[MDEV-21369] rpl.rpl_timezone fails with valgrind: Use of uninitialised value Created: 2019-12-17  Updated: 2023-11-27  Resolved: 2019-12-23

Status: Closed
Project: MariaDB Server
Component/s: Server, Tests
Affects Version/s: 10.2
Fix Version/s: 10.2.31

Type: Bug Priority: Critical
Reporter: Elena Stepanova Assignee: Alexey Botchkov
Resolution: Fixed Votes: 0
Labels: None


 Description   

10.4-enterprise 809261c6

$ perl ./mtr rpl.rpl_timezone --valgrind
 
==14430== Thread 13:
==14430== Use of uninitialised value of size 8
==14430==    at 0xBA4D9F: inline_mysql_mutex_trylock (mysql_thread.h:728)
==14430==    by 0xBA9251: Table_cache_instance::lock_and_check_contention(unsigned int, unsigned int) (table_cache.cc:168)
==14430==    by 0xBA5A18: tc_acquire_table(THD*, TDC_element*) (table_cache.cc:417)
==14430==    by 0xBA76A2: tdc_acquire_share(THD*, TABLE_LIST*, unsigned int, TABLE**) (table_cache.cc:872)
==14430==    by 0x8645B9: open_table(THD*, TABLE_LIST*, Open_table_context*) (sql_base.cc:1944)
==14430==    by 0x86866A: open_and_process_table(THD*, TABLE_LIST*, unsigned int*, unsigned int, Prelocking_strategy*, bool, Open_table_context*) (sql_base.cc:3850)
==14430==    by 0x869913: open_tables(THD*, DDL_options_st const&, TABLE_LIST**, unsigned int*, unsigned int, Prelocking_strategy*) (sql_base.cc:4324)
==14430==    by 0x86BA8F: open_and_lock_tables(THD*, DDL_options_st const&, TABLE_LIST*, bool, unsigned int, Prelocking_strategy*) (sql_base.cc:5217)
==14430==    by 0x860467: open_and_lock_tables(THD*, TABLE_LIST*, bool, unsigned int, Prelocking_strategy*) (sql_base.h:271)
==14430==    by 0x86B310: open_n_lock_single_table(THD*, TABLE_LIST*, thr_lock_type, unsigned int, Prelocking_strategy*) (sql_base.cc:5060)
==14430==    by 0x8C494F: Delayed_insert::open_and_lock_table() (sql_insert.cc:2990)
==14430==    by 0x8C4C4D: handle_delayed_insert (sql_insert.cc:3093)
==14430==    by 0x14F87EC: pfs_spawn_thread (pfs.cc:1862)
==14430==    by 0x4E3F4A3: start_thread (pthread_create.c:456)
==14430==    by 0x6938D0E: clone (clone.S:97)

Not reproducible on 10.3-enterprise or on 10.4 CS (both fail with "blocks still reachable", which is probably another matter, many tests fail with it.



 Comments   
Comment by Elena Stepanova [ 2019-12-17 ]

Raised priority for initial analysis/evaluation

Comment by Julien Fritsch [ 2019-12-17 ]

tturenko and abychko ^^

Comment by Alexey Bychko (Inactive) [ 2019-12-17 ]

we don't have valgrind build for ES atm. with GCloud or another cloud we will allocate resources for valgrind build, will be discussed in 2020

Comment by Oleksandr Byelkin [ 2019-12-17 ]

if rollback 9f90e0ffe7d0d79a329ce2bad93fbba780640ccd:

git revert -n 9f90e0ffe7d0d79a329ce2bad93fbba780640ccd

valgrind do not report the problem any more (at least for me).

Generated at Thu Feb 08 09:06:37 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.