On 2021/9/29 14:45, Nikolay Borisov wrote:
On 29.09.21 г. 3:44, Qu Wenruo wrote:
The planned fix is titled "btrfs-progs: receive: fallback to buffered
copy if clone failed".
The test case itself will create two send streams, and the 2nd stream is
an incremental stream with a clone command in it.
Using different mount options we are able to create a situation where
clone source and destination have different NODATASUM flags, which is
prohibited inside btrfs.
The planned fix will make btrfs receive to fall back to buffered write
to copy the data from the source file.
Signed-off-by: Qu Wenruo <wqu@xxxxxxxx>
If this is fixed by btrfs-progs fix shouldn't the test be part of
btrfs-progs and not the kernel test suite?
Makes sense.
Would move it to btrfs-progs misc-tests, which can be more flex there.
THanks,
Qu
