Good identification Mr. Left,
Just adding some addtional information
file:= \mysql-5.6.7-rc\mysys\mf_iocache.c Line:= 262
Mutex will be never be cleared as sizeof(info) will return 4 as info
is pointer to IO_CACHE, so the first four bytes of
info->append_buffer_lock will be set to 0;
Shaukat Mahmood Ahmad
On Wed, Oct 31, 2012 at 6:06 PM, Jon Olav Hauglid
> On 10/27/2012 07:33 AM, Mr. Left wrote:
>> Hi, all
>> I am reading the source code of mysql 5.6.6-m9, and i found the code below
>> in function:init_io_cache
>> /* Clear mutex so that safe_mutex will notice that it's not initialized */
>> memset(&info->append_buffer_lock, 0, sizeof(info));
>> I want to know why use sizeof(info) instead of
>> sizeof(info->append_buffer_lock) ?
> Good question :-)
> It certainly looks like a bug.
> Thanks for making us aware of this issue!
> I've reported it as http://bugs.mysql.com/bug.php?id=67437
> --- Jon Olav Hauglid, Principal Software Engineer MySQL, Oracle.
> MySQL Internals Mailing List
> For list archives: http://lists.mysql.com/internals
> To unsubscribe: http://lists.mysql.com/internals