| commit | 5b845de636224ef3e065be8e1c7d2df3389aa175 | [log] [tgz] |
|---|---|---|
| author | David Benjamin <davidben@google.com> | Sat Jan 07 23:21:52 2023 -0800 |
| committer | Boringssl LUCI CQ <boringssl-scoped@luci-project-accounts.iam.gserviceaccount.com> | Tue May 16 18:53:51 2023 +0000 |
| tree | ac11056d78e1be7bf38ab48fa94e44226a99ce0b | |
| parent | dd9ee6068667ca58c8d6f1c1cea617fd69452ecf [diff] |
Use Windows Interlocked* APIs for refcounts when C11 isn't available Right now, MSVC has to fallback to refcount_lock.c, which uses a single, global lock for all refcount operations. Instead, use the Interlocked* APIs to implement them. The motivation is two-fold. First, this removes a performance cliff when building for Windows on a non-Clang compiler. (Although I've not been able to measure it in an end-to-end EVP benchmark, only a synthetic refcount-only benchmark.) More importantly, it gets us closer to assuming atomics support on all non-NO_THREADS configurations. (The next CL will clear through that.) That, in turn, will make it easier to add an atomics-like abstractions to some of our hotter synchronization points. (Even in newer glibc, with its better rwlock, read locks fundamentally need to write to memory, so we have some cacheline contention on shared locks.) Annoyingly, the Windows atomic_load replacement is not quite right. I've used a "no-op" InterlockedCompareExchange(p, 0, 0) which, empirically, still results in a write. But a write to the refcount cacheline is surely better than taking a global exclusive lock. See comments in file for details. OpenSSL uses InterlockedOr(p, 0), but that actually results in even worse code. (InterlockedOr needs a retry loop when the underlying cmpxchg fails, whereas InterlockedCompareExchange is a single cmpxchg.) Hopefully, in the future (perhaps when we require VS 2022's successor, based on [1]), this can be removed in favor of C11 atomics everywhere. [1] https://devblogs.microsoft.com/cppblog/c11-atomics-in-visual-studio-2022-version-17-5-preview-2/ Bug: 570 Cq-Include-Trybots: luci.boringssl.try:linux_clang_rel_tsan Change-Id: I125da139e2fd3ae51e54309309fda16ba97ccf20 Reviewed-on: https://boringssl-review.googlesource.com/c/boringssl/+/59846 Commit-Queue: David Benjamin <davidben@google.com> Reviewed-by: Adam Langley <agl@google.com>
BoringSSL is a fork of OpenSSL that is designed to meet Google's needs.
Although BoringSSL is an open source project, it is not intended for general use, as OpenSSL is. We don't recommend that third parties depend upon it. Doing so is likely to be frustrating because there are no guarantees of API or ABI stability.
Programs ship their own copies of BoringSSL when they use it and we update everything as needed when deciding to make API changes. This allows us to mostly avoid compromises in the name of compatibility. It works for us, but it may not work for you.
BoringSSL arose because Google used OpenSSL for many years in various ways and, over time, built up a large number of patches that were maintained while tracking upstream OpenSSL. As Google's product portfolio became more complex, more copies of OpenSSL sprung up and the effort involved in maintaining all these patches in multiple places was growing steadily.
Currently BoringSSL is the SSL library in Chrome/Chromium, Android (but it's not part of the NDK) and a number of other apps/programs.
Project links:
There are other files in this directory which might be helpful: