On Oct 21, 2015, at 8:40 AM, connman-requ...@connman.net wrote:
> From: Marcel Holtmann
> Subject: Re: [PATCH] gdbus: move typedefs for interwork with strict compilers
> Date: October 21, 2015 at 3:35:48 AM PDT
> To: connman@connman.net
> Reply-To: connman@connman.net
>
> Hi
Move enumeration type defintions AFTER the enumerations themselves are declared
and defined such that the header works with strict compilers.
---
gdbus/gdbus.h | 9 +
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/gdbus/gdbus.h b/gdbus/gdbus.h
index d99c254..7a6a18f 100644
On Oct 13, 2015, at 11:01 PM, Patrik Flykt wrote:
> On Tue, 2015-10-13 at 14:42 -0700, Naveen Singh wrote:
>
>> There is nothing that application knows that connman does not know.
>> In fact application gets to know through connman that connection did
>> not go
On Jul 8, 2014, at 5:39 AM, Tomasz Bursztyka tomasz.burszt...@linux.intel.com
wrote:
Thanks Grant for the reply. The config option sounds right to me as a
general solution. Connman folks, thoughts?
I don't think we will ever put this PassPhrase property back again.
And we don't do #ifdef
On Jun 27, 2014, at 12:00 PM, Jukka Rissanen jukka.rissa...@linux.intel.com
wrote:
On to, 2014-06-26 at 16:52 -0700, Drew Stebbins wrote:
Hi folks,
I'm working on a project where devices on a WiFi network must occasionally
provide an application with the WiFi network's security credentials.
On Nov 4, 2013, at 12:20 AM, Patrik Flykt wrote:
On Fri, 2013-11-01 at 13:54 -0700, Andrew LeCain wrote:
We do indeed have a sleep manager that will put our system to deep
sleep/suspend. This sleep manager also uses the RTC to schedule a wake
for
future tasks. However, the wifi chip we use
Patrik:
I am debugging an issue in which there are reports of not being able to connect
to or see a particular SSID scanned (let me know if you want specific details)
for an extended period (several hours).
I've been able to correlate them back to this check-in:
commit
On Thu May 16 05:32:29 PDT 2013, Tomasz Bursztyka
tomasz.burszt...@linux.intel.com wrote:
Currently a scan will only scan the first found stored hidden network.
This patch fixes it: it will scan all, taking into account the limit of
scan parameters the driver can take.
---
@@ -548,32 +640,21
Per MEEGO-25999 https://bugs.meego.com/show_bug.cgi?id=25999, there
exists a race between scan_callback and wifi_disable such that a care-
fully-timed deassertion of the technology Powered property between
throw_wifi_scan and scan_callback will lead to a device reference
count underflow. A second
This terminates the body of stop_autoscan if autoscan is
not active. Per MEEGO-25987 http://bugs.meego.com/show_bug.cgi?id=25987,
running stop_autoscan when changing connected services can prematurely and
incorrectly terminate a new, in-flight service connection and result in a
failure of that
On Dec 27, 2012, at 9:54 AM, Marcel Holtmann mar...@holtmann.org wrote:
Is there any reason why a Gateway property is missing from the
service IPv6 dictionary? Is this assumed to be implicit from the
address prefix and therefore unnecessary?
it is documented in the API. If it is not returned
I've been chasing down for the past few days an issue in which I thought the
kernel, but it now appears to be connman, was periodically and incorrectly
disabling IPv6 support. For those interested in following the thread, more
at:
https://bugs.meego.com/show_bug.cgi?id=25929
Best,
Grant
Is there any reason why a Gateway property is missing from the service
IPv6 dictionary? Is this assumed to be implicit from the address prefix
and therefore unnecessary?
-Grant
___
connman mailing list
connman@connman.net
In order to allow clients to update the property in a manner similar to
other service properties, signal changes in the service Error property
key/value pair.
The value is used when no error is asserted because D-Bus property
changed convention has no semantics for indicating a key/value pair
This patch adds support for specifying a main configuration file
different from the default value of /etc/connman/main.conf. This
is valuable during debug to specify alternate debug-only
configurations.
---
v1: Initial submission.
v2: Incorporated feedback from Marcel about not propagating the
This patch adds support for specifying a main configuration file
different from the default value of /etc/connman/main.conf. This
is valuable during debug to specify alternate debug-only
configurations.
---
v1: Initial submission.
v2: Incorporated feedback from Marcel about not propagating the
On Aug 6, 2012, at 9:07 AM, Marcel Holtmann wrote:
@@ -350,6 +365,9 @@ static GOptionEntry options[] = {
Specify debug options to enable, DEBUG },
{ device, 'i', 0, G_OPTION_ARG_STRING, option_device,
Specify networking device or
In order to allow clients to update the property in a manner similar to
other service properties, signal changes in the service Error property
key/value pair.
The value none is used when no error is asserted because D-Bus property
changed convention has no semantics for indicating a key/value
In order to allow clients to update the property in a manner similar to
other service properties, signal changes in the service Error property
key/value pair.
The value none is used when no error is asserted because D-Bus property
changed convention has no semantics for indicating a key/value
This patch adds support for specifying a main configuration file
different from the default value of /etc/connman/main.conf. This
is valuable during debug to specify alternate debug-only
configurations.
---
Makefile.am |3 ++-
src/main.c | 23 ++-
2 files changed, 20
This patch adds support for specifying a main configuration file
different from the default value of /etc/connman/main.conf. This
is valuable during debug to specify alternate debug-only
configurations.
---
v1: Initial submission.
v2: Incorporated feedback from Marcel about not propagating the
---
src/ntp.c | 30 ++
1 file changed, 30 insertions(+)
diff --git a/src/ntp.c b/src/ntp.c
index 8c31fdb..7555618 100644
--- a/src/ntp.c
+++ b/src/ntp.c
@@ -69,6 +69,36 @@ struct ntp_msg {
#define LOGTOD(a) ((a) 0 ? 1. / (1L -(a)) : 1L (int)(a))
+#define
intervals before giving up and moving on to the next
server.
2) The resolver will not allow unsuccessful A or response status to
override successful status when performing a wildcard lookup
(i.e. the address family is unspecified).
Grant Erickson (13):
ntp: add mnemonics
Convert a magic number into equivalent mnemonics for the flags (leap,
version, mode) packet field.
---
src/ntp.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/ntp.c b/src/ntp.c
index 7555618..c7224a8 100644
--- a/src/ntp.c
+++ b/src/ntp.c
@@ -114,7 +114,7 @@ static
This patch sets the non-in-sync flag of the leap subfield of the
packet flags field. This reflects behavior implemented by ntpdate or
ntpd in one-shot mode.
---
src/ntp.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/ntp.c b/src/ntp.c
index 1916a72..7ef072c 100644
---
Use mnemonics when checking reply flags and print out information in
debug mode about why the reply flags were rejected.
---
src/ntp.c | 14 +-
1 file changed, 13 insertions(+), 1 deletion(-)
diff --git a/src/ntp.c b/src/ntp.c
index 7ef072c..e4267ea 100644
--- a/src/ntp.c
+++
---
src/ntp.c |7 +++
1 file changed, 7 insertions(+)
diff --git a/src/ntp.c b/src/ntp.c
index e4267ea..3e90741 100644
--- a/src/ntp.c
+++ b/src/ntp.c
@@ -99,6 +99,13 @@ struct ntp_msg {
#define NTP_FLAGS_VN_DECODE(flags)((uint8_t)(((flags)
NTP_FLAG_VN_SHIFT) NTP_FLAG_VN_MASK))
This patch uses a mnemonic to specify the request precision and also adds a
comment how a future implementation might pass along the actual system
wall clock precision.
---
src/ntp.c |7 +++
1 file changed, 7 insertions(+)
diff --git a/src/ntp.c b/src/ntp.c
index 3e90741..e15bd08 100644
This patch causes the internal NTP client to retry a given server address up to
three times (four total attempts) at a two second interval before
falling back to another server name or server address.
---
src/ntp.c | 66 +
1 file
---
src/timeserver.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/timeserver.c b/src/timeserver.c
index 439d6f5..59b299d 100644
--- a/src/timeserver.c
+++ b/src/timeserver.c
@@ -129,7 +129,7 @@ void __connman_timeserver_sync_next()
/* if its a IP , directly
This patch adds an analagous debug message when starting NTP against a
resolved server as when starting against a numbered server.
---
src/timeserver.c |2 ++
1 file changed, 2 insertions(+)
diff --git a/src/timeserver.c b/src/timeserver.c
index 59b299d..f934b58 100644
--- a/src/timeserver.c
In both parse_response and query_timeout there exists logic that checks to
ensure that both an A and lookups
have either been responded to or timed out before processing and returning
results to the caller. In query_timeout,
there was a typo in the condition check such that it did not match
When performing a resolver lookup from timeserver or wpad, both
perform queries with an unspecified address family. This means that
both A and record queries are issued. In cases where a valid,
successful A response comes back but where the query results in a
timeout, do not smash the
Use predefined mnemonics from arpa/nameserv.h for the return value
from ns_msg_getflag rather than magic numbers.
---
gweb/gresolv.c | 12 ++--
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/gweb/gresolv.c b/gweb/gresolv.c
index d96f345..79abc9b 100644
---
In both parse_response and query_timeout there exists logic that checks to
ensure that both an A and lookups
have either been responded to or timed out before processing and returning
results to the caller. In query_timeout,
there was a typo in the condition check such that it did not match
When performing a resolver lookup from timeserver or wpad, both
perform queries with an unspecified address family. This means that
both A and record queries are issued. In cases where a valid,
successful A response comes back but where the query results in a
timeout, do not smash the
Use predefined mnemonics from arpa/nameserv.h for the return value
from ns_msg_getflag rather than magic numbers.
---
gweb/gresolv.c | 12 ++--
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/gweb/gresolv.c b/gweb/gresolv.c
index d96f345..79abc9b 100644
---
---
src/ntp.c | 30 ++
1 file changed, 30 insertions(+)
diff --git a/src/ntp.c b/src/ntp.c
index 91a8832..87017ed 100644
--- a/src/ntp.c
+++ b/src/ntp.c
@@ -69,6 +69,36 @@ struct ntp_msg {
#define LOGTOD(a) ((a) 0 ? 1. / (1L -(a)) : 1L (int)(a))
+#define
---
v1: Initial submission.
v2: Removed trailing white space.
src/ntp.c | 30 ++
1 file changed, 30 insertions(+)
diff --git a/src/ntp.c b/src/ntp.c
index 91a8832..87017ed 100644
--- a/src/ntp.c
+++ b/src/ntp.c
@@ -69,6 +69,36 @@ struct ntp_msg {
#define
Convert a magic number into equivalent mnemonics for the flags (leap,
version, mode) packet field.
---
src/ntp.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/ntp.c b/src/ntp.c
index 87017ed..f257cda 100644
--- a/src/ntp.c
+++ b/src/ntp.c
@@ -114,7 +114,7 @@ static
This patch adds two changes that make the NTP request packet sent by
connman more consistent with those sent from ntpdate or ntpd in
one-shot mode.
Grant Erickson (2):
ntp: send a transmit time value that reflects current system time
ntp: specify leap-not-in-sync in flags
--
1.7.10
From
This sets the ntp request packet transmit time to the current system
time to reflect behavior implemented by ntpdate and ntpd in one-shot
mode.
---
src/ntp.c |5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/src/ntp.c b/src/ntp.c
index f257cda..88c8b8e 100644
---
This patch sets the non-in-sync flag of the leap subfield of the
packet flags field. This reflects behavior implemented by ntpdate or
ntpd in one-shot mode.
---
src/ntp.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/ntp.c b/src/ntp.c
index 88c8b8e..623c21e 100644
Use mnemonics when checking reply flags and print out information in
debug mode about why the reply flags were rejected.
---
src/ntp.c | 14 +-
1 file changed, 13 insertions(+), 1 deletion(-)
diff --git a/src/ntp.c b/src/ntp.c
index f257cda..d646161 100644
--- a/src/ntp.c
+++
Use mnemonics when checking reply flags and print out information in
debug mode about why the reply flags were rejected.
---
v1: Initial submission.
v2: Fixed typo that prevented compilation.
src/ntp.c | 14 +-
1 file changed, 13 insertions(+), 1 deletion(-)
diff --git
Use mnemonics when checking reply flags and print out information in
debug mode about why the reply flags were rejected.
---
v1: Initial submission.
v2: Fixed typo that prevented compilation.
v3: More typo corrections.
src/ntp.c | 14 +-
1 file changed, 13 insertions(+), 1
This patch causes the internal NTP client to retry a given server address up to
three times (four total attempts) at a two second interval before
falling back to another server name or server address.
---
src/ntp.c | 66 +
1 file
---
src/timeserver.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/timeserver.c b/src/timeserver.c
index 467da52..3603c5a 100644
--- a/src/timeserver.c
+++ b/src/timeserver.c
@@ -123,7 +123,7 @@ void __connman_timeserver_sync_next()
/* if its a IP , directly
This patch adds an analagous debug message when starting NTP against a
resolved server as when starting against a numbered server.
---
src/timeserver.c |2 ++
1 file changed, 2 insertions(+)
diff --git a/src/timeserver.c b/src/timeserver.c
index 3603c5a..df4de3c 100644
--- a/src/timeserver.c
On Jun 27, 2012, at 10:57 PM, Patrik Flykt wrote:
On Wed, 2012-06-27 at 17:21 -0700, Grant Erickson wrote:
- char *default_auto_connect[] = {
+ static char *default_auto_connect[] = {
This isn't strictly necessary as default_auto_connect[] is only used
locally as input
On Jun 27, 2012, at 11:01 PM, Marcel Holtmann wrote:
I ran connman-1.2 through it's paces today. I noticed two significant issues
right away:
1) An empty or commented out NetworkInterfaceBlacklist setting in main.conf
does not work. Patch submitted (see
connmand[1771]: ...
connmand[1771]: src/rtnl.c:connman_rtnl_remove_watch() id 2
connmand[1771]: plugins/wifi.c:network_removed() name WiFi Activation
connmand[1771]: plugins/wifi.c:network_removed() name Rock
connmand[1771]: plugins/wifi.c:network_removed() name Scissors
connmand[1771]:
The default configuration arrays default_auto_connect and
default_blacklist must be static-qualified. Otherwise, their
stack-allocated storage will go out of scope after parse_config
is finished and empty or commented-out DefaultAutoConnectTechnologies
or NetworkInterfaceBlacklist config entries
I ran connman-1.2 through it's paces today. I noticed two significant issues
right away:
1) An empty or commented out NetworkInterfaceBlacklist setting in main.conf
does not work. Patch submitted (see
http://lists.connman.net/pipermail/connman/2012-June/010241.html).
2) Named (e.g.
Patrik Flykt patrik.flykt at linux.intel.com
On Wed Jun 13 00:36:52 PDT 2012, Patrik Flykt wrote:
On Wed, 2012-06-06 at 23:15 -0700, Grant Erickson wrote:
In order to allow clients to update the property in a manner similar
to other service properties, signal changes in the service Error
In order to allow clients to update the property in a manner similar to other
service properties, signal changes
in the service Error property key/value pair.
The value none is used when no error is asserted because D-Bus property
changed convention has no semantics
for indicating a key/value
On 6/6/12 12:30 AM, Daniel Wagner wrote:
Thanks for the logs. I have uploaded it here for the rest of us to enjoy
(it a bit too large for the ml):
http://www.monom.org/logs/wrong-wifi-pointer.log
The second log is easier to understand, but
I think the same problem bis in the first log.
---
src/service.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/service.c b/src/service.c
index de28264..d6bd870 100644
--- a/src/service.c
+++ b/src/service.c
@@ -4852,7 +4852,7 @@ int __connman_service_clear_error(struct connman_service
*service)
If a device scan request immediately returns -ELAREADY or -EINPROGRESS from
device_scan, do not filter the error and effect a
successful pending return. Instead, immediately process and return the pending
error reply.
Failure to do this results in periodic D-Bus daemon error assertions about
Over the weekend I identified a few add'l regressions in functionality in
connman-1.x relative to connman-0.7x that I am going to be diving into today:
1) Connman does not automatically reconnect to idle services following
a technology powered or offline mode assertion / deassertion
On Wed, 2012-05-23 at 15:24 +0300, patrik.fl...@linux.intel.com wrote:
After looking over the first patch once more with Jukka, we came to the
conclusion that the service should not go to failure state after the
-ENOKEY checking part. If the service is set to failure, it cannot be
---
src/ntp.c |2 ++
1 file changed, 2 insertions(+)
diff --git a/src/ntp.c b/src/ntp.c
index 9bc548d..91a8832 100644
--- a/src/ntp.c
+++ b/src/ntp.c
@@ -196,6 +196,8 @@ static void decode_msg(void *base, size_t len, struct
timeval *tv)
poll_id =
Picking up a thread from IRC, I've been working on getting agent-based service
connections working with connman-1.0. However, what I've found is that rather
than performing the following sequence of operations:
1) Issue .Service.Connect call.
2) Receive properties dictionary
On May 22, 2012, at 2:03 PM, Grant Erickson wrote:
Picking up a thread from IRC, I've been working on getting agent-based
service connections working with connman-1.0. However, what I've found is
that rather than performing the following sequence of operations:
1) Issue
This patch ensures that [device_type] Powered=true|false device enabled
/ powered state migrates from 0.76-era versions of default.profile.
---
src/storage.c | 34 +++
1 file changed, 29 insertions(+), 5 deletions(-)
diff --git a/src/storage.c b/src/storage.c
index
On 2012-05-21, Grant Erickson wrote:
Following up on an IRC discussion, I had noted on upgrading connman from
0.76+ to 1.0 that my previously-set resource limits of:
rlimit-as=5242880
rlimit-core=0
rlimit-data=3145728
rlimit-nofile=768
rlimit-stack=262144
were
This fixes raw PSK support by first converting the raw PSK from
a 64-character hexadecimal string to 32-byte array before appending
it to the network dictionary.
---
gsupplicant/supplicant.c | 42 --
1 file changed, 36 insertions(+), 6 deletions(-)
diff
This fixes raw PSK support by first converting the raw PSK from
a 64-character hexadecimal string to 32-byte array before appending
it to the network dictionary.
---
v1: Initial submission.
v2: Fixed white space.
gsupplicant/supplicant.c | 42 --
1
On May 15, 2012, at 8:09 AM, Marcel Holtmann wrote:
This fixes raw PSK support by first converting the raw PSK from
a 64-character hexadecimal string to 32-byte array before appending
it to the network dictionary.
---
v1: Initial submission.
v2: Fixed white space.
This fixes raw PSK support by first converting the raw PSK from
a 64-character hexadecimal string to 32-byte array before appending
it to the network dictionary.
---
v1: Initial submission.
v2: Fixed white space.
gsupplicant/supplicant.c | 42 --
1
This resolves several post-commit review coding style comments for raw PSK
support.
---
gsupplicant/supplicant.c | 26 ++
1 file changed, 14 insertions(+), 12 deletions(-)
diff --git a/gsupplicant/supplicant.c b/gsupplicant/supplicant.c
index 829ceb7..0ae5540 100644
Has anyone recently run a regression with connman against full, 64-character
raw WPA/WPA2 PSKs? I found the following results today:
http://pastebin.com/y73zH074
While the following non-upstreamed wpa_s patch looks relevant
On May 11, 2012, at 2:42 AM, connman-requ...@connman.net wrote:
From: Marcel Holtmann mar...@holtmann.org
Subject: Re: Wifi won't connect
Date: May 11, 2012 1:56:04 AM PDT
Hi Patrik,
Adding a global section seems to change nothing.
[global]
Reading the code in src/config.c,
On 4/25/12 8:15 AM, Ogor Jean-Jacques (AQL) wrote:
I noticed that changing Wifi security configuration (from Wep to WPA or WPA to
Wep) is not dynamically corrected handled by connman 0.78
I experienced inconsistencies when changing security protection on a fly on
wifi AP.
Jean-Jacques:
On Apr 3, 2012, at 5:59 AM, connman-requ...@connman.net wrote:
Hi Julien,
In the current Connman version the state of the manager API is
either offline or online.
If only one device is marked as connected the global state is online.
So the global online state doesn't have the same
On Apr 3, 2012, at 9:51 AM, Marcel Holtmann wrote:
In the current Connman version the state of the manager API is
either offline or online.
If only one device is marked as connected the global state is online.
So the global online state doesn't have the same signification for the
service
On 3/28/12 12:54 PM, Mario Domenech Goulart wrote:
On Wed, 28 Mar 2012 12:26:20 -0700 Grant Erickson maratho...@gmail.com
wrote:
On 3/28/12 11:35 AM, Mario Domenech Goulart wrote:
I've been facing a weird behavior on the system I'm working. I'll
try to describe it here.
I'm using a regular
Pass NULL rather than 0 when calling time(2) to follow prefered
stylistic convention.
--
diff --git a/src/dhcpv6.c b/src/dhcpv6.c
index aab5868..462b78d 100644
--- a/src/dhcpv6.c
+++ b/src/dhcpv6.c
@@ -606,7 +606,7 @@ static int check_restart(struct connman_dhcpv6 *dhcp)
: Initial submission.
v1: Break initial patch into three parts.
Grant Erickson (3)
dhcpv4-client: Refactor alloc_dhcp_option to accept string or raw data.
dhcpv4-client: Add free function to add client-id option to sent packets.
dhcpv4-client: Specify the client-id option to be added
This patch adds a RFC 1533- and 2132-compliant DHCP client-id option
(61) to DHCPv4 packets.
---
gdhcp/client.c | 50 +++---
gdhcp/gdhcp.h |2 ++
src/dhcp.c |2 ++
3 files changed, 47 insertions(+), 7 deletions(-)
diff --git
On 10/24/11 8:26 AM, Jouni Malinen wrote:
On Mon, Oct 24, 2011 at 10:05:04AM -0500, Dan Williams wrote:
I used to have a problem with this, but I now see where it makes sense
for the State property alone. Most properties *should* be coalesced to
reduce dbus traffic, but the State property
Following up on this thread:
http://lists.connman.net/pipermail/connman/2011-July/005641.html
and this one:
http://lists.shmoo.com/pipermail/hostap/2010-September/021714.html
I have further explored this and after applying this patch series:
FYI. I'm working on running down a bug that has appears to have been
with connman at least since 0.73. The issue is that a change of a PSK
passphrase on the access point and subsequent disconnection, connman
falls into a dead-end loop.
Then, when attempting to change the passphrase from the UI
Some access points incorrectly advertise WPS even when they are
configured as open (i.e. no security), so filter adding WPS to the
security array appropriately.
v2: Incorporated default vs. full-enumeration feedback.
---
src/service.c | 25 +
1 files changed, 21
On Jul 26, 2011, at 6:12 PM, Grant Erickson wrote:
Is it expected that connman should receive and process all supplicant state
changes? If so, I am seeing cases where the supplicant fails to signal or
connman misses GROUP_HANDSHAKE always and occasionally ASSOCIATED,
4WAY_HANDSHAKE
Is it expected that connman should receive and process all supplicant state
changes? If so, I am seeing cases where the supplicant fails to signal or
connman misses GROUP_HANDSHAKE always and occasionally ASSOCIATED,
4WAY_HANDSHAKE, and DISCONNECTED as below:
2011-07-26 21:34:22.00
On 7/22/11 2:32 AM, Jukka Rissanen wrote:
Issue with possible unaligned memory access is still there
as I have no hardware to check whether there are any real
problems with the code.
Jukka:
Catching issues here without the appropriate hardware might be as simple as
adding an assert or warning
On 7/18/11 1:12 AM, Marcel Holtmann wrote:
When generating symbolic links for local headers, ensure that they are
handled correctly when building from either colocated or non-colocated
source and build directories.
diff --git a/Makefile.am b/Makefile.am
index ac3999b..c96e73a 100644
---
Ensure ipconfig data and operations are cleared out when a service is
freed.
This is at least a partial fix and maybe a full fix for issues 21000
http://bugs.meego.com/show_bug.cgi?id=21000 and 21285
http://bugs.meego.com/show_bug.cgi?id=21285.
---
src/service.c |4
1 files changed, 4
When generating symbolic links for local headers, ensure that they are
handled correctly when building from either colocated or non-colocated
source and build directories.
---
Makefile.am | 13 +++--
1 files changed, 7 insertions(+), 6 deletions(-)
diff --git a/Makefile.am
In certain, as yet root-caused, circumstances connman may lose during
its operation the connman- or kernel-assigned loopback interface
address.
When this occurs and when connman is acting as a DNS proxy, all
subsequent DNS resolution requests will fail, along with other system
services dependent
Return a more descriptive -errno rather than -1 when we fail to create
a socket.
---
plugins/loopback.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/plugins/loopback.c b/plugins/loopback.c
index 6cd5b85..ba8f51a 100644
--- a/plugins/loopback.c
+++
It is possible to end up in situations in which the loopback interface
is up but has no valid address. In that case, we expect SIOCGIFADDR
will return EADDRNOTAVAIL and valid_loopback should return FALSE such
that connman can add a loopback address back to the interface.
This is a partial fix for
On 7/15/11 1:34 PM, Marcel Holtmann wrote:
In certain, as yet root-caused, circumstances connman may lose during
its operation the connman- or kernel-assigned loopback interface
address.
The set improves the error handling and reporting when there are
loopback problems to be resolved during
On Jul 8, 2011, at 1:56 PM, Grant Erickson wrote:
On Jul 7, 2011, at 10:12 PM, Grant Erickson wrote:
On Jul 7, 2011, at 9:56 PM, Grant Erickson wrote:
I checked out the latest top-of-tree Service State NG patches today and
while they improve the fidelity and accuracy of the service state
Return and pass a consistent set of informative error codes and
display a consistent set of error messages for
connman_inet_modify_address and connman_{clear,set}_*_address.
v2: Incorporate feedback from Samuel Ortiz on format and severity
of the error messages reported.
---
src/inet.c
Check for NULL dhcp pointers in dhcp_free and dhcp_release.
---
src/dhcp.c | 9
1 files changed, 8 insertions(+)
diff --git a/src/dhcp.c b/src/dhcp.c
index ab8ee5a..cddee25 100644
--- a/src/dhcp.c
+++ b/src/dhcp.c
@@ -50,6 +50,11 @@ static GHashTable *network_table;
static void
I mentioned this briefly on IRC today; however, as promised here is a log
illustrating the issue:
OK, here's an interesting case that seemed mostly fixed by
http://git.kernel.org/?p=network/connman/connman.git;a=commit;h=39825846a219b47b216d89cc036cf9640aaeead4
Reproduction steps:
1)
Given the following reproduction steps:
1) Set up an access point with a Class B IP address, say
10.2.40.254.
2) Set up the access point to provide DHCP leases over a range
10.2.40.2 through 10.2.40.253.
3) Set up a short lease renewal period of 5-10 minutes.
4)
On 5/30/11 12:50 PM, Samuel Ortiz wrote:
On Wed, May 25, 2011 at 10:10:17AM -0700, Grant Erickson wrote:
In the Connection Manager, completion of a valid IP configuration
excites the service state machine to move from the configuration to
the ready state.
However, the existing
1 - 100 of 165 matches
Mail list logo