Debian Bug report logs - #484265
vde_plug does not connect to socket

Toggle useless messages

Message #3 received at submit@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Debian Bug Tracking System <submit@bugs.debian.org>

Subject: vde_plug does not connect to socket

Date: Tue, 3 Jun 2008 12:19:18 +0200

[Message part 1 (text/plain, inline)]

Package: vde2
Version: 2.2.0-pre2-1
Severity: normal
File: /usr/bin/vde_plug

There's some weird stuff happening in vde_plug: I have a vde_switch
started with ctlsock directory in ./vde0, and now I am trying to
plug into it:

+lapse:~|master|.tmp/testnet.k13727% ls vde0                              #1839
ctl  vde_switch.pid
+lapse:~|master|.tmp/testnet.k13727% strace -f vde_plug vde0/ctl          #1840
[...]
connect(5, {sa_family=AF_FILE, path="vde0/ctl/ctl"}, 110) = -1 ENOTDIR (Not a directory)
[...]
send(5, "\316\372\355\376\3\0\0\0\0\0\0\0\1\0vde0/ctl.14357-000"..., 183, 0) = -1 ENOTCONN (Transport endpoint is not connected)

Note how the manpage says control socket path, and yet the connect()
seems to take it as directory. And even though the
vde0/ctl.14357-000 control socket gets created (it now exists and is
never removed), the send() fails.

Full vde_plug.strace attached.

Let's see what happens if we pass the directory instead of the
socket path.

+lapse:~|master|.tmp/testnet.k13727% strace -f vde_plug vde0              #1841
[...]
connect(5, {sa_family=AF_FILE, path="vde0/ctl"}, 110) = 0
bind(6, {sa_family=AF_FILE, path="vde0.14393-00000"}, 110) = 0
[...]
chmod("", 0700)                         = -1 ENOENT (No such file or directory)

There's stuff broken here too, the chmod doesn't get a proper path.
Also, if I pass ./vde0/, the connect() uses ./vde0//ctl and the bind
then tries ./vde0/.14393-00000.

Full vde_plug.strace2 attached.

This all makes me think that there's a need for some time with the
code...

-- System Information:
Debian Release: lenny/sid
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable'), (1, 'experimental')
Architecture: i386 (i686)

Kernel: Linux 2.6.24-1+scoflowctrl.1-686 (SMP w/1 CPU core)
Locale: LANG=en_GB, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash

Versions of packages vde2 depends on:
ii  adduser                     3.107        add and remove users and groups
ii  libc6                       2.7-11       GNU C Library: Shared libraries
ii  libpcap0.8                  0.9.8-3      system interface for user-level pa
ii  libvdemgmt0                 2.2.0-pre2-1 Virtual Distributed Ethernet - Man
ii  libvdeplug2                 2.2.0-pre2-1 Virtual Distributed Ethernet - Plu

Versions of packages vde2 recommends:
ii  daemon                        0.6.3-1    turns other processes into daemons

-- no debconf information


-- 
 .''`.   martin f. krafft <madduck@debian.org>
: :'  :  proud Debian developer, author, administrator, and user
`. `'`   http://people.debian.org/~madduck - http://debiansystem.info
  `-  Debian - when you have better things to do than fixing systems

[vde_plug.strace.bz2 (application/octet-stream, attachment)]

[vde_plug.strace2.bz2 (application/octet-stream, attachment)]

[digital_signature_gpg.asc (application/pgp-signature, inline)]

Message #8 received at 484265@bugs.debian.org (full text, mbox, reply):

From: Filippo Giunchedi <filippo@debian.org>

To: martin f krafft <madduck@debian.org>, 484265@bugs.debian.org

Subject: Re: Bug#484265: vde_plug does not connect to socket

Date: Sun, 15 Jun 2008 10:32:43 +0200

On Tue, Jun 03, 2008 at 12:19:18PM +0200, martin f krafft wrote:
> There's some weird stuff happening in vde_plug: I have a vde_switch
> started with ctlsock directory in ./vde0, and now I am trying to
> plug into it:
> 
> +lapse:~|master|.tmp/testnet.k13727% ls vde0                              #1839
> ctl  vde_switch.pid
> +lapse:~|master|.tmp/testnet.k13727% strace -f vde_plug vde0/ctl          #1840
> [...]
> connect(5, {sa_family=AF_FILE, path="vde0/ctl/ctl"}, 110) = -1 ENOTDIR (Not a directory)
> [...]
> send(5, "\316\372\355\376\3\0\0\0\0\0\0\0\1\0vde0/ctl.14357-000"..., 183, 0) = -1 ENOTCONN (Transport endpoint is not connected)
> 
> Note how the manpage says control socket path, and yet the connect()
> seems to take it as directory. And even though the
> vde0/ctl.14357-000 control socket gets created (it now exists and is
> never removed), the send() fails.

can you try with absolute paths? I guess the problem is there, it should be
fixed in upstream svn and in the next release which is upcoming.

thanks,
filippo
--
Filippo Giunchedi - http://esaurito.net
PGP key: 0x6B79D401
random quote follows:

Debugging is twice as hard as writing the code in the first place.
Therefore, if you write the code as cleverly as possible, you are, by
definition, not smart enough to debug it.
-- Brian W. Kernighan

Message #17 received at 484265-done@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Luca Bigliardi <shammash@artha.org>, Filippo Giunchedi <filippo@debian.org>

Cc: 484265-done@bugs.debian.org

Subject: vde "failure to connect to switch" bug fixed in 2.2.1-1

Date: Tue, 8 Jul 2008 15:27:50 +0200

[Message part 1 (text/plain, inline)]

Version: 2.2.1-1

also sprach Luca Bigliardi <shammash@artha.org> [2008.06.12.0022 +0200]:
> can you try with absolute paths? I guess the problem is there, it
> should be fixed in upstream svn and in the next release which is
> upcoming.

I can confirm that #484265 is not reproducible even with relative
paths using 2.2.1-1.

-- 
 .''`.   martin f. krafft <madduck@debian.org>
: :'  :  proud Debian developer, author, administrator, and user
`. `'`   http://people.debian.org/~madduck - http://debiansystem.info
  `-  Debian - when you have better things to do than fixing systems
 
"we did rate the microsoft security researcher as
 less-bad than the people who prepare the carcasses for dissection in
 biology laboratories."
             -- michael moyer, executive editor of _popular science_

[digital_signature_gpg.asc (application/pgp-signature, inline)]

Send a report that this bug log contains spam.