[PATCH][updated] Force SDmode function args into FP registers per the ABI
Paolo Bonzini
paolo.bonzini@lu.unisi.ch
Tue Jun 19 15:38:00 GMT 2007
>> /* Reuse a stack slot if possible. */
>> else if (spill_stack_slot[from_reg] != 0
>> && spill_stack_slot_width[from_reg] >= total_size
>> && (GET_MODE_SIZE (GET_MODE (spill_stack_slot[from_reg]))
>> >= inherent_size)
>
> Umm, are you sure? I would think if we spilled the register as
> SFmode first and then as an SDmode, that we would fail the
> "spill_stack_slot_width[from_reg] >= total_size" test.
> OTOH, if we spilled it as SDmode first and then an SFmode, then
> we would reuse the stack slot, but that's probably alright.
You are right. I am still a little confused by the way big-endian
adjustment is done; but I guess these adjustments would produce wrong
code if inherent_size had to be 8.
Paolo
More information about the Gcc-patches
mailing list