On Sunday, 8 October 2006 08:47, Pavel Machek wrote:
> On Sat 2006-10-07 22:46:45, Rafael J. Wysocki wrote:
> > On Saturday, 7 October 2006 22:39, Stefan Seyfried wrote:
> > > On Sat, Oct 07, 2006 at 10:29:25PM +0200, Rafael J. Wysocki wrote:
> > > > On Saturday, 7 October 2006 22:16, Stefan Seyfried wrote:
> > > > > the additional printf() after the suspend_to_ram() seems to "fix" it.
> > > > > At
> > > > > least it did not crash on me for ~30 suspend cycles now.
> > > > > However, i'd rather like to understand what the real problem is
> > > > > instead
> > > > > of engaging in such cargo cult programming ;-)
> > > >
> > > > It evidently adds some tiny delay here ...
> > >
> > > yes, maybe. However, i am running memtest on the machine over night,
> > > maybe it
> > > is just a broken memory module :-)
> > >
> > > > > > Anyway, have you tried to replace suspend_to_ram(snapshot_fd) in
> > > > > > s2ram_do
> > > > > > with a noop and see what happens?
> > > > >
> > > > > Then it does not segfault.
> > > >
> > > > Hm, I'm still wondering if s2both has the same problem ...
> > >
> > > I tried for ~10 cycles and it had no problem.
> >
> > So the hardware may be the right guess ...
>
> I'd say it is way too repeatable for that. Some kernel code may be
> scribbling over memory, or something.... suspend is not exactly the
> "very common, very well tested" codepath.
Generally a segfault means there was a page fault when we accessed certain
address and it turned out to be not in our page tables (or the page was not
present) and the kernel was unable to map it, whatever the reason. What
gives?
Rafael
--
You never change things by fighting the existing reality.
R. Buckminster Fuller
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Suspend-devel mailing list
Suspend-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/suspend-devel
