"Shawn O. Pearce" [off-list ref] writes:
quoted
+static void try_to_free_from_threads(size_t size)
+{
+ int self = pthread_equal(read_mutex_owner, pthread_self());
+ if (!self)
+ read_lock();
+ release_pack_memory(size, -1);
+ if (!self)
+ read_unlock();
+}
Is there any concern that a partially unset read_mutex_owner might
look like the current thread's identity?
That is, memset() can be setting the bytes one by one. If the lock
is being released we might observe the current owner as ourselves
if we see only part of that release, and our identity is the same
as another thread, only with the lower-address bytes unset.
Yuck. I hope that it doesn't mean we need another mutex to protect the
owner data.