[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Unaligned accesses (was Re: debian on ruffian)

On Mon, Sep 14, 1998 at 10:35:44AM -0400, Christopher C Chimelis wrote:
> On Mon, 14 Sep 1998, Andrew Isaacson wrote:
> > The X server unaligned accesses are apparently fixable, and there
> > might be a fix in the next X release.  I'll see about getting a patch
> > if anyone is interested.
> Yes :)

There is a patch floating around that's supposed to fix some unaligned
accesses in XFree86, but it doesn't seem to get 'em all (at least not
on my Millennium II).  I'll be looking into this in more depth RSN.

> > In the meantime, I've got a kernel patch that allows a bit more
> > control over how the kernel logs unaligned accesses.  This patch
> > creates a new /proc file, /proc/sys/kernel/log-unaligned.  If this
> > file contains '0', unaligned accesses are not logged.  If this file
> > contains '1', unaligned accesses are logged.
> > 
> > I'd appreciate knowing if anybody finds this useful.
> Actually, this will be very useful.  Do you know if it'll apply cleanly to
> the devel kernels?  I'm not at my alpha now, otherwise I would check.

It looks like someone added much better capabilities along these lines
to the 2.1 series.  2.1.121 apparently lets you control on a
per-process basis what the kernel does with unaligned accesses:
 - silently fix 'em up
 - log and fix (the old behavior)
 - deliver SIGBUS to the offending process
 - don't fix /* Not sure why you'd want to use this, but... */

I don't yet know how to access these capabilities, but I'll look into
it and get back to the list on it.  It looks like it has something to
do with osf_getsysinfo; is this interface documented anywhere?

So, my patch has been superseded in the 2.1 kernels.

Andy Isaacson adisaacs@mtu.edu adi@acm.org    Fight Spam, join CAUCE:
http://www.csl.mtu.edu/~adisaacs/              http://www.cauce.org/

Reply to: