Subject: Re: FreSSH and bounds checking
To: <>
From: David Laight <david@l8s.co.uk>
List: current-users
Date: 03/12/2002 09:03:26
> I don't think it would have helped one bit with this bug, for that
> you would need a malloc implementation that mprotect()'d
> each side of the allocated area, so the program would die if the heap
> overflowed or was accessed, but this is very inefficient and not so
> portable and might fail too. (I think it would add 8192 bytes to each
> malloc on i386.)

Probably 4k...

I ran a SVR4 system with a kernel malloc that only ever gave out
pages while trying to find one corruption (on a test system
during heavy soak testing).

Couldn't run the TCP stack at all!

(bug turned out to be hardware)

	David

-- 
David Laight: david@l8s.co.uk