Details
Major Description
--source include/have_innodb.inc
|
|
|
--error ER_WRONG_VALUE_FOR_VAR
|
SET GLOBAL innodb_monitor_reset_all= '%', innodb_compression_algorithm= foo; |
|
10.5 080522dc ASAN non-debug |
2020-09-29 14:20:40 0 [Note] /data/bld/10.5-rel-asan-nightly/bin/mariadbd: Shutdown complete
|
|
|
Warning: Memory not freed: 32
|
|
|
=================================================================
|
==2843902==ERROR: LeakSanitizer: detected memory leaks
|
|
|
Direct leak of 32 byte(s) in 1 object(s) allocated from:
|
#0 0x7f3944b01bc8 in malloc (/lib/x86_64-linux-gnu/libasan.so.5+0x10dbc8)
|
#1 0x562a14886463 in my_malloc /data/src/10.5/mysys/my_malloc.c:88
|
#2 0x562a148869f2 in my_strdup /data/src/10.5/mysys/my_malloc.c:231
|
#3 0x562a13f59d4f in innodb_monitor_validate /data/src/10.5/storage/innobase/handler/ha_innodb.cc:17827
|
#4 0x562a12e1c242 in sys_var_pluginvar::do_check(THD*, set_var*) /data/src/10.5/sql/sql_plugin.cc:3546
|
#5 0x562a12b9e851 in sys_var::check(THD*, set_var*) /data/src/10.5/sql/set_var.cc:246
|
#6 0x562a12ba2134 in set_var::check(THD*) /data/src/10.5/sql/set_var.cc:811
|
#7 0x562a12ba313e in sql_set_variables(THD*, List<set_var_base>*, bool) /data/src/10.5/sql/set_var.cc:739
|
#8 0x562a12e11af5 in mysql_execute_command(THD*) /data/src/10.5/sql/sql_parse.cc:5009
|
#9 0x562a12dd3b1c in mysql_parse(THD*, char*, unsigned int, Parser_state*, bool, bool) /data/src/10.5/sql/sql_parse.cc:7994
|
#10 0x562a12dffce0 in dispatch_command(enum_server_command, THD*, char*, unsigned int, bool, bool) /data/src/10.5/sql/sql_parse.cc:1867
|
#11 0x562a12e05ac4 in do_command(THD*) /data/src/10.5/sql/sql_parse.cc:1348
|
#12 0x562a1319fd2c in do_handle_one_connection(CONNECT*, bool) /data/src/10.5/sql/sql_connect.cc:1410
|
#13 0x562a131a08ac in handle_one_connection /data/src/10.5/sql/sql_connect.cc:1312
|
#14 0x562a13d81648 in pfs_spawn_thread /data/src/10.5/storage/perfschema/pfs.cc:2201
|
#15 0x7f394497e608 in start_thread /build/glibc-ZN95T4/glibc-2.31/nptl/pthread_create.c:477
|
|
|
SUMMARY: AddressSanitizer: 32 byte(s) leaked in 1 allocation(s).
|
|
10.5 080522dc ASAN debug |
Warning: Memory not freed: 32
|
Warning: 32 bytes lost at 0x60e000044ad0, allocated by T@0 at 0x55b0b289a69d, mysys/my_malloc.c:88, mysys/my_malloc.c:231, handler/ha_innodb.cc:17827, sql/sql_plugin.cc:3546, sql/set_var.cc:246, sql/set_var.cc:739, sql/sql_parse.cc:5009
|
Memory lost: 32 bytes in 1 chunks
|
Warning: 32 bytes lost at 0x60e000044ad0, allocated by T@0 at mysys/my_malloc.c:88, mysys/my_malloc.c:231, handler/ha_innodb.cc:17827, sql/sql_plugin.cc:3546, sql/set_var.cc:246, sql/set_var.cc:811, sql/set_var.cc:739, sql/sql_parse.cc:5009
|
Memory lost: 32 bytes in 1 chunks
|
|
10.2 84261653 Valgrind |
Warning: Memory not freed: 16
|
==2844360== 16 bytes in 1 blocks are definitely lost in loss record 1 of 2
|
==2844360== at 0x483B7F3: malloc (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
|
==2844360== by 0x1180925: my_malloc (my_malloc.c:101)
|
==2844360== by 0x1181024: my_strdup (my_malloc.c:241)
|
==2844360== by 0xBBBFAD: innodb_monitor_validate(THD*, st_mysql_sys_var*, void*, st_mysql_value*) (ha_innodb.cc:18655)
|
==2844360== by 0x71F62F: sys_var_pluginvar::do_check(THD*, set_var*) (sql_plugin.cc:3452)
|
==2844360== by 0x62BFCC: sys_var::check(THD*, set_var*) (set_var.cc:248)
|
==2844360== by 0x62D9D1: set_var::check(THD*) (set_var.cc:789)
|
==2844360== by 0x62D6B2: sql_set_variables(THD*, List<set_var_base>*, bool) (set_var.cc:731)
|
==2844360== by 0x70467D: mysql_execute_command(THD*) (sql_parse.cc:4583)
|
==2844360== by 0x70E550: mysql_parse(THD*, char*, unsigned int, Parser_state*, bool, bool) (sql_parse.cc:7733)
|
==2844360== by 0x6FC85F: dispatch_command(enum_server_command, THD*, char*, unsigned int, bool, bool) (sql_parse.cc:1823)
|
==2844360== by 0x6FB35A: do_command(THD*) (sql_parse.cc:1377)
|
==2844360== by 0x85AA12: do_handle_one_connection(CONNECT*) (sql_connect.cc:1336)
|
==2844360== by 0x85A777: handle_one_connection (sql_connect.cc:1241)
|
==2844360== by 0x10CDE41: pfs_spawn_thread (pfs.cc:1869)
|
==2844360== by 0x4941608: start_thread (pthread_create.c:477)
|
Reproducible on 10.1-10.5, with different verbosity of the final diagnostics.
Attachments
Activity
elenst, I have seen in the past that SAFEMALLOC has hidden some leaks from AddressSanitizer. I would recommend cmake -DWITH_SAFEMALLOC=OFF -DWITH_ASAN=ON.
When it comes to this bug, it is obvious when looking at innodb_monitor_validate():
/* monitor_name could point to memory from MySQL |
or buff[]. Always dup the name to memory allocated
|
by InnoDB, so we can access it in another callback
|
function innodb_monitor_update() and free it appropriately */
|
if (name) { |
monitor_name = my_strdup(PSI_INSTRUMENT_ME,
|
name, MYF(0));
|
} else { |
return(1); |
}
|
|
|
ret = innodb_monitor_valid_byname(save, monitor_name);
|
|
|
if (ret) { |
/* Validation failed */ |
my_free(monitor_name);
|
} else { |
/* monitor_name will be freed in separate callback function |
innodb_monitor_update(). Assert "save" point to
|
the "monitor_name" variable */
|
ut_ad(*static_cast<char**>(save) == monitor_name); |
}
|
This logic is bound to fail if a subsequent assignment is flagged invalid.
The bug is not very important in itself, but it would be useful for further testing to understand why LeakSanitizer only reports the leak on a non-debug build with ASAN. Or maybe it will be different on different machines/builds.