[ARM] cache align destination pointer when copying memory for some processors

The implementation for memory copy functions on ARM had a (disabled)
provision for aligning the source pointer before loading registers with
data.  Turns out that aligning the _destination_ pointer is much more
useful, as the read side is already sufficiently helped with the use of
preload.

So this changes the definition of the CALGN() macro to target the
destination pointer instead, and turns it on for Feroceon processors
where the gain is very noticeable.

Signed-off-by: Nicolas Pitre <nico@marvell.com>
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>

3 files changed, 19 insertions, 20 deletions

diff --git a/arch/arm/lib/copy_template.S b/arch/arm/lib/copy_template.S
index cab355c0c1f7..139cce646055 100644
--- a/arch/arm/lib/copy_template.S
+++ b/arch/arm/lib/copy_template.S
@@ -13,14 +13,6 @@
  */
 
 /*
- * This can be used to enable code to cacheline align the source pointer.
- * Experiments on tested architectures (StrongARM and XScale) didn't show
- * this a worthwhile thing to do.  That might be different in the future.
- */
-//#define CALGN(code...)        code
-#define CALGN(code...)
-
-/*
  * Theory of operation
  * -------------------
  *
@@ -82,7 +74,7 @@
                 stmfd   sp!, {r5 - r8}
                 blt     5f
 
-        CALGN(  ands    ip, r1, #31             )
+        CALGN(  ands    ip, r0, #31             )
         CALGN(  rsb     r3, ip, #32             )
         CALGN(  sbcnes  r4, r3, r2              )  @ C is always set here
         CALGN(  bcs     2f                      )
@@ -168,7 +160,7 @@
                 subs    r2, r2, #28
                 blt     14f
 
-        CALGN(  ands    ip, r1, #31             )
+        CALGN(  ands    ip, r0, #31             )
         CALGN(  rsb     ip, ip, #32             )
         CALGN(  sbcnes  r4, ip, r2              )  @ C is always set here
         CALGN(  subcc   r2, r2, ip              )
diff --git a/arch/arm/lib/memmove.S b/arch/arm/lib/memmove.S
index 018522c3ff26..2e301b7bd8f1 100644
--- a/arch/arm/lib/memmove.S
+++ b/arch/arm/lib/memmove.S
@@ -13,14 +13,6 @@
 #include <linux/linkage.h>
 #include <asm/assembler.h>
 
-/*
- * This can be used to enable code to cacheline align the source pointer.
- * Experiments on tested architectures (StrongARM and XScale) didn't show
- * this a worthwhile thing to do.  That might be different in the future.
- */
-//#define CALGN(code...)        code
-#define CALGN(code...)
-
                 .text
 
 /*
@@ -55,7 +47,7 @@ ENTRY(memmove)
                 stmfd   sp!, {r5 - r8}
                 blt     5f
 
-        CALGN(  ands    ip, r1, #31             )
+        CALGN(  ands    ip, r0, #31             )
         CALGN(  sbcnes  r4, ip, r2              )  @ C is always set here
         CALGN(  bcs     2f                      )
         CALGN(  adr     r4, 6f                  )
@@ -139,7 +131,7 @@ ENTRY(memmove)
                 subs    r2, r2, #28
                 blt     14f
 
-        CALGN(  ands    ip, r1, #31             )
+        CALGN(  ands    ip, r0, #31             )
         CALGN(  sbcnes  r4, ip, r2              )  @ C is always set here
         CALGN(  subcc   r2, r2, ip              )
         CALGN(  bcc     15f                     )
diff --git a/include/asm-arm/assembler.h b/include/asm-arm/assembler.h
index fce832820825..911393b2c6f0 100644
--- a/include/asm-arm/assembler.h
+++ b/include/asm-arm/assembler.h
@@ -56,6 +56,21 @@
 #endif
 
 /*
+ * This can be used to enable code to cacheline align the destination
+ * pointer when bulk writing to memory.  Experiments on StrongARM and
+ * XScale didn't show this a worthwhile thing to do when the cache is not
+ * set to write-allocate (this would need further testing on XScale when WA
+ * is used).
+ *
+ * On Feroceon there is much to gain however, regardless of cache mode.
+ */
+#ifdef CONFIG_CPU_FEROCEON
+#define CALGN(code...) code
+#else
+#define CALGN(code...)
+#endif
+
+/*
  * Enable and disable interrupts
  */
 #if __LINUX_ARM_ARCH__ >= 6