Testing 1394 cameras with WLK 1.2

Has anyone ever done any DTM testing on 1394 cameras (IIDC)? We are trying to do that and get some errors which definitely look silly. We are talking about a 1394a camera and of course using a 1394a host adapter on Vista x86.

To start with we get the usual file signature verification errors, that have been reported in other posts, like:
File 1394HCT.exe did not pass Signature Verification
File 1394crom.dll did not pass Signature Verification

All these for Microsoft provided executables.

The real problem though is the so called 1394 Configuration ROM test. This test checks the config rom of all devices on the bus (currently only the camera and the PC with the Microsoft 1394 drivers).
We get the following for the camera (actually ANY iidc camera):
Assertion 19.18.2 Failed: Root Directory must contain a valid Vendor ID Textual Descriptor
Assertion 19.18.2 Failed: Unit Directory must contain a valid Model ID Textual Descriptor

I don’t really know who came up with the first assertion, but it is obviously a misinterpretation of IEEE1212 paragraph 7.6.1 which states that only the Vendor ID value is mandatory in the root directory, NOT its textual descriptor. ALL 1394 cameras we could get our hands on seem to have the textual descriptor in the Unit Dependent Directory.

The second assertion is also wrong, since IEEE1212 paragraph 7.6.3 does NOT list the Model ID or its textual descriptor as mandatory.

And then we move to the silly part… The test examines the configuration ROM of the PC itself. As you might not know, 1394 host adapters don’t have any configuration ROM data stored in an EEPROM or something, it is up to the driver to publish anything it wants as the configuration ROM. So the Microsoft 1394 driver publishes something that the Microsoft DTM finds incorrect with the following error:

Assertion 19.18.1 Failed: One or more Unit directories must be located in the Configuration ROM. Please examine IEEE 1212 specification for guidance.

Unlike the situation with the signature failures where the test is a pass, the 1394 configuration test is NOT a pass. It is a FAIL…

The camera configuration ROMs are abdolutely legal and obviously cannot change for the sake of buggy DTM (and of course we are talking about many different camera models).
On the other hand, the PC configuration ROM is under the control of Microsoft, so we are at a dead end here.

I really wonder how on earth did they test those tests in the first place?
Does anyone have ANY idea what we could possibly do other than trying to talk to the WinQual/WLK guys?

Warm Regards
Dimitris Staikos
Unibrain

> We get the following for the camera (actually ANY iidc camera):

Assertion 19.18.2 Failed: Root Directory must contain a valid Vendor ID
Textual
Descriptor
Assertion 19.18.2 Failed: Unit Directory must contain a valid Model ID
Textual
Descriptor

I don’t really know who came up with the first assertion, but it is obviously
a
misinterpretation of IEEE1212 paragraph 7.6.1 which states that only the
Vendor
ID value is mandatory in the root directory, NOT its textual descriptor.

Exactly so, when I was busy with 1394 in around 1999-2000, I saw devices with
no textual leaves for model/vendor ID, only the codes.


Maxim Shatskih, Windows DDK MVP
StorageCraft Corporation
xxxxx@storagecraft.com
http://www.storagecraft.com