Send connman mailing list submissions to
        [email protected]

To subscribe or unsubscribe via the World Wide Web, visit
        https://lists.01.org/mailman/listinfo/connman
or, via email, send a message with subject or body 'help' to
        [email protected]

You can reach the person managing the list at
        [email protected]

When replying, please edit your Subject line so it is more specific
than "Re: Contents of connman digest..."


Today's Topics:

   1. Re: [PATCH 1/2] ntp: Add option TimeUpdateThreshold
      (Marcel Holtmann)
   2. Re: [PATCH 2/2] ntp: Add option TimeserversFromDHCP
      (Marcel Holtmann)
   3. Re: [RFC v0 0/3] Refactoring firewall and iptables code
      (Marcel Holtmann)
   4. Re: [PATCH 1/2] ntp: Add option TimeUpdateThreshold
      (Boszormenyi Zoltan)


----------------------------------------------------------------------

Message: 1
Date: Mon, 7 Mar 2016 09:19:47 -0800
From: Marcel Holtmann <[email protected]>
To: Zolt?n B?sz?rm?nyi <[email protected]>
Cc: [email protected]
Subject: Re: [PATCH 1/2] ntp: Add option TimeUpdateThreshold
Message-ID: <[email protected]>
Content-Type: text/plain; charset=utf-8

Hi Zoltan,

> Add option TimeUpdateThreshold (type: double units: seconds)
> to make the threshold for using settimeofday() vs adjtime()
> configurable. The RTC in some machines is so bad that with the
> current hardcoded 0.128 seconds, connman occasionally decides
> to use settimeofday() instead of adjtime().
> 
> Signed-off-by: Zolt?n B?sz?rm?nyi <[email protected]>
> 
> ---
> doc/clock-api.txt |  5 +++++
> src/clock.c       | 39 ++++++++++++++++++++++++++++++++++++++-
> src/connman.h     |  1 +
> src/dbus.c        |  3 +++
> src/ntp.c         |  6 +++---
> 5 files changed, 50 insertions(+), 4 deletions(-)
> 
> diff --git a/doc/clock-api.txt b/doc/clock-api.txt
> index 6818f5a..06e14f0 100644
> --- a/doc/clock-api.txt
> +++ b/doc/clock-api.txt
> @@ -46,20 +46,25 @@ Properties        uint64 Time [readonly or readwrite]  
> [experimental]
> 
>               string TimeUpdates [readwrite]  [experimental]
> 
>                       Possible values are "manual" and "auto" to indicate
>                       time update policy.
> 
>                       With the "auto" setting the system tries to use as
>                       many sources as possible to determine the correct
>                       and updated time.
> 
> +             double TimeUpdateThreshold [readwrite]  [experimental]
> +
> +                     Threshold value in milliseconds for settimeofday()
> +                     versus adjtime(). Default 0.128.
> +

this is something that belongs in main.conf and not in a D-Bus property. Since 
you will not make runtime changes to this value. It is specific to the platform 
you deploy.

Regards

Marcel



------------------------------

Message: 2
Date: Mon, 7 Mar 2016 09:20:55 -0800
From: Marcel Holtmann <[email protected]>
To: Zolt?n B?sz?rm?nyi <[email protected]>
Cc: [email protected]
Subject: Re: [PATCH 2/2] ntp: Add option TimeserversFromDHCP
Message-ID: <[email protected]>
Content-Type: text/plain; charset=utf-8

Hi Zoltan,

> Add option TimeserversFromDHCP (type boolean) to be able to disable
> the NTP servers published by the DHCP server. In some use cases the
> DHCP server is not under your control and you have to use a specific
> NTP server, not the DHCP-provided one.
> 
> Adding the gateway as a timeserver now also depends on this option.
> 
> Signed-off-by: Zolt?n B?sz?rm?nyi <[email protected]>
> 
> ---
> doc/clock-api.txt |  5 +++++
> src/clock.c       | 45 +++++++++++++++++++++++++++++++++++++++++++++
> src/connman.h     |  1 +
> src/dhcp.c        | 47 +++++++++++++++++++++++++----------------------
> src/dhcpv6.c      | 44 ++++++++++++++++++++++----------------------
> src/timeserver.c  |  2 +-
> 6 files changed, 99 insertions(+), 45 deletions(-)
> 
> diff --git a/doc/clock-api.txt b/doc/clock-api.txt
> index 06e14f0..be174f4 100644
> --- a/doc/clock-api.txt
> +++ b/doc/clock-api.txt
> @@ -83,10 +83,15 @@ Properties        uint64 Time [readonly or readwrite]  
> [experimental]
>               array{string} Timeservers [readwrite]  [experimental]
> 
>                       List of global default NTP servers. The list should
>                       be sorted in order of preference.
> 
>                       If a service configuration provides NTP servers,
>                       then they are preferred over the global ones.
> 
>                       This list of servers is used when TimeUpdates is set
>                       to auto.
> +
> +             boolean TimeserversFromDHCP [readwrite]  [experimental]
> +
> +                     Whether the time server(s) provided by DHCP should
> +                     be used. Default true.

same here. I think you want to make this a main.conf option.

Regards

Marcel



------------------------------

Message: 3
Date: Mon, 7 Mar 2016 09:23:49 -0800
From: Marcel Holtmann <[email protected]>
To: Daniel Wagner <[email protected]>
Cc: [email protected], Daniel Wagner <[email protected]>
Subject: Re: [RFC v0 0/3] Refactoring firewall and iptables code
Message-ID: <[email protected]>
Content-Type: text/plain; charset=us-ascii

Hi Daniel,

> here is the initial refactoring part for getting nftables support.
> 
> As you notice the arguments for the nat API is not as we discussed.
> The main reason for this is that if we use struct connman_ipconfig as
> argument tools/iptables-unit.c will pull in the complete ConnMan src
> directory. I dropped the firewall tests just to get it compiling
> again. Maybe we just delete those tests and have only iptables 'unit'
> tests. In this case the nat API with struct connman_ipconfig could
> just work fine.
> 
> All is just compile tested only. I just wanted to hear if this
> direction is okay with you guys. And if yes, I'd like to get this in
> before doing the nftables implementation.
> 
> What do you think?

I am actually totally okay with removing iptables support. I mean why keep 
trying to chase some old configuration method. The support for nftables made it 
upstream a long time ago and distributions will enable it.

Also keep in mind that iptables always was a stop-gap measure. Its API is racy 
and problematic at best. I do not really want to know what problems hide 
underneath that we yet have to discover when you do dynamic changes.

Regards

Marcel



------------------------------

Message: 4
Date: Mon, 7 Mar 2016 18:47:58 +0100
From: Boszormenyi Zoltan <[email protected]>
To: Marcel Holtmann <[email protected]>
Cc: [email protected]
Subject: Re: [PATCH 1/2] ntp: Add option TimeUpdateThreshold
Message-ID: <[email protected]>
Content-Type: text/plain; charset=utf-8; format=flowed

Hi,

2016-03-07 18:19 keltez?ssel, Marcel Holtmann ?rta:
> Hi Zoltan,
>
>> +            double TimeUpdateThreshold [readwrite]  [experimental]
>> +
>> +                    Threshold value in milliseconds for settimeofday()
>> +                    versus adjtime(). Default 0.128.
>> +
> this is something that belongs in main.conf and not in a D-Bus property. 
> Since you will not make runtime changes to this value. It is specific to the 
> platform you deploy.
>
> Regards
>
> Marcel

thanks for your comments, I will rework the patches.

Best regards,
Zolt?n B?sz?rm?nyi



------------------------------

Subject: Digest Footer

_______________________________________________
connman mailing list
[email protected]
https://lists.01.org/mailman/listinfo/connman


------------------------------

End of connman Digest, Vol 5, Issue 7
*************************************

Reply via email to