[PATCH 0/3] btrfs-progs: fix btrfs/249 failure

[Qu Wenruo <wqu@suse.com>]

Test case btrfs/249 fails from its introduction, as the fix is not
merged.

The root cause can be found in the first patch, which also introduced an
extra safe net to properly exclude the missing seed device.

However that safe net can only work with root privilege (as it goes
tree-search ioctl), that is good enough for the test case, but not the
best solution.

Instead a new kernel patch is introduced to expand the ability of dev
info ioctl (in a compatible way).

This patch would utilize that new member (and can fallback to the
tree-search if kernel doesn't support), so with newer kernel, fi-usage
can work without root privilege.

This patchset relies on the kernel patch titled:

  btrfs: ioctl: allow dev info ioctl to return fsid of a device

I really hope this won't cause extra delays, as the test case btrfs/249
is really causing too much hassle than it should.

Qu Wenruo (3):
  btrfs-progs: filesystem-usage: handle missing seed device properly
  btrfs-progs: sync ioctl from kernel
  btrfs-progs: filesystem-usage: use btrfs_ioctl_dev_info_args::fsid to
    determine if a device is seed

 cmds/filesystem-usage.c | 129 ++++++++++++++++++++++++++++++++++++----
 ioctl.h                 |  13 +++-
 libbtrfs/ioctl.h        |  13 +++-
 3 files changed, 141 insertions(+), 14 deletions(-)

-- 
2.39.1