On Sat, Nov 14, 2020 at 10:47:53AM +0000, Christoph Hellwig wrote: > Instead of the loop to check if the attr existed can't we just check > for the ENODATA return value from removexattr? Yes. > Also I think with this series we can drop the libattr dependency > for xfstests, can't we? I'll do that in a subsequent patch. --D
