Hi,
I've attached a diff against the current CVS version. Only
modules/owlib/src/c/ow_net.c needs to be changed. (not sure this is
correct input for patch, please give me a commandline for cvs diff if
you need another format...).
Regards,
Wim.
Paul Alfille wrote:
> I'd love a patch!
>
> On 4/10/06, *Wim Heirman* <[EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]>> wrote:
>
> Hi,
>
> > Are you using a standard i386 with glib? I'm having trouble
> seeing what
> > has changed to make a code change needed.
>
> Yes, it's a Plain Old Dell i686 running a stock Fedora Core 5 with
> kernel 2.6.16 and glibc 2.4. It only started going wrong when I upgraded
> my kernel to 2.6.16, do you have that same kernel version too?
>
> > The "embedded functions" were mainly a convenience. Without them,
> a lot
> > of custom structures, pointer manipulation and extra functions are
> needed.
>
> I've noticed that ;) Actually you only need one extra struct to pass
> out2 and HandlerRoutine to AcceptThread through the thread creation.
> For
> ConnectionThread it would be worse, but it's still embedded here and it
> works, so the problem may be only in the double embedding.
>
> > If it's needed, I'll do it.
>
> I can send a patch, just need to clean up some. Is anyone else running
> 2.6.16 already? Does it happen there too?
>
>
> Regards,
> Wim
>
>
> > On 4/10/06, *Wim Heirman* <[EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]>
> > <mailto: [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>>> wrote:
> >
> > Hi,
> >
> > I updated to the latest CVS, no change. However, I started digging
> > around some more myself, I believe the problem is that
> pthread_create()
> > doesn't like running an embedded function. When I get rid of the
> > embedding it seems to work again.
> > Anyone else has some experience/comments on this, or should I
> just clean
> > this up into a patch you can apply?
> >
> > Regards,
> > Wim.
> >
> >
> > Paul Alfille wrote:
> > > I've been testing (and correcting) OWFS extensively. No
> segfaults, no
> > > problems with threading. Want to try a fresh pull and retest?
> > >
> > > Paul
> > >
> > > On 4/9/06, *Wim Heirman* <[EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]>
> > <mailto: [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>>
> > > <mailto: [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>
> <mailto:[EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>>>> wrote:
> > >
> > > I've grabbed some more gdb output:
> > >
> > > (gdb) run -u --foreground -p localhost:owfs
> > > <...>
> > > Program received signal SIGSEGV, Segmentation fault.
> > > [Switching to Thread 23919520 (LWP 23820)]
> > > 0x001163af in start_thread () from /lib/libpthread.so.0
> > > (gdb) bt
> > > #0 0x001163af in start_thread () from /lib/libpthread.so.0
> > > #1 0x00ae133e in clone () from /lib/libc.so.6
> > > (gdb) info
> > > threads
> > > * 2 Thread 23919520 (LWP 23820) 0x001163af in
> start_thread ()
> > from
> > > /lib/libpthread.so.0
> > > 1 Thread -1208199504 (LWP 23796) 0x009fb402 in
> > __kernel_vsyscall ()
> > > (gdb) thread 1
> > > [Switching to thread 1 (Thread -1208199504 (LWP
> > > 23796))]#0 0x009fb402 in
> > > __kernel_vsyscall ()
> > > (gdb) bt
> > > #0 0x009fb402 in __kernel_vsyscall ()
> > > #1 0x0011b7ae in __lll_mutex_lock_wait () from
> > /lib/libpthread.so.0
> > > #2 0x0011816c in _L_mutex_lock_70 () from
> /lib/libpthread.so.0
> > > #3 0x00117fb8 in pthread_mutex_lock () from
> /lib/libpthread.so.0
> > > #4 0x00cb5e3f in ConnectionThread.6627 () at ow_net.c:314
> > > #5 0x00cb5fab in ServerProcess (HandlerRoutine=0x8049a20
> > <Handler>,
> > > Exit=0x8049020 <ow_exit>) at ow_net.c:353
> > > #6 0x08049192 in main (argc=5, argv=0xbfeddce4) at
> owserver.c:558
> > >
> > >
> > > I also tried using owpython without owserver (so with
> > ow.init('u')),
> > > this
> > > works regardless of whether thread-support is compiled in.
> > >
> > > Regards,
> > > Wim
> > >
> > >
> > > On Fri, 7 Apr 2006, Wim Heirman wrote:
> > >
> > > > Hi,
> > > >
> > > > Apparently my new (not yet installed) owserver was still
> > linking
> > > to the
> > > > libow in /lib, not the freshly compiled one, so now at
> least
> > I got the
> > > > --disable-mt version working. With multithreading enabled
> > the problem
> > > > remains however.
> > > >
> > > > owfs is defined as port 10123 in /etc/services.
> > > >
> > > > Regards,
> > > > Wim
> > > >
> > > > Alfille, Paul H.,M.D. wrote:
> > > >> Wim,
> > > >>
> > > >> I ran into a similar problem a little while ago, fixed by
> > making
> > > sure that libow
> > > >> and libfuse were compiled and updated.
> > > >>
> > > >> By the way, is "owfs" defined as a port?
> > > >>
> > > >> Paul
> > > >>
> > > >>
> > > >> -----Original Message-----
> > > >> From: [EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]>
> > <mailto: [EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]>>
> > > <mailto:[EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]>
> > <mailto: [EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]>>> on behalf of
> > > Wim Heirman
> > > >> Sent: Fri 4/7/2006 5:12 AM
> > > >> To: [email protected]
> <mailto:[email protected]>
> > <mailto:[email protected]
> <mailto:[email protected]>>
> > > <mailto: [email protected]
> <mailto:[email protected]>
> > <mailto:[email protected]
> <mailto:[email protected]>>>
> > > >> Subject: [Owfs-developers] owserver segfaulting on
> startup
> > > >>
> > > >> Hello,
> > > >>
> > > >> Has anyone seen the following problem? When starting
> owserver:
> > > >>
> > > >> root# ./owserver -u --foreground -p localhost:owfs
> > > >> Segmentation fault
> > > >>
> > > >> GDB traceback is included at the bottom. This is on a
> > Fedora Core 5
> > > >> machine with kernel 2.6.16. Yesterday, with kernel
> 2.6.15,
> > > owserver ran
> > > >> normally. Since the segfault is in libpthread I tried
> > recompiling
> > > owfs
> > > >> with --disable-mt, but this did not change the
> result. I'm
> > using
> > > the CVS
> > > >> version of owfs from today (2006-04-07).
> > > >>
> > > >> Regards,
> > > >> Wim.
> > > >>
> > > >>
> > > >> root# gdb ./owserver
> > > >> GNU gdb Red Hat Linux (6.3.0.0-1.122rh)
> > > >> Copyright 2004 Free Software Foundation, Inc.
> > > >> GDB is free software, covered by the GNU General Public
> > License,
> > > and you are
> > > >> welcome to change it and/or distribute copies of it
> under
> > certain
> > > >> conditions.
> > > >> Type "show copying" to see the conditions.
> > > >> There is absolutely no warranty for GDB. Type "show
> warranty"
> > > for details.
> > > >> This GDB was configured as
> "i386-redhat-linux-gnu"...Using host
> > > >> libthread_db library "/lib/libthread_db.so.1".
> > > >>
> > > >> (gdb) run -u --foreground -p localhost:owfs
> > > >> Starting program:
> > > >>
> /usr/src/owfs_cvs/owfs/module/owserver/src/c/.libs/owserver -u
> > > >> --foreground -p localhost:owfs
> > > >> Reading symbols from shared object read from target
> > memory...done.
> > > >> Loaded system supplied DSO at 0x9fb000
> > > >> [Thread debugging using libthread_db enabled]
> > > >> [New Thread -1208432976 (LWP 30428)]
> > > >> [New Thread 23866272 (LWP 30431)]
> > > >>
> > > >> Program received signal SIGSEGV, Segmentation fault.
> > > >> [Switching to Thread 23866272 (LWP 30431)]
> > > >> 0x00cb33af in start_thread () from /lib/libpthread.so.0
> > > >> (gdb) bt
> > > >> #0 0x00cb33af in start_thread () from
> /lib/libpthread.so.0
> > > >> #1 0x00ae133e in clone () from /lib/libc.so.6
> > > >> (gdb)
> > > >>
> > > >>
> > > >>
> > > >>
> > > >>
> > > >
> > > >
> > >
> > > --
> > > ir. Wim Heirman,
> > > ELIS Department, Ghent University, Belgium
> > > Phone: +32-9-264.95.27
> > > E-mail: [EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]>
> > <mailto:[EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]>>
> <mailto:[EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>
> > <mailto:[EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]>>>
> > > http://www.elis.UGent.be/~wheirman/
> > >
> > >
> > > -------------------------------------------------------
> > > This SF.Net email is sponsored by xPML, a groundbreaking
> > scripting
> > > language
> > > that extends applications into web and mobile media. Attend
> > the live
> > > webcast
> > > and join the prime developer group breaking into this
> new coding
> > > territory!
> > >
> >
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642
> <http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642>
> >
> <http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642
>
> <http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642>>
> > > <
> >
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642
> <http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642>
> > <
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642
>
> <http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642>>>
> > > _______________________________________________
> > > Owfs-developers mailing list
> > > [email protected]
> <mailto:[email protected]>
> > <mailto:[email protected]
> <mailto:[email protected]>>
> > > <mailto:[email protected]
> <mailto:[email protected]>
> > <mailto:[email protected]
> <mailto:[email protected]>>>
> > > https://lists.sourceforge.net/lists/listinfo/owfs-developers
> > >
> > >
> >
> > --
> > ir. Wim Heirman,
> > ELIS Department, Ghent University, Belgium
> > Phone: +32-9-264.95.27
> > E-mail: [EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]> <mailto:
> [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>>
> > http://www.elis.UGent.be/~wheirman/
> >
> >
> > -------------------------------------------------------
> > This SF.Net email is sponsored by xPML, a groundbreaking scripting
> > language
> > that extends applications into web and mobile media. Attend
> the live
> > webcast
> > and join the prime developer group breaking into this new coding
> > territory!
> >
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642
> <http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642>
> >
> <http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642
>
> <http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642>>
> > _______________________________________________
> > Owfs-developers mailing list
> > [email protected]
> <mailto:[email protected]>
> > <mailto: [email protected]
> <mailto:[email protected]>>
> > https://lists.sourceforge.net/lists/listinfo/owfs-developers
> >
> >
>
> --
> ir. Wim Heirman,
> ELIS Department, Ghent University, Belgium
> Phone: +32-9-264.95.27
> E-mail: [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>
> http://www.elis.UGent.be/~wheirman/
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by xPML, a groundbreaking scripting
> language
> that extends applications into web and mobile media. Attend the live
> webcast
> and join the prime developer group breaking into this new coding
> territory!
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642
> <http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642>
> _______________________________________________
> Owfs-developers mailing list
> [email protected]
> <mailto:[email protected]>
> https://lists.sourceforge.net/lists/listinfo/owfs-developers
>
>
--
ir. Wim Heirman,
ELIS Department, Ghent University, Belgium
Phone: +32-9-264.95.27
E-mail: [EMAIL PROTECTED]
http://www.elis.UGent.be/~wheirman/
RCS file: /cvsroot/owfs/owfs/module/owlib/src/c/ow_net.c,v
retrieving revision 1.36
diff -r1.36 ow_net.c
253d252
< Doubly embedded function.
266a266,292
>
> void RunAccepted( int rafd, void (*HandlerRoutine)(int fd) ) {
> if ( rafd>=0 ) {
> HandlerRoutine( rafd ) ;
> close( rafd ) ;
> } else {
> STAT_ADD1(NET_accept_errors);
> LEVEL_CONNECT("accept() error %d [%s]\n", errno, strerror(errno));
> }
> }
>
> struct AcceptThread_data {
> struct connection_out *o2;
> void (*HandlerRoutine)(int fd);
> };
> void * AcceptThread( void * v2 ) {
> struct AcceptThread_data * data = (struct AcceptThread_data *)v2;
> int acceptfd = accept( data->o2->fd, NULL, NULL ) ;
> ACCEPTUNLOCK(data->o2);
> RunAccepted( acceptfd, data->HandlerRoutine ) ;
> free(data);
> #ifndef VALGRIND
> pthread_exit((void *)0);
> #endif /* VALGRIND */
> return NULL;
> }
>
284,294d309
< /* embedded function */
< void RunAccepted( int rafd ) {
< if ( rafd>=0 ) {
< HandlerRoutine( rafd ) ;
< close( rafd ) ;
< } else {
< STAT_ADD1(NET_accept_errors);
< LEVEL_CONNECT("accept() error %d [%s]\n", errno, strerror(errno));
< }
< }
<
300,310d314
< /* Doubly Embedded function */
< void * AcceptThread( void * v2 ) {
< struct connection_out *o2 = (struct connection_out *)v2;
< int acceptfd = accept( o2->fd, NULL, NULL ) ;
< ACCEPTUNLOCK(o2);
< RunAccepted( acceptfd ) ;
< #ifndef VALGRIND
< pthread_exit((void *)0);
< #endif /* VALGRIND */
< return NULL;
< }
315c319,324
< if ( pthread_create( &thread2, attr_p, AcceptThread, out2 ) ) Exit(1) ;
---
> /* Start a thread running AcceptThread(), all necessary parameters are passed
> using a struct AcceptThread_data */
> struct AcceptThread_data * data = malloc(sizeof(struct AcceptThread_data));
> data->o2 = out2;
> data->HandlerRoutine = HandlerRoutine;
> if ( pthread_create( &thread2, attr_p, AcceptThread, data ) ) Exit(1) ;
