- developers - lists.mariadb.org

[Maria-developers] Patch for xtradb compilation on visual studio
by Kristian Nielsen 15 Dec '11

15 Dec '11

Here is a patch for XtraDB (from Percona Server 5.5.17-rel22.1). We needed this in MariaDB 5.5 to compile on Windows. The problem is that this compiler apparently does not like having #ifdef inside macro invocations. I thought Percona might want to include it upstream, in case there are other compilers that also do not like this. - Kristian. === modified file 'storage/innobase/handler/ha_innodb.cc' --- storage/innobase/handler/ha_innodb.cc 2011-12-15 09:34:14 +0000 +++ storage/innobase/handler/ha_innodb.cc 2011-12-15 12:35:48 +0000 @@ -12369,14 +12369,16 @@ static MYSQL_SYSVAR_ULONG(concurrency_ti "Number of times a thread is allowed to enter InnoDB within the same SQL query after it has once got the ticket", NULL, NULL, 500L, 1L, ~0L, 0); -static MYSQL_SYSVAR_LONG(kill_idle_transaction, srv_kill_idle_transaction, - PLUGIN_VAR_RQCMDARG, #ifdef EXTENDED_FOR_KILLIDLE - "If non-zero value, the idle session with transaction which is idle over the value in seconds is killed by InnoDB.", +#define TMP_STR "If non-zero value, the idle session with transaction which is idle over the value in seconds is killed by InnoDB." #else - "No effect for this build.", +#define TMP_STR "No effect for this build." #endif +static MYSQL_SYSVAR_LONG(kill_idle_transaction, srv_kill_idle_transaction, + PLUGIN_VAR_RQCMDARG, + TMP_STR, NULL, NULL, 0, 0, LONG_MAX, 0); +#undef TMP_STR static MYSQL_SYSVAR_LONG(file_io_threads, innobase_file_io_threads, PLUGIN_VAR_RQCMDARG | PLUGIN_VAR_READONLY | PLUGIN_VAR_NOSYSVAR,

2 1

Re: [Maria-developers] Please fix: innodb_plugin.innodb_gis fails on 'gcov' builder
by Sergei Petrunia 14 Dec '11

14 Dec '11

Hi Alexey, We have a repeatable failure like this: https://internal.askmonty.org/buildbot/builders/gcov/builds/830/steps/test_… innodb_plugin.innodb_gis w3 [ fail ] Test ended at 2011-12-14 06:52:31 CURRENT_TEST: innodb_plugin.innodb_gis mysqltest: At line 1: query 'SET storage_engine=innodb' failed: 1286: Unknown table engine 'innodb' The result from queries just before the failure was: SET storage_engine=innodb; - saving '/home/buildbot/buildbot_basedir/perpush-gcov/build/mysql-test/var/3/log/innodb_plugin.innodb_gis/' to '/home/buildbot/buildbot_basedir/perpush-gcov/build/mysql-test/var/log/innodb_plugin.innodb_gis/' It's in GIS, so I assume it is on your turf. Could you please fix this? BR Sergei -- Sergei Petrunia, Software Developer Monty Program AB, http://askmonty.org Blog: http://s.petrunia.net/blog

3 2

Re: [Maria-developers] Please do a test of 5.3 with these new @@optimizer_switch settings
by Sergei Petrunia 14 Dec '11

14 Dec '11

Hi Elena, On Thu, Dec 08, 2011 at 06:29:01AM +0400, Elena Stepanova wrote: > I tried to set them run overnight, but it isn't going too well. Most of > tests are crashing, see https://bugs.launchpad.net/maria/+bug/901506 . > This bug is finally fixed and pushed. Could you please run more tests on the fix asap? BR Sergei -- Sergei Petrunia, Software Developer Monty Program AB, http://askmonty.org Blog: http://s.petrunia.net/blog

1 0

Re: [Maria-developers] Compilation problem on an old 32-bit machine
by Kristian Nielsen 12 Dec '11

12 Dec '11

Hi Paul, It seems a patch is needed for PBXT to build on 32-bit x86 in some cases: Elena Stepanova <elenst(a)montyprogram.com> writes: > lock_xt.h: Assembler messages: > lock_xt.h:70: Error: bad register name `%sil' > After some re-installation and long digging, I found the ancient bug > report http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10153 which gets > re-filed regularly, the last time was a few months ago. They say it's > a code problem, and suggest replacing the '=r' constraint with '=q'. I > did so in all 4 lines that gcc complained about, and finally got it > work. It seems correct to me that the "q" constraint must be used rather than "r" for the 8-bit case. Below patch should do it (we will apply this in MariaDB). - Kristian. === modified file 'storage/pbxt/src/lock_xt.h' --- storage/pbxt/src/lock_xt.h 2010-05-06 12:57:15 +0000 +++ storage/pbxt/src/lock_xt.h 2011-12-12 11:34:54 +0000 @@ -67,9 +67,9 @@ inline void xt_atomic_inc1(volatile xtWo #elif defined(XT_ATOMIC_GNUC_X86) xtWord1 val; - asm volatile ("movb %1,%0" : "=r" (val) : "m" (*mptr) : "memory"); + asm volatile ("movb %1,%0" : "=q" (val) : "m" (*mptr) : "memory"); val++; - asm volatile ("xchgb %1,%0" : "=r" (val) : "m" (*mptr), "0" (val) : "memory"); + asm volatile ("xchgb %1,%0" : "=q" (val) : "m" (*mptr), "0" (val) : "memory"); #elif defined(XT_ATOMIC_SOLARIS_LIB) atomic_inc_8(mptr); #else @@ -91,9 +91,9 @@ inline xtWord1 xt_atomic_dec1(volatile x #elif defined(XT_ATOMIC_GNUC_X86) xtWord1 val2; - asm volatile ("movb %1, %0" : "=r" (val) : "m" (*mptr) : "memory"); + asm volatile ("movb %1, %0" : "=q" (val) : "m" (*mptr) : "memory"); val--; - asm volatile ("xchgb %1,%0" : "=r" (val2) : "m" (*mptr), "0" (val) : "memory"); + asm volatile ("xchgb %1,%0" : "=q" (val2) : "m" (*mptr), "0" (val) : "memory"); /* Should work, but compiler makes a mistake? * asm volatile ("xchgb %1, %0" : : "r" (val), "m" (*mptr) : "memory"); */

2 1

Re: [Maria-developers] Please fix: gis-precise crashes on solaris
by Alexey Botchkov 12 Dec '11

12 Dec '11

Hi, Adam. Yes i need the access to the machine. Wasn't able to figure out what's wrong by looking at the traces. The crashing query is probably the most trivial of them all. Regards HF 10.12.2011 15:29, Adam M Dutko wrote: > > Let me know if you need access to the machine and I will coordinate > off-list. > > -Adam > >

1 0

[Maria-developers] WL#246 New (by Wlad): Threadpool
by worklog-noreply＠askmonty.org 10 Dec '11

10 Dec '11

----------------------------------------------------------------------- WORKLOG TASK -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- TASK...........: Threadpool CREATION DATE..: Fri, 09 Dec 2011, 22:23 SUPERVISOR.....: IMPLEMENTOR....: Wlad COPIES TO......: CATEGORY.......: Server-BackLog TASK ID........: 246 (http://askmonty.org/worklog/?tid=246) VERSION........: Maria-2.0 STATUS.........: In-Progress PRIORITY.......: 60 WORKED HOURS...: 0 ESTIMATE.......: 300 (hours remain) ORIG. ESTIMATE.: 300 PROGRESS NOTES: DESCRIPTION: Implement a dynamic thread pool, which provides good (superior compared to thread-per-connection) performance for short queries and many threads and does not block on on long queries. ESTIMATED WORK TIME ESTIMATED COMPLETION DATE ----------------------------------------------------------------------- WorkLog (v4.0.0)

1 0

[Maria-developers] Please fix: gis-precise crashes on solaris
by Sergei Petrunia 09 Dec '11

09 Dec '11

Hi Alexey, I'm looking at 5.3 tree and see main.gis-precise to repeatably crash on solaris: http://buildbot.askmonty.org/buildbot/builders/sol-sparc-32/builds/1518/ste… We don't have a solaris machine but the test failure has a stack trace. Could you please look into this? BR Sergei -- Sergei Petrunia, Software Developer Monty Program AB, http://askmonty.org Blog: http://s.petrunia.net/blog

1 0

[Maria-developers] Please do a test of 5.3 with these new @@optimizer_switch settings
by Sergei Petrunia 08 Dec '11

08 Dec '11

Hi Elena, We're going to make the following enabled by default: @@optimizer_switch='semijoin_with_cache=on,outer_join_with_cache=on' @@join_cache_level=2 Could you please make a test run with these settings? It is important to test revision>=3337 (just pushed it), because it has a change in the optimizer which makes it choose semijoin+join_cache execution plans much more often. BR Sergei -- Sergei Petrunia, Software Developer Monty Program AB, http://askmonty.org Blog: http://s.petrunia.net/blog

1 0

[Maria-developers] WL#245 New (by Sanja): EXISTS to IN transformation
by worklog-noreply＠askmonty.org 07 Dec '11

07 Dec '11

----------------------------------------------------------------------- WORKLOG TASK -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- TASK...........: EXISTS to IN transformation CREATION DATE..: Wed, 07 Dec 2011, 12:51 SUPERVISOR.....: IMPLEMENTOR....: COPIES TO......: CATEGORY.......: Client-BackLog TASK ID........: 245 (http://askmonty.org/worklog/?tid=245) VERSION........: Benchmarks-3.0 STATUS.........: Un-Assigned PRIORITY.......: 60 WORKED HOURS...: 0 ESTIMATE.......: 0 (hours remain) ORIG. ESTIMATE.: 0 PROGRESS NOTES: DESCRIPTION: Transform queries like: ... WHERE EXISTS (SELECT 1 FROM inner_table WHERE inner_table.field = 2*outer_table.field AND maybe_something_else)... and ... WHERE NOT EXISTS (SELECT 1 FROM inner_table WHERE inner_table.field = 2*outer_table.field AND maybe_something_else)... into ... WHERE 2*outer_table.field IN (SELECT inner_table.field FROM inner_table WHERE 1 = 1 AND maybe_something_else).. ... WHERE 2*outer_table.field IS NULL OR 2*outer_table.field IN (SELECT inner_table.field FROM inner_table WHERE inner_table.field IS NOT NULL AND maybe_something_else).. To allow optimizations made for IN/ALL/ANY subqueries. Conversion is possible only if: 1)real NULL is not important (top element of WHERE/ON AND/OR list, i.e. NULL equal to FALSE) 2)the subquery has the only dependence which we bring out of it 3)the subquery is simple (has no aggregate function, GROUp BY, ORDER BY, LIMIT HAVING and so on) For NOT EXISTS conversion the subquery should be marked that its left part can't be NULL. ESTIMATED WORK TIME ESTIMATED COMPLETION DATE ----------------------------------------------------------------------- WorkLog (v4.0.0)

1 0

[Maria-developers] No coverage for cases where materialization failed to create temptable?
by Sergei Petrunia 06 Dec '11

06 Dec '11

Hi Timour, I was working on BUG#868908, which involves having different kind of handling for cases where Materialization is not applicable because temp table had the wrong properties. So, I've set a goal to catch such cases: === modified file 'sql/item_subselect.cc' --- sql/item_subselect.cc 2011-12-06 21:03:00 +0000 +++ sql/item_subselect.cc 2011-12-06 21:11:16 +0000 @@ -4182,6 +4182,8 @@ bool subselect_hash_sj_engine::init(List */ if (tmp_table->s->keys == 0) { + //fprintf(stderr, "Q: %s\n", current_thd->query()); + DBUG_ASSERT(0); DBUG_ASSERT( tmp_table->s->uniques || tmp_table->key_info->key_length >= tmp_table->file->max_key_length() || Surprisingly, running mysql-test-run t/subselect*.test with this didn't produce any failures. Is this poor test coverage? Or, this branch is now deadcode? BR Sergei -- Sergei Petrunia, Software Developer Monty Program AB, http://askmonty.org Blog: http://s.petrunia.net/blog

1 0