ZTE MF683

[Shawn J. Goff]

I previously said we were using little-endian; I was wrong; it's big-endian.

Looks like it's qmi; I did it twice because the first time, it gave me
other stuff before the 01120080... response:

c2890e20 811070979 S Ii:1:005:7 -115:16 64 <
c2890d20 811071305 S Co:1:005:0 s 21 00 0000 0005 000c 12 = 010b0000
00000001 27000000
c2890e20 811073223 C Ii:1:005:7 0:16 8 = a1010000 05000000
c2890da0 811073260 S Ci:1:005:0 s a1 01 0000 0005 0200 512 <
c2890e20 811073300 S Ii:1:005:7 -115:16 64 <
c2890d20 811075108 C Co:1:005:0 0 12 >
c2890e20 811075772 C Ii:1:005:7 -2:16 0
c2890da0 811075864 C Ci:1:005:0 0 12 = 010b0080 00000200 27000000
c2890e20 818498491 S Ii:1:005:7 -115:16 64 <
c2890e20 818499214 C Ii:1:005:7 0:16 8 = a1010000 05000000
c2890da0 818499251 S Ci:1:005:0 s a1 01 0000 0005 0200 512 <
c2890e20 818499296 S Ii:1:005:7 -115:16 64 <
c2890da0 818499826 C Ci:1:005:0 0 12 = 010b0080 00000200 27000000
c2890e20 818501250 C Ii:1:005:7 0:16 8 = a1010000 05000000
c2890da0 818501277 S Ci:1:005:0 s a1 01 0000 0005 0200 512 <
c2890e20 818501319 S Ii:1:005:7 -115:16 64 <
c2890da0 818501827 C Ci:1:005:0 0 12 = 010b0080 00000200 27000000
c2890e20 818503222 C Ii:1:005:7 0:16 8 = a1010000 05000000
c2890da0 818503249 S Ci:1:005:0 s a1 01 0000 0005 0200 512 <
c2890e20 818503291 S Ii:1:005:7 -115:16 64 <
c2890da0 818503831 C Ci:1:005:0 0 19 = 01120080 00000101 27000700
02040000 000000
c2890e20 867559891 C Ii:1:005:7 -2:16 0
c2890e20 869704434 S Ii:1:005:7 -115:16 64 <
c2890d20 869704776 S Co:1:005:0 s 21 00 0000 0005 000c 12 = 010b0000
00000001 27000000
c2890d20 869705842 C Co:1:005:0 0 12 >
c2890e20 869706589 C Ii:1:005:7 -2:16 0
c2890e20 872003785 S Ii:1:005:7 -115:16 64 <
c2890e20 872005214 C Ii:1:005:7 0:16 8 = a1010000 05000000
c2890da0 872005252 S Ci:1:005:0 s a1 01 0000 0005 0200 512 <
c2890e20 872005297 S Ii:1:005:7 -115:16 64 <
c2890da0 872005829 C Ci:1:005:0 0 19 = 01120080 00000101 27000700
02040000 000000
c2890e20 894220746 C Ii:1:005:7 -2:16 0



And here is the assertion failure:

# qmicli -d /dev/cdc-wdm1 --nas-get-signal-strength
c2890e20 1071749128 S Ii:1:005:7 -115:16 64 <
c2890d20 1071814844 S Co:1:005:0 s 21 00 0000 0005 0010 16 = 010f0000
00000001 22000400 01010003
c2890d20 1071817356 C Co:1:005:0 0 16 >
c2890e20 1071821236 C Ii:1:005:7 0:16 8 = a1010000 05000000
c2890da0 1071821273 S Ci:1:005:0 s a1 01 0000 0005 0200 512 <
c2890e20 1071821315 S Ii:1:005:7 -115:16 64 <
c2890da0 1071821857 C Ci:1:005:0 0 24 = 01170080 00000101 22000c00
02040000 00000001 02000301
c2890d20 1071838300 S Co:1:005:0 s 21 00 0000 0005 0012 18 = 01110000
03010001 00200005 00100200 ef00
c2890d20 1071840855 C Co:1:005:0 0 18 >
c2890e20 1071843237 C Ii:1:005:7 0:16 8 = a1010000 05000000
c2890da0 1071843273 S Ci:1:005:0 s a1 01 0000 0005 0200 512 <
c2890e20 1071843315 S Ii:1:005:7 -115:16 64 <
c2890da0 1071844090 C Ci:1:005:0 0 39 = 01260080 03010201 0020001a
00020400 00000000 0102009a 05110400 01006605
**
ERROR:qmi-utils.c:72:qmi_utils_read_guint8_from_buffer: assertion
failed: (*buffer_size >= 1)
Aborted
# c2890e20 1071967308 C Ii:1:005:7 -2:16 0


On Tue, Sep 18, 2012 at 6:20 AM, Bjørn Mork <bjorn at mork.no> wrote:
> Marius Kotsbak <marius.kotsbak at gmail.com> writes:
>> 2012/9/18 Bjørn Mork <bjorn at mork.no>
>>
>>> BTW, is this a big endian platform (sorry for my complete ignorance wrt
>>> Atmel chips...)?
>>>
>>>
>> I think the Atmel AVR32 is the opposite of x86, but isn't the driver code
>> endian safe?
>
> It is supposed to be, yes.  And given that it doesn't touch the contents
> of the embedded QMI, I am pretty sure it mostly is too.  Would be nice
> to have that confirmed.
>
> But how about the libqmi tools?  Do they work on big endian systems?
>
> Of course, you can always do the manual command line, tool-less and
> endian safe, QMI_CTL SYNC test:
>
>    bjorn at nemi:~$ echo -ne "\x01\x0b\x00\x00\x00\x00\x00\x01\x27\x00\x00\x00" >/dev/cdc-wdm0
>
>    bjorn at nemi:~$ hexdump -C /dev/cdc-wdm0
>    00000000  01 12 00 80 00 00 01 01  27 00 07 00 02 04 00 00  |........'.......|
>
>
> If you get that reply, then there is a QMI device there.  If you don't
> then there is not.
>
>
> Bjørn