[MDEV-19536] Server crash or ASAN heap-use-after-free in is_temporary_table / read_statistics_for_tables_if_needed Created: 2019-05-21  Updated: 2019-10-07  Resolved: 2019-10-07

Status: Closed
Project: MariaDB Server
Component/s: Optimizer
Affects Version/s: 10.1, 10.2, 10.3, 10.4
Fix Version/s: 10.2.28, 10.1.42, 10.3.19, 10.4.9

Type: Bug Priority: Critical
Reporter: Elena Stepanova Assignee: Sergey Vojtovich
Resolution: Fixed Votes: 0
Labels: affects-tests


 Description   

Please note that it's a non-deterministic concurrent test.

SET use_stat_tables= PREFERABLY;
 
 
 
CREATE TABLE t1 (a INT);
 
CREATE OR REPLACE VIEW v1 AS SELECT * FROM t1;
 
 
 
--connect (con1,localhost,root,,)
 
--let $conid= `SELECT CONNECTION_ID()`
 
--delimiter $
 
--send
 
    LOOP
 
        CREATE OR REPLACE TABLE t1 (b INT);
 
    END LOOP $
 
--delimiter ;
 
 
 
--connection default
 
--let $run= 10000
 
while ($run)
 
{
 
    --error ER_VIEW_INVALID
 
    SHOW INDEX FROM v1;
 
    --dec $run
 
}
 
 
 
# Cleanup
 
--eval KILL $conid
 
DROP VIEW v1;
 
DROP TABLE IF EXISTS t1;

10.1 91efcc63

 
#3  <signal handler called>
 
#4  0x00005555964598fc in is_temporary_table (tl=0x7f153a9a8658) at /data/src/10.1/sql/sql_base.h:675
 
#5  0x0000555596573414 in read_statistics_for_tables_if_needed (thd=0x7f153c7cd070, tables=0x7f153a84a868) at /data/src/10.1/sql/sql_statistics.cc:3275
 
#6  0x000055559655ca05 in fill_schema_table_by_open (thd=0x7f153c7cd070, is_show_fields_or_keys=true, table=0x7f153a8ff088, schema_table=0x555597627920 <schema_tables+1792>, orig_db_name=0x7f1544f38790, orig_table_name=0x7f1544f387e0, open_tables_state_backup=0x7f1544f38800, can_deadlock=false) at /data/src/10.1/sql/sql_show.cc:4290
 
#7  0x000055559655da5b in get_all_tables (thd=0x7f153c7cd070, tables=0x7f153a846960, cond=0x0) at /data/src/10.1/sql/sql_show.cc:4814
 
#8  0x000055559656ceeb in get_schema_tables_result (join=0x7f153a8483a8, executed_place=PROCESSED_BY_JOIN_EXEC) at /data/src/10.1/sql/sql_show.cc:8280
 
#9  0x000055559650c5f0 in JOIN::exec_inner (this=0x7f153a8483a8) at /data/src/10.1/sql/sql_select.cc:2714
 
#10 0x000055559650bc9b in JOIN::exec (this=0x7f153a8483a8) at /data/src/10.1/sql/sql_select.cc:2562
 
#11 0x000055559650f229 in mysql_select (thd=0x7f153c7cd070, rref_pointer_array=0x7f153c7d1560, tables=0x7f153a846960, wild_num=0, fields=..., conds=0x0, og_num=0, order=0x0, group=0x0, having=0x0, proc_param=0x0, select_options=2684619520, result=0x7f153a848388, unit=0x7f153c7d0bb8, select_lex=0x7f153c7d12b8) at /data/src/10.1/sql/sql_select.cc:3499
 
#12 0x0000555596504913 in handle_select (thd=0x7f153c7cd070, lex=0x7f153c7d0af0, result=0x7f153a848388, setup_tables_done_option=0) at /data/src/10.1/sql/sql_select.cc:376
 
#13 0x00005555964d40e2 in execute_sqlcom_select (thd=0x7f153c7cd070, all_tables=0x7f153a846960) at /data/src/10.1/sql/sql_parse.cc:5951
 
#14 0x00005555964ca443 in mysql_execute_command (thd=0x7f153c7cd070) at /data/src/10.1/sql/sql_parse.cc:3038
 
#15 0x00005555964d7d5b in mysql_parse (thd=0x7f153c7cd070, rawbuf=0x7f153a845088 "SHOW INDEX FROM v1", length=18, parser_state=0x7f1544f3a1e0) at /data/src/10.1/sql/sql_parse.cc:7469
 
#16 0x00005555964c6455 in dispatch_command (command=COM_QUERY, thd=0x7f153c7cd070, packet=0x7f153e7f0071 "SHOW INDEX FROM v1", packet_length=18) at /data/src/10.1/sql/sql_parse.cc:1499
 
#17 0x00005555964c5213 in do_command (thd=0x7f153c7cd070) at /data/src/10.1/sql/sql_parse.cc:1131
 
#18 0x00005555966006c3 in do_handle_one_connection (thd_arg=0x7f153c7cd070) at /data/src/10.1/sql/sql_connect.cc:1330
 
#19 0x0000555596600409 in handle_one_connection (arg=0x7f153c7cd070) at /data/src/10.1/sql/sql_connect.cc:1242
 
#20 0x0000555596a33a3e in pfs_spawn_thread (arg=0x7f1542c39c70) at /data/src/10.1/storage/perfschema/pfs.cc:1861
 
#21 0x00007f1544bb84a4 in start_thread (arg=0x7f1544f3b700) at pthread_create.c:456
 
#22 0x00007f15435ced0f in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:97

Reproducible on 10.1-10.4.

 Comments   
Comment by Alice Sherepa [ 2019-07-05 ] 

SET use_stat_tables= PREFERABLY;
 
 
CREATE TABLE t1 (a INT);
 
CREATE OR REPLACE VIEW v1 AS SELECT * FROM t1;
 
 
--connect (con1,localhost,root,,)
 
--let $conid= `SELECT CONNECTION_ID()`
 
--delimiter $
 
--send
 
    LOOP
 
        CREATE OR REPLACE TABLE t1 (b INT);
 
    END LOOP $
 
--delimiter ;
 
 
--connection default
 
--let $run= 10000
 
while ($run)
 
{
 
    SELECT * FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = 'v1';
 
    --dec $run
 
}
 
 
# Cleanup
 
--eval KILL $conid
 
DROP VIEW v1;
 
DROP TABLE IF EXISTS t1;

Comment by Sergey Vojtovich [ 2019-10-02 ]

Test provided by Alice covers another problem, likely MDEV-19061. Although it didn't fail for me.

Comment by Sergey Vojtovich [ 2019-10-02 ]

igor, please review three patches in bb-10.1-svoj-MDEV-19536:
https://github.com/MariaDB/server/commit/e43791d4dc8feb2c02a08ce73c0bb0e2c320018c
https://github.com/MariaDB/server/commit/1b1b9013a5045983bbb3be68a2e8bc8f81186aea
https://github.com/MariaDB/server/commit/9374fe3d33b267bf12556a26119d6496625f659c

Comment by Alice Sherepa [ 2019-10-02 ]

svoj, Just checked the test, fails for me, but it is sporadic, please try with --repeat=N

10.4 97d82c34293375aa0ddf4

 
Version: '10.4.9-MariaDB-debug-log'  socket: '/10.4/mysql-test/var/tmp/mysqld.1.sock'  port: 16000  Source distribution
 
=================================================================
 
==29349==ERROR: AddressSanitizer: heap-use-after-free on address 0x6200000770f0 at pc 0x0000008287a2 bp 0x7f26ab2290a0 sp 0x7f26ab229090
 
READ of size 8 at 0x6200000770f0 thread T27
 
    #0 0x8287a1 in is_temporary_table(TABLE_LIST*) /10.4/sql/sql_base.h:622
 
    #1 0xb9d08e in statistics_for_tables_is_needed /10.4/sql/sql_statistics.cc:3215
 
    #2 0xb9dca4 in read_statistics_for_tables_if_needed(THD*, TABLE_LIST*) /10.4/sql/sql_statistics.cc:3344
 
    #3 0xb553b4 in fill_schema_table_by_open /10.4/sql/sql_show.cc:4653
 
    #4 0xb58b07 in get_all_tables(THD*, TABLE_LIST*, Item*) /10.4/sql/sql_show.cc:5292
 
    #5 0xb88fcb in get_schema_tables_result(JOIN*, enum_schema_table_state) /10.4/sql/sql_show.cc:8899
 
    #6 0xa73c21 in JOIN::exec_inner() /10.4/sql/sql_select.cc:4343
 
    #7 0xa71cd3 in JOIN::exec() /10.4/sql/sql_select.cc:4168
 
    #8 0xa7593a in mysql_select(THD*, TABLE_LIST*, unsigned int, List<Item>&, Item*, unsigned int, st_order*, st_order*, Item*, st_order*, unsigned long long, select_result*, st_select_lex_unit*, st_select_lex*) /10.4/sql/sql_select.cc:4600
 
    #9 0xa4b899 in handle_select(THD*, LEX*, select_result*, unsigned long) /10.4/sql/sql_select.cc:420
 
    #10 0x9cb79f in execute_sqlcom_select /10.4/sql/sql_parse.cc:6352
 
    #11 0x9b960e in mysql_execute_command(THD*) /10.4/sql/sql_parse.cc:3894
 
    #12 0x9d3f66 in mysql_parse(THD*, char*, unsigned int, Parser_state*, bool, bool) /10.4/sql/sql_parse.cc:7912
 
    #13 0x9acefa in dispatch_command(enum_server_command, THD*, char*, unsigned int, bool, bool) /10.4/sql/sql_parse.cc:1841
 
    #14 0x9a9d1e in do_command(THD*) /10.4/sql/sql_parse.cc:1359
 
    #15 0xd3caf0 in do_handle_one_connection(CONNECT*) /10.4/sql/sql_connect.cc:1412
 
    #16 0xd3c496 in handle_one_connection /10.4/sql/sql_connect.cc:1316
 
    #17 0x21eb0e2 in pfs_spawn_thread /10.4/storage/perfschema/pfs.cc:1862
 
    #18 0x7f26c2fe46b9 in start_thread (/lib/x86_64-linux-gnu/libpthread.so.0+0x76b9)
 
    #19 0x7f26c1bc741c in clone (/lib/x86_64-linux-gnu/libc.so.6+0x10741c)

Comment by Sergey Vojtovich [ 2019-10-02 ]

alice, thanks! Looks like I was wrong and your test case covering exactly this issue.

Comment by Igor Babaev [ 2019-10-07 ]

It's ok to push all 3 patches to 10.1

Generated at Thu Feb 08 08:52:26 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.