connman: revisit patches and (again) try upstream the wifi patch about venus HOT 9 CLOSED

Regarding 1. hwclock: read this to see ConnMans maintainers opinion on syncing to the hwclock:
https://bugs.tizen.org/jira/si/jira.issueviews:issue-html/TC-165/TC-165.html

Also, rtc and all related datetime mgmt works fine on the ccgx. No complaints and no issues. So just adding a few explanation lines on that patch, for example a link to this issue, is more than enough.

@jhofstee: you'll be happy to know that systemd is coming and all will be better. See Tizen link 😄.

from venus.

Some extra info:

about 1) "normally" the hwclock is saved during shutdown / reboot. Since
a color control can just be removed from power, we always update the rtc as well.

about 3) g-web multi header:

it belongs to this thread:

• https://www.mail-archive.com/[email protected]/msg15637.html
• https://www.mail-archive.com/[email protected]/msg15731.html

from venus.

Some findings:

1) hwclock

The patch is not suitable for upstream, but it's small and should be easy to maintain. An alternative could be a cron job that periodically syncs the hardware clock.

2) ntp max poll

From RFC5905:

Poll: 8-bit signed integer representing the maximum interval between successive messages, in log2 seconds. Suggested default limits for minimum and maximum poll intervals are 6 and 10, respectively.

Increasing the value from 10 to 16 changes the maximum interval from 17 minutes to 18 hours although there actual intervals may be shorter depending on network conditions and clock stability.

The patch as is would probably not be accepted upstream as it changes behaviour without a good reason. A new option allowing this value to be set in the config file might have a better chance.

3) gweb multi-header

An attempt to ask about this on the connman mailing list has received no response, and the message to the list appears to have been black-holed.

from venus.

@mansr

1. hwclock: let's leave that as a patch

2. ntp max poll: I think the rationale is to reduce network traffic. Let's leave as a patch as well

3. gweb multi-header: have you looked into reproducing that yet?

4. wifi: ignore invalid key on AR failure:
   you already made the patch apply to the latest Connman version right? Is there much work involved in testing its effectiveness?

I'd prefer to first update our Connman to latest version including these patches; except possibly 3. And then see about mainlining patch (4).

from venus.

The gweb patch is now upstream: http://git.kernel.org/cgit/network/connman/connman.git/commit/?id=22901212105055b24f73504a74f5a57eee809777

from venus.

@mansr how is the upstreaming going?

from venus.

just for record keeping, the last attempt at upstreaming the 'dont ever give up' wifi patch is here:
https://lists.01.org/pipermail/connman/2017-January/021470.html. And here the discussion ended.

And the patch is also here. It is scheduled to replace what we have in meta-victronenergy/master now soon.

from venus.

I've sent connman a reminder email: https://lists.01.org/pipermail/connman/2017-August/021993.html

from venus.

Since upstreaming the remaining patches seems impossible, I am closing this issue.

from venus.