Re: NULL pointer dereference in swsusp_free with 3.17-rc5

[Joerg Roedel <jroedel@suse.de>]

On Tue, Sep 23, 2014 at 10:28:00PM +0200, Rafael J. Wysocki wrote:
> On Tuesday, September 23, 2014 07:27:06 PM Bjørn Mork wrote:
> > 
> > --=-=-=
> > Content-Type: text/plain; charset=utf-8
> > Content-Transfer-Encoding: quoted-printable
> > 
> > "Rafael J. Wysocki" <rjw@rjwysocki.net> writes:
> > 
> > 
> > > I would suspect one of these commits:
> > >
> > > 84c91b7ae07c PM / hibernate: avoid unsafe pages in e820 reserved regions
> > > 0f7d83e85dbd PM / Hibernate: Touch Soft Lockup Watchdog in rtree_next_node
> > > 9047eb629e5c PM / Hibernate: Remove the old memory-bitmap implementation
> > > 6efde38f0769 PM / Hibernate: Iterate over set bits instead of PFNs in sws=
> > usp_free()
> > > 3a20cb177961 PM / Hibernate: Implement position keeping in radix tree
> > > 07a338236fdc PM / Hibernate: Add memory_rtree_find_bit function
> > > f469f02dc6fa PM / Hibernate: Create a Radix-Tree to store memory bitmap
> > >
> > > so I guess you can start from checking them (the topmpost one is the late=
> > st).
> > 
> > Thanks.  Yes, you were correct. The bad commit is
> > 
> >  6efde38f0769 PM / Hibernate: Iterate over set bits instead of PFNs in swsu=
> > sp_free()
> > 
> > I have confirmed that reverting only this commit on top of a clean
> > v3.17-rc6 fixes the problem.  I am attaching the context-modified revert
> > patch I used.
> 
> OK, thanks!
> 
> Joerg, can you please have a look at this?

Yes, I have a look.


	Joerg