[PATCH] RISC-V: Use binvi to cover more immediates than with xori alone

[PATCH] RISC-V: Use binvi to cover more immediates than with xori alone

[Philipp Tomsich]

Sequences of the form "a ^ C" with C being the positive half of a
signed immediate's range with one extra bit set in addtion are mapped
to xori and one binvi to avoid using a temporary (and a multi-insn
sequence to load C into that temporary).

gcc/ChangeLog:

	* config/riscv/bitmanip.md (*binvi<mode>_extrabit): New pattern.

gcc/testsuite/ChangeLog:

	* gcc.target/riscv/zbs-binvi.c: New test.

Signed-off-by: Philipp Tomsich <philipp.tomsich@vrull.eu>
---
- Depends on a predicate posted in "RISC-V: Optimize branches testing
  a bit-range or a shifted immediate".  Depending on the order of
  applying these, I'll take care to pull that part out of the other
  patch if needed.

 gcc/config/riscv/bitmanip.md               | 19 +++++++++++++++++++
 gcc/testsuite/gcc.target/riscv/zbs-binvi.c | 22 ++++++++++++++++++++++
 2 files changed, 41 insertions(+)
 create mode 100644 gcc/testsuite/gcc.target/riscv/zbs-binvi.c

diff --git a/gcc/config/riscv/bitmanip.md b/gcc/config/riscv/bitmanip.md
index 436ff4ba958..7fa8461bb71 100644
--- a/gcc/config/riscv/bitmanip.md
+++ b/gcc/config/riscv/bitmanip.md
@@ -577,6 +577,25 @@
   "binvi\t%0,%1,%S2"
   [(set_attr "type" "bitmanip")])
 
+; Catch those cases where we can use a binvi + xori or binvi + binvi
+; instead of a lui + addi + xor sequence.
+(define_insn_and_split "*binvi<mode>_extrabit"
+  [(set (match_operand:X 0 "register_operand" "=r")
+	(xor:X (match_operand:X 1 "register_operand" "r")
+	       (match_operand:X 2 "uimm_extra_bit_operand" "i")))]
+  "TARGET_ZBS"
+  "#"
+  "&& reload_completed"
+  [(set (match_dup 0) (xor:X (match_dup 1) (match_dup 3)))
+   (set (match_dup 0) (xor:X (match_dup 0) (match_dup 4)))]
+{
+	unsigned HOST_WIDE_INT bits = UINTVAL (operands[2]);
+	unsigned HOST_WIDE_INT topbit = HOST_WIDE_INT_1U << floor_log2 (bits);
+
+	operands[3] = GEN_INT (bits &~ topbit);
+	operands[4] = GEN_INT (topbit);
+})
+
 (define_insn "*bext<mode>"
   [(set (match_operand:X 0 "register_operand" "=r")
 	(zero_extract:X (match_operand:X 1 "register_operand" "r")
diff --git a/gcc/testsuite/gcc.target/riscv/zbs-binvi.c b/gcc/testsuite/gcc.target/riscv/zbs-binvi.c
new file mode 100644
index 00000000000..c2d6725b53b
--- /dev/null
+++ b/gcc/testsuite/gcc.target/riscv/zbs-binvi.c
@@ -0,0 +1,22 @@
+/* { dg-do compile } */
+/* { dg-options "-march=rv64gc_zbs -mabi=lp64" } */
+/* { dg-skip-if "" { *-*-* } { "-O0" "-Og" } } */
+
+long long f3(long long a)
+{
+  return a ^ 0x1100;
+}
+
+long long f4 (long long a)
+{
+  return a ^ 0x80000000000000ffull;
+}
+
+long long f5 (long long a)
+{
+  return a ^ 0x8000001000000000ull;
+}
+
+/* { dg-final { scan-assembler-times "binvi\t" 4 } } */
+/* { dg-final { scan-assembler-times "xori\t" 2 } } */
+
-- 
2.34.1

Re: [PATCH] RISC-V: Use binvi to cover more immediates than with xori alone

[Jeff Law]

On 11/10/22 14:35, Philipp Tomsich wrote:
> Sequences of the form "a ^ C" with C being the positive half of a
> signed immediate's range with one extra bit set in addtion are mapped
> to xori and one binvi to avoid using a temporary (and a multi-insn
> sequence to load C into that temporary).
>
> gcc/ChangeLog:
>
> 	* config/riscv/bitmanip.md (*binvi<mode>_extrabit): New pattern.
>
> gcc/testsuite/ChangeLog:
>
> 	* gcc.target/riscv/zbs-binvi.c: New test.

Could this have been potentially done by extending the ior pattern with 
a code iterator that covered IOR/XOR?


Not a big deal, but if it'd work, consider to avoid the pattern duplication.


OK.

jeff

Re: [PATCH] RISC-V: Use binvi to cover more immediates than with xori alone

[Philipp Tomsich]

On Wed, 16 Nov 2022 at 04:45, Jeff Law <jeffreyalaw@gmail.com> wrote:
>
>
> On 11/10/22 14:35, Philipp Tomsich wrote:
> > Sequences of the form "a ^ C" with C being the positive half of a
> > signed immediate's range with one extra bit set in addtion are mapped
> > to xori and one binvi to avoid using a temporary (and a multi-insn
> > sequence to load C into that temporary).
> >
> > gcc/ChangeLog:
> >
> >       * config/riscv/bitmanip.md (*binvi<mode>_extrabit): New pattern.
> >
> > gcc/testsuite/ChangeLog:
> >
> >       * gcc.target/riscv/zbs-binvi.c: New test.
>
> Could this have been potentially done by extending the ior pattern with
> a code iterator that covered IOR/XOR?
>
>
> Not a big deal, but if it'd work, consider to avoid the pattern duplication.

We'll rework to using an iterator and a mapping in v2 for the
following cases, folding into a single patch:
 - X | C => bseti
 - X & ~C => bclri
 - X ^ C => binvi

Thanks.
Philipp.