This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: Bug in gas from binutils-2.13.2.1
- From: Daniel Jacobowitz <drow at false dot org>
- To: Jeff Rubin <jbr at keyresearch dot com>
- Cc: bug-gnu-utils at gnu dot org, binutils at sources dot redhat dot com
- Date: Fri, 3 Jan 2003 14:33:01 -0500
- Subject: Re: Bug in gas from binutils-2.13.2.1
- References: <138AA78F80DCE84B8EE424399FFBF9C902BFE1@exchange.ad.skymv.com>
On Thu, Jan 02, 2003 at 10:23:40PM -0800, Jeff Rubin wrote:
> This is a mips-specific bug report.
> In the file gas/config/tc-mips.c
> in the routine "macro" for the case
> M_DLA_AB, with mips_big_got == 1,
> mips_pic == SVR4_PIC, and HAVE_NEWABI
> (in my case -64). As the comment beginning
> "For NEWABI, we want ..." indicates, the relocation
> BFD_RELOC_MIPS_CALL16 should be used in the
> tempreg == PIC_CALL_REG case only for an external
> symbol. But the code uses that relocation without
> creating a frag_var for the local symbol case.
> (Now I'm starting to get on thin ice ...) It also
> doesn't seem to call frag_grow before the first
> call to macro_build. I have a patch that I am
> testing, but I am very unsure if what I am doing
> is correct. Let me know if you want it.
>
> Also, I noticed that mips_big_got is set to 1
> unconditionally for HAVE_NEWABI. I don't read
> the discussion lists so I may have missed a discussion
> about that. But is seems wrong to me. I have removed
> that code in my version.
I don't know much about the current state of the NEWABI code, but I
know that it's changed a lot since 2.13.2.1; if you're using PIC in
particular you should probably try a CVS head snapshot rather than
a 2.13 release.
--
Daniel Jacobowitz
MontaVista Software Debian GNU/Linux Developer