Bug#671866: gmp: causes gcc ICEs on amd64 and kfreebsd-amd64

[Steve M. Robbins]

On Tue, May 08, 2012 at 10:44:57AM +0200, Torbjorn Granlund wrote:

> "Steve M. Robbins" <steve at sumost.ca> writes:

>   Interesting.  It could well be the configure option.  For the past 6
>   years, Debian built using --disable-fat on amd64.  I changed it
>   because there are both bug workarounds and code speedups that are
>   processor-based.
> 
> Changed it to --disable-fat a couple of years ago with that aim, or
> changed it away from --disable-fat now with that aim?

The latter.  I changed it this weekend TO --enable-fat with the aim to
take advantage of the run-time processor detection.


> If you are aware of any bugs that will exist only in fat or only in
> non-fat GMP builds, the GMP project would like to know about that.

What we have learned in this bug thread is this: when built with
--enable-fat on an amd64 machine and then run on an amd64 *virtual*
machine, the GMP shared lib causes an internal compiler error when gcc
compiles a simple file.


> No idea why your emulator doesn't emulate hardware to GMP's
> satisfaction.  Some instruction is probably not emulated correctly.
> I've seen cases where emulators claim 64-bittyness and non-64-bittyness
> at the same time (such as 486 with AMD64 insns), and even had suggested
> patch to GMP a while ago with a workaround...

Agreed that we still don't know whether there is a bug in GMP or in
the emulator.  I hope the others will chime in with details of the
emulator used.

Torbjorn: which emulator caused a problem for GMP in the past?  What
was the symptom?  How does the run-time detection get confused by it?
Was the GMP patch applied?

Thanks,
-Steve
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 190 bytes
Desc: Digital signature
URL: <http://gmplib.org/list-archives/gmp-bugs/attachments/20120508/450c8448/attachment.bin>