Re: [PATCH 1/2] tmpfs mempolicy: fix /proc/mounts corrupting memory

[Hugh Dickins <hughd@google.com>]

On Wed, 2 Jan 2013, Christoph Lameter wrote:
> On Wed, 2 Jan 2013, Hugh Dickins wrote:
> 
> > Recent NUMA enhancements are not to blame: this dates back to 2.6.35,
> > when commit e17f74af351c "mempolicy: don't call mpol_set_nodemask()
> > when no_context" skipped mpol_parse_str()'s call to mpol_set_nodemask(),
> > which used to initialize v.preferred_node, or set MPOL_F_LOCAL in flags.
> > With slab poisoning, you can then rely on mpol_to_str() to set the bit
> > for node 0x6b6b, probably in the next page above the caller's stack.
> 
> Ugly. But 2.6.35 means that the patch was not included in several
> enterprise linux releases.

Thanks, that's some relief.  I forgot to mention that a good test for
whether your particular kernel (with who knows what additional patches
applied) is affected, is to

mount -o remount,mpol=local /dev/shm # which should be a tmpfs
grep /dev/shm /proc/mounts

If that says "mpol=prefer" then you're affected and need the fix; if
it says "mpol=local" (like 2.6.34 or after this fix) then you're safe.

(Conversely, setting "mpol=prefer" shows up as "mpol=local" after the,
fix, since that's what prefer without a node specification amounts to.)

> 
> > I don't understand why MPOL_LOCAL is described as a pseudo-policy:
> > it's a reasonable policy which suffers from a confusing implementation
> > in terms of MPOL_PREFERRED with MPOL_F_LOCAL.  I believe this would be
> > much more robust if MPOL_LOCAL were recognized in switch statements
> > throughout, MPOL_F_LOCAL deleted, and MPOL_PREFERRED use the (possibly
> > empty) nodes mask like everyone else, instead of its preferred_node
> > variant (I presume an optimization from the days before MPOL_LOCAL).
> > But that would take me too long to get right and fully tested.
> 
> The current approaches to implementing NUMA scheduling are making
> MPOL_LOCAL an explicit policy. See
> https://patchwork.kernel.org/patch/1703641/.

It's a good step in the right direction.

> 
> Does that address the concerns?

It makes no difference to this bug, and does not go far enough to
remove all the MPOL_F_LOCAL MPOL_PREFERRED MPOL_LOCAL twistiness.

Hugh

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>