Date: Sun, 30 Dec 2007 10:34:33 +0100 From: Bernd Walter <ticso@cicely12.cicely.de> To: Dag-Erling =?iso-8859-1?Q?Sm=F8rgrav?= <des@des.no> Cc: Wilko Bulte <wb@freebie.xs4all.nl>, ivoras@freebsd.org, freebsd-hackers@freebsd.org Subject: Re: Architectures with strict alignment? Message-ID: <20071230093432.GF31522@cicely12.cicely.de> In-Reply-To: <86prwp161k.fsf@ds4.des.no> References: <fl4c8o$vpu$1@ger.gmane.org> <20071229.122221.-432830441.imp@bsdimp.com> <20071229193034.GA73845@freebie.xs4all.nl> <86prwp161k.fsf@ds4.des.no>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, Dec 29, 2007 at 11:37:27PM +0100, Dag-Erling Smørgrav wrote: > Wilko Bulte <wb@freebie.xs4all.nl> writes: > > In the past the alpha port had it too. > > No, it was optional and defaulted to off. It was never optional, since no alpha CPU can do missaligned access Alphas can fix missaligned access by software trap handlers in the same way as most other strong alignment architectures can and we did this for userland, but not in kernel. Nevertheless it is horribly slow to do this, but was required since there was so many crappy software that days - fortunately this has changed over time, although I still see aligment traps on new software as well. Sadly said we never implemented missaligment traps for x86 so developers without alphas could see their faults and made us alpha people a hard live by introducing new missalignments bugs on a regular basis so that many finaly gave up on that loop. On the other hand people should keep in mind that even modern x86 are not very good when it comes to missaligned access. They handle it in hardware, but it is not that optimized as handling alligned access, so you still see a major performance penalty. -- B.Walter http://www.bwct.de http://www.fizon.de bernd@bwct.de info@bwct.de support@fizon.de
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20071230093432.GF31522>