| Bug #47923 | New "mf_keycache.c" requires thread support | ||
|---|---|---|---|
| Submitted: | 8 Oct 2009 18:42 | Modified: | 18 Dec 2009 11:43 |
| Reporter: | Joerg Bruehe | Email Updates: | |
| Status: | Closed | Impact on me: | |
| Category: | MySQL Server: Compiling | Severity: | S3 (Non-critical) |
| Version: | 5.0.87 | OS: | Any |
| Assigned to: | Joerg Bruehe | CPU Architecture: | Any |
[8 Oct 2009 19:59]
Bugs System
A patch for this bug has been committed. After review, it may be pushed to the relevant source trees for release in the next version. You can access the patch from: http://lists.mysql.com/commits/86246 2817 Joerg Bruehe 2009-10-08 Fix bug#47923 New "mf_keycache.c" requires thread support The bug is a compilation issue: Function "find_key_block()" had thread operations which were not guarded by "#if THREAD", add that now. @ mysys/mf_keycache.c Without thread-support, we are in a single-threaded world, so there is no parallelism, and the keycache cannot be within a resize operation while this request is being processed.
[8 Oct 2009 20:09]
Joerg Bruehe
Running local tests with various "configure" combinations.
[9 Oct 2009 20:15]
Bugs System
A patch for this bug has been committed. After review, it may be pushed to the relevant source trees for release in the next version. You can access the patch from: http://lists.mysql.com/commits/86460 3639 Joerg Bruehe 2009-10-09 [merge] Upmerge the fixes for bug#47923 and bug#47957 from 5.1.
[15 Oct 2009 17:24]
Bugs System
A patch for this bug has been committed. After review, it may be pushed to the relevant source trees for release in the next version. You can access the patch from: http://lists.mysql.com/commits/87019 3639 Joerg Bruehe 2009-10-15 [merge] Upmerge the fixes for bug#47923 and bug#47957 from 5.1. This includes further changes which are needed in 6.0, see the comments for "storage/archive/Makefile.am" and "unittest/backup/Makefile.am". @ configure.in Upmerge from 5.1 Specific in 6.0, we need a conditional "SERVER" which is used in "unittest/backup/" to not build some server-related tests if the build is "--without-server". @ mysys/mf_keycache.c Upmerge from 5.1 @ storage/archive/Makefile.am See comment in "concurrency_test.c": That test cannot be used unless there is thread support in the client library. The code treated this as a runtime issue, but that is wrong: It is even a linkage issue, because the test program needs both "libmysqlclient" (from "concurrency_test.c") and "libmysyslt" (from "azio.c"), and these two will conflict unless it is the threaded client library. The solution is to make the build of this test program depend on "THREAD_SAFE_CLIENT". @ unittest/backup/Makefile.am The necessary "libbackupstream" will not be built unless the build includes the server, so building the test program must be restricted to that case.
[16 Oct 2009 10:21]
Joerg Bruehe
Patch is pushed to the bugteam trees, currently set at versions 5.0.88 and 5.1.41.
[16 Oct 2009 12:37]
Bugs System
Pushed into 5.0.88 (revid:joerg@mysql.com-20091016122941-rf6z0keqvmlgjfto) (version source revid:joerg@mysql.com-20091016122941-rf6z0keqvmlgjfto) (merge vers: 5.0.88) (pib:13)
[16 Oct 2009 17:46]
Paul DuBois
Noted in 5.0.88 changelog. mysys/mf_keycache.c requires threading, but no test was made for thread support. Setting report to NDI pending push to 5.1.x+.
[22 Oct 2009 6:33]
Bugs System
Pushed into 6.0.14-alpha (revid:alik@sun.com-20091022063126-l0qzirh9xyhp0bpc) (version source revid:alik@sun.com-20091019135554-s1pvptt6i750lfhv) (merge vers: 6.0.14-alpha) (pib:13)
[22 Oct 2009 7:05]
Bugs System
Pushed into 5.5.0-beta (revid:alik@sun.com-20091022060553-znkmxm0g0gm6ckvw) (version source revid:alik@sun.com-20091019132831-5j6xs19ld16m7r8j) (merge vers: 5.5.0-beta) (pib:13)
[22 Oct 2009 19:48]
Paul DuBois
Noted in 5.1.41, 5.5.0, 6.0.14 changelogs.
[18 Dec 2009 10:27]
Bugs System
Pushed into 5.1.41-ndb-7.1.0 (revid:jonas@mysql.com-20091218102229-64tk47xonu3dv6r6) (version source revid:jonas@mysql.com-20091218095730-26gwjidfsdw45dto) (merge vers: 5.1.41-ndb-7.1.0) (pib:15)
[18 Dec 2009 10:43]
Bugs System
Pushed into 5.1.41-ndb-6.2.19 (revid:jonas@mysql.com-20091218100224-vtzr0fahhsuhjsmt) (version source revid:jonas@mysql.com-20091217101452-qwzyaig50w74xmye) (merge vers: 5.1.41-ndb-6.2.19) (pib:15)
[18 Dec 2009 10:45]
MC Brown
Added entry to 5.1.40-ndb-7.1.0
[18 Dec 2009 10:58]
Bugs System
Pushed into 5.1.41-ndb-6.3.31 (revid:jonas@mysql.com-20091218100616-75d9tek96o6ob6k0) (version source revid:jonas@mysql.com-20091217154335-290no45qdins5bwo) (merge vers: 5.1.41-ndb-6.3.31) (pib:15)
[18 Dec 2009 11:11]
MC Brown
Added entry to 5.1.41-ndb-6.2.19 and 5.1.42-ndb-6.3.31
[18 Dec 2009 11:12]
Bugs System
Pushed into 5.1.41-ndb-7.0.11 (revid:jonas@mysql.com-20091218101303-ga32mrnr15jsa606) (version source revid:jonas@mysql.com-20091218064304-ezreonykd9f4kelk) (merge vers: 5.1.41-ndb-7.0.11) (pib:15)
[18 Dec 2009 11:43]
MC Brown
Already in 5.1.41
Description: The current code of "mf_keycache.c" will not compile unless there is thread support, the error is: gcc -DDEFAULT_BASEDIR=\"/usr/local\" -DMYSQL_DATADIR="\"/usr/local/var\"" -DDEFAULT_CHARSET_HOME="\"/usr/local\"" -DSHAREDIR="\"/usr/local/share/mysql\"" -DDEFAULT_HOME_ENV=MYSQL_HOME -DDEFAULT_GROUP_SUFFIX_ENV=MYSQL_GROUP_SUFFIX -DDEFAULT_SYSCONFDIR="\"/usr/local/etc\"" -DHAVE_CONFIG_H -I. -I../include -I../include -I../include -I. -O3 -DDBUG_OFF -MT mf_keycache.o -MD -MP -MF .deps/mf_keycache.Tpo -c -o mf_keycache.o mf_keycache.c mf_keycache.c: In function ‘find_key_block’: mf_keycache.c:1782: error: ‘my_thread_var’ undeclared (first use in this function) mf_keycache.c:1782: error: (Each undeclared identifier is reported only once mf_keycache.c:1782: error: for each function it appears in.) mf_keycache.c:1783: error: dereferencing pointer to incomplete type mf_keycache.c:1789: error: dereferencing pointer to incomplete type mf_keycache.c:1791: error: dereferencing pointer to incomplete type mf_keycache.c:1792: error: dereferencing pointer to incomplete type make[2]: *** [mf_keycache.o] Error 1 make[2]: Leaving directory `/MySQL/REPO/V50/platform/mysql-5.0.87/mysys' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/MySQL/REPO/V50/platform/mysql-5.0.87' make: *** [all] Error 2 This is no issue in a server build (which has thread support, so it doesn't happen), but it happens (in 5.0.87) when you specify "--without-server". While I see no use for this module in "libmysys" unless we build a server, it still is contained. How to repeat: Configure "--without-server", try to build. Suggested fix: Enclose the whole section dealing with a key cache resize in "#ifdef THREAD": If there is no thread support (single-threaded), a resize operation cannot be in progress when another request is processed in the key cache. In 6.0, this would be the patch; 5.0 is identical (except for a slight line number shift): --- /V60/push-6.0/mysys/mf_keycache.c 2009-09-28 15:47:56.000000000 +0200 +++ mysys/mf_keycache.c 2009-10-05 12:13:50.000000000 +0200 @@ -1757,6 +1757,7 @@ - block assigned but not yet read from file (invalid data). */ +#if THREAD if (keycache->in_resize) { /* This is a request during a resize operation */ @@ -1998,6 +1999,9 @@ } DBUG_RETURN(0); } +#else /* THREAD */ +DBUG_ASSERT(!keycache->in_resize); +#endif if (page_status == PAGE_READ && (block->status & (BLOCK_IN_EVICTION | BLOCK_IN_SWITCH | In the long run, a restructuring of the server file grouping that puts this module into a server-only library would be even better.