X-Git-Url: https://git.camperquake.de/gitweb.cgi?a=blobdiff_plain;f=module%2Fzfs%2Fzil.c;h=5296b38be726e7fac83855c5a680587d816a75fc;hb=ebe7e575eae1e03b1faa545a424f008faeac589d;hp=fd4d768214280f7aa1ed76769f3a7be5ae22a7ff;hpb=b8864a233c569edcc57c686f3ea8cd1ae3b89153;p=zfs.git

diff --git a/module/zfs/zil.c b/module/zfs/zil.c
index fd4d768..5296b38 100644
--- a/module/zfs/zil.c
+++ b/module/zfs/zil.c
@@ -20,6 +20,7 @@
  */
 /*
  * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2011 by Delphix. All rights reserved.
  */
 
 /* Portions Copyright 2010 Robert Milkowski */
@@ -74,7 +75,7 @@ int zil_replay_disable = 0;    /* disable intent logging replay */
  * zfs_nocacheflush will cause corruption on power loss if a volatile
  * out-of-order write cache is enabled.
  */
-boolean_t zfs_nocacheflush = B_FALSE;
+int zfs_nocacheflush = 0;
 
 static kmem_cache_t *zil_lwb_cache;
 
@@ -296,6 +297,8 @@ zil_parse(zilog_t *zilog, zil_parse_blk_func_t *parse_blk_func,
 	char *lrbuf, *lrp;
 	int error = 0;
 
+	bzero(&next_blk, sizeof(blkptr_t));
+
 	/*
 	 * Old logs didn't record the maximum zh_claim_lr_seq.
 	 */
@@ -317,7 +320,7 @@ zil_parse(zilog_t *zilog, zil_parse_blk_func_t *parse_blk_func,
 	for (blk = zh->zh_log; !BP_IS_HOLE(&blk); blk = next_blk) {
 		uint64_t blk_seq = blk.blk_cksum.zc_word[ZIL_ZC_SEQ];
 		int reclen;
-		char *end;
+		char *end = NULL;
 
 		if (blk_seq > claim_blk_seq)
 			break;
@@ -560,7 +563,7 @@ zil_destroy(zilog_t *zilog, boolean_t keep_first)
 
 	if (!list_is_empty(&zilog->zl_lwb_list)) {
 		ASSERT(zh->zh_claim_txg == 0);
-		ASSERT(!keep_first);
+		VERIFY(!keep_first);
 		while ((lwb = list_head(&zilog->zl_lwb_list)) != NULL) {
 			list_remove(&zilog->zl_lwb_list, lwb);
 			if (lwb->lwb_buf != NULL)
@@ -1073,7 +1076,8 @@ zil_itx_create(uint64_t txtype, size_t lrsize)
 
 	lrsize = P2ROUNDUP_TYPED(lrsize, sizeof (uint64_t), size_t);
 
-	itx = kmem_alloc(offsetof(itx_t, itx_lr) + lrsize, KM_SLEEP);
+	itx = kmem_alloc(offsetof(itx_t, itx_lr) + lrsize,
+	    KM_PUSHPAGE | KM_NODEBUG);
 	itx->itx_lr.lrc_txtype = txtype;
 	itx->itx_lr.lrc_reclen = lrsize;
 	itx->itx_sod = lrsize; /* if write & WR_NEED_COPY will be increased */
@@ -1421,7 +1425,7 @@ zil_commit_writer(zilog_t *zilog)
 	}
 
 	DTRACE_PROBE1(zil__cw1, zilog_t *, zilog);
-	while (itx = list_head(&zilog->zl_itx_commit_list)) {
+	while ((itx = list_head(&zilog->zl_itx_commit_list))) {
 		txg = itx->itx_lr.lrc_txg;
 		ASSERT(txg);
 
@@ -1662,21 +1666,11 @@ zil_alloc(objset_t *os, zil_header_t *zh_phys)
 void
 zil_free(zilog_t *zilog)
 {
-	lwb_t *head_lwb;
 	int i;
 
 	zilog->zl_stop_sync = 1;
 
-	/*
-	 * After zil_close() there should only be one lwb with a buffer.
-	 */
-	head_lwb = list_head(&zilog->zl_lwb_list);
-	if (head_lwb) {
-		ASSERT(head_lwb == list_tail(&zilog->zl_lwb_list));
-		list_remove(&zilog->zl_lwb_list, head_lwb);
-		zio_buf_free(head_lwb->lwb_buf, head_lwb->lwb_sz);
-		kmem_cache_free(zil_lwb_cache, head_lwb);
-	}
+	ASSERT(list_is_empty(&zilog->zl_lwb_list));
 	list_destroy(&zilog->zl_lwb_list);
 
 	avl_destroy(&zilog->zl_vdev_tree);
@@ -1716,6 +1710,10 @@ zil_open(objset_t *os, zil_get_data_t *get_data)
 {
 	zilog_t *zilog = dmu_objset_zil(os);
 
+	ASSERT(zilog->zl_clean_taskq == NULL);
+	ASSERT(zilog->zl_get_data == NULL);
+	ASSERT(list_is_empty(&zilog->zl_lwb_list));
+
 	zilog->zl_get_data = get_data;
 	zilog->zl_clean_taskq = taskq_create("zil_clean", 1, minclsyspri,
 	    2, 2, TASKQ_PREPOPULATE);
@@ -1729,7 +1727,7 @@ zil_open(objset_t *os, zil_get_data_t *get_data)
 void
 zil_close(zilog_t *zilog)
 {
-	lwb_t *tail_lwb;
+	lwb_t *lwb;
 	uint64_t txg = 0;
 
 	zil_commit(zilog, 0); /* commit all itx */
@@ -1741,9 +1739,9 @@ zil_close(zilog_t *zilog)
 	 * destroy the zl_clean_taskq.
 	 */
 	mutex_enter(&zilog->zl_lock);
-	tail_lwb = list_tail(&zilog->zl_lwb_list);
-	if (tail_lwb != NULL)
-		txg = tail_lwb->lwb_max_txg;
+	lwb = list_tail(&zilog->zl_lwb_list);
+	if (lwb != NULL)
+		txg = lwb->lwb_max_txg;
 	mutex_exit(&zilog->zl_lock);
 	if (txg)
 		txg_wait_synced(zilog->zl_dmu_pool, txg);
@@ -1751,6 +1749,19 @@ zil_close(zilog_t *zilog)
 	taskq_destroy(zilog->zl_clean_taskq);
 	zilog->zl_clean_taskq = NULL;
 	zilog->zl_get_data = NULL;
+
+	/*
+	 * We should have only one LWB left on the list; remove it now.
+	 */
+	mutex_enter(&zilog->zl_lock);
+	lwb = list_head(&zilog->zl_lwb_list);
+	if (lwb != NULL) {
+		ASSERT(lwb == list_tail(&zilog->zl_lwb_list));
+		list_remove(&zilog->zl_lwb_list, lwb);
+		zio_buf_free(lwb->lwb_buf, lwb->lwb_sz);
+		kmem_cache_free(zil_lwb_cache, lwb);
+	}
+	mutex_exit(&zilog->zl_lock);
 }
 
 /*
@@ -1937,7 +1948,7 @@ zil_replay(objset_t *os, void *arg, zil_replay_func_t *replay_func[TX_MAX_TYPE])
 	zr.zr_replay = replay_func;
 	zr.zr_arg = arg;
 	zr.zr_byteswap = BP_SHOULD_BYTESWAP(&zh->zh_log);
-	zr.zr_lr = kmem_alloc(2 * SPA_MAXBLOCKSIZE, KM_SLEEP);
+	zr.zr_lr = vmem_alloc(2 * SPA_MAXBLOCKSIZE, KM_SLEEP);
 
 	/*
 	 * Wait for in-progress removes to sync before starting replay.
@@ -1949,7 +1960,7 @@ zil_replay(objset_t *os, void *arg, zil_replay_func_t *replay_func[TX_MAX_TYPE])
 	ASSERT(zilog->zl_replay_blks == 0);
 	(void) zil_parse(zilog, zil_incr_blks, zil_replay_log_record, &zr,
 	    zh->zh_claim_txg);
-	kmem_free(zr.zr_lr, 2 * SPA_MAXBLOCKSIZE);
+	vmem_free(zr.zr_lr, 2 * SPA_MAXBLOCKSIZE);
 
 	zil_destroy(zilog, B_FALSE);
 	txg_wait_synced(zilog->zl_dmu_pool, zilog->zl_destroy_txg);
@@ -1992,3 +2003,11 @@ zil_vdev_offline(const char *osname, void *arg)
 	dmu_objset_rele(os, FTAG);
 	return (error);
 }
+
+#if defined(_KERNEL) && defined(HAVE_SPL)
+module_param(zil_replay_disable, int, 0644);
+MODULE_PARM_DESC(zil_replay_disable, "Disable intent logging replay");
+
+module_param(zfs_nocacheflush, int, 0644);
+MODULE_PARM_DESC(zfs_nocacheflush, "Disable cache flushes");
+#endif