Export of internal Abseil changes

--
e7f53dfbf809812e84770217777f81b6308a3084 by Abseil Team <absl-team@google.com>:

Add a parameter pack to absl profile to allow profiles to separate
dynamic data from static data that is available at constructor-time.

Background: `inline_element_size` is effectively constant, but there
is a data race between its initialization and its access. We had fixed that race by making
inline_element_size atomic.  This CL changes `inline_element_size`
back to a non-atomic integer, and provides a way for all profiles to
provide Register()-time values.
PiperOrigin-RevId: 413960559

--
70234c5943f8e37e17c1d9c54d8ed61d39880abf by Chris Kennelly <ckennelly@google.com>:

Document that absl::FunctionRef does not allocate.

PiperOrigin-RevId: 413946831

--
3308ae571412c4be3cc32d088c6edac98ff2d1ed by Samuel Benzaquen <sbenza@google.com>:

Internal change

PiperOrigin-RevId: 413933619

--
1617093a730d055edcf7bc04fdd6509783f5f75d by Martijn Vels <mvels@google.com>:

Internal Change

PiperOrigin-RevId: 413778735

--
03ad683f059c806a6c8b04f5b79b2662c3df8c73 by Evan Brown <ezb@google.com>:

Unify btree erase_if definitions and optimize them so that we only do rebalancing once per leaf node.

PiperOrigin-RevId: 413757280

--
5ba402f70801938178e486617063f01c7862525d by Martijn Vels <mvels@google.com>:

Cleanup up cord sampling internals

PiperOrigin-RevId: 413755011

--
522da8f9d3e0f11630d89fb41952004742bc335a by Evan Brown <ezb@google.com>:

Add b-tree benchmark for erase_if.

Since this benchmark doesn't work for std:: containers before C++20, disable it for them.

PiperOrigin-RevId: 413740844

--
a690ea42de8ed4a761d00235d8b2fb7548ba9732 by Andy Getzendanner <durandal@google.com>:

Import of CCTZ from GitHub.

PiperOrigin-RevId: 413735737
GitOrigin-RevId: e7f53dfbf809812e84770217777f81b6308a3084
Change-Id: I4f9f9039ba92831bc48971964aa063244c9fed72

1 files changed, 3 insertions, 13 deletions

diff --git a/absl/container/internal/hashtablez_sampler.h b/absl/container/internal/hashtablez_sampler.h
index 0fd9349f..6738786c 100644
--- a/absl/container/internal/hashtablez_sampler.h
+++ b/absl/container/internal/hashtablez_sampler.h
@@ -67,7 +67,8 @@ struct HashtablezInfo : public profiling_internal::Sample<HashtablezInfo> {
 
   // Puts the object into a clean state, fills in the logically `const` members,
   // blocking for any readers that are currently sampling the object.
-  void PrepareForSampling() ABSL_EXCLUSIVE_LOCKS_REQUIRED(init_mu);
+  void PrepareForSampling(size_t inline_element_size_value)
+      ABSL_EXCLUSIVE_LOCKS_REQUIRED(init_mu);
 
   // These fields are mutated by the various Record* APIs and need to be
   // thread-safe.
@@ -82,18 +83,6 @@ struct HashtablezInfo : public profiling_internal::Sample<HashtablezInfo> {
   std::atomic<size_t> hashes_bitwise_xor;
   std::atomic<size_t> max_reserve;
 
-  // One could imagine that inline_element_size could be non-atomic, since it
-  // *almost* follows the rules for the fields that are set by
-  // `PrepareForSampling`.  However, TSAN reports a race (see b/207323922) in
-  // which
-  //   A: Thread 1: Register() returns, unlocking init_mu.
-  //   B: Thread 2: Iterate() is called, locking init_mu.
-  //   C: Thread 1: inlined_element_size is stored.
-  //   D: Thread 2: inlined_element_size is accessed (a race).
-  // A simple solution is to make inline_element_size atomic so that we treat it
-  // at as we do the other atomic fields.
-  std::atomic<size_t> inline_element_size;
-
   // All of the fields below are set by `PrepareForSampling`, they must not be
   // mutated in `Record*` functions.  They are logically `const` in that sense.
   // These are guarded by init_mu, but that is not externalized to clients,
@@ -103,6 +92,7 @@ struct HashtablezInfo : public profiling_internal::Sample<HashtablezInfo> {
   absl::Time create_time;
   int32_t depth;
   void* stack[kMaxStackDepth];
+  size_t inline_element_size;  // How big is the slot?
 };
 
 inline void RecordRehashSlow(HashtablezInfo* info, size_t total_probe_length) {