From: Nikolay Borisov <nborisov@suse.com>
To: Qu Wenruo <wqu@suse.com>, linux-btrfs@vger.kernel.org
Subject: Re: [PATCH 1/3] btrfs-progs: check/lowmem: Check and repair root generation
Date: Fri, 9 Aug 2019 19:10:07 +0300 [thread overview]
Message-ID: <4e94c953-2ecb-8941-1b1c-d1a2dd6a080e@suse.com> (raw)
In-Reply-To: <20190809065320.22702-2-wqu@suse.com>
On 9.08.19 г. 9:53 ч., Qu Wenruo wrote:
> Since kernel is going to reject any root item which is newer than super
> block generation, we need to provide a way to fix such problem in
> btrfs-check.
>
> This patch addes the ability to report and repair root generation in
> lowmem mode.
>
> This is done by cowing the root node, so we will update the root
> generation along with the root node generation at commit transaction
> time.
>
> Signed-off-by: Qu Wenruo <wqu@suse.com>
> ---
> check/mode-common.c | 29 +++++++++++++++++++++++++++++
> check/mode-common.h | 1 +
> check/mode-lowmem.c | 16 ++++++++++++++++
> check/mode-lowmem.h | 1 +
> 4 files changed, 47 insertions(+)
>
> diff --git a/check/mode-common.c b/check/mode-common.c
> index d5f6c8ef97b1..a5b86a0ac32a 100644
> --- a/check/mode-common.c
> +++ b/check/mode-common.c
> @@ -924,3 +924,32 @@ int check_repair_free_space_inode(struct btrfs_fs_info *fs_info,
> }
> return ret;
> }
> +
> +int repair_root_generation(struct btrfs_root *root)
> +{
> + struct btrfs_trans_handle *trans;
> + struct btrfs_path path;
> + struct btrfs_key key;
> + int ret;
> +
> + trans = btrfs_start_transaction(root, 1);
> + if (IS_ERR(trans))
> + return PTR_ERR(trans);
> +
> + key.objectid = 0;
> + key.type = 0;
> + key.offset = 0;
> + btrfs_init_path(&path);
> +
> + /* Here we only CoW the tree root to update the geneartion */
> + path.lowest_level = btrfs_header_level(root->node);
> + root->node->flags |= EXTENT_BAD_TRANSID;
Why do you set EXTENT_BAD_TRANSID? This flag is currently used only in
read_tree_block to link blocks which have failed ordinary validation to
the recow_ebs list, and this only happens in case we have a single copy
for this eb. This list is then processed in the main check logic.
So repair_root_generation could possibly be as simple as just linking
root->node to the fs_info->recow_ebs and leave the rest to the existing
logic?
> +
> + ret = btrfs_search_slot(trans, root, &key, &path, 0, 1);
> + if (ret < 0)
> + return ret;
> +
> + btrfs_release_path(&path);
> + ret = btrfs_commit_transaction(trans, root);
> + return ret;
> +}
> diff --git a/check/mode-common.h b/check/mode-common.h
> index 4c169c6e3b29..4a3abeb02c81 100644
> --- a/check/mode-common.h
> +++ b/check/mode-common.h
> @@ -155,4 +155,5 @@ static inline bool is_valid_imode(u32 imode)
> return true;
> }
>
> +int repair_root_generation(struct btrfs_root *root);
> #endif
> diff --git a/check/mode-lowmem.c b/check/mode-lowmem.c
> index a2be0e6d7034..bf3b57f5ad2d 100644
> --- a/check/mode-lowmem.c
> +++ b/check/mode-lowmem.c
> @@ -4957,6 +4957,7 @@ static int check_btrfs_root(struct btrfs_root *root, int check_all)
> struct btrfs_path path;
> struct node_refs nrefs;
> struct btrfs_root_item *root_item = &root->root_item;
> + u64 super_generation = btrfs_super_generation(root->fs_info->super_copy);
> int ret;
> int level;
> int err = 0;
> @@ -4978,6 +4979,21 @@ static int check_btrfs_root(struct btrfs_root *root, int check_all)
> level = btrfs_header_level(root->node);
> btrfs_init_path(&path);
>
> + if (btrfs_root_generation(root_item) > super_generation + 1) {
> + error(
> + "invalid root generation for root %llu, have %llu expect (0, %llu)",
> + root->root_key.objectid, btrfs_root_generation(root_item),
> + super_generation + 1);
> + err |= INVALID_GENERATION;
> + if (repair) {
> + ret = repair_root_generation(root);
> + if (!ret) {
> + err &= ~INVALID_GENERATION;
> + printf("Reset generation for root %llu\n",
> + root->root_key.objectid);
> + }
> + }
> + }
> if (btrfs_root_refs(root_item) > 0 ||
> btrfs_disk_key_objectid(&root_item->drop_progress) == 0) {
> path.nodes[level] = root->node;
> diff --git a/check/mode-lowmem.h b/check/mode-lowmem.h
> index d2983fd12eb4..0361fb3382b1 100644
> --- a/check/mode-lowmem.h
> +++ b/check/mode-lowmem.h
> @@ -47,6 +47,7 @@
> #define INODE_FLAGS_ERROR (1<<23) /* Invalid inode flags */
> #define DIR_ITEM_HASH_MISMATCH (1<<24) /* Dir item hash mismatch */
> #define INODE_MODE_ERROR (1<<25) /* Bad inode mode */
> +#define INVALID_GENERATION (1<<26) /* Generation is too new */
>
> /*
> * Error bit for low memory mode check.
>
next prev parent reply other threads:[~2019-08-09 16:10 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-08-09 6:53 [PATCH 0/3] btrfs-progs: Check and repair invalid root item generation Qu Wenruo
2019-08-09 6:53 ` [PATCH 1/3] btrfs-progs: check/lowmem: Check and repair root generation Qu Wenruo
2019-08-09 16:10 ` Nikolay Borisov [this message]
2019-08-10 0:30 ` Qu Wenruo
2019-08-10 6:12 ` Nikolay Borisov
2019-08-10 9:24 ` Qu Wenruo
2019-08-12 5:57 ` Qu Wenruo
2019-08-09 6:53 ` [PATCH 2/3] btrfs-progs: check/original: Check and repair root item geneartion Qu Wenruo
2019-08-09 6:53 ` [PATCH 3/3] btrfs-progs: fsck-tests: Add test case for invalid root generation Qu Wenruo
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=4e94c953-2ecb-8941-1b1c-d1a2dd6a080e@suse.com \
--to=nborisov@suse.com \
--cc=linux-btrfs@vger.kernel.org \
--cc=wqu@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).