/connman_device_ifindex and
using the info from rtnl instead. Either with or without a device struct
as argument.
Any takers...?
Cheers,
Patrik
Patrik Flykt (3):
wifi: Don't fetch interface name from device structure
device: Remove unused device-devname
device: Don't store interface name
As no code requests the interface name from the device struct,
remove the attribute and associated code.
---
include/device.h | 2 --
src/device.c | 26 --
2 files changed, 28 deletions(-)
diff --git a/include/device.h b/include/device.h
index 57b925c..d97f84d 100644
---
src/device.c | 5 -
1 file changed, 5 deletions(-)
diff --git a/src/device.c b/src/device.c
index f0b7a4f..a97d790 100644
--- a/src/device.c
+++ b/src/device.c
@@ -60,7 +60,6 @@ struct connman_device {
char *interface;
char *ident;
char *path;
- char
Hi,
On Tue, 2014-04-29 at 20:18 -0400, Eduardo Abinader wrote:
The following two patches are related somehow in the
scenario where wifi dongle is removed and inserted again,
and systemd-udev renames device.
Sytemd-udev may have predefined device names, which does
not necessarily
On Tue, 2014-04-29 at 20:18 -0400, Eduardo Abinader wrote:
If a pending scan_callback is called after a dongle removal,
invalid reads occur in connman_device_reset_scanning, for
instance. This patch cleans the callback, before dbus remove
interface is returned.
Applied, thanks!
On Thu, 2014-05-01 at 08:05 +, Marcel Mulder wrote:
I experience reproducible connman crashes with version v1.20, v1.21,
v1.22, v1.23 and HEAD of git. Older version I did not try.
See if dd96d41ea182f0d1e0b55686adf793fa72e0e4d3 helps. At least it was
an attempt to solve a very similar
On Wed, 2014-04-09 at 16:54 +0300, Marko Saukko wrote:
Related systemd documentation:
http://www.freedesktop.org/wiki/Software/systemd/syslog/
Newer systemd versions (v35+) do not support non-socket-activated
syslog daemons anymore and we do no longer recommend people to order
their units
On Tue, 2014-04-22 at 11:21 +0300, Patrik Flykt wrote:
Hi,
Patch 2/3 fixes a situation where the lease expires but the structure
is not properly removed. This results in a crash, so it needs fixing.
Patch 3/3 ensures that refcounts are properly handled even if the
caller decides
removes an unused variable.
Cheers,
Patrik
Patrik Flykt (3):
gdhcp: Remove unnecessary listener channel variable
dhcp: Keep DHCP running until an explicit stop is called
dhcp: Restart existing DHCP client with __connman_dhcp_start()
gdhcp/client.c | 8 +--
src/dhcp.c | 69
In order to recover from a lost lease, gdhcp will restart the address
acquisition process. As the dhcp structures are removed from the hash
and freed in dhcp_invalidate() without notifying gdhcp, the following
crash can happen:
connmand[28337]: DHCP index 3: lease expired
connmand[28337]: DHCP
On Thu, 2014-04-10 at 11:55 +0800, leonew wrote:
so what can I do if client device want to connect the AP using WPS
method(both PBC and PIN)?
WPS is not supported on the tethering interface. It requires some amount
of planning and work to implement.
Cheers,
Patrik
On Thu, 2014-04-10 at 10:20 +0200, Tim Sander wrote:
It would be nevertheless nice to have the information connman posseses
available at the dbus. Like the DHCP information or ntp round trip
times. But as far as i remember you don't want these due to
maintenance burden. Alas.
That, and a
On Tue, 2014-04-08 at 08:51 +0200, Tim Sander wrote:
Hi Patrik and list
On Thu, 2014-04-03 at 15:35 +0200, Tim Sander wrote:
I currently have a connman 1.22 set up with time TimeUpdates set to
manual. This is due to the fact that i use ntpd for time
synchronisation due to the better
Hi,
On Thu, 2014-04-03 at 15:35 +0200, Tim Sander wrote:
I currently have a connman 1.22 set up with time TimeUpdates set to
manual. This is due to the fact that i use ntpd for time
synchronisation due to the better accuracy. If connman now is set to
dhcp i have a hard time
On Mon, 2014-04-07 at 14:19 +0100, Homam Dabis wrote:
I have a and RT5370 WiFi driver that supports wext. I would like to set it
up as an AP using connman 1.20.
I run connmand -d -n --wifi=wext
I get the following error when I try to enable tethering:
connmand[2550]:
On Fri, 2014-04-04 at 11:01 +0300, Tomasz Bursztyka wrote:
P2P technology exists if only wifi technology exists and does not own
any devices. Thus it's unnecessary to refcount it more than once.
Reported by Eduardo Abinader eduardo.abina...@openbossa.org
Applied, thanks!
Patrik
Hi,
On Thu, 2014-04-03 at 10:07 +0300, Tomasz Bursztyka wrote:
Identify which part does not support P2P: wpa_supplicant or the
interface itself.
---
gsupplicant/supplicant.c | 9 -
1 file changed, 8 insertions(+), 1 deletion(-)
diff --git a/gsupplicant/supplicant.c
On Thu, 2014-04-03 at 12:39 +0300, Jukka Rissanen wrote:
Hi,
I was trying to find the memory leak in DHCP and accidentally
opened a can of worms which resulted some important fixes
how DHCP handles the network hash.
All patches applied, thanks!
Patrik
On Thu, 2014-04-03 at 10:35 +0300, Tomasz Bursztyka wrote:
Identify which part does not support P2P: wpa_supplicant or the
interface itself.
Patch has been applied, thanks!
Patrik
___
connman mailing list
connman@connman.net
On Tue, 2014-04-01 at 16:55 +0300, Patrik Flykt wrote:
ConnMan creates a /var/run/connman state directory even though nothing
writes files there. ConnMan VPNd uses the directory for l2tp and ppp
configuration files. Fix this by removing the state directory creation
from ConnMan and modifying
On Wed, 2014-04-02 at 13:37 +0300, Jukka Rissanen wrote:
Reported by valgrind
Applied, thanks!
Patrik
___
connman mailing list
connman@connman.net
https://lists.connman.net/mailman/listinfo/connman
On Tue, 2014-04-01 at 10:54 +0300, Tomasz Bursztyka wrote:
New version of the patchset, changes against:
v2:
- no need of p2p_enabled in src/technology.c, just use existing
enable_persistent
- small compilation fix (device.h needs to include service.h
as early as
Hi,
On Mon, 2014-03-31 at 10:01 -0400, Glenn Schmottlach wrote:
The global routing table used by everybody else will have its default
route updated (or not) according to normal rules, i.e. the new service
if the new service can go online and the existing service was in state
On Mon, 2014-03-31 at 23:47 +0200, Michael Trimarchi wrote:
The best available frequency for tethering should be automatically
requested by ConnMan from wpa_supplicant. wpa_supplicant in turn may
unfortunately need some selective fixing to make this a working
solution.
We can proceed
ConnMan creates a /var/run/connman state directory even though nothing
writes files there. ConnMan VPNd uses the directory for l2tp and ppp
configuration files. Fix this by removing the state directory creation
from ConnMan and modifying ConnMan VPNd to create a state directory
named
On Sat, 2014-03-29 at 14:46 +0100, Eric BOUXIROT wrote:
Yes, your patch looks very sane to me.
cool !
how i must do to submit this patch to the connman project for the next
release ?
Add a decent commit message to your patch and send the patch to the
connman mailing list. Using git:
-
Hi,
On Fri, 2014-03-28 at 11:07 -0400, Glenn Schmottlach wrote:
After reading the docs on Sessions, I have several questions that I
don't believe were directly addressed:
1) Let's say I have two session policies:
[policy_auser]
uid = auser
AllowedBearers = wifi
On Thu, 2014-03-27 at 17:08 +0200, Patrik Flykt wrote:
Hi,
In a similar manner as ethernet networks were removed when tethering
was enabled, remove also networks when the gadget technology tethering
is enabled.
Glenn, could you verify that this patch removes gadget services when
Hi,
On Tue, 2014-03-25 at 17:04 +0100, Michael Trimarchi wrote:
This is just an idea to add setFreq to access point. Can
you comment on it?
A patch like this certainly sets the frequency. Exposing a frequency
setting API has not been provided so far, as the user is hardly ever
On Thu, 2014-03-27 at 14:31 -0400, Glenn Schmottlach wrote:
Just be to clear, the behavior you are expecting to see is that all
Gadget related services disappear when tethering is enabled,
correct?
Yes.
I haven't used the offered tethering feature before (e.g. my
patches allow ConnMan to
Hi,
This patch fails compilation with:
GENsrc/builtin.h
make --no-print-directory all-am
CC tools/iptables-test.o
CC src/log.o
CC src/dbus.o
CC src/ippool.o
CC unit/test-ippool.o
CC tools/session-test.o
CC tools/session-utils.o
CC
Remove connman_device_disconnect_service() as it is not used anymore.
---
include/device.h | 1 -
src/device.c | 17 -
2 files changed, 18 deletions(-)
diff --git a/include/device.h b/include/device.h
index 4c014d1..6c93645 100644
--- a/include/device.h
+++
When tethering is enabled, remove the network associated with
this gadget device. Removing the network will in turn remove
the corresponding gadget service. Thus no gadget services
are visible when tethering has been enabled.
This is an identical change as commit
Hi,
On Mon, 2014-03-24 at 17:06 +0100, Simon Busch wrote:
I know there was once ago a functional test suite for connman as part of
the Meego project. By searching through the web I just found the old
meego references. Is there still a up to date and used functional test
suite for
On Thu, 2014-03-27 at 10:53 +0800, zhengguang.zh...@intel.com wrote:
From: Zhang zhengguang zhengguang.zh...@intel.com
In bluez 4.x plugin, when bluetooth tethering is enabled,
set_tethering() returns 0, while in bluez 5.x plugin,
it returns -EINPROGRESS, which is not compatiable for
If the callback stops and removes the DHCP transaction, the dhcp
structure is no longer valid when returning from the callback. Fix
this by restarting the DHCP transaction before calling the callback.
---
gdhcp/client.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git
On Wed, 2014-03-26 at 09:42 +0200, Patrik Flykt wrote:
If the callback stops and removes the DHCP transaction, the dhcp
structure is no longer valid when returning from the callback. Fix
this by restarting the DHCP transaction before calling the callback.
Applied,
Patrik
On Wed, 2014-03-26 at 15:14 +0800, Guoqiang Liu wrote:
From: Guoqiang Liu guoqiang@archermind.com
Make sure that all the features of the bridge are kept enabled as
long as needed.
Applied both patches. Fixed up commit messages also. In the future
_please_ add a cover letter to your patch
Once a Bluetooth adapter changes its Powered property, PAN networks
need to be created or removed. Factor out device and PAN creation
and removal in order to call it when the adapter changes or the
device is enabled or disabled. PAN network creation and removal
needs to be done explicitely, as the
On Mon, 2014-03-24 at 11:32 +0200, Patrik Flykt wrote:
Hi,
A potential crash has been fixed on DHCPv4 expiry.
In addition to that, network disconnection and service connection have
been updated. Network such that a missing driver disconnect function is
considered indicating
On Fri, 2014-03-21 at 12:21 -0400, Ernast Sevo wrote:
Is this script posted somewhere? It would be great to get access to
it.
Not at the moment. I'd have to dig through my laptop to find it.
The script was very simple, just grepping for 'Modified' from the save
file, printing out the value and
On Mon, 2014-03-24 at 16:19 +0200, Jukka Rissanen wrote:
Hi,
the fixes are needed when IPv4 method is set to DHCP by oFono.
In patch 1, we do not break too soon as if the oFono starts to
send the properties in different order, then we might miss the
index value if method is dhcp.
Patch
On Thu, 2014-03-20 at 10:51 +0800, Chengyi Zhao wrote:
From: Chengyi Zhao chengyix.z...@gmail.com
Based on the following routing table of the openvpn,
when disabled the VPN, ConnMan can't remove
the VPN's host routing 27.17.16.90 192.168.1.1 255.255.255.255,
in fact, ConnMan should remove
Call dhcp_release() before dhcp_invalidate(), as the latter function
frees the dhcp struct. Also do not free the structure twice.
---
src/dhcp.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/src/dhcp.c b/src/dhcp.c
index d98f42d..c1dc5ec 100644
--- a/src/dhcp.c
+++
. On clearing
the 'Error' property it is not considered good behavior to force the
service states to idle either...
Cheers,
Patrik
Patrik Flykt (5):
dhcp: Release dhcp structure before freeing it
network: A network is disconnected except on -EINPROGRESS
service: Clear up connect
If an error happens at connect time and there is still a pending call
present indicating the error has not been handled, return error. With
no pending call present, the error has already been signalled.
---
src/service.c | 6 ++
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git
If the driver does not have a disconnect function or the disconnect
function returns anything else than -EINPROGRESS, consider the
network to be disconnected.
---
src/network.c | 11 +--
1 file changed, 5 insertions(+), 6 deletions(-)
diff --git a/src/network.c b/src/network.c
index
When removing a service, let the service disconnect take the service
to the idle state. Also, when unsetting the 'Error' property, don't
force the service to idle.
---
src/service.c | 15 ++-
1 file changed, 2 insertions(+), 13 deletions(-)
diff --git a/src/service.c b/src/service.c
When connected via D-Bus, iterate only over the beginning of the list
where the services are either connected or connecting. If the interface
is not available, disconnect the other service(s) which use the
interface. Once all disconnections are progressing, bail out if any
of them was
Hi,
On Mon, 2014-03-24 at 18:20 +0800, Chengyi Zhao wrote:
ConnMan can't set the hardware clock from the current system time,
when user has modified the system time, the time is unable to be saved
after reboot.
That is correct.
Can the functionality of setting the hardware clock be
Add a check also for an IPv6 address before stating that the interface
has an invalid index or address.
---
Hi,
Adding a check also for IPv6 address. Only if neither IPv4 or IPv6 addresses
are present, consider the connection failed.
Ismail, please test,
Patrik
Hi,
On Fri, 2014-03-21 at 14:05 +0200, Tomasz Bursztyka wrote:
@@ -1093,6 +1096,9 @@ static struct connman_technology *technology_get(enum
connman_service_type type)
technology-type = type;
technology-path = g_strdup_printf(%s/technology/%s,
Hi,
On Mon, 2014-03-24 at 12:31 +, Ismail YILDIZ wrote:
I have an issue with the DHCP address retrieval.
What DHCP address retrieval? AFAIK cellular connections come configured
via oFono and should just work.
There was a peculiar thing in your log, though. How is the cellular
On Thu, 2014-03-20 at 21:45 +0400, Daniel Wagner wrote:
Hi,
Here is a bugfix which was anyoing me for a while. The first patch is
just a cleanup I found while digging into the problem.
And applied! Thanks!
Patrik
___
connman mailing list
On Thu, 2014-03-20 at 11:14 +0200, Tomasz Bursztyka wrote:
- technology_get() has a possible memory leak: if dbus registration
fails, the tech_drivers list is not freed.
- removing useless NULL pointer setting since memory is zeroed
- call technology_put in case of dbus registration failure,
On Thu, 2014-03-20 at 11:15 +0200, Tomasz Bursztyka wrote:
[service] means connman's dbus service.
Applied, thanks!
Patrik
___
connman mailing list
connman@connman.net
https://lists.connman.net/mailman/listinfo/connman
On Wed, 2014-03-19 at 10:09 +0200, Patrik Flykt wrote:
Hi,
While testing Guoqiang's patch, I discovered a few other things that
didn't work out as anticipated.
Applied, updated commit message for patch 1/1.
Patrik
___
connman
Hi,
On Thu, 2014-03-20 at 13:37 +, Ismail YILDIZ wrote:
Anybody has an idea ? Maybe a configuration between connman and
ofono ?
'connmand -n -d plugins/ofono.c' will provide the debug output for the
ofono plugin in ConnMan.
Cheers,
Patrik
On Thu, 2014-03-20 at 14:19 +, Ismail YILDIZ wrote:
Thanks you for your reply. Do you know which process launch again
connman when I kill it ?
Depends on your system and how it's configured. Perhaps systemd, if
that's in use, perhaps it can also have been configured not to restart.
In the
ARPHDR_PHONET_PIPE and ARPHRD_PPP call ipconfig newlink, thus they
need also to call ipconfig dellink.
---
src/rtnl.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/src/rtnl.c b/src/rtnl.c
index 2a594da..be1bce9 100644
--- a/src/rtnl.c
+++ b/src/rtnl.c
@@ -520,6 +520,8 @@ static void
---
plugins/ethernet.c | 2 --
1 file changed, 2 deletions(-)
diff --git a/plugins/ethernet.c b/plugins/ethernet.c
index d0efd4a..b8e52ce 100644
--- a/plugins/ethernet.c
+++ b/plugins/ethernet.c
@@ -298,8 +298,6 @@ static void eth_tech_disable_tethering(struct
connman_technology *technology,
Hi,
While testing Guoqiang's patch, I discovered a few other things that
didn't work out as anticipated.
Cheers,
Patrik
Patrik Flykt (3):
plugins: Remove existing ethernet service when enabling tethering
plugins: Ethernet index is already down after disabling tethering
Remove existing service using the already existing helper function.
---
plugins/ethernet.c | 8 ++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/plugins/ethernet.c b/plugins/ethernet.c
index 20248d4..d0efd4a 100644
--- a/plugins/ethernet.c
+++ b/plugins/ethernet.c
@@ -263,14
On Tue, 2014-03-18 at 16:10 -0400, Eduardo Abinader wrote:
fix typo on p2p_service_flush
---
test/p2p-on-supplicant | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
And applied. Thanks!
Patrik
___
connman mailing list
Hi,
On Tue, 2014-03-18 at 12:00 +0800, Guoqiang Liu wrote:
From: Guoqiang Liu guoqiang@archermind.com
ConnMan use a volatile variable tethering_enabled to count
the tether bridge users. and only destroy the tether bridge
when tethering_enabled is 0.
If two or more
On Tue, 2014-03-18 at 13:33 +, Homam Dabis wrote:
1. If I have 2 active interfaces (ppp and ethernet) is it possible to
specify tethering between wifi and the ppp interface?
One only specifies which interfaces are used for tethering, not the
direction. The default route out from the device
On Fri, 2014-03-14 at 12:58 +0200, Patrik Flykt wrote:
Don't set up readline or listen to standard in when in non-interactive
mode. Some shells seem to send garbage when redirected, and if the
device is not a tty readline goes into a CPU consuming loop. Also
print out an error if agent
On Fri, 2014-03-14 at 13:32 +0200, Jukka Rissanen wrote:
This is related to fix at commit d7d8f8bdb7c1ddae098893e10608fcb42034a9ff
Some corner cases we missed when the code checked the error value
from __connman_service_disconnect(). So we can ignore the error value
and just proceed to connect
Hi,
On Fri, 2014-03-14 at 02:49 +, Lorn Potter wrote:
One problem is that I believe connman assumes pre-provisioned wpa
enterprise configurations. These configs are kept in a root access
only area (for good reason), so any process running as user cannot
create or edit them.
A
Don't set up readline or listen to standard in when in non-interactive
mode. Some shells seem to send garbage when redirected, and if the
device is not a tty readline goes into a CPU consuming loop. Also
print out an error if agent or vpnagent is attempted when run
non-interactively.
This is a
If you have any suggestions then great.
Homam
As this is happening in oFono, let's take the discussion to the ofono
ml.
Cheers,
Patrik
Date: Wed, 12 Mar 2014 09:59:54 +0200
From: Patrik Flykt patrik.fl...@linux.intel.com
To: connman@connman.net
Subject: Re: Unable
Hi,
On Thu, 2014-03-13 at 10:17 +0200, Stefano Mosconi wrote:
We are currently running on a somewhat old version (1.15 with some
patches of ours (https://github.com/mer-packages/connman)) and would
need to update to latest.
That would be an awsome target. Release notes are
On Thu, 2014-03-13 at 13:50 +0200, Stefano Mosconi wrote:
Autoconnecting works as a charm with the following line added
to /etc/connman/main.conf and a restart of ConnMan:
PreferredTechnologies = wifi,cellular
You mean in 1.21?
0.80
Cheers,
Patrik
Hi,
On Wed, 2014-03-12 at 16:11 +0100, Daniel Wagner wrote:
It would be awesome if we could agree on something like patch 1 to
6. The plugin doesn't need necessarily to be applied to ConnMan. This
could easily live in the pold repository.
I agreed with patches 1-6 and applied them.
Hi,
On Mon, 2014-03-10 at 16:21 +, Homam Dabis wrote:
Hi,
I am using connman 1.20, ofono 1.14 and linux 2.6.35.
I have a Huawei 3G USB dongle that I use to connect to a particular APN.
I can connect OK. However, if I remove the dongle and reinsert it, The
process fails at
On Tue, 2014-03-11 at 09:45 +0800, Chengyi Zhao wrote:
Would you like to explain what scenario a application needs to create
two sessions, and how to create them?
No idea, depends on the use case. According to the documentation in
doc/session-api.txt one could set up another session by calling
On Tue, 2014-03-11 at 10:23 +0200, Tomasz Bursztyka wrote:
The second array as a paremeter was missing.
Applied, thanks!
Patrik
___
connman mailing list
connman@connman.net
https://lists.connman.net/mailman/listinfo/connman
Hi,
On Fri, 2014-03-07 at 15:18 +1000, Aaron McCarthy wrote:
The previous data counters used unsigned 32 bit types. This is
probably adequate for most of the data fields, but for the bytes
received and bytes transmitted it means that the counters wrap after
4GB of data. Change to use
Hi,
On Fri, 2014-03-07 at 15:18 +1000, Aaron McCarthy wrote:
The net.connman.Counter interface has not been marked as experimental
but the RegisterCounter/UnregisterCounter methods of
net.connman.Manager interface have.
Indeed. Checked with Marcel and net.connman.Counter is
On Tue, 2014-03-04 at 09:42 +0100, Daniel Wagner wrote:
The attached patch changes it to use 64 bit data. Currently it makes
no attempt to maintain compatibility as both the format of the data
file and the DBus API changes. I see little advantage in maintaining
compatibility with the data
On Mon, 2014-03-03 at 08:44 +0100, Daniel Wagner wrote:
service_state_changed gets called for READY and ONLINE state. struct
connman_service_info *info needs only to be updated and not created on
the second call (transition from READY - ONLINE).
Applied, thanks!
Patrik
On Mon, 2014-03-03 at 16:22 +0100, Daniel Wagner wrote:
When something goes wrong during session creation, we end up in
cleanup_routing_table() and try to remove a non existing routing
table. Current kernel version seem to get very confused by it and
remove the built in and default tables
On Mon, 2014-03-03 at 13:47 +0100, Daniel Wagner wrote:
That is debatable if you ask me. Marcel suggested to go with state
less behavior which means applications need to keep track of the
states. e.g. it wants to stay online, it needs to call Connect() again
and again. That is also pretty bad.
Hi,
On Tue, 2014-03-04 at 09:02 +, Maneesh Jain wrote:
Description: This patch fix the sync-up issue between wpa_supplicant
and connman and allow connman to receive the WPA failure event.
Signed-off-by: Maneesh Jain maneesh.j...@samsung.com
Applied your patch, thanks!
In the
On Fri, 2014-02-28 at 14:56 +0200, Tomasz Bursztyka wrote:
Discovers, add and delete services methods. Again, loosely copy/pasted
against wpa_cli tool.
Some changes were needed in the checkarg() decorator so it allows non
fixed amount of arguments etc...
Applied, thanks!
Patrik
On Fri, 2014-02-28 at 16:22 +0200, Tomasz Bursztyka wrote:
If available, the p2p technology will be instanciated. Note that it's
fully dependant over wifi technology: wifi is the parent technology of
p2p.
If present, it will be possible to run a p2p peers discovery through
p2p technology's
Hi,
On Fri, 2014-02-28 at 16:10 +0100, Daniel Wagner wrote:
BTW, the wifi bearer gets connected altough the ethernet was still
online when I issue a Connect().
So if AllowedBearers is set to [ ethernet, wifi ], in free-ride mode the
session is notified if either is connected. Is it
Hi,
On Mon, 2014-03-03 at 09:27 +0100, Daniel Wagner wrote:
Is it expected that issuing
a Connect on the already connected session connects everything in
AllowedBearers or is it assumed that other bearers, in this case
wifi, stays unconnected?
IIRC we did not really discuss
Hi,
On Mon, 2014-03-03 at 08:16 +0100, Daniel Wagner wrote:
Create() - Connect() - Disconnect() - Change AllowedBearers
Well, I think it would be okay to go here online again after changing
AllowedBearers in the end. It would be more consisted after all. After
creating we also go
Hi,
On Mon, 2014-03-03 at 13:47 +0100, Daniel Wagner wrote:
Should the previous Connect() be remembered and ConnMan told to start
connecting the new beares?
That is debatable if you ask me. Marcel suggested to go with state less
behavior which means applications need to keep
Hi,
On Thu, 2014-02-27 at 15:18 +0100, Daniel Wagner wrote:
Last time we agreed on that both function are only
indication to ConnMan.
I suppose that was indication as in running an autoconnect, with the end
result that it's not guaranteed that a service becomes connected always.
The
On the first call to session Connect update service autoconnection
with the desired bearer types. On consecutive calls only run
autoconnect. When disconnecting, remove the bearer types from
autoconnect and get a disconnected signal sent to the session.
---
src/session.c | 23
(patch 02). Patch 03 is a trivial bug fix and patch 04
allows connmanctl to issue multiple connects and disconnects. Finally
patch 05 updates the session only when needed and not always when its
deactivated.
Cheers,
Patrik
Patrik Flykt (5):
session: Allow Session Connect and Disconnect
When a configuration or session is changed, remove all session
bearers from autoconnect.
---
src/session.c | 22 --
1 file changed, 12 insertions(+), 10 deletions(-)
diff --git a/src/session.c b/src/session.c
index f9e1c6e..c245bbd 100644
--- a/src/session.c
+++
---
client/commands.c | 11 ---
1 file changed, 11 deletions(-)
diff --git a/client/commands.c b/client/commands.c
index 6b555f7..14c92de 100644
--- a/client/commands.c
+++ b/client/commands.c
@@ -1733,11 +1733,6 @@ static int cmd_session(char *args[], int num, struct
connman_option
Don't unconditionally send updates when the session is deactivated
as it will send a disconnect and immediately a connect when the
session bearers are updated.
---
src/session.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/src/session.c b/src/session.c
index
Hi,
As we agreed that it's not very application writer friendly to expose
individual services via D-Bus or to write extensive amounts of code
implementing them in core ConnMan, the next best thing is perhaps to
keep them inside the session policy.
The following patch is approximately
Hi,
On Fri, 2014-02-28 at 15:21 +0100, Daniel Wagner wrote:
Hi Patrik,
On 02/28/2014 12:09 PM, Patrik Flykt wrote:
On the first call to session Connect update service autoconnection
with the desired bearer types. On consecutive calls only run
autoconnect. When disconnecting
On Fri, 2014-02-28 at 15:30 +0100, Daniel Wagner wrote:
On 02/28/2014 12:09 PM, Patrik Flykt wrote:
When a configuration or session is changed, remove all session
bearers from autoconnect.
Okay, but why?
When removing and active session we know the bearers have been supplied
When a session is created, session_activate() is called to set up the
initial session state. As the session is not yet connected, don't
update the service autoconnect with information regarding the newly
created session. When the session is disconnected or has been removed,
remember to clean up
801 - 900 of 2960 matches
Mail list logo
