A list for people who work on the MySQL code

Guillaume Morin | 1 Aug 2002 13:35

using my_thread_* in a library

Hi folks,

I am the maintainer of NSS MySQL which is, as you may have guessed, a
nss module using MySQL as backend. It shares a mysql connection between
threads. As explained in the documentation, I use my_thread_init and
my_thread_end functions to properly manage that situation.

As you may know, a nss module provides UNIX user lookup functions such
as getgrnam. That means that basically any applications can call the nss
modules function even the ones which uses MySQL too. And there lies the
issue. Indeed, if the application follows this path :

- calls mysql_connect or mysql_thread_init
...
- calls a function like getgrnam 
- does MySQL stuff

The application will segfault since the nss module has called
my_thread_end. I do not see any solutions to this because if I remove
calls to my_thread_end I create a memory leak. I've read the source of
my_thread_init, it seems that there is no way to know if the MySQL data
was already allocated (in this case, I would not call my_thread_end).

Does anyone have an idea ? Maybe, my_thread_init should be modified to
return a special error code if the MySQL data was already allocated...

TIA.

PS: sorry if it is the wrong list, I already tried to ask on the
general list and got no answer.

(Continue reading)

Sinisa Milivojevic | 1 Aug 2002 14:10

Picon

Re: using my_thread_* in a library

On Thu, 1 Aug 2002 13:35:38 +0200
Guillaume Morin <guillaume <at> morinfr.org> wrote:

> Hi folks,
> 
> I am the maintainer of NSS MySQL which is, as you may have guessed, a
> nss module using MySQL as backend. It shares a mysql connection between
> threads. As explained in the documentation, I use my_thread_init and
> my_thread_end functions to properly manage that situation.
> 
> As you may know, a nss module provides UNIX user lookup functions such
> as getgrnam. That means that basically any applications can call the nss
> modules function even the ones which uses MySQL too. And there lies the
> issue. Indeed, if the application follows this path :
> 
> - calls mysql_connect or mysql_thread_init
> ...
> - calls a function like getgrnam 
> - does MySQL stuff
> 
> The application will segfault since the nss module has called
> my_thread_end. I do not see any solutions to this because if I remove
> calls to my_thread_end I create a memory leak. I've read the source of
> my_thread_init, it seems that there is no way to know if the MySQL data
> was already allocated (in this case, I would not call my_thread_end).
> 
> Does anyone have an idea ? Maybe, my_thread_init should be modified to
> return a special error code if the MySQL data was already allocated...
> 
> TIA.

(Continue reading)

Sinisa | 1 Aug 2002 14:19

Picon

bk commit into 4.0 tree

Below is the list of changes that have just been committed into a
4.0 repository of Sinisa. When Sinisa does a push these changes will
be propagated to the main repository and, within 24 hours after the
push, to the public repository.
For information on how to access the public repository
see http://www.mysql.com/doc/I/n/Installing_source_tree.html

ChangeSet <at> 1.1138, 2002-08-01 15:19:06+03:00, Sinisa <at> sinisa.nasamreza.org
  few small fixes

  BitKeeper/etc/ignore
    1.125 02/08/01 15:19:04 Sinisa <at> sinisa.nasamreza.org +1 -0
    Added mysys/ste5KbMa to the ignore list

  include/my_semaphore.h
    1.4 02/08/01 15:19:04 Sinisa <at> sinisa.nasamreza.org +2 -2
    A small fix for BSDi

  sql/sql_update.cc
    1.59 02/08/01 15:19:04 Sinisa <at> sinisa.nasamreza.org +1 -1
    Just a proper error message

# This is a BitKeeper patch.  What follows are the unified diffs for the
# set of deltas contained in the patch.  The rest of the patch, the part
# that BitKeeper cares about, is below these diffs.
# User:	Sinisa
# Host:	sinisa.nasamreza.org
# Root:	/mnt/work/mysql-4.0

--- 1.58/sql/sql_update.cc	Tue Jul 23 18:31:18 2002

(Continue reading)

Guillaume Morin | 1 Aug 2002 14:52

Re: using my_thread_* in a library

Dans un message du 01 Aug à 15:10, Sinisa Milivojevic écrivait :
> The answer depends very much on the version of MySQL C API that you
> are using.

I code for soname "10", ie. Mysql 3.23.x C API. I really do not know if
there is another kind of versionning.

> It also depens on your application.
> But for multi-threaded applications you have to call mysql_init and
> mysql_thread_init at start and mysql_thread_ned at the end of each of
> your threads.

I think I was kinda unclear, sorry for that. I code the nss module.
Therefore, I have to find a solution for all applications. I chose to
call my_pthread_end at the end of the NSS functions of the module (I of
course call my_thread_init at the beginning of the functions).

But some apps which uses the client library such as the MySQL server itself 
calls the functions you've mentionned, then calls a NSS lookup function
(which will do a my_thread_init and my_pthread_end) and then uses some
other MySQL C functions again which will make the app segfault since
my_pthread_end has been called.

If I remove the calls to my_pthread_end in the NSS module, I'll create a
mem leak for the apps which does not use the client library as described
above.

Regards,

--

--

(Continue reading)

paul | 1 Aug 2002 16:55

bk commit into 4.0 tree

Below is the list of changes that have just been committed into a
4.0 repository of paul. When paul does a push these changes will
be propagated to the main repository and, within 24 hours after the
push, to the public repository.
For information on how to access the public repository
see http://www.mysql.com/doc/I/n/Installing_source_tree.html

ChangeSet <at> 1.1139, 2002-08-01 09:55:26-05:00, paul <at> teton.kitebird.com
  manual.texi	fixups for mysqld_multi section

  Docs/manual.texi
    1.1070 02/08/01 09:54:56 paul <at> teton.kitebird.com +62 -46
    fixups for mysqld_multi section

# This is a BitKeeper patch.  What follows are the unified diffs for the
# set of deltas contained in the patch.  The rest of the patch, the part
# that BitKeeper cares about, is below these diffs.
# User:	paul
# Host:	teton.kitebird.com
# Root:	/home/paul/mysql-4.0

--- 1.1069/Docs/manual.texi	Wed Jul 31 17:41:49 2002
+++ 1.1070/Docs/manual.texi	Thu Aug  1 09:54:56 2002
 <at>  <at>  -20825,7 +20825,7  <at>  <at> 
  <at> menu
 * Server-Side Overview::        Overview of the Server-Side Scripts and Utilities
 * safe_mysqld::                  <at> code{safe_mysqld}, The Wrapper Around  <at> code{mysqld}
-* mysqld_multi::                 <at> code{mysqld_multi}, Program for Managing Multiple MySQL Servers
+* mysqld_multi::                 <at> code{mysqld_multi}, A Program for Managing Multiple MySQL Servers
 * myisampack::                   <at> code{myisampack}, The MySQL Compressed Read-only Table Generator

(Continue reading)

Tim Bunce | 1 Aug 2002 19:11

Picon

Re: bk commit into 4.0 tree

On Thu, Aug 01, 2002 at 09:55:32AM -0500, paul <at> kitebird.com wrote:
> +using  <at> code{safe_mysqld} instead of  <at> code{mysqld} directly here is,
> +that  <at> code{safe_mysqld} ``guards'' every  <at> code{mysqld} process and will

Now that safe_mysqld has been renamed mysqld_safe in v4.0.0, is
there a policy for which the docs should refer to?

Will future v3.23.x releases ship with both to allow migration
(and for the docs to be updated to only refer to the new name)?

Tim.

---------------------------------------------------------------------
Before posting, please check:
   http://www.mysql.com/manual.php   (the manual)
   http://lists.mysql.com/           (the list archive)

To request this thread, e-mail internals-thread4236 <at> lists.mysql.com
To unsubscribe, e-mail <internals-unsubscribe <at> lists.mysql.com>

paul | 1 Aug 2002 23:47

bk commit into 4.0 tree

Below is the list of changes that have just been committed into a
4.0 repository of paul. When paul does a push these changes will
be propagated to the main repository and, within 24 hours after the
push, to the public repository.
For information on how to access the public repository
see http://www.mysql.com/doc/I/n/Installing_source_tree.html

ChangeSet <at> 1.1141, 2002-08-01 16:47:10-05:00, paul <at> teton.kitebird.com
  manual.texi	tweaks

  Docs/manual.texi
    1.1071 02/08/01 16:46:52 paul <at> teton.kitebird.com +15 -15
    tweaks

# This is a BitKeeper patch.  What follows are the unified diffs for the
# set of deltas contained in the patch.  The rest of the patch, the part
# that BitKeeper cares about, is below these diffs.
# User:	paul
# Host:	teton.kitebird.com
# Root:	/home/paul/mysql-4.0

--- 1.1070/Docs/manual.texi	Thu Aug  1 09:54:56 2002
+++ 1.1071/Docs/manual.texi	Thu Aug  1 16:46:52 2002
 <at>  <at>  -13663,7 +13663,7  <at>  <at> 
  <at> subsection Using  <at> code{AUTO_INCREMENT}

  <at> cindex AUTO_INCREMENT
-The  <at> code{AUTO_INCREMENT} attribute can be used to generate an unique
+The  <at> code{AUTO_INCREMENT} attribute can be used to generate a unique
 identity for new rows:

(Continue reading)

paul | 1 Aug 2002 23:58

bk commit into 4.0 tree

Below is the list of changes that have just been committed into a
4.0 repository of paul. When paul does a push these changes will
be propagated to the main repository and, within 24 hours after the
push, to the public repository.
For information on how to access the public repository
see http://www.mysql.com/doc/I/n/Installing_source_tree.html

ChangeSet <at> 1.1142, 2002-08-01 16:57:59-05:00, paul <at> teton.kitebird.com
  manual.texi	tweaks

  Docs/manual.texi
    1.1072 02/08/01 16:57:45 paul <at> teton.kitebird.com +7 -5
    tweaks

# This is a BitKeeper patch.  What follows are the unified diffs for the
# set of deltas contained in the patch.  The rest of the patch, the part
# that BitKeeper cares about, is below these diffs.
# User:	paul
# Host:	teton.kitebird.com
# Root:	/home/paul/mysql-4.0

--- 1.1071/Docs/manual.texi	Thu Aug  1 16:46:52 2002
+++ 1.1072/Docs/manual.texi	Thu Aug  1 16:57:45 2002
 <at>  <at>  -5464,7 +5464,7  <at>  <at> 
  <at> subsubsection Preparing the Windows MySQL Environment

 Starting with MySQL 3.23.38, the Windows distribution includes
-both the normal and the  <at> strong{MySQL-Max} server binaries.
+both the normal and the MySQL-Max server binaries.
 Here is a list of the different MySQL servers you can use:

(Continue reading)

jani | 2 Aug 2002 02:24

Picon

bk commit into 4.0 tree

Below is the list of changes that have just been committed into a
4.0 repository of jani. When jani does a push these changes will
be propagated to the main repository and, within 24 hours after the
push, to the public repository.
For information on how to access the public repository
see http://www.mysql.com/doc/I/n/Installing_source_tree.html

ChangeSet
  1.1143 02/08/02 03:24:02 jani <at> rhols221.adsl.netsonic.fi +1 -0
  Portability fix for QNX 6.2

  include/my_global.h
    1.25 02/08/02 03:23:57 jani <at> rhols221.adsl.netsonic.fi +2 -0
    Passed a bug in QNX 6.2 library that affected compiling.

# This is a BitKeeper patch.  What follows are the unified diffs for the
# set of deltas contained in the patch.  The rest of the patch, the part
# that BitKeeper cares about, is below these diffs.
# User:	jani
# Host:	rhols221.adsl.netsonic.fi
# Root:	/home/my/bk/mysql-4.0

--- 1.24/include/my_global.h	Sun Jun 30 23:04:09 2002
+++ 1.25/include/my_global.h	Fri Aug  2 03:23:57 2002
 <at>  <at>  -127,6 +127,8  <at>  <at> 
 #undef  HAVE_SYS_UN_H
 #undef  HAVE_FINITE
 #undef  HAVE_RINT
+#undef  LONGLONG_MIN            /* These get wrongly defined in QNX 6.2 */
+#undef  LONGLONG_MAX            /* standard system library 'limits.h' */

(Continue reading)

Tim Bunce | 2 Aug 2002 15:25

Picon

Database server tuning: The Oracle Wait Interface

Monty, we spoke at the conference about the mechanisms Oracle
provides to do database performance tuning.

I said I'd send you some information...

http://www.quest-pipelines.com/newsletter-v3/0402_C.htm
http://www.quest.com/whitepapers/TuningOracleWithoutCacheHit.pdf

Details of the views that form the "Oracle Wait Interface":
http://download-west.oracle.com/otndoc/oracle9i/901_doc/server.901/a90190/apa.htm#264187

Tim.

---------------------------------------------------------------------
Before posting, please check:
   http://www.mysql.com/manual.php   (the manual)
   http://lists.mysql.com/           (the list archive)

To request this thread, e-mail internals-thread4240 <at> lists.mysql.com
To unsubscribe, e-mail <internals-unsubscribe <at> lists.mysql.com>

Group

gmane.comp.db.mysql.devel.

Advertisement

Search Archive

Page

1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 37

Articles in period: 360

August 2002

Language

Change language

Options

Current view: Threads only / Showing only 20 lines / Not hiding cited text.
Change to All messages, whole messages, or hide cited text.

Post a message
NNTP Newsgroup
Classic Gmane web interface
XML

XML

RSS Feed
List Information

About Gmane

Recent entries

Patch for mysqldump (5 May 2013)
outputting varchars to a file to be read by load data infile (5 May 2013)
adding an index of my own type/code (23 Apr 2013)
after cross compile, mysql table hidden while operation on it still wo (10 Mar 2013)
inconsistency in source-code/comments in mf_keycache (4 Mar 2013)
effects of load data local on RSS (28 Feb 2013)
properly detecting cases where handler can silently overwrite rows (26 Feb 2013)
[Maria-developers] Fwd: some question on bundled libedit in mysql (25 Feb 2013)
Fwd: some question on bundled libedit in mysql (25 Feb 2013)
Are handlerton->commit calls serialized in MySQL 5.6? (21 Feb 2013)
API version for STORAGE ENGINE plugin is too different (15 Feb 2013)
debugging how query plans are made (12 Feb 2013)
when propagate_cond_constants will be called? (10 Feb 2013)
[PATCH] use fallocate to create table (8 Feb 2013)
source does not compile (5 Feb 2013)
reducing fsyncs during handlerton->prepare and handlerton->commit in 5 (30 Jan 2013)
what pros/cons of storing binary log in an InnoDB table? (28 Jan 2013)
making a system variable that is a float or double (18 Jan 2013)
making a system variable that is a float or double (18 Jan 2013)
detecting select SQL statements that are functions (18 Jan 2013)
detecting select SQL statements that are functions (18 Jan 2013)

Archives

4	May 2013
6	April 2013
12	March 2013
27	February 2013
28	January 2013
39	December 2012
21	November 2012
33	October 2012
10	September 2012
21	August 2012
18	July 2012
23	June 2012
9	May 2012
24	April 2012
21	March 2012
28	February 2012
18	January 2012
13	December 2011
16	November 2011
3	October 2011
5	September 2011
2	August 2011
23	July 2011
18	June 2011
21	May 2011
33	April 2011
23	March 2011
18	February 2011
10	January 2011
13	December 2010
45	November 2010
45	October 2010
69	September 2010
46	August 2010
27	July 2010
54	June 2010
34	May 2010
48	April 2010
89	March 2010
75	February 2010
61	January 2010
48	December 2009
106	November 2009
121	October 2009
53	September 2009
44	August 2009
189	July 2009
223	June 2009
245	May 2009
93	April 2009
177	March 2009
99	February 2009
35	January 2009
69	December 2008
40	November 2008
103	October 2008
64	September 2008
58	August 2008
76	July 2008
137	June 2008
88	May 2008
82	April 2008
73	March 2008
98	February 2008
46	January 2008
59	December 2007
44	November 2007
49	October 2007
107	September 2007
67	August 2007
134	July 2007
105	June 2007
83	May 2007
129	April 2007
126	March 2007
64	February 2007
60	January 2007
60	December 2006
109	November 2006
100	October 2006
78	September 2006
94	August 2006
63	July 2006
104	June 2006
60	May 2006
52	April 2006
37	March 2006
66	February 2006
71	January 2006
439	December 2005
1175	November 2005
1113	October 2005
1450	September 2005
1359	August 2005
1189	July 2005
1142	June 2005
924	May 2005
955	April 2005
1034	March 2005
1204	February 2005
1151	January 2005
956	December 2004
1059	November 2004
845	October 2004
797	September 2004
863	August 2004
576	July 2004
667	June 2004
646	May 2004
466	April 2004
771	March 2004
540	February 2004
431	January 2004
422	December 2003
345	November 2003
512	October 2003
364	September 2003
415	August 2003
450	July 2003
415	June 2003
382	May 2003
441	April 2003
523	March 2003
430	February 2003
509	January 2003
408	December 2002
636	November 2002
361	October 2002
243	September 2002
360	August 2002
221	July 2002
116	June 2002

Design

Zawodny | Right Menu | Left Menu

Your Own Design

Paste the URL of your CSS below. Download CSS template