This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Fwd: question about register pairs
- From: Joern Rennecke <joern dot rennecke at embecosm dot com>
- To: GCC <gcc at gcc dot gnu dot org>
- Date: Fri, 25 Oct 2013 13:53:28 +0100
- Subject: Fwd: question about register pairs
- Authentication-results: sourceware.org; auth=none
- References: <201310240659 dot r9O6x4jo006247 at greed dot delorie dot com> <CAMqJFCpoeVT6uWFusG6m9AfihEHcLTNwBWYRAi6R-eHmcaFCAw at mail dot gmail dot com> <201310250415 dot r9P4FEZt020568 at greed dot delorie dot com> <CAMqJFCqGiuQnE2Rv0b3WEOOzQYg9iEJcOyEq=0V_wqzaqCJ6jg at mail dot gmail dot com>
>From the discussion on gcc-patches - I wonder how far should the port go
here to paper over basic brokenness in gcc to handle multi-hard-register
pointers.
On 25 October 2013 13:35, Joern Rennecke <joern.rennecke@embecosm.com> wrote:
> Well, it's not really a lie if you map hardware registers 22 and 23 to
> a single register for the purposes of gcc internals. Although it
> does make some other things more awkward, e.g. when you
> copy fp, and this gets split so you have an insn that copies the
> highpart of fp to another register.
As an aside, I wonder how much it'd take to make gcc properly
handle multi-register stack/frame/arg pointers.
Do we also have issues with this and static frame pointers?
Should we have REG_INC notes for each hard register affected,
or should HARD_REGNO_NREGS of the reg inside been taken
into account? Maybe with some info stashed in the mode of the
note to speed things up?