K5160

[nietzsche at lysator.liu.se]

Cool - Here you can see my raw dump from usbmon attached.

//M
> nietzsche at lysator.liu.se writes:
>
>> OK, this is the pcap dump of the start and connect sequence on Windows
>> 10
>> with Telenor SE.
>
> Thanks!  Educating stuff.  A quick analysis:
>
> The Windows driver sends a "reset function" request after the "get NTB
> parameters" request.  The Linux driver doesn't do that.  Might make a
> difference.  And it sends another "reset function" request just when
> you'd think the setup was complete, restarting it all.  Don't know why
> it does that.  Makes no sense to me.  But we could try it just for
> completeness.
>
> The Windows driver sets a sane NTB input size as well, just like Linux.
> Unfortunately it seems all request data is missing in this dump?  But
> the response size is 32768, so that must be what was requested. Not too
> far away from our default (16384).  But we could try the exact same
> value to elimininate this unnecessary difference.
>
> The Windows driver requests 32bit mode, even with the reduced buffers.
> We cannot do that without adding a bit of code to the driver.  But it is
> worth trying as a last resort if we cannot get this to work in 16bit
> mode.
>
> The Windows driver issues "set NTB input size" before "set NTB format".
> We should definitely try that.  Trying to set 16bit format while the
> input size is > 2^16 might be considered invalid by the firmware. An
> error response would bave been nice, but experience has taught us that
> we cannot expect that.  Silent failures seems to be the rule here...
>
>
> Bjørn
> _______________________________________________
> libmbim-devel mailing list
> libmbim-devel at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/libmbim-devel
>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: dump3.txt
URL: <https://lists.freedesktop.org/archives/libmbim-devel/attachments/20160301/8cb88b0d/attachment.txt>