From mboxrd@z Thu Jan 1 00:00:00 1970 Content-Type: multipart/mixed; boundary="===============4235894412278195975==" MIME-Version: 1.0 From: Denis Kenzior Subject: Re: [PATCH] rilmodem: Fix GPRS feature inavailable issue Date: Thu, 14 Jan 2016 09:52:35 -0600 Message-ID: <5697C443.7090609@gmail.com> In-Reply-To: List-Id: To: ofono@ofono.org --===============4235894412278195975== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Hi Caiwen, Please don't top-post on this list. Thanks for your understanding. Now as to your question: > > In fact I have ever considered the way you mentioned. But l found in most= plugins, atoms (gprs, gprs_context, sms, > call_forwarding ...) are created at post_sim phrase. It seems post_online= phrase is more reasonable. There should > be some reason that the atom creation is brought forward. Does anyone kno= w the reason? > It really depends on what the particular atom supports and what the = hardware supports. The modem driver has some leeway to place atoms in = the appropriate state callback (.pre_sim, .post_sim, .post_online). Generally, anything that accesses the SIM can safely go into .post_sim = if the hardware supports this mode properly, e.g. sim on, rx/tx radio = circuits off. sim and gprs atoms also support being in post_sim. This = is done so that settings can be tweaked while offline. In the case of the rilmodem gprs driver, the querying of max_cids is = needed in order to tell oFono core the id range. This operation needs = to be done once before the atom is registered. If this operation is not = available on the RIL in 'post_sim' (not online) state, then it is safer = to move the gprs atom into post_online callback than trying to follow = radio state changes and (repeatedly) query max_cids. If you really need gprs to be in post_sim, then we can discuss some = alternative solutions. Regards, -Denis --===============4235894412278195975==--