If a client uses a call_id > 0xffff, we don't find fragmented calls based on the call_id. This happens for example in a very long running winbindd process and causes wbinfo -u to return nothing. This was noticed by a failing this: [1825(12537)/2242 at 3h1m13s] samba4.winbind.struct(ad_dc:local) Running WINBINDD_LIST_USERS (struct based) WARNING!: ../source4/torture/winbind/struct_based.c:615: WINBINDD_LIST_USERS returned -1, expected 1 UNEXPECTED(failure): samba4.winbind.struct.list_users(ad_dc:local) REASON: Exception: Exception: ../source4/torture/winbind/struct_based.c:644: Expression `ok' failed: failed to get user list ../source3/rpc_client/cli_pipe.c:568: RPC fault code DCERPC_NCA_S_PROTO_ERROR received from host addc!
Created attachment 13994 [details] Patch for 4.8 cherry-picked from master
Pushed to autobuild-v4-8-test.
Pushed to v4-8-test. Closing out bug report. Thanks!
Sorry, I missed to say: we need this also in older versions
Re-assigning to Metze for backports.
Created attachment 14149 [details] Patches for v4-7-test
Reassigning to Karolin for inclusion in 4.7 (is already in 4.8).
(In reply to Ralph Böhme from comment #7) Pushed to autobuild-v4-7-test.
(In reply to Karolin Seeger from comment #8) Pushed to v4-7-test. Closing out bug report. Thanks!