Crash at creation of bcache if caching size > backing size

Hi,

I found a bug that crashes the machine if you create a bcache where the caching device size is bigger than backing device.

This case may happen if you create a bcache with a plan to add a bigger disk/partition/raid later to it.

To reproduce it follow this steps:

start a vm with utopic-amd64-desktop iso, it must have 2 disks, one of 16G and on of 8G

  apt-get update && apt-get install bcache-tools
  /dev/sda is 16G caching device
  /dev/sdb is 8G backing device

  make-bcache --writeback --discard -C /dev/sda -B /dev/sdb

now the machine is hanged!

if you reboot and reinstall bcache-tools
the bcache device has been created

  ************************************
  bcache-super-show -f /dev/sda:
  ==============================
  sb.magic ok
  sb.first_sector 8 [match]
  sb.csum 45C96B38A79275A7 [match]
  sb.version 0 [cache device]

  dev.label (empty)
  dev.uuid 155ab65e-e74a-4ca8-91da-467ef696db63
  dev.sectors_per_block 1
  dev.sectors_per_bucket 1024
  dev.cache.first_sector 1024
  dev.cache.cache_sectors 33553408
  dev.cache.total_sectors 33554432
  dev.cache.ordered no
  dev.cache.discard yes
  dev.cache.pos 0
  dev.cache.replacement 0 [lru]

  cset.uuid bcaa2595-bbc8-43a7-8c44-5d0aca0f3fc8

  *****************************************************
  bcache-super-show /dev/sdb:
  =============================================
  sb.magic ok
  sb.first_sector 8 [match]
  sb.csum F37A1BB08A7F5DE7 [match]
  sb.version 1 [backing device]

  dev.label (empty)
  dev.uuid 341f1c83-1797-4157-aa9e-524107db5606
  dev.sectors_per_block 1
  dev.sectors_per_bucket 1024
  dev.data.first_sector 16
  dev.data.cache_mode 1 [writeback]
  dev.data.cache_state 0 [detached]

  cset.uuid bcaa2595-bbc8-43a7-8c44-5d0aca0f3fc8
  *****************************************************

you can format it:
  mkfs.ext4 /dev/bcache0
  Rejet des blocs de périphérique : complété
  Creating filesystem with 2097150 4k blocks and 524288 inodes
  Filesystem UUID: b99a09de-3a92-4a44-82cd-29f75f8bf0c9
  Superblocs de secours stockés sur les blocs :
   32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632

  Allocation des tables de groupe : complété
  Écriture des tables d'i-noeuds : complété
  Création du journal (32768 blocs) : complété
  Écriture des superblocs et de l'information de comptabilité du système de
  fichiers :

restarting the format gave me an other bug (X.org crashing):
  free from cli:
        total used free shared buffers cached
  Mem: 1017168 814040 203128 195568 91264 513064
  -/+ buffers/cache: 209712 807456
  Swap: 0 0 0

  tail /var/log/syslog:
  [ 568.016481] [ 4983] 0 4983 3332 121 11 0 0 mdadm
  [ 568.016483] [ 5259] 0 5259 6703 300 19 0 0 mkfs.ext4
  [ 568.016484] Out of memory: Kill process 1947 (Xorg) score 38 or sacrifice child
  [ 568.016485] Killed process 1947 (Xorg) total-vm:313160kB, anon-rss:35976kB, file-rss:4232kB
  [ 568.707107] systemd-logind[1396]: Failed to start unit user@112.service: Unknown unit: user@112.service
  [ 568.707113] systemd-logind[1396]: Failed to start user service: Unknown unit: user@112.service
  [ 568.709876] systemd-logind[1396]: New session c8 of user lightdm.
  [ 568.709890] systemd-logind[1396]: Linked /tmp/.X11-unix/X0 to /run/user/112/X11-display.

You can unregister the bcache device,but to re-use all disks, you'll need a reboot (cf bug 1377142):
  [ 830.811150] bcache: cached_dev_detach_finish() Caching disabled for sdb
  [ 830.949898] bcache: cache_set_free() Cache set bcaa2595-bbc8-43a7-8c44-5d0aca0f3fc8 unregistered