bdb thread locking info is same now for both rdonly and rdwr

[dorinhogea]

/plugin-branch onebdblock

[roborivers]

Cbuild submission: Success ✓.
Regression testing: Success ✓.

The first 10 failing tests are:
sirace [setup failed with core dumped]
catchup_idle [setup failed with core dumped]
cinsert_linearizable [setup failed with core dumped]
maxtable [setup failed with core dumped]
socksql_master_swings [setup failed with core dumped]
truncatesc [setup failed with core dumped] **quarantined**
phys_rep_tiered [setup failed with core dumped]
phys_rep_tiered_nosource_generated [setup failed with core dumped]
phys_rep_tiered_firstfile_generated [setup failed with core dumped]
truncatesc_offline_generated [setup failed with core dumped] **quarantined**

[roborivers]

Cbuild submission: Success ✓.
Regression testing: Success ✓.

The first 10 failing tests are:
consumer_non_atomic_default_consumer_generated **quarantined**
remsql_locks_rte_connect_generated **quarantined**
remsql_locks **quarantined**
reco-ddlk-sql [timeout] **quarantined**

[dorinhogea]

In case it is not clear (look at bdb_thread_event), the changed code does not LOCK, it only prepares the thread for later locking. Originally, we used to do things different depending on the flag. That is not true anymore.
The PR does the following:

• replace 4 COMDB2_THR_EVENT_START/DONE_RDONLY/RDWR with just two flags COMDB2_THR_EVENT_START/DONE; these flags are used with backend_thread_event
• remove BDBTHR_EVENT_START/DONE_RDONLY/RDWR with just to flags BDBTHR_EVENT_START/DONE; these flags are used with bdb_thread_event
• fix cases when we use literals 0,1 instead of flag
• fix cases when bdbthr_thread_event uses backend_thread_event flags
• fix cases when backend_thread_event uses bdbthr_threadevent flags

[dorinhogea]

This is the important part; rdonly and rdwr are identical