[Ovmsdev] SSL Support

Mark Webb-Johnson mark at webb-johnson.net
Tue Feb 18 15:31:27 HKT 2020


Yes, unfortunately we cannot provide sdkconfig with the git, as everyone has their own small changes. So we provide the default sdkconfig in support directly. We also try to let people know (via this list) if there are changes suggested.

The diff command I gave is a good way of seeing if you are up-to-date.

Regards, Mark

> On 18 Feb 2020, at 3:28 PM, Chris van der Meijden <chris at arachnon.de> wrote:
> 
> Hi Mark,
> 
> it is compiling again :-)
> 
> The build system was OK, status and submodules too. It was the sdkconfig.
> 
> When you check out "a blank canvas" and start the make you get some minimal sdkconfig. With the previous code, that sdkconfig compiled after adding some CONFIG_PPP_ parameters. Now, with the new code just adding the CONFIG_PPP_ parameters was not enough.
> 
> I now took the sdkconfig.default.hw31 and voila, everything runs fine :-)
> 
> Thanx for your help.
> 
> Regards
> 
> Chris
> 
> Am Dienstag, den 18.02.2020, 09:05 +0800 schrieb Mark Webb-Johnson:
>> 
>> It sounds like your build system is not up to date, or sdkconfig parameters incorrect. Most likely MBEDTLS is not enabled.
>> 
>> Can you make sure you have ‘git pull’, and ‘git submodule update’, and then provide:
>> 
>> git status -v
>> git branch -v
>> git describe
>> git submodule status
>> diff -u support/sdkconfig.default.hw31 sdkconfig
>> 
>> Can you also try a ‘make clean’, and then ‘make’?
>> 
>> Regards, Mark.
>> 
>>> On 18 Feb 2020, at 8:41 AM, Chris van der Meijden <chris at arachnon.de <mailto:chris at arachnon.de>> wrote:
>>> Today I checked out the new code and tried to compile OVMS.V3.
>>> 
>>> No success :-(
>>> 
>>> I get this error:
>>> 
>>> ovms_server_v2.cpp:791:10: error: 'struct mg_connect_opts' has no member named 'ssl_ca_cert'
>>>      opts.ssl_ca_cert = MyOvmsTLS.GetTrustedList();
>>>           ^
>>> ovms_server_v2.cpp:792:10: error: 'struct mg_connect_opts' has no member named 'ssl_server_name'
>>>      opts.ssl_server_name = m_server.c_str();
>>> 
>>> I tried to figure out what goes wrong for about two hours now, but I'm stuck ...
>>> 
>>> Can you give me an hint on what goes wrong here?
>>> 
>>> Thanx
>>> 
>>> Chris
>>> 
>>> Am Montag, den 17.02.2020, 15:34 +0800 schrieb Mark Webb-Johnson:
>>>> I forgot to answer one question:
>>>> 
>>>>>> Maybe we can also provide a similar option for the Hologram registration, or at least some way to hand over the ICCID for registration.
>>>> 
>>>> Hologram has an API. Authentication is either by API key or username+password (from hologram.io <http://hologram.io/>). So, we could ask for the hologram  credentials and then request to activate the SIM:
>>>> 
>>>> POST https://dashboard.hologram.io/api/1/links/cellular/bulkclaim <https://dashboard.hologram.io/api/1/links/cellular/bulkclaim>
>>>> 
>>>> (the data is JSON containing things like the SIM ICCID, plan, zone, orgid (from user account), etc)
>>>> 
>>>> For example, to retrieve your own account information:
>>>> 
>>>> $ curl --verbose GET 'https://dashboard.hologram.io/api/1/users/me <https://dashboard.hologram.io/api/1/users/me>' -u <emailaddress>:<password>
>>>> 
>>>> or, using an API key:
>>>> 
>>>> $ curl --verbose GET 'https://dashboard.hologram.io/api/1/users/me <https://dashboard.hologram.io/api/1/users/me>' -u apikey:<apikey>
>>>> 
>>>> I don’t see any API to create a hologram account.
>>>> 
>>>> Regards, Mark.
>>>> 
>>>>> On 17 Feb 2020, at 3:04 PM, Mark Webb-Johnson <mark at webb-johnson.net <mailto:mark at webb-johnson.net>> wrote:
>>>>> I guess about 50% of new user support requests are related to not understanding ’server password’, with the other 50% being vehicle ID registration and modification.
>>>>> 
>>>>> From a user’s point of view, they are very aware of web service usernames and passwords. They know how to register for an account, and are used to it. They know that when they get an App, they put in the corresponding website credentials and it simply works.
>>>>> 
>>>>>> On auth: I think #1 will not help that much. #2 is an option. Most needed is a way to bypass the need to register user & vehicle on the server first. After filling in the module ID & password, there should be an option to create a new server + vehicle account and retreive the credentials simply by clicking a button (as outlined previously: https://github.com/openvehicles/Open-Vehicle-Monitoring-System-3/issues/182 <https://github.com/openvehicles/Open-Vehicle-Monitoring-System-3/issues/182>).
>>>>> 
>>>>> We have a to be careful, because many users have more than one vehicle on their account (at least on my server). I think from the user interface point of view it should be server username/email and password. We may implement it differently under the hood, but keep it as simple as possible for the user. Note that v3 MQTT works off username+password as well.
>>>>> 
>>>>> An API so that this could be done from the server / app makes good sense. Not too hard to do (in a standard way, with plugins for different backend servers). Something like:
>>>>> 
>>>>> Which server would you like to use?
>>>>> Do you already have an account?
>>>>> Yes: Enter username and password, then validate
>>>>> No: Provide a form to register a new account, then create
>>>>> What ID would you like to use for this vehicle?
>>>>> 
>>>>> I too would like to do away with pre-registered vehicle IDs, and was thinking about ways around this last week. The way I see this working is:
>>>>> 
>>>>> When the module logs on with a particular user+password and vehicle ID, if the vehicle record doesn’t exist we can simply add it. This is similar to the way MQTT does it.
>>>>> For the Apps, we can provide an API to enumerate the available vehicle IDs for user selection. I also like your suggestion for a QR code on the module for quick App setup.
>>>>> The server can continue to show the registered vehicles on each account.
>>>>> However, there is one issue here in that for v2 the vehicle ID must be unique on that server. For MQTT, it only needs to be unique for each user.
>>>>> We should change the database structure, and ovms_server.pl code, to add owner ID as a key field (along with Vehicle ID), to address this. The change is not too complex.
>>>>> 
>>>>> I did spend some time reviewing the v2 protocol, and it seems easy to add another protocol scheme. Our current scheme 0x30 is for RC4 encryption based on the server password. We can add say 0x40 for just plaintext within SSL. The MP-C and MP-A methods currently pass the encryption scheme, followed by some secure tokens and the vehicle ID for authentication. The 0x40 scheme can simply be passed the desired vehicle ID, plus the server username/email and password (or authentication token).
>>>>> 
>>>>> The other thing to think about is the password itself. If we simply use server username/email and password, then it becomes an issue if the password is changed. Suddenly, three different systems break. The module in particular is worrying (as a password change on the server would break the module). So perhaps best to use tokens, and have a way for a module/app to request a persistent token. We could present that on the server as authorised apps/modules, and provide a facility to cancel (delete) a token. That is a lot more secure as the user can easily see what is authorised. My suggestion is that authentication works based on:
>>>>> 
>>>>> Either server username OR server eMail address
>>>>> Either server password OR authentication token
>>>>> 
>>>>> Whatever we do, the concentration should be on v3 MQTT protocol. We need to eventually migrate the Apps to that, as it is the future. The other option is to forget about supporting this in v2, and concentrate solely on v3 MQTT. Get this new arrangement working with v3, upgrade the Apps to support it, and then migrate to that.
>>>>> 
>>>>> Regards, Mark.
>>>>> 
>>>>>> On 16 Feb 2020, at 8:04 PM, Michael Balzer <dexter at expeedo.de <mailto:dexter at expeedo.de>> wrote:
>>>>>> Deployed on my server as well.
>>>>>> 
>>>>>> I've also added enabling TLS by default to the setup wizard, so new users by default will have strong encryption.
>>>>>> 
>>>>>> The V2 connection performance is much better than with HTTP.Request(), even the initial connect is done within ~3 seconds normally. TX/RX afterwards has no significant delay.
>>>>>> 
>>>>>> On auth: I think #1 will not help that much. #2 is an option. Most needed is a way to bypass the need to register user & vehicle on the server first. After filling in the module ID & password, there should be an option to create a new server + vehicle account and retreive the credentials simply by clicking a button (as outlined previously: https://github.com/openvehicles/Open-Vehicle-Monitoring-System-3/issues/182 <https://github.com/openvehicles/Open-Vehicle-Monitoring-System-3/issues/182>).
>>>>>> 
>>>>>> Maybe we can also provide a similar option for the Hologram registration, or at least some way to hand over the ICCID for registration.
>>>>>> 
>>>>>> Filling in credentials manually must still be possible of course, but the standard case of a new user would be simplified as far as possible.
>>>>>> 
>>>>>> OAuth could simplify the App connection, but not substantially -- it would only allow to skip entering the password. A better option here would be to display the full App credentials as a QR code during the setup (and on the server page), so the user just needs to scan that to configure the App.
>>>>>> 
>>>>>> Regards,
>>>>>> Michael
>>>>>> 
>>>>>> 
>>>>>> Am 11.02.20 um 07:27 schrieb Mark Webb-Johnson:
>>>>>>> 
>>>>>>> I looked into #6 (iOS App) and #5 (OVMS Server) and switching to use SSL is trivial in both.
>>>>>>> 
>>>>>>> The commit for the server side is already done and pushed. Running as tcp/6870 on www.openvehicles.com <http://www.openvehicles.com/> server. Perhaps @michael can do the same on his (so the servers are at least ready for this)?
>>>>>>> 
>>>>>>> But before doing the iOS App, I think we should discuss the bigger issue of authentication and passwords. Once we SSL encrypt (and authenticate the server), we no longer need to use the password for encryption (although we do need to use something for authentication). Looking at a recent intro video for OVMS reminded me of how many passwords we current have:
>>>>>>> 
>>>>>>> Website user account and password
>>>>>>> Vehicle ID
>>>>>>> Server password
>>>>>>> Module password
>>>>>>> Hologram account + password
>>>>>>> 
>>>>>>> I think we must have website user account and password. Some identifier for the vehicle is also probably necessary (so we can have more than one vehicle per account). If we use the website user+password for authentication, we don’t need the server password (which is the one that causes the most confusion). The module password perhaps needs to stay for local admin. Same for hologram (unless we offer packages with credits purchased under openvehicles.com <http://openvehicles.com/> or the server - which is something I have been reluctant to do).
>>>>>>> 
>>>>>>> We can remove the requirement for server password in one of two ways:
>>>>>>> 
>>>>>>> Simply extend/replace the MP-C / MP-A calls to have an authentication mechanism passing the website user, website password, and vehicle ID. The server would then validate the user and password provided, and ensure that the provided Vehicle ID matches a registered vehicle for that user. No more ’server password’. This is close to what the mqtt v3 protocol does already (except it doesn’t need the vehicleid as everyone’s namespace is kept separate).
>>>>>>> 
>>>>>>> Move to an authentication token system. Have some API (HTTP or within the v2 protocol) to authenticate website user+password and return a token (stored persistently in a database). Subsequent API calls can simply pass that token to authenticate themselves. A page on the website (similar to our current vehicles page) could show the authenticated tokens (apps/cars) and allow them to be cleared. There are various standard mechanisms for this, and it is considered better suited to API style usages with third party apps.
>>>>>>> 
>>>>>>> There is also the whole OAUTH thing we could just move to (probably in combination with option 2).
>>>>>>> 
>>>>>>> What do people think?
>>>>>>> 
>>>>>>> Regards, Mark.
>>>>>>> 
>>>>>>>> On 10 Feb 2020, at 11:24 AM, Mark Webb-Johnson <mark at webb-johnson.net <mailto:mark at webb-johnson.net>> wrote:
>>>>>>>> 
>>>>>>>> Given that we use the mongoose library for most of our stuff, adding SSL support should not be hard. This would finally bring strong encryption and server side authentication. Given the number of attacks now on IoT devices that would not be a bad thing.
>>>>>>>> 
>>>>>>>> I think what we need is:
>>>>>>>> 
>>>>>>>> A set of helper functions to make it easier for components to use SSL. Build on top of mongoose.
>>>>>>>> 
>>>>>>>> A way to manage a list of trusted Certificate Authorities, including adding to the trusted list via:
>>>>>>>> Components providing Certificate Authorities in firmware.
>>>>>>>> Certificate Authorities in configuration.
>>>>>>>> 
>>>>>>>> Extensions to ovms_server_v2 to support an SSL connection option.
>>>>>>>> 
>>>>>>>> Extensions to ovms_server_v3 to support an SSL connection option.
>>>>>>>> 
>>>>>>>> Extensions to the Ovms Server v2 code to support an SSL connection listener.
>>>>>>>> 
>>>>>>>> Extensions to the iOS App to support an SSL connection option.
>>>>>>>> 
>>>>>>>> Extensions to the Android App to support an SSL connection option.
>>>>>>>> 
>>>>>>>> Migration of any components already supporting SSL to this new standardised approach.
>>>>>>>> 
>>>>>>>> Then we can open up the discussion of the whole thing of passwords. We have far too many of these at the moment (user account+password, vehicle ID, server password, module password, hologram account+password, etc). Once we have an encrypted connection, we don’t need to use the password for encryption, but merely for authentication. That simplifies things, as we can perhaps just use the user account+password for most things (giving access to all vehicles registered under that user account - in a similar way to MQTT does it already for ovms_server_v3).
>>>>>>>> 
>>>>>>>> I will take on the majority of this project. I can do #1, #2, #3, #4, #5, and #6).
>>>>>>>> 
>>>>>>>> If anyone has any feedback on requirements, please let me know.
>>>>>>>> 
>>>>>>>> Regards, Mark.
>>>>>>>> 
>>>>>>>> _______________________________________________
>>>>>>>> OvmsDev mailing list
>>>>>>>> OvmsDev at lists.openvehicles.com <mailto:OvmsDev at lists.openvehicles.com>
>>>>>>>> http://lists.openvehicles.com/mailman/listinfo/ovmsdev <http://lists.openvehicles.com/mailman/listinfo/ovmsdev>
>>>>>>> 
>>>>>>> 
>>>>>>> _______________________________________________
>>>>>>> OvmsDev mailing list
>>>>>>> OvmsDev at lists.openvehicles.com <mailto:OvmsDev at lists.openvehicles.com>
>>>>>>> http://lists.openvehicles.com/mailman/listinfo/ovmsdev <http://lists.openvehicles.com/mailman/listinfo/ovmsdev>
>>>>>> 
>>>>>> _______________________________________________
>>>>>> OvmsDev mailing list
>>>>>> OvmsDev at lists.openvehicles.com <mailto:OvmsDev at lists.openvehicles.com>
>>>>>> http://lists.openvehicles.com/mailman/listinfo/ovmsdev <http://lists.openvehicles.com/mailman/listinfo/ovmsdev>
>>>>> 
>>>> 
>>>> _______________________________________________
>>>> OvmsDev mailing list
>>>> OvmsDev at lists.openvehicles.com <mailto:OvmsDev at lists.openvehicles.com>
>>>> http://lists.openvehicles.com/mailman/listinfo/ovmsdev <http://lists.openvehicles.com/mailman/listinfo/ovmsdev>
>>> _______________________________________________
>>> OvmsDev mailing list
>>> OvmsDev at lists.openvehicles.com <mailto:OvmsDev at lists.openvehicles.com>
>>> http://lists.openvehicles.com/mailman/listinfo/ovmsdev
>> 
>> _______________________________________________
>> OvmsDev mailing list
>> OvmsDev at lists.openvehicles.com <mailto:OvmsDev at lists.openvehicles.com>
>> http://lists.openvehicles.com/mailman/listinfo/ovmsdev <http://lists.openvehicles.com/mailman/listinfo/ovmsdev>
> _______________________________________________
> OvmsDev mailing list
> OvmsDev at lists.openvehicles.com
> http://lists.openvehicles.com/mailman/listinfo/ovmsdev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openvehicles.com/pipermail/ovmsdev/attachments/20200218/4bc2a522/attachment-0001.html>


More information about the OvmsDev mailing list