Pawel, please add your analysis to the bug.

Its not clear to me if we are all talking about same things. Is Tyler running 
the C sample or C++ sample?

But otherwise, I agree with the idea that if the issue is limited to sample 
only, we should address it with a relnote.

Thanks,
Arvind

From: Josh Spain [mailto:[email protected]]
Sent: Tuesday, November 8, 2016 4:02 PM
To: Lioy, Marcello <[email protected]>
Cc: Pawel Winogrodzki <[email protected]>; Kevin Kane 
<[email protected]>; Arvind Padole <[email protected]>; Carrie 
Hertel <[email protected]>; [email protected]
Subject: Re: Release note ASACORE-3464

Yes, if it is only the sample it is not blocking. I would call it critical 
though since it would be much better if we didn't have samples with incorrect 
code.

-Josh

[https://s3.amazonaws.com/ucwebapp.wisestamp.com/12898986-6b24-4785-b392-dfd3cd5cdd09/Affinegylogo201450px.format_png.resize_200x.png#logo]

Josh Spain
Director of Engineering Affinegy, Inc.
t.512-535-1700 x1006<tel:512-535-1700%20x1006>
a. 1705 S Capital of Texas Hwy, Suite 310, Austin, TX 78746 USA
Website<http://affinegy.com> Email<mailto:[email protected]>
[https://s3.amazonaws.com/images.wisestamp.com/icons_32/linkedin.png]<http://www.linkedin.com/in/joshspain>
 [https://s3.amazonaws.com/images.wisestamp.com/icons_32/twitter.png] 
<http://twitter.com/affinegy>

[https://s3.amazonaws.com/images.wisestamp.com/icons/twitter.png]Latest 
Tweet:<https://twitter.com/Affinegy> It's not all about 
#TheCloud.<https://twitter.com/TheCloud.> Intelligence at 
#TheEdge<https://twitter.com/TheEdge> and in 
#TheFog<https://twitter.com/TheFog> matter. Distributed computing will continue 
a… https://t.co/6gwm8lsyIK

Read More<https://twitter.com/Affinegy/statuses/794704825367887874>

[https://s3.amazonaws.com/images.wisestamp.com/email-apps/twitter_button/twitter-white.png]<https://twitter.com/Affinegy>

On Tue, Nov 8, 2016 at 5:49 PM, Lioy, Marcello 
<[email protected]<mailto:[email protected]>> wrote:
So it sounds like the issue is with the C door sample?  If so that doesn’t seem 
like it is release blocking, but worth calling out the bug in a release note.  
Pawel, could you add your analysis to the issue indicating what the problem is 
so that the release notes give a sense of what should actually be happening.

Josh, Arvind, do you agree it’s not blocking but we do want to add to the 
release notes?

From: Pawel Winogrodzki 
[mailto:[email protected]<mailto:[email protected]>]
Sent: Tuesday, November 08, 2016 2:59 PM
To: Kevin Kane <[email protected]<mailto:[email protected]>>; Arvind Padole 
<[email protected]<mailto:[email protected]>>; Lioy, 
Marcello <[email protected]<mailto:[email protected]>>; Josh Spain 
<[email protected]<mailto:[email protected]>>; Carrie Hertel 
<[email protected]<mailto:[email protected]>>

Cc: 
'[email protected]<mailto:[email protected]>'
 
<[email protected]<mailto:[email protected]>>
Subject: RE: Release note ASACORE-3464

I’ve looked into the Security Manager’s code a bit and it seems to be doing the 
right thing. The C samples on the other hand (so my code, unfortunately) seem 
to have a bug – they wait for their status to change to “CLAIMED” and proceed 
with modifying the accepted authentication methods. They do this in two steps:

QStatus status = alljoyn_busattachment_enablepeersecurity(doorData->bus, "", 
nullptr, nullptr, QCC_TRUE);
    if (ER_OK != status) {
        fprintf(stderr, "Failed to clear peer security - status (%s)\n", 
QCC_StatusText(status));
        return status;
    }

    status = alljoyn_busattachment_enablepeersecurity(doorData->bus, 
KEYX_ECDHE_DSA, doorData->authListener, nullptr, QCC_FALSE);
    if (ER_OK != status) {
        fprintf(stderr, "Failed to set peer security for claimed mode - status 
(%s)\n", QCC_StatusText(status));
        return status;
    }

In the meantime the Security Manager will try to establish another connection 
and call “EndManagement()”, so it might happen, that the connection is 
established before or in between these two API calls.

I’ll try to change the samples locally and verify this fixed the issue. The C++ 
samples should be fine, because they are waiting for an event set in the 
“EndManagement” callback.

From: Kevin Kane
Sent: Tuesday, November 8, 2016 14:29
To: Arvind Padole 
<[email protected]<mailto:[email protected]>>; Lioy, 
Marcello <[email protected]<mailto:[email protected]>>; Josh Spain 
<[email protected]<mailto:[email protected]>>; Carrie Hertel 
<[email protected]<mailto:[email protected]>>; Pawel 
Winogrodzki <[email protected]<mailto:[email protected]>>
Cc: 
'[email protected]<mailto:[email protected]>'
 
<[email protected]<mailto:[email protected]>>
Subject: RE: Release note ASACORE-3464

This looks like a real bug. The “Old application state/New application state” 
output is from the security manager. I’m not as familiar with this code but I’m 
guessing it’s changing its internal database that tracks the work it has to do, 
but when it attempts to actually claim the door provider, it fails to do so. 
0x9032 is ER_BUS_REPLY_IS_ERROR_MESSAGE.

SecurityApplicationProxy.cc:1035 is calling EndManagement, which suggests the 
Claim call itself was successful, and EndManagement for some reason is failing. 
Pawel, I seem to recall you changed Claim such that the app was in the 
managementStarted state, so this shouldn’t be failing. Any ideas?

From: Arvind Padole
Sent: Tuesday, November 8, 2016 12:54 PM
To: Lioy, Marcello <[email protected]<mailto:[email protected]>>; 
Josh Spain <[email protected]<mailto:[email protected]>>; Carrie Hertel 
<[email protected]<mailto:[email protected]>>; Kevin 
Kane <[email protected]<mailto:[email protected]>>; Pawel Winogrodzki 
<[email protected]<mailto:[email protected]>>
Cc: 
'[email protected]<mailto:[email protected]>'
 
<[email protected]<mailto:[email protected]>>
Subject: RE: Release note ASACORE-3464

We should probably investigate and understand what’s going on here before we 
can decide if it’s a blocking bug that must be fixed or just relnoted. From the 
output, it looks like a minor issue in the secmgr sample in that the provider 
is claimed but somehow the error is displayed.

Adding Pawel as well. Kevin/Pawel – can you take a look to see if my theory 
above is correct or if there is a different issue here?


Please enter the password provided by the application:
A02D04
Claiming application ...
110.300 ****** ERROR ALLJOYN_OBJ lepDisp2_0 .../router/AllJoynObj.cc:4466 | 
0x9029
110.300 ****** ERROR ALLJOYN_OBJ lepDisp2_0 .../router/AllJoynObj.cc:4466 | 
0x9029
110.300 ****** ERROR ALLJOYN_OBJ lepDisp2_0 .../router/AllJoynObj.cc:4466 | 
0x9029
>> Old application state : OnlineApplication: Busname: :1E7cCe4y.40, Claim 
>> state: CLAIMABLE, Sync state: SYNC_UNMANAGED
>> New application state : OnlineApplication: Busname: :1E7cCe4y.40, Claim 
>> state: CLAIMED, Sync state: SYNC_UNMANAGED
>> Application id : 9d3d951a5784efa7f84a2eafeb963f5d
>> Application name : DoorProvider (CASS-IOT-WIN10)

135.375 ****** ERROR ALLJOYN_SECURITY external ...tyApplicationProxy.cc:1035 | 
0x9032
135.375 ****** ERROR SECMGR_AGENT external .../ProxyObjectManager.cc:340 | 
0x9032
135.375 ****** ERROR SECMGR_AGENT external ...c/SecurityAgentImpl.cc:516 | 
0x9032
Failed to claim application ...


From: Lioy, Marcello [mailto:[email protected]]
Sent: Tuesday, November 8, 2016 10:49 AM
To: Arvind Padole 
<[email protected]<mailto:[email protected]>>; Josh 
Spain <[email protected]<mailto:[email protected]>>; Carrie Hertel 
<[email protected]<mailto:[email protected]>>; Kevin 
Kane <[email protected]<mailto:[email protected]>>
Cc: 
'[email protected]<mailto:[email protected]>'
 
<[email protected]<mailto:[email protected]>>
Subject: Release note ASACORE-3464
Importance: High

Do people thing that we should include 
ASACORE-3464<https://jira.allseenalliance.org/browse/ASACORE-3464> (Security 
2.0 Door Provider on windows 10 must be claimed 
first<https://jira.allseenalliance.org/browse/ASACORE-3464>) in the release 
notes? Or is that expected behavior?


_______________________________________________
Allseen-core mailing list
[email protected]
https://lists.allseenalliance.org/mailman/listinfo/allseen-core

Reply via email to