This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: RFA: Fix debug info for MIPS doubles
- From: Daniel Jacobowitz <drow at false dot org>
- To: gcc-patches at gcc dot gnu dot org
- Cc: Eric Christopher <echristo at apple dot com>, rsandifo at nildram dot co dot uk
- Date: Wed, 17 Oct 2007 11:41:28 -0400
- Subject: Re: RFA: Fix debug info for MIPS doubles
- References: <20071017125452.GA6253@caradoc.them.org> <87ir55c0dr.fsf@firetop.home>
On Wed, Oct 17, 2007 at 02:54:24PM +0100, Richard Sandiford wrote:
> This condition seems overly complex. Did you come across a case that
> forced you to handle SCmode? I wouldn't have expected to see a hard
> SCmode FPR: the MIPS function hooks should return pairs of registers
> instead, and SCmode variables ought to be CONCATs of two SFmode registers.
> If we need to handle SCmode FPRs, we'd need to buff up other parts
> of the port too.
No, I was just being defensive. If we'll never get a reg:SC in a
register, then it's not necessary.
> Also, did you find a condition that needs words == 4 or words == 8?
Ditto; I was surprised to learn that there would never be a reg:TF in
a hard register.
Your changes seem reasonable to me; I'm running tests now (along with
the alternate FAKE_CALL_REGNO change).
--
Daniel Jacobowitz
CodeSourcery