vfs: be less eager to call uma_reclaim(UMA_RECLAIM_DRAIN)

In face of vnode shortage the count very easily can go few units above
the limit before going back down.

Calling uma_reclaim results in massive amount of work which in this case
is not warranted.

Sponsored by:	Rubicon Communications, LLC ("Netgate")

This commit is contained in:

Mateusz Guzik

2023-10-10 16:15:53 +00:00

parent e49c7cd677

commit 1bf55a739e

1 changed files with 6 additions and 1 deletions

									
										sys/kern/vfs_subr.c
									
		+6
		-1
	
												View File
												
				@@ -1674,7 +1674,12 @@ vnlru_proc(void)

						target = target / 10 + 1;

						done = vlrureclaim(reclaim_nc_src, trigger, target);

						mtx_unlock(&vnode_list_mtx);

						if (onumvnodes > desiredvnodes && numvnodes <= desiredvnodes)

						/*

						 * Total number of vnodes can transiently go slightly above the

						 * limit (see vn_alloc_hard), no need to call uma_reclaim if

						 * this happens.

						 */

						if (onumvnodes + 1000 > desiredvnodes && numvnodes <= desiredvnodes)

							uma_reclaim(UMA_RECLAIM_DRAIN);

						if (done == 0) {

							if (force == 0 || force == 1) {