This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Regclass cleanups


> 
> I didn't intend to approve it, but since you have asked
> I have studied it a bit more and I find no problems with
> it.
Thanks a lot. I didn't meant to urge you. I really wan't sure
if it is approval or not :)
> 
> You may install the patch with the minor formatting fix.
> 
> The only piece I questioned was the ASM_INPUT part, I tried to dig
> into ChangeLogs and CVS histories to figure out if anyone explicitly
> put it there for a reason.  Unfortunately Bernd's original cleanups in
> regclass (creating scan_one_insn) made it difficult to find out who
> changed what parts before that point in time so I stopped there.
It is already present in gcc 2.7.1 source tree (the oldes I have on my
HDD).
But while looking for the history I got idea why it is correct - the
asm appears as ASM_INPUT only when it is w/o operands. ASM with one
input operant is represented as ASM_OPERANDS with nested ASM_INPUT.

So I will install the patch w/o this bit.

Thanks
   Honza
> 
> Later,
> David S. Miller
> davem@redhat.com

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]