All of lore.kernel.org
 help / color / mirror / Atom feed
From: robbieko <robbieko@synology.com>
To: linux-btrfs@vger.kernel.org
Cc: Robbie Ko <robbieko@synology.com>
Subject: [PATCH v2 2/6] Btrfs: incremental send, add generation check for the inode waiting for rmdir operation.
Date: Fri, 28 Oct 2016 09:40:46 +0800	[thread overview]
Message-ID: <1477618850-12922-3-git-send-email-robbieko@synology.com> (raw)
In-Reply-To: <1477618850-12922-1-git-send-email-robbieko@synology.com>

From: Robbie Ko <robbieko@synology.com>

Example scenario:
Parent snapshot:
|---- dir258/ (ino 258, gen 15, dir)
    |--- dir257/ (ino 257, gen 15, dir)
|---- dir259/ (ino 259, gen 15, dir)

Send snapshot:
|---- file258 (ino 258, gen 21, file)
|---- new_dir259/ (ino 259, gen 21, dir)
    |--- dir257/ (ino 257, gen 15, dir)

utimes
mkdir o259-21-0
rename dir258 -> o258-15-0
utimes
mkfile o258-21-0
rename o258-21-0 -> file258
utimes
truncate o258-21-0 size=0
ERROR: truncate o258-21-0 failed: No such file or directory

While computing the send stream the following steps happen:

1) While processing inode 257, we create o259-21-0 and delay
   dir257 rename operation, because its new parent in the send
   snapshot, inode 259, was not yet processed and therefore not
   yet renamed.

2) Later when processing inode 258, we delay rmdir operation for dir258
   because it's not empty, and then orphanize it.

3) After we create o258-21-0 and rename it to file258, we get ENOENT on
   truncate file258. The reason is that the dir258 with inode 258 is
   waiting for rmdir operation and file258's inode is also 258, then
   get_current_path called before truncate will return a unique name.

Fix this by adding generation check for the inode waiting for rmdir
operation.

Signed-off-by: Robbie Ko <robbieko@synology.com>
---
 fs/btrfs/send.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c
index 2060e75..22eca86 100644
--- a/fs/btrfs/send.c
+++ b/fs/btrfs/send.c
@@ -311,7 +311,7 @@ static int is_waiting_for_move(struct send_ctx *sctx, u64 ino);
 static struct waiting_dir_move *
 get_waiting_dir_move(struct send_ctx *sctx, u64 ino);
 
-static int is_waiting_for_rm(struct send_ctx *sctx, u64 dir_ino);
+static int is_waiting_for_rm(struct send_ctx *sctx, u64 dir_ino, u64 dir_gen);
 
 static int need_send_hole(struct send_ctx *sctx)
 {
@@ -2292,7 +2292,7 @@ static int get_cur_path(struct send_ctx *sctx, u64 ino, u64 gen,
 
 		fs_path_reset(name);
 
-		if (is_waiting_for_rm(sctx, ino)) {
+		if (is_waiting_for_rm(sctx, ino, gen)) {
 			ret = gen_unique_name(sctx, ino, gen, name);
 			if (ret < 0)
 				goto out;
@@ -2899,11 +2899,11 @@ get_orphan_dir_info(struct send_ctx *sctx, u64 dir_ino)
 	return NULL;
 }
 
-static int is_waiting_for_rm(struct send_ctx *sctx, u64 dir_ino)
+static int is_waiting_for_rm(struct send_ctx *sctx, u64 dir_ino, u64 dir_gen)
 {
 	struct orphan_dir_info *odi = get_orphan_dir_info(sctx, dir_ino);
 
-	return odi != NULL;
+	return (odi != NULL && odi->gen == dir_gen);
 }
 
 static void free_orphan_dir_info(struct send_ctx *sctx,
@@ -3166,7 +3166,7 @@ static int path_loop(struct send_ctx *sctx, struct fs_path *name,
 	while (ino != BTRFS_FIRST_FREE_OBJECTID) {
 		fs_path_reset(name);
 
-		if (is_waiting_for_rm(sctx, ino))
+		if (is_waiting_for_rm(sctx, ino, gen))
 			break;
 		if (is_waiting_for_move(sctx, ino)) {
 			if (*ancestor_ino == 0)
-- 
1.9.1


  parent reply	other threads:[~2016-10-28  1:41 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-10-28  1:40 [PATCH v2 0/6] Btrfs: incremental send, fix serval case for inode 256 and generation robbieko
2016-10-28  1:40 ` [PATCH v2 1/6] Btrfs: incremental send, do not skip generation inconsistency check for inode 256 robbieko
2016-10-28  1:40 ` robbieko [this message]
2016-10-28  1:40 ` [PATCH v2 3/6] Btrfs: incremental send, add generation for waiting_dir_move struct and check it in can_rmdir robbieko
2016-10-28 15:13   ` Filipe Manana
2016-10-28  1:40 ` [PATCH v2 4/6] Btrfs: incremental send, skip check overwritten if parents' generation are different robbieko
2016-10-28  1:40 ` [PATCH v2 5/6] Btrfs: incremental send, add generation check for inode is waiting for move robbieko
2016-10-28  1:40 ` [PATCH v2 6/6] Btrfs: incremental send, add generation check in existence demtermination for the parent directory robbieko

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=1477618850-12922-3-git-send-email-robbieko@synology.com \
    --to=robbieko@synology.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.