Re: [FSO M5] - Answering a Call

2009-02-18 Thread SCarlson

Thanks, I will apply/test the patch and also file a ticket tonight. (eastern
standard)

-SCarlson

On Tue, Feb 17, 2009 at 7:41 PM, Michael 'Mickey' Lauer (via Nabble) 
ml-user+1966-127261...@n2.nabble.comml-user%2b1966-127261...@n2.nabble.com
 wrote:

 Am Tuesday 17 February 2009 20:46:20 schrieb SCarlson:
   Does anyone know a good fix for the answer call delay.. When I get a
  call, I hit the answer button, and it takes approx. 3-5 seconds for the
  phone to actually pick the call up??? What the heck is it doing? heh...
 Has
  anyone tweaked around with this yet?

 This is a (temporary) result of fixing bug FSO #205 as described in
 http://trac.freesmartphone.org/ticket/205

 At the moment, we serialize all dbus signals initiated by oeventsd, meaning
 we
 wait until the result has been parsed, before we send the next one. In
 particular, this means we probably wait until the ring tone has stopped
 before actually taking the call and switching the audio scenario. Thanks
 to
 gstreamer and its audio buffers, this can take up to 3 seconds.

 This behaviour is slightly suboptimal for calls to different subsystems,
 which
 do not need to be serialized.

 Please open a ticket with this mail, so we do not forget the issue and I'll

 try to come up with a speed-up.

 In the meantime, please check whether applying the attached patch improves
 it
 for you.

 Cheers,
 --
 :M:

 [attachment removed]
 ___
 Openmoko community mailing list
 commun...@...http://n2.nabble.com/user/SendEmail.jtp?type=nodenode=2344225i=0
 http://lists.openmoko.org/mailman/listinfo/community


 --
  This email is a reply to your post @
 http://n2.nabble.com/-FSO-M5Answering-a-Call-tp2342862p2344225.html
 You can reply by email or by visting the link above.



-- 
View this message in context: 
http://n2.nabble.com/-FSO-M5Answering-a-Call-tp2342862p2346805.html
Sent from the Openmoko Community mailing list archive at Nabble.com.


___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community


Re: [FSO M5] - Answering a Call

2009-02-18 Thread Helge Hafting
Michael 'Mickey' Lauer wrote:
 Am Tuesday 17 February 2009 20:46:20 schrieb SCarlson:
  Does anyone know a good fix for the answer call delay.. When I get a
 call, I hit the answer button, and it takes approx. 3-5 seconds for the
 phone to actually pick the call up??? What the heck is it doing? heh... Has
 anyone tweaked around with this yet?
 
 This is a (temporary) result of fixing bug FSO #205 as described in 
 http://trac.freesmartphone.org/ticket/205
 
 At the moment, we serialize all dbus signals initiated by oeventsd, meaning 
 we 
 wait until the result has been parsed, before we send the next one. In 
 particular, this means we probably wait until the ring tone has stopped 
 before actually taking the call and switching the audio scenario. Thanks to 
 gstreamer and its audio buffers, this can take up to 3 seconds.

Not taking the call until the ringtone stops is actually a good idea, 
the caller is probably not interested in hearing the last parts of the 
ringtone.

Instead, there is a need for a ringtone player that can be switched
off in 0.02s or so. A standalone app can be turned off with kill -9.
If using a library, switch to something that support real-time sound 
manipulation.

dbus can run with high priority, so it doesn't become a bottleneck.

Helge Hafting

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community


Re: [FSO M5] - Answering a Call

2009-02-18 Thread Michael 'Mickey' Lauer
Am Wednesday 18 February 2009 16:20:43 schrieb Helge Hafting:
 Not taking the call until the ringtone stops is actually a good idea,
 the caller is probably not interested in hearing the last parts of the
 ringtone.

Hehe, true, although on the Neo hardware this won't be a problem since 
switching the audio scenarios to handset turns off the stereoout.

 Instead, there is a need for a ringtone player that can be switched
 off in 0.02s or so.

Yes, the Neo is a bit on the too slow side for the gstreamer infrastructure. I 
guess we might look into something that just dumps out a raw file 
via /dev/adsp.

 dbus can run with high priority, so it doesn't become a bottleneck.

Sure thing, eventually we will have different one process per framework 
subsystem which will have different task priorities anyways.

Cheers,
-- 
:M:

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community


Re: [FSO M5] - Answering a Call

2009-02-18 Thread Esben Stien
Michael 'Mickey' Lauer mic...@openmoko.org writes:

 guess we might look into something that just dumps out a raw file 
 via /dev/adsp.

A better solution is using something like supercollider, a synth, to
generate the ring tones and every other sound we need. A synth should
also be an integral part of any desktop system. A programmable synth
can also adapt to every situation, which a static file cannot.

-- 
Esben Stien is b...@e s  a 
 http://www. s tn m
  irc://irc.  b  -  i  .   e/%23contact
   sip:b0ef@   e e 
   jid:b0ef@n n

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community


Re: [FSO M5] - Answering a Call

2009-02-18 Thread Michal Brzozowski
2009/2/17 SCarlson scottrcarl...@gmail.com



  Hello All --

  Does anyone know a good fix for the answer call delay.. When I get a
 call, I hit the answer button, and it takes approx. 3-5 seconds for the
 phone to actually pick the call up??? What the heck is it doing? heh... Has
 anyone tweaked around with this yet?

 -Scott


I noticed this on Om2008.12 and others, so not really specific to FSO.
___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community


[FSO M5] - Answering a Call

2009-02-17 Thread SCarlson


 Hello All --

 Does anyone know a good fix for the answer call delay.. When I get a
call, I hit the answer button, and it takes approx. 3-5 seconds for the
phone to actually pick the call up??? What the heck is it doing? heh... Has
anyone tweaked around with this yet?  

-Scott
-- 
View this message in context: 
http://n2.nabble.com/-FSO-M5Answering-a-Call-tp2342862p2342862.html
Sent from the Openmoko Community mailing list archive at Nabble.com.


___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community


Re: [FSO M5] - Answering a Call

2009-02-17 Thread ivvmm
SCarlson wrote:
 
  Hello All --
 
  Does anyone know a good fix for the answer call delay.. When I get a
 call, I hit the answer button, and it takes approx. 3-5 seconds for the
 phone to actually pick the call up??? What the heck is it doing? heh... Has
 anyone tweaked around with this yet?  
 
 -Scott

My idea is that dbus messages are still just too slow, and it takes time
for them to do that task.



signature.asc
Description: OpenPGP digital signature
___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community


Re: [FSO M5] - Answering a Call

2009-02-17 Thread Michael 'Mickey' Lauer
Am Tuesday 17 February 2009 20:46:20 schrieb SCarlson:
  Does anyone know a good fix for the answer call delay.. When I get a
 call, I hit the answer button, and it takes approx. 3-5 seconds for the
 phone to actually pick the call up??? What the heck is it doing? heh... Has
 anyone tweaked around with this yet?

This is a (temporary) result of fixing bug FSO #205 as described in 
http://trac.freesmartphone.org/ticket/205

At the moment, we serialize all dbus signals initiated by oeventsd, meaning we 
wait until the result has been parsed, before we send the next one. In 
particular, this means we probably wait until the ring tone has stopped 
before actually taking the call and switching the audio scenario. Thanks to 
gstreamer and its audio buffers, this can take up to 3 seconds.

This behaviour is slightly suboptimal for calls to different subsystems, which 
do not need to be serialized.

Please open a ticket with this mail, so we do not forget the issue and I'll 
try to come up with a speed-up.

In the meantime, please check whether applying the attached patch improves it 
for you.

Cheers,
-- 
:M:
diff --git a/framework/subsystems/oeventsd/fso_actions.py b/framework/subsystems/oeventsd/fso_actions.py
index a3a2de4..c20e859 100644
--- a/framework/subsystems/oeventsd/fso_actions.py
+++ b/framework/subsystems/oeventsd/fso_actions.py
@@ -13,13 +13,13 @@ Module: fso_actions
 
 
 
-__VERSION__ = 0.4.2
+__VERSION__ = 0.4.3
 MODULE_NAME = oeventsd
 
 import framework.patterns.tasklet as tasklet
 
 from action import Action
-from action import QueuedDBusAction as DBusAction
+from action import QueuedDBusAction, DBusAction
 from framework.controller import Controller
 from framework.config import installprefix
 
@@ -75,7 +75,7 @@ class AudioAction(Action):
 self.length = length
 
 def trigger(self, **kargs):
-DBusAction(
+QueuedDBusAction(
 dbus.SystemBus(),
 'org.freesmartphone.odeviced',
 '/org/freesmartphone/Device/Audio',
@@ -83,7 +83,7 @@ class AudioAction(Action):
 'PlaySound', self.path, self.loop, self.length).trigger()
 
 def untrigger(self, **kargs):
-DBusAction(
+QueuedDBusAction(
 dbus.SystemBus(),
 'org.freesmartphone.odeviced',
 '/org/freesmartphone/Device/Audio',
___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community