On Thu, Dec 13, 2018 at 9:18 PM <fdmanana@xxxxxxxxxx> wrote:
>
> From: Filipe Manana <fdmanana@xxxxxxxx>
>
> Several places allocate a device while holding the device list mutex. This
> can result in a deadlock if reclaim happens because the device, and its
> flush bio, are allocated using GFP_KERNEL mode (by __alloc_device() which
> is used by btrfs_alloc_device()). A transaction commit, which reclaim can
> trigger, needs to lock the device list mutex in its critical section, done
> at btrfs_update_commit_device_size().
>
> Some of these places are device_list_add(), which ends up being called
> through the device scan ioctl, and btrfs_close_one_device(), which ends up
> being called through the device remove ioctl.
>
> Since all the places that add elements to the list of resized devices (the
> device grow and shrink functions) only lock the chunk mutex before adding
> a device to the list, drop the need to acquire the device list mutex from
> btrfs_update_commit_device_size(), which is the only other place that uses
> this list and it already locks the chunk mutex.
>
> Fixes: 78f2c9e6dbb14 ("btrfs: device add and remove: use GFP_KERNEL")
> Fixes: e0ae999414238 ("btrfs: preallocate device flush bio")
> Signed-off-by: Filipe Manana <fdmanana@xxxxxxxx>
Ping.
> ---
> fs/btrfs/volumes.c | 8 ++------
> fs/btrfs/volumes.h | 1 +
> 2 files changed, 3 insertions(+), 6 deletions(-)
>
> diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
> index c872adfc939e..74c4ed29e36e 100644
> --- a/fs/btrfs/volumes.c
> +++ b/fs/btrfs/volumes.c
> @@ -176,7 +176,8 @@ static int __btrfs_map_block(struct btrfs_fs_info *fs_info,
> * chunk_mutex
> * -----------
> * protects chunks, adding or removing during allocation, trim or when a new
> - * device is added/removed
> + * device is added/removed, and the list of resized devices at struct
> + * btrfs_fs_info::fs_devices::resized_devices
> *
> * cleaner_mutex
> * -------------
> @@ -7298,10 +7299,6 @@ void btrfs_update_commit_device_size(struct btrfs_fs_info *fs_info)
> struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
> struct btrfs_device *curr, *next;
>
> - if (list_empty(&fs_devices->resized_devices))
> - return;
> -
> - mutex_lock(&fs_devices->device_list_mutex);
> mutex_lock(&fs_info->chunk_mutex);
> list_for_each_entry_safe(curr, next, &fs_devices->resized_devices,
> resized_list) {
> @@ -7309,7 +7306,6 @@ void btrfs_update_commit_device_size(struct btrfs_fs_info *fs_info)
> curr->commit_total_bytes = curr->disk_total_bytes;
> }
> mutex_unlock(&fs_info->chunk_mutex);
> - mutex_unlock(&fs_devices->device_list_mutex);
> }
>
> /* Must be invoked during the transaction commit */
> diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h
> index aefce895e994..362574b9c37a 100644
> --- a/fs/btrfs/volumes.h
> +++ b/fs/btrfs/volumes.h
> @@ -229,6 +229,7 @@ struct btrfs_fs_devices {
> struct mutex device_list_mutex;
> struct list_head devices;
>
> + /* protected by struct btrfs_fs_info::chunk_mutex */
> struct list_head resized_devices;
> /* devices not currently being allocated */
> struct list_head alloc_list;
> --
> 2.11.0
>
