Re: gta02, 2007.2: frequent wakeups -- creating workaround -- scripts for apm
On Fri, Aug 15, 2008 at 08:35:20PM +0200, arne anka wrote: > ok, here they are, teh scripts i use to make the baby sleep. > the fr still wkaes up from an gsm event after a few hours (6 or 10), but > still the battery lasts a lot longer now. > thanks to Stevhen**, who pointed me to the qtopia sources executing the at > commands on suspend/resume. > > now, if someone could shed some light on what the do and which values are > sensible/allowed besides 0 and 2? and if there's a way to know which event > still escapes the guard and how to catch it? With OM2008.8, at least in its current configuration, the complete communication with the GSM modem is written to the log. Maybe you could do something similar in OM2007.2 to be able to read what the modem sent when it woke up the system (if it sent something)? Regards, Thomas ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround -- scripts for apm
On Fri, Aug 15, 2008 at 03:16:58PM -0500, Steven ** wrote: > Thanks should be directed to Thomas B. <[EMAIL PROTECTED]>. He's the > one that found the relevant Qtopia source. Thanks Thomas! Glad that I could help :-) Regards, Thomas ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround
> Well, you could turn the GSM chip off via the power menu... that's not really a solution. yers ago i had i cellphone working in the 1800 band only. since my provider did not cover the entire area i was moving in sometimes the phone did not find connection and scanned wildly burning a lot of power. for exactly those situations one could disable scanning, the phone did not try to find a new tower once it lost connection, but when back in coverage it worked. > I have noticed once or twice that the signal strength doesn't report > right after a resume with these commands. It may have been because I > did a suspend while without signal and a resume with. I don't really > know... i think the comments or something i found on the net suggested that one of the commands disabled updating of signal strength, thus wrong numbers are to expect but should be ok soon (depending on the intervall) ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround -- scripts for apm
Thanks should be directed to Thomas B. <[EMAIL PROTECTED]>. He's the one that found the relevant Qtopia source. Thanks Thomas! -Steven On Fri, Aug 15, 2008 at 1:35 PM, arne anka <[EMAIL PROTECTED]> wrote: > ok, here they are, teh scripts i use to make the baby sleep. > the fr still wkaes up from an gsm event after a few hours (6 or 10), but > still the battery lasts a lot longer now. > thanks to Stevhen**, who pointed me to the qtopia sources executing the at > commands on suspend/resume. > > now, if someone could shed some light on what the do and which values are > sensible/allowed besides 0 and 2? and if there's a way to know which event > still escapes the guard and how to catch it? > > in /etc/apm/suspend.d/ > > #!/bin/sh > PATH=/usr/bin:/bin > > echo -e "AT+CREG=0" | libgsmd-tool -m atcmd 2>&1 > /dev/null > echo -e "AT+CGREG=0" | libgsmd-tool -m atcmd 2>&1 > /dev/null > echo -e "AT%CREG=0" | libgsmd-tool -m atcmd 2>&1 > /dev/null > echo -e "AT%CGREG=0" | libgsmd-tool -m atcmd 2>&1 > /dev/null > echo -e "AT%CSQ=0" | libgsmd-tool -m atcmd 2>&1 > /dev/null > > > in /etc/apm/resume.d/ > > #!/bin/sh > PATH=/usr/bin:/bin > > echo -e "AT+CREG=2" | libgsmd-tool -m atcmd 2>&1 > /dev/null > echo -e "AT+CGREG=2" | libgsmd-tool -m atcmd 2>&1 > /dev/null > echo -e "AT%CREG=2" | libgsmd-tool -m atcmd 2>&1 > /dev/null > echo -e "AT%CGREG=2" | libgsmd-tool -m atcmd 2>&1 > /dev/null > echo -e "AT%CSQ=1" | libgsmd-tool -m atcmd 2>&1 > /dev/null > > > ___ > support mailing list > support@lists.openmoko.org > https://lists.openmoko.org/mailman/listinfo/support > ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround
Well, you could turn the GSM chip off via the power menu... I haven't noticed a large power drain when without signal. It suspends just fine now and will sit in suspend consistently. Suspended for an hour uses 1% or so of the battery. Perhaps one of the GSM commands in the suspend script is disabling that search. I have noticed once or twice that the signal strength doesn't report right after a resume with these commands. It may have been because I did a suspend while without signal and a resume with. I don't really know... -Steven On Fri, Aug 15, 2008 at 1:13 PM, arne anka <[EMAIL PROTECTED]> wrote: >> it's worth mentioning that running GSM without association to a network >> probably drains battery like hell, as GSM-chipsets tend to constantly >> search >> the whole frequency range for some BS to register. I've experienced >> factor 5 >> to 10 for power consumption. > > i know, i had a phone once that allowed to disable that -- is there a way > to tell the gsm modem not to do so? > > ___ > support mailing list > support@lists.openmoko.org > https://lists.openmoko.org/mailman/listinfo/support > ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround -- scripts for apm
ok, here they are, teh scripts i use to make the baby sleep. the fr still wkaes up from an gsm event after a few hours (6 or 10), but still the battery lasts a lot longer now. thanks to Stevhen**, who pointed me to the qtopia sources executing the at commands on suspend/resume. now, if someone could shed some light on what the do and which values are sensible/allowed besides 0 and 2? and if there's a way to know which event still escapes the guard and how to catch it? in /etc/apm/suspend.d/ #!/bin/sh PATH=/usr/bin:/bin echo -e "AT+CREG=0" | libgsmd-tool -m atcmd 2>&1 > /dev/null echo -e "AT+CGREG=0" | libgsmd-tool -m atcmd 2>&1 > /dev/null echo -e "AT%CREG=0" | libgsmd-tool -m atcmd 2>&1 > /dev/null echo -e "AT%CGREG=0" | libgsmd-tool -m atcmd 2>&1 > /dev/null echo -e "AT%CSQ=0" | libgsmd-tool -m atcmd 2>&1 > /dev/null in /etc/apm/resume.d/ #!/bin/sh PATH=/usr/bin:/bin echo -e "AT+CREG=2" | libgsmd-tool -m atcmd 2>&1 > /dev/null echo -e "AT+CGREG=2" | libgsmd-tool -m atcmd 2>&1 > /dev/null echo -e "AT%CREG=2" | libgsmd-tool -m atcmd 2>&1 > /dev/null echo -e "AT%CGREG=2" | libgsmd-tool -m atcmd 2>&1 > /dev/null echo -e "AT%CSQ=1" | libgsmd-tool -m atcmd 2>&1 > /dev/null ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround
> it's worth mentioning that running GSM without association to a network > probably drains battery like hell, as GSM-chipsets tend to constantly > search > the whole frequency range for some BS to register. I've experienced > factor 5 > to 10 for power consumption. i know, i had a phone once that allowed to disable that -- is there a way to tell the gsm modem not to do so? ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround
Am Mi 13. August 2008 schrieb arne anka: > > I finally got around to checking this. The script runs instantly when > > I have cell registration. So, that explains why my script was taking > > so long. But what can I do about it? > > you could check if you have registration, someway along > echo -n "csr" | libgsmd-tool -m shell > and read the return. > > > > If I've suspended my Neo, I > > wouldn't want it waking up (and corrupting the SD card) as soon as I > > move into a cell tower's range... it's worth mentioning that running GSM without association to a network probably drains battery like hell, as GSM-chipsets tend to constantly search the whole frequency range for some BS to register. I've experienced factor 5 to 10 for power consumption. /jOERG signature.asc Description: This is a digitally signed message part. ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 |>> To what are you referring? Did you see that in the linked qtopia code? |> no. i think you said that -- umount/mount on suspend/resume was a |> workaround except when resuming from gsm event. am i wrong? | Oh, yes. That is what it seemed like. Perhaps I was just unlucky the | first time. I'll mess with it again after work to confirm that that | truly is the difference. But right now, I haven't corrupted my SD | card for a few days. I'm pretty sure this will make you immune for getting crapped on. But for SD Card rootfs case, it isn't really a solution. - -Andy -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org iEYEARECAAYFAkijGY4ACgkQOjLpvpq7dMrCSQCglCwDj9FwEP3k7gDZzC/NrOgj GEYAn0lvdZw4ZdCqm4V4yYD1w4xcEyov =Ijzp -END PGP SIGNATURE- ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround
Oh, yes. That is what it seemed like. Perhaps I was just unlucky the first time. I'll mess with it again after work to confirm that that truly is the difference. But right now, I haven't corrupted my SD card for a few days. -Steven On Wed, Aug 13, 2008 at 9:42 AM, arne anka <[EMAIL PROTECTED]> wrote: >> To what are you referring? Did you see that in the linked qtopia code? > > no. i think you said that -- umount/mount on suspend/resume was a > workaround except when resuming from gsm event. am i wrong? > > ___ > support mailing list > support@lists.openmoko.org > https://lists.openmoko.org/mailman/listinfo/support > ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround
> To what are you referring? Did you see that in the linked qtopia code? no. i think you said that -- umount/mount on suspend/resume was a workaround except when resuming from gsm event. am i wrong? ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround
To what are you referring? Did you see that in the linked qtopia code? -Steven On Wed, Aug 13, 2008 at 8:08 AM, arne anka <[EMAIL PROTECTED]> wrote: > well, i am still curios inhowfar resume_reason!=GSM is different from > resume_reason==GSM ... ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround
> I finally got around to checking this. The script runs instantly when > I have cell registration. So, that explains why my script was taking > so long. But what can I do about it? you could check if you have registration, someway along echo -n "csr" | libgsmd-tool -m shell and read the return. > If I've suspended my Neo, I > wouldn't want it waking up (and corrupting the SD card) as soon as I > move into a cell tower's range... well, i am still curios inhowfar resume_reason!=GSM is different from resume_reason==GSM ... ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround
I finally got around to checking this. The script runs instantly when I have cell registration. So, that explains why my script was taking so long. But what can I do about it? If I've suspended my Neo, I wouldn't want it waking up (and corrupting the SD card) as soon as I move into a cell tower's range... Oh well, it's a good start. -Steven On Tue, Aug 12, 2008 at 9:52 AM, arne anka <[EMAIL PROTECTED]> wrote: >> Yup. OM2007.2 used here. > > check when you got registration? > > ___ > support mailing list > support@lists.openmoko.org > https://lists.openmoko.org/mailman/listinfo/support > ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround
> Yup. OM2007.2 used here. check when you got registration? ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround
Yup. OM2007.2 used here. -Steven On Tue, Aug 12, 2008 at 9:42 AM, arne anka <[EMAIL PROTECTED]> wrote: >> I'm concerned then. Running the commands with "echo -e ..." takes >> even longer. I gave up and killed the script after 6 minutes. Maybe >> it's because I'm not receiving cell signals right now? Why else would >> mine take so long when yours goes quick? > > sounds likely. but then, i do absolutely not know how the libsmd-tool > stuff works internally. > do you use 2007.2? > > ___ > support mailing list > support@lists.openmoko.org > https://lists.openmoko.org/mailman/listinfo/support > ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround
> I'm concerned then. Running the commands with "echo -e ..." takes > even longer. I gave up and killed the script after 6 minutes. Maybe > it's because I'm not receiving cell signals right now? Why else would > mine take so long when yours goes quick? sounds likely. but then, i do absolutely not know how the libsmd-tool stuff works internally. do you use 2007.2? ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround
On Tue, Aug 12, 2008 at 7:44 AM, arne anka <[EMAIL PROTECTED]> wrote: >> Well, I think I figured it out a little myself. Maybe I'm doing it >> wrong, but those chat commands take a long time to run on my Neo. > > i do > echo -e "AT..." | libgsmd-tool -m atcmd > and there's no significant delay. I'm concerned then. Running the commands with "echo -e ..." takes even longer. I gave up and killed the script after 6 minutes. Maybe it's because I'm not receiving cell signals right now? Why else would mine take so long when yours goes quick? > i am still not sure if my script in /etc/apm/suspend.d/ is read on suspend > (had to run it by hand), so i'd like to check that before posting. My apm definitely runs the script I have there before suspend. I have it unmounting the SD card and stopping the audio services (trying to work around other suspend bugs). -Steven ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround
> Well, I think I figured it out a little myself. Maybe I'm doing it > wrong, but those chat commands take a long time to run on my Neo. i do echo -e "AT..." | libgsmd-tool -m atcmd and there's no significant delay. > There's no way they were run before the device entered suspend when I > put them in the /etc/apm/suspend.d/ folder. Does your suspend script > take several minutes to run? Each chat command I run takes about 45 > seconds. see above. i am still not sure if my script in /etc/apm/suspend.d/ is read on suspend (had to run it by hand), so i'd like to check that before posting. ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround
Well, I think I figured it out a little myself. Maybe I'm doing it wrong, but those chat commands take a long time to run on my Neo. There's no way they were run before the device entered suspend when I put them in the /etc/apm/suspend.d/ folder. Does your suspend script take several minutes to run? Each chat command I run takes about 45 seconds. I ran each of the commands and then ran "apm -s". I still get woken up in under a minute almost every time. So, there has so be some other GSM event affecting me. -Steven On Tue, Aug 12, 2008 at 6:52 AM, Steven ** <[EMAIL PROTECTED]> wrote: > Would you mind posting those scripts? I'd like to try them out. > > -Steven > > On Tue, Aug 12, 2008 at 4:13 AM, arne anka <[EMAIL PROTECTED]> wrote: http://git.openmoko.org/?p=qtopia.git;a=blob;f=devices/ficgta01/src/plugins/phonevendors/ficgta01/vendor_ficgta01.cpp;hb=HEAD#l603 >> >> i put it into /etc/apm/suspend.d/ (and the accoring reset into >> /etc/apam/resume.d/) and after running the script manually (apm seemed not >> to pick it up automatically, needs further inverstigation) at 00:48 it >> slept until 10:27 -- then i was somewhere on my way into the office. >> looks very promising though there's apparently still a gsm event not >> catched. >> does somebody know how to check the state of CSQ, CREG, CGREG? i'd like to >> see if the scripts have effect or if it's only coincidence. >> >> ___ >> support mailing list >> support@lists.openmoko.org >> https://lists.openmoko.org/mailman/listinfo/support >> > ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround
Would you mind posting those scripts? I'd like to try them out. -Steven On Tue, Aug 12, 2008 at 4:13 AM, arne anka <[EMAIL PROTECTED]> wrote: >>> http://git.openmoko.org/?p=qtopia.git;a=blob;f=devices/ficgta01/src/plugins/phonevendors/ficgta01/vendor_ficgta01.cpp;hb=HEAD#l603 > > i put it into /etc/apm/suspend.d/ (and the accoring reset into > /etc/apam/resume.d/) and after running the script manually (apm seemed not > to pick it up automatically, needs further inverstigation) at 00:48 it > slept until 10:27 -- then i was somewhere on my way into the office. > looks very promising though there's apparently still a gsm event not > catched. > does somebody know how to check the state of CSQ, CREG, CGREG? i'd like to > see if the scripts have effect or if it's only coincidence. > > ___ > support mailing list > support@lists.openmoko.org > https://lists.openmoko.org/mailman/listinfo/support > ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround
>> http://git.openmoko.org/?p=qtopia.git;a=blob;f=devices/ficgta01/src/plugins/phonevendors/ficgta01/vendor_ficgta01.cpp;hb=HEAD#l603 i put it into /etc/apm/suspend.d/ (and the accoring reset into /etc/apam/resume.d/) and after running the script manually (apm seemed not to pick it up automatically, needs further inverstigation) at 00:48 it slept until 10:27 -- then i was somewhere on my way into the office. looks very promising though there's apparently still a gsm event not catched. does somebody know how to check the state of CSQ, CREG, CGREG? i'd like to see if the scripts have effect or if it's only coincidence. ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround
> http://git.openmoko.org/?p=qtopia.git;a=blob;f=devices/ficgta01/src/plugins/phonevendors/ficgta01/vendor_ficgta01.cpp;hb=HEAD#l603 sounds good, i give it a try tonight. ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: gta02, 2007.2: frequent wakeups -- creating workaround
On Mon, Aug 11, 2008 at 12:14:10PM +0200, arne anka wrote: > i logged the resume_reason for the past days and all undesired wakeups are > caused by EINT01_GSM -- as expected. > so, i'd like to catch that events and send the phone back to sleep > immediately. > but, and it's a big but, EINT01_GSM is triggered even when a call comes in > (and probably an sms, too). so it's not sufficient. Another idea: While poking through the Qtopia source code, I found something that might be interesting. Upon suspend, Qtopia sends commands to the modem to disable notifications about cell broadcasts, signal quality, etc. Upon wakeup the settings are restored. If you want to see how this is done, you can look here, for example: http://git.openmoko.org/?p=qtopia.git;a=blob;f=devices/ficgta01/src/plugins/phonevendors/ficgta01/vendor_ficgta01.cpp;hb=HEAD#l603 I don't know whether OM2007.2 does something similar at the moment (I haven't looked), but if it doesn't then adding it might help. Regards, Thomas ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
gta02, 2007.2: frequent wakeups -- creating workaround
i logged the resume_reason for the past days and all undesired wakeups are caused by EINT01_GSM -- as expected. so, i'd like to catch that events and send the phone back to sleep immediately. but, and it's a big but, EINT01_GSM is triggered even when a call comes in (and probably an sms, too). so it's not sufficient. how do i the following three things in sh/bash: - how do i check if a call is coming in or ongoing? - how do i check if an sms causing the wake up? - how do i disable the touchscreen temporarily? someone told that raster is creating a daemon handling this kind of things, but ther was no answer to my inquiry if it would be asu specific, and anyway with raster leaving and the 2008.8 just released i don't expect it to be finished in a near future. the phone waking up every 6 or 7 minutes as it hours on end does _is_ annoying and i like to create a workaround at least. ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support