This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Results for egcs-2.92.23 19981122 (gcc2 ss-980609 experimental)^M testsuite on i386-unknown-openbsd2.4
- To: Joern Rennecke <amylaar at cygnus dot co dot uk>
- Subject: Re: Results for egcs-2.92.23 19981122 (gcc2 ss-980609 experimental)^M testsuite on i386-unknown-openbsd2.4
- From: Jeffrey A Law <law at cygnus dot com>
- Date: Sat, 28 Nov 1998 10:04:47 -0700
- cc: schwab at issan dot informatik dot uni-dortmund dot de (Andreas Schwab), Marc dot Espie at liafa1 dot liafa dot jussieu dot fr, egcs at cygnus dot com, egcs-bugs at cygnus dot com
- Reply-To: law at cygnus dot com
In message <199811271938.TAA07708@phal.cygnus.co.uk>you write:
> > This is a bug in reload, it kills the REG_UNUSED notes, which cannot be
> > recomputed accurately. As a result the second stage compiler is
> > miscompiled (more exactly the function div_and_round_double which is
> > called by choose_multiplier). Here is a patch:
>
> I don't see why REG_UNUSED notes should be any more reliable than REG_DEAD
> notes at the end of reload.
> Neither do I see why it should be impossible to re-calculate them
> accurately.
> Can you please describe in more detail why flow failed to recompute the
> notes accurately.
I've already discussed this with Andreas, and I've already indicated that the
patch is conceptually OK. See older messages in egcs-patches.
There is still the problem that flow could add a second REG_UNUSED note to an
insn that I asked Andreas to resolve before we added his patch. Or at least
explain why such a note is OK and would not cause problems.
jeff