Re: memory corruption [Re: try KMGUARD]

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

On Thu, Mar 15, 2012 at 03:59:24PM +0200, Andreas Gustafsson wrote:
> Manuel Bouyer wrote:
> > And another one. It looks like all corruptions occurs in bss for me ...
> 
> Let's see if src/sys/netinet/rfc6056.c 1.5 fixes it.

that may well be it. I took my previous alanysis (I didn't post
all of them :), in all but one case the corrupted address is after dhtable
in bss and in the range of a uint16_t index. the case that doesn't fit
is the first one; I don't have the kernel for this case any more but it's
possible that different compile options have shifted the addresses in
bss.

Running the atf tests again now with an updated kernel; it will need a few
hours to make sure the problem is fixed.

-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--