[PATCH v3 0/2] vfs-6.19.iomap commit 51311f045375 fixups

[Joanne Koong <joannelkoong@gmail.com>]

These are two fixups for commit 51311f045375 ("iomap: track pending read
bytes more optimally") in the vfs-6.19.iomap branch. It would be great
if these could get folded into that original commit, if possible.

The fix for the race was locally tested by running generic/051 in a loop on an
xfs filesystem with 1k block size, as reported by Brian in [1].

Thanks,
Joanne

[1] https://lore.kernel.org/linux-fsdevel/20250926002609.1302233-1-joannelkoong@gmail.com/T/#t 

Changelog:
v2 -> v3:
Fix the race by adding a bias instead of returning from iomap_read_end() early.

v2: https://lore.kernel.org/linux-fsdevel/20251027181245.2657535-1-joannelkoong@gmail.com/
v1: https://lore.kernel.org/linux-fsdevel/20251024215008.3844068-1-joannelkoong@gmail.com/#t

Joanne Koong (2):
  iomap: rename bytes_pending/bytes_accounted to
    bytes_submitted/bytes_not_submitted
  iomap: fix race when reading in all bytes of a folio

 fs/iomap/buffered-io.c | 75 +++++++++++++++++++++++++++++++-----------
 1 file changed, 56 insertions(+), 19 deletions(-)

-- 
2.47.3