[MIPS] Improve branch prediction in ll/sc atomic operations.
Now that finally all supported versions of binutils have functioning
support for .subsection use .subsection to tweak the branch prediction
I did not modify the R10000 errata variants because it seems unclear if
this will invalidate the workaround which actually relies on the cheesy
prediction of branch likely to cause a misspredict if the sc was
successful.
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
diff --git a/include/asm-mips/atomic.h b/include/asm-mips/atomic.h
index c1a2409..8578869 100644
--- a/include/asm-mips/atomic.h
+++ b/include/asm-mips/atomic.h
@@ -69,7 +69,10 @@
"1: ll %0, %1 # atomic_add \n"
" addu %0, %2 \n"
" sc %0, %1 \n"
- " beqz %0, 1b \n"
+ " beqz %0, 2f \n"
+ " .subsection 2 \n"
+ "2: b 1b \n"
+ " .previous \n"
" .set mips0 \n"
: "=&r" (temp), "=m" (v->counter)
: "Ir" (i), "m" (v->counter));
@@ -111,7 +114,10 @@
"1: ll %0, %1 # atomic_sub \n"
" subu %0, %2 \n"
" sc %0, %1 \n"
- " beqz %0, 1b \n"
+ " beqz %0, 2f \n"
+ " .subsection 2 \n"
+ "2: b 1b \n"
+ " .previous \n"
" .set mips0 \n"
: "=&r" (temp), "=m" (v->counter)
: "Ir" (i), "m" (v->counter));
@@ -155,8 +161,11 @@
"1: ll %1, %2 # atomic_add_return \n"
" addu %0, %1, %3 \n"
" sc %0, %2 \n"
- " beqz %0, 1b \n"
+ " beqz %0, 2f \n"
" addu %0, %1, %3 \n"
+ " .subsection 2 \n"
+ "2: b 1b \n"
+ " .previous \n"
" .set mips0 \n"
: "=&r" (result), "=&r" (temp), "=m" (v->counter)
: "Ir" (i), "m" (v->counter)
@@ -204,8 +213,11 @@
"1: ll %1, %2 # atomic_sub_return \n"
" subu %0, %1, %3 \n"
" sc %0, %2 \n"
- " beqz %0, 1b \n"
+ " beqz %0, 2f \n"
" subu %0, %1, %3 \n"
+ " .subsection 2 \n"
+ "2: b 1b \n"
+ " .previous \n"
" .set mips0 \n"
: "=&r" (result), "=&r" (temp), "=m" (v->counter)
: "Ir" (i), "m" (v->counter)
@@ -267,10 +279,13 @@
" bltz %0, 1f \n"
" sc %0, %2 \n"
" .set noreorder \n"
- " beqz %0, 1b \n"
+ " beqz %0, 2f \n"
" subu %0, %1, %3 \n"
" .set reorder \n"
"1: \n"
+ " .subsection 2 \n"
+ "2: b 1b \n"
+ " .previous \n"
" .set mips0 \n"
: "=&r" (result), "=&r" (temp), "=m" (v->counter)
: "Ir" (i), "m" (v->counter)
@@ -429,7 +444,10 @@
"1: lld %0, %1 # atomic64_add \n"
" addu %0, %2 \n"
" scd %0, %1 \n"
- " beqz %0, 1b \n"
+ " beqz %0, 2f \n"
+ " .subsection 2 \n"
+ "2: b 1b \n"
+ " .previous \n"
" .set mips0 \n"
: "=&r" (temp), "=m" (v->counter)
: "Ir" (i), "m" (v->counter));
@@ -471,7 +489,10 @@
"1: lld %0, %1 # atomic64_sub \n"
" subu %0, %2 \n"
" scd %0, %1 \n"
- " beqz %0, 1b \n"
+ " beqz %0, 2f \n"
+ " .subsection 2 \n"
+ "2: b 1b \n"
+ " .previous \n"
" .set mips0 \n"
: "=&r" (temp), "=m" (v->counter)
: "Ir" (i), "m" (v->counter));
@@ -515,8 +536,11 @@
"1: lld %1, %2 # atomic64_add_return \n"
" addu %0, %1, %3 \n"
" scd %0, %2 \n"
- " beqz %0, 1b \n"
+ " beqz %0, 2f \n"
" addu %0, %1, %3 \n"
+ " .subsection 2 \n"
+ "2: b 1b \n"
+ " .previous \n"
" .set mips0 \n"
: "=&r" (result), "=&r" (temp), "=m" (v->counter)
: "Ir" (i), "m" (v->counter)
@@ -564,8 +588,11 @@
"1: lld %1, %2 # atomic64_sub_return \n"
" subu %0, %1, %3 \n"
" scd %0, %2 \n"
- " beqz %0, 1b \n"
+ " beqz %0, 2f \n"
" subu %0, %1, %3 \n"
+ " .subsection 2 \n"
+ "2: b 1b \n"
+ " .previous \n"
" .set mips0 \n"
: "=&r" (result), "=&r" (temp), "=m" (v->counter)
: "Ir" (i), "m" (v->counter)
@@ -627,10 +654,13 @@
" bltz %0, 1f \n"
" scd %0, %2 \n"
" .set noreorder \n"
- " beqz %0, 1b \n"
+ " beqz %0, 2f \n"
" dsubu %0, %1, %3 \n"
" .set reorder \n"
"1: \n"
+ " .subsection 2 \n"
+ "2: b 1b \n"
+ " .previous \n"
" .set mips0 \n"
: "=&r" (result), "=&r" (temp), "=m" (v->counter)
: "Ir" (i), "m" (v->counter)