Details
-
Bug
-
Status: Closed (View Workflow)
-
Critical
-
Resolution: Fixed
-
5.5.30, 5.5(EOL), 10.0(EOL), 10.1(EOL), 10.2(EOL), 10.3(EOL)
-
Linux x64
Description
During a generic DML stress test of Aria tables, after a while a debug build crashed a few times with the same stack. There is no exact testcase to trigger a problem.
Version: '5.5.30-MariaDB-debug' Source distribution
|
Thread 18:
|
Invalid read of size 8
|
at: make_lock_and_pin (ma_pagecache.c:2587)
|
by: flush_cached_blocks (ma_pagecache.c:4378)
|
by: flush_pagecache_blocks_int (ma_pagecache.c:4689)
|
by: flush_pagecache_blocks_with_filter (ma_pagecache.c:4805)
|
by: ma_checkpoint_background (ma_checkpoint.c:666)
|
by: start_thread (pthread_create.c:309)
|
by: clone (clone.S:115)
|
Address 0x10 is not stack'd, malloc'd or (recently) free'd
|
The line of code in question is a DBUG_ASSERT.
2587 DBUG_ASSERT(block->hash_link->block == block);
|
(gdb) p block
|
$1 = (PAGECACHE_BLOCK_LINK *) 0x27dbafa0
|
(gdb) p block->hash_link
|
$2 = (struct st_pagecache_hash_link *) 0x0
|
Attachments
Issue Links
- relates to
-
-
- Closed
-
-
-
- Confirmed
-