From: Anand Jain <anand.jain@oracle.com>
To: linux-btrfs@vger.kernel.org
Subject: [PATCH] btrfs: handle the BUG_ON in btrfs_num_devices()
Date: Fri, 3 Aug 2018 20:45:26 +0800 [thread overview]
Message-ID: <20180803124526.18497-2-anand.jain@oracle.com> (raw)
In-Reply-To: <20180803124526.18497-1-anand.jain@oracle.com>
Its a logical bug if we hit fs_devices::num_devices == 1 and if the
replace is running because, as fs_devices::num_devices counts the in memory
devices, so it should include the replace target which is running as
indicated by the flag. If this happens return the -EINVAL back.
Suggested-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Anand Jain <anand.jain@oracle.com>
---
Hi,
As it fixes the BUG_ON I have spun a new patch for this.
Instead of -EINVAL should we use ASSERT?
fs/btrfs/volumes.c | 27 ++++++++++++++++++++-------
1 file changed, 20 insertions(+), 7 deletions(-)
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index 7359596ac8eb..ed2399caff80 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -1855,9 +1855,11 @@ void btrfs_assign_next_active_device(struct btrfs_device *device,
}
/* Returns btrfs_fs_devices::num_devices minus replace device if any */
-static u64 btrfs_num_devices(struct btrfs_fs_info *fs_info)
+static int btrfs_num_devices(struct btrfs_fs_info *fs_info, u64 *num_devices)
{
- u64 num_devices = fs_info->fs_devices->num_devices;
+ int ret = 0;
+
+ *num_devices = fs_info->fs_devices->num_devices;
/*
* balance and replace co-exists in a scenario as below..
@@ -1867,12 +1869,13 @@ static u64 btrfs_num_devices(struct btrfs_fs_info *fs_info)
*/
btrfs_dev_replace_read_lock(&fs_info->dev_replace);
if (btrfs_dev_replace_is_ongoing(&fs_info->dev_replace)) {
- BUG_ON(num_devices < 1);
- num_devices--;
+ if (*num_devices < 1)
+ ret = -EINVAL;
+ (*num_devices)--;
}
btrfs_dev_replace_read_unlock(&fs_info->dev_replace);
- return num_devices;
+ return ret;
}
int btrfs_rm_device(struct btrfs_fs_info *fs_info, const char *device_path,
@@ -1886,7 +1889,12 @@ int btrfs_rm_device(struct btrfs_fs_info *fs_info, const char *device_path,
mutex_lock(&uuid_mutex);
- num_devices = btrfs_num_devices(fs_info);
+ ret = btrfs_num_devices(fs_info, &num_devices);
+ if (ret) {
+ btrfs_err(fs_info, "logical bug num_devices %llu < 0",
+ num_devices);
+ return ret;
+ }
ret = btrfs_check_raid_min_devices(fs_info, num_devices - 1);
if (ret)
@@ -3755,7 +3763,12 @@ int btrfs_balance(struct btrfs_fs_info *fs_info,
}
}
- num_devices = btrfs_num_devices(fs_info);
+ ret = btrfs_num_devices(fs_info, &num_devices);
+ if (ret) {
+ btrfs_err(fs_info, "hits a logical bug num_devices %llu < 0",
+ num_devices);
+ return ret;
+ }
allowed = BTRFS_AVAIL_ALLOC_BIT_SINGLE | BTRFS_BLOCK_GROUP_DUP;
if (num_devices > 1)
--
2.7.0
next prev parent reply other threads:[~2018-08-03 14:38 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-07-26 6:53 [PATCH v2 0/4] Misc volume patch set part2 Anand Jain
2018-07-26 6:53 ` [PATCH v4 1/4] btrfs: drop uuid_mutex in btrfs_free_extra_devids() Anand Jain
2018-07-26 6:53 ` [PATCH v2 2/4] btrfs: fix race between free_stale_devices and close_fs_devices Anand Jain
2018-08-01 14:29 ` David Sterba
2018-08-02 9:29 ` Anand Jain
2018-08-07 14:59 ` David Sterba
2018-08-08 9:51 ` Anand Jain
2018-07-26 6:53 ` [PATCH 3/4] btrfs: bug_on for num_devices below 0 Anand Jain
2018-07-26 6:53 ` [PATCH v2 4/4] btrfs: add helper btrfs_num_devices() to deduce num_devices Anand Jain
2018-08-01 14:41 ` David Sterba
2018-08-02 10:09 ` Anand Jain
2018-08-02 10:09 ` [PATCH v3 " Anand Jain
2018-08-02 10:11 ` Nikolay Borisov
2018-08-02 12:21 ` David Sterba
2018-08-02 13:07 ` Anand Jain
2018-08-07 15:02 ` David Sterba
2018-08-07 22:43 ` Anand Jain
2018-08-03 12:45 ` [PATCH v4 " Anand Jain
2018-08-03 12:45 ` Anand Jain [this message]
2018-08-03 13:33 ` [PATCH] btrfs: handle the BUG_ON in btrfs_num_devices() Nikolay Borisov
2018-08-06 8:57 ` Anand Jain
2018-08-07 17:09 ` David Sterba
2018-08-07 22:51 ` Anand Jain
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=20180803124526.18497-2-anand.jain@oracle.com \
--to=anand.jain@oracle.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).