linux-btrfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Omar Sandoval <osandov@osandov.com>
To: linux-btrfs@vger.kernel.org
Cc: kernel-team@fb.com, Chris Mason <clm@fb.com>,
	Josef Bacik <josef@toxicpanda.com>,
	Nikolay Borisov <nborisov@suse.com>
Subject: [PATCH v3 08/11] Btrfs: refactor btrfs_evict_inode() reserve refill dance
Date: Fri, 11 May 2018 00:56:13 -0700	[thread overview]
Message-ID: <01accf9b0787fafe377d0dd766d04d741fcf69fc.1526025007.git.osandov@fb.com> (raw)
In-Reply-To: <cover.1526025007.git.osandov@fb.com>
In-Reply-To: <cover.1526025007.git.osandov@fb.com>

From: Omar Sandoval <osandov@fb.com>

The truncate loop in btrfs_evict_inode() does two things at once:

- It refills the temporary block reserve, potentially stealing from the
  global reserve or committing
- It calls btrfs_truncate_inode_items()

The tangle of continues hides the fact that these two steps are actually
separate. Split the first step out into a separate function both for
clarity and so that we can reuse it in a later patch.

Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Omar Sandoval <osandov@fb.com>
---
 fs/btrfs/inode.c | 113 ++++++++++++++++++-----------------------------
 1 file changed, 42 insertions(+), 71 deletions(-)

diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 73bc66d153ef..7ca55af8aa17 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -5189,13 +5189,52 @@ static void evict_inode_truncate_pages(struct inode *inode)
 	spin_unlock(&io_tree->lock);
 }
 
+static struct btrfs_trans_handle *evict_refill_and_join(struct btrfs_root *root,
+							struct btrfs_block_rsv *rsv,
+							u64 min_size)
+{
+	struct btrfs_fs_info *fs_info = root->fs_info;
+	struct btrfs_block_rsv *global_rsv = &fs_info->global_block_rsv;
+	int failures = 0;
+
+	for (;;) {
+		struct btrfs_trans_handle *trans;
+		int ret;
+
+		ret = btrfs_block_rsv_refill(root, rsv, min_size,
+					     BTRFS_RESERVE_FLUSH_LIMIT);
+
+		if (ret && ++failures > 2) {
+			btrfs_warn(fs_info,
+				   "could not allocate space for a delete; will truncate on mount");
+			return ERR_PTR(-ENOSPC);
+		}
+
+		trans = btrfs_join_transaction(root);
+		if (IS_ERR(trans) || !ret)
+			return trans;
+
+		/*
+		 * Try to steal from the global reserve if there is space for
+		 * it.
+		 */
+		if (!btrfs_check_space_for_delayed_refs(trans, fs_info) &&
+		    !btrfs_block_rsv_migrate(global_rsv, rsv, min_size, 0))
+			return trans;
+
+		/* If not, commit and try again. */
+		ret = btrfs_commit_transaction(trans);
+		if (ret)
+			return ERR_PTR(ret);
+	}
+}
+
 void btrfs_evict_inode(struct inode *inode)
 {
 	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
 	struct btrfs_trans_handle *trans;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
-	struct btrfs_block_rsv *rsv, *global_rsv;
-	int steal_from_global = 0;
+	struct btrfs_block_rsv *rsv;
 	u64 min_size;
 	int ret;
 
@@ -5248,85 +5287,17 @@ void btrfs_evict_inode(struct inode *inode)
 	}
 	rsv->size = min_size;
 	rsv->failfast = 1;
-	global_rsv = &fs_info->global_block_rsv;
 
 	btrfs_i_size_write(BTRFS_I(inode), 0);
 
-	/*
-	 * This is a bit simpler than btrfs_truncate since we've already
-	 * reserved our space for our orphan item in the unlink, so we just
-	 * need to reserve some slack space in case we add bytes and update
-	 * inode item when doing the truncate.
-	 */
 	while (1) {
-		ret = btrfs_block_rsv_refill(root, rsv, min_size,
-					     BTRFS_RESERVE_FLUSH_LIMIT);
-
-		/*
-		 * Try and steal from the global reserve since we will
-		 * likely not use this space anyway, we want to try as
-		 * hard as possible to get this to work.
-		 */
-		if (ret)
-			steal_from_global++;
-		else
-			steal_from_global = 0;
-		ret = 0;
-
-		/*
-		 * steal_from_global == 0: we reserved stuff, hooray!
-		 * steal_from_global == 1: we didn't reserve stuff, boo!
-		 * steal_from_global == 2: we've committed, still not a lot of
-		 * room but maybe we'll have room in the global reserve this
-		 * time.
-		 * steal_from_global == 3: abandon all hope!
-		 */
-		if (steal_from_global > 2) {
-			btrfs_warn(fs_info,
-				   "Could not get space for a delete, will truncate on mount %d",
-				   ret);
-			btrfs_orphan_del(NULL, BTRFS_I(inode));
-			btrfs_free_block_rsv(fs_info, rsv);
-			goto no_delete;
-		}
-
-		trans = btrfs_join_transaction(root);
+		trans = evict_refill_and_join(root, rsv, min_size);
 		if (IS_ERR(trans)) {
 			btrfs_orphan_del(NULL, BTRFS_I(inode));
 			btrfs_free_block_rsv(fs_info, rsv);
 			goto no_delete;
 		}
 
-		/*
-		 * We can't just steal from the global reserve, we need to make
-		 * sure there is room to do it, if not we need to commit and try
-		 * again.
-		 */
-		if (steal_from_global) {
-			if (!btrfs_check_space_for_delayed_refs(trans, fs_info))
-				ret = btrfs_block_rsv_migrate(global_rsv, rsv,
-							      min_size, 0);
-			else
-				ret = -ENOSPC;
-		}
-
-		/*
-		 * Couldn't steal from the global reserve, we have too much
-		 * pending stuff built up, commit the transaction and try it
-		 * again.
-		 */
-		if (ret) {
-			ret = btrfs_commit_transaction(trans);
-			if (ret) {
-				btrfs_orphan_del(NULL, BTRFS_I(inode));
-				btrfs_free_block_rsv(fs_info, rsv);
-				goto no_delete;
-			}
-			continue;
-		} else {
-			steal_from_global = 0;
-		}
-
 		trans->block_rsv = rsv;
 
 		ret = btrfs_truncate_inode_items(trans, root, inode, 0, 0);
-- 
2.17.0


  parent reply	other threads:[~2018-05-11  7:56 UTC|newest]

Thread overview: 24+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-05-11  7:56 [PATCH v3 00/11] Btrfs: orphan and truncate fixes Omar Sandoval
2018-05-11  7:56 ` [PATCH v3 01/11] Btrfs: remove stale comment referencing vmtruncate() Omar Sandoval
2018-05-11 10:19   ` David Sterba
2018-05-11 17:16     ` Omar Sandoval
2018-05-11  7:56 ` [PATCH v3 02/11] Btrfs: fix error handling in btrfs_truncate_inode_items() Omar Sandoval
2018-05-11  7:56 ` [PATCH v3 03/11] Btrfs: don't BUG_ON() " Omar Sandoval
2018-05-11  7:56 ` [PATCH v3 04/11] Btrfs: stop creating orphan items for truncate Omar Sandoval
2018-05-11 14:17   ` Josef Bacik
2018-05-11  7:56 ` [PATCH v3 05/11] Btrfs: get rid of BTRFS_INODE_HAS_ORPHAN_ITEM Omar Sandoval
2018-05-11  9:22   ` Nikolay Borisov
2018-05-11 10:06   ` David Sterba
2018-05-11 16:10     ` Josef Bacik
2018-05-11 16:51       ` David Sterba
2018-05-11 17:16         ` Omar Sandoval
2018-05-11  7:56 ` [PATCH v3 06/11] Btrfs: delete dead code in btrfs_orphan_commit_root() Omar Sandoval
2018-05-11 14:19   ` Josef Bacik
2018-05-11  7:56 ` [PATCH v3 07/11] Btrfs: don't return ino to ino cache if inode item removal fails Omar Sandoval
2018-05-11 14:20   ` Josef Bacik
2018-05-11  7:56 ` Omar Sandoval [this message]
2018-05-11  7:56 ` [PATCH v3 09/11] Btrfs: fix ENOSPC caused by orphan items reservations Omar Sandoval
2018-05-11  9:47   ` Nikolay Borisov
2018-05-11  7:56 ` [PATCH v3 10/11] Btrfs: get rid of unused orphan infrastructure Omar Sandoval
2018-05-11  9:31   ` Nikolay Borisov
2018-05-11  7:56 ` [PATCH v3 11/11] Btrfs: reserve space for O_TMPFILE orphan item deletion Omar Sandoval

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=01accf9b0787fafe377d0dd766d04d741fcf69fc.1526025007.git.osandov@fb.com \
    --to=osandov@osandov.com \
    --cc=clm@fb.com \
    --cc=josef@toxicpanda.com \
    --cc=kernel-team@fb.com \
    --cc=linux-btrfs@vger.kernel.org \
    --cc=nborisov@suse.com \
    /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).