[MDEV-7398] mysqld segfaults on FreeBSD 10.1 i386 when built with clang 3.4 Created: 2014-12-31  Updated: 2015-10-29  Resolved: 2015-06-17

Status: Closed
Project: MariaDB Server
Component/s: Compiling
Affects Version/s: 10.0.15
Fix Version/s: 10.0.20

Type: Bug Priority: Major
Reporter: Bernard Spil Assignee: Sergei Golubchik
Resolution: Fixed Votes: 1
Labels: verified
Environment:

FreeBSD i386bsd 10.1-RELEASE FreeBSD 10.1-RELEASE #0 r274401: Tue Nov 11 22:51:51 UTC 2014 root@releng1.nyi.freebsd.org:/usr/obj/usr/src/sys/GENERIC i386

Issue Links:
Duplicate
duplicates MDEV-8054 core dumps in test: main Closed
Sprint: 10.0.20

 Description   

From: spil.oss@gmail.com
To: maria-developers@lists.launchpad.net
subject: mysqld 10.0.15 segfaults on FreeBSD i386 clang 3.4

>Description:
When MariaDB is built with clang 3.4 on FreeBSD i386 (which is
the default compiler) the server will segfault the moment a
client connects to it. Same behaviour is observed with clang
3.5. Built with clang 3.3 on 10.1 runs without segfaults.
Built on FreeBSD 10.0 (which comes with clang 3.3) runs OK.
>How-To-Repeat:
Use FreeBSD 10.1 i386
Use port to build MariaDB 10.0 or 5.5
Connect to server using client

>Fix:
Build with clang 3.3 or build using Pourdiere (uses 10.0 jail)

>Submitter-Id: <submitter ID>
>Originator: Bernard Spil
>Organization:
FreeBSD MariaDB 10.0 port committer
>MySQL support: none
>Synopsis: MariaDB segfaults on i386 FreeBSD
>Severity: non-critical
>Priority: low
>Category: mysql
>Class: sw-bug
>Release: mysql-10.0.15 (FreeBSD Ports)

>C compiler: clang 3.4
>C++ compiler: clang 3.4
>Environment:
FreeBSD 10.1 GENERIC i386 Celeron U4100 Dual Core 4GB
System: FreeBSD i386bsd 10.1-RELEASE FreeBSD 10.1-RELEASE #0 r274401: Tue Nov 11 22:51:51 UTC 2014 root@releng1.nyi.freebsd.org:/usr/obj/usr/src/sys/GENERIC i386

Some paths: /usr/bin/perl /usr/bin/make /usr/local/bin/gmake /usr/bin/cc

Compilation info (call): CC='/usr/bin/cc' CFLAGS='-O2 -pipe -fstack-protector -fno-strict-aliasing -O2 -g -DNDEBUG -DDBUG_OFF' CXX='/usr/bin/c++' CXXFLAGS='-O2 -pipe -fstack-protector -fno-strict-aliasing -O2 -g -DNDEBUG -DDBUG_O
FF' LDFLAGS='' ASFLAGS=''
Compilation info (used): CC='/usr/bin/cc' CFLAGS='-O2 -pipe -fstack-protector -fno-strict-aliasing -O2 -g -DNDEBUG -DDBUG_OFF' CXX='/usr/bin/c++' CXXFLAGS='-O2 -pipe -fstack-protector -fno-strict-aliasing -O2 -g -DNDEBUG -DDBUG_O
FF' LDFLAGS='' ASFLAGS=''
LIBC:
rrr- 1 root wheel 1427444 Nov 11 23:52 /lib/libc.so.7
rrr- 1 root wheel 2833712 Nov 11 23:52 /usr/lib/libc.a
rrr- 1 root wheel 166 Nov 11 23:52 /usr/lib/libc.so

Perl: This is perl 5, version 18, subversion 4 (v5.18.4) built for i386-freebsd-thread-multi-64int

 Comments   
Comment by Bernard Spil [ 2014-12-31 ]

Created bug 196399 on FreeBSD bugzilla
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=196399

Comment by Bernard Spil [ 2014-12-31 ]

Uploaded mariadb-bug-clang34segfaultMDEV-7398-2.tgz to ftp.askmonty.org/private
$ tar cvzf mariadb-bug-clang34segfaultMDEV-7398.tgz mysqld.core README i386bsd.err /usr/local/libexec/mysqld gdb.txt
a mysqld.core
a README
a i386bsd.err
tar: Removing leading '/' from member names
a usr/local/libexec/mysqld
a gdb.txt

Comment by Elena Stepanova [ 2015-02-12 ]

I'm getting a crash also, with just as useless coredump; but apparently, we should wait for the FreeBSD fix?

Comment by Bernard Spil [ 2015-02-16 ]

This has now been confirmed to occur with clang 3.6.0rc3 as well
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=196399#c5

Can someone try this on a different (i.e. non-FreeBSD) platform as well?

Comment by Daniel Black [ 2015-04-25 ]

spil https://travis-ci.org/openquery/mariadb-server/builds/59973991 10.0 git head, with patches on gcc-4.9 and clang-3.4 on ubuntu 12.04

Having a cause/test case would obviously help validate this more.

Update: url change - ABI checks under clang were faulty

Comment by Daniel Black [ 2015-04-25 ]

Seemed to be getting enough random failures. While the link the previous comment, with patches applied, survived running the test suite until the 50 minute timeout killed it.

On the 10.0 trunk failures occurred quite quickly in the test suite https://travis-ci.org/openquery/mariadb-server/jobs/59991034

From the failed test run mysqld in debug mode and the var directory including 10 failures has been in the private uploads as travis-ci.org-59991034#117.2-3a193ec2491b...e47656b2abf2-openquery-mariadb-server.tar.bz2

Comment by Daniel Black [ 2015-04-25 ]

https://github.com/MariaDB/server/pull/51 is the same patches from https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=196399

Have these made it into ports? Any feed back on them?

Comment by Daniel Black [ 2015-04-26 ]

Split test cases out so that they might finish: with patches: https://travis-ci.org/openquery/mariadb-server/builds/60093722 regarding 141.1 - I saw sum_distinct-big fail in something else may not be related.

without patches: https://travis-ci.org/openquery/mariadb-server/builds/60083380 lots of clang failures.

Comment by Sergei Golubchik [ 2015-06-06 ]

spil, could you please see if the patch is for fine for you? I've tested it manually, but we don't have working freebsd in buildbot at the moment. It is here. Note that it's 5.5 patch, C11 atomic operations are only used in 10.0 and up.

Comment by Sergei Golubchik [ 2015-06-06 ]

spil, and this is the 10.0 patch for atomics.

Comment by Bernard Spil [ 2015-06-06 ]

Thanks! Testing... Now where's my i386 VM...

Comment by Bernard Spil [ 2015-06-06 ]

Seems to have improved a bit but still segfaults.
Build 10.0.19 on FreeBSD-10.1-i386 OK
Start OK
Connect client OK
use mysql -> segfault and daemon restarts
(before the daemon would not restart)

Comment by Sergei Golubchik [ 2015-06-08 ]

Trying 9.2 with clang 3.4. mariadb-5.5 doesn't crash for me even without any patches.
mariadb-10.0 does crash on startup if innodb or xtradb are enabled. with the patches it seems to work, tests pass.
I'll remove 5.5 from "affected versions".

Comment by Bernard Spil [ 2015-06-09 ]

Does the system produce build logs? I'd like to look at the build script you're using as well.

Furthermore, you should update the build host, FreeBSD 9.2 has been End-of-Life for quite a long time
https://www.freebsd.org/security/security.html#sup

Generated at Thu Feb 08 07:19:17 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.