Hi, See below a patch to address PR 83009.
Tested with aarch64-linux-gnu bootstrap and regtests for c, c++ and fortran. Ran the adjusted testcase for -mabi=ilp32. Is this OK for gcc-9? Cheers, Andre PR target/83009: Relax strict address checking for store pair lanes The operand constraint for the memory address of store/load pair lanes was enforcing strictly hardware registers be allowed as memory addresses. We want to relax that such that these patterns can be used by combine. During register allocation the register constraint will enforce the correct register is chosen. gcc 2018-05-xx Andre Vieira <andre.simoesdiasvie...@arm.com> PR target/83009 * config/aarch64/predicates.md (aarch64_mem_pair_lanes_operand): Make address check not strict. gcc/testsuite 2018-05-xx Andre Vieira <andre.simoesdiasvie...@arm.com> PR target/83009 * gcc/target/aarch64/store_v2vec_lanes.c: Add extra tests.
diff --git a/gcc/config/aarch64/predicates.md b/gcc/config/aarch64/predicates.md index 5d41d4350402b2a9e5941f160c6ab6f933bfff90..f29bc8e74f0070589014ac87fd22a95723ba9be8 100644 --- a/gcc/config/aarch64/predicates.md +++ b/gcc/config/aarch64/predicates.md @@ -222,7 +222,7 @@ ;; as a 128-bit vec_concat. (define_predicate "aarch64_mem_pair_lanes_operand" (and (match_code "mem") - (match_test "aarch64_legitimate_address_p (DFmode, XEXP (op, 0), 1, + (match_test "aarch64_legitimate_address_p (DFmode, XEXP (op, 0), 0, ADDR_QUERY_LDP_STP)"))) (define_predicate "aarch64_prefetch_operand" diff --git a/gcc/testsuite/gcc.target/aarch64/store_v2vec_lanes.c b/gcc/testsuite/gcc.target/aarch64/store_v2vec_lanes.c index 990aea32de6f8239effa95a081950684c6e11386..3296d04da14149d26d19da785663b87bd5ad8994 100644 --- a/gcc/testsuite/gcc.target/aarch64/store_v2vec_lanes.c +++ b/gcc/testsuite/gcc.target/aarch64/store_v2vec_lanes.c @@ -22,10 +22,32 @@ construct_lane_2 (long long *y, v2di *z) z[2] = x; } +void +construct_lane_3 (double **py, v2df **pz) +{ + double *y = *py; + v2df *z = *pz; + double y0 = y[0] + 1; + double y1 = y[1] + 2; + v2df x = {y0, y1}; + z[2] = x; +} + +void +construct_lane_4 (long long **py, v2di **pz) +{ + long long *y = *py; + v2di *z = *pz; + long long y0 = y[0] + 1; + long long y1 = y[1] + 2; + v2di x = {y0, y1}; + z[2] = x; +} + /* We can use the load_pair_lanes<mode> pattern to vec_concat two DI/DF values from consecutive memory into a 2-element vector by using a Q-reg LDR. */ -/* { dg-final { scan-assembler-times "stp\td\[0-9\]+, d\[0-9\]+" 1 { xfail ilp32 } } } */ -/* { dg-final { scan-assembler-times "stp\tx\[0-9\]+, x\[0-9\]+" 1 { xfail ilp32 } } } */ -/* { dg-final { scan-assembler-not "ins\t" { xfail ilp32 } } } */ +/* { dg-final { scan-assembler-times "stp\td\[0-9\]+, d\[0-9\]+" 2 } } */ +/* { dg-final { scan-assembler-times "stp\tx\[0-9\]+, x\[0-9\]+" 2 } } */ +/* { dg-final { scan-assembler-not "ins\t" } } */