Hi,
During testing of the most recent version of this patch we've hit a kernel bug
in alloc_vmap_area():
}
412 BUG_ON(first->va_start < addr); <- Bug triggers here
413 if (addr + cached_hole_size < first->va_start)
414 cached_hole_size = first->va_start - addr;
415 }
This appears to be caused by a call from:
[<c050f870>] ? kmem_cache_alloc_notrace+0xa0/0xb0
[<c0503695>] ? __get_vm_area_node+0xc5/0x1c0
[<c0458625>] ? walk_system_ram_range+0xa5/0x1c0
[<c050383e>] ? get_vm_area_caller+0x4e/0x60
[<f827328c>] ? intel_opregion_init+0xbc/0x510 [i915]
[<c0430ce0>] ? __ioremap_caller+0x2b0/0x420
[<f827328c>] ? intel_opregion_init+0xbc/0x510 [i915]
[<c0430f98>] ? ioremap_nocache+0x18/0x20
[<f827328c>] ? intel_opregion_init+0xbc/0x510 [i915]
[<f827328c>] ? intel_opregion_init+0xbc/0x510 [i915]
[<c04cdc7c>] ? delayed_slow_work_enqueue+0xcc/0xf0
[<f80f7c20>] ? drm_kms_helper_poll_init+0xc0/0x130 [drm_kms_helper]
[<f8248ec7>] ? i915_driver_load+0x757/0x10c0 [i915]
[<f82471b0>] ? i915_vga_set_decode+0x0/0x20 [i915]
[<f81c5675>] ? drm_sysfs_device_add+0x75/0xa0 [drm]
[<f81c33d7>] ? drm_get_dev+0x2b7/0x4c0 [drm]
[<c05f6986>] ? pci_match_device+0x16/0xc0
[<c05f68ab>] ? local_pci_probe+0xb/0x10
[<c05f76b1>] ? pci_device_probe+0x61/0x80
[<c06a34f7>] ? driver_probe_device+0x87/0x290
etc.
So I guess that there might be something odd about that ioremap. It
triggers on every boot of the machine in question. It looks to me as if
perhaps its found a cached entry which doesn't fit in some way, but I'm
not certain of that yet,
Steve.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxxx For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>