[jcifs] Andx bug w/ IBM cluster server (was Bug: Error in
SmbTree.java??)
Michael B Allen
ioplex at gmail.com
Sun Feb 22 20:19:58 GMT 2009
Correction. Try this instead:
http://www.ioplex.com/~miallen/jcifs-1.3.4b.jar
http://www.ioplex.com/~miallen/jcifs-1.3.4b.tgz
http://www.ioplex.com/~miallen/jcifs-1.3.4b.zip
The 1.3.4a had a signature failure problen. Turns out there was a
reason for the ANDX response not being decoded like it should have.
It's probably not worth going into but for posterity all of this
stemmed from the offline files / client side caching (csc) wordCount
goof in the SMB_COM_NT_CREATE_ANDX response. It should all be right
now.
On Sat, Feb 21, 2009 at 2:00 PM, Michael B Allen <ioplex at gmail.com> wrote:
> Johannes,
>
> Does this fix it?
>
<snip links to bad release>
>
> The SMB_COM_TREE_CONNECT_ANDX response in your screenshot shows
> additional "parameter words" (7 instead of the usual 3) that JCIFS was
> not decoding correctly. I've never seen that before (must be an IBM
> cluster thing) but I have fixed JCIFS so that it should properly
> advance the buffer by the number parameter words as opposed to the
> content of the parameter words and as a result move on to decoding the
> "parameter bytes" correctly. The service string is in the parameter
> bytes so it was not being decoded properly and thus the trans2 issue
> was just an artifact of the message not being decoded properly.
>
> For future reference, your initial message should probably have
> mentioned that you were using an IBM cluster server. Obviously if no
> one else reported this issue there is a good chance it's
> environmental.
>
> Mike
>
> On Sat, Feb 21, 2009 at 5:18 AM, Johannes Bredner
> <johannes.bredner at googlemail.com> wrote:
>> Hi,
>>
>> i now figured out where the problem occurs. See working-call.jpg and
>> not-working-call.jpg. Working-call is a request to the ibm hosted server and
>> the other from a request of the other provider. The error occurs when the
>> response is received. Out of this the variable service is taken. This
>> response does not contain an "A:" for the service that is set by receiving
>> the response. I don`t know wich packet to inspect to look for the content of
>> the response for service because the information is not readable for me.
>> Could this happen caused by clustering the servers?
>>
>> Does this help you? Do you need more information?
>>
>> Best regards,
>>
>> Johannes
>>
>>
>> 2009/2/19 Johannes Bredner <johannes.bredner at googlemail.com>
>>>
>>> The programm is more than simple.
>>>
>>> Set wins, user, password, domain in jcifs conffig, new SMBFile and call a
>>> SmbFile.listFiles();
>>>
>>> This Programm worked fine for our IBM hosted servers, but we have another
>>> provier that has clustered win 2k3 servers where the problem occurs. I think
>>> in case of the ibm servers two non "A:" are called with
>>> command = 50
>>> subCommand = 16
>>> and
>>> command = 50
>>> subCommand = 32
>>>
>>> Im not at work now, but will debug the code tomorrow and hope to find out
>>> why the calls are different.
>>>
>>> Johannes
>>>
>>>
>>>
>>> 2009/2/19 Michael B Allen <ioplex at gmail.com>
>>>>
>>>> I have not looked at this very carefully but I think the service for
>>>> TRANS2_FIND_FIRST2 should be "A:" so your code change should be
>>>> unnecessary.
>>>>
>>>> What are the conditions under which you get this error?
>>>>
>>>> Can you provide a small test program that replicates the error?
>>>>
>>>> Mike
>>>>
>>>> On Thu, Feb 19, 2009 at 8:29 AM, Johannes Bredner
>>>> <johannes.bredner at googlemail.com> wrote:
>>>> > Hi all there,
>>>> >
>>>> > i think i`ve found a bug in SmbTree.java. Here the Problem:
>>>> >
>>>> > In the constructor of Trans2FindFirst2.java
>>>> >
>>>> > Trans2FindFirst2(String filename, String wildcard, int
>>>> > searchAttributes){
>>>> > ........
>>>> > command = SMB_COM_TRANSACTION2;
>>>> > subCommand = TRANS2_FIND_FIRST2;
>>>> > .....
>>>> > }
>>>> >
>>>> > This subcommand results in an exception in SmbTree.java
>>>> >
>>>> >
>>>> >
>>>> >
>>>> > ------------------------------------------------------------------------------------------------------------------------
>>>> > void send( ServerMessageBlock request,
>>>> > ServerMessageBlock response ) throws
>>>> > SmbException {
>>>> > if( response != null ) {
>>>> > response.received = false;
>>>> > }
>>>> > treeConnect( request, response );
>>>> > if( request == null || (response != null && response.received ))
>>>> > {
>>>> > return;
>>>> > }
>>>> > if( service.equals( "A:" ) == false ) {
>>>> > switch( request.command ) {
>>>> > case ServerMessageBlock.SMB_COM_OPEN_ANDX:
>>>> > case ServerMessageBlock.SMB_COM_NT_CREATE_ANDX:
>>>> > case ServerMessageBlock.SMB_COM_READ_ANDX:
>>>> > case ServerMessageBlock.SMB_COM_WRITE_ANDX:
>>>> > case ServerMessageBlock.SMB_COM_CLOSE:
>>>> > case ServerMessageBlock.SMB_COM_TREE_DISCONNECT:
>>>> > break;
>>>> > case ServerMessageBlock.SMB_COM_TRANSACTION:
>>>> > case ServerMessageBlock.SMB_COM_TRANSACTION2:
>>>> > switch( ((SmbComTransaction)request).subCommand &
>>>> > 0xFF
>>>> > ) {
>>>> > case SmbComTransaction.NET_SHARE_ENUM:
>>>> > case SmbComTransaction.NET_SERVER_ENUM2:
>>>> > case SmbComTransaction.NET_SERVER_ENUM3:
>>>> > case SmbComTransaction.TRANS_PEEK_NAMED_PIPE:
>>>> > case SmbComTransaction.TRANS_WAIT_NAMED_PIPE:
>>>> > case SmbComTransaction.TRANS_CALL_NAMED_PIPE:
>>>> > case
>>>> > SmbComTransaction.TRANS_TRANSACT_NAMED_PIPE:
>>>> > case SmbComTransaction.TRANS2_GET_DFS_REFERRAL:
>>>> > break;
>>>> > default:
>>>> > throw new SmbException( "Invalid operation
>>>> > for " + service + " service" );
>>>> > }
>>>> > break;
>>>> > default:
>>>> > throw new SmbException( "Invalid operation for " +
>>>> > service + " service" + request );
>>>> > }
>>>> > }
>>>> > request.tid = tid;
>>>> >
>>>> > ------------------------------------------------------------------------------------------------------------------------
>>>> >
>>>> > throws the Exception "Invalid operation for service". Here my idea:
>>>> >
>>>> >
>>>> >
>>>> > ------------------------------------------------------------------------------------------------------------------------
>>>> > void send( ServerMessageBlock request,
>>>> > ServerMessageBlock response ) throws
>>>> > SmbException {
>>>> > if( response != null ) {
>>>> > response.received = false;
>>>> > }
>>>> > treeConnect( request, response );
>>>> > if( request == null || (response != null && response.received ))
>>>> > {
>>>> > return;
>>>> > }
>>>> > if( service.equals( "A:" ) == false ) {
>>>> > switch( request.command ) {
>>>> > case ServerMessageBlock.SMB_COM_OPEN_ANDX:
>>>> > case ServerMessageBlock.SMB_COM_NT_CREATE_ANDX:
>>>> > case ServerMessageBlock.SMB_COM_READ_ANDX:
>>>> > case ServerMessageBlock.SMB_COM_WRITE_ANDX:
>>>> > case ServerMessageBlock.SMB_COM_CLOSE:
>>>> > case ServerMessageBlock.SMB_COM_TREE_DISCONNECT:
>>>> > break;
>>>> > case ServerMessageBlock.SMB_COM_TRANSACTION:
>>>> > case ServerMessageBlock.SMB_COM_TRANSACTION2:
>>>> > switch( ((SmbComTransaction)request).subCommand &
>>>> > 0xFF
>>>> > ) {
>>>> > case SmbComTransaction.NET_SHARE_ENUM:
>>>> > case SmbComTransaction.NET_SERVER_ENUM2:
>>>> > case SmbComTransaction.NET_SERVER_ENUM3:
>>>> > case SmbComTransaction.TRANS_PEEK_NAMED_PIPE:
>>>> > case SmbComTransaction.TRANS_WAIT_NAMED_PIPE:
>>>> > case SmbComTransaction.TRANS_CALL_NAMED_PIPE:
>>>> > case
>>>> > SmbComTransaction.TRANS_TRANSACT_NAMED_PIPE:
>>>> > case SmbComTransaction.TRANS2_GET_DFS_REFERRAL:
>>>> > case SmbComTransaction.TRANS2_FIND_FIRST2:
>>>> > break;
>>>> > default:
>>>> > throw new SmbException( "Invalid operation
>>>> > for " + service + " service" );
>>>> > }
>>>> > break;
>>>> > default:
>>>> > throw new SmbException( "Invalid operation for " +
>>>> > service + " service" + request );
>>>> > }
>>>> > }
>>>> > request.tid = tid;
>>>> >
>>>> >
>>>> > ------------------------------------------------------------------------------------------------------------------------
>>>> >
>>>> > If i add the subcommand line for TRANS2_FIND_FIRST2 everything works
>>>> > fine,
>>>> > but i don`t really know if this is the right way to fix the problem.
>>>> > Just
>>>> > seems, that this command is missing in the list.
>>>> >
>>>> > Do you agree with my fix? Is this the right way to report bugs?
>>>> >
>>>> > Please answer to this email, because i`m not on the list at the moment.
>>>> >
>>>> > Best regards,
>>>> >
>>>> > Johannes
>>>> >
>>>>
>>>>
>>>>
>>>> --
>>>> Michael B Allen
>>>> Java Active Directory Integration
>>>> http://www.ioplex.com/
>>>
>>
>>
>
>
>
> --
> Michael B Allen
> Java Active Directory Integration
> http://www.ioplex.com/
>
--
Michael B Allen
Java Active Directory Integration
http://www.ioplex.com/
More information about the jcifs
mailing list