Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 09 Jun 2004 20:24:05 +0900
From:      Rob <stopspam@users.sourceforge.net>
To:        freebsd-stable@freebsd.org
Subject:   Re: Stable-CVS commit on June 8: hostname lookup failures !?!
Message-ID:  <40C6F355.8050209@users.sourceforge.net>
In-Reply-To: <20040609150946.B776@woozle.rinet.ru>
References:  <40C6E665.1020104@users.sourceforge.net> <20040609150946.B776@woozle.rinet.ru>

next in thread | previous in thread | raw e-mail | index | archive | help
Dmitry Morozovsky wrote:
> On Wed, 9 Jun 2004, Rob wrote:
> 
> R> I upgraded my kernel & world from cvs2.freebsd.org (time stamp about
> R> June 9th 8:00 UTC). After that, the browsers mozilla and opera hang
> R> when looking up hostnames (but konqueror worked fine, strange, isn't it?).
> R> 
> R> I booted up with kernel.old, to no avail. So I concluded the problem
> R> is in the the world upgrade.
> R> 
> R> With up- and downgrading, I narrowed down the offending commit to CVS:
> R> 
> R> Establishing multiplexed-mode data connection
> R> Running
> R> Updating collection src-all/cvs
> R>  Edit src/bin/ps/ps.c
> R>   Add delta 1.30.2.15 2004.06.08.00.11.26 gad
> R>  Edit src/lib/libc/stdio/fclose.c
> R>   Add delta 1.8.2.1 2004.06.08.13.15.01 hmp
> R>  Edit src/sys/net/bpf.h
> R>   Add delta 1.21.2.5 2004.06.08.21.16.42 dwmalone
> R> Shutting down connection to server
> R> Finished successfully
> R> 
> R> Does anyone have an idea what is the connection between my mozilla/opera's
> R> behaviour and these commits? Is this a bug in Stable?
> 
> It seems reverting revision 1.8.2.1 of src/lib/libc/stdio/fclose.c should 
> band-aid the problem:
> 
> marck@woozle:/usr/src/lib/libc/stdio> cvs diff fclose.c
> Index: fclose.c
> ===================================================================
> RCS file: /home/ncvs/src/lib/libc/stdio/fclose.c,v
> retrieving revision 1.8.2.1
> diff -u -r1.8.2.1 fclose.c
> --- fclose.c    8 Jun 2004 13:15:01 -0000       1.8.2.1
> +++ fclose.c    9 Jun 2004 11:13:00 -0000
> @@ -68,9 +68,9 @@
>                 FREEUB(fp);
>         if (HASLB(fp))
>                 FREELB(fp);
> +       FUNLOCKFILE(fp);
>         fp->_file = -1;
>         fp->_r = fp->_w = 0;    /* Mess up if reaccessed. */
>         fp->_flags = 0;         /* Release this FILE for reuse. */
> -       FUNLOCKFILE(fp);
>         return (r);
>  }

Yes, indeed this is the offending commit.
I hope someone will revert this in CVS.

Regards,
Rob.




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?40C6F355.8050209>