[PATCH] Update tcache double-free check

Carlos O'Donell carlos@redhat.com
Mon Aug 10 13:12:53 GMT 2020


On 8/10/20 9:07 AM, Eyal Itkin wrote:
> Updated the patch to perform an atomic update operation on the global
> entropy state, so to avoid races when multiple threads are initialized
> simultaneously. The patch now accumulates entropy between threads,
> while still using the backup case of ~tcache to take care of cases in
> which no entropy was yet to be available.
> 
> As Carlos mentioned earlier, I guess you will want to discuss this
> patch before integrating it. Also, feel free to update the patch if
> needed in case I missed some whitespace / long line coding style.

Thank you for putting this together. I need to spend some time thinking
more deeply on this and considering where the right balance might lie
between a per-process value and a per-thread value. Particularly with
respect to the tradeoff between maintaining the code and security.

Do you have any strong opinions on the use of a per-thread vs. per-process
value?

This patch is third on my queue.

My queue is currently:
- NSS configuration reloading (DJ Delorie)
- DSO sorting DFS (Chung-Lin Tang)
- Tcache double-free check (Eyal Itkin)

-- 
Cheers,
Carlos.



More information about the Libc-alpha mailing list