From d3d7a8ef09b6fa79109e8930aaba7a677f8b24ac Mon Sep 17 00:00:00 2001
From: mgrice <mgrice@fb.com>
Date: Fri, 8 Mar 2019 11:50:30 -0800
Subject: [PATCH] remove compare and branch in fast path for c++ operator
 delete[]

Summary: sdallocx is checking a flag that will never be set (at least in the provided C++ destructor implementation).  This branch will probably only rarely be mispredicted however it removes two instructions in sdallocx and one at the callsite (to zero out flags).
---
 bin/jeprof.in                                      |  1 +
 .../jemalloc/internal/jemalloc_internal_externs.h  |  1 +
 src/jemalloc.c                                     | 14 +++++++++++++-
 src/jemalloc_cpp.cpp                               |  4 ++--
 4 files changed, 17 insertions(+), 3 deletions(-)

diff --git a/bin/jeprof.in b/bin/jeprof.in
index 588c6b43..16a76c6c 100644
--- a/bin/jeprof.in
+++ b/bin/jeprof.in
@@ -2909,6 +2909,7 @@ sub RemoveUninterestingFrames {
                       '@JEMALLOC_PREFIX@xallocx',
                       '@JEMALLOC_PREFIX@dallocx',
                       '@JEMALLOC_PREFIX@sdallocx',
+                      '@JEMALLOC_PREFIX@sdallocx_noflags',
                       'tc_calloc',
                       'tc_cfree',
                       'tc_malloc',
diff --git a/include/jemalloc/internal/jemalloc_internal_externs.h b/include/jemalloc/internal/jemalloc_internal_externs.h
index b7843623..cdbc33a2 100644
--- a/include/jemalloc/internal/jemalloc_internal_externs.h
+++ b/include/jemalloc/internal/jemalloc_internal_externs.h
@@ -51,5 +51,6 @@ void jemalloc_prefork(void);
 void jemalloc_postfork_parent(void);
 void jemalloc_postfork_child(void);
 bool malloc_initialized(void);
+void je_sdallocx_noflags(void *ptr, size_t size);
 
 #endif /* JEMALLOC_INTERNAL_EXTERNS_H */
diff --git a/src/jemalloc.c b/src/jemalloc.c
index c8afa9c4..7bc7b957 100644
--- a/src/jemalloc.c
+++ b/src/jemalloc.c
@@ -2732,7 +2732,7 @@ bool free_fastpath(void *ptr, size_t size, bool size_hint) {
 	tcache_t *tcache = tsd_tcachep_get(tsd);
 
 	alloc_ctx_t alloc_ctx;
-	/* 
+	/*
 	 * If !config_cache_oblivious, we can check PAGE alignment to
 	 * detect sampled objects.  Otherwise addresses are
 	 * randomized, and we have to look it up in the rtree anyway.
@@ -3522,6 +3522,18 @@ je_sdallocx(void *ptr, size_t size, int flags) {
 	LOG("core.sdallocx.exit", "");
 }
 
+void JEMALLOC_NOTHROW
+je_sdallocx_noflags(void *ptr, size_t size) {
+	LOG("core.sdallocx.entry", "ptr: %p, size: %zu, flags: 0", ptr,
+		size);
+
+	if (!free_fastpath(ptr, size, true)) {
+		sdallocx_default(ptr, size, 0);
+	}
+
+	LOG("core.sdallocx.exit", "");
+}
+
 JEMALLOC_EXPORT size_t JEMALLOC_NOTHROW
 JEMALLOC_ATTR(pure)
 je_nallocx(size_t size, int flags) {
diff --git a/src/jemalloc_cpp.cpp b/src/jemalloc_cpp.cpp
index f0ceddae..da0441a7 100644
--- a/src/jemalloc_cpp.cpp
+++ b/src/jemalloc_cpp.cpp
@@ -128,14 +128,14 @@ operator delete(void *ptr, std::size_t size) noexcept {
 	if (unlikely(ptr == nullptr)) {
 		return;
 	}
-	je_sdallocx(ptr, size, /*flags=*/0);
+	je_sdallocx_noflags(ptr, size);
 }
 
 void operator delete[](void *ptr, std::size_t size) noexcept {
 	if (unlikely(ptr == nullptr)) {
 		return;
 	}
-	je_sdallocx(ptr, size, /*flags=*/0);
+	je_sdallocx_noflags(ptr, size);
 }
 
 #endif  // __cpp_sized_deallocation