On Mon, 20 Feb 2023 19:38:35 GMT, Daniel Jeliński <[email protected]> wrote:

> Please review this fix for a race in KeepAliveCache.
> 
> The sweeper thread (`KeepAliveCache.run()`) terminates when the cache is 
> empty. The check for empty cache was performed  without holding the cache 
> lock. As a result, there was a small window of time where a new connection 
> could be added to the cache while the sweeper thread was stopping. The added 
> connection would then be removed from cache without closing when a new 
> sweeper thread was started.
> 
> As an additional observation, the check for empty cache was performed after 
> sweeping. The cache could be empty because all connections were closed, or 
> because all connections were busy. In the latter case, a new thread was 
> created soon after the old one terminated.
> 
> This patch addresses both these issues. The sweeper thread makes the decision 
> to terminate while holding the lock, and other threads are aware of that when 
> they acquire the lock. Also, the sweeper thread only terminates if the cache 
> is empty AND there was no cache activity in the last `LIFETIME` (5 seconds).
> 
> Tier1-3 clean. No new tests, the issue was very hard to reproduce without 
> adding delays in production code..

You're right, all accesses to ClientVector use the cacheLock. I removed the 
vector lock.

-------------

PR: https://git.openjdk.org/jdk/pull/12676

Reply via email to