Re: [RFC PATCH 2/4] mm: make consistent use of PG_slab flag

From: Christoph Lameter
Date: Fri Jul 06 2012 - 09:53:22 EST

On Fri, 6 Jul 2012, Jiang Liu wrote:

> This patch is not for hotplug, but is to fix some issues in current
> kernel, such as:
> 1) make show_mem() on ARM and unicore32 report consistent information
> no matter which slab allocator is used.

The information is only different because allocations do not go through
the slab allocators for SLUB/SLOB.

> 2) make /proc/kpagecount and /proc/kpageflags return accurate information.

Fix the compound handling in those and the numbers will be correct. This
is also good for other issues that may arise because the flags in the
compound head are not considered.

> 3) Get rid of risks in mm/memory_failure.c and arch/ia64/kernel/mca_drv.c

Assuming that a slab allocation fits into a page is a dangerous
assumption. There are arches with much large page sizes. Please fix the

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at