81 lines
3.8 KiB
C++
81 lines
3.8 KiB
C++
/**
|
|
* @file bug653.cpp regression case for bug 653 ("Memory corruption when users with long hostnames that can no the resolved are loaded into MaxScale")
|
|
*
|
|
* - CREATE USER 'user_with_very_long_hostname'@'very_long_hostname_that_can_not_be_resolved_and_it_probably_caused_crash.com.net.org' IDENTIFIED BY 'old';
|
|
* - try to connect using user 'user_with_very_long_hostname'
|
|
* - DROP USER 'user_with_very_long_hostname'@'very_long_hostname_that_can_not_be_resolved_and_it_probably_caused_crash.com.net.org'
|
|
* - check MaxScale is alive
|
|
*/
|
|
|
|
/*
|
|
Mark Riddoch 2014-12-16 13:17:25 UTC
|
|
Program received signal SIGSEGV, Segmentation fault.
|
|
0x00007ffff49385ac in free () from /lib64/libc.so.6
|
|
Missing separate debuginfos, use: debuginfo-install glibc-2.12-1.47.el6_2.12.x86_64 keyutils-libs-1.4-4.el6.x86_64 krb5-libs-1.10.3-10.el6_4.2.x86_64 libaio-0.3.107-10.el6.x86_64 libcom_err-1.41.12-14.el6.x86_64 libgcc-4.4.7-4.el6.x86_64 libselinux-2.0.94-5.3.el6_4.1.x86_64 libstdc++-4.4.7-4.el6.x86_64 nss-pam-ldapd-0.7.5-14.el6_2.1.x86_64 nss-softokn-freebl-3.14.3-10.el6_5.x86_64 openssl-1.0.1e-16.el6_5.15.x86_64 zlib-1.2.3-29.el6.x86_64
|
|
(gdb) where
|
|
#0 0x00007ffff49385ac in free () from /lib64/libc.so.6
|
|
#1 0x000000000041d421 in add_mysql_users_with_host_ipv4 (users=0x72c4c0,
|
|
user=0x739030 "u3", host=0x739033 "aver.log.hostname.to.overflow.the.buffer",
|
|
passwd=0x73905c "", anydb=0x739089 "Y", db=0x0) at dbusers.c:291
|
|
#2 0x000000000041e302 in getUsers (service=0x728ef0, users=0x72c4c0)
|
|
at dbusers.c:742
|
|
#3 0x000000000041cf97 in load_mysql_users (service=0x728ef0) at dbusers.c:99
|
|
#4 0x00000000004128c7 in serviceStartPort (service=0x728ef0, port=0x729b70)
|
|
at service.c:227
|
|
#5 0x0000000000412e27 in serviceStart (service=0x728ef0) at service.c:365
|
|
#6 0x0000000000412f00 in serviceStartAll () at service.c:413
|
|
#7 0x000000000040b592 in main (argc=2, argv=0x7fffffffe108) at gateway.c:1750
|
|
Comment 1 Mark Riddoch 2014-12-16 13:18:09 UTC
|
|
The problem is a buffer overrun in normalise_hostname. Fix underway.
|
|
Comment 2 Mark Riddoch 2014-12-16 15:45:59 UTC
|
|
Increased buffer size to prevent overrun issue
|
|
Comment 3 Timofey Turenko 2014-12-22 15:39:32 UTC
|
|
I'm not sure I understand the bug correctly.
|
|
But 60-chars long host name does not cause problem (longer is not possible "String 'very_long_hostname_that_can_not_be_resolved_and_it_probably_caused_cra' is too long for host name (should be no longer than 60)"
|
|
*/
|
|
|
|
|
|
#include <iostream>
|
|
#include "testconnections.h"
|
|
|
|
int main(int argc, char *argv[])
|
|
{
|
|
TestConnections * Test = new TestConnections(argc, argv);
|
|
Test->set_timeout(50);
|
|
Test->connect_maxscale();
|
|
|
|
Test->tprintf("Creating user with old style password\n");
|
|
Test->try_query(Test->conn_rwsplit,
|
|
(char *) "CREATE USER 'user_long_host11'@'very_long_hostname_that_probably_caused_crashhh.com.net.org' IDENTIFIED BY 'old'");
|
|
Test->try_query(Test->conn_rwsplit,
|
|
(char *) "GRANT ALL PRIVILEGES ON *.* TO 'user_long_host11'@'very_long_hostname_that_probably_caused_crashhh.com.net.org' WITH GRANT OPTION");
|
|
sleep(10);
|
|
|
|
Test->tprintf("Trying to connect using user with old style password\n");
|
|
MYSQL * conn = open_conn(Test->rwsplit_port, Test->maxscale_IP, (char *) "user_long_host11", (char *) "old",
|
|
Test->ssl);
|
|
|
|
if ( mysql_errno(conn) != 0 )
|
|
{
|
|
Test->tprintf("Connections is not open as expected\n");
|
|
}
|
|
else
|
|
{
|
|
Test->add_result(1, "Connections is open for the user with bad host\n");
|
|
}
|
|
if (conn != NULL)
|
|
{
|
|
mysql_close(conn);
|
|
}
|
|
|
|
Test->try_query(Test->conn_rwsplit,
|
|
(char *) "DROP USER 'user_long_host11'@'very_long_hostname_that_probably_caused_crashhh.com.net.org'");
|
|
Test->close_maxscale_connections();
|
|
|
|
Test->check_maxscale_alive();
|
|
|
|
int rval = Test->global_result;
|
|
delete Test;
|
|
return rval;
|
|
}
|