[patch] sms ack

Jim Rayner openmoko at jrayner.com
Tue Oct 30 17:15:29 CET 2007

Hi Erin

As I said is up to you if you want to use the patch

In a perfect world I agree CSMS=0 will work fine

The difficulty with CSMS=0 is that the baseband manages the messages,
the baseband only produces the +CMT once, therefore you need to
introduce a range of measures to check the message store to ensure
that you haven't missed a message, if for any reason the application
processor is unable to process it, during restart of gsmd or similar

Best regards,

On 10/30/07, Erin Yueh <erin_yueh at openmoko.com> wrote:
> Hi Jim,
> Based on 3GPP TS 27.005  Section 3.4.1, you have to send ACK when
> when +CSMS <service> value equals 1. If you would just like to store
> all incoming SMS, you may just send "AT+CNMI=2,2,2,1,0" command and
> still set "AT+CSMS=0". Then you don't need to care about "AT+CNMA=0".
> The service center won't resend it again, and you can get the status
> report too. It will be easier and no need ACK.
> Cheers,
> Erin
> Jim Rayner wrote:
> > Hi
> >
> > Attached patch to enable applications to govern the acking of sms's
> >
> > If you are interested in storing all messages in your own messaging
> > store, removing your dependency on the size of the SIM then you may
> > wish to use this patch.
> >
> > Acking informs the message service centre that the message has been
> > received. If the message is nak'd then the service centre will try a
> > resend some period later (unless your stack is capable of sending a
> > resume sending command - unsure if this exists in practice).
> >
> > Mileage may vary due to differences in implementation of the CNMA
> > command between different stacks (for example the Siemens TC65 appears
> > to have a poor implementation, allowing only CNMA=0, to nak a message
> > you have to allow the protocol stack timer to expire)
> >
> > Patch generated in comparison to version on openmoko svn 26th October
> >
> > Best regards,
> > Jim

More information about the gsmd-devel mailing list