View source for Seed-device

=== Seed-device use example ===
The original filesystem and devices are included as a readonly starting point to the new FS. All modifications go onto different devices and the COW machinery makes sure the original is unchanged.

 # dd if=/dev/zero of=/mnt/gentoo/btrfs-test-1 count=2 bs=1G
 # dd if=/dev/zero of=/mnt/gentoo/btrfs-test-2 count=2 bs=1G

 # losetup /dev/loop0 /mnt/gentoo/btrfs-test-1
 # losetup /dev/loop1 /mnt/gentoo/btrfs-test-2

 # mkfs.btrfs /dev/loop0

 # mount /dev/loop0 /mnt/test/
 # echo a > /mnt/test/a
 # echo b > /mnt/test/b

 # umount /mnt/test

Now we will use this filesystem as a seed device:

 # [[Btrfstune|btrfstune]] -S 1 /dev/loop0

 # mount /dev/loop0 /mnt/test
 mount: block device /dev/loop0 is write-protected, mounting read-only
 # btrfs device add /dev/loop1 /mnt/test
 # ls /mnt/test
 a  b
 # echo c > /mnt/test/c
 bash: c: Read-only file system

 # mount -o remount,rw /mnt/test
 OR
 # umount /mnt/test
 # mount /dev/loop1 /mnt/test

 # echo c > /mnt/test/c
 # echo d > /mnt/test/d
 # umount /mnt/test

 # mount /dev/loop0 /mnt/test
 mount: block device /dev/loop0 is write-protected, mounting read-only
 # ls /mnt/test
 a  b
 # umount /mnt/test
 # mount /dev/loop1 /mnt/test
 # ls /mnt/test
 a  b  c  d
 # cat /mnt/test/c
 c

It's also worth noting that seed devices can be recursive (Tested on 3.3 with the btrfs/for-linus branch that went into 3.4-final merged in):

 # dd if=/dev/zero of=seed1.img bs=1 seek=1G count=1
 1+0 records in
 1+0 records out
 1 byte (1 B) copied, 7.4136e-05 s, 13.5 kB/s
 
 # dd if=/dev/zero of=seed2.img bs=1 seek=1G count=1
 1+0 records in
 1+0 records out
 1 byte (1 B) copied, 5.8166e-05 s, 17.2 kB/s
 
 # dd if=/dev/zero of=seed3.img bs=1 seek=1G count=1
 1+0 records in
 1+0 records out
 1 byte (1 B) copied, 8.4253e-05 s, 11.9 kB/s
 
 # sudo losetup /dev/loop0 seed1.img
 
 # sudo losetup /dev/loop1 seed2.img
 
 # sudo losetup /dev/loop2 seed3.img
 
 # sudo mkdir /mnt/btrseed
 
 # sudo mkfs.btrfs -d single -m single /dev/loop0
 
 WARNING! - Btrfs Btrfs v0.19-dirty IS EXPERIMENTAL
 WARNING! - see http://btrfs.wiki.kernel.org before using
 
 fs created label (null) on /dev/loop0
         nodesize 4096 leafsize 4096 sectorsize 4096 size 1.00GB
 Btrfs Btrfs v0.19-dirty
 
 # sudo mount /dev/loop0 /mnt/btrseed
 
 # sudo touch /mnt/btrseed/seed1
 
 # sudo umount /mnt/btrseed
 
 # sudo btrfstune -S 1 /dev/loop0
 
 # sudo mount /dev/loop0 /mnt/btrseed
 mount: block device /dev/loop0 is write-protected, mounting read-only
 
 # sudo btrfs dev add /dev/loop1 /mnt/btrseed
 
 # sudo mount -o remount,rw /mnt/btrseed
 
 # sudo touch /mnt/btrseed/seed2
 
 # ls /mnt/btrseed
 seed1  seed2
 
 # sudo umount /mnt/btrseed
 
 # sudo btrfstune -S 1 /dev/loop1
 
 # sudo mount /dev/loop1 /mnt/btrseed
 mount: block device /dev/loop1 is write-protected, mounting read-only
 
 # sudo btrfs dev add /dev/loop2 /mnt/btrseed
 
 # sudo mount -o remount,rw /mnt/btrseed
 
 # sudo touch /mnt/btrseed/seed3
 
 # ls /mnt/btrseed
 seed1  seed2  seed3
 
 # sudo umount /mnt/btrseed
 
 # sudo mount /dev/loop1 /mnt/btrseed
 mount: block device /dev/loop1 is write-protected, mounting read-only
 
 # ls /mnt/btrseed
 seed1  seed2
 
 # sudo mount /dev/loop0 /mnt/btrseed
 mount: block device /dev/loop0 is write-protected, mounting read-only
 
 # ls /mnt/btrseed
 seed1