On Tue, Nov 30, 2010 at 8:22 PM, Denis Kenzior <denkenz(a)gmail.com> wrote:
> While debugging a connection issue using huawei EM770W modem, I
> a DBG code to print out the status value received in at_cgreg_cb() in
> drivers/atmodem/gprs.c, I found the status got from the
> at_util_parse_reg() is incorrect sometimes, because ofono does not set
> vendor ID in the gprs_data, and it looks to me the vendor code is
> required so that the at_util_parse_reg() will read unquoted strings in
> the solicited events for huawei modem. For example, when we got
> "+CGREG: 1, 2833, 1231B60" followed by "+CGREG: 2,1, 2733, 1B60"
> unsolicited codes, the status parsed from this code is 2833 (i.e. the
> 1st code), instead of the expected 1 in the 2nd code.
I suggest you complain to the vendor that they do not follow standards,
and ask them to fix their firmware.
Any possibility that this is a race condition? Just while ofono
sending AT+CGREG? to poll the CGREG state, modem already sent out
unsolicited CGREG? I saw at_util_parse_reg() also tries to skip
unsolicited CREG/CGREG, so it looks like this is valid concern?
> It looks like all modem plugins invoke
> with vendor=0, is this on purpose? then how could at_util_parse_reg()
> parses strings correctly without knowing vendor code?
This can be added to the huawei driver, however this won't help you
completely as the firmware still reports bizarre values. Care to send a
Sure I will patch and send out the huawei plugins patch, but since I
am not familiar with the overall ofono code, so I'd like to verify
what my understanding so far -- modem plugins should send the vendor
ID while invoke ofono_gprs_context_create()? Would you please