The MDS/client are already accidentally doing the right thing unless
they are running on a big-endian machine.
Credit to Venky Shankar for originally hypothesizing an endianness issue
with the frag_t.
Fixes: https://tracker.ceph.com/issues/73792
Signed-off-by: Patrick Donnelly <pdonnell@ibm.com>
}
void frag_t::encode(ceph::buffer::list& bl) const {
- ceph::encode_raw(_enc, bl);
+ ceph::encode(_enc, bl);
}
void frag_t::decode(ceph::buffer::list::const_iterator& p) {
- __u32 v;
- ceph::decode_raw(v, p);
- _enc = v;
+ ceph::decode(_enc, p);
}
void frag_t::dump(ceph::Formatter *f) const {