Add _verify_du_delta() checker which is used to check that real disk
usage delta meets the expectations. For now we use it for checking that
subcluster-based discard/unmap operations lead to actual disk usage
decrease (i.e. PUNCH_HOLE operation is performed).
Also add separate test case for disca
On 11/3/23 17:59, Hanna Czenczek wrote:
> On 03.11.23 16:51, Hanna Czenczek wrote:
>> On 20.10.23 23:56, Andrey Drobyshev wrote:
>
> [...]
>
>>> @@ -528,6 +543,14 @@ for use_backing_file in yes no; do
>>> else
>>> _make_test_img -o extended_l2=on 1M
>>> fi
>>> + # Write
On 11/3/23 17:51, Hanna Czenczek wrote:
> On 20.10.23 23:56, Andrey Drobyshev wrote:
>> Add _verify_du_delta() checker which is used to check that real disk
>> usage delta meets the expectations. For now we use it for checking that
>> subcluster-based discard/unmap operations lead to actual disk u
On 20.10.23 23:56, Andrey Drobyshev wrote:
Add _verify_du_delta() checker which is used to check that real disk
usage delta meets the expectations. For now we use it for checking that
subcluster-based discard/unmap operations lead to actual disk usage
decrease (i.e. PUNCH_HOLE operation is perfo
On 03.11.23 16:51, Hanna Czenczek wrote:
On 20.10.23 23:56, Andrey Drobyshev wrote:
[...]
@@ -528,6 +543,14 @@ for use_backing_file in yes no; do
else
_make_test_img -o extended_l2=on 1M
fi
+ # Write cluster #0 and discard its subclusters #0-#3
+ $QEMU_IO -c 'write