On Fri, Sep 04, 2020 at 10:07:22AM -0400, Chuck Lever wrote: > My primary concern is that the result of a file copy operation should > look the same on NFS/TCP (with READ_PLUS) and NFS/RDMA (with SEEK_DATA/HOLE). I'm not sure what you mean. I don't see the spec providing any guarantee of consistency between READ_PLUS and SEEK. It also doesn't guarantee that the results tell you anything about how the file is actually stored--a returned "hole" could represent an unallocated segment, or a fully allocated segment that's filled with zeroes, or some combination. So, for example, if you implemented an optimized copy that used ALLOCATE, DEALLOCATE, SEEK and/or READ_PLUS to avoid reading and writing a lot of zeroes--there's no guarantee that the target file would end up allocated in the same way as the source. --b.
