[Devel] [PATCH hci-8.1 v9 00/10] dm-qcow2: truncate unused tail of image on suspend

[Konstantin Khorenko]

I won't apply this version, please send v10 with the changes below.

On 10/28/26 16:06, Alexander Ivanov wrote:
> Patch 1 adds qcow2_md_page_read_sync() helper, which is needed by the
> scan. Patch 2 is a small cleanup of duplicate prefixes in QC_ERR/QC_INFO
> messages. Patches 3-9 fix existing cases where the refcounts or the
> dirty bit on disk don't reflect real cluster usage, or where such an
> image is not flagged for check, which the scan relies on. Patch 10 does
> the actual truncation.

- Add a Fixes tag for 222b6f1291c1c ("dm-qcow2: Introduce driver to
  create block devices over QCOW2 files") to patches 3, 4, 5, 7, 8
  and 9.
- Put a blank line before Signed-off-by in patches 4, 5 and 6.
- Describe patch 7 on its own here.
- Say whether the ENOSPC retry of md writeback is planned.

Patches 3-9 fix bugs that came with the driver itself. Other dm-qcow2
fixes in the tree carry the tag, these should too.

Patch 7 is not about refcounts or the dirty bit. It fixes a relocation
that keeps failing after one error, and pages freed while they are
still on the writeback list.

My v5 cover reply asked for three fixes. Patch 3 does place_r2(), and
patch 9 handles the failed L1/L2 write (see my reply there). The ENOSPC
retry of md writeback is not done: a refblock write that fails on
ENOSPC still blocks resume until a reload, and the "FIXME: call
md_make_dirty() and try once again?" is still there.

-- 
Konstantin Khorenko,
Virtuozzo Linux Kernel Team