherent(struct device *hw
 	 * instead, or use ZONE_DMA32 (ia64 overloads ZONE_DMA to be a ~32
 	 * bit range instead of a 16MB one).
 	 */
-	flags |= GFP_DMA;
+	if (!(flags & GFP_DMA32))
+		flags |= GFP_DMA;
 
 	ret = (void *)__get_free_pages(flags, order);
 	if (ret && address_needs_mapping(hwdev, virt_to_phys(ret))) {

Re: Kernel BUG at include/linux/gfp.h:80

[Carsten Otto]

[-- Attachment #1: Type: text/plain, Size: 549 bytes --]

On Fri, Jan 20, 2006 at 12:16:22PM +0200, Pekka J Enberg wrote:
> Does the following patch fix your problem?

Yes, it does (I moved to 2.6.15.4 in the same step, though).

Unfortunately my PC freezes as soon as I start a 3D game (Enemy
Territory True Combat Elite) when I see the first few 3D ingame frames.
Running glxgears is fine, as is running multiple instances of VLC to
output several audio streams through one device.
Switching off the sound in ET:TCE does not help at all.

Bye,
-- 
Carsten Otto
c-otto@gmx.de
www.c-otto.de

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Kernel BUG at include/linux/gfp.h:80

[Andrew Morton]

Carsten Otto <c-otto@gmx.de> wrote:
>
> Unfortunately my PC freezes as soon as I start a 3D game (Enemy
>  Territory True Combat Elite) when I see the first few 3D ingame frames.
>  Running glxgears is fine, as is running multiple instances of VLC to
>  output several audio streams through one device.
>  Switching off the sound in ET:TCE does not help at all.

Is that new behaviour?  If so, which was the most recent kernel which
worked OK?

What sort of video card are you using?

Thanks.

Re: Kernel BUG at include/linux/gfp.h:80

[Carsten Otto]

[-- Attachment #1: Type: text/plain, Size: 397 bytes --]

On Mon, Feb 13, 2006 at 08:04:29PM -0800, Andrew Morton wrote:
> Is that new behaviour?

Yes.

> If so, which was the most recent kernel which
> worked OK?

2.6.15.1 (but without this patch and therefore without sound from the
emu10k1).

> What sort of video card are you using?

Geforce 4 4600 (AGP, NVIDIA). Driver: 8178

Thanks,
-- 
Carsten Otto
c-otto@gmx.de
www.c-otto.de

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Kernel BUG at include/linux/gfp.h:80

[Pekka J Enberg]

On Fri, Jan 20, 2006 at 12:16:22PM +0200, Pekka J Enberg wrote:
> > Does the following patch fix your problem?
 
On Mon, 13 Feb 2006, Carsten Otto wrote:
> Yes, it does (I moved to 2.6.15.4 in the same step, though).

This is also fixed in current git (bit differently by Andi):

http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=6bca52b544489b626c7d0db801df6b4aa3d5adb5

				Pekka

Kernel BUG at include/linux/gfp.h:80

[Carsten Otto]

[-- Attachment #1: Type: text/plain, Size: 712 bytes --]

Hi there!

My kernel panics at every boot, see screenshot[1].

I upgraded my CPU to an Opteron 175 (Dual Core) and enabled SMP in the
kernel. With "nosmp" the kernel boots (but has problems finding one hard
disk later on, some IRQ timeout?). The kernel version is 2.6.15.1,
config see here[2].

With my current non-SMP-kernel (2.6.14.4) everything works.

Please tell me what that BUG message means and how I can get my system
running.

Addition information regarding the hardware can be found here[3].

Thanks a lot,
Carsten

[1]: http://c-otto.de/panic.jpg
[2]: http://c-otto.de/config.txt
[3]: http://c-otto.de/index.php?x=hardware#teile
-- 
Carsten Otto
c-otto@gmx.de
www.c-otto.de

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Kernel BUG at include/linux/gfp.h:80

[Pekka Enberg]

Hi Carsten,

On 1/19/06, Carsten Otto <c-otto@gmx.de> wrote:
> My kernel panics at every boot, see screenshot[1].
>
> I upgraded my CPU to an Opteron 175 (Dual Core) and enabled SMP in the
> kernel. With "nosmp" the kernel boots (but has problems finding one hard
> disk later on, some IRQ timeout?). The kernel version is 2.6.15.1,
> config see here[2].
>
> With my current non-SMP-kernel (2.6.14.4) everything works.
>
> Please tell me what that BUG message means and how I can get my system
> running.
>
> Addition information regarding the hardware can be found here[3].
>
> Thanks a lot,
> Carsten
>
> [1]: http://c-otto.de/panic.jpg
> [2]: http://c-otto.de/config.txt
> [3]: http://c-otto.de/index.php?x=hardware#teile

I think the problem is related to the following changeset where we
potentially allocate with GFP_DMA and GFP_DMA32 set at the same time.
Andi?

http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=47492d3667ec519172ab978bd8231b8c7152fa9d

                                   Pekka

Re: Kernel BUG at include/linux/gfp.h:80

[Carsten Otto]

[-- Attachment #1: Type: text/plain, Size: 2406 bytes --]

On Thu, Jan 19, 2006 at 05:10:33PM +0100, Carsten Otto wrote:
> My kernel panics at every boot, see screenshot[1].

As indicated the sounddriver seems to have some sort of problem. I
disabled it and now I can boot (without problems).

modprobe snd_emu10k1 gives:

----------- [cut here ] --------- [please bite here ] ---------
Kernel BUG at include/linux/gfp.h:80
invalid operand: 0000 [1] SMP 
CPU 1 
Modules linked in: snd_emu10k1 snd_rawmidi snd_ac97_codec snd_ac97_bus snd_util_mem snd_hwdep nvidia
Pid: 11886, comm: modprobe Tainted: P      2.6.15.1 #5
RIP: 0010:[<ffffffff8015185e>] <ffffffff8015185e>{__get_free_pages+16}
RSP: 0018:ffff8100b5619d08  EFLAGS: 00010202
RAX: 0000000000000005 RBX: 0000000000000003 RCX: 00000000000052d4
RDX: ffff8100b802e090 RSI: 0000000000000003 RDI: 00000000000052d5
RBP: 0000000000008000 R08: 0000000000000000 R09: 0000000300000001
R10: 0000000300000001 R11: 0000000000000001 R12: ffff81012ff14870
R13: ffff8100b802e090 R14: 0000000000008000 R15: ffff8100b802e090
FS:  0000000000859d80(0000) GS:ffffffff806ae880(0000) knlGS:000000000804aba0
CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 00002aaaaac4cad0 CR3: 00000000b5596000 CR4: 00000000000006e0
Process modprobe (pid: 11886, threadinfo ffff8100b5618000, task ffff8100b6cc0a40)
Stack: 00000000000052d4 ffffffff802ad905 0000000000000000 000000007fffffff 
       ffff81012ff14870 ffffffff80000000 ffff8100b802e078 ffffffff803e160b 
       ffffffff884d924e 000000038014cb00 
Call Trace:<ffffffff802ad905>{swiotlb_alloc_coherent+47} <ffffffff803e160b>{snd_dma_alloc_pages+327}
       <ffffffff884cac0d>{:snd_emu10k1:snd_emu10k1_create+1028}
       <ffffffff803cff4d>{snd_info_register+63} <ffffffff884ca0e2>{:snd_emu10k1:snd_card_emu10k1_probe+226}
       <ffffffff80429870>{udp_poll+0} <ffffffff802b1e41>{pci_device_probe+86}
       <ffffffff803291cc>{driver_probe_device+65} <ffffffff80329288>{__driver_attach+0}
       <ffffffff803292be>{__driver_attach+54} <ffffffff80328906>{bus_for_each_dev+67}
       <ffffffff80328d46>{bus_add_driver+116} <ffffffff802b1cb8>{__pci_register_driver+143}
       <ffffffff80149027>{sys_init_module+246} <ffffffff8010d84a>{system_call+126}
       

Code: 0f 0b 68 c7 ab 4b 80 c2 50 00 48 63 d0 48 6b d2 28 48 81 c2 
RIP <ffffffff8015185e>{__get_free_pages+16} RSP <ffff8100b5619d08>
-- 
Carsten Otto
c-otto@gmx.de
www.c-otto.de

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]