Re: "Exception at ..." with 2.1.9 and 2.1.10 kernels

Kenneth Albanowski (kjahds@kjahds.com)
Tue, 26 Nov 1996 01:56:09 -0500 (EST)


On 24 Nov 1996, H. Peter Anvin wrote:

> > > > Btw, wouldn't it be a nice thing for us RISC people if on Intel
> > > > the kernel would enable the 486+ alignment exceptions and throw
> > > > warnings ;-)
> I suspect the point is to use it as a debugging device to find
> unaligned accesses in the kernel (that aren't portable to RISC
> architectures). Don't enable it in user space, though -- legitimate
> programs will break.

As long as it was just a warning, not a crash, user programs wouldn't
suffer to badly.

And let me add my vote for more stringent alignment. If I'm writing
something that I want to be portable, I'd gladly make use of any available
alignment strictures.

--
Kenneth Albanowski (kjahds@kjahds.com, CIS: 70705,126)