Da Zheng, le Tue 22 Dec 2009 22:29:08 +0800, a écrit : > Samuel Thibault wrote: > > And the concurrency-safety of Hurd code has never been tested. Yes, it > > does matter: threaded code can work on UP and fail on SMP because a > > couple of new scheduling scenarii appear and a lot of rare scheduling > > scenarii become common. > It's really bad to hear that:-(
That's why parallel programming is difficult. > In the multi-core era, Hurd should be upgraded to use multiple processors. It's quite ready already since most servers are multithreaded. However that has never been tested, so expect quite a few bugs :) Also, there are probably scalability issues. A couple of cores shouldn't be bad, more would seen problems with the few central locks I have noticed in various servers (like the renamedirlock in libdiskfs) Samuel