linux-kernel - Re: NCPFS and brittle connections

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <459E01B2.50309@vc.cvut.cz>
Date:	Thu, 04 Jan 2007 23:43:46 -0800
From:	Petr Vandrovec <vandrove@...cvut.cz>
To:	Pierre Ossman <drzeus-list@...eus.cx>
CC:	LKML <linux-kernel@...r.kernel.org>
Subject: Re: NCPFS and brittle connections

Pierre Ossman wrote:
> Petr Vandrovec wrote:
>> Nobody is working on it (at least to my knowledge), and to me it is
>> feature - it always worked this way, like smbfs did back in the past -
>> if you send signal 9 to process using mount point, and there is some
>> transaction in progress, nobody can correctly finish that transaction
>> anymore.  Fixing it would require non-trivial amount of code, and
>> given that NCP itself is more or less dead protocol I do not feel that
>> it is necessary.
>>
> 
> Someone needs to tell our customers then so they'll stop using it. :)

When I asked at Brainshare 2001 when support for files over 4GB files 
will be added to NCP, they told me that I should switch to CIFS or 
NFS...  Years after that confirmed it - only NW6.5SP3 finally got NCPs 
to support for files over 4GB, although you could have such files even 
on NW5.

>> If you want to fix it, feel free.  Culprit is RQ_INPROGRESS handling
>> in ncp_abort_request - it just aborts whole connection so it does not
>> have to provide temporary buffers and special handling for reply - as
>> buffers currently specified as reply buffers are owned by caller, so
>> after aborting request you cannot use them anymore.
> 
> Do you have any pointers to how it was solved with smbfs? Relevant
> patches perhaps? Provided a similar solution can be applied here.

I believe it was fixed when smbiod was introduced.  When find_request() 
returns failure, it simple throws away data received from network.

Unfortunately NCP does not run on top of TCP stream, but on top of 
IPX/UDP, and so dropping reply is not sufficient - you must continue 
resending request (so you must buffer it somewhere...) until you get 
result from server - after you receive answer from server, you can 
finally throw away both request & reply, and move on.
							Petr


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/