[Patch v2] align crash_notes allocation to make it be inside one physical page

Minfei Huang mhuang at redhat.com
Tue Aug 11 01:01:08 PDT 2015


On 08/11/15 at 02:33pm, Baoquan He wrote:
> Hi Andrew,
> 
> On 08/03/15 at 03:04pm, Andrew Morton wrote:
> > On Mon,  3 Aug 2015 20:50:43 +0800 Baoquan He <bhe at redhat.com> wrote:
> > And I think the WARN_ON can be replaced with a
> > BUILD_BUG_ON(sizeof>PAGE_SIZE)?  That would avoid adding runtime
> > overhead.
> 
> Rethink about this, you are right. Using BUILD_BUG_ON is better.
> Anyone who found this compiling break should check if his/her code
> changes increase the crash_notes size. If possible that increase need be
> avoidable. Otherwise he should report this to upstream why it's
> unavoidable to increase crash_notes size, then let's consider the
> redesign the crash_notes data structure. 
> 
> So I will use BUILD_BUG_ON and repost.

Baoquan.

If the size of notes never be exceeded to PAGE_SIZE, I think we can
revert below patch, since the situation which describes in patch does
not happen.

    commit 38dfac843cb6d7be1874888839817404a15a6b3c
    Author: Greg Pearson <greg.pearson at hp.com>
    Date:   Mon Feb 10 14:25:36 2014 -0800
    
        vmcore: prevent PT_NOTE p_memsz overflow during header update

What do you think about this?

Thanks
Minfei



More information about the kexec mailing list