From mboxrd@z Thu Jan 1 00:00:00 1970 Content-Type: multipart/mixed; boundary="===============3828094613022748571==" MIME-Version: 1.0 From: Lei Yu Subject: Re: [PATCH] doc: Add modem Type property Date: Fri, 21 Jan 2011 14:19:42 -0800 Message-ID: <4D3A067E.1070501@nokia.com> In-Reply-To: <4D3A02E9.30807@gmail.com> List-Id: To: ofono@ofono.org --===============3828094613022748571== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Hi Denis, On 01/21/2011 02:04 PM, ext Denis Kenzior wrote: > Hi Lei, > > On 01/21/2011 03:48 PM, Lei Yu wrote: >> Hi Denis, >> >> On 01/21/2011 01:04 PM, ext Denis Kenzior wrote: >>> Hi Lei, >>> >>>>>> + string Type [readonly, optional, experimental] >>>>>> + >>>>>> + String representing the type of the modem device. >>>>>> + >>>>>> + The possible values are: >>>>>> + "gsm" Modem supports one or more of the 3GPP >>>>>> + family of technologies including GSM, >>>>>> + EDGE, UMTS, HSPA, LTE. >>>>>> + "cdma" Modem supports one or more of the 3GPP2 >>>>>> + family of technologies including CDMA 1x, >>>>>> + EVDO. >>>>>> + "hfp" Bluetooth hands-free profile (HFP) modem >>>>>> + >>>>> >>>>> So one pointed question: >>>>> >>>>> How are we planning to support dual-mode devices with this property? = I >>>>> believe during the CDMA workshop we decided that the clients should >>>>> simply look at the set of interfaces, not a semi-static property... >>>> >>>> We understand that the decision during the CDMA workshop but as we >>>> explained one use case in the cover letter >>>> (http://lists.ofono.org/pipermail/ofono/2011-January/007890.html) for >>>> this RFC, there will be cases where atom's D-Bus I/F is not suitable (= or >>>> as a reliable way) for telling what is current mode (CDMA or GSM). E.g= ., >>>> in the case of retrieving serial number from modem. Currently both CDMA >>>> and GSM shares the same I/F (ofono.org.modem) rather than having >>>> separate (ofono.org.cdma.modem). And, further if no other atom has been >>>> instantiated, then, there will be no other way to tell which mode (CDMA >>>> or GSM) the modem is. And, further for displaying serial number, upper >>>> layer will need to know it is CDMA or GSM. In case of CDMA, the serial >>>> number is MEID and GSM it is IMEI. As explained in cover letter, we >>>> anticipate more cases like this. >>> >>> So let us say I have a dual-mode modem. Are there now two serial >>> numbers? One for MEID and one for IMEI? Are there now two >>> Manufacturers? Two Revisions? >> >> Try to clarify what you try to ask first. Let's start with the dual-mode >> case where the modem is integrated modem (one piece of HW) where it >> supports both CDMA and GSM. Then, there is case where the phone actually >> has both MEID and IMEI. And, these two numbers are actually the same >> (note, per IMEI's assignment there is a pre-arranged segment for this >> kind of world phone). And, in which case, one manufacture, one revision. >> Per the other thread, in which case the "Type" will be something like >> "gsm_cdma". > > So how does the modem type help you here if you have 'gsm_cdma'? Is the > modem giving you a MEID? An IMEI? I'm not convinced the Type helps you > in any way. It might help you for HFP, but certainly not the dual mode > case. > First, our original use case is really for a single mode case, not dual = mode. Please see our cover letter in our original RFC. Please do let us = know if you agree or disagree whether there is a problem needing to be = solved for single mode case. Second, in case of the example I gave above, modem will provide one = serial number which is both MEID and IMEI since they are identical. In = which case based on "Type" setting to "cdma_gsm", it will be upper = layer's choice to do the processing. E.g., upper layer can do something = like: MEID/IMEI: xxxxxxx. >> Of course, there can be case where two physical modems are in the device >> in which case there will be two separate serial numbers and two >> manufacturers and revisions. I am assuming you are not asking for this >> case. > > Yes, I'm ignoring the true dual-mode dual-radio case right now. > > Regards, > -Denis Regards, -Lei --===============3828094613022748571==--