https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102517
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Assignee|rguenth at gcc dot gnu.org |rearnsha at gcc dot
gnu.org
CC| |rguenth at gcc dot gnu.org
--- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> ---
So gcc.target/aarch64/cpymem-q-reg_1.c remains - since it all started with your
patch Richard, can you please see how to adjust the testcase? For me it now
produces
foo:
.LFB0:
.cfi_startproc
adrp x1, src
add x1, x1, :lo12:src
adrp x0, dst
add x0, x0, :lo12:dst
ld1 {v0.16b - v1.16b}, [x1]
st1 {v0.16b - v1.16b}, [x0]
ret
using OImode moves with "uint256_t" type.
_2 = MEM <uint256_t> [(char * {ref-all})&src];
MEM <uint256_t> [(char * {ref-all})&dst] = _2;
previously the C frontend didn't have such type and thus
langhooks.type_for_size returned NULL. But now we do have the OImode integer
mode and support for moving it [but the actual move ends up using vector
registers].
