summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* improve LTE datacall setupreplicant-6.0Joey Hewitt2021-01-041-1/+2
| | | | | | It just started reporting LTE after not doing so the last few days... I'm not sure if I did something to fix it or if the phone/network just decided to.
* add NITZ time sync supportJoey Hewitt2021-01-045-5/+167
|
* address some timing issues; data seems more reliable nowJoey Hewitt2021-01-044-9/+16
|
* clean up property-setting a bitJoey Hewitt2021-01-037-25/+37
|
* add RilWrapper to git, to aid understandingJoey Hewitt2021-01-033-42/+2630
|
* make data more reliably workingJoey Hewitt2021-01-034-46/+59
| | | | I think Deactivate and Reset was overkill that was breaking it
* add CSD clientJoey Hewitt2020-03-254-2/+151
|
* data is now intermittently working on t0lteJoey Hewitt2020-02-1712-62/+321
| | | | | It's flaky. There may be some wrong interactions between us and Android for setting up IP addrs, routes, and DNS.
* add service config to be imported by device treesJoey Hewitt2019-10-043-0/+90
|
* Update README.mdJoey Hewitt2019-08-171-1/+1
|
* update project statusJoey Hewitt2017-10-131-26/+16
|
* extract auxiliary classes and fix rawtypesJoey Hewitt2017-10-064-16/+57
|
* clean up BaseCommands' empty implementationsJoey Hewitt2017-10-0610-9/+211
| | | | | Identify empty impls, mark them abstract, then provide exception-throwing methods to implement them.
* fix SIM status on Replicant 6Joey Hewitt2017-10-063-1/+10
| | | | | | Due to empty implementation of iccOpenLogicalChannel() in BaseCommands, the framework was waiting forever for UiccCarrierPrivilegeRules to load, so the SIM would never appear as present.
* fix early crashes on Replicant 6Joey Hewitt2017-10-063-1/+6
| | | | | Wrapper class was not public, and Proguard was removing an enclosing class that the dbus lib requires.
* fix build on Replicant 6Joey Hewitt2017-10-062-0/+27
| | | | implement new CommandsInterface methods
* (un)load modules in response to oFono's interfacesJoey Hewitt2017-09-2616-283/+452
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | * restructure inter-module dependencies to support this * dispatch all signals through RilOfono class so we don't have to deal with registering/unregistering signals as modules come and go * store module references in RilOfono (rather than RilWrapper) with internal interface types, so we can dispatch signals * remove _unknown registration state (not a real oFono string), use "none" instead * remove VoiceRadioTechnologyGetter, because the code using it ended up needing more netreg interaction, to the point we may as well just use the oFono stuff directly * allow suppressing propchange logs, because two modules are now tracking netreg props * returning RADIO_NOT_AVAILABLE in the RilWrapper, when modules are null, seemed to prevent the framework from coming online, so just use GENERIC_FAILURE with a modified log message * work around issue where ConnManager never reports Attached. It seems to be a new issue, I guess due to timing differences with this new implementation. The workaround is to watch NetworkRegistration for Status changes, and reload ConnectionManager's props in response.
* extract NetworkRegistrationModule from ModemJoey Hewitt2017-09-218-167/+232
|
* dependency injection and signal unregisteringJoey Hewitt2017-09-218-54/+109
| | | | | | | | | Main object passes oFono interfaces into modules, instead of them querying the interfaces themselves. Explicitly unregister signals instead of throwing away the dbus connection. Preparing to dynamically create/destroy individual modules in response to oFono interfaces coming (un)available.
* fix PDU length bug (casting to integer)Joey Hewitt2017-09-153-5/+6
| | | | Update status, and move test execution to modem-up.
* update feature statusJoey Hewitt2017-09-151-3/+2
|
* ofono: merge identical path-and-props structsJoey Hewitt2017-09-1510-87/+33
|
* handle modem up/down properlyJoey Hewitt2017-09-156-101/+200
| | | | | | | | | | | | My test case was to restart ofonod. Before this change, the RIL wouldn't bring the modem up again. proguard: allow dbus exception classes to survive (created by reflection) Remove hardcoded modem path, take the "first" modem. Make the error report for some semi-common dbus errors be briefer.
* refactor CallState. use enum like similar areas doJoey Hewitt2017-09-143-40/+39
|
* fixup incoming calls with DIALING stateJoey Hewitt2017-09-141-1/+8
| | | | | This is what oFono reports (maybe QMI voicecall bug), but it breaks Android's expectations, so map that to INCOMING.
* simplify build, refine proguardJoey Hewitt2017-09-143-19/+8
| | | | | | | | Remove intermediate buildtool build step, because we can just let Proguard strip out the build tool in the final APK. Change some -keep to -keepclassmembers, as we only need the member if the class is ever used.
* enable proguardJoey Hewitt2017-09-132-3/+25
|
* implement voicemail alert: put count in EF_MWISJoey Hewitt2017-09-092-32/+61
| | | | | Fix bugs and cleanup the SIM file refresh implementation since it was hard to test in the last commit.
* refactor SimFiles to notify of changesJoey Hewitt2017-09-094-45/+106
|
* improve datacall implementationJoey Hewitt2017-09-095-30/+157
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Bugfixes, work with oFono driver other than rilmodem. * Notify VoiceNetworkStateRegistrants when we attach to the data network -- seems to be the way to get the framework to re-poll data registration state and start any datacalls it wants to. This should fix the problems where the framework wouldn't even try to start a datacall. * Fall back on voice radio technology value if oFono doesn't provide data technology. Do it by dependency injection of a limited interface to minimize coupling. * Return even inactive datacalls (not sure that it matters much). * Return the netmask, in e.g. "/30" notation as expected by framework code. * Setup/teardown the network interface, using INetworkManagementService. I guess the oFono rilmodem plugin was doing this for us as a side-effect. I found the network management service by searching for Java code that could do that stuff, but we could switch to something else if needed. Currently relying on oFono to give us the interface name (patched my gobidev plugin to specify rmnet0), but I wonder if interface assignment is something better done outside oFono. Does the modem care which interface is used for which context? It seems weird, in this usecase at least, for oFono to give us an interface, if it has no relation to the modem. Could investigate rild's message "Interface rmnet1 will be used for CID(2)" to see how it decides. Some errors I was seeing early on may have come from oFono not having an interface assigned, so if we do leave that unset on their end maybe there's something to work around.
* update log levels for messagesJoey Hewitt2017-09-094-6/+5
|
* implement limited DTMF supportJoey Hewitt2017-09-095-40/+45
| | | | | | | oFono has no way for us to support the full Android API of start/stop tones on demand. The implementation seems good enough to navigate my voicemail. There was some flakiness, likely caused by my quick hacky implementation in oFono for DTMF on QMI devices.
* improve SELinux policyJoey Hewitt2017-09-093-5/+52
| | | | | Gets rid of 'init' warning us it's starting daemons in the init domain, and allows the debugger to strace ofonod if it crashes.
* update statusJoey Hewitt2017-09-051-8/+8
|
* voicecall: fix number/name presentationJoey Hewitt2017-09-031-6/+24
|
* wrapper selectively sends method calls to dbus threadJoey Hewitt2017-09-0310-71/+197
| | | | | | | Methods can mark themselves as safe to run on main thread, in which case they will not be sent to the dbus thread. A lot are safe for main thread, because they either always return UNSUPPORTED, or they just return some props we've been keeping up-to-date.
* exclude dbus test classes from buildJoey Hewitt2017-09-031-1/+1
|
* dynamically build the wrapper classJoey Hewitt2017-09-0321-2779/+635
|
* fix warnings, such as raw typesJoey Hewitt2017-09-0215-74/+80
|
* ignore spurious prop changes (value didn't actually change)Joey Hewitt2017-09-023-12/+20
| | | | netreg CellID seems especially prone to this
* notifications cleanupJoey Hewitt2017-09-028-73/+58
| | | | | * Removed unnecessary cross-thread posting (Messages will find their own way to their thread) * Centralize logging of notifications and facade Android's RegistrantList under my own
* make exception handling hopefully cleanerJoey Hewitt2017-09-024-50/+44
| | | | | | Such as: * Let catch-alls stop us from crashing unnecessarily * Log and drop things that aren't severe
* implemented signal strengthJoey Hewitt2017-09-024-21/+21
|
* refactor: wrapper class that handles the async partsJoey Hewitt2017-09-0213-776/+3301
| | | | | | | | | | Should remove repetitive code, and chances for errors like doing stuff on the wrong thread, or an exception that crashes the phone process or never responds to the caller after the error occurrred. One downside is that now *everything* is shuttled to another thread, even the trivial ones that just return some props. This can be addressed later by something like a @SafeForMainThread annotation.
* switch oFono to Gobi driver!Joey Hewitt2017-09-018-25/+59
| | | | | | | | | | Basic network info, and SMS seem to be working. No calls or USSD, as they are not supported by the oFono mainline. * update init.rc patches * fix some assumptions I made (ConnManager was sometimes not available in the QMI version, causing crash; and incoming SMSs seem to have no SMSC) * change hardcoded (in one place) dbus path to the modem
* switch to raw SMS PDU signal I added to oFonoJoey Hewitt2017-08-305-112/+91
|
* switch to raw SMS PDU method I added to oFonoJoey Hewitt2017-08-303-37/+27
|
* add system patches neededJoey Hewitt2017-08-305-6/+126
|
* update README with current statusJoey Hewitt2017-08-301-12/+35
|
* fixes to work with new StrictMode enforcementJoey Hewitt2017-08-302-15/+38
|
generated by cgit v1.2.3 (git 2.25.1) at 2026-01-28 21:56:20 +0000