Reset l2ad_hand and l2ad_first in l2arc_evict

Increasing l2arc_write_size or l2arc_write_boost can result in
l2arc_write_buffers() not having enough space to perform its writes and
panic zio_write_phys().

Instead of resetting l2ad_hand to l2ad_start at the end of
l2arc_write_buffers() and not taking into account a possible
user-mediated increase of l2arc_write_max, we do this in l2arc_evict(),
right after l2arc_write_size() has run. If there is not enough space to
evict (ie we will exceed l2ad_end) we evict to the end of the device,
reset l2ad_hand to l2ad_start, set l2ad_first to 0 and iterate
l2arc_evict(). We avoid infinite iteration of l2arc_evict() by making
sure in l2arc_write_size() that l2ad_start + size does not exceed
l2ad_end.

Reviewed-by: Brian Behlendorf <[email protected]>
Reviewed-by: Ryan Moeller <[email protected]>
Signed-off-by: George Amanakis <[email protected]>
Closes openzfs#10154

module/zfs/arc.c

@@ -7467,9 +7467,9 @@ l2arc_write_eligible(uint64_t spa_guid, arc_buf_hdr_t *hdr)
 }
 
 static uint64_t
-l2arc_write_size(void)
+l2arc_write_size(l2arc_dev_t *dev)
 {
-	uint64_t size;
+	uint64_t size, dev_size;
 
 	/*
 	 * Make sure our globals have meaningful values in case the user
@@ -7486,6 +7486,23 @@ l2arc_write_size(void)
 	if (arc_warm == B_FALSE)
 		size += l2arc_write_boost;
 
+	/*
+	 * Make sure the write size does not exceed the size of the cache
+	 * device. This is important in l2arc_evict(), otherwise infinite
+	 * iteration can occur.
+	 */
+	dev_size = dev->l2ad_end - dev->l2ad_start;
+	if (size >= dev_size) {
+		cmn_err(CE_NOTE, "l2arc_write_max or l2arc_write_boost "
+		    "exceeds the size of the cache device (guid %llu), "
+		    "resetting them to the default (%d)",
+		    dev->l2ad_vdev->vdev_guid, L2ARC_WRITE_SIZE);
+		size = l2arc_write_max = l2arc_write_boost = L2ARC_WRITE_SIZE;
+
+		if (arc_warm == B_FALSE)
+			size += l2arc_write_boost;
+	}
+
 	return (size);
 
 }
@@ -8008,30 +8025,38 @@ l2arc_evict(l2arc_dev_t *dev, uint64_t distance, boolean_t all)
 	arc_buf_hdr_t *hdr, *hdr_prev;
 	kmutex_t *hash_lock;
 	uint64_t taddr;
+	boolean_t rerun;
 
 	buflist = &dev->l2ad_buflist;
 
-	if (!all && dev->l2ad_first) {
-		/*
-		 * This is the first sweep through the device.  There is
-		 * nothing to evict.
-		 */
-		return;
-	}
-
-	if (dev->l2ad_hand >= (dev->l2ad_end - (2 * distance))) {
+top:
+	rerun = B_FALSE;
+	if (dev->l2ad_hand >= (dev->l2ad_end - distance)) {
 		/*
-		 * When nearing the end of the device, evict to the end
-		 * before the device write hand jumps to the start.
+		 * When there is no space to accomodate upcoming writes,
+		 * evict to the end. Then bump the write hand to the start
+		 * and iterate. This iteration does not happen indefinitely
+		 * as we make sure in l2arc_write_size() that when l2ad_hand
+		 * is reset, the write size does not exceed the end of the
+		 * device.
 		 */
+		rerun = B_TRUE;
 		taddr = dev->l2ad_end;
 	} else {
 		taddr = dev->l2ad_hand + distance;
 	}
 	DTRACE_PROBE4(l2arc__evict, l2arc_dev_t *, dev, list_t *, buflist,
 	    uint64_t, taddr, boolean_t, all);
 
-top:
+	if (!all && dev->l2ad_first) {
+		/*
+		 * This is the first sweep through the device.  There is
+		 * nothing to evict.
+		 */
+		goto out;
+	}
+
+retry:
 	mutex_enter(&dev->l2ad_mtx);
 	for (hdr = list_tail(buflist); hdr; hdr = hdr_prev) {
 		hdr_prev = list_prev(buflist, hdr);
@@ -8052,7 +8077,7 @@ l2arc_evict(l2arc_dev_t *dev, uint64_t distance, boolean_t all)
 			mutex_exit(&dev->l2ad_mtx);
 			mutex_enter(hash_lock);
 			mutex_exit(hash_lock);
-			goto top;
+			goto retry;
 		}
 
 		/*
@@ -8101,6 +8126,17 @@ l2arc_evict(l2arc_dev_t *dev, uint64_t distance, boolean_t all)
 		mutex_exit(hash_lock);
 	}
 	mutex_exit(&dev->l2ad_mtx);
+
+out:
+	if (rerun) {
+		/*
+		 * Bump device hand to the device start if it is approaching the
+		 * end. l2arc_evict() has already evicted ahead for this case.
+		 */
+		dev->l2ad_hand = dev->l2ad_start;
+		dev->l2ad_first = B_FALSE;
+		goto top;
+	}
 }
 
 /*
@@ -8448,15 +8484,6 @@ l2arc_write_buffers(spa_t *spa, l2arc_dev_t *dev, uint64_t target_sz)
 	ARCSTAT_INCR(arcstat_l2_lsize, write_lsize);
 	ARCSTAT_INCR(arcstat_l2_psize, write_psize);
 
-	/*
-	 * Bump device hand to the device start if it is approaching the end.
-	 * l2arc_evict() will already have evicted ahead for this case.
-	 */
-	if (dev->l2ad_hand >= (dev->l2ad_end - target_sz)) {
-		dev->l2ad_hand = dev->l2ad_start;
-		dev->l2ad_first = B_FALSE;
-	}
-
 	dev->l2ad_writing = B_TRUE;
 	(void) zio_wait(pio);
 	dev->l2ad_writing = B_FALSE;
@@ -8539,7 +8566,7 @@ l2arc_feed_thread(void *unused)
 
 		ARCSTAT_BUMP(arcstat_l2_feeds);
 
-		size = l2arc_write_size();
+		size = l2arc_write_size(dev);
 
 		/*
 		 * Evict L2ARC buffers that will be overwritten.

tests/runfiles/common.run

@@ -49,7 +49,7 @@ post =
 [tests/functional/cache]
 tests = ['cache_001_pos', 'cache_002_pos', 'cache_003_pos', 'cache_004_neg',
     'cache_005_neg', 'cache_006_pos', 'cache_007_neg', 'cache_008_neg',
-    'cache_009_pos', 'cache_010_neg', 'cache_011_pos']
+    'cache_009_pos', 'cache_010_neg', 'cache_011_pos', 'cache_012_pos']
 tags = ['functional', 'cache']
 
 [tests/functional/cachefile]

tests/zfs-tests/include/libtest.shlib

@@ -4053,3 +4053,16 @@ function ls_xattr # path
 		;;
 	esac
 }
+
+function get_arcstat # stat
+{
+	if is_linux; then
+		typeset stat=$1
+		typeset zfs_arcstats="/proc/spl/kstat/zfs/arcstats"
+		[[ -f "$zfs_arcstats" ]] || return 1
+		grep $stat $zfs_arcstats | awk '{print $3}'
+		return $?
+	else
+		return 1
+	fi
+}

tests/zfs-tests/include/tunables.cfg

@@ -35,6 +35,9 @@ DISABLE_IVSET_GUID_CHECK	disable_ivset_guid_check	zfs_disable_ivset_guid_check
 INITIALIZE_CHUNK_SIZE		initialize_chunk_size		zfs_initialize_chunk_size
 INITIALIZE_VALUE		initialize_value		zfs_initialize_value
 KEEP_LOG_SPACEMAPS_AT_EXPORT	keep_log_spacemaps_at_export	zfs_keep_log_spacemaps_at_export
+L2ARC_NOPREFETCH		l2arc.noprefetch		l2arc_noprefetch
+L2ARC_WRITE_BOOST		l2arc.write_boost		l2arc_write_boost
+L2ARC_WRITE_MAX			l2arc.write_max			l2arc_write_max
 LIVELIST_CONDENSE_NEW_ALLOC	livelist.condense.new_alloc	zfs_livelist_condense_new_alloc
 LIVELIST_CONDENSE_SYNC_CANCEL	livelist.condense.sync_cancel	zfs_livelist_condense_sync_cancel
 LIVELIST_CONDENSE_SYNC_PAUSE	livelist.condense.sync_pause	zfs_livelist_condense_sync_pause

tests/zfs-tests/tests/functional/cache/Makefile.am

@@ -12,7 +12,8 @@ dist_pkgdata_SCRIPTS = \
 	cache_008_neg.ksh \
 	cache_009_pos.ksh \
 	cache_010_neg.ksh \
-	cache_011_pos.ksh
+	cache_011_pos.ksh \
+	cache_012_pos.ksh
 
 dist_pkgdata_DATA = \
 	cache.cfg \

tests/zfs-tests/tests/functional/cache/cache_012_pos.ksh

@@ -0,0 +1,110 @@
+#!/bin/ksh -p
+#
+# CDDL HEADER START
+#
+# This file and its contents are supplied under the terms of the
+# Common Development and Distribution License ("CDDL"), version 1.0.
+# You may only use this file in accordance with the terms of version
+# 1.0 of the CDDL.
+#
+# A full copy of the text of the CDDL should have accompanied this
+# source.  A copy of the CDDL is also available via the Internet at
+# http://www.illumos.org/license/CDDL.
+#
+# CDDL HEADER END
+#
+
+#
+# Copyright (c) 2020, George Amanakis. All rights reserved.
+#
+
+. $STF_SUITE/tests/functional/cache/cache.cfg
+. $STF_SUITE/tests/functional/cache/cache.kshlib
+
+#
+# DESCRIPTION:
+#	Looping around a cache device with l2arc_write_size exceeding
+#	the device size succeeds.
+#
+# STRATEGY:
+#	1. Create pool with a cache device.
+#	2. Set l2arc_write_max to a value larger than the cache device.
+#	3. Create a file larger than the cache device and random read
+#		for 10 sec.
+#	4. Verify that l2arc_write_max is set back to the default.
+#	5. Set l2arc_write_max to a value less than the cache device size but
+#		larger than the default (64MB).
+#	6. Record the l2_size.
+#	7. Random read for 1 sec.
+#	8. Record the l2_size again.
+#	9. If (6) <= (8) then we have not looped around yet.
+#	10. If (6) > (8) then we looped around. Break out of the loop and test.
+#	11. Destroy pool.
+#
+
+verify_runnable "global"
+
+log_assert "Looping around a cache device succeeds."
+
+function cleanup
+{
+	if poolexists $TESTPOOL ; then
+		destroy_pool $TESTPOOL
+	fi
+
+	log_must set_tunable32 L2ARC_WRITE_MAX $write_max
+	log_must set_tunable32 L2ARC_NOPREFETCH $noprefetch
+}
+log_onexit cleanup
+
+typeset write_max=$(get_tunable L2ARC_WRITE_MAX)
+typeset noprefetch=$(get_tunable L2ARC_NOPREFETCH)
+log_must set_tunable32 L2ARC_NOPREFETCH 0
+
+typeset VDEV="$VDIR/vdev.disk"
+typeset VDEV_SZ=$(( 4 * 1024 * 1024 * 1024 ))
+typeset VCACHE="$VDIR/vdev.cache"
+typeset VCACHE_SZ=$(( $VDEV_SZ / 2 ))
+
+typeset fill_mb=$(( floor($VDEV_SZ * 3 / 4 ) ))
+export DIRECTORY=/$TESTPOOL
+export NUMJOBS=4
+export RUNTIME=10
+export PERF_RANDSEED=1234
+export PERF_COMPPERCENT=66
+export PERF_COMPCHUNK=0
+export BLOCKSIZE=128K
+export SYNC_TYPE=0
+export DIRECT=1
+export FILE_SIZE=$(( floor($fill_mb / $NUMJOBS) ))
+
+log_must set_tunable32 L2ARC_WRITE_MAX $(( $VCACHE_SZ * 2 ))
+
+log_must truncate -s $VCACHE_SZ $VCACHE
+log_must truncate -s $VDEV_SZ $VDEV
+
+log_must zpool create -f $TESTPOOL $VDEV cache $VCACHE
+
+log_must fio $FIO_SCRIPTS/mkfiles.fio
+log_must fio $FIO_SCRIPTS/random_reads.fio
+
+typeset write_max2=$(get_tunable L2ARC_WRITE_MAX)
+
+log_must test $write_max2 -eq $write_max
+
+log_must set_tunable32 L2ARC_WRITE_MAX $(( 64 * 1024 * 1024 ))
+export RUNTIME=1
+
+typeset do_once=true
+while $do_once || [[ $l2_size1 -le $l2_size2 ]]; do
+	typeset l2_size1=$(get_arcstat l2_size)
+	log_must fio $FIO_SCRIPTS/random_reads.fio
+	typeset l2_size2=$(get_arcstat l2_size)
+	do_once=false
+done
+
+log_must test $l2_size1 -gt $l2_size2
+
+log_must zpool destroy $TESTPOOL
+
+log_pass "Looping around a cache device succeeds."