stable.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Greg KH <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org, stable@vger.kernel.org
Cc: torvalds@linux-foundation.org, akpm@linux-foundation.org,
	alan@lxorguk.ukuu.org.uk, Jan Schmidt <list.btrfs@jan-o-sch.net>,
	Chris Mason <chris.mason@oracle.com>
Subject: [ 01/75] Btrfs: fix regression in scrub path resolving
Date: Thu, 19 Apr 2012 14:03:04 -0700	[thread overview]
Message-ID: <20120419210303.221572189@linuxfoundation.org> (raw)
In-Reply-To: <20120419210322.GA6478@kroah.com>

3.3-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Jan Schmidt <list.btrfs@jan-o-sch.net>

commit 7a3ae2f8c8c8432e65467b7fc84d5deab04061a0 upstream.

In commit 4692cf58 we introduced new backref walking code for btrfs. This
assumes we're searching live roots, which requires a transaction context.
While scrubbing, however, we must not join a transaction because this could
deadlock with the commit path. Additionally, what scrub really wants to do
is resolving a logical address in the commit root it's currently checking.

This patch adds support for logical to path resolving on commit roots and
makes scrub use that.

Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 fs/btrfs/backref.c |  115 ++++++++++++++++++++++++++++++-----------------------
 fs/btrfs/backref.h |    5 +-
 fs/btrfs/ioctl.c   |    4 -
 fs/btrfs/scrub.c   |    4 -
 4 files changed, 73 insertions(+), 55 deletions(-)

--- a/fs/btrfs/backref.c
+++ b/fs/btrfs/backref.c
@@ -116,6 +116,7 @@ add_parent:
  * to a logical address
  */
 static int __resolve_indirect_ref(struct btrfs_fs_info *fs_info,
+					int search_commit_root,
 					struct __prelim_ref *ref,
 					struct ulist *parents)
 {
@@ -131,6 +132,7 @@ static int __resolve_indirect_ref(struct
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
+	path->search_commit_root = !!search_commit_root;
 
 	root_key.objectid = ref->root_id;
 	root_key.type = BTRFS_ROOT_ITEM_KEY;
@@ -188,6 +190,7 @@ out:
  * resolve all indirect backrefs from the list
  */
 static int __resolve_indirect_refs(struct btrfs_fs_info *fs_info,
+				   int search_commit_root,
 				   struct list_head *head)
 {
 	int err;
@@ -212,7 +215,8 @@ static int __resolve_indirect_refs(struc
 			continue;
 		if (ref->count == 0)
 			continue;
-		err = __resolve_indirect_ref(fs_info, ref, parents);
+		err = __resolve_indirect_ref(fs_info, search_commit_root,
+					     ref, parents);
 		if (err) {
 			if (ret == 0)
 				ret = err;
@@ -586,6 +590,7 @@ static int find_parent_nodes(struct btrf
 	struct btrfs_delayed_ref_head *head;
 	int info_level = 0;
 	int ret;
+	int search_commit_root = (trans == BTRFS_BACKREF_SEARCH_COMMIT_ROOT);
 	struct list_head prefs_delayed;
 	struct list_head prefs;
 	struct __prelim_ref *ref;
@@ -600,6 +605,7 @@ static int find_parent_nodes(struct btrf
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
+	path->search_commit_root = !!search_commit_root;
 
 	/*
 	 * grab both a lock on the path and a lock on the delayed ref head.
@@ -614,35 +620,39 @@ again:
 		goto out;
 	BUG_ON(ret == 0);
 
-	/*
-	 * look if there are updates for this ref queued and lock the head
-	 */
-	delayed_refs = &trans->transaction->delayed_refs;
-	spin_lock(&delayed_refs->lock);
-	head = btrfs_find_delayed_ref_head(trans, bytenr);
-	if (head) {
-		if (!mutex_trylock(&head->mutex)) {
-			atomic_inc(&head->node.refs);
-			spin_unlock(&delayed_refs->lock);
-
-			btrfs_release_path(path);
-
-			/*
-			 * Mutex was contended, block until it's
-			 * released and try again
-			 */
-			mutex_lock(&head->mutex);
-			mutex_unlock(&head->mutex);
-			btrfs_put_delayed_ref(&head->node);
-			goto again;
-		}
-		ret = __add_delayed_refs(head, seq, &info_key, &prefs_delayed);
-		if (ret) {
-			spin_unlock(&delayed_refs->lock);
-			goto out;
+	if (trans != BTRFS_BACKREF_SEARCH_COMMIT_ROOT) {
+		/*
+		 * look if there are updates for this ref queued and lock the
+		 * head
+		 */
+		delayed_refs = &trans->transaction->delayed_refs;
+		spin_lock(&delayed_refs->lock);
+		head = btrfs_find_delayed_ref_head(trans, bytenr);
+		if (head) {
+			if (!mutex_trylock(&head->mutex)) {
+				atomic_inc(&head->node.refs);
+				spin_unlock(&delayed_refs->lock);
+
+				btrfs_release_path(path);
+
+				/*
+				 * Mutex was contended, block until it's
+				 * released and try again
+				 */
+				mutex_lock(&head->mutex);
+				mutex_unlock(&head->mutex);
+				btrfs_put_delayed_ref(&head->node);
+				goto again;
+			}
+			ret = __add_delayed_refs(head, seq, &info_key,
+						 &prefs_delayed);
+			if (ret) {
+				spin_unlock(&delayed_refs->lock);
+				goto out;
+			}
 		}
+		spin_unlock(&delayed_refs->lock);
 	}
-	spin_unlock(&delayed_refs->lock);
 
 	if (path->slots[0]) {
 		struct extent_buffer *leaf;
@@ -679,7 +689,7 @@ again:
 	if (ret)
 		goto out;
 
-	ret = __resolve_indirect_refs(fs_info, &prefs);
+	ret = __resolve_indirect_refs(fs_info, search_commit_root, &prefs);
 	if (ret)
 		goto out;
 
@@ -1074,8 +1084,7 @@ int tree_backref_for_extent(unsigned lon
 	return 0;
 }
 
-static int iterate_leaf_refs(struct btrfs_fs_info *fs_info,
-				struct btrfs_path *path, u64 logical,
+static int iterate_leaf_refs(struct btrfs_fs_info *fs_info, u64 logical,
 				u64 orig_extent_item_objectid,
 				u64 extent_item_pos, u64 root,
 				iterate_extent_inodes_t *iterate, void *ctx)
@@ -1143,35 +1152,38 @@ static int iterate_leaf_refs(struct btrf
  * calls iterate() for every inode that references the extent identified by
  * the given parameters.
  * when the iterator function returns a non-zero value, iteration stops.
- * path is guaranteed to be in released state when iterate() is called.
  */
 int iterate_extent_inodes(struct btrfs_fs_info *fs_info,
-				struct btrfs_path *path,
 				u64 extent_item_objectid, u64 extent_item_pos,
+				int search_commit_root,
 				iterate_extent_inodes_t *iterate, void *ctx)
 {
 	int ret;
 	struct list_head data_refs = LIST_HEAD_INIT(data_refs);
 	struct list_head shared_refs = LIST_HEAD_INIT(shared_refs);
 	struct btrfs_trans_handle *trans;
-	struct ulist *refs;
-	struct ulist *roots;
+	struct ulist *refs = NULL;
+	struct ulist *roots = NULL;
 	struct ulist_node *ref_node = NULL;
 	struct ulist_node *root_node = NULL;
 	struct seq_list seq_elem;
-	struct btrfs_delayed_ref_root *delayed_refs;
-
-	trans = btrfs_join_transaction(fs_info->extent_root);
-	if (IS_ERR(trans))
-		return PTR_ERR(trans);
+	struct btrfs_delayed_ref_root *delayed_refs = NULL;
 
 	pr_debug("resolving all inodes for extent %llu\n",
 			extent_item_objectid);
 
-	delayed_refs = &trans->transaction->delayed_refs;
-	spin_lock(&delayed_refs->lock);
-	btrfs_get_delayed_seq(delayed_refs, &seq_elem);
-	spin_unlock(&delayed_refs->lock);
+	if (search_commit_root) {
+		trans = BTRFS_BACKREF_SEARCH_COMMIT_ROOT;
+	} else {
+		trans = btrfs_join_transaction(fs_info->extent_root);
+		if (IS_ERR(trans))
+			return PTR_ERR(trans);
+
+		delayed_refs = &trans->transaction->delayed_refs;
+		spin_lock(&delayed_refs->lock);
+		btrfs_get_delayed_seq(delayed_refs, &seq_elem);
+		spin_unlock(&delayed_refs->lock);
+	}
 
 	ret = btrfs_find_all_leafs(trans, fs_info, extent_item_objectid,
 				   extent_item_pos, seq_elem.seq,
@@ -1188,7 +1200,7 @@ int iterate_extent_inodes(struct btrfs_f
 		while (!ret && (root_node = ulist_next(roots, root_node))) {
 			pr_debug("root %llu references leaf %llu\n",
 					root_node->val, ref_node->val);
-			ret = iterate_leaf_refs(fs_info, path, ref_node->val,
+			ret = iterate_leaf_refs(fs_info, ref_node->val,
 						extent_item_objectid,
 						extent_item_pos, root_node->val,
 						iterate, ctx);
@@ -1198,8 +1210,11 @@ int iterate_extent_inodes(struct btrfs_f
 	ulist_free(refs);
 	ulist_free(roots);
 out:
-	btrfs_put_delayed_seq(delayed_refs, &seq_elem);
-	btrfs_end_transaction(trans, fs_info->extent_root);
+	if (!search_commit_root) {
+		btrfs_put_delayed_seq(delayed_refs, &seq_elem);
+		btrfs_end_transaction(trans, fs_info->extent_root);
+	}
+
 	return ret;
 }
 
@@ -1210,6 +1225,7 @@ int iterate_inodes_from_logical(u64 logi
 	int ret;
 	u64 extent_item_pos;
 	struct btrfs_key found_key;
+	int search_commit_root = path->search_commit_root;
 
 	ret = extent_from_logical(fs_info, logical, path,
 					&found_key);
@@ -1220,8 +1236,9 @@ int iterate_inodes_from_logical(u64 logi
 		return ret;
 
 	extent_item_pos = logical - found_key.objectid;
-	ret = iterate_extent_inodes(fs_info, path, found_key.objectid,
-					extent_item_pos, iterate, ctx);
+	ret = iterate_extent_inodes(fs_info, found_key.objectid,
+					extent_item_pos, search_commit_root,
+					iterate, ctx);
 
 	return ret;
 }
--- a/fs/btrfs/backref.h
+++ b/fs/btrfs/backref.h
@@ -22,6 +22,8 @@
 #include "ioctl.h"
 #include "ulist.h"
 
+#define BTRFS_BACKREF_SEARCH_COMMIT_ROOT ((struct btrfs_trans_handle *)0)
+
 struct inode_fs_paths {
 	struct btrfs_path		*btrfs_path;
 	struct btrfs_root		*fs_root;
@@ -44,9 +46,8 @@ int tree_backref_for_extent(unsigned lon
 				u64 *out_root, u8 *out_level);
 
 int iterate_extent_inodes(struct btrfs_fs_info *fs_info,
-				struct btrfs_path *path,
 				u64 extent_item_objectid,
-				u64 extent_offset,
+				u64 extent_offset, int search_commit_root,
 				iterate_extent_inodes_t *iterate, void *ctx);
 
 int iterate_inodes_from_logical(u64 logical, struct btrfs_fs_info *fs_info,
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -3066,8 +3066,8 @@ static long btrfs_ioctl_logical_to_ino(s
 		goto out;
 
 	extent_item_pos = loi->logical - key.objectid;
-	ret = iterate_extent_inodes(root->fs_info, path, key.objectid,
-					extent_item_pos, build_ino_list,
+	ret = iterate_extent_inodes(root->fs_info, key.objectid,
+					extent_item_pos, 0, build_ino_list,
 					inodes);
 
 	if (ret < 0)
--- a/fs/btrfs/scrub.c
+++ b/fs/btrfs/scrub.c
@@ -352,8 +352,8 @@ static void scrub_print_warning(const ch
 		} while (ret != 1);
 	} else {
 		swarn.path = path;
-		iterate_extent_inodes(fs_info, path, found_key.objectid,
-					extent_item_pos,
+		iterate_extent_inodes(fs_info, found_key.objectid,
+					extent_item_pos, 1,
 					scrub_print_warning_inode, &swarn);
 	}
 



  reply	other threads:[~2012-04-19 21:03 UTC|newest]

Thread overview: 76+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-04-19 21:03 [ 00/75] 3.3.3-stable review Greg KH
2012-04-19 21:03 ` Greg KH [this message]
2012-04-19 21:03 ` [ 02/75] drm/radeon/kms: fix DVO setup on some r4xx chips Greg KH
2012-04-19 21:03 ` [ 03/75] drm/i915: Removed IVB forced enable of sprite dest key Greg KH
2012-04-19 21:03 ` [ 04/75] drm/i915/ringbuffer: Exclude last 2 cachlines of ring on 845g Greg KH
2012-04-19 21:03 ` [ 05/75] drm/radeon: only add the mm i2c bus if the hw_i2c module param is set Greg KH
2012-04-19 21:03 ` [ 06/75] drm/i915: properly compute dp dithering for user-created modes Greg KH
2012-04-19 21:03 ` [ 07/75] drm/i915: make rc6 module parameter read-only Greg KH
2012-04-19 21:03 ` [ 08/75] rtlwifi: Preallocate USB read buffers and eliminate kalloc in read routine Greg KH
2012-04-19 21:03 ` [ 09/75] rtlwifi: Add missing DMA buffer unmapping for PCI drivers Greg KH
2012-04-19 21:03 ` [ 10/75] ARM: 7379/1: DT: fix atags_to_fdt() second call site Greg KH
2012-04-19 21:03 ` [ 11/75] ARM: 7384/1: ThumbEE: Disable userspace TEEHBR access for !CONFIG_ARM_THUMBEE Greg KH
2012-04-19 21:03 ` [ 12/75] md/raid1,raid10: Fix calculation of vcnt when processing error recovery Greg KH
2012-04-19 21:03 ` [ 13/75] md/bitmap: prevent bitmap_daemon_work running while initialising bitmap Greg KH
2012-04-19 21:03 ` [ 14/75] [PATCH] Bluetooth: uart-ldisc: Fix memory leak Greg KH
2012-04-19 21:03 ` [ 15/75] Bluetooth: hci_ldisc: fix NULL-pointer dereference on tty_close Greg KH
2012-04-19 21:03 ` [ 16/75] Bluetooth: hci_core: fix NULL-pointer dereference at unregister Greg KH
2012-04-19 21:03 ` [ 17/75] Bluetooth: Remove unneeded locking Greg KH
2012-04-19 21:03 ` [ 18/75] Revert "Btrfs: increase the global block reserve estimates" Greg KH
2012-04-19 21:03 ` [ 19/75] ALSA: hda/realtek - Add a fixup entry for Acer Aspire 8940G Greg KH
2012-04-19 21:03 ` [ 20/75] ext4: address scalability issue by removing extent cache statistics Greg KH
2012-04-19 21:03 ` [ 21/75] ia64: fix futex_atomic_cmpxchg_inatomic() Greg KH
2012-04-19 21:03 ` [ 22/75] panic: fix stack dump print on direct call to panic() Greg KH
2012-04-19 21:03 ` [ 23/75] drivers/rtc/rtc-pl031.c: enable clock on all ST variants Greg KH
2012-04-19 21:03 ` [ 24/75] hugetlb: fix race condition in hugetlb_fault() Greg KH
2012-04-19 21:03 ` [ 25/75] staging: iio: hmc5843: Fix crash in probe function Greg KH
2012-04-19 21:03 ` [ 26/75] Revert "serial/8250_pci: init-quirk msi support for kt serial controller" Greg KH
2012-04-19 21:03 ` [ 27/75] serial: samsung: fix omission initialize ulcon in reset port fn() Greg KH
2012-04-19 21:03 ` [ 28/75] Revert "serial/8250_pci: setup-quirk workaround for the kt serial controller" Greg KH
2012-04-19 21:03 ` [ 29/75] serial/8250_pci: add a "force background timer" flag and use it for the "kt" serial port Greg KH
2012-04-19 21:03 ` [ 30/75] tty: serial: altera_uart: Check for NULL platform_data in probe Greg KH
2012-04-19 21:03 ` [ 31/75] sparc64: Eliminate obsolete __handle_softirq() function Greg KH
2012-04-19 21:03 ` [ 32/75] sparc64: Fix bootup crash on sun4v Greg KH
2012-04-19 21:03 ` [ 33/75] cciss: Initialize scsi host max_sectors for tape drive support Greg KH
2012-04-19 21:03 ` [ 34/75] cciss: Fix scsi tape io with more than 255 scatter gather elements Greg KH
2012-04-19 21:03 ` [ 35/75] perf hists: Catch and handle out-of-date hist entry maps Greg KH
2012-04-19 21:03 ` [ 36/75] video:uvesafb: Fix oops that uvesafb try to execute NX-protected page Greg KH
2012-04-19 21:03 ` [ 37/75] IB/srpt: Set srq_type to IB_SRQT_BASIC Greg KH
2012-04-19 21:03 ` [ 38/75] nohz: Fix stale jiffies update in tick_nohz_restart() Greg KH
2012-04-19 21:03 ` [ 39/75] pch_uart: Fix MSI setting issue Greg KH
2012-04-19 21:03 ` [ 40/75] x86: Use correct byte-sized register constraint in __xchg_op() Greg KH
2012-04-19 21:03 ` [ 41/75] x86: Use correct byte-sized register constraint in __add() Greg KH
2012-04-19 21:03 ` [ 42/75] USB: serial: fix race between probe and open Greg KH
2012-04-19 21:03 ` [ 43/75] USB: pl2303: fix DTR/RTS being raised on baud rate change Greg KH
2012-04-19 21:03 ` [ 44/75] USB: option: re-add NOVATELWIRELESS_PRODUCT_HSPA_HIGHSPEED to option_id array Greg KH
2012-04-19 21:03 ` [ 45/75] USB: ftdi_sio: fix status line change handling for TIOCMIWAIT and TIOCGICOUNT Greg KH
2012-04-19 21:03 ` [ 46/75] USB: ftdi_sio: fix race condition in TIOCMIWAIT, and abort of TIOCMIWAIT when the device is removed Greg KH
2012-04-19 21:03 ` [ 47/75] USB: sierra: add support for Sierra Wireless MC7710 Greg KH
2012-04-19 21:03 ` [ 48/75] USB: dont clear urb->dev in scatter-gather library Greg KH
2012-04-19 21:03 ` [ 49/75] USB: dont ignore suspend errors for root hubs Greg KH
2012-04-19 21:03 ` [ 50/75] xhci: dont re-enable IE constantly Greg KH
2012-04-19 21:03 ` [ 51/75] xhci: Dont write zeroed pointers to xHC registers Greg KH
2012-04-19 21:03 ` [ 52/75] xhci: Restore event ring dequeue pointer on resume Greg KH
2012-04-19 21:03 ` [ 53/75] USB: fix bug of device descriptor got from superspeed device Greg KH
2012-04-19 21:03 ` [ 54/75] xHCI: add XHCI_RESET_ON_RESUME quirk for VIA xHCI host Greg KH
2012-04-19 21:03 ` [ 55/75] xHCI: Correct the #define XHCI_LEGACY_DISABLE_SMI Greg KH
2012-04-19 21:03 ` [ 56/75] [S390] fix tlb flushing for page table pages Greg KH
2012-04-19 21:04 ` [ 57/75] memcg: fix Bad page state after replace_page_cache Greg KH
2012-04-19 21:04 ` [ 58/75] serial: PL011: clear pending interrupts Greg KH
2012-04-19 21:04 ` [ 59/75] serial: PL011: move interrupt clearing Greg KH
2012-04-19 21:04 ` [ 60/75] fcaps: clear the same personality flags as suid when fcaps are used Greg KH
2012-04-19 21:04 ` [ 61/75] xhci: Fix register save/restore order Greg KH
2012-04-19 21:04 ` [ 62/75] usb: gadget: pch_udc: Fix disconnect issue Greg KH
2012-04-19 21:04 ` [ 63/75] usb: gadget: pch_udc: Fix wrong return value Greg KH
2012-04-19 21:04 ` [ 64/75] usb: gadget: pch_udc: Fix USB suspend issue Greg KH
2012-04-19 21:04 ` [ 65/75] usb: gadget: pch_udc: Fix usb/gadget/pch_udc: Fix ether gadget connect/disconnect issue Greg KH
2012-04-19 21:04 ` [ 66/75] usb: gadget: pch_udc: Reduce redundant interrupt Greg KH
2012-04-19 21:04 ` [ 67/75] security: fix compile error in commoncap.c Greg KH
2012-04-19 21:04 ` [ 68/75] spi-topcliff-pch: fix -Wuninitialized warning Greg KH
2012-04-19 21:04 ` [ 69/75] Bluetooth: Adding USB device 13d3:3375 as an Atheros AR3012 Greg KH
2012-04-19 21:04 ` [ 70/75] Bluetooth: Add Atheros maryann PIDVID support Greg KH
2012-04-19 21:04 ` [ 71/75] futex: Do not leak robust list to unprivileged process Greg KH
2012-04-19 21:04 ` [ 72/75] drm/i915: Hold mode_config lock whilst changing mode for lastclose() Greg KH
2012-04-19 21:04 ` [ 73/75] drm/radeon/kms: fix the regression of DVI connector check Greg KH
2012-04-19 21:04 ` [ 74/75] drm/radeon: disable MSI on RV515 Greg KH
2012-04-19 21:04 ` [ 75/75] drm/radeon: fix load detect on rn50 with hardcoded EDIDs Greg KH

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20120419210303.221572189@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=akpm@linux-foundation.org \
    --cc=alan@lxorguk.ukuu.org.uk \
    --cc=chris.mason@oracle.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=list.btrfs@jan-o-sch.net \
    --cc=stable@vger.kernel.org \
    --cc=torvalds@linux-foundation.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).