FreeBSD SMPng Interview with Scott Long

animus9 writes "There's an interesting interview with Scott Long over at the ONLamp.com. Scott explains the difference between the various locking methods, and the current status of SMP in FreeBSD 5. He also compares the new SMP implementation with that of FreeBSD 4.x, NetBSD, DragonFly, and Linux. Other items touched upon include scalability, the status of KSE & ULE, and much more."

Re:Sleep locking, spin locking

[Anonymous Coward]

Linux uses spinlocks but only when it is certain that it will be released very quickly. It boils down to efficiency. That is because if a lock is held for a very short amount of time, it is more efficient to wait for it than to switch tasks. The Linux design is really to minimize lock hold times by doing as much work as possible without holding locks, and then checking to make sure that things are still right. This technique has allowed Linux to scale linearly up to hundreds of processors. In practice, Linux's SMP implementation has proved to be one of the best.

Re:Sleep locking, spin locking

[glasn0st]

If I recall correctly, FreeBSD also uses this strategy when the kernel is compiled with ADAPTIVE_MUTEXES which is now the default in 5.3.