Subject: Re: ARM 6/7 late abort fixup (was Re: Kernel copyin/out optimizations for ARM...)
To: Chris Gilbert <chris@paradox.demon.co.uk>
From: Reinoud Zandijk <reinoud@netbsd.org>
List: port-arm
Date: 03/16/2002 21:39:53
On Sat, Mar 16, 2002 at 12:02:53PM -0000, Chris Gilbert wrote:
> > i've just fixed a nasty ARM 6/7 specific
> > problem ... an instruction that couldn't be fixed up in f.e. the late
> abort
> > handler would _allways_ panic the kernel ... even when the instruction was
> > in userland :( ... appears to have been the case since 1994 when Mark
> > Brinicombe started on it ...
> 
> Have you got any test code that triggered it?  would it be worth putting it
> into a regression test (make sure we don't lose/break it again?)

Sure! allready added the patch :) ... the simplest way is to include an assember
instruction that allways fails die to segmentation fault but can't be fixed up
by the early nor late abort handler.

The code is now in src/regress/sys/arch/arm/abort-fixup

Cheers,
Reinoud