[patch] Merge cfo-branch, RTL sequence abstraction (part 1)
Daniel Jacobowitz
drow@false.org
Sat Jan 7 18:59:00 GMT 2006
On Sat, Jan 07, 2006 at 11:12:25AM -0700, Roger Sayle wrote:
> This allows target backends to provide size information for their
> functionality. On CISC architectures such as x86, vax, etc...
> this is the usual way that this size information is provided to
> the middle-end. In recent years, the "length" insn attribute has
> been available for recognized instructions, but rtx_costs is used
> early during the compiler to estimate sizes prior to instruction
> selection or even RTL expansion. i.e. rtx_cost(op) should the
> speed/size of operation op, even if it isn't representable in a
> single instruction.
Of course I'm aware of this (and most of the rest of your message);
that wasn't my point. By the time we're doing CFO oughtn't we to have
established insns and their lengths? We're after reload, we've
selected variants of everything by now.
Again: CFO is a size optimization.
> Gabor's current approach of assuming constant length for each INSN_P
I didn't suggest that. I suggested using get_attr_length.
--
Daniel Jacobowitz
CodeSourcery
More information about the Gcc-patches
mailing list