ext4: check superblock mapped prior to committing
authorJon Derrick <jonathan.derrick@intel.com>
Mon, 2 Jul 2018 22:45:18 +0000 (18:45 -0400)
committerTheodore Ts'o <tytso@mit.edu>
Mon, 2 Jul 2018 22:45:18 +0000 (18:45 -0400)
commita17712c8e4be4fa5404d20e9cd3b2b21eae7bc56
tree5e7904cf0721942898da5e47ff811091252644f4
parentbfe0a5f47ada40d7984de67e59a7d3390b9b9ecc
ext4: check superblock mapped prior to committing

This patch attempts to close a hole leading to a BUG seen with hot
removals during writes [1].

A block device (NVME namespace in this test case) is formatted to EXT4
without partitions. It's mounted and write I/O is run to a file, then
the device is hot removed from the slot. The superblock attempts to be
written to the drive which is no longer present.

The typical chain of events leading to the BUG:
ext4_commit_super()
  __sync_dirty_buffer()
    submit_bh()
      submit_bh_wbc()
        BUG_ON(!buffer_mapped(bh));

This fix checks for the superblock's buffer head being mapped prior to
syncing.

[1] https://www.spinics.net/lists/linux-ext4/msg56527.html

Signed-off-by: Jon Derrick <jonathan.derrick@intel.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: stable@kernel.org
fs/ext4/super.c