From: Mattias Jonsson
Date: September 9 2008 7:11pm
Subject: Re: bzr commit into mysql-5.1 branch (ingo.struewing:2735) Bug#37958
List-Archive: http://lists.mysql.com/commits/53642
Message-Id: <48C6CA4B.3060709@sun.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 8bit

Hi,

It still crashes on the 'SELECT * FROM t1' see below.

Ingo, I can try to setup an account on my machine for you, if you want? 
(or try to get some time to investigate myself...)

/Mattias

Output from test:

main.plugin                    [ fail ]

mysqltest: At line 17: query 'SELECT * FROM t1' failed: 2013: Lost 
connection to MySQL server during query

The result from queries just before the failure was:
CREATE TABLE t1(a int) ENGINE=EXAMPLE;
Warnings:
Warning	1286	Unknown table engine 'EXAMPLE'
Warning	1266	Using storage engine MyISAM for table 't1'
DROP TABLE t1;
INSTALL PLUGIN example SONAME 'ha_example.so';
INSTALL PLUGIN EXAMPLE SONAME 'ha_example.so';
ERROR HY000: Function 'EXAMPLE' already exists
UNINSTALL PLUGIN example;
INSTALL PLUGIN example SONAME 'ha_example.so';
CREATE TABLE t1(a int) ENGINE=EXAMPLE;
SELECT * FROM t1;

More results from queries before failure can be found in 
/Users/mattiasj/clones/bzrroot/test-51-bugteam/mysql-test/var/log/plugin.log

Here is a bt at the crash:
Program received signal EXC_ARITHMETIC, Arithmetic exception.
0x91bc2466 in hash_search ()
(gdb) bt
#0  0x91bc2466 in hash_search ()
#1  0x00ee40ee in get_share (table_name=0x10724a8 "./test/t1", 
table=0x1071a18) at ha_example.cc:177
#2  0x00ee4476 in ha_example::open (this=0x1074628, name=0x10724a8 
"./test/t1", mode=2, test_if_locked=2) at ha_example.cc:294
#3  0x00232241 in handler::ha_open (this=0x1074628, table_arg=0x1071a18, 
name=0x10724a8 "./test/t1", mode=2, test_if_locked=2) at handler.cc:2025
#4  0x00167a77 in open_table_from_share (thd=0x1026418, share=0x1072228, 
alias=0x1069b50 "t1", db_stat=39, prgflag=44, ha_open_flags=0, 
outparam=0x1071a18, is_create_table=false) at table.cc:1858
#5  0x0015a2a5 in open_unireg_entry (thd=0x1026418, entry=0x1071a18, 
table_list=0x1069b58, alias=0x1069b50 "t1", cache_key=0xb00616bd "test", 
cache_key_length=8, mem_root=0xb00618e4, flags=0) at sql_base.cc:3881
#6  0x0015cf39 in open_table (thd=0x1026418, table_list=0x1069b58, 
mem_root=0xb00618e4, refresh=0xb0061927, flags=0) at sql_base.cc:2904
#7  0x0015daba in open_tables (thd=0x1026418, start=0xb00619b4, 
counter=0xb0061998, flags=0) at sql_base.cc:4515
#8  0x0015e30d in open_and_lock_tables_derived (thd=0x1026418, 
tables=0x1069b58, derived=true) at sql_base.cc:4915
#9  0x00117aa0 in open_and_lock_tables (thd=0x1026418, tables=0x1069b58) 
at mysql_priv.h:1530
#10 0x00104245 in execute_sqlcom_select (thd=0x1026418, 
all_tables=0x1069b58) at sql_parse.cc:4791
#11 0x0010b75a in mysql_execute_command (thd=0x1026418) at sql_parse.cc:2134
#12 0x00114c42 in mysql_parse (thd=0x1026418, inBuf=0x1069a28 "SELECT * 
FROM t1", length=16, found_semicolon=0xb0062df4) at sql_parse.cc:5717
#13 0x00115f1c in dispatch_command (command=COM_QUERY, thd=0x1026418, 
packet=0xce6019 "SELECT * FROM t1", packet_length=16) at sql_parse.cc:1152
#14 0x001171f8 in do_command (thd=0x1026418) at sql_parse.cc:809
#15 0x001026fd in handle_one_connection (arg=0x1026418) at 
sql_connect.cc:1115
#16 0x91b566f5 in _pthread_start ()
#17 0x91b565b2 in thread_start ()

my quick guess is an error in the example handler...

Ingo Strüwing wrote:
> Hi Sergey,
> 
> Sergey Vojtovich, 09.09.2008 16:34:
> ...
>> On Tue, Sep 09, 2008 at 04:29:57PM +0200, Ingo Strüwing wrote:
> ...
>>> Sergey, thanks for the review. Why do you think it's a different problem?
>> At least because it is not obvious how hton2plugin mapping could affect
>> EXAMPLE share. Of course unless it is some kind of tricky memory corruption.
> 
> I hope, the array overrun was the cause for this.
> 
> There is one doubt left though. One run of plugin.test was not
> sufficient to overrun the array. OTOH, on my machine it crashed only on
> slot 20 of 15. So it didn't crash on the first overrun. Three test runs
> were required to come to slot 20.
> 
> On Mac it might crash on the first overrun. If some "INSTALL PLUGIN"
> have been executed in the same server life time before plugin.test, the
> crash could still result from the overrun.
> 
> Regards
> Ingo