Re: [PATCH 3.2 099/125] mm: fix crashes from mbind() merging vmas

[Hugh Dickins]

On Wed, 9 Jul 2014, Ben Hutchings wrote:
> On Tue, 2014-07-08 at 14:26 -0700, Hugh Dickins wrote:
> > On Tue, 8 Jul 2014, Hugh Dickins wrote:
> > > On Tue, 8 Jul 2014, Ben Hutchings wrote:
> > > 
> > > > 3.2.61-rc1 review patch.  If anyone has any objections, please let me know.
> > > > 
> > > > ------------------
> > > > 
> > > > From: Hugh Dickins <hughd@xxxxxxxxxx>
> > > > 
> > > > commit d05f0cdcbe6388723f1900c549b4850360545201 upstream.
> > > 
> > > I don't think that's correct, I can't see a change to the
> > > "vma = check_range(" in migrate_to_node() which 3.2 and 3.4 trees have.
> > > Does it build (without warning) with CONFIG_NUMA and CONFIG_MIGRATION?
> > > I expect this version I sent yesterday for 3.4.98 will be good for 3.2:
> > 
> > Hah, I now see that 098/125 in your series
> 
> Which, for anyone who missed it, was commit 082708072a42, 'mm: revert
> 0def08e3 ("mm/mempolicy.c: check return code of  check_range")'.
> 
> > removes precisely the "vma = "
> > I was calling attention to.  So all should be okay and you can ignore me.
> 
> Right - of course I'm assuming that the reasoning given for that revert
> is valid for 3.2.

It should be fine: both reverter 082708072a42 and revertee 0def08e3
were cleanups of little importance, it just happens that the revert
made your backport of mine slightly easier, hence you included it.

The revert does add a VM_BUG_ON, at first a worry for -stable,
but in fact it looks safe to me.

(I wish I had done a much simpler patch for mine, but too late for that!)

Hugh
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/