cvs commit: src/lib/libc/gdtoa Makefile.inc _hdtoa.c src/lib/libc/stdio vfprintf.c

[Marcel Moolenaar]

On Sun, Jan 18, 2004 at 03:03:13PM -0800, David Schultz wrote:
> On Sun, Jan 18, 2004, David Schultz wrote:
> > The correct result is 4.940656e-324, which is denormalized.  Some
> > architectures generate a trap for denormals and expect them to be
> > handled in software, but I didn't think that IA64 was one of them.
> > I'm looking over the IA64 architecture manual now to see if I can
> > figure out what's going on...
> 
> Upon further examination, it looks like IA64 *does* delegate the
> edge cases of the IEEE 754 standard to software, unlike the x86.

Correct.

> This implies that there may be a problem with the software assist
> routines that the kernel calls when it gets a trap.  I don't have
> any IA64 hardware to muck around with, though, so I can't tell you
> anything more.

Since we have infrequent SIGFPEs in top(1) as well, I knew something
was not quite right. I just never got around analyzing what goes
wrong exactly. I'll look into it.

Thanks,

-- 
 Marcel Moolenaar	  USPA: A-39004		 marcel at xcllnt.net