linux-kernel - Description of complete() is either misleading or there is a bug

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [day] [month] [year] [list]

Date:	Sat, 18 Jun 2016 15:17:50 +0300
From:	roy rosen <roy.1rosen@...il.com>
To:	linux-kernel@...r.kernel.org
Subject: Description of complete() is either misleading or there is a bug

The Description says "Threads will be awakened in the same order in
which they were queued."
Reading this the programmer may thing that it implies that the first
thread to get to wait_for_completion_timeout would also be the first
thread to proceed and this is incorrect, at least at the following
scenario:

thread A calls wait_for_completion_timeout and goes to sleep. when
complete is called it wakes up and completes execution of timeout =
action(timeout); in function do_wait_for_common. Now lets say that
another thread B calls wait_for_completion_timeout. now x->done is 1
so this thread sets x->done to 0 and goes on with its execution. Now
thread A continues its execution, it goes to the next line:

} while (!x->done && timeout);

and goes back to sleep. And so thread B executed first although thread
A started waiting first.

So either the Description is wrong or the implementation of
wait_for_completion_timeout and the functions it calls is wrong.