Btrfs-progs: detect when scrub is started twice

Check whether any involved device is already busy running a
scrub. This would cause damaged status messages and the state
"aborted" without the explanation that a scrub was already
running. Therefore check it first, prevent it and give some
feedback to the user if scrub is already running.
Note that if scrub is started with a block device as the
parameter, only that particular block device is checked. It
is a normal mode of operation to start scrub on multiple
single devices, there is no reason to prevent this.

Here is an example:
/mnt2 is the mountpoint of a filesystem.
/dev/sdk and /dev/sdl are the block devices for that filesystem.

case 1:
btrfs scrub start /mnt2
btrfs scrub start /mnt2
-> complain

case 1:
btrfs scrub start /dev/sdk
btrfs scrub start /dev/sdk
-> complain

case 3:
btrfs scrub start /dev/sdk
btrfs scrub start /dev/sdl
-> don't complain

case 4:
btrfs scrub start /dev/sdk
btrfs scrub start /mnt2
-> complain

case 5:
btrfs scrub start /mnt2
btrfs scrub start /dev/sdk
-> complain if the scrub on /dev/sdk is still running.
-> don't complain if the scrub on /dev/sdk is finished, the
   status messages will be fine.

Reported-by: David Sterba <dsterba@suse.cz>
Signed-off-by: Stefan Behrens <sbehrens@giantdisaster.de>
Signed-off-by: David Sterba <dsterba@suse.cz>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>

diff --git a/cmds-scrub.c b/cmds-scrub.c
index c0dc584..95dfee3 100644 (file)
--- a/cmds-scrub.c
+++ b/cmds-scrub.c
@@ -1025,6 +1025,27 @@ int mkdir_p(char *path)
        return 0;
 }
 
+static int is_scrub_running_on_fs(struct btrfs_ioctl_fs_info_args *fi_args,
+                                 struct btrfs_ioctl_dev_info_args *di_args,
+                                 struct scrub_file_record **past_scrubs)
+{
+       int i;
+
+       if (!fi_args || !di_args || !past_scrubs)
+               return 0;
+
+       for (i = 0; i < fi_args->num_devices; i++) {
+               struct scrub_file_record *sfr =
+                       last_dev_scrub(past_scrubs, di_args[i].devid);
+
+               if (!sfr)
+                       continue;
+               if (!(sfr->stats.finished || sfr->stats.canceled))
+                       return 1;
+       }
+       return 0;
+}
+
 static const char * const cmd_scrub_start_usage[];
 static const char * const cmd_scrub_resume_usage[];
 
@@ -1162,6 +1183,27 @@ static int scrub_start(int argc, char **argv, int resume)
                close(fdres);
        }
 
+       /*
+        * check whether any involved device is already busy running a
+        * scrub. This would cause damaged status messages and the state
+        * "aborted" without the explanation that a scrub was already
+        * running. Therefore check it first, prevent it and give some
+        * feedback to the user if scrub is already running.
+        * Note that if scrub is started with a block device as the
+        * parameter, only that particular block device is checked. It
+        * is a normal mode of operation to start scrub on multiple
+        * single devices, there is no reason to prevent this.
+        */
+       if (is_scrub_running_on_fs(&fi_args, di_args, past_scrubs)) {
+               ERR(!do_quiet,
+                   "ERROR: scrub is already running.\n"
+                   "To cancel use 'btrfs scrub cancel %s'.\n"
+                   "To see the status use 'btrfs scrub status [-d] %s'.\n",
+                   path, path);
+               err = 1;
+               goto out;
+       }
+
        t_devs = malloc(fi_args.num_devices * sizeof(*t_devs));
        sp = calloc(fi_args.num_devices, sizeof(*sp));
        spc.progress = calloc(fi_args.num_devices * 2, sizeof(*spc.progress));