[PATCH][GCC][Arm] Add pattern for bswap + rotate -> rev16 [Bug 108933]

[PATCH][GCC][Arm] Add pattern for bswap + rotate -> rev16 [Bug 108933]

[Matthieu Longo]

[-- Attachment #1: Type: text/plain, Size: 639 bytes --]

rev16 pattern was not recognised anymore as a change in the bswap tree
pass was introducing a new GIMPLE form, not recognized by the assembly
final transformation pass.

More details in https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108933

gcc/ChangeLog:

         PR target/108933
         * config/arm/arm.md (*arm_rev16si2_alt3): new pattern to convert
           a bswap + rotate by 16 bits into rev16

gcc/testsuite/ChangeLog:

         PR target/108933
         * gcc.target/arm/rev16.c: Moved to...
         * gcc.target/arm/rev16_1.c: ...here.
         * gcc.target/arm/rev16_2.c: New test to check that rev16 is
           emitted.

[-- Attachment #2: 108933 --]
[-- Type: text/plain, Size: 1763 bytes --]

diff --git a/gcc/config/arm/arm.md b/gcc/config/arm/arm.md
index 4a98f2d7b6251da940806b26d4c310a7f7af927b..330c0a5ce2926439760466746a68fd5f6c5b1b09 100644
--- a/gcc/config/arm/arm.md
+++ b/gcc/config/arm/arm.md
@@ -12601,6 +12601,18 @@
    (set_attr "type" "rev")]
 )
 
+;; Similar pattern to match (rotate (bswap) 16)
+(define_insn "*arm_rev16si2_alt3"
+  [(set (match_operand:SI 0 "register_operand" "=l,r")
+        (rotate:SI (bswap:SI (match_operand:SI 1 "register_operand" "l,r"))
+                     (const_int 16)))]
+  "arm_arch6"
+  "rev16\\t%0, %1"
+  [(set_attr "arch" "t,32")
+   (set_attr "length" "2,4")
+   (set_attr "type" "rev")]
+)
+
 (define_expand "arm_rev16si2"
   [(set (match_operand:SI 0 "s_register_operand")
 	(bswap:SI (match_operand:SI 1 "s_register_operand")))]
diff --git a/gcc/testsuite/gcc.target/arm/rev16.c b/gcc/testsuite/gcc.target/arm/rev16_1.c
similarity index 100%
rename from gcc/testsuite/gcc.target/arm/rev16.c
rename to gcc/testsuite/gcc.target/arm/rev16_1.c
diff --git a/gcc/testsuite/gcc.target/arm/rev16_2.c b/gcc/testsuite/gcc.target/arm/rev16_2.c
new file mode 100644
index 0000000000000000000000000000000000000000..90213f9b49f45340ced4f29c31446971dbc88ecd
--- /dev/null
+++ b/gcc/testsuite/gcc.target/arm/rev16_2.c
@@ -0,0 +1,20 @@
+/* { dg-options "-O2" } */
+/* { dg-do compile } */
+
+typedef unsigned int __u32;
+
+__u32
+__rev16_32_alt (__u32 x)
+{
+  return (((__u32)(x) & (__u32)0xff00ff00UL) >> 8)
+         | (((__u32)(x) & (__u32)0x00ff00ffUL) << 8);
+}
+
+__u32
+__rev16_32 (__u32 x)
+{
+  return (((__u32)(x) & (__u32)0x00ff00ffUL) << 8)
+         | (((__u32)(x) & (__u32)0xff00ff00UL) >> 8);
+}
+
+/* { dg-final { scan-assembler-times {rev16\tr[0-9]+, r[0-9]+} 2 } } */
\ No newline at end of file

Re: [PATCH][GCC][Arm] Add pattern for bswap + rotate -> rev16 [Bug 108933]

[Richard Earnshaw (lists)]

On 22/01/2024 12:18, Matthieu Longo wrote:
> rev16 pattern was not recognised anymore as a change in the bswap tree
> pass was introducing a new GIMPLE form, not recognized by the assembly
> final transformation pass.
> 
> More details in https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108933
> 
> gcc/ChangeLog:
> 
>         PR target/108933
>         * config/arm/arm.md (*arm_rev16si2_alt3): new pattern to convert
>           a bswap + rotate by 16 bits into rev16

ChangeLog entries need to be written as sentences, so start with a capital letter and end with a full stop; continuation lines should start in column 8 (one hard tab, don't use spaces).  But in this case, "New pattern." is sufficient.

> 
> gcc/testsuite/ChangeLog:
> 
>         PR target/108933
>         * gcc.target/arm/rev16.c: Moved to...
>         * gcc.target/arm/rev16_1.c: ...here.
>         * gcc.target/arm/rev16_2.c: New test to check that rev16 is
>           emitted.


+;; Similar pattern to match (rotate (bswap) 16)
+(define_insn "*arm_rev16si2_alt3"
+  [(set (match_operand:SI 0 "register_operand" "=l,r")
+        (rotate:SI (bswap:SI (match_operand:SI 1 "register_operand" "l,r"))
+                     (const_int 16)))]
+  "arm_arch6"
+  "rev16\\t%0, %1"
+  [(set_attr "arch" "t,32")
+   (set_attr "length" "2,4")
+   (set_attr "type" "rev")]
+)
+

Unfortunately, this is insufficient.  When generating Arm or Thumb2 code (but not thumb1) we also have to handle conditional execution: we need to have '%?' in the output template at the point where a condition code might be needed.  That means we need separate output templates for all three alternatives (as we need a 16-bit variant for thumb2 that's conditional and a 16-bit for thumb1 that isn't).  See the output of arm_rev16 for a guide of what is really needed.

I note that the arm_rev16si2_alt1, and arm_rev16si2_alt2 patterns are incorrect in this regard as well; that will need fixing.

I also see that arm_rev16si2 currently expands to the alt1 variant above; given that the preferred canonical form would now appear to use bswap + rotate, we should change that as well.  In fact, we can merge your new pattern with the expand entirely and eliminate the need to call gen_arm_rev16si2_alt1.  Something like:

(define_insn "arm_rev16si2"
  [(set (match_operand:SI 0 "s_register_operand")
        (rotate:SI (bswap:SI (match_operand:SI 1 "s_register_operand")) (const_int 16))]
  "arm_arch6"
  "@
  rev16...
  ...


R.

Re: [PATCH][GCC][Arm] Add pattern for bswap + rotate -> rev16 [Bug 108933]

[Matthieu Longo]

[-- Attachment #1: Type: text/plain, Size: 3431 bytes --]

Hi Richard,

Please find below the new patch where I addressed your comments and 
updated the changelog.

rev16 pattern was not recognised anymore as a change in the bswap tree
pass was introducing a new GIMPLE form, not recognized by the assembly
final transformation pass.

More details in https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108933

gcc/ChangeLog:

	PR target/108933
	* config/arm/arm.md (arm_rev16si2): Convert to define_insn.
	Correct generated RTL.
	(arm_rev16si2_alt1): Correctly handle conditional execution.
	(arm_rev16si2_alt2): Likewise.

gcc/testsuite/ChangeLog:

	PR target/108933
	* gcc.target/arm/rev16.c: Moved to...
	* gcc.target/arm/rev16_1.c: ...here.
	* gcc.target/arm/rev16_2.c: New test to check that rev16 is
	emitted.

On 2024-01-22 16:25, Richard Earnshaw (lists) wrote:
> On 22/01/2024 12:18, Matthieu Longo wrote:
>> rev16 pattern was not recognised anymore as a change in the bswap tree
>> pass was introducing a new GIMPLE form, not recognized by the assembly
>> final transformation pass.
>>
>> More details in https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108933
>>
>> gcc/ChangeLog:
>>
>>          PR target/108933
>>          * config/arm/arm.md (*arm_rev16si2_alt3): new pattern to convert
>>            a bswap + rotate by 16 bits into rev16
> 
> ChangeLog entries need to be written as sentences, so start with a capital letter and end with a full stop; continuation lines should start in column 8 (one hard tab, don't use spaces).  But in this case, "New pattern." is sufficient.
> 
>>
>> gcc/testsuite/ChangeLog:
>>
>>          PR target/108933
>>          * gcc.target/arm/rev16.c: Moved to...
>>          * gcc.target/arm/rev16_1.c: ...here.
>>          * gcc.target/arm/rev16_2.c: New test to check that rev16 is
>>            emitted.
> 
> 
> +;; Similar pattern to match (rotate (bswap) 16)
> +(define_insn "*arm_rev16si2_alt3"
> +  [(set (match_operand:SI 0 "register_operand" "=l,r")
> +        (rotate:SI (bswap:SI (match_operand:SI 1 "register_operand" "l,r"))
> +                     (const_int 16)))]
> +  "arm_arch6"
> +  "rev16\\t%0, %1"
> +  [(set_attr "arch" "t,32")
> +   (set_attr "length" "2,4")
> +   (set_attr "type" "rev")]
> +)
> +
> 
> Unfortunately, this is insufficient.  When generating Arm or Thumb2 code (but not thumb1) we also have to handle conditional execution: we need to have '%?' in the output template at the point where a condition code might be needed.  That means we need separate output templates for all three alternatives (as we need a 16-bit variant for thumb2 that's conditional and a 16-bit for thumb1 that isn't).  See the output of arm_rev16 for a guide of what is really needed.
> 
> I note that the arm_rev16si2_alt1, and arm_rev16si2_alt2 patterns are incorrect in this regard as well; that will need fixing.
> 
> I also see that arm_rev16si2 currently expands to the alt1 variant above; given that the preferred canonical form would now appear to use bswap + rotate, we should change that as well.  In fact, we can merge your new pattern with the expand entirely and eliminate the need to call gen_arm_rev16si2_alt1.  Something like:
> 
> (define_insn "arm_rev16si2"
>    [(set (match_operand:SI 0 "s_register_operand")
>          (rotate:SI (bswap:SI (match_operand:SI 1 "s_register_operand")) (const_int 16))]
>    "arm_arch6"
>    "@
>    rev16...
>    ...
> 
> 
> R.
> 

[-- Attachment #2: 108933 --]
[-- Type: text/plain, Size: 2781 bytes --]

diff --git a/gcc/config/arm/arm.md b/gcc/config/arm/arm.md
index 4a98f2d7b6251da940806b26d4c310a7f7af927b..5816409f86f1106b410c5e21d77e599b485f85f2 100644
--- a/gcc/config/arm/arm.md
+++ b/gcc/config/arm/arm.md
@@ -12578,7 +12578,10 @@ (define_insn "arm_rev16si2_alt1"
   "arm_arch6
    && aarch_rev16_shleft_mask_imm_p (operands[3], SImode)
    && aarch_rev16_shright_mask_imm_p (operands[2], SImode)"
-  "rev16\\t%0, %1"
+  "@
+   rev16\t%0, %1
+   rev16%?\t%0, %1
+   rev16%?\t%0, %1"
   [(set_attr "arch" "t1,t2,32")
    (set_attr "length" "2,2,4")
    (set_attr "type" "rev")]
@@ -12595,22 +12598,28 @@ (define_insn "*arm_rev16si2_alt2"
   "arm_arch6
    && aarch_rev16_shleft_mask_imm_p (operands[3], SImode)
    && aarch_rev16_shright_mask_imm_p (operands[2], SImode)"
-  "rev16\\t%0, %1"
+  "@
+   rev16\t%0, %1
+   rev16%?\t%0, %1
+   rev16%?\t%0, %1"
   [(set_attr "arch" "t1,t2,32")
    (set_attr "length" "2,2,4")
    (set_attr "type" "rev")]
 )
 
-(define_expand "arm_rev16si2"
-  [(set (match_operand:SI 0 "s_register_operand")
-	(bswap:SI (match_operand:SI 1 "s_register_operand")))]
+;; Similar pattern to match (rotate (bswap) 16)
+(define_insn "arm_rev16si2"
+  [(set (match_operand:SI 0 "register_operand" "=l,l,r")
+        (rotate:SI (bswap:SI (match_operand:SI 1 "register_operand" "l,l,r"))
+                   (const_int 16)))]
   "arm_arch6"
-  {
-    rtx left = gen_int_mode (HOST_WIDE_INT_C (0xff00ff00ff00ff00), SImode);
-    rtx right = gen_int_mode (HOST_WIDE_INT_C (0xff00ff00ff00ff), SImode);
-    emit_insn (gen_arm_rev16si2_alt1 (operands[0], operands[1], right, left));
-    DONE;
-  }
+  "@
+   rev16\t%0, %1
+   rev16%?\t%0, %1
+   rev16%?\t%0, %1"
+  [(set_attr "arch" "t1,t2,32")
+   (set_attr "length" "2,2,4")
+   (set_attr "type" "rev")]
 )
 
 (define_expand "bswaphi2"
diff --git a/gcc/testsuite/gcc.target/arm/rev16.c b/gcc/testsuite/gcc.target/arm/rev16_1.c
similarity index 100%
rename from gcc/testsuite/gcc.target/arm/rev16.c
rename to gcc/testsuite/gcc.target/arm/rev16_1.c
diff --git a/gcc/testsuite/gcc.target/arm/rev16_2.c b/gcc/testsuite/gcc.target/arm/rev16_2.c
new file mode 100644
index 0000000000000000000000000000000000000000..90213f9b49f45340ced4f29c31446971dbc88ecd
--- /dev/null
+++ b/gcc/testsuite/gcc.target/arm/rev16_2.c
@@ -0,0 +1,20 @@
+/* { dg-options "-O2" } */
+/* { dg-do compile } */
+
+typedef unsigned int __u32;
+
+__u32
+__rev16_32_alt (__u32 x)
+{
+  return (((__u32)(x) & (__u32)0xff00ff00UL) >> 8)
+         | (((__u32)(x) & (__u32)0x00ff00ffUL) << 8);
+}
+
+__u32
+__rev16_32 (__u32 x)
+{
+  return (((__u32)(x) & (__u32)0x00ff00ffUL) << 8)
+         | (((__u32)(x) & (__u32)0xff00ff00UL) >> 8);
+}
+
+/* { dg-final { scan-assembler-times {rev16\tr[0-9]+, r[0-9]+} 2 } } */
\ No newline at end of file

Re: [PATCH][GCC][Arm] Add pattern for bswap + rotate -> rev16 [Bug 108933]

[Richard Earnshaw]

On 29/01/2024 14:14, Matthieu Longo wrote:
> Hi Richard,
> 
> Please find below the new patch where I addressed your comments and 
> updated the changelog.
> 
> rev16 pattern was not recognised anymore as a change in the bswap tree
> pass was introducing a new GIMPLE form, not recognized by the assembly
> final transformation pass.
> 
> More details in https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108933
> 
> gcc/ChangeLog:
> 
>      PR target/108933
>      * config/arm/arm.md (arm_rev16si2): Convert to define_insn.
>      Correct generated RTL.
>      (arm_rev16si2_alt1): Correctly handle conditional execution.
>      (arm_rev16si2_alt2): Likewise.
> 
> gcc/testsuite/ChangeLog:
> 
>      PR target/108933
>      * gcc.target/arm/rev16.c: Moved to...
>      * gcc.target/arm/rev16_1.c: ...here.
>      * gcc.target/arm/rev16_2.c: New test to check that rev16 is
>      emitted.

Thanks.  I've tweaked the commit message very slightly and pushed this.

Could you please prepare backports for gcc-11 thru 13?  It should just 
be a matter of cherry-picking the commit.

R.

> 
> On 2024-01-22 16:25, Richard Earnshaw (lists) wrote:
>> On 22/01/2024 12:18, Matthieu Longo wrote:
>>> rev16 pattern was not recognised anymore as a change in the bswap tree
>>> pass was introducing a new GIMPLE form, not recognized by the assembly
>>> final transformation pass.
>>>
>>> More details in https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108933
>>>
>>> gcc/ChangeLog:
>>>
>>>          PR target/108933
>>>          * config/arm/arm.md (*arm_rev16si2_alt3): new pattern to 
>>> convert
>>>            a bswap + rotate by 16 bits into rev16
>>
>> ChangeLog entries need to be written as sentences, so start with a 
>> capital letter and end with a full stop; continuation lines should 
>> start in column 8 (one hard tab, don't use spaces).  But in this case, 
>> "New pattern." is sufficient.
>>
>>>
>>> gcc/testsuite/ChangeLog:
>>>
>>>          PR target/108933
>>>          * gcc.target/arm/rev16.c: Moved to...
>>>          * gcc.target/arm/rev16_1.c: ...here.
>>>          * gcc.target/arm/rev16_2.c: New test to check that rev16 is
>>>            emitted.
>>
>>
>> +;; Similar pattern to match (rotate (bswap) 16)
>> +(define_insn "*arm_rev16si2_alt3"
>> +  [(set (match_operand:SI 0 "register_operand" "=l,r")
>> +        (rotate:SI (bswap:SI (match_operand:SI 1 "register_operand" 
>> "l,r"))
>> +                     (const_int 16)))]
>> +  "arm_arch6"
>> +  "rev16\\t%0, %1"
>> +  [(set_attr "arch" "t,32")
>> +   (set_attr "length" "2,4")
>> +   (set_attr "type" "rev")]
>> +)
>> +
>>
>> Unfortunately, this is insufficient.  When generating Arm or Thumb2 
>> code (but not thumb1) we also have to handle conditional execution: we 
>> need to have '%?' in the output template at the point where a 
>> condition code might be needed.  That means we need separate output 
>> templates for all three alternatives (as we need a 16-bit variant for 
>> thumb2 that's conditional and a 16-bit for thumb1 that isn't).  See 
>> the output of arm_rev16 for a guide of what is really needed.
>>
>> I note that the arm_rev16si2_alt1, and arm_rev16si2_alt2 patterns are 
>> incorrect in this regard as well; that will need fixing.
>>
>> I also see that arm_rev16si2 currently expands to the alt1 variant 
>> above; given that the preferred canonical form would now appear to use 
>> bswap + rotate, we should change that as well.  In fact, we can merge 
>> your new pattern with the expand entirely and eliminate the need to 
>> call gen_arm_rev16si2_alt1.  Something like:
>>
>> (define_insn "arm_rev16si2"
>>    [(set (match_operand:SI 0 "s_register_operand")
>>          (rotate:SI (bswap:SI (match_operand:SI 1 
>> "s_register_operand")) (const_int 16))]
>>    "arm_arch6"
>>    "@
>>    rev16...
>>    ...
>>
>>
>> R.
>>