public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
From: Kugan <kugan.vivekanandarajah@linaro.org>
To: Richard Earnshaw <Richard.Earnshaw@foss.arm.com>,
	 "gcc-patches@gcc.gnu.org" <gcc-patches@gcc.gnu.org>
Cc: Ramana Radhakrishnan <ramrad01@arm.com>,
	 Richard Earnshaw <Richard.Earnshaw@arm.com>,
	Kyrill Tkachov <kyrylo.tkachov@arm.com>
Subject: Re: [ARM][PR65768] Keep constants in register when expanding
Date: Sun, 26 Apr 2015 23:08:00 -0000	[thread overview]
Message-ID: <553D6FCD.8010808@linaro.org> (raw)
In-Reply-To: <55326FC2.3060904@foss.arm.com>

[-- Attachment #1: Type: text/plain, Size: 1424 bytes --]


> While I support your goals, I think your approach needs some refinement.
> 
> In particular, we DO NOT want another function that starts looking at
> constant values and tries to decide, on a case by case basis, what to do
> with that value.  We need to keep the logic for that, as much as
> possible, in one small set of functions so that the compiler cannot end
> up with conflicting decisions coming from different bits of code.
> 
> So const_ok_for_split has to go.  Instead you should be using
> const_ok_for op (an existing routine) and a simple macro that
> encapsulates "optimize >= 2 && can_create_pseudo_p ()" as the gate for
> when to use a separate scratch register.
> 

Thanks for the review. Please find the patch attached that does this.
Regression tested on qemu for arm-none-linux-gnueab with no new
regressions. Bootstrapped successfully on chromebook. Is this OK for trunk?

Thanks,
Kugan


gcc/ChangeLog:

2015-04-26  Kugan Vivekanandarajah  <kuganv@linaro.org>
	    Zhenqiang Chen  <zhenqiang.chen@linaro.org>

	PR target/65768
	* config/arm/arm.h (CONST_OK_FOR_SPLIT): New macro.
	* config/arm/arm.md (subsi3, andsi3, iorsi3, xorsi3, movsi): Keep some
	 large constants in register instead of splitting them.

gcc/testsuite/ChangeLog:

2015-04-26  Kugan Vivekanandarajah  <kuganv@linaro.org>
	    Zhenqiang Chen  <zhenqiang.chen@linaro.org>

	PR target/65768
	* gcc.target/arm/maskdata.c: New test.

[-- Attachment #2: arm_split2.txt --]
[-- Type: text/plain, Size: 4795 bytes --]

diff --git a/gcc/config/arm/arm.h b/gcc/config/arm/arm.h
index 8c10ea3..89f0a52 100644
--- a/gcc/config/arm/arm.h
+++ b/gcc/config/arm/arm.h
@@ -389,6 +389,11 @@ extern void (*arm_lang_output_object_attributes_hook)(void);
 /* Should NEON be used for 64-bits bitops.  */
 #define TARGET_PREFER_NEON_64BITS (prefer_neon_for_64bits)
 
+/* Should constant I be slplit for OP.  */
+#define CONST_OK_FOR_SPLIT(i, op)	(optimize < 2 \
+					 || !can_create_pseudo_p () \
+					 || const_ok_for_op (i, op))
+
 /* True iff the full BPABI is being used.  If TARGET_BPABI is true,
    then TARGET_AAPCS_BASED must be true -- but the converse does not
    hold.  TARGET_BPABI implies the use of the BPABI runtime library,
diff --git a/gcc/config/arm/arm.md b/gcc/config/arm/arm.md
index 164ac13..b499885 100644
--- a/gcc/config/arm/arm.md
+++ b/gcc/config/arm/arm.md
@@ -1164,10 +1164,16 @@
     {
       if (TARGET_32BIT)
         {
-          arm_split_constant (MINUS, SImode, NULL_RTX,
-	                      INTVAL (operands[1]), operands[0],
-	  		      operands[2], optimize && can_create_pseudo_p ());
-          DONE;
+	  if (!CONST_OK_FOR_SPLIT (INTVAL (operands[1]), MINUS))
+	    operands[1] = force_reg (SImode, operands[1]);
+	  else
+	    {
+	      arm_split_constant (MINUS, SImode, NULL_RTX,
+				  INTVAL (operands[1]), operands[0],
+				  operands[2],
+				  optimize && can_create_pseudo_p ());
+	      DONE;
+	    }
 	}
       else /* TARGET_THUMB1 */
         operands[1] = force_reg (SImode, operands[1]);
@@ -2078,14 +2084,19 @@
 	      operands[1] = convert_to_mode (QImode, operands[1], 1);
 	      emit_insn (gen_thumb2_zero_extendqisi2_v6 (operands[0],
 							 operands[1]));
+	      DONE;
 	    }
+	  else if (!CONST_OK_FOR_SPLIT (INTVAL (operands[2]), AND))
+	    operands[2] = force_reg (SImode, operands[2]);
 	  else
-	    arm_split_constant (AND, SImode, NULL_RTX,
-				INTVAL (operands[2]), operands[0],
-				operands[1],
-				optimize && can_create_pseudo_p ());
+	    {
+	      arm_split_constant (AND, SImode, NULL_RTX,
+				  INTVAL (operands[2]), operands[0],
+				  operands[1],
+				  optimize && can_create_pseudo_p ());
 
-          DONE;
+	      DONE;
+	    }
         }
     }
   else /* TARGET_THUMB1 */
@@ -2884,10 +2895,16 @@
     {
       if (TARGET_32BIT)
         {
-          arm_split_constant (IOR, SImode, NULL_RTX,
-	                      INTVAL (operands[2]), operands[0], operands[1],
-			      optimize && can_create_pseudo_p ());
-          DONE;
+	  if (!CONST_OK_FOR_SPLIT (INTVAL (operands[2]), IOR))
+	    operands[2] = force_reg (SImode, operands[2]);
+	  else
+	    {
+	      arm_split_constant (IOR, SImode, NULL_RTX,
+				  INTVAL (operands[2]), operands[0],
+				  operands[1],
+				  optimize && can_create_pseudo_p ());
+	      DONE;
+	    }
 	}
       else /* TARGET_THUMB1 */
         {
@@ -3054,10 +3071,16 @@
     {
       if (TARGET_32BIT)
         {
-          arm_split_constant (XOR, SImode, NULL_RTX,
-	                      INTVAL (operands[2]), operands[0], operands[1],
-			      optimize && can_create_pseudo_p ());
-          DONE;
+	  if (!CONST_OK_FOR_SPLIT (INTVAL (operands[2]), XOR))
+	    operands[2] = force_reg (SImode, operands[2]);
+	  else
+	    {
+	      arm_split_constant (XOR, SImode, NULL_RTX,
+				  INTVAL (operands[2]), operands[0],
+				  operands[1],
+				  optimize && can_create_pseudo_p ());
+	      DONE;
+	    }
 	}
       else /* TARGET_THUMB1 */
         {
@@ -5554,10 +5577,18 @@
           && !(const_ok_for_arm (INTVAL (operands[1]))
                || const_ok_for_arm (~INTVAL (operands[1]))))
         {
-           arm_split_constant (SET, SImode, NULL_RTX,
-	                       INTVAL (operands[1]), operands[0], NULL_RTX,
-			       optimize && can_create_pseudo_p ());
-          DONE;
+	   if (!CONST_OK_FOR_SPLIT (INTVAL (operands[1]), SET))
+	     {
+		emit_insn (gen_rtx_SET (VOIDmode, operands[0], operands[1]));
+		DONE;
+	     }
+	  else
+	     {
+		arm_split_constant (SET, SImode, NULL_RTX,
+	                            INTVAL (operands[1]), operands[0], NULL_RTX,
+			            optimize && can_create_pseudo_p ());
+		DONE;
+	     }
         }
     }
   else /* TARGET_THUMB1...  */
diff --git a/gcc/testsuite/gcc.target/arm/maskdata.c b/gcc/testsuite/gcc.target/arm/maskdata.c
index e69de29..6d6bb39 100644
--- a/gcc/testsuite/gcc.target/arm/maskdata.c
+++ b/gcc/testsuite/gcc.target/arm/maskdata.c
@@ -0,0 +1,15 @@
+/* { dg-do compile } */
+/* { dg-options " -O2 -fno-gcse " }  */
+/* { dg-require-effective-target arm_thumb2_ok } */
+
+#define MASK 0xff00ff
+void maskdata (int * data, int len)
+{
+   int i = len;
+   for (; i > 0; i -= 2)
+    {
+      data[i] &= MASK;
+      data[i + 1] &= MASK;
+    }
+}
+/* { dg-final { scan-assembler-not "65280" } } */

  reply	other threads:[~2015-04-26 23:08 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-04-15  7:48 Kugan
2015-04-15  8:21 ` Kyrill Tkachov
2015-04-15  8:28   ` Kugan
2015-04-18 14:52 ` Richard Earnshaw
2015-04-26 23:08   ` Kugan [this message]
2015-05-14  5:47     ` Kugan
2015-05-14  9:26     ` Richard Earnshaw

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=553D6FCD.8010808@linaro.org \
    --to=kugan.vivekanandarajah@linaro.org \
    --cc=Richard.Earnshaw@arm.com \
    --cc=Richard.Earnshaw@foss.arm.com \
    --cc=gcc-patches@gcc.gnu.org \
    --cc=kyrylo.tkachov@arm.com \
    --cc=ramrad01@arm.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).