[PATCH 0/2] xfs: fix eof zeroing i_size race

[Brian Foster]

Hi all,

This addresses an EOF zeroing race that's been reproduced by a somewhat
obscure virtual machine workload on qcow2 vdisks with an increasing
i_size. Patch 1 contains the fix and patch 2 just adds a tracepoint that
helped identify the problem. Thoughts, reviews, flames appreciated.

Brian

Brian Foster (2):
  xfs: always drain dio before extending aio write submission
  xfs: add an xfs_zero_eof() tracepoint

 fs/xfs/xfs_file.c  | 17 +++++++++++------
 fs/xfs/xfs_trace.h |  1 +
 2 files changed, 12 insertions(+), 6 deletions(-)

-- 
2.1.0