Merging upstream branch-2.16 [RH git: 897937f6d3]

Merging upstream branch-2.16 [RH git: e4d6d108a3]
netdev: Fix a race condition in deletion of offloaded flows.
netdev: Move port flush after datapath reconfiguration.

Merging upstream branch-2.16 [RH git: 970214133d]

Merging upstream branch-2.16 [RH git: ac5da61d03]
windows: Fix NXM_OF_IP_TOS issue

Merging upstream branch-2.16 [RH git: b2df459e49]
idl: Fix use-after-free when destroying an IDL loop.

Merging upstream branch-2.16 [RH git: bba08b5363]
netdev-dpcls: Make subtable reprobe thread-safe.
packet: Ensure packet base is always non-NULL.
idlc: Avoid accessing member within NULL idl index cursors.

Merging upstream branch-2.16 [RH git: 7b6570c65f]

Merging upstream branch-2.16 [RH git: c5ad7f71c5]

Merging upstream branch-2.16 [RH git: 4541c91b99]
netdev: fix vlan and ipv4 parsing in avx512

Merging upstream branch-2.16 [RH git: 9d51785142]
util: Fix tunnel key attr for GTP-U.
idl: Only process successful txn in ovsdb_idl_loop_run.

Merging upstream branch-2.16 [RH git: 6e6f66ffd0]
dpif-upcall: Fix n_revalidators on upcall show.

Merging upstream branch-2.16 [RH git: 513117cbb0]

ovsdb: transaction: Keep one entry in the transaction history. [RH git: 7665f42d12] (#2044621)
connect
downloading of the whole database, since the
sync, because this last transaction will not be used
by: Mike Pattrick <mkp@redhat.com>
by: Han Zhou <hzhou@ovn.org>
off-by: Ilya Maximets <i.maximets@ovn.org>
at: https://bugzilla.redhat.com/2044621
off-by: Ilya Maximets <i.maximets@redhat.com>

Merging upstream branch-2.16 [RH git: d202cd6da1]
idl: ovsdb_idl_loop_destroy must also destroy the committing txn.
cs: Clear last_id on reconnect if condition changes in-flight.
flow: Skip flow reply if it exceeds the maximum message size.
cs: Fix ignoring of the last id from the initial monitor reply. (#2044624)

ovsdb: storage: Randomize should_snapshot checks when the minimum time passed. [RH git: abe61535ca] (#2044614)
20 minutes. It's a random value
15 seconds for OVN_Southbound database in a larger scale
scheduling a quick retry if the minimal time already
2 min delay
schedules are only used for the error cases, and
schedule after the successful compaction.
off-by: Ilya Maximets <i.maximets@ovn.org>
by: Han Zhou <hzhou@ovn.org>
by: Dumitru Ceara <dceara@redhat.com>
at: https://bugzilla.redhat.com/2044614
off-by: Ilya Maximets <i.maximets@redhat.com>

raft: Only allow followers to snapshot. [RH git: 915efc8c00] (#2044614)
server cluster
node clusters for which we
by: Han Zhou <hzhou@ovn.org>
off-by: Dumitru Ceara <dceara@redhat.com>
off-by: Ilya Maximets <i.maximets@ovn.org>
at: https://bugzilla.redhat.com/2044614
off-by: Ilya Maximets <i.maximets@redhat.com>