This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: [patch] MIPS64: A 32-bit address range fix
- From: "Maciej W. Rozycki" <macro at ds2 dot pg dot gda dot pl>
- To: Eric Christopher <echristo at redhat dot com>
- Cc: binutils at sources dot redhat dot com
- Date: Tue, 9 Jul 2002 02:27:46 +0200 (MET DST)
- Subject: Re: [patch] MIPS64: A 32-bit address range fix
- Organization: Technical University of Gdansk
- Reply-to: "Maciej W. Rozycki" <macro at ds2 dot pg dot gda dot pl>
On 8 Jul 2002, Eric Christopher wrote:
> > No. Should I? Especially with gcc ~2.95.4 I'm using? The code
> > generated now is definitely broken for the case I described.
>
> Probably, as long as there aren't any regressions in the gcc that you
> are using that'll be fine for me. I agree that it is broken, and a
> binutils testcase would be nice, but a sanity check with gcc is always
> appreciated as well.
I don't see any relevance here, sorry. The range of addresses that is
affected, i.e. 0xffffffff7fff8000 to 0xffffffff7fffffff, is invalid on
MIPS64 -- references to it always generate an address error exception.
How can gcc make any use of it?
I ran the test anyway and the only failure relevant to binutils is
because of a "la used to load 64-bit address" gas warning. Most tests
fail because of missing crt1.o -- I have no libc and I cannot do anything
about it at this stage.
OK to apply, then?
Maciej
--
+ Maciej W. Rozycki, Technical University of Gdansk, Poland +
+--------------------------------------------------------------+
+ e-mail: macro@ds2.pg.gda.pl, PGP key available +