Double check lock lab.

[dhanna11]

I suspect that there's a deadlock condition that the tests aren't triggering.

[dhanna11]

Actually, I suspect that a circular wait and hold condition is broken by checking for a free buffer in the first bucket we lock. If there's a free block within the cache, this check has to eventually succeed, unblocking any threads that would be waiting on this bucket lock.

xv6/kernel/bio.c

Line 115 in 3d4c39c

b = get_lru(bcache.bufs[bucket]);