diff options
author | MichaĆ Chojnowski <michal.chojnowski@scylladb.com> | 2022-07-07 13:54:51 +0300 |
---|---|---|
committer | Vladislav Khmelevsky <och95@yandex.ru> | 2022-07-07 14:27:29 +0300 |
commit | bd301a418bf2a866c7c17a5cdc82a1d7d13e156f (patch) | |
tree | b143e5626daf91818d581d8ff210000636f6cd02 /bolt/runtime | |
parent | 8e03bfc368f72a4f97129b0837396e69ae2ba1fe (diff) | |
download | llvm-bd301a418bf2a866c7c17a5cdc82a1d7d13e156f.tar.gz |
[BOLT] Fix concurrent hash table modification in the instrumentation runtime
`__bolt_instr_data_dump()` does not lock the hash tables when iterating
over them, so the iteration can happen concurrently with a modification
done in another thread, when the table is in an inconsistent state. This
also has been observed in practice, when it caused a segmentation fault.
We fix this by locking hash tables during iteration. This is done by taking
the lock in `forEachElement()`.
The only other site of iteration, `resetCounters()`, has been correctly
locking the table even before this patch. This patch removes its `Lock`
because the lock is now taken in the inner `forEachElement()`.
Reviewed By: maksfb, yota9
Differential Revision: https://reviews.llvm.org/D129089
Diffstat (limited to 'bolt/runtime')
-rw-r--r-- | bolt/runtime/instr.cpp | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/bolt/runtime/instr.cpp b/bolt/runtime/instr.cpp index 8dd84ffcb4e8..1a8bcf500362 100644 --- a/bolt/runtime/instr.cpp +++ b/bolt/runtime/instr.cpp @@ -289,6 +289,7 @@ public: /// Traverses all elements in the table template <typename... Args> void forEachElement(void (*Callback)(MapEntry &, Args...), Args... args) { + Lock L(M); if (!TableRoot) return; return forEachElement(Callback, InitialSize, TableRoot, args...); @@ -378,7 +379,6 @@ template <typename T> void resetIndCallCounter(T &Entry) { template <typename T, uint32_t X, uint32_t Y> void SimpleHashTable<T, X, Y>::resetCounters() { - Lock L(M); forEachElement(resetIndCallCounter); } |