Re: threaded programs report bogus process size ?

[mlelstv%serpens.de@localhost (Michael van Elst)]

plunky%rya-online.net@localhost (Iain Hibbert) writes:

>can you get "almost" the same effect without looping by using LIFO for the
>deadqueue rather than FIFO? (eg PTQ_INSERT_TAIL() in pthread__reap())

I think so. However, in a situation where threads are frequently
created and destroyed this will have negative effect on the cache
because more threads and more thread stacks are actually in use.

N.B. the kernel race occurs very rarely and only in this case the
looping generates additional overhead.

-- 
-- 
                                Michael van Elst
Internet: mlelstv%serpens.de@localhost
                                "A potential Snark may lurk in every tree."