home *** CD-ROM | disk | FTP | other *** search
- Path: cegelecproj.co.uk!usenet
- From: kt1@cegelecproj.co.uk
- Newsgroups: comp.sys.m68k
- Subject: Re: Bus Error Handler for 68040
- Date: 5 Feb 1996 12:44:22 GMT
- Organization: Cegelec Projects Ltd.
- Message-ID: <4f4u36$okc@jupiter.sdd.cegelecproj.co.uk>
- References: <MRK.96Jan31162911@tlcpmh19.elex.co.il> <3110FC52.474F@infinet.com> <4f3303$bff@gazette.tandem.com>
- Reply-To: kt1@cegelecproj.co.uk
- NNTP-Posting-Host: bob.cegelecproj.co.uk
- X-Newsreader: IBM NewsReader/2 v1.2
-
- In <4f3303$bff@gazette.tandem.com>, leeb@tandem.com (britton_lee) writes:
- >CPM2 Mark Rasin 3821 p.o.b.025 wrote:
- > The problem I've had with trying to recover from a Bus Error on the
- >68040 (and 68060) is that, unlike the 68030 and below, the processor
- >hardware always retries the failing instruction after the RTE.
- > ....
- >
-
- If you know which accesses may cause problems, you could write a 'safe access'
- function that temporarily replaces the bus error vector with a location further
- on in your code (you must save the current stack pointer and old vector etc).
- Then if you get a bus error, you jump to your code, restore the stack, and carry
- on in an appropriate way ...
-
- Kevin Taylor
- (kt1@cegelecproj.co.uk).
-
-