[PATCH 0/2] arch_hweight implementation for MIPS

[PATCH 0/2] arch_hweight implementation for MIPS

[David Daney]

Recently the core hweight implementation was reworked to make it
easier for an architecture to implement optimized hweight()
operations.

For MIPS, I get lazy and fall back to letting GCC emit the optimized
code if it can.  If ARCH_HAS_USABLE_BUILTIN_POPCOUNT is defined, we
let GCC do it, otherwise, the fallback library call is used.

The second patch turns this on for OCTEON as it has POP and DPOP
instructions that GCC knows about.

David Daney (2):
  MIPS: Create and use asm/arch_hweight.h
  MIPS: Octeon: Define ARCH_HAS_USABLE_BUILTIN_POPCOUNT for OCTEON.

 arch/mips/include/asm/arch_hweight.h               |   38 ++++++++++++++++++++
 arch/mips/include/asm/bitops.h                     |    5 ++-
 .../asm/mach-cavium-octeon/cpu-feature-overrides.h |    8 ++++
 3 files changed, 50 insertions(+), 1 deletions(-)
 create mode 100644 arch/mips/include/asm/arch_hweight.h

[PATCH 1/2] MIPS: Create and use asm/arch_hweight.h

[David Daney]

Some MIPS ISA processor varients can do hweight operations
efficiently.

Split arch_hweight.h into a seperate file, and implement the
operations with __builtin_popcount{,ll} if supported.

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
---
 arch/mips/include/asm/arch_hweight.h |   38 ++++++++++++++++++++++++++++++++++
 arch/mips/include/asm/bitops.h       |    5 +++-
 2 files changed, 42 insertions(+), 1 deletions(-)
 create mode 100644 arch/mips/include/asm/arch_hweight.h

diff --git a/arch/mips/include/asm/arch_hweight.h b/arch/mips/include/asm/arch_hweight.h
new file mode 100644
index 0000000..712a744
--- /dev/null
+++ b/arch/mips/include/asm/arch_hweight.h
@@ -0,0 +1,38 @@
+/*
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License.  See the file "COPYING" in the main directory of this archive
+ * for more details.
+ *
+ */
+#ifndef _ASM_ARCH_HWEIGHT_H
+#define _ASM_ARCH_HWEIGHT_H
+
+#ifdef ARCH_HAS_USABLE_BUILTIN_POPCOUNT
+
+#include <asm/types.h>
+
+static inline unsigned int __arch_hweight32(unsigned int w)
+{
+	return __builtin_popcount(w);
+}
+
+static inline unsigned int __arch_hweight16(unsigned int w)
+{
+	return __builtin_popcount(w & 0xffff);
+}
+
+static inline unsigned int __arch_hweight8(unsigned int w)
+{
+	return __builtin_popcount(w & 0xff);
+}
+
+static inline unsigned long __arch_hweight64(__u64 w)
+{
+	return __builtin_popcountll(w);
+}
+
+#else
+#include <asm-generic/bitops/arch_hweight.h>
+#endif
+
+#endif /* _ASM_ARCH_HWEIGHT_H */
diff --git a/arch/mips/include/asm/bitops.h b/arch/mips/include/asm/bitops.h
index 9255cfb..b0ce7ca 100644
--- a/arch/mips/include/asm/bitops.h
+++ b/arch/mips/include/asm/bitops.h
@@ -700,7 +700,10 @@ static inline int ffs(int word)
 #ifdef __KERNEL__
 
 #include <asm-generic/bitops/sched.h>
-#include <asm-generic/bitops/hweight.h>
+
+#include <asm/arch_hweight.h>
+#include <asm-generic/bitops/const_hweight.h>
+
 #include <asm-generic/bitops/ext2-non-atomic.h>
 #include <asm-generic/bitops/ext2-atomic.h>
 #include <asm-generic/bitops/minix.h>
-- 
1.6.6.1

[PATCH 2/2] MIPS: Octeon: Define ARCH_HAS_USABLE_BUILTIN_POPCOUNT for OCTEON.

[David Daney]

OCTEON implements __builtin_popcount with a single instruction, so
lets use it.

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
---
 .../asm/mach-cavium-octeon/cpu-feature-overrides.h |    8 ++++++++
 1 files changed, 8 insertions(+), 0 deletions(-)

diff --git a/arch/mips/include/asm/mach-cavium-octeon/cpu-feature-overrides.h b/arch/mips/include/asm/mach-cavium-octeon/cpu-feature-overrides.h
index e412777..b952fc7 100644
--- a/arch/mips/include/asm/mach-cavium-octeon/cpu-feature-overrides.h
+++ b/arch/mips/include/asm/mach-cavium-octeon/cpu-feature-overrides.h
@@ -66,6 +66,14 @@
 #define spin_lock_prefetch(x) prefetch(x)
 #define PREFETCH_STRIDE 128
 
+#ifdef __OCTEON__
+/*
+ * All gcc versions that have OCTEON support define __OCTEON__ and have the
+ *  __builtin_popcount support.
+ */
+#define ARCH_HAS_USABLE_BUILTIN_POPCOUNT 1
+#endif
+
 static inline int octeon_has_saa(void)
 {
 	int id;
-- 
1.6.6.1

Re: [PATCH 1/2] MIPS: Create and use asm/arch_hweight.h

[Ralf Baechle]

Thanks, queued for 2.6.36.

  Ralf

Re: [PATCH 2/2] MIPS: Octeon: Define ARCH_HAS_USABLE_BUILTIN_POPCOUNT for OCTEON.

[Ralf Baechle]

Thanks, queued for 2.6.36.

  Ralf