Do not stop kernel shrinker on lock contention

Lock contention, by itself, shouldn't indicate a stop condition to the
kernel's slab shrinker. Doing so can cause stalls when the kernel is
trying to free large parts of the cache such as is done by drop_caches

Also, perhaps arc_reclaim_lock should be a spinlock, and this code
eliminated.

AKAMAI: zfs: CR 3593801
Reviewed-by: Tim Chase <[email protected]>
Reviewed-by: Richard Yao <[email protected]>
Reviewed-by: Brian Behlendorf <[email protected]>
Signed-off-by: Debabrata Banerjee <[email protected]>
Issue #6035

module/zfs/arc.c

@@ -4416,7 +4416,7 @@ __arc_shrinker_func(struct shrinker *shrink, struct shrink_control *sc)
 	/* Reclaim in progress */
 	if (mutex_tryenter(&arc_reclaim_lock) == 0) {
 		ARCSTAT_INCR(arcstat_need_free, ptob(sc->nr_to_scan));
-		return (SHRINK_STOP);
+		return (0);
 	}
 
 	mutex_exit(&arc_reclaim_lock);