From: Qu Wenruo <wqu@suse.com>
To: linux-btrfs@vger.kernel.org
Subject: [PATCH v2 4/5] btrfs-progs: check: handle csum generation properly for `--init-csum-tree --init-extent-tree`
Date: Fri, 14 Jan 2022 08:51:22 +0800 [thread overview]
Message-ID: <20220114005123.19426-5-wqu@suse.com> (raw)
In-Reply-To: <20220114005123.19426-1-wqu@suse.com>
When using `--init-csum-tree` with `--init-extent-tree`, csum tree
population will be done by iterating all file extent items.
This allow us to skip preallocated extents, but it still has the
folllwing problems:
- Inodes with NODATASUM
- Hole file extents
- Written preallocated extents
We will generate csum for the whole extent, while other part may still
be unwritten.
Make it to follow the same behavior of recently introduced
fill_csum_for_file_extent(), so we can generate correct csum.
Signed-off-by: Qu Wenruo <wqu@suse.com>
---
check/mode-common.c | 53 +++++++++++++++++++++++++++++++++++++++++----
1 file changed, 49 insertions(+), 4 deletions(-)
diff --git a/check/mode-common.c b/check/mode-common.c
index 4c6dca2d5973..e487f2f0bb8e 100644
--- a/check/mode-common.c
+++ b/check/mode-common.c
@@ -1225,6 +1225,7 @@ static int fill_csum_tree_from_one_fs_root(struct btrfs_trans_handle *trans,
struct extent_buffer *node;
struct btrfs_file_extent_item *fi;
char *buf = NULL;
+ u64 skip_ino = 0;
u64 start = 0;
u64 len = 0;
int slot = 0;
@@ -1243,24 +1244,68 @@ static int fill_csum_tree_from_one_fs_root(struct btrfs_trans_handle *trans,
goto out;
/* Iterate all regular file extents and fill its csum */
while (1) {
+ u8 type;
+
btrfs_item_key_to_cpu(path.nodes[0], &key, path.slots[0]);
- if (key.type != BTRFS_EXTENT_DATA_KEY)
+ if (key.type != BTRFS_EXTENT_DATA_KEY &&
+ key.type != BTRFS_INODE_ITEM_KEY)
+ goto next;
+
+ /* This item belongs to an inode with NODATASUM, skip it */
+ if (key.objectid == skip_ino)
goto next;
+
+ if (key.type == BTRFS_INODE_ITEM_KEY) {
+ struct btrfs_inode_item *ii;
+
+ ii = btrfs_item_ptr(path.nodes[0], path.slots[0],
+ struct btrfs_inode_item);
+ /* Check if the inode has NODATASUM flag */
+ if (btrfs_inode_flags(path.nodes[0], ii) &
+ BTRFS_INODE_NODATASUM)
+ skip_ino = key.objectid;
+ goto next;
+ }
node = path.nodes[0];
slot = path.slots[0];
fi = btrfs_item_ptr(node, slot, struct btrfs_file_extent_item);
- if (btrfs_file_extent_type(node, fi) != BTRFS_FILE_EXTENT_REG)
+ type = btrfs_file_extent_type(node, fi);
+
+ /* Skip inline extents */
+ if (type == BTRFS_FILE_EXTENT_INLINE)
goto next;
- start = btrfs_file_extent_disk_bytenr(node, fi);
- len = btrfs_file_extent_disk_num_bytes(node, fi);
+ start = btrfs_file_extent_disk_bytenr(node, fi);
+ /* Skip holes */
+ if (start == 0)
+ goto next;
+ /*
+ * Always generate the csum for the whole preallocated/regular
+ * first, then remove the csum for preallocated range.
+ *
+ * This is to handle holes on regular extents like:
+ * xfs_io -f -c "pwrite 0 8k" -c "sync" -c "punch 0 4k".
+ *
+ * This behavior will cost extra IO/CPU time, but there is
+ * not other way to ensure the correctness.
+ */
csum_root = btrfs_csum_root(gfs_info, start);
+ len = btrfs_file_extent_disk_num_bytes(node, fi);
ret = populate_csum(trans, csum_root, buf, start, len);
if (ret == -EEXIST)
ret = 0;
if (ret < 0)
goto out;
+
+ /* Delete the csum for the preallocated range */
+ if (type == BTRFS_FILE_EXTENT_PREALLOC) {
+ start += btrfs_file_extent_offset(node, fi);
+ len = btrfs_file_extent_num_bytes(node, fi);
+ ret = btrfs_del_csums(trans, start, len);
+ if (ret < 0)
+ goto out;
+ }
next:
/*
* TODO: if next leaf is corrupted, jump to nearest next valid
--
2.34.1
next prev parent reply other threads:[~2022-01-14 0:51 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-01-14 0:51 [PATCH v2 0/5] btrfs-progs: check: properly generate csum for various complex combinations Qu Wenruo
2022-01-14 0:51 ` [PATCH v2 1/5] btrfs-progs: backref: properly queue indirect refs Qu Wenruo
2022-01-14 0:51 ` [PATCH v2 2/5] btrfs-progs: check: move csum tree population into mode-common.[ch] Qu Wenruo
2022-01-14 0:51 ` [PATCH v2 3/5] btrfs-progs: check: don't calculate csum for preallocated file extents Qu Wenruo
2022-01-14 0:51 ` Qu Wenruo [this message]
2022-01-14 0:51 ` [PATCH v2 5/5] btrfs-progs: fsck-tests: add test case for init-csum-tree Qu Wenruo
2022-01-14 16:40 ` David Sterba
2022-01-14 23:40 ` Qu Wenruo
2022-01-17 13:50 ` David Sterba
2022-01-14 16:50 ` [PATCH v2 0/5] btrfs-progs: check: properly generate csum for various complex combinations 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=20220114005123.19426-5-wqu@suse.com \
--to=wqu@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).