Bug on PA-RISC

[John David Anglin]

>  > It's possible that the pa64 problems are not related to the DFA code.
>  > I know there are other bugs.  For example, there is a miscompilation
>  > of ld.  I have started looking at it but I haven't found the exact
>  > problem yet.  However, I don't think it will be particularly hard
>  > to find the problem.
> It's certainly odd that PA64 is going nuts while PA32 genattrtab runs in less
> than an minute on my machines.

In ld, it looks as if the problem is in __udivmoddi4.  The results
of the '%' operation are wrong.  This appears to be causing a problem with
the hash algorithm.

Time to write a little test program.

Dave
-- 
J. David Anglin                                  dave.anglin@nrc.ca
National Research Council of Canada              (613) 990-0752 (FAX: 952-6605)