https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111154
--- Comment #3 from Tomas Chang <changyp6 at gmail dot com> ---
(In reply to Richard Biener from comment #2)
> Confirmed. Reproduces with -O3 -march=armv8.3-a and
>
> unsigned char desta[8];
> void copya(unsigned char *src, int size)
> {
> for (int i = 0; i < size; i++)
> desta[i] = src[i];
> }
>
> we end up with a peeled epilogue:
>
> if (_12 == niters_vector_mult_vf.6_39)
> goto <bb 26>; [12.50%]
> else
> goto <bb 6>; [87.50%]
>
> <bb 6> [local count: 73583527]:
> _96 = MEM[(unsigned char *)src_8(D) + 8B];
> desta[_36] = _96;
> if (size_7(D) > 9)
> goto <bb 7>; [85.71%]
> else
> goto <bb 27>; [14.29%]
>
> <bb 27> [local count: 10511932]:
> goto <bb 21>; [100.00%]
>
> <bb 7> [local count: 63071596]:
> _103 = MEM[(unsigned char *)src_8(D) + 9B];
> desta[9] = _103;
> if (size_7(D) != 10)
> goto <bb 8>; [85.71%]
> else
> goto <bb 28>; [14.29%]
>
> <bb 28> [local count: 9010228]:
> goto <bb 21>; [100.00%]
>
> <bb 8> [local count: 54061368]:
> _110 = MEM[(unsigned char *)src_8(D) + 10B];
> desta[10] = _110;
> if (size_7(D) != 11)
> goto <bb 9>; [85.71%]
> else
> goto <bb 29>; [14.29%]
> ...
Thanks for confirming this bug so quickly!
Does this issue affect the actually generated code ?
Are there any clues about how to fix this issue ?
If there are patches that can fix, I can help test the patch.
