10-modem.fdi

[Dan Williams]

On Sun, 2008-06-01 at 14:52 +0100, Matthew Garrett wrote:
> On Sun, Jun 01, 2008 at 09:46:46AM -0400, Dan Williams wrote:
> > On Sun, 2008-06-01 at 14:28 +0100, Matthew Garrett wrote:
> > > I'm not sure we want to blindly try a stored PIN even once, to be honest 
> > > - if the user plugs and unplugs their phone 3 times, we don't want to 
> > > trigger the PUK locking. Does AT+CGSN require a PIN first?
> > 
> > Remember, you want AT+GCAP though...  that's what lists the standards
> > that the phone supports and allows us to detect plain modems from GSM
> > from CDMA.
> 
> AT+CGSN should give us a unique identifier regardless of whether it's 
> CDMA or GSM. Worst case on normal modems is that it gives an error, but 

It only gives me an error on a Sierra 595U.  Do you mean "AT+GSN"
instead?  That's defined in IS-707-A, but CGSN is not.  I have also
found some references that talk about AT+CGSN for CDMA devices though,
but again I've got an example right here that doesn't support CGSN but
does support GSN.

> we probably aren't worried about a PIN in that case.
> 
> > But yes, that was Marcel's point, that there are some that simply don't
> > respond to anything until AT+CPIN or whatever has happened.
> 
> Presumably they'll at least give us the PIN count? Possibly in that case 
> (where we can't identify the phone uniquely) we should just refuse to 
> store the PIN. More inconvenient, but much safer.

Yeah, pick and choose who you piss off :)

dan