suspend / resume
Hi, I am using fso on debian. If I suspend the freerunner with apm -s no problem. But if the gsm is active then the freerunner will wakeup about in a minute ... without I am receiving a call or sms. Any idea? Where can I find the event that wake up the phone? Any log? ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: suspend / resume
On Mon, Mar 09, 2009 at 06:58:56PM +0100, Davide wrote: > Hi, > > I am using fso on debian. > > If I suspend the freerunner with apm -s no problem. > > But if the gsm is active then the freerunner will wakeup > about in a minute ... without I am receiving a call or sms. > > Any idea? > > Where can I find the event that wake up the phone? Any log? Hi, 1. You will have to suspend the GSM modem. It would be the easiest to do this via DBus. Here is a small Python script, which will suspend the GSM modem *and* your system: - #!/usr/bin/python import dbus, sys, os, gobject # Setup mainloop and dbus from dbus.mainloop.glib import DBusGMainLoop DBusGMainLoop(set_as_default=True) mainloop = gobject.MainLoop() bus = dbus.SystemBus() def getObject( busname, objectpath ): return bus.get_object( busname, objectpath, follow_name_owner_changes=True ) def getInterface( busname, objectpath, interface ): return dbus.Interface(getObject(busname, objectpath), interface) usageiface = getInterface( "org.freesmartphone.ousaged", "/org/freesmartphone/Usage", "org.freesmartphone.Usage" ) usageiface.Suspend() - 2. You will get the wakeup reason by: $ cat /sys/class/i2c-adapter/i2c-0/0-0073/neo1973-resume.0/resume_reason | grep "*" In your case it should be "EINT01_GSM" 3. Please be aware of the fact, that the current (2.6.28) debian kernel will resume with an WSOD, because some of Openmoko's recent patches are missing. -- Sebastian Reichel signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: suspend / resume
YES!!! YOU ARE RIGHT!!! With fso suspend all ok! both system and gsm are suspended! And when a call comes the phone will wake up! 2009/3/9 Sebastian Reichel : > On Mon, Mar 09, 2009 at 06:58:56PM +0100, Davide wrote: >> Hi, >> >> I am using fso on debian. >> >> If I suspend the freerunner with apm -s no problem. >> >> But if the gsm is active then the freerunner will wakeup >> about in a minute ... without I am receiving a call or sms. >> >> Any idea? >> >> Where can I find the event that wake up the phone? Any log? > > Hi, > > 1. You will have to suspend the GSM modem. It would be the easiest > to do this via DBus. Here is a small Python script, which will > suspend the GSM modem *and* your system: > > - > #!/usr/bin/python > import dbus, sys, os, gobject > > # Setup mainloop and dbus > from dbus.mainloop.glib import DBusGMainLoop > DBusGMainLoop(set_as_default=True) > mainloop = gobject.MainLoop() > bus = dbus.SystemBus() > > def getObject( busname, objectpath ): > return bus.get_object( busname, objectpath, > follow_name_owner_changes=True ) > def getInterface( busname, objectpath, interface ): > return dbus.Interface(getObject(busname, objectpath), interface) > > usageiface = getInterface( "org.freesmartphone.ousaged", > "/org/freesmartphone/Usage", "org.freesmartphone.Usage" ) > > usageiface.Suspend() > - > > > 2. You will get the wakeup reason by: > $ cat /sys/class/i2c-adapter/i2c-0/0-0073/neo1973-resume.0/resume_reason | > grep "*" > In your case it should be "EINT01_GSM" > > 3. Please be aware of the fact, that the current (2.6.28) debian kernel > will resume with an WSOD, because some of Openmoko's recent patches > are missing. > > -- Sebastian Reichel > > -BEGIN PGP SIGNATURE- > Version: GnuPG v1.4.9 (GNU/Linux) > > iEYEARECAAYFAkm1YHEACgkQH0JwilpTmKhV/wCeOPc43qBc7LlULTC2LIEYduyW > EGwAn1v3YLOUHy0YDAOqWnf1YezbPbRe > =KWsS > -END PGP SIGNATURE- > > ___ > Smartphones-userland mailing list > Smartphones-userland@linuxtogo.org > http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland > > ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: suspend / resume
Sebastian Reichel writes: > to do this via DBus. Here is a small Python script, which will I use mdbus -s org.freesmartphone.ogsmd /org/freesmartphone/GSM/Device org.freesmartphone.Resource.Suspend sleep 4 apm -s sleep 4 mdbus -s org.freesmartphone.ogsmd /org/freesmartphone/GSM/Device org.freesmartphone.Resource.Resume in my suspend script. > 3. Please be aware of the fact, that the current (2.6.28) debian kernel > will resume with an WSOD, because some of Openmoko's recent patches > are missing. Which ones? I am running recent andy-tracking and I need the "sleep 4" there or Xglamo will hit http://docs.openmoko.org/trac/ticket/1315 ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: suspend / resume
On Mon, Mar 09, 2009 at 09:39:38PM +0200, Timo Juhani Lindfors wrote: > > 3. Please be aware of the fact, that the current (2.6.28) debian kernel > > will resume with an WSOD, because some of Openmoko's recent patches > > are missing. > > Which ones? I am running recent andy-tracking and I need the "sleep 4" > there or Xglamo will hit http://docs.openmoko.org/trac/ticket/1315 Well - I think I tracked my WSOD down to the fact, that it always happens when I suspend while the display is disabled. I'm also using qi. This sounds like [1], so I hope it will be fixed when the 2.6.29 Debian kernel arrives :) -- Sebastian [1] http://git.openmoko.org/?p=kernel.git;a=commit;h=83cf37799009f6a3db4fef141c798c876124954f signature.asc Description: Digital signature ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
HTC Dream Suspend/Resume (was: HTC Dream status ?)
More updates here... > The truth is, obviously in our 2.6.32 for some reason that is unknown, > we don't get the "freezing tasks... CPU halted" messages, which probably > means the earlysuspend (android invention :/) or the wakelocks (android > invention :/) are giving us problems / aborting the suspend / whatever. Once I removed the usb vbus wakelock in htc_battery.c, the freezing tasks started to appear (which is good). The bad thing is it is then stuck in a suspend/resume loop -- it immediately resumes after suspend. (Yes, I even removed all wakelocks from the defconfig to make sure they don't interfere). I'm lucky I got so far, but now I really need some help from kernel people. -- :M: ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
make suspend/resume not kill ssh connections with wlan
Hi, I realized that automatic suspend would be much more useful if it did not kill my ssh connections over wlan. It seems that kernel does not restore the state of wlan networking on resume. It seems I can do this in userland. I added a few lines to a suspend hook: essid="$(iwconfig eth0 | tr ' ' '\n' | grep ESSID | cut -d':' -f2|sed 's/^.//;s/.$//;')" ap="$(iwconfig eth0 | grep 'Access Point'|cut -d':' -f4-|cut -d' ' -f2)" ip="$(ip addr show eth0 | grep inet | awk '{print $2}'|cut -d'/' -f1)" ip route | grep eth0 > /tmp/eth0-routes and a few lines to resume hook: iwconfig eth0 essid "$essid" iwconfig eth0 ap "$ap" ifconfig eth0 "$ip" cat /tmp/eth0-routes | while read route; do ip route replace $route done and now my ssh connections don't die if I suspend the phone for a while. Is there some already existing software that should do this? Should kernel restore these parameters? -Timo ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: HTC Dream Suspend/Resume (was: HTC Dream status ?)
Hello, Selon Michael 'Mickey' Lauer : > More updates here... > > > The truth is, obviously in our 2.6.32 for some reason that is unknown, > > we don't get the "freezing tasks... CPU halted" messages, which probably > > means the earlysuspend (android invention :/) or the wakelocks (android > > invention :/) are giving us problems / aborting the suspend / whatever. > > Once I removed the usb vbus wakelock in htc_battery.c, the freezing > tasks started to appear (which is good). The bad thing is it is then > stuck in a suspend/resume loop -- it immediately resumes after suspend. > (Yes, I even removed all wakelocks from the defconfig to make sure they > don't interfere). I have only removed the wake_lock from kernel config and can see the suspend/resume operation, but it also resumes after suspend saying : msm_sleep(): mode 0 delay 0 idle 0 msm_sleep(): enter A11S_CLK_SLEEP_EN 0, A11S_PWRDOWN 1, smsm_get_state 10ca9 msm_sleep(): vector ea0e 35392e30 -> e51ff004 10030a28 msm_pm_collapse(): returned 0 but this last line seems to say the the suspend was not done (according to pm.c it should be 1). However this part is in assembler and I do not understand how the return value is choosen. Moreover on openezx there was a register giving the argument of resume. I have not found it. By the way, I do not succeed in making wiki work : I do not see the loading in /sys/class/firmware/ and in dmesg, I see a lot of mmc0 error. (I use the recent gnu-conform kernel and an "old" rootfs) Thanks, Alain ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: HTC Dream Suspend/Resume (was: HTC Dream status ?)
Hello, sorry partial answer to myself about wifi (nothing about suspend) > > By the way, I do not succeed in making wiki work : I do not see the loading > in > /sys/class/firmware/ and in dmesg, I see a lot of mmc0 error. > the bug was in the rootfs (build some times ago from OE) : in /etc/udev there was two rules for firmware.sh With one rules, it works (with some kernel oops) only loading modules. Thanks, alain ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: HTC Dream Suspend/Resume (was: HTC Dream status ?)
Hi Alain, > the bug was in the rootfs (build some times ago from OE) : in > /etc/udev there was two rules for firmware.sh > With one rules, it works (with some kernel oops) only loading modules. It's not an OOPS, just a WARNING. For what it's worth, fsodeviced has now a firmwareloader plugin as well, so if you don't want udev, you can use fsodevice and enable the loader with [fsodevice.kernel26_firmwareloader] in fsodeviced.conf. Ok, back to Wifi, it's good and bad now. It works fine until you ifconfig wlan0 down, then it crashes and will subsequently lead to a full hang (can't reboot or turn off the device until you take the battery out). Suspend/Resume is still under investigation, GNUtoo (Denis) is working on it. If you don't mind, we talk a lot about that on #htc-linux and #openmoko-cdevel, feel invited to join IRC, if you want. Cheers, -- :M: ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: HTC Dream Suspend/Resume (was: HTC Dream status ?)
Hello, Selon Michael 'Mickey' Lauer : > Hi Alain, > > > the bug was in the rootfs (build some times ago from OE) : in > > /etc/udev there was two rules for firmware.sh > > With one rules, it works (with some kernel oops) only loading modules. > > It's not an OOPS, just a WARNING. For what it's worth, fsodeviced has > now a firmwareloader plugin as well, so if you don't want udev, you can > use fsodevice and enable the loader with > > [fsodevice.kernel26_firmwareloader] in fsodeviced.conf. Ok, I still use an old image. I will try do build an recent fso2/zhone2. > > Ok, back to Wifi, it's good and bad now. It works fine until you > ifconfig wlan0 down, then it crashes and will subsequently lead to a > full hang (can't reboot or turn off the device until you take the > battery out). I succeed in logging with ssh and do a /sbin/halt (which by the way does not halt the phone...) > > Suspend/Resume is still under investigation, GNUtoo (Denis) is working > on it. If you don't mind, we talk a lot about that on #htc-linux and > #openmoko-cdevel, feel invited to join IRC, if you want. > This morning I have read the last 3 days of irc-log of #htc-linux. I'm connected also since 2 hours. Thanks, Alain ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: make suspend/resume not kill ssh connections with wlan
Hi, Am Freitag, den 23.10.2009, 13:03 +0300 schrieb Timo Juhani Lindfors: > Is there some already existing software that should do this? > Should kernel restore these parameters? what if you move between WLANs while suspending? I’d say a program like wpasupplicant should handle this intelligently, e.g. after resume check if the same WLAN is in range and then quickly re-connect, and otherwise just look for new WLANs. But I’m not sure if it does this already. At least on my laptop, I often have to re-load the WLAN driver after suspend anyways, to get back in quickly. Greetings, Joachim -- Joachim "nomeata" Breitner mail: m...@joachim-breitner.de | ICQ# 74513189 | GPG-Key: 4743206C JID: nome...@joachim-breitner.de | http://www.joachim-breitner.de/ Debian Developer: nome...@debian.org signature.asc Description: Dies ist ein digital signierter Nachrichtenteil ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: make suspend/resume not kill ssh connections with wlan
Joachim Breitner writes: > what if you move between WLANs while suspending? Good point. I guess I could add logic to use "iwlist" to check if the $ap is still within range? > I’d say a program like wpasupplicant should handle this intelligently, > e.g. after resume check if the same WLAN is in range and then quickly > re-connect, and otherwise just look for new WLANs. But I’m not sure if But if "re-connect" means new DHCP query then the IP could change? -Timo ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland
Re: make suspend/resume not kill ssh connections with wlan
> But if "re-connect" means new DHCP query then the IP could change? Yes, tho many DHCP clients are clever enough to send the previously used IP to the DHCP server, so the server can give them the same IP as last time (and some DHCP servers try to do that on their own based on the mac address). I.e. if your IP changes every time you connect, maybe there's a way to address this directly by changing your DHCP server or client. Stefan ___ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland