2.6.22-rc2 amd64 slab allocator bug reported at boot time

2.6.22-rc2 amd64 slab allocator bug reported at boot time

[Andrew Paprocki]

Snippet from dmesg output (the full dmesg output was attached to my
previous e-mail right before this one regarding 2.6.22-rc2 amd64+4gb
ram agpgart bug):

PCI: Using ACPI for IRQ routing
PCI: If a device doesn't work, try "pci=routeirq".  If it helps, post a report
NET: Registered protocol family 8
NET: Registered protocol family 20
BUG: at /usr/src/linux-2.6.22-rc2/mm/slab.c:777 __find_general_cachep()

Call Trace:
 [<ffffffff802743dc>] __kmalloc+0x3e/0xbe
 [<ffffffff8021d65f>] cache_k8_northbridges+0x7f/0xf0
 [<ffffffff8052bb53>] gart_iommu_init+0x13/0x4f8
 [<ffffffff802d15e1>] kobject_shadow_add+0xef/0x1a2
 [<ffffffff802d13fb>] kobject_get+0x12/0x17
 [<ffffffff8033b2a4>] class_register+0x10a/0x13d
 [<ffffffff80527964>] pci_iommu_init+0x9/0x12
 [<ffffffff8052060e>] kernel_init+0x167/0x2d7
 [<ffffffff8020aba8>] child_rip+0xa/0x12
 [<ffffffff802fb7e4>] acpi_ds_init_one_object+0x0/0x7c
 [<ffffffff805204a7>] kernel_init+0x0/0x2d7
 [<ffffffff8020ab9e>] child_rip+0x0/0x12

PCI-GART: No AMD northbridge found.
hpet0: at MMIO 0xfed00000, IRQs 2, 8, 0
hpet0: 3 64-bit timers, 14318180 Hz

-Andrew

Re: 2.6.22-rc2 amd64 slab allocator bug reported at boot time

[Stefan Richter]

Andrew Paprocki wrote:
...
> BUG: at /usr/src/linux-2.6.22-rc2/mm/slab.c:777 __find_general_cachep()
> 
> Call Trace:
>  [<ffffffff802743dc>] __kmalloc+0x3e/0xbe
>  [<ffffffff8021d65f>] cache_k8_northbridges+0x7f/0xf0
>  [<ffffffff8052bb53>] gart_iommu_init+0x13/0x4f8
>  [<ffffffff802d15e1>] kobject_shadow_add+0xef/0x1a2
>  [<ffffffff802d13fb>] kobject_get+0x12/0x17
>  [<ffffffff8033b2a4>] class_register+0x10a/0x13d
>  [<ffffffff80527964>] pci_iommu_init+0x9/0x12
...

This is a known "zero allocation" problem, patch available.
http://kernelnewbies.org/known_regressions#line-249
-- 
Stefan Richter
-=====-=-=== -=-= ==---
http://arcgraph.de/sr/

Re: 2.6.22-rc2 amd64 slab allocator bug reported at boot time

[Andi Kleen]

Stefan Richter <stefanr@s5r6.in-berlin.de> writes:

> Andrew Paprocki wrote:
> ...
> > BUG: at /usr/src/linux-2.6.22-rc2/mm/slab.c:777 __find_general_cachep()
> > 
> > Call Trace:
> >  [<ffffffff802743dc>] __kmalloc+0x3e/0xbe
> >  [<ffffffff8021d65f>] cache_k8_northbridges+0x7f/0xf0
> >  [<ffffffff8052bb53>] gart_iommu_init+0x13/0x4f8
> >  [<ffffffff802d15e1>] kobject_shadow_add+0xef/0x1a2
> >  [<ffffffff802d13fb>] kobject_get+0x12/0x17
> >  [<ffffffff8033b2a4>] class_register+0x10a/0x13d
> >  [<ffffffff80527964>] pci_iommu_init+0x9/0x12
> ...
> 
> This is a known "zero allocation" problem, patch available.
> http://kernelnewbies.org/known_regressions#line-249

It should be already fixed in mainline

-Andi