[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#565801: src:firebird2.5: FTBFS on kfreebsd-amd64

Hi Petr,

I have some movement to report.

-=| Petr Salinger, Tue, Feb 02, 2010 at 07:46:44PM +0100 |=-
>> firebird2.5 behaved the same:
>>    ../gen/firebird/bin/gbak_static -MODE read_only -R
>>    ../builds/misc/help.gbak ../gen/firebird/help/help.fdb
>>    make[3]: *** [../gen/firebird/help/help.fdb] Segmentation fault
>>    (core dumped)
>> ˙˙ and built fine on asdfasdf.debian.net. Is it possible to get the
>> core dump? Maybe looking into it can sched some light.
> Yes and no.
> Line 472 is call to "dbb_functions(*p)"
> #3  0x00000000005248b6 in Database (this=0x843d5fe28, p=0x843d52020) at  
> ../src/jrd/../jrd/../jrd/Database.h:472

Thanks for looking!

The next upstream release candidate now reliably fails on 
kfreebsf/amd64 (and builds fine on /i386). The stacktrace is (obtained 
on asdfasdf.d.n):
(gdb) run -MODE read_only -R ../builds/misc/help.gbak ../gen/firebird/help/help.fdb
Starting program: /home/dmn/firebird2.5- -MODE read_only -R ../builds/misc/help.gbak ../gen/firebird/help/help.fdb

Program received signal ?, Unknown signal.
0x00000008019347d7 in __pthread_sigsuspend () from /lib/libpthread.so.0
(gdb) bt
#0  0x00000008019347d7 in __pthread_sigsuspend () from /lib/libpthread.so.0
#1  0x0000000801933668 in __pthread_wait_for_restart_signal ()
   from /lib/libpthread.so.0
#2  0x00000008019346e0 in pthread_create@@GLIBC_2.3 ()
   from /lib/libpthread.so.0
#3  0x000000000063d1a8 in ThreadStart::start (
    routine=0x462a72 <shutdownThread>, arg=0x0, 
    priority_arg=<value optimized out>, thd_id=0x800a456a0)
    at ../src/jrd/ThreadStart.cpp:171
#4  0x000000000063d29d in gds__thread_start (entrypoint=0x7fffffffa4d0, 
    arg=0x1f, priority=0, thd_id=0x94016030) at ../src/jrd/ThreadStart.cpp:73
#5  0x0000000000454ad6 in CtrlCHandler (this=0x7fffffffa940, primary=0x0)
    at ../src/jrd/why.cpp:1004
#6  GlobalPtr (this=0x7fffffffa940, primary=0x0)
    at ../src/jrd/../common/classes/init.h:127
#7  YEntry (this=0x7fffffffa940, primary=0x0) at ../src/jrd/why.cpp:1036
#8  0x00000000004610d7 in isc_attach_database (
    user_status=<value optimized out>, file_length=0, file_name=0x0, 
    public_handle=0x0, dpb_length=24624, dpb=0x0) at ../src/jrd/why.cpp:1373
#9  0x000000000042f083 in open_files (uSvc=0x800a457a8)
    at ../src/burp/burp.cpp:1905
#10 gbak (uSvc=0x800a457a8) at ../src/burp/burp.cpp:1062
#11 0x00000000004320c2 in main (argc=<value optimized out>, 
---Type <return> to continue, or q <return> to quit---
    argv=<value optimized out>) at ../src/burp/burpMain.cpp:47

I posted this to upstream devel list[1] and they responded:

    [1] http://sourceforge.net/mailarchive/forum.php?thread_name=20100309131707.GF3666%40ktnx.net&forum_name=firebird-devel

> At the first glance this looks like broken kFreebsd/amd64. This is 
> the first thread, started by us (it happens is yValve in 
> CtrlCHandler, in ctor of static instance of this class). Started 
> thread does nothing except wait for posix semaphore. How did Unknown 
> signal arriave here - absolutely no idea.
> PS. On freebsd/amd64 I have successfull build.

How likely does it seem to you that the problem is in threading stuff 
on kfreebsd/amd64? Maybe they are right, or maybe my patches[2] adding 
support to kfreebsd are not quite right.

    [2] http://git.debian.org/?p=pkg-firebird/2.5.git;a=tree;f=debian/patches;hb=HEAD
        fix-kfreebsd_hurd-detection, freebsd_targets, 
        install_freebsd_as_linux, maybe also optional_posix_fadv

Thank you for your (and anyone else's) help on this! The bug is quite 
important for the release as the reverse dependencies include such 
things as php5 and qt4-x11 :|

Attachment: signature.asc
Description: Digital signature

Reply to: