Re: [GCC][PATCH v2][Aarch64] Exploiting BFXIL when OR-ing two AND-operations with appropriate bitmasks

[Sam Tebbs <sam dot tebbs at arm dot com>]

On 09/11/2018 04:20 PM, James Greenhalgh wrote:
> On Tue, Sep 04, 2018 at 10:13:43AM -0500, Sam Tebbs wrote:
> > Hi James,
> >
> > Thanks for the feedback. Here is an update with the changes you proposed
> > and an updated changelog.
> >
> > gcc/
> > 2018-09-04  Sam Tebbs  <sam.tebbs@arm.com>
> >
> >           PR target/85628
> >           * config/aarch64/aarch64.md (*aarch64_bfxil):
> >           Define.
> >           * config/aarch64/constraints.md (Ulc): Define
> >           * config/aarch64/aarch64-protos.h (aarch64_high_bits_all_ones_p):
> >           Define.
> >           * config/aarch64/aarch64.c (aarch64_high_bits_all_ones_p): New function.
> >
> > gcc/testsuite
> > 2018-09-04  Sam Tebbs  <sam.tebbs@arm.com>
> >
> >           PR target/85628
> >           * gcc.target/aarch64/combine_bfxil.c: New file.
> >           * gcc.target/aarch64/combine_bfxil_2.c: New file.
> <snip>
>
> > +/* Return true if I's bits are consecutive ones from the MSB.  */
> > +bool
> > +aarch64_high_bits_all_ones_p (HOST_WIDE_INT i)
> > +{
> > +  return exact_log2(-i) != HOST_WIDE_INT_M1;
> > +}
> You need a space in here between the function name and the bracket:
>
>    exact_log2 (-i)
>
>
> > +extern void abort(void);
> The same comment applies multiple places in this file.
>
> Likewise; if (
>
> Otherwise, OK, please apply with those fixes.
>
> Thanks,
> James

Thanks for noticing that, here's the fixed version.

Sam

diff --git a/gcc/config/aarch64/aarch64-protos.h b/gcc/config/aarch64/aarch64-protos.h
index aae1db45ed69c14e306ccce056861a58d9acd834..b26e46f81a414bf71762527f84fd9ac38b81b829 100644
--- a/gcc/config/aarch64/aarch64-protos.h
+++ b/gcc/config/aarch64/aarch64-protos.h
@@ -624,4 +624,6 @@ rtl_opt_pass *make_pass_tag_collision_avoidance (gcc::context *);
 
 poly_uint64 aarch64_regmode_natural_size (machine_mode);
 
+bool aarch64_high_bits_all_ones_p (HOST_WIDE_INT);
+
 #endif /* GCC_AARCH64_PROTOS_H */
diff --git a/gcc/config/aarch64/aarch64.c b/gcc/config/aarch64/aarch64.c
index d088ef7ee0f256ad0d4f59d2735121de2dd67eba..34acc58510110fbc2cb4abd19ec9d7a04bad3f4c 100644
--- a/gcc/config/aarch64/aarch64.c
+++ b/gcc/config/aarch64/aarch64.c
@@ -1432,6 +1432,13 @@ aarch64_hard_regno_caller_save_mode (unsigned regno, unsigned,
     return SImode;
 }
 
+/* Return true if I's bits are consecutive ones from the MSB.  */
+bool
+aarch64_high_bits_all_ones_p (HOST_WIDE_INT i)
+{
+  return exact_log2 (-i) != HOST_WIDE_INT_M1;
+}
+
 /* Implement TARGET_CONSTANT_ALIGNMENT.  Make strings word-aligned so
    that strcpy from constants will be faster.  */
 
diff --git a/gcc/config/aarch64/aarch64.md b/gcc/config/aarch64/aarch64.md
index 955769a64d2030839cdb337321a808626188190e..88f66104db31320389f05cdd5d161db9992a77b8 100644
--- a/gcc/config/aarch64/aarch64.md
+++ b/gcc/config/aarch64/aarch64.md
@@ -5336,6 +5336,31 @@
   [(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")
+		    (match_operand:GPI 3 "const_int_operand" "n, Ulc"))
+	    (and:GPI (match_operand:GPI 2 "register_operand" "0,r")
+		    (match_operand:GPI 4 "const_int_operand" "Ulc, n"))))]
+  "(INTVAL (operands[3]) == ~INTVAL (operands[4]))
+  && (aarch64_high_bits_all_ones_p (INTVAL (operands[3]))
+    || aarch64_high_bits_all_ones_p (INTVAL (operands[4])))"
+  {
+    switch (which_alternative)
+    {
+      case 0:
+	operands[3] = GEN_INT (ctz_hwi (~INTVAL (operands[3])));
+	return "bfxil\\t%<w>0, %<w>1, 0, %3";
+      case 1:
+	operands[3] = GEN_INT (ctz_hwi (~INTVAL (operands[4])));
+	return "bfxil\\t%<w>0, %<w>2, 0, %3";
+      default:
+	gcc_unreachable ();
+    }
+  }
+  [(set_attr "type" "bfm")]
+)
+
 ;; There are no canonicalisation rules for the position of the lshiftrt, ashift
 ;; operations within an IOR/AND RTX, therefore we have two patterns matching
 ;; each valid permutation.
diff --git a/gcc/config/aarch64/constraints.md b/gcc/config/aarch64/constraints.md
index 72cacdabdac52dcb40b480f7a5bfbf4997c742d8..31fc3eafd8bba03cc773e226223a6293c6dde8d4 100644
--- a/gcc/config/aarch64/constraints.md
+++ b/gcc/config/aarch64/constraints.md
@@ -172,6 +172,13 @@
   A constraint that matches the immediate constant -1."
   (match_test "op == constm1_rtx"))
 
+(define_constraint "Ulc"
+ "@internal
+ A constraint that matches a constant integer whose bits are consecutive ones
+ from the MSB."
+ (and (match_code "const_int")
+      (match_test "aarch64_high_bits_all_ones_p (ival)")))
+
 (define_constraint "Usv"
   "@internal
    A constraint that matches a VG-based constant that can be loaded by
diff --git a/gcc/testsuite/gcc.target/aarch64/combine_bfxil.c b/gcc/testsuite/gcc.target/aarch64/combine_bfxil.c
new file mode 100644
index 0000000000000000000000000000000000000000..adb0582ed9d8207f7b52c8912d03345369747448
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/combine_bfxil.c
@@ -0,0 +1,98 @@
+/* { dg-do run } */
+/* { dg-options "-O2 --save-temps" } */
+
+extern void abort (void);
+
+unsigned long long
+combine_balanced (unsigned long long a, unsigned long long b)
+{
+  return (a & 0xffffffff00000000ll) | (b & 0x00000000ffffffffll);
+}
+
+unsigned long long
+combine_minimal (unsigned long long a, unsigned long long b)
+{
+  return (a & 0xfffffffffffffffe) | (b & 0x0000000000000001);
+}
+
+unsigned long long
+combine_unbalanced (unsigned long long a, unsigned long long b)
+{
+  return (a & 0xffffffffff000000ll) | (b & 0x0000000000ffffffll);
+}
+
+unsigned int
+combine_balanced_int (unsigned int a, unsigned int b)
+{
+  return (a & 0xffff0000ll) | (b & 0x0000ffffll);
+}
+
+unsigned int
+combine_unbalanced_int (unsigned int a, unsigned int b)
+{
+  return (a & 0xffffff00ll) | (b & 0x000000ffll);
+}
+
+__attribute__ ((noinline)) void
+foo (unsigned long long a, unsigned long long b, unsigned long long *c,
+  unsigned long long *d)
+{
+  *c = combine_minimal (a, b);
+  *d = combine_minimal (b, a);
+}
+
+__attribute__ ((noinline)) void
+foo2 (unsigned long long a, unsigned long long b, unsigned long long *c,
+  unsigned long long *d)
+{
+  *c = combine_balanced (a, b);
+  *d = combine_balanced (b, a);
+}
+
+__attribute__ ((noinline)) void
+foo3 (unsigned long long a, unsigned long long b, unsigned long long *c,
+  unsigned long long *d)
+{
+  *c = combine_unbalanced (a, b);
+  *d = combine_unbalanced (b, a);
+}
+
+void
+foo4 (unsigned int a, unsigned int b, unsigned int *c, unsigned int *d)
+{
+  *c = combine_balanced_int (a, b);
+  *d = combine_balanced_int (b, a);
+}
+
+void
+foo5 (unsigned int a, unsigned int b, unsigned int *c, unsigned int *d)
+{
+  *c = combine_unbalanced_int (a, b);
+  *d = combine_unbalanced_int (b, a);
+}
+
+int
+main (void)
+{
+  unsigned long long a = 0x0123456789ABCDEF, b = 0xFEDCBA9876543210, c, d;
+  foo3 (a, b, &c, &d);
+  if (c != 0x0123456789543210) abort ();
+  if (d != 0xfedcba9876abcdef) abort ();
+  foo2 (a, b, &c, &d);
+  if (c != 0x0123456776543210) abort ();
+  if (d != 0xfedcba9889abcdef) abort ();
+  foo (a, b, &c, &d);
+  if (c != 0x0123456789abcdee) abort ();
+  if (d != 0xfedcba9876543211) abort ();
+
+  unsigned int a2 = 0x01234567, b2 = 0xFEDCBA98, c2, d2;
+  foo4 (a2, b2, &c2, &d2);
+  if (c2 != 0x0123ba98) abort ();
+  if (d2 != 0xfedc4567) abort ();
+  foo5 (a2, b2, &c2, &d2);
+  if (c2 != 0x01234598) abort ();
+  if (d2 != 0xfedcba67) abort ();
+  return 0;
+}
+
+/* { dg-final { scan-assembler-times "bfxil\\t" 10 } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/combine_bfxil_2.c b/gcc/testsuite/gcc.target/aarch64/combine_bfxil_2.c
new file mode 100644
index 0000000000000000000000000000000000000000..0fc140443bc67bcf12b93d72b7970e095620021e
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/combine_bfxil_2.c
@@ -0,0 +1,16 @@
+/* { dg-do compile } */
+/* { dg-options "-O2" } */
+
+unsigned long long
+combine_non_consecutive (unsigned long long a, unsigned long long b)
+{
+  return (a & 0xfffffff200f00000ll) | (b & 0x00001000ffffffffll);
+}
+
+void
+foo4 (unsigned long long a, unsigned long long b, unsigned long long *c,
+  unsigned long long *d) {
+  /* { dg-final { scan-assembler-not "bfxil\\t" } } */
+  *c = combine_non_consecutive (a, b);
+  *d = combine_non_consecutive (b, a);
+}