This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PING] Target hook for rewriting inline asm constraints
- From: "Andreas Krebbel" <Andreas dot Krebbel at de dot ibm dot com>
- To: Michael Meissner <michael dot meissner at amd dot com>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Wed, 7 Nov 2007 10:09:34 +0100
- Subject: Re: [PING] Target hook for rewriting inline asm constraints
- References: <20071030115424.GA6864@homer.boeblingen.de.ibm.com> <20071106181838.GA31748@mmeissner-gold.amd.com>
Hi Michael,
thanks for your review. I'll address the points you mentioned.
> Now, getting on the broader scope issues about the patch, I can sympathize with
> the desire, but I think this is really opening up a can of worms if we let the
> back end rewrite asm constraints like this. I'm also not sure I understand
> what the exact problem is. I would imagine if you have address modes that are
> allowed in some cases, but not others, this is better expressed in
> GO_IF_MODE_DEPENDENT_ADDRESS instead of GO_IF_LEGITIMATE_ADDRESS.
In order to support new address formats there is no other way than
extending the GO_IF_LEGITIMATE_ADDRESS hook. All addresses not
accepted by this hook will be reloaded into a register.
The GO_IF_MODE_DEPENDENT_ADDRESS tells the middle end if an address
should be interpreted depending on the mode of the memory access.
How do you think this can help me here?
Bye,
-Andreas-