* [PATCH][committed] aarch64: Use standard RTL codes for __rev16 intrinsic expansion
@ 2023-04-18 14:07 Kyrylo Tkachov
0 siblings, 0 replies; only message in thread
From: Kyrylo Tkachov @ 2023-04-18 14:07 UTC (permalink / raw)
To: gcc-patches
[-- Attachment #1: Type: text/plain, Size: 744 bytes --]
Hi all,
I noticed for the expansion of the __rev16* arm_acle.h intrinsics we don't need to use an unspec just because it doesn't match neatly to a bswap code.
We have organic combine patterns for it that we can reuse.
This patch removes the define_insn using UNSPEC_REV (should it have been an UNSPEC_REV16?) and adds an expander to emit
the patterns we have for rev16 using standard RTL codes.
Bootstrapped and tested on aarch64-none-linux-gnu.
Pushing to trunk.
Thanks,
Kyrill
gcc/ChangeLog:
* config/aarch64/aarch64.md (@aarch64_rev16<mode>): Change to
define_expand.
(rev16<mode>2): Rename to...
(aarch64_rev16<mode>2_alt1): ... This.
(rev16<mode>2_alt): Rename to...
(*aarch64_rev16<mode>2_alt2): ... This.
[-- Attachment #2: rev16.patch --]
[-- Type: application/octet-stream, Size: 2286 bytes --]
diff --git a/gcc/config/aarch64/aarch64.md b/gcc/config/aarch64/aarch64.md
index 671843c64af7ac3d98428e2d0f999520c026291b..036f0067b6dfd3df10c3d57119bab8e0ed7ecba5 100644
--- a/gcc/config/aarch64/aarch64.md
+++ b/gcc/config/aarch64/aarch64.md
@@ -6089,13 +6089,6 @@ (define_insn "bswaphi2"
[(set_attr "type" "rev")]
)
-(define_insn "@aarch64_rev16<mode>"
- [(set (match_operand:GPI 0 "register_operand" "=r")
- (unspec:GPI [(match_operand:GPI 1 "register_operand" "r")] UNSPEC_REV))]
- ""
- "rev16\\t%<w>0, %<w>1"
- [(set_attr "type" "rev")])
-
(define_insn "*aarch64_bfxil<mode>"
[(set (match_operand:GPI 0 "register_operand" "=r,r")
(ior:GPI (and:GPI (match_operand:GPI 1 "register_operand" "r,0")
@@ -6152,7 +6145,7 @@ (define_insn "*aarch64_bfxilsi_uxtw"
;; operations within an IOR/AND RTX, therefore we have two patterns matching
;; each valid permutation.
-(define_insn "rev16<mode>2"
+(define_insn "aarch64_rev16<mode>2_alt1"
[(set (match_operand:GPI 0 "register_operand" "=r")
(ior:GPI (and:GPI (ashift:GPI (match_operand:GPI 1 "register_operand" "r")
(const_int 8))
@@ -6166,7 +6159,7 @@ (define_insn "rev16<mode>2"
[(set_attr "type" "rev")]
)
-(define_insn "rev16<mode>2_alt"
+(define_insn "*aarch64_rev16<mode>2_alt2"
[(set (match_operand:GPI 0 "register_operand" "=r")
(ior:GPI (and:GPI (lshiftrt:GPI (match_operand:GPI 1 "register_operand" "r")
(const_int 8))
@@ -6180,6 +6173,21 @@ (define_insn "rev16<mode>2_alt"
[(set_attr "type" "rev")]
)
+;; Expander for __rev16 intrinsics. We have organic RTL patterns for rev16 above.
+;; Use this expander to just create the shift constants needed.
+(define_expand "@aarch64_rev16<mode>"
+ [(match_operand:GPI 0 "register_operand")
+ (match_operand:GPI 1 "register_operand")]
+ ""
+ {
+ rtx left = gen_int_mode (HOST_WIDE_INT_C (0xff00ff00ff00ff00), <MODE>mode);
+ rtx right = gen_int_mode (HOST_WIDE_INT_C (0xff00ff00ff00ff), <MODE>mode);
+ emit_insn (gen_aarch64_rev16<mode>2_alt1 (operands[0], operands[1],
+ right, left));
+ DONE;
+ }
+)
+
;; zero_extend version of above
(define_insn "*bswapsi2_uxtw"
[(set (match_operand:DI 0 "register_operand" "=r")
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2023-04-18 14:08 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-04-18 14:07 [PATCH][committed] aarch64: Use standard RTL codes for __rev16 intrinsic expansion Kyrylo Tkachov
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).