This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: PATCH [mainline] Altivec PPC-DARWIN ABI Parameter Passing with Varargs
>>>>> "Fariborz" == Fariborz Jahanian <fjahanian@apple.com> writes:
>> work
>> with your patch:
>>
>> printf ("%vx %vx %vx", v1, v2, v3);
Testcase? Preferably one without printf extensions which most OSes
don't have.
> ! else if (ALTIVEC_VECTOR_MODE(mode) && align_words <= (GP_ARG_NUM_REG - 1))
Space after ALTIVEC_VECTOR_MODE.
> ! /* R9-R10 */
/* R9 - R10. */
> ! + align_words+1),
align_words + 1
> ! GEN_INT(4))));
Space after GEN_INT.
> ! /* R5-R8 */
/* R5 - R8. */
> ! int k;
> ! int size = int_size_in_bytes (type);
> ! int no_units = ((size - 1) / 4) + 1;
> ! int max_no_words = GP_ARG_NUM_REG - align_words;
> ! int rtlvec_len = no_units < max_no_words ? no_units : max_no_words;
> ! rtx *rtlvec = (rtx *) alloca (rtlvec_len * sizeof (rtx));
> ! memset ((char *) rtlvec, 0, rtlvec_len * sizeof
Indentation. Also, this same code appears in the BLKmode case. Can
you abstract it somehow to avoid code duplicity?
David, Alan. Are you guys ok with this patch wrt AIX and PPC64?
Aldy