Hi Denis, On Thu, Dec 13, 2018 at 6:31 PM Denis Kenzior <[email protected]> wrote: > > Hi Giacinto, > > > on the other hand, the sessions will be lost anyway, downloads and > > streaming interrupted, due to the change of mobile IP, even if > > the reboot would be fast. > > Hey, I'm not fully convinced this is a good idea either ;) > > But then to play devil's advocate: > - TCP can already deal with network interruptions
not if the IP changes. > - Sane networks should issue the same IP not at all. They are handled in DHCP, and with million users, most likely it will be different the next time. There are recommendations not to reuse the same IP on some networks, and to reassign the same on others. Most of the networks do not guarantee it, unless there is a special contract for fix IP (for servers over mobile network). > - This has already been done before with good success on CS networks mainly. On PS the situation evolves every day, and old solutions might not work even today. > > Whether this is a good idea to do on the application processor is > another question. > > > > >> > >> So if we do proceed with this feature, it should try pretty hard to > >> comply with the spirit of the cited section, even if it isn't complying > >> with it in a literal sense. > > > > I don't agree with this. > > In the original intents, the PIN is shared through DBus, > > with quite some risk to disclose it to still other applications. > > Please stop going off on tangents. There is no more risk of exposing > the PIN through D-Bus with a properly setup D-Bus security policy than > any other IPC mechanism you can think of. In fact it is probably safer > than any IPC you can roll on your own. If you want to discuss D-Bus > security issues, take it to the D-Bus mailing list. > > > > > And then, if you want to restore the status as before the reboot, > > maybe there were other PINs or keys presented. > > Should they be stored and managed too? > > Maybe? That is a question for the modem guys. The SIM, apart from the authentication and USIM files, deals with user sensitive contents that are completely transparent to the modem and to ofono, and only tunneled end-to-end. > > > > > And in case of USB devices, they disappear and re-enum later. > > Should we think about storing the PIN non-volatile > > for supporting the feature for these devices??? > > > > And should we have a timeout for the reboot? > > > > timeouts are generally a bad idea since you can't control how long > something takes. But yes, maybe the cache should also have an > expiration time set when the device is removed as well as the IMEI of > the device such that only one IMEI/ICCID/PIN combination can exist. > > Regards, > -Denis Regards, Giacinto _______________________________________________ ofono mailing list [email protected] https://lists.ofono.org/mailman/listinfo/ofono
