Fwd: Cinterion plugin patch

[Andrew Bird]

On Fri, 25 Nov 2016 21:27:57 +0100
Reinhard Speyerer <rspmn at arcor.de> wrote:

> On Fri, Nov 25, 2016 at 09:55:41AM +0000, Andrew Bird wrote:
> > On Fri, 25 Nov 2016 10:13:08 +0100
> > Aleksander Morgado <aleksander at aleksander.es> wrote:
> > 
> > > Hey Reinhard & Matthew,
> > 
> > <snip>
> > 
> > > > One further difference I noticed is that USSD responses for AT+CSCS="GSM"
> > > > contain a non-standard \xx escaping for non-printable characters (unless
> > > > I missed some recent TS 27.007 changes).
> > > >
> > > > PLS8-E REVISION 01.090:
> > > > +CUSD: 1,"Hauptmen~:
> > > > 1 Guthabenkonto
> > > > 2 Guthaben Verf~gbarkeit
> > > > 3 Gutschein einl|sen
> > > > 4 Pack Manager
> > > > 7 Tarifinfo
> > > > 8 Hilfe",15
> > > >
> > > > PLS8-E REVISION 02.011:
> > > > +CUSD: 1,"Hauptmen~:\0A1 Guthabenkonto\0A2 Guthaben Verf~gbarkeit\0A3 Gutschein einl|sen\0A4 Pack Manager\0A7 Tarifinfo\0A8 Hilfe",15
> > > 
> > > Any idea if this has any impact on how ModemManager handles USSD in
> > > this device? Or is this only an issue to the consumer of the USSD
> > > responses?
> > 
> > What I think curious is that 0x0a is a valid GSM0338 character and should have been interpreted as linefeed
> > 
> > ftp://ftp.unicode.org/Public/MAPPINGS/ETSI/GSM0338.TXT
> > 
> > Were both modem sessions captured on the same system, as I'm just wondering if what you are seeing is an artifact of the terminal?
> 
> Hi Andrew,
> both modem sessions were captured on the same system with only the mPCIe
> modules being exchanged.
> 
> I also checked with strace that the \xx escapes are not artefacts of the
> terminal emulation used (<ESC>e = Euro sign):
> 
> PLS8-E REVISION 01.090:
> 
> read(10, "\r\n+CUSD: 0,\"Ihr Guthaben betr{gt: 42,00\33e. Jetzt auch Ihr Guthaben aufladen: einfach *103*Aufladenummer# und die H|rertaste eingeben.\",15\r\n", 4096) = 139
> 
> PLS8-E REVISION 03.017:
> read(10, "\r\n+CUSD: 2,\"Ihr Guthaben betr{gt: 42,00\\1Be. Jetzt auch Ihr Guthaben aufladen: einfach *103*Aufladenummer# und die H|rertaste eingeben.\",15\r\n", 4096) = 141
> 
> Regards,
> Reinhard

Hi Reinhard,
   On your strace capture I don't see the linefeed difference, but it's a different msg of course. Bear in mind that strace is not showing raw binary, but its interpretation of it (see \r\n etc). So assuming strace is displaying binary as octal or known escapes, I agree the v1 output \33e should generate 0x1b, 0x65 i.e. a Euro symbol in GSM escape table. Whereas the v3 '\\1Be' is really odd, a backslash followed by 1Be. The read return value confirms two more characters were read, so I think my interpretation is correct. Looks like a firmware issue to me.

Best regards,


-- 
Andrew Bird <ajb at spheresystems.co.uk>