[gcc(refs/users/marxin/heads/marxin-gcc-benchmark-branch)] aarch64: Add an extra sbfiz pattern [PR87763]
Martin Liska
marxin@gcc.gnu.org
Mon Mar 30 10:50:32 GMT 2020
https://gcc.gnu.org/g:b65a1eb3fae53f2e1ea1ef8c1164f490d55855a1
commit b65a1eb3fae53f2e1ea1ef8c1164f490d55855a1
Author: Richard Sandiford <richard.sandiford@arm.com>
Date: Mon Feb 3 21:12:35 2020 +0000
aarch64: Add an extra sbfiz pattern [PR87763]
This patch matches another form of sbfiz, in which the input
has DImode and the output has SImode. It fixes a regression
in gcc.target/aarch64/lsl_asr_sbfiz.c from GCC 8.
2020-02-06 Richard Sandiford <richard.sandiford@arm.com>
gcc/
PR rtl-optimization/87763
* config/aarch64/aarch64.md (*ashiftsi_extvdi_bfiz): New pattern.
Diff:
---
gcc/ChangeLog | 5 +++++
gcc/config/aarch64/aarch64.md | 15 +++++++++++++++
2 files changed, 20 insertions(+)
diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index 1fe29d337cd..efbbbf08225 100644
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,3 +1,8 @@
+2020-02-06 Richard Sandiford <richard.sandiford@arm.com>
+
+ PR rtl-optimization/87763
+ * config/aarch64/aarch64.md (*ashiftsi_extvdi_bfiz): New pattern.
+
2020-02-06 Delia Burduv <delia.burduv@arm.com>
* config/aarch64/aarch64-simd-builtins.def
diff --git a/gcc/config/aarch64/aarch64.md b/gcc/config/aarch64/aarch64.md
index 4f5898185f5..90eebce85c0 100644
--- a/gcc/config/aarch64/aarch64.md
+++ b/gcc/config/aarch64/aarch64.md
@@ -5771,6 +5771,21 @@
[(set_attr "type" "bfx")]
)
+(define_insn "*ashiftsi_extvdi_bfiz"
+ [(set (match_operand:SI 0 "register_operand" "=r")
+ (ashift:SI
+ (match_operator:SI 4 "subreg_lowpart_operator"
+ [(sign_extract:DI
+ (match_operand:DI 1 "register_operand" "r")
+ (match_operand 2 "aarch64_simd_shift_imm_offset_si")
+ (const_int 0))])
+ (match_operand 3 "aarch64_simd_shift_imm_si")))]
+ "IN_RANGE (INTVAL (operands[2]) + INTVAL (operands[3]),
+ 1, GET_MODE_BITSIZE (SImode) - 1)"
+ "sbfiz\\t%w0, %w1, %3, %2"
+ [(set_attr "type" "bfx")]
+)
+
;; When the bit position and width of the equivalent extraction add up to 32
;; we can use a W-reg LSL instruction taking advantage of the implicit
;; zero-extension of the X-reg.
More information about the Gcc-cvs
mailing list