Re: [PATC] Fix to prevent AT-command interference

Hi Matti,

looks reasonable to me, commited to cvs.

Thanks,
Alex

Matti Ärmänen schrieb:
> Hello,
> If a memory check is performed right after AT+CNMA-command, an error
> message is generated and modem connection is reopened.
> 
> Here is an example:
> 2008-09-28 17:57:59 [7962] [16] DEBUG: AT2[12]: <-- +CMT: ,20
> 2008-09-28 17:57:59 [7962] [16] DEBUG: AT2[12]: <-- ****DELETED****
> 2008-09-28 17:57:59 [7962] [16] DEBUG: AT2[12]: received message from
> SMSC: ****DELETED****
> 2008-09-28 17:57:59 [7962] [16] DEBUG: AT2[12]: Numeric sender
> (international) <****DELETED****>
> 2008-09-28 17:57:59 [7962] [16] DEBUG: AT2[12]: User data length read as 
> (1)
> 2008-09-28 17:57:59 [7962] [16] DEBUG: AT2[12]: Udh decoding done len=1
> udhi=0 udhlen=0 udh=''
> 2008-09-28 17:57:59 [7962] [16] DEBUG: AT2[12]: --> AT+CNMA^M
> 2008-09-28 17:58:00 [7962] [16] DEBUG: AT2[12]: --> AT+CPMS?^M
> 2008-09-28 17:58:00 [7962] [16] DEBUG: AT2[12]: <-- OK
> 2008-09-28 17:58:00 [7962] [16] DEBUG: AT2[12]: no correct header for
> CPMS response.
> 2008-09-28 17:58:00 [7962] [16] DEBUG: AT2[12]: memory check error
> 2008-09-28 17:58:00 [7962] [16] INFO: AT2[12]: Closing device
> 2008-09-28 17:58:00 [7962] [16] ERROR: AT2[12]: Couldn't connect
> (retrying in 10 seconds).
> 
> This can be fixed, if after issuing the AT+CNMA-command, the OK-answer
> is waited before continuing. As the OK-answer is waited, the logic with
> "expect_extra_ok"-counter is no longer necessary.
> 
> We have run this in 3 kannel instances for 2 months without problems.
> 
> Regards,
> Matti Ärmänen
> 
>