v2-0001-Use-direct-function-calls-for-pg_popcount-32-64-o.patch

From 9d25c79719859b87d2137b0cd888d1f1a7c4d3cc Mon Sep 17 00:00:00 2001
From: John Naylor <john.naylor@postgresql.org>
Date: Wed, 11 Aug 2021 21:26:59 -0400
Subject: [PATCH v2 1/2] Use direct function calls for pg_popcount{32,64} on
 non-x86 platforms

Previously, all pg_popcount{32,64} calls were indirected through
a function pointer, even though we had no fast implementation for
non-x86 platforms. Instead, use wrappers around the
pg_popcount{32,64}_slow functions.
---
 src/include/port/pg_bitutils.h | 38 +++++++++++++++++++++-
 src/port/pg_bitutils.c         | 58 ++++++++++++----------------------
 2 files changed, 57 insertions(+), 39 deletions(-)

diff --git a/src/include/port/pg_bitutils.h b/src/include/port/pg_bitutils.h
index 086bd08132..21db364984 100644
--- a/src/include/port/pg_bitutils.h
+++ b/src/include/port/pg_bitutils.h
@@ -253,9 +253,45 @@ pg_ceil_log2_64(uint64 num)
 		return pg_leftmost_one_pos64(num - 1) + 1;
 }
 
-/* Count the number of one-bits in a uint32 or uint64 */
+/*
+ * With MSVC on x86_64 builds, try using native popcnt instructions via the
+ * __popcnt and __popcnt64 intrinsics.  These don't work the same as GCC's
+ * __builtin_popcount* intrinsic functions as they always emit popcnt
+ * instructions.
+ */
+#if defined(_MSC_VER) && defined(_M_AMD64)
+#define HAVE_X86_64_POPCNTQ
+#endif
+
+/*
+ * On x86_64, we can use the hardware popcount instruction, but only if
+ * we can verify that the CPU supports it via the cpuid instruction.
+ *
+ * Otherwise, we fall back to a hand-rolled implementation.
+ */
+#ifdef HAVE_X86_64_POPCNTQ
+#if defined(HAVE__GET_CPUID) || defined(HAVE__CPUID)
+#define TRY_POPCNT_FAST 1
+#endif
+#endif
+
+#ifdef TRY_POPCNT_FAST
+/* Use the POPCNT instruction, but perform a runtime check first. */
+extern int	pg_popcount32_slow(uint32 word);
+extern int	pg_popcount64_slow(uint64 word);
 extern int	(*pg_popcount32) (uint32 word);
 extern int	(*pg_popcount64) (uint64 word);
+extern int	pg_popcount32_fast(uint32 word);
+extern int	pg_popcount64_fast(uint64 word);
+
+#else
+/* Use a portable implementation. */
+extern int	pg_popcount32_slow(uint32 word);
+extern int	pg_popcount64_slow(uint64 word);
+extern int	pg_popcount32(uint32 word);
+extern int	pg_popcount64(uint64 word);
+
+#endif							/* TRY_POPCNT_FAST */
 
 /* Count the number of one-bits in a byte array */
 extern uint64 pg_popcount(const char *buf, int bytes);
diff --git a/src/port/pg_bitutils.c b/src/port/pg_bitutils.c
index 10676b285c..b5a62def6f 100644
--- a/src/port/pg_bitutils.c
+++ b/src/port/pg_bitutils.c
@@ -103,47 +103,13 @@ const uint8 pg_number_of_ones[256] = {
 	4, 5, 5, 6, 5, 6, 6, 7, 5, 6, 6, 7, 6, 7, 7, 8
 };
 
-/*
- * With MSVC on x86_64 builds, try using native popcnt instructions via the
- * __popcnt and __popcnt64 intrinsics.  These don't work the same as GCC's
- * __builtin_popcount* intrinsic functions as they always emit popcnt
- * instructions.
- */
-#if defined(_MSC_VER) && defined(_M_AMD64)
-#define HAVE_X86_64_POPCNTQ
-#endif
-
-/*
- * On x86_64, we can use the hardware popcount instruction, but only if
- * we can verify that the CPU supports it via the cpuid instruction.
- *
- * Otherwise, we fall back to __builtin_popcount if the compiler has that,
- * or a hand-rolled implementation if not.
- */
-#ifdef HAVE_X86_64_POPCNTQ
-#if defined(HAVE__GET_CPUID) || defined(HAVE__CPUID)
-#define TRY_POPCNT_FAST 1
-#endif
-#endif
-
-static int	pg_popcount32_slow(uint32 word);
-static int	pg_popcount64_slow(uint64 word);
-
 #ifdef TRY_POPCNT_FAST
 static bool pg_popcount_available(void);
 static int	pg_popcount32_choose(uint32 word);
 static int	pg_popcount64_choose(uint64 word);
-static int	pg_popcount32_fast(uint32 word);
-static int	pg_popcount64_fast(uint64 word);
 
 int			(*pg_popcount32) (uint32 word) = pg_popcount32_choose;
 int			(*pg_popcount64) (uint64 word) = pg_popcount64_choose;
-#else
-int			(*pg_popcount32) (uint32 word) = pg_popcount32_slow;
-int			(*pg_popcount64) (uint64 word) = pg_popcount64_slow;
-#endif							/* TRY_POPCNT_FAST */
-
-#ifdef TRY_POPCNT_FAST
 
 /*
  * Return true if CPUID indicates that the POPCNT instruction is available.
@@ -208,7 +174,7 @@ pg_popcount64_choose(uint64 word)
  * pg_popcount32_fast
  *		Return the number of 1 bits set in word
  */
-static int
+int
 pg_popcount32_fast(uint32 word)
 {
 #ifdef _MSC_VER
@@ -225,7 +191,7 @@ __asm__ __volatile__(" popcntl %1,%0\n":"=q"(res):"rm"(word):"cc");
  * pg_popcount64_fast
  *		Return the number of 1 bits set in word
  */
-static int
+int
 pg_popcount64_fast(uint64 word)
 {
 #ifdef _MSC_VER
@@ -245,7 +211,7 @@ __asm__ __volatile__(" popcntq %1,%0\n":"=q"(res):"rm"(word):"cc");
  * pg_popcount32_slow
  *		Return the number of 1 bits set in word
  */
-static int
+int
 pg_popcount32_slow(uint32 word)
 {
 #ifdef HAVE__BUILTIN_POPCOUNT
@@ -267,7 +233,7 @@ pg_popcount32_slow(uint32 word)
  * pg_popcount64_slow
  *		Return the number of 1 bits set in word
  */
-static int
+int
 pg_popcount64_slow(uint64 word)
 {
 #ifdef HAVE__BUILTIN_POPCOUNT
@@ -291,6 +257,22 @@ pg_popcount64_slow(uint64 word)
 #endif							/* HAVE__BUILTIN_POPCOUNT */
 }
 
+#ifndef TRY_POPCNT_FAST
+
+/* call these directly rather than through a function pointer */
+int
+pg_popcount32(uint32 word)
+{
+	return pg_popcount32_slow(word);
+}
+
+int
+pg_popcount64(uint64 word)
+{
+	return pg_popcount64_slow(word);
+}
+
+#endif							/* !TRY_POPCNT_FAST */
 
 /*
  * pg_popcount
-- 
2.31.1