Re: [shr-u] wakeuptime, was Re: 2.6.31 is out,where is my ultimate kernel?
how does the end user get to that point? I think there are some who will point to Qi as a part of the solution, but I understand that this is a faith-based discussion. ;) well, i am not 100% sure, but i think indeed it is qi -- not at least because it a) does less on wakeup (might be wrong here) b) is configurable to print considerable less messages while resuming, those print outs take a lot of time additionally the recently mentioned debugfs hack _seems_to improve things. i tested a bit yesterday with and without the debugfs hack (only 6,7 calls) and with the debugfs hack, one ring was enough. the second ring the caller heard was the first one the fr made. w/ the hack it took usually two or sometimes three (which i knew from before already). ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: [shr-u] wakeuptime, was Re: 2.6.31 is out,where is my ultimate kernel?
I'm using QI, have to wait one ring to get the screen, press accept and than another ring to have the other person actually hear me. The delay is very confusing, for both parties.(Seems like they hear that someone picked up the phone, but can't hear me in the beginning. -- have to acknowledge that). Trying the debugfs hack today afternoon. On Tue, Sep 15, 2009 at 09:44, arne anka openm...@ginguppin.de wrote: how does the end user get to that point? I think there are some who will point to Qi as a part of the solution, but I understand that this is a faith-based discussion. ;) well, i am not 100% sure, but i think indeed it is qi -- not at least because it a) does less on wakeup (might be wrong here) b) is configurable to print considerable less messages while resuming, those print outs take a lot of time additionally the recently mentioned debugfs hack _seems_to improve things. i tested a bit yesterday with and without the debugfs hack (only 6,7 calls) and with the debugfs hack, one ring was enough. the second ring the caller heard was the first one the fr made. w/ the hack it took usually two or sometimes three (which i knew from before already). ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: [shr-u] wakeuptime, was Re: 2.6.31 is out, where is my ultimate kernel?
On Tuesday 15 September 2009, Russell Dwiggins wrote: I'm still a bit confused. I'm in the same boat with others here where it takes between 3 and 4 rings before I can take any action on the incoming call. (shr-u and u-boot, both recently updated) What is the condition that allows one to take action on a call ~1 ring, and how does the end user get to that point? I think there are some who will point to Qi as a part of the solution, but I understand that this is a faith-based discussion. ;) I'd prefer to remain with u-boot for now until the suggested kexec kernel shows up to allow me to directly select the partition I wish to boot from. You need to reduce the log level. The best way is to do it in the kernel parameters as this will also help with the boot time, but it can be done at runtime as well, so you could add it to an init script. Log level 4 gets rid of almost everything but should still give something useful with errors, while 0 will stop it entirely. You could pick something in between too. For the kernel parameters you need to edit your uboot environment. Add either 'quiet' or 'loglevel=n' where n is the desired loglevel. quiet gives loglevel 4. See: http://wiki.openmoko.org/wiki/Bootloader_commands#Environment The runtime method is described on the shr tweaks page: http://www.shr- project.org/trac/wiki/Tweaks#majorreallymajorspeedupofsuspendandwakeup ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: [shr-u] wakeuptime, was Re: 2.6.31 is out,where is my ultimate kernel?
2009/9/15, Russell Dwiggins undrwa...@verizon.net: I'm still a bit confused. I'm in the same boat with others here where it takes between 3 and 4 rings before I can take any action on the incoming call. (shr-u and u-boot, both recently updated) What is the condition that allows one to take action on a call ~1 ring, and how does the end user get to that point? I think there are some who will point to Qi as a part of the solution, but I understand that this is a faith-based discussion. ;) I'd prefer to remain with u-boot for now until the suggested kexec kernel shows up to allow me to directly select the partition I wish to boot from. Russell Dwiggins ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community This is documented in the SHR User Manual[1] and on the SHR wiki[2]. [1] http://wiki.openmoko.org/wiki/SHR_User_Manual#Speedup_of_suspend_and_wake_up [2] http://shr-project.org/trac/wiki/Tweaks#majorreallymajorspeedupofsuspendandwakeup ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: [shr-u] wakeuptime, was Re: 2.6.31 is out, where is my ultimate kernel?
On Mon, Sep 14, 2009 at 08:03:38AM +0200, Matthias Huber wrote: Paul Fertser schrieb: Matthias Huber matthias.hu...@wollishausen.de writes: Paul Fertser schrieb: What messages do you see with loglevel=4 that shouldn't be there? For me normal suspend/resume doesn't add anything to console. ... i cannot tell exact, but i saw this messages. test it yourself. I did. And sent some patches to the kernel to hide inappropriate messages, they were applied long time ago. Currently i don't see any single message on boot (if g_ether is compiled in) and no single message on suspend/resume (if everything is working properly). hmmm, now we have to define over what we are talking,. i mean: Linux om-gta02 2.6.29-rc3 #1 PREEMPT Tue Sep 1 23:03:59 CEST 2009 armv4tl unknown Yes, Sep 1 is modern enough. where does a normal user get g_ether compiled in, when he uses the standard shr-u - kernel ? With g_ether not compiled in i see only 4 messages on boot and i'm sure they can't slow down the boot considerably. i will return later with exact messages on suspend/resume. dont have time now. but maybe this is g_ether. Nah, it doesn't print anything on suspend/resume. ... if nothing else then loglevel=0 will solve this, i will make it zero. Go ahead, but i very much hope you wouldn't be able to convince anybody to make it default. and speed is also bootup-speed, (althoug this wasn't the trhread) where i see a lot of for an non-kernel-developer sensless messages, wich nobody cant read without a magnifying glass. I can read messages on LCM without a magnifying glass, so nobody is obviously a little bit inaccurate here. what do you mean with LCM ? On moko, i cant read this. its too small for me. LCM is LCD module (because it's not only LCD but also backlight and touchscreen bundled together) You != everybody is this exact enough for you ? but please don't say, it's unsharp :-)) I won't. 285 dpi that our displays have can't be unsharp. for me, who is/must/wants use this phone as daily phone totally senseless. I do use my gta02 as my only cell phone since November. And i think that if some messages have inappropriate loglevels, that should be fixed in kernel instead of hiding them. with that, i agree, although my knoledge of linux says me, that, with any loglevel 0 you will get some output at all. Somehow i do not (except when something is going wrong, but it happens only once, after that the messages don't take time to be outputted, because no scrolling is involved). So i'm waiting for something that can in fact help to find and fix the messages in case they're really inappropriate. -- Be free, use free (http://www.gnu.org/philosophy/free-sw.html) software! mailto:fercer...@gmail.com ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: [shr-u] wakeuptime, was Re: 2.6.31 is out, where is my ultimate kernel?
first, let me say sorry for the post this moring. i had the finger to long on the power knob. in fact, i had loglevel 8 while i took the picture. with loglevel 4 i get (meanwhile) the four messages you told (g_ether) and two more messages of my own. -- View this message in context: http://n2.nabble.com/2-6-31-is-out-where-is-my-ultimate-kernel-tp3619631p3640633.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: [shr-u] wakeuptime, was Re: 2.6.31 is out, where is my ultimate kernel?
matzehuber matthias.hu...@wollishausen.de writes: with loglevel 4 i get (meanwhile) the four messages you told (g_ether) and two more messages of my own. Ok, np. -- Be free, use free (http://www.gnu.org/philosophy/free-sw.html) software! mailto:fercer...@gmail.com ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
RE: [shr-u] wakeuptime, was Re: 2.6.31 is out,where is my ultimate kernel?
I'm still a bit confused. I'm in the same boat with others here where it takes between 3 and 4 rings before I can take any action on the incoming call. (shr-u and u-boot, both recently updated) What is the condition that allows one to take action on a call ~1 ring, and how does the end user get to that point? I think there are some who will point to Qi as a part of the solution, but I understand that this is a faith-based discussion. ;) I'd prefer to remain with u-boot for now until the suggested kexec kernel shows up to allow me to directly select the partition I wish to boot from. Russell Dwiggins ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: [shr-u] wakeuptime, was Re: 2.6.31 is out, where is my ultimate kernel?
Matthias Huber matthias.hu...@wollishausen.de writes: and i suggest to change qi, because, as you said above, it is: 1. easily chanchable upon runtime for the developer 2. faster for the normal daily use 3. qi can with power-key hold give you a loglevel of 8. What messages do you see with loglevel=4 that shouldn't be there? For me normal suspend/resume doesn't add anything to console. -- Be free, use free (http://www.gnu.org/philosophy/free-sw.html) software! mailto:fercer...@gmail.com ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: [shr-u] wakeuptime, was Re: 2.6.31 is out, where is my ultimate kernel?
Paul Fertser schrieb: Matthias Huber matthias.hu...@wollishausen.de writes: and i suggest to change qi, because, as you said above, it is: 1. easily chanchable upon runtime for the developer 2. faster for the normal daily use 3. qi can with power-key hold give you a loglevel of 8. What messages do you see with loglevel=4 that shouldn't be there? For me normal suspend/resume doesn't add anything to console. first: i cant see anything without a magnifying glass :-) that's not the point although there are some messages. and if you thing logically, you will give me right. i cannot tell exact, but i saw this messages. test it yourself. and speed is also bootup-speed, (althoug this wasn't the trhread) where i see a lot of for an non-kernel-developer sensless messages, wich nobody cant read without a magnifying glass. for i made /var/log/messages permanent, there is/was writing into the flash which seemed to slow down things. this slows of course also the process down (also the process of getting asleep and waking up) normally one must think on other things, like the extensive logging of frameworkd and ophonekit. for me, who is/must/wants use this phone as daily phone totally senseless. sorry for my maybe provicative answers. ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: [shr-u] wakeuptime, was Re: 2.6.31 is out, where is my ultimate kernel?
Matthias Huber matthias.hu...@wollishausen.de writes: Paul Fertser schrieb: What messages do you see with loglevel=4 that shouldn't be there? For me normal suspend/resume doesn't add anything to console. ... i cannot tell exact, but i saw this messages. test it yourself. I did. And sent some patches to the kernel to hide inappropriate messages, they were applied long time ago. Currently i don't see any single message on boot (if g_ether is compiled in) and no single message on suspend/resume (if everything is working properly). and speed is also bootup-speed, (althoug this wasn't the trhread) where i see a lot of for an non-kernel-developer sensless messages, wich nobody cant read without a magnifying glass. I can read messages on LCM without a magnifying glass, so nobody is obviously a little bit inaccurate here. for me, who is/must/wants use this phone as daily phone totally senseless. I do use my gta02 as my only cell phone since November. And i think that if some messages have inappropriate loglevels, that should be fixed in kernel instead of hiding them. -- Be free, use free (http://www.gnu.org/philosophy/free-sw.html) software! mailto:fercer...@gmail.com ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
[shr-u] wakeuptime, was Re: 2.6.31 is out, where is my ultimate kernel?
Warren Baird schrieb: On Thu, Sep 10, 2009 at 2:15 PM, Matthias Huber matthias.hu...@wollishausen.de mailto:matthias.hu...@wollishausen.de wrote: Hmmm it takes three rings until i can get the call (heear a ring and get a GREEN Button. Don't think, this is fast. Which distro? I used to have this problem with OM2008 - I missed a bunch of calls because of it and then switched to QTEI. I've since tried OM2009 and am now using SHR-U, without this issue. If I call myself from a land line using SHR-U, I see the screen turn on at exactly the same time as I hear the ring on the land-line, and the FR starts to ring before the first ring on the landline finishes. I don't have a stop-watch, but it certainly seems fast enough. Distro is shr-u, with all updates of yesterday ~17:00h GMT with my GREEN BUTTONS and my own Ringtone. What firmware do you use ? ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
[shr-u] wakeuptime, was Re: 2.6.31 is out, where is my ultimate kernel?
Today i did another test, because i patched qi with ghex for giving the kernel loglevel=0, and my result today (look at the bottom for version) was: ONE Ring until i get shr-todays slider to see. That is ok for me. On that there is the question: Why is qi delivered on the repo with debuglevel 4 ? Warren Baird schrieb: On Thu, Sep 10, 2009 at 2:15 PM, Matthias Huber matthias.hu...@wollishausen.de mailto:matthias.hu...@wollishausen.de wrote: Hmmm it takes three rings until i can get the call (heear a ring and get a GREEN Button. Don't think, this is fast. Which distro? I used to have this problem with OM2008 - I missed a bunch of calls because of it and then switched to QTEI. I've since tried OM2009 and am now using SHR-U, without this issue. If I call myself from a land line using SHR-U, I see the screen turn on at exactly the same time as I hear the ring on the land-line, and the FR starts to ring before the first ring on the landline finishes. I don't have a stop-watch, but it certainly seems fast enough. Distro is shr-u, with all updates of yesterday ~17:00h GMT with my GREEN BUTTONS and my own Ringtone. What firmware do you use ? ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: [shr-u] wakeuptime, was Re: 2.6.31 is out, where is my ultimate kernel?
On Friday 11 September 2009, Matthias Huber wrote: Today i did another test, because i patched qi with ghex for giving the kernel loglevel=0, and my result today (look at the bottom for version) was: ONE Ring until i get shr-todays slider to see. That is ok for me. On that there is the question: Why is qi delivered on the repo with debuglevel 4 ? Because it's a fair tradeoff between speed and getting some debug information if something goes wrong? If booting from SD the loglevel is easily changed anyway. It would be just as easy for NAND if Qi could read jffs2. You can also change the loglevel at runtime as suggested in the shr tweak/customisation guides. This will be just as effective for resume, but not as fast for boot. Warren Baird schrieb: On Thu, Sep 10, 2009 at 2:15 PM, Matthias Huber matthias.hu...@wollishausen.de mailto:matthias.hu...@wollishausen.de wrote: Hmmm it takes three rings until i can get the call (heear a ring and get a GREEN Button. Don't think, this is fast. Which distro? I used to have this problem with OM2008 - I missed a bunch of calls because of it and then switched to QTEI. I've since tried OM2009 and am now using SHR-U, without this issue. If I call myself from a land line using SHR-U, I see the screen turn on at exactly the same time as I hear the ring on the land-line, and the FR starts to ring before the first ring on the landline finishes. I don't have a stop-watch, but it certainly seems fast enough. Distro is shr-u, with all updates of yesterday ~17:00h GMT with my GREEN BUTTONS and my own Ringtone. What firmware do you use ? ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: [shr-u] wakeuptime, was Re: 2.6.31 is out, where is my ultimate kernel?
Al Johnson schrieb: On Friday 11 September 2009, Matthias Huber wrote: Today i did another test, because i patched qi with ghex for giving the kernel loglevel=0, and my result today (look at the bottom for version) was: ONE Ring until i get shr-todays slider to see. That is ok for me. On that there is the question: Why is qi delivered on the repo with debuglevel 4 ? Because it's a fair tradeoff between speed and getting some debug information if something goes wrong? If booting from SD the loglevel is easily changed anyway. It would be just as easy for NAND if Qi could read jffs2. You can also change the loglevel at runtime as suggested in the shr tweak/customisation guides. This will be just as effective for resume, but not as fast for boot. i stated the question as advicatus diaboli for the normal user, which wants to use the phone as daily. and i suggest to change qi, because, as you said above, it is: 1. easily chanchable upon runtime for the developer 2. faster for the normal daily use 3. qi can with power-key hold give you a loglevel of 8. ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
