Hi Jeremiah, Please find my comments below inline:
Regards, Abhishek ------- Original Message ------- Sender : Jeremiah Foster<[email protected]> Date : Apr 23, 2014 21:04 (GMT+09:00) Title : Re: Re: Proposal for new component in Tizen IVI On Wed, Apr 23, 2014 at 1:11 PM, Abhishek Sharma <[email protected]> wrote: Hi Jeremiah, Till date, we have been primarily involved with Tizen IVI release for Pandaboard. Where does this release live? I see this: https://wiki.tizen.org/wiki/IVI/IVI_Platforms and https://wiki.tizen.org/wiki/Tizen_IVI_Getting_Started_Guide_For_PandaBoard but that does not seem to have any repos connected with it. Also, its looks like you're pulling in some Android drivers to do stuff like WLAN. What is the deal with Bluetooth? Are you not using bluez? <Abhishek> Pandaboard is another reference board for Tizen IVI. For hardware adaptation firmware driver for WLAN, bluetooth corresponding to that board was used. Bluez is present as well. All the details related to the Tizen IVI releases can be found at https://wiki.tizen.org/wiki/IVI/IVI_Nov_25,_2013_2.0_ARM_release. Please check the same for more info on packages/ downloads. Based on requirements for Tizen IVI, we started looking into contributing for System Health Manager. As mentioned in previous communications, the main idea of System Health Manager is to provide for a minimum fault tolerance capability by identifying faults/failure at runtime and take corrective (recovery) actions accordingly. Did you look at Node Health Monitor? It likely does what you need. <Abhishek> yes we are analysing it and will propose a way forward. Systemd was the obvious choice for the same as it typically starts most of the critical services and receives signals from them. The System Health Manager is simple standalone system component which captures the state of monitored critical services using Systemd dbus signals and interfaces, saves them in the database as fault-data of services and calls a pre-defined recovery client for the detected/diagnosed failure. Please find some of the use cases which warrants components like System Health Manager in an IVI platform. Case 1: As times, it is observed that when a particular infotainment application is installed and run, it affects functioning of some of the other existing working applications. Typically, error message displayed is "running out of memory". In such a scenario, if there is out of memory detected for a particular service, HM can initiate (a) Garbage collection mechanism like GC in android to recover the unused memory besides restarting the service. (b) termination of high memory consuming, less priority background services. Case 2: HM can maintain a threshold value of the memory for a given service, If memory consumed exceeds the threshold value and it is detected by resourced, HM shall notify the same to the corresponding service. Case 3: If an abrupt termination of a service is detected due to signal or core-dump, HM shall apply a corresponding recovery strategy like CheckPointing/Restart, service restart, node restart, deferred restart etc. Case 4: Update, maintain fault data which can be uploaded to server for remote diagnostics purposes. It could also transfer this data on a CAN bus using AMBD to the OBD port. Case 5: If a service/app has been scheduled for a deferred/ delayed start, the system shall keep a track of this. @Jean-Pierre Thanks for providing the details and links for the code base. We are currently going through that and we propose on how we could have a synergy to co-develop this component. We are looking into Node State Controller in details and will take a decision to co-work to see how we could potentially integrate that solution into Tizen IVI. Will propose on that shortly. This is good news. We suggest to proceed with the following steps. 1. Discuss and share the use cases and requirements (we have already started that). 2. Understand similarity and differences and how to address them. GENIVI has done significantly work the last few years in defining a middleware stack that is suitable for an IVI unit. One of the things we've gone through, for example, is bluez which we've worked hard to align with. This is something that Tizen also uses and is an example of how we as an industry can collaborate on a common code base. You appear to be moving away from bluez (judging by the kernel configs you recommend in the wiki.) Can you tell me what the plan is for having bluetooth on your system? <Abhishek> Pandaboard release is based on Tizen IVI releases. However, hardware specific components naturally will be different. Please see the link mentioned earlier for release details. 3. Address the issues like modularity, dependency on other modules and their availability in Tizen IVI platform. You appear to have a lot of dependencies on Android components. Can you explain why that is? <Abhishek> Our proposal in discussion does not have anything to do with Andorid. A passing reference is made to Android in explaining one of the use cases. 4. Leverage the existing Tizen IVI platform components like system-server, SQLite, Systemd, dbus-binding etc. Tizen IVI has worked hard to become GENIVI compliant and Intel has contributed a great deal of high-quality software to that effort. Do you plan to reuse that as well? Its seems like you're not really using Tizen IVI but just cherry-picking some packages, which is fine of course, it will likely limit the contribution you get and other's interest in your software. <Abhishek> Tizen IVI for pandaboard has some specific hardware adaptation changes; All other packages remain the same as Tizen IVI release. Please do note that Tizen IVI 1.0 for C210 device is Genivi compliant. 5. How to seemlessly integrate this with Tizen IVI. Seamless integration is hard, let's go shopping! :-) Thank you very much for your considered and detailed response in this thread, it is much appreciated. Regards, Jeremiah Regards Abhishek Note: Sorry for the delay in response. Due to holidays, could not respond to the mails earlier. ------- Original Message ------- Sender : Jeremiah Foster Date : Apr 22, 2014 16:51 (GMT+09:00) Title : Re: Re: Proposal for new component in Tizen IVI On Wed, Apr 16, 2014 at 2:49 PM, Stoppa, Igor wrote: Hello Abhishek, On 16 April 2014 15:33, Abhishek Sharma wrote: It's a typo. Please ignore the same. We had to remove those to reduce the file size for posting on the mailing list. I think it would be good if you could share those use cases even in text format I agree with Igor, sharing your use cases will go a long way towards describing the relevance of your new software. From what I understand Samsung is not a car maker, at least not directly, and there is already Open Source software which does a lot of what you're stating you'll do. Perhaps detailed use cases might be a way to help explicate your needs? Have you taken a look at the other projects that do something similar to yours, like the Node State Controller in GENIVI? It would be good to hear about your plans, this thread has gone quiet and that is not necessarily good for momentum. Cheers, Jeremiah -- ============================================= Jeremiah C. Foster GENIVI Community Manager Pelagicore AB Ekelundsgatan 4, 6tr, SE-411 18 Gothenburg, Sweden Mobile: +46 (0)730 93 0506 E-Mail: [email protected] ============================================= === NOTE === The information contained in this E-mail message is intended only for use of the individual or entity named above. If the reader of this message is not the intended recipient, or the employee or agent responsible to deliver it to the intended recipient, you are hereby notified that any dissemination, distribution or copying of this communication is strictly prohibited. ============= _______________________________________________ IVI mailing list [email protected] https://lists.tizen.org/listinfo/ivi
