This fixed bug #2369 [1] This patch is a new implementation for a previous (non-merged) patch to make measure useful to calculate the new size of the target image after commit. It is very useful to have this kind of calculation if you do qcow2 on block storage (like iSCSi). This because you need to be able to size the target volume big enough to handle the committed image.
Instead of modifying the existing measure logic, a new logic was added that really looks into the allocated clusters and use the (local) next_cluster_index to calculate the target image size. It also takes into account discards from the top into base to lower the index when a cluster is freed. New test was added to verify the code. When testing the new code on the impacted images in bug #2369 [1], this gives the following results: Start: <filesize> <image> 6174015488 top 13421772800 base calculate with discard-no-unref=on: required -> 13624475648 Commited image size: 13610385408 When calculating with discard-no-unref=off: required -> 13520404480 Commited image size: 13504806912 Let me know if I made some mistakes :) [1]: https://gitlab.com/qemu-project/qemu/-/issues/2369 Jean-Louis Dupond (3): block: add for_commit option to measure qcow2: make measure for_commit aware iotests/290: add test case for for_commit measure block/qcow2.c | 137 +++++++++++++++++++++++++++++-- include/block/block_int-common.h | 4 + qapi/block-core.json | 28 +++++++ qemu-img.c | 40 +++++++-- tests/qemu-iotests/290 | 45 ++++++++++ tests/qemu-iotests/290.out | 61 ++++++++++++++ 6 files changed, 304 insertions(+), 11 deletions(-) -- 2.49.0
