Re: [PATCH 1/3] btrfs: qgroups, fix rescan worker running races

[Nikolay Borisov <nborisov@suse.com>]

On 26.04.2018 22:23, jeffm@suse.com wrote:
> From: Jeff Mahoney <jeffm@suse.com>
> 
> Commit d2c609b834d6 (Btrfs: fix qgroup rescan worker initialization)
> fixed the issue with BTRFS_IOC_QUOTA_RESCAN_WAIT being racy, but
> ended up reintroducing the hang-on-unmount bug that the commit it
> intended to fix addressed.
> 
> The race this time is between qgroup_rescan_init setting
> ->qgroup_rescan_running = true and the worker starting.  There are
> many scenarios where we initialize the worker and never start it.  The
> completion btrfs_ioctl_quota_rescan_wait waits for will never come.
> This can happen even without involving error handling, since mounting
> the file system read-only returns between initializing the worker and
> queueing it.
> 
> The right place to do it is when we're queuing the worker.  The flag
> really just means that btrfs_ioctl_quota_rescan_wait should wait for
> a completion.
> 
> This patch introduces a new helper, queue_rescan_worker, that handles
> the ->qgroup_rescan_running flag, including any races with umount.
> 
> While we're at it, ->qgroup_rescan_running is protected only by the
> ->qgroup_rescan_mutex.  btrfs_ioctl_quota_rescan_wait doesn't need
> to take the spinlock too.
> 
> Fixes: d2c609b834d6 (Btrfs: fix qgroup rescan worker initialization)
> Signed-off-by: Jeff Mahoney <jeffm@suse.com>


LGTM.

Reviewed-by: Nikolay Borisov <nborisov@suse.com>

> ---
>  fs/btrfs/ctree.h  |  1 +
>  fs/btrfs/qgroup.c | 40 ++++++++++++++++++++++++++++------------
>  2 files changed, 29 insertions(+), 12 deletions(-)
> 
> diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
> index da308774b8a4..dbba615f4d0f 100644
> --- a/fs/btrfs/ctree.h
> +++ b/fs/btrfs/ctree.h
> @@ -1045,6 +1045,7 @@ struct btrfs_fs_info {
>  	struct btrfs_workqueue *qgroup_rescan_workers;
>  	struct completion qgroup_rescan_completion;
>  	struct btrfs_work qgroup_rescan_work;
> +	/* qgroup rescan worker is running or queued to run */
>  	bool qgroup_rescan_running;	/* protected by qgroup_rescan_lock */
>  
>  	/* filesystem state */
> diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c
> index aa259d6986e1..be491b6c020a 100644
> --- a/fs/btrfs/qgroup.c
> +++ b/fs/btrfs/qgroup.c
> @@ -2072,6 +2072,30 @@ int btrfs_qgroup_account_extents(struct btrfs_trans_handle *trans,
>  	return ret;
>  }
>  
> +static void queue_rescan_worker(struct btrfs_fs_info *fs_info)
> +{
> +	mutex_lock(&fs_info->qgroup_rescan_lock);
> +	if (btrfs_fs_closing(fs_info)) {
> +		mutex_unlock(&fs_info->qgroup_rescan_lock);
> +		return;
> +	}
> +	if (WARN_ON(fs_info->qgroup_rescan_running)) {
> +		btrfs_warn(fs_info, "rescan worker already queued");
> +		mutex_unlock(&fs_info->qgroup_rescan_lock);
> +		return;
> +	}
> +
> +	/*
> +	 * Being queued is enough for btrfs_qgroup_wait_for_completion
> +	 * to need to wait.
> +	 */
> +	fs_info->qgroup_rescan_running = true;
> +	mutex_unlock(&fs_info->qgroup_rescan_lock);
> +
> +	btrfs_queue_work(fs_info->qgroup_rescan_workers,
> +			 &fs_info->qgroup_rescan_work);
> +}
> +
>  /*
>   * called from commit_transaction. Writes all changed qgroups to disk.
>   */
> @@ -2123,8 +2147,7 @@ int btrfs_run_qgroups(struct btrfs_trans_handle *trans,
>  		ret = qgroup_rescan_init(fs_info, 0, 1);
>  		if (!ret) {
>  			qgroup_rescan_zero_tracking(fs_info);
> -			btrfs_queue_work(fs_info->qgroup_rescan_workers,
> -					 &fs_info->qgroup_rescan_work);
> +			queue_rescan_worker(fs_info);
>  		}

So here it's not possible to race, since if qgroup_rescan_init returns 0
then we are guaranteed to queue the rescan.

>  		ret = 0;
>  	}
> @@ -2713,7 +2736,6 @@ qgroup_rescan_init(struct btrfs_fs_info *fs_info, u64 progress_objectid,
>  		sizeof(fs_info->qgroup_rescan_progress));
>  	fs_info->qgroup_rescan_progress.objectid = progress_objectid;
>  	init_completion(&fs_info->qgroup_rescan_completion);
> -	fs_info->qgroup_rescan_running = true;
>  
>  	spin_unlock(&fs_info->qgroup_lock);
>  	mutex_unlock(&fs_info->qgroup_rescan_lock);
> @@ -2785,9 +2807,7 @@ btrfs_qgroup_rescan(struct btrfs_fs_info *fs_info)
>  
>  	qgroup_rescan_zero_tracking(fs_info);
>  
> -	btrfs_queue_work(fs_info->qgroup_rescan_workers,
> -			 &fs_info->qgroup_rescan_work);
> -
> +	queue_rescan_worker(fs_info);

Which leaves this to be the only problematic case, in case transaction
joining/commit fails, right?

>  	return 0;
>  }
>  
> @@ -2798,9 +2818,7 @@ int btrfs_qgroup_wait_for_completion(struct btrfs_fs_info *fs_info,
>  	int ret = 0;
>  
>  	mutex_lock(&fs_info->qgroup_rescan_lock);
> -	spin_lock(&fs_info->qgroup_lock);
>  	running = fs_info->qgroup_rescan_running;
> -	spin_unlock(&fs_info->qgroup_lock);
>  	mutex_unlock(&fs_info->qgroup_rescan_lock);
>  
>  	if (!running)
> @@ -2819,12 +2837,10 @@ int btrfs_qgroup_wait_for_completion(struct btrfs_fs_info *fs_info,
>   * this is only called from open_ctree where we're still single threaded, thus
>   * locking is omitted here.
>   */
> -void
> -btrfs_qgroup_rescan_resume(struct btrfs_fs_info *fs_info)
> +void btrfs_qgroup_rescan_resume(struct btrfs_fs_info *fs_info)
>  {
>  	if (fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN)
> -		btrfs_queue_work(fs_info->qgroup_rescan_workers,
> -				 &fs_info->qgroup_rescan_work);
> +		queue_rescan_worker(fs_info);
>  }
>  
>  /*
>