This is the mail archive of the
mailing list for the binutils project.
Re: [PATCH 5/6] Add Visium support to gas
- From: Eric Botcazou <ebotcazou at adacore dot com>
- To: Alan Modra <amodra at gmail dot com>
- Cc: binutils at sourceware dot org
- Date: Fri, 05 Dec 2014 13:27:38 +0100
- Subject: Re: [PATCH 5/6] Add Visium support to gas
- Authentication-results: sourceware.org; auth=none
- References: <cover dot 1417646850 dot git dot ebotcazou at gcc dot gnu dot org> <fee07d8c32de19e9586f02a00a656e3628d5f68f dot 1417646850 dot git dot ebotcazou at gcc dot gnu dot org> <20141205022551 dot GC20383 at bubble dot grove dot modra dot org>
> Heh. I'd like an explanation too. :) If the clearing is necessary
> it indicates to me that you have a bug somewhere in md_assemble
> setting the field non-zero..
> BTW, since your target is rela, you don't need to write to the field
> if you're emitting a reloc (which will happen if fx_done is clear on
> exit from md_apply_fix). It's cleaner to leave the field zero rather
> than writing in the addend.
> Also, I don't see any check on fx_subsy. You should emit an error if
> fx_subsy is present and not handled.
Neither tc-i386.c nor tc-sparc.c does that though.
> > + /* Are we finished with this relocation now? */
> > + if (fixP->fx_addsy == 0 && !fixP->fx_pcrel)
> > + fixP->fx_done = 1;
> That !fixP->fx_pcrel lools suspicious. Why punt all pcrel to the
Because tc-sparc.c does exactly the same? Removed.