linux-btrfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Nikolay Borisov <nborisov@suse.com>
To: linux-btrfs@vger.kernel.org
Cc: Nikolay Borisov <nborisov@suse.com>
Subject: [PATCH 3/3] btrfs: Ensure btrfs_pin_reserved_extent is always called with valid transaction
Date: Mon,  2 Dec 2019 11:40:15 +0200	[thread overview]
Message-ID: <20191202094015.19444-4-nborisov@suse.com> (raw)
In-Reply-To: <20191202094015.19444-1-nborisov@suse.com>

btrfs_pin_reserved_extent only makes sense when called with a valid
transaction handle. At the moment all cases when it's called and
the transaction handle passed to its caller are from error context in
which 'btrfs_fs_info::pinned_extents' is not going to be processed e.g.

btrfs_Free_fs_roots <--called from open_ctree error label or
close_ctree, after all transaction processing has ceased.
 btrfs_free_log_root_tree
  free_log_tree
   walk_log_tree

btrfs_drop_and_free_fs_root
  btrfs_Free_log <-- called if fs is flagged BTRFS_FS_STATE_ERROR, all
  IO is stopped.
   free_log_tree
     walk_log_tree

To make this explicit in the code rework the 3 call sites where
btrfs_pin_reserved_extent is called. This is in preparation to moving
pinned extent machinery to btrfs_transaction.

Signed-off-by: Nikolay Borisov <nborisov@suse.com>
---
 fs/btrfs/tree-log.c | 49 ++++++++++++++++++++++-----------------------
 1 file changed, 24 insertions(+), 25 deletions(-)

diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
index 6ccf260b313e..b76150cce0fe 100644
--- a/fs/btrfs/tree-log.c
+++ b/fs/btrfs/tree-log.c
@@ -2712,20 +2712,18 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
 					free_extent_buffer(next);
 					return ret;
 				}
-
-				if (trans) {
-					btrfs_tree_lock(next);
-					btrfs_set_lock_blocking_write(next);
-					btrfs_clean_tree_block(next);
-					btrfs_wait_tree_block_writeback(next);
-					btrfs_tree_unlock(next);
-				} else {
+				if (!trans) {
 					if (test_and_clear_bit(EXTENT_BUFFER_DIRTY, &next->bflags))
 						clear_extent_buffer_dirty(next);
+					return ret;
 				}
 
-				WARN_ON(root_owner !=
-					BTRFS_TREE_LOG_OBJECTID);
+				btrfs_tree_lock(next);
+				btrfs_set_lock_blocking_write(next);
+				btrfs_clean_tree_block(next);
+				btrfs_wait_tree_block_writeback(next);
+				btrfs_tree_unlock(next);
+				WARN_ON(root_owner != BTRFS_TREE_LOG_OBJECTID);
 				ret = btrfs_pin_reserved_extent(fs_info,
 							bytenr, blocksize);
 				if (ret) {
@@ -2791,18 +2789,18 @@ static noinline int walk_up_log_tree(struct btrfs_trans_handle *trans,
 				struct extent_buffer *next;
 
 				next = path->nodes[*level];
-
-				if (trans) {
-					btrfs_tree_lock(next);
-					btrfs_set_lock_blocking_write(next);
-					btrfs_clean_tree_block(next);
-					btrfs_wait_tree_block_writeback(next);
-					btrfs_tree_unlock(next);
-				} else {
+				if (!trans) {
 					if (test_and_clear_bit(EXTENT_BUFFER_DIRTY, &next->bflags))
 						clear_extent_buffer_dirty(next);
+					return ret;
 				}
 
+				btrfs_tree_lock(next);
+				btrfs_set_lock_blocking_write(next);
+				btrfs_clean_tree_block(next);
+				btrfs_wait_tree_block_writeback(next);
+				btrfs_tree_unlock(next);
+
 				WARN_ON(root_owner != BTRFS_TREE_LOG_OBJECTID);
 				ret = btrfs_pin_reserved_extent(fs_info,
 						path->nodes[*level]->start,
@@ -2873,17 +2871,18 @@ static int walk_log_tree(struct btrfs_trans_handle *trans,
 
 			next = path->nodes[orig_level];
 
-			if (trans) {
-				btrfs_tree_lock(next);
-				btrfs_set_lock_blocking_write(next);
-				btrfs_clean_tree_block(next);
-				btrfs_wait_tree_block_writeback(next);
-				btrfs_tree_unlock(next);
-			} else {
+			if (!trans) {
 				if (test_and_clear_bit(EXTENT_BUFFER_DIRTY, &next->bflags))
 					clear_extent_buffer_dirty(next);
+				goto out;
 			}
 
+			btrfs_tree_lock(next);
+			btrfs_set_lock_blocking_write(next);
+			btrfs_clean_tree_block(next);
+			btrfs_wait_tree_block_writeback(next);
+			btrfs_tree_unlock(next);
+
 			ret = btrfs_pin_reserved_extent(fs_info, next->start,
 							next->len);
 			if (ret)
-- 
2.17.1


  parent reply	other threads:[~2019-12-02  9:40 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-12-02  9:40 [PATCH 0/3] Cleanups from pinned rework try Nikolay Borisov
2019-12-02  9:40 ` [PATCH 1/3] btrfs: Remove WARN_ON in walk_log_tree Nikolay Borisov
2019-12-03 17:21   ` David Sterba
2019-12-02  9:40 ` [PATCH 2/3] btrfs: Remove redundant WARN_ON in walk_down_log_tree Nikolay Borisov
2019-12-02  9:40 ` Nikolay Borisov [this message]
2019-12-03 17:42 ` [PATCH 0/3] Cleanups from pinned rework try David Sterba

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=20191202094015.19444-4-nborisov@suse.com \
    --to=nborisov@suse.com \
    --cc=linux-btrfs@vger.kernel.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).