This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Live range splitting in new allocator
- To: meissner at cygnus dot com (Michael Meissner)
- Subject: Re: Live range splitting in new allocator
- From: Joe Buck <jbuck at racerx dot synopsys dot com>
- Date: Wed, 31 Jan 2001 11:41:43 -0800 (PST)
- Cc: chris at lslsun dot epfl dot ch (Christian Iseli), meissner at cygnus dot com, Marc dot Espie at liafa dot jussieu dot fr, aoliva at redhat dot com, espie at quatramaran dot ens dot fr, dewar at gnat dot com, gcc at gcc dot gnu dot org
On Wed, Jan 31, 2001 at 05:11:16PM +0100, Christian Iseli wrote:
> > For the registers, I guess the function prologue could set them to
> > the specified value at the same time as the stack...
Michael Meissner writes:
> That assumes there is a convenient method to set the call clobbered registers
> (and call saved registers that are used) to a given value. Some registers are
> harder to set (for example fp registers on machines where you can't copy from
> int registers to fp registers without going through memory).
You're attacking the problem at too low a level. Front ends can insert
initializations for all automatic variables easily enough. This then
allows dead code elimination to kill unused initializations.