On 31.03.2017 00:36, Eric Blake wrote: > Tweak 097 and 176 to operate on an image that is not cluster-aligned, > to give further coverage of clearing out an entire image. > > Tested for qcow (97) and qcow2 (97 and 176). > > The fact that there is a subtle difference between the expected > outputs of 97 and 176 on pass 0 is evidence that we still don't > always clear out the full image in qcow2. Commit a3e1505 fixed > the case of clusters stranded at the 2G mark, but did not fix > stranding an unaligned tail. That will be the next patch, kept > separate to make it easier to see the impact.
I'm not sure I like this very much. The reference output should always be what we actually expect it to be. Therefore, usually the issue is fixed first and the test is added only afterwards. If I want to see the impact I can just do what I always do in those cases: Go to the base commit, compile qemu, step forward to commits to the test, run the test, see it fail, compile qemu again, run the test, see it pass. So would it be possible to switch this and patch 3? (Apart from that, this patch looks good.) Max > Signed-off-by: Eric Blake <ebl...@redhat.com> > > --- > v7: also write data in the unaligned cluster, and document bug it exposes > v6: new patch > --- > tests/qemu-iotests/097 | 17 +++++++++----- > tests/qemu-iotests/097.out | 55 > ++++++++++++++++++++++++++++++++++++---------- > tests/qemu-iotests/176 | 17 +++++++++----- > tests/qemu-iotests/176.out | 55 > ++++++++++++++++++++++++++++++++++++---------- > 4 files changed, 108 insertions(+), 36 deletions(-)
signature.asc
Description: OpenPGP digital signature
