[PATCH v2 5/5] loc: add definitions for more messages

Bjørn Mork bjorn at mork.no
Sun Feb 11 21:23:15 UTC 2018

It might be worded differently - saying that mandatory TLVs are numbered < 0x10. But, yes it's there. So is the endianness. I guess IP addresses are considered structured byte sequences and not numbers....

Look in the top level qmux document. Noone follows it to the letter (probably because few have seen it and lived to tell), but it is still a very useful guide to understand the protocol design. 

Just too bad we can't quote from it or use any of the info in any way that would disclose the contents. 


On February 11, 2018 9:48:41 PM GMT+01:00, Aleksander Morgado <aleksander at aleksander.es> wrote:
>On Sun, Feb 11, 2018 at 9:28 PM, Bjørn Mork <bjorn at mork.no> wrote:
>> Aleksander Morgado <aleksander at aleksander.es> writes:
>>>> I was not aware of the rule that TLV < 0x10 should be mandatory.
>>>> Maybe we could express the default logic in the code generator?
>>> Not that I've seen that rule in written anywhere... I just got that
>>> impression from all the other messages I've encoded myself :)
>>> Maybe we could have a warning issued by the generator if it finds
>>> inconsitencies with that rule, but I wouldn't hardcode the logic to
>>> applied automatically... unless we see it in written somewhere! :)
>> It sure would be nice if Qualcomm published some of their now ancient
>> docs under a less restrictive agreement...  Then we could have
>> both this question and the little endian question.
>> But I guess we just have to pretend we don't know the answers :-)
>> Maybe Qualcomm will apply for a patent on little endian?  Or maybe
>> already did?  Who knows.  I am sure it is important that we don't
>> disclose this company critical information.
>Wait, is there anywhere in the docs where it said anything about TLVs
>< 0x10 being mandatory?  I wasn't really trying to play any game, I
>honestly don't recall having read that.

More information about the libqmi-devel mailing list