9fans - fans of the OS Plan 9 from Bell Labs
 help / color / mirror / Atom feed
* [9fans] Aquarela faulting
@ 2010-03-01 21:39 Georg Lehner
  2010-03-02 21:22 ` erik quanstrom
  0 siblings, 1 reply; 11+ messages in thread
From: Georg Lehner @ 2010-03-01 21:39 UTC (permalink / raw)
  To: Fans of the OS Plan 9 from Bell Labs

Hello!

I'm trying to connect to a Plan9 file server via SMB, but cannot get a connection.

On plan9, as user glenda i run:

% bind /usr/glenda /n/glenda
% aquarela -p -n allcmds -w MAGMA

and get the following output (shortened to the last view lines) when
connecting e.g. from Windows XP Home with

> net use \\acp\glenda password /user:glenda

----
...
andxcommand: SMB_COM_TREE_CONNECT_ANDX offset 242
client.maxbuffersize: 61440
client.maxmpxcount: 1
vcnumber: 0
sessionkey: 0x00000004
caseinsensitivepasswordlength: 24
casesensitivepasswordlength: 24
clientcapabilities: 0x000000d4
account: glenda
primarydomain: TOSHIBA
nativeos: Windows 2002 Service Pack 3 2600
nativelanman: Windows 2002 5.1
aquarela 3927: suicide: sys: trap: fault write addr=0x0 pc=0x8dcb
----

Any ideas where to start from?
Any best practices or pitfalls for this setup?

Regards,

	Jorge-Le�n





^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [9fans] Aquarela faulting
  2010-03-01 21:39 [9fans] Aquarela faulting Georg Lehner
@ 2010-03-02 21:22 ` erik quanstrom
  2010-03-05 11:00   ` Georg Lehner
  0 siblings, 1 reply; 11+ messages in thread
From: erik quanstrom @ 2010-03-02 21:22 UTC (permalink / raw)
  To: 9fans

> Any ideas where to start from?
> Any best practices or pitfalls for this setup?

perhaps the new cifs server would be better?

if you're looking for this bug, it's clear that there's
something out-of-order with the authentication.
the server is especting some auth info, and there
isn't any.  i suppose that this is also a security bug.

- erik



^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [9fans] Aquarela faulting
  2010-03-02 21:22 ` erik quanstrom
@ 2010-03-05 11:00   ` Georg Lehner
  2010-03-05 13:13     ` erik quanstrom
  2010-03-10  5:22     ` cinap_lenrek
  0 siblings, 2 replies; 11+ messages in thread
From: Georg Lehner @ 2010-03-05 11:00 UTC (permalink / raw)
  To: Fans of the OS Plan 9 from Bell Labs

erik quanstrom wrote:
> perhaps the new cifs server would be better?
>

A pointer please.

Already searched mailing list, (8) man pages, and the contrib packages
but only found mention to aquarela as a cifs server.

Thanks, and

Regards,

    Jorge-Le�n



^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [9fans] Aquarela faulting
  2010-03-05 11:00   ` Georg Lehner
@ 2010-03-05 13:13     ` erik quanstrom
  2010-03-05 13:29       ` Richard Miller
  2010-03-06 21:46       ` Georg Lehner
  2010-03-10  5:22     ` cinap_lenrek
  1 sibling, 2 replies; 11+ messages in thread
From: erik quanstrom @ 2010-03-05 13:13 UTC (permalink / raw)
  To: 9fans

> erik quanstrom wrote:
> > perhaps the new cifs server would be better?
> >
>
> A pointer please.
>
> Already searched mailing list, (8) man pages, and the contrib packages
> but only found mention to aquarela as a cifs server.

it's in the distribution.

cifs(4),  /sys/src/cmd/cifs, /$objtype/bin/aux/cifs, /$objtype/bin/cifscmd

- erik



^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [9fans] Aquarela faulting
  2010-03-05 13:13     ` erik quanstrom
@ 2010-03-05 13:29       ` Richard Miller
  2010-03-05 13:46         ` erik quanstrom
  2010-03-06 21:46       ` Georg Lehner
  1 sibling, 1 reply; 11+ messages in thread
From: Richard Miller @ 2010-03-05 13:29 UTC (permalink / raw)
  To: 9fans

> it's in the distribution.
>
> cifs(4),  /sys/src/cmd/cifs, /$objtype/bin/aux/cifs, /$objtype/bin/cifscmd

Looks like the lookman index should be updated:

term% lookman cifs
man 8 aquarela # aquarela(8)




^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [9fans] Aquarela faulting
  2010-03-05 13:29       ` Richard Miller
@ 2010-03-05 13:46         ` erik quanstrom
  2010-03-05 15:21           ` Richard Miller
  0 siblings, 1 reply; 11+ messages in thread
From: erik quanstrom @ 2010-03-05 13:46 UTC (permalink / raw)
  To: 9fans

> > it's in the distribution.
> >
> > cifs(4),  /sys/src/cmd/cifs, /$objtype/bin/aux/cifs, /$objtype/bin/cifscmd
>
> Looks like the lookman index should be updated:
>
> term% lookman cifs
> man 8 aquarela # aquarela(8)

if you install even one man page that's not in the distribution,
you should do this yourself.  see the realitively
new /cron/sys/cron:

# update manual indices
0 4 * * 0	local	cd /sys/man && mk indices

i have this installed in the hostowners account,
but commented out.  i run it by hand on an as-needed
basis.

- erik



^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [9fans] Aquarela faulting
  2010-03-05 13:46         ` erik quanstrom
@ 2010-03-05 15:21           ` Richard Miller
  2010-03-05 19:29             ` geoff
  0 siblings, 1 reply; 11+ messages in thread
From: Richard Miller @ 2010-03-05 15:21 UTC (permalink / raw)
  To: 9fans

> see the realitively
> new /cron/sys/cron:

No /cron/sys/cron on my system.  It's not in /dist/replica/plan9.proto
so it doesn't get distributed from sources.  Should it?




^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [9fans] Aquarela faulting
  2010-03-05 15:21           ` Richard Miller
@ 2010-03-05 19:29             ` geoff
  0 siblings, 0 replies; 11+ messages in thread
From: geoff @ 2010-03-05 19:29 UTC (permalink / raw)
  To: 9fans

[-- Attachment #1: Type: text/plain, Size: 53 bytes --]

/cron/sys/cron is now in /dist/replica/plan9.proto.

[-- Attachment #2: Type: message/rfc822, Size: 2011 bytes --]

From: Richard Miller <9fans@hamnavoe.com>
To: 9fans@9fans.net
Subject: Re: [9fans] Aquarela faulting
Date: Fri, 5 Mar 2010 15:21:39 +0000
Message-ID: <755049dee21ba921776978317ac49d6f@hamnavoe.com>

> see the realitively
> new /cron/sys/cron:

No /cron/sys/cron on my system.  It's not in /dist/replica/plan9.proto
so it doesn't get distributed from sources.  Should it?


^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [9fans] Aquarela faulting
  2010-03-05 13:13     ` erik quanstrom
  2010-03-05 13:29       ` Richard Miller
@ 2010-03-06 21:46       ` Georg Lehner
  1 sibling, 0 replies; 11+ messages in thread
From: Georg Lehner @ 2010-03-06 21:46 UTC (permalink / raw)
  To: Fans of the OS Plan 9 from Bell Labs

As far as i understand, cifs(4) is a cifs client, which makes remote
Windows shares
available to a Plan9 machine, while aquarela(8) is a cifs server, which
makes (lokal)
Plan9 filesystems available as Windows shares to remote clients.

So cifs(4) is the complent of aquarela(8), not a replacement.

Did i miss something?

Regards,

    Jorge-Le�n

erik quanstrom wrote:
>> erik quanstrom wrote:
>>
>>> perhaps the new cifs server would be better?
>>>
>>>
>> A pointer please.
>>
>> Already searched mailing list, (8) man pages, and the contrib packages
>> but only found mention to aquarela as a cifs server.
>>
>
> it's in the distribution.
>
> cifs(4),  /sys/src/cmd/cifs, /$objtype/bin/aux/cifs, /$objtype/bin/cifscmd
>
> - erik
>
>




^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [9fans] Aquarela faulting
  2010-03-05 11:00   ` Georg Lehner
  2010-03-05 13:13     ` erik quanstrom
@ 2010-03-10  5:22     ` cinap_lenrek
  2010-03-26 18:24       ` Georg Lehner
  1 sibling, 1 reply; 11+ messages in thread
From: cinap_lenrek @ 2010-03-10  5:22 UTC (permalink / raw)
  To: 9fans

[-- Attachment #1: Type: text/plain, Size: 1728 bytes --]

just got a aquarela crash when trying to mount it from windows xp.
w2k worked fine.

the offending code (with fix) was this:

/n/dump/2010/0307/sys/src/cmd/aquarela/smbcomsessionsetupandx.c:119,124 - smbcomsessionsetupandx.c:119,130
  			smbseterror(s, ERRSRV, ERRbadpw);
  			goto errordone;
  		}
+ 		if(s->cs == nil){
+ 			smblogprint(h->command,
+ 				"smbcomsessionsetupandx: s->cs == nil\n");
+ 			smbseterror(s, ERRSRV, ERRbadpw);
+ 			goto errordone;
+ 		}
  		memcpy(&s->client.mschapreply, mschapreply, sizeof(s->client.mschapreply));
  		s->cs->user = accountname;
  		s->cs->resp = &s->client.mschapreply;

turned out, s->cs was nil and i got the following message when turning on debugging:

		s->cs = auth_challenge("proto=mschap role=server");
		if (s->cs == nil) {
			smblogprint(h->command, "smbnegotiate: couldn't get mschap challenge\n");
			return SmbProcessResultMisc;
		}

so auth failed.  i'v diffed the authserver and factotum code and
noticed that here where some mschap related changes so i applied them,
build factotum and authsrv and made a new kernel and it all works now.

not sure if this was the same problem here...  its easy to diagnose...
touch /sys/log/aquarela and turn on debugging.  also when something
crashes you make ps, find the broken process pid and run acid pid, then
type lstk(). then you copy and paste the output of acid and put them in
your troubleshooting mail.

another thing...  could we replace the ascii art wich looks like a mix
of c-comments, lisp, brainfuck and hitler that the windows network drive
property dialog shows on the volume size of the aquarela share to
something else?

"Capacity:	/*-(, bytes -16384"

--
cinap

[-- Attachment #2: Type: message/rfc822, Size: 2917 bytes --]

From: Georg Lehner <jorge-plan9@magma.com.ni>
To: Fans of the OS Plan 9 from Bell Labs <9fans@9fans.net>
Subject: Re: [9fans] Aquarela faulting
Date: Fri, 05 Mar 2010 12:00:45 +0100
Message-ID: <4B90E45D.6080701@magma.com.ni>

erik quanstrom wrote:
> perhaps the new cifs server would be better?
>

A pointer please.

Already searched mailing list, (8) man pages, and the contrib packages
but only found mention to aquarela as a cifs server.

Thanks, and

Regards,

    Jorge-Le�n

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [9fans] Aquarela faulting
  2010-03-10  5:22     ` cinap_lenrek
@ 2010-03-26 18:24       ` Georg Lehner
  0 siblings, 0 replies; 11+ messages in thread
From: Georg Lehner @ 2010-03-26 18:24 UTC (permalink / raw)
  To: Fans of the OS Plan 9 from Bell Labs

Hello!

Thanks for the fix, aquarela now does not die on me anymore.

The hard way i also learnt, that i have to modify /rc/bin/9fs to get
anything useful working,
i added e.g.
- - -
  case usr
            bind /usr /n/usr
- - -
to be able to access \\acp\usr over the network, else aquarela would try
to dial net!usr!9fs
and fail on it.

Since i still can not get my setup to work satisfactorily here some
problems i have:

There is a Plan9 cpu/file/auth server named 'acp' with hostowner named
glenda.
 From the Plan9 server console i run:

  aquarela -u 1 -p -d sessions -w MAGMA

 From Windows XP Home commandline i can run successfully

  net use L: \\acp\acp password /user:glenda

and then browse with Windows Explorer the L: drive to L:\usr\glenda and
create/delete files and directories
there.

The same works from Windows Explorer menu bar: Tools -> Map Network drive.

On the Plan9 server also exists another user: jorge

Whenever i try to

  net use L:\\acp\acp password /user:jorge

or 'Map Network drive' with this user

i get the following error message on the Plan9 console:

  srv acp: mount failed: fossil authCheck: auth protocol not finished

Also another usage szenario fails:

Introducing '\\acp\acp' in the Windows Explorer Address bar, or in the
Start Menu's 'Run'
dialog gives me first a password dialog, which i answer with 'glenda'
etc., and then yields
an unresponsive Windows Explorer Windows and the following error
messages on the
Plan9 console:

  netsharegetinfo: service acp unimplemented
  srv: timeout establishing connection to net!.hg!9fs
  netsharegetinfo: service .hg unimplemented
  ....

(Note: i have TortoiseHG installed, which obviously tries to find a .hg
directory)

After some minutes i get the root-directory listing, but already lost
patience to try out if
i can walk to usr\glenda and kill Windows Explorer.  The same happens
with user 'jorge'
and with other shares, e.g.  'usr'.


Since i have also Linux clients in my network, i tried to connect from
them, but in vain:

  jorge@tink:~$ smbclient  -d 1 -W MAGMA -U glenda   //acp/dump
  Enter glenda's password:
  Domain=[MAGMA] OS=[Plan 9 4th edition] Server=[Aquarela]
  smb: \> SMBecho failed. Maybe server has closed the connection


Some hints?

Best Regards,

    Jorge-Le�n


cinap_lenrek@gmx.de wrote:
> just got a aquarela crash when trying to mount it from windows xp.
> w2k worked fine.
>
> the offending code (with fix) was this:
>
> /n/dump/2010/0307/sys/src/cmd/aquarela/smbcomsessionsetupandx.c:119,124 - smbcomsessionsetupandx.c:119,130
>   			smbseterror(s, ERRSRV, ERRbadpw);
>   			goto errordone;
>   		}
> + 		if(s->cs == nil){
> + 			smblogprint(h->command,
> + 				"smbcomsessionsetupandx: s->cs == nil\n");
> + 			smbseterror(s, ERRSRV, ERRbadpw);
> + 			goto errordone;
> + 		}
>   		memcpy(&s->client.mschapreply, mschapreply, sizeof(s->client.mschapreply));
>   		s->cs->user = accountname;
>   		s->cs->resp = &s->client.mschapreply;
>
> turned out, s->cs was nil and i got the following message when turning on debugging:
>
> 		s->cs = auth_challenge("proto=mschap role=server");
> 		if (s->cs == nil) {
> 			smblogprint(h->command, "smbnegotiate: couldn't get mschap challenge\n");
> 			return SmbProcessResultMisc;
> 		}
>
> so auth failed.  i'v diffed the authserver and factotum code and
> noticed that here where some mschap related changes so i applied them,
> build factotum and authsrv and made a new kernel and it all works now.
>
> not sure if this was the same problem here...  its easy to diagnose...
> touch /sys/log/aquarela and turn on debugging.  also when something
> crashes you make ps, find the broken process pid and run acid pid, then
> type lstk(). then you copy and paste the output of acid and put them in
> your troubleshooting mail.
>
> another thing...  could we replace the ascii art wich looks like a mix
> of c-comments, lisp, brainfuck and hitler that the windows network drive
> property dialog shows on the volume size of the aquarela share to
> something else?
>
> "Capacity:	/*-(, bytes -16384"
>
> --
> cinap
>




^ permalink raw reply	[flat|nested] 11+ messages in thread

end of thread, other threads:[~2010-03-26 18:24 UTC | newest]

Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2010-03-01 21:39 [9fans] Aquarela faulting Georg Lehner
2010-03-02 21:22 ` erik quanstrom
2010-03-05 11:00   ` Georg Lehner
2010-03-05 13:13     ` erik quanstrom
2010-03-05 13:29       ` Richard Miller
2010-03-05 13:46         ` erik quanstrom
2010-03-05 15:21           ` Richard Miller
2010-03-05 19:29             ` geoff
2010-03-06 21:46       ` Georg Lehner
2010-03-10  5:22     ` cinap_lenrek
2010-03-26 18:24       ` Georg Lehner

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).