On Mon, 20 Jan 2025 at 15:09, Lionel Cons <lionelcons1972@xxxxxxxxx> wrote:
>
> Does anyone know why the per-fileobjet attribute FATTR4_WORD0_FSID
> always returns { .minor=0, .major=0 } for a Linux nfsd?
> This does not sound right.
>
> Ref https://datatracker.ietf.org/doc/html/rfc5661#section-5.8.1.9
Never mind. The colleague found a bug in our XDR/cache code. Wireshark
shows that fsid contains non-zero data in the .major field (why not
.minor ?!), and it was our fault of not decoding it into the correct
struct member.
No action required on Linux nfsd side. Sorry
Lionel
