Skip to contentSkip to conversation

Problem with connecting to SMB server in 3.9.11 kernel.

DISCUSSION Ben Greear 6 participants 14 messages Jan 18, 2014 – Feb 3, 2014 9 messages not in the archive

Subsystems (2)
lore
Share & tools
lore

Ben Greear

author
#0 lore raw

Hello!

A customer reported problems connecting our CIFS traffic generation test gear to their SMB server. We are using the 3.9.11+ kernel, and though it is patched, we do not have any patches to cifs.

The OS is Fedora 14, 64-bit.

A similar system on Fedora 14, with a 3.7.10+ kernel was working, but when we tried the 3.7.10+ kernel on the broken machine, it also failed to work. So, it could be the SMB server itself is having issues. (In the 3.7.10+ failure, the server just failed to respond after "Setup andX Request AUTH message", though we did get the TCP ack so it looks like the message was received by the server.)

I'm attaching a capture taken on the SMB server.

From looking at this page:

http://msdn.microsoft.com/en-us/library/ff469913.aspx

It appears the problem (STATUS_UNSUCCESSFUL) is:

"The size of the extended attribute list is not correct. Check the EaErrorOffset field for the address of the SMB_GEA structure at which the error was detected."

I did not see anything about extended attribute list in the capture, but if someone else with more knowledge could take a look and see if they notice any problems I would be grateful.

Thanks!
Ben

Signature · 2 lines
Ben Greear <greearb-my8/[email protected]>
Candela Technologies Inc  http://www.candelatech.com
#2 lore raw
In reply to a message not in the archive
Quoted 22 lines · On Sat, 18 Jan 2014 09:17:32 -0800 Ben Greear…

Hello!

A customer reported problems connecting our CIFS traffic generation test gear to their SMB server. We are using the 3.9.11+ kernel, and though it is patched, we do not have any patches to cifs.

The OS is Fedora 14, 64-bit.

A similar system on Fedora 14, with a 3.7.10+ kernel was working, but when we tried the 3.7.10+ kernel on the broken machine, it also failed to work. So, it could be the SMB server itself is having issues. (In the 3.7.10+ failure, the server just failed to respond after "Setup andX Request AUTH message", though we did get the TCP ack so it looks like the message was received by the server.)

I'm attaching a capture taken on the SMB server.

From looking at this page:

http://msdn.microsoft.com/en-us/library/ff469913.aspx

It appears the problem (STATUS_UNSUCCESSFUL) is:

Quoted 9 lines

"The size of the extended attribute list is not correct. Check the EaErrorOffset field for the address of the SMB_GEA structure at which the error was detected."

I did not see anything about extended attribute list in the capture, but if someone else with more knowledge could take a look and see if they notice any problems I would be grateful.

Thanks!
Ben

There's no EA list on this call so that description isn't valid here.

AIUI, NT_STATUS_UNSUCCESSFUL is basically a generic "something went wrong" error (sort of like EIO on POSIX). It looks like the server doesn't like something about the request being sent, but it's tough to know what it is.

I cc'ed Shirish though as he wrote most of that code. Maybe he has some idea?

Signature · 1 line
Jeff Layton <[email protected]>
#4 lore raw
In reply to a message not in the archive
Quoted 49 lines

-----Original Message-----
From: [email protected] [mailto:linux-cifs- [email protected]] On Behalf Of Jeff Layton
Sent: Monday, January 20, 2014 7:57 AM
To: Ben Greear
Cc: [email protected]; Amit Haval (India Bangalore); Shirish Pargaonkar
Subject: Re: Problem with connecting to SMB server in 3.9.11 kernel.

Quoted 3 lines · On Sat, 18 Jan 2014 09:17:32 -0800 Ben Greear…

Hello!

A customer reported problems connecting our CIFS traffic generation test

gear to their SMB server.

Quoted 7 lines

We are using the 3.9.11+ kernel, and though it is patched, we do not have any patches to cifs.

The OS is Fedora 14, 64-bit.

A similar system on Fedora 14, with a 3.7.10+ kernel was working, but when we tried the 3.7.10+ kernel on the broken machine, it also failed to

work.

Quoted 7 lines

So, it could be the SMB server itself is having issues. (In the 3.7.10+ failure, the server just failed to respond after "Setup andX Request AUTH message", though we did get the TCP ack so it looks like the message was received by the server.)

I'm attaching a capture taken on the SMB server.

Quoted 8 lines

From looking at this page:

http://msdn.microsoft.com/en-us/library/ff469913.aspx

It appears the problem (STATUS_UNSUCCESSFUL) is:

"The size of the extended attribute list is not correct. Check the EaErrorOffset field for the address of the SMB_GEA structure at which the

error was detected."

Quoted 3 lines

I did not see anything about extended attribute list in the capture, but if someone else with more knowledge could take a look and see if they

notice any problems I would be grateful.

Quoted 4 lines

Thanks!
Ben

Quoted 5 lines

There's no EA list on this call so that description isn't valid here.

AIUI, NT_STATUS_UNSUCCESSFUL is basically a generic "something went wrong" error (sort of like EIO on POSIX). It looks like the server doesn't like something about the request being sent, but it's tough to know what it is.

The trace Ben attached shows an authentication failure, and I don't see any TRANS2_QUERY_FILE_INFORMATION in the trace at all as Ben references. The client is still sending its SESSION_SETUP attempt. Looks like the client is attempting NTLMSSP/NTLMv2 for NICVALIDATION\administrator, and this status is resulting from the authentication and not the SMB protocol.

Microsoft Message Analyzer shows no problem with the packet structure itself, in any case. So this doesn't appear to be a malformed request.

What type of server is this? The NativeOS string is "SpinStream2", which I don't recognize.

#6 lore raw
In reply to a message not in the archive

Adding our team to comment/help.

-----Original Message-----
From: Tom Talpey [mailto:[email protected]]
Sent: Monday, January 20, 2014 7:17 PM
To: Jeff Layton; Ben Greear
Cc: [email protected]; Amit Haval (India Bangalore); Shirish Pargaonkar
Subject: RE: Problem with connecting to SMB server in 3.9.11 kernel.

Quoted 53 lines

-----Original Message-----
From: [email protected] [mailto:linux-cifs- [email protected]] On Behalf Of Jeff Layton
Sent: Monday, January 20, 2014 7:57 AM
To: Ben Greear
Cc: [email protected]; Amit Haval (India Bangalore); Shirish Pargaonkar
Subject: Re: Problem with connecting to SMB server in 3.9.11 kernel.

Quoted 4 lines · On Sat, 18 Jan 2014 09:17:32 -0800 Ben Greear…

Hello!

A customer reported problems connecting our CIFS traffic generation test

gear to their SMB server.

Quoted 8 lines

We are using the 3.9.11+ kernel, and though it is patched, we do not have any patches to cifs.

The OS is Fedora 14, 64-bit.

A similar system on Fedora 14, with a 3.7.10+ kernel was working, but when we tried the 3.7.10+ kernel on the broken machine, it also failed to

work.

Quoted 9 lines

So, it could be the SMB server itself is having issues. (In the 3.7.10+ failure, the server just failed to respond after "Setup andX Request AUTH message", though we did get the TCP ack so it looks like the message was received by the server.)

I'm attaching a capture taken on the SMB server.

From looking at this page:

Quoted 7 lines

http://msdn.microsoft.com/en-us/library/ff469913.aspx

It appears the problem (STATUS_UNSUCCESSFUL) is:

"The size of the extended attribute list is not correct. Check the EaErrorOffset field for the address of the SMB_GEA structure at which the

error was detected."

Quoted 4 lines

I did not see anything about extended attribute list in the capture, but if someone else with more knowledge could take a look and see if they

notice any problems I would be grateful.

Quoted 4 lines

Thanks!
Ben

Quoted 5 lines

There's no EA list on this call so that description isn't valid here.

AIUI, NT_STATUS_UNSUCCESSFUL is basically a generic "something went wrong" error (sort of like EIO on POSIX). It looks like the server doesn't like something about the request being sent, but it's tough to know what it is.

The trace Ben attached shows an authentication failure, and I don't see any TRANS2_QUERY_FILE_INFORMATION in the trace at all as Ben references. The client is still sending its SESSION_SETUP attempt. Looks like the client is attempting NTLMSSP/NTLMv2 for NICVALIDATION\administrator, and this status is resulting from the authentication and not the SMB protocol.

Microsoft Message Analyzer shows no problem with the packet structure itself, in any case. So this doesn't appear to be a malformed request.

What type of server is this? The NativeOS string is "SpinStream2", which I don't recognize.

______________________________________________________________________
This email has been scanned by the Boundary Defense for Email Security System. For more information please visit http://www.apptix.com/email-security/antispam-virus
______________________________________________________________________

______________________________________________________________________
This email has been scanned by the Boundary Defense for Email Security System. For more information please visit http://www.apptix.com/email-security/antispam-virus
______________________________________________________________________

#7 lore raw
In reply to a message not in the archive

yes, this looks like a authentication failure.
It probably is a NetApp server. Can any other client (e.g. Windows, smbclient etc.) authenticate using ntlmv2/ntlmssp to this server?

SMB servers do not typically point_out/log what part of the request is invalid, so it is little harder to figure out how the request is incorrect.

Quoted 43 lines · Ben Greear

Hello!

A customer reported problems connecting our CIFS traffic generation test gear to their SMB server.
We are using the 3.9.11+ kernel, and though it is patched, we do not have any
patches to cifs.

The OS is Fedora 14, 64-bit.

A similar system on Fedora 14, with a 3.7.10+ kernel was working, but when we tried the 3.7.10+ kernel on the broken machine, it also failed to work. So, it could be the SMB server itself is having issues. (In the 3.7.10+ failure,
the server just failed to respond after "Setup andX Request AUTH message", though we
did get the TCP ack so it looks like the message was received by the server.)

I'm attaching a capture taken on the SMB server.

From looking at this page:

http://msdn.microsoft.com/en-us/library/ff469913.aspx

It appears the problem (STATUS_UNSUCCESSFUL) is:

"The size of the extended attribute list is not correct. Check the EaErrorOffset field for the
address of the SMB_GEA structure at which the error was detected."

I did not see anything about extended attribute list in the capture, but if someone else
with more knowledge could take a look and see if they notice any problems I would be grateful.

Thanks!
Ben

--
Ben Greear <greearb-my8/[email protected]> Candela Technologies Inc http://www.candelatech.com

#9 lore raw
In reply to a message not in the archive

Adding nic-dl...

-----Original Message-----
From: Shirish Pargaonkar [mailto:[email protected]]
Sent: Monday, January 20, 2014 9:24 PM
To: Ben Greear
Cc: linux-cifs; Amit Haval (India Bangalore)
Subject: Re: Problem with connecting to SMB server in 3.9.11 kernel.

yes, this looks like a authentication failure.
It probably is a NetApp server. Can any other client (e.g. Windows, smbclient etc.) authenticate using ntlmv2/ntlmssp to this server?

SMB servers do not typically point_out/log what part of the request is invalid, so it is little harder to figure out how the request is incorrect.

Quoted 33 lines · Ben Greear

Hello!

A customer reported problems connecting our CIFS traffic generation test gear to their SMB server.
We are using the 3.9.11+ kernel, and though it is patched, we do not have any patches to cifs.

The OS is Fedora 14, 64-bit.

A similar system on Fedora 14, with a 3.7.10+ kernel was working, but when we tried the 3.7.10+ kernel on the broken machine, it also failed to work. So, it could be the SMB server itself is having issues. (In the 3.7.10+ failure, the server just failed to respond after "Setup andX Request AUTH message", though we did get the TCP ack so it looks like the message was received by the
server.)

I'm attaching a capture taken on the SMB server.

From looking at this page:

http://msdn.microsoft.com/en-us/library/ff469913.aspx

It appears the problem (STATUS_UNSUCCESSFUL) is:

"The size of the extended attribute list is not correct. Check the EaErrorOffset field for the address of the SMB_GEA structure at which the error was detected."

I did not see anything about extended attribute list in the capture, but if someone else with more knowledge could take a look and see if they notice any problems I would be grateful.

Quoted 7 lines

Thanks!
Ben

--
Ben Greear <greearb-my8/[email protected]> Candela Technologies Inc http://www.candelatech.com

______________________________________________________________________
This email has been scanned by the Boundary Defense for Email Security System. For more information please visit http://www.apptix.com/email-security/antispam-virus
______________________________________________________________________

______________________________________________________________________
This email has been scanned by the Boundary Defense for Email Security System. For more information please visit http://www.apptix.com/email-security/antispam-virus
______________________________________________________________________

#11 lore raw
In reply to a message not in the archive

I will spend more time later tonight/tomorrow but it could be related to cifs client not using Target received in NTLMSSP Challenge response from the server (since the bit Target Type Domain in flags is set) while constructing NTLMSSP Authenticate message. Just a guess.

Quoted 64 lines · Amit Haval (India Bangalore)

Adding nic-dl...

-----Original Message-----
From: Shirish Pargaonkar [mailto:[email protected]]
Sent: Monday, January 20, 2014 9:24 PM
To: Ben Greear
Cc: linux-cifs; Amit Haval (India Bangalore)
Subject: Re: Problem with connecting to SMB server in 3.9.11 kernel.

yes, this looks like a authentication failure.
It probably is a NetApp server. Can any other client (e.g. Windows, smbclient etc.) authenticate using ntlmv2/ntlmssp to this server?

SMB servers do not typically point_out/log what part of the request is invalid, so it is little harder to figure out how the request is incorrect.

Quoted 33 lines · Ben Greear

Hello!

A customer reported problems connecting our CIFS traffic generation test gear to their SMB server.
We are using the 3.9.11+ kernel, and though it is patched, we do not have any patches to cifs.

The OS is Fedora 14, 64-bit.

A similar system on Fedora 14, with a 3.7.10+ kernel was working, but when we tried the 3.7.10+ kernel on the broken machine, it also failed to work. So, it could be the SMB server itself is having issues. (In the 3.7.10+ failure, the server just failed to respond after "Setup andX Request AUTH message", though we did get the TCP ack so it looks like the message was received by the
server.)

I'm attaching a capture taken on the SMB server.

From looking at this page:

http://msdn.microsoft.com/en-us/library/ff469913.aspx

It appears the problem (STATUS_UNSUCCESSFUL) is:

"The size of the extended attribute list is not correct. Check the EaErrorOffset field for the address of the SMB_GEA structure at which the error was detected."

I did not see anything about extended attribute list in the capture, but if someone else with more knowledge could take a look and see if they notice any problems I would be grateful.

Quoted 7 lines

Thanks!
Ben

--
Ben Greear <greearb-my8/[email protected]> Candela Technologies Inc http://www.candelatech.com

______________________________________________________________________
This email has been scanned by the Boundary Defense for Email Security System. For more information please visit http://www.apptix.com/email-security/antispam-virus
______________________________________________________________________

______________________________________________________________________
This email has been scanned by the Boundary Defense for Email Security System. For more information please visit http://www.apptix.com/email-security/antispam-virus
______________________________________________________________________

author
#13 lore raw
In reply to a message not in the archive
Quoted 4 lines · Shirish Pargaonkar

I will spend more time later tonight/tomorrow but it could be related to cifs client not using Target received in NTLMSSP Challenge response from the server (since the bit Target Type Domain in flags is set) while constructing NTLMSSP Authenticate message. Just a guess.

Did you have any luck with this?

Thanks,
Ben

Signature · 2 lines
Ben Greear <greearb-my8/[email protected]>
Candela Technologies Inc  http://www.candelatech.com
#15 lore raw
In reply to a message not in the archive

I want to code to use Target Name instead of using domain name from the avpair info received in type 2 ntlmssp challenge message from the server to calculate ntlmv2 response within ntlmssp.
Will post a preliminary patch to try out with that code change (hopefully in a day or two). That is one code change I think might work.

Quoted 18 lines · Ben Greear
Quoted 5 lines · Shirish Pargaonkar

I will spend more time later tonight/tomorrow but it could be related to cifs client not using Target received in NTLMSSP Challenge response from the server (since the bit Target Type Domain in flags is set) while constructing NTLMSSP Authenticate message. Just a guess.

Did you have any luck with this?

Thanks,

Ben

--
Ben Greear <greearb-my8/[email protected]> Candela Technologies Inc http://www.candelatech.com

#16 lore raw
In reply to a message not in the archive

I Was thinking perhaps client not setting either NTLMSSP_TARGET_TYPE_SERVER or NTLMSSP_TARGET_TYPE_DOMAIN or NTLMSSP_TARGET_TYPE_SHARE
(the one that server sets in type 2/challenge message) in type 3 message is a problem but it is not (as per page 33 in ms-nlmp document).

cifs client should change to code Target Name instead of NetBIOS domain name from Target Info to calculate ntlmv2 response but that is not the problem here (because they happen to be the exact same).

Is there a way to check what kind of response this SMB server expects i.e. NTLMv1 or NTLMv2? There may be a conf file on the server effecting that.

Also, if you can provide wireshark trace for any other client besides cifs client successfully authenticating with this server, that would be useful too.

Quoted 43 lines · Ben Greear

Hello!

A customer reported problems connecting our CIFS traffic generation test gear to their SMB server.
We are using the 3.9.11+ kernel, and though it is patched, we do not have any
patches to cifs.

The OS is Fedora 14, 64-bit.

A similar system on Fedora 14, with a 3.7.10+ kernel was working, but when we tried the 3.7.10+ kernel on the broken machine, it also failed to work. So, it could be the SMB server itself is having issues. (In the 3.7.10+ failure,
the server just failed to respond after "Setup andX Request AUTH message", though we
did get the TCP ack so it looks like the message was received by the server.)

I'm attaching a capture taken on the SMB server.

From looking at this page:

http://msdn.microsoft.com/en-us/library/ff469913.aspx

It appears the problem (STATUS_UNSUCCESSFUL) is:

"The size of the extended attribute list is not correct. Check the EaErrorOffset field for the
address of the SMB_GEA structure at which the error was detected."

I did not see anything about extended attribute list in the capture, but if someone else
with more knowledge could take a look and see if they notice any problems I would be grateful.

Thanks!
Ben

--
Ben Greear <greearb-my8/[email protected]> Candela Technologies Inc http://www.candelatech.com

#17 lore raw
In reply to a message not in the archive

Can you try this patch and see if it works?

fs/cifs/sess.c +2 -3
index e87387d..ac14d71 100644
@@ -450,9 +450,8 @@ int build_ntlmssp_auth_blob(unsigned char *pbuffer,
        sec_blob->WorkstationName.MaximumLength = 0;
        tmp += 2;
 
-       if (((ses->ntlmssp->server_flags & NTLMSSP_NEGOTIATE_KEY_XCH) ||
-               (ses->ntlmssp->server_flags & NTLMSSP_NEGOTIATE_EXTENDED_SEC))
-                       && !calc_seckey(ses)) {
+       if ((ses->ntlmssp->server_flags & NTLMSSP_NEGOTIATE_KEY_XCH) &&
+                       !calc_seckey(ses)) {
                memcpy(tmp, ses->ntlmssp->ciphertext, CIFS_CPHTXT_SIZE);
                sec_blob->SessionKey.BufferOffset = cpu_to_le32(tmp - pbuffer);
                sec_blob->SessionKey.Length = cpu_to_le16(CIFS_CPHTXT_SIZE);
Quoted 43 lines · Ben Greear

Hello!

A customer reported problems connecting our CIFS traffic generation test gear to their SMB server.
We are using the 3.9.11+ kernel, and though it is patched, we do not have any
patches to cifs.

The OS is Fedora 14, 64-bit.

A similar system on Fedora 14, with a 3.7.10+ kernel was working, but when we tried the 3.7.10+ kernel on the broken machine, it also failed to work. So, it could be the SMB server itself is having issues. (In the 3.7.10+ failure,
the server just failed to respond after "Setup andX Request AUTH message", though we
did get the TCP ack so it looks like the message was received by the server.)

I'm attaching a capture taken on the SMB server.

From looking at this page:

http://msdn.microsoft.com/en-us/library/ff469913.aspx

It appears the problem (STATUS_UNSUCCESSFUL) is:

"The size of the extended attribute list is not correct. Check the EaErrorOffset field for the
address of the SMB_GEA structure at which the error was detected."

I did not see anything about extended attribute list in the capture, but if someone else
with more knowledge could take a look and see if they notice any problems I would be grateful.

Thanks!
Ben

--
Ben Greear <greearb-my8/[email protected]> Candela Technologies Inc http://www.candelatech.com

author
#19 lore raw
In reply to a message not in the archive
Quoted 1 line · Shirish Pargaonkar

Can you try this patch and see if it works?

We have verified that this works in our test systems. We are just passing user=foo,password=bar in our mount options, so hopefully that is sufficient to exercise this code?

Thanks,
Ben

Quoted 16 lines
fs/cifs/sess.c +2 -3
index e87387d..ac14d71 100644
@@ -450,9 +450,8 @@ int build_ntlmssp_auth_blob(unsigned char *pbuffer,
        sec_blob->WorkstationName.MaximumLength = 0;
        tmp += 2;
 
-       if (((ses->ntlmssp->server_flags & NTLMSSP_NEGOTIATE_KEY_XCH) ||
-               (ses->ntlmssp->server_flags & NTLMSSP_NEGOTIATE_EXTENDED_SEC))
-                       && !calc_seckey(ses)) {
+       if ((ses->ntlmssp->server_flags & NTLMSSP_NEGOTIATE_KEY_XCH) &&
+                       !calc_seckey(ses)) {
                memcpy(tmp, ses->ntlmssp->ciphertext, CIFS_CPHTXT_SIZE);
                sec_blob->SessionKey.BufferOffset = cpu_to_le32(tmp - pbuffer);
                sec_blob->SessionKey.Length = cpu_to_le16(CIFS_CPHTXT_SIZE);
Signature · 2 lines
Ben Greear <greearb-my8/[email protected]>
Candela Technologies Inc  http://www.candelatech.com
#21 lore raw
In reply to a message not in the archive

username=foo is preferred

Quoted 36 lines · Ben Greear
Quoted 1 line · Shirish Pargaonkar

Can you try this patch and see if it works?

We have verified that this works in our test systems. We are just passing user=foo,password=bar in our mount options, so hopefully that is sufficient to exercise this code?

Thanks,
Ben

Quoted 6 lines
fs/cifs/sess.c +0 -0
index e87387d..ac14d71 100644
@@ -450,9 +450,8 @@ int build_ntlmssp_auth_blob(unsigned char *pbuffer,
        sec_blob->WorkstationName.MaximumLength = 0;

fs/cifs/sess.c int build_ntlmssp_auth_blob(unsigned char *pbuffer, · line 457

Quoted 10 lines
fs/cifs/sess.c +2 -3
        tmp += 2;
 
-       if (((ses->ntlmssp->server_flags & NTLMSSP_NEGOTIATE_KEY_XCH) ||
-               (ses->ntlmssp->server_flags & NTLMSSP_NEGOTIATE_EXTENDED_SEC))
-                       && !calc_seckey(ses)) {
+       if ((ses->ntlmssp->server_flags & NTLMSSP_NEGOTIATE_KEY_XCH) &&
+                       !calc_seckey(ses)) {
                memcpy(tmp, ses->ntlmssp->ciphertext, CIFS_CPHTXT_SIZE);
                sec_blob->SessionKey.BufferOffset = cpu_to_le32(tmp - pbuffer);
                sec_blob->SessionKey.Length = cpu_to_le16(CIFS_CPHTXT_SIZE);

--
Ben Greear <greearb-my8/[email protected]> Candela Technologies Inc http://www.candelatech.com

--
To unsubscribe from this list: send the line "unsubscribe linux-cifs" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

Signature · 3 lines
Thanks,

Steve
#22 lore raw
In reply to a message not in the archive

Thanks. Will post a patch on the mailing list.

Regards,

Shirish

Quoted 32 lines · Ben Greear
Quoted 1 line · Shirish Pargaonkar

Can you try this patch and see if it works?

We have verified that this works in our test systems. We are just passing user=foo,password=bar in our mount options, so hopefully that is sufficient to exercise this code?

Thanks,
Ben

Quoted 6 lines
fs/cifs/sess.c +0 -0
index e87387d..ac14d71 100644
@@ -450,9 +450,8 @@ int build_ntlmssp_auth_blob(unsigned char *pbuffer,
        sec_blob->WorkstationName.MaximumLength = 0;

fs/cifs/sess.c int build_ntlmssp_auth_blob(unsigned char *pbuffer, · line 457

Quoted 10 lines
fs/cifs/sess.c +2 -3
        tmp += 2;
 
-       if (((ses->ntlmssp->server_flags & NTLMSSP_NEGOTIATE_KEY_XCH) ||
-               (ses->ntlmssp->server_flags & NTLMSSP_NEGOTIATE_EXTENDED_SEC))
-                       && !calc_seckey(ses)) {
+       if ((ses->ntlmssp->server_flags & NTLMSSP_NEGOTIATE_KEY_XCH) &&
+                       !calc_seckey(ses)) {
                memcpy(tmp, ses->ntlmssp->ciphertext, CIFS_CPHTXT_SIZE);
                sec_blob->SessionKey.BufferOffset = cpu_to_le32(tmp - pbuffer);
                sec_blob->SessionKey.Length = cpu_to_le16(CIFS_CPHTXT_SIZE);

--
Ben Greear <greearb-my8/[email protected]> Candela Technologies Inc http://www.candelatech.com