![]() |
Re: [Announce] kernel-power unstable v49
Quote:
my lawyer is calling me. |
Re: [Announce] kernel-power unstable v49
Dude, relax.
First of all it wasn't russian, trust me on that one. Second, as already been told by freemangordon, drop this subject, it's offtopic |
Re: [Announce] kernel-power unstable v49
Quote:
|
Re: [Announce] kernel-power unstable v49
Quote:
Will try smth else later. EDIT btw i actually left it WITHOUT charger |
Re: [Announce] kernel-power unstable v49
Quote:
EDIT: Seems most of the devices cannot handle voltage bellow 20(850mv). So it does not make sense to enable SR to go bellow that. If someone wants he can always make a custom profile with whatever voltages he want :) |
Re: [Announce] kernel-power unstable v49
Quote:
im stupid coz im indian. |
Re: [Announce] kernel-power unstable v49
Quote:
|
Re: [Announce] kernel-power unstable v49
I can confirm that voltages lower than 20 will cause reboots sometimes.
|
Re: [Announce] kernel-power unstable v49
hello freemangordon...
i had a reboot at 900MHz during a video playback and also under some normal usage but not very frequently... so i have set freq limit to 850 & dsp to 560.. can please suggest some changes..? thanks |
Re: [Announce] kernel-power unstable v49
Hello, I uploaded new KP build. If it is stable it will go to extras-devel!
|
Re: [Announce] kernel-power unstable v49
Quote:
EDIT: @all - this is the changelog for latest build compared to previous 1.VDD1 efuse calibration values support up to 900MHz 2. Fixed bug (introduced by me) causing incorrect efuse calibration value to be used for VDD1 SR (i.e. instead calibration value for OPP(n), calibration value for OPP(n-1) was used) 3. efuse calibration values for all VDD1 OPPs recalculated as stock ones(for 125MHz and 250MHz) were keeping too high voltages 4. added several entries under /sys/power sr_vdd1_voltage sr_vdd2_voltage efuse_vdd2 sr_vdd1_voltage, sr_vdd2_voltage reflect current voltages for VDD1 and VDD2 respectively when SmartReflex is active 5. /sys/power/Efuse renamed to efuse_vdd1 6. When SmartReflex is active, current calculated voltage is stored as default for current frequency when frequency switch occurs. That way next time this frequency becomes active, AVS starts from last voltage(+1 to be on a safe side) instead from profile value(no overvoltage happens on frequency switch). 7. In case DSP is overclocked for a particular frequency a little voltage boost is added to SR calibration for that frequency(12.5 mV if 430<freq<=520 and 25mV if freq>520). 8. SmartReflex AVS step reduced to 1(12.5mV) from 4(50mV) to prevent undervoltage |
Re: [Announce] kernel-power unstable v49
Sorry for most likely dumb question, but where latest devel-candidate version was uploaded? I don't want to accidentally test older version. Of course tried searching for a while, but again, I'm lost when it comes to non-repos testing places (no accusation here, it's probably just me).
Also, for testing purposes, should we use only vdd1, or vdd2 too? I remember, that you thought about possible need for recalculating vdd2 to, so common sense tells me "use vdd1 only". Confirm/Deny? /Estel |
Re: [Announce] kernel-power unstable v49
Quote:
|
Re: [Announce] kernel-power unstable v49
Thanks, I thought about that, but wasn't sure if it is "latest-latest", or "one-before-latest" ;)
/Estel |
Re: [Announce] kernel-power unstable v49
Ok, I've been trying this latest kernel (nov.16), and I have had no reboots, took a 720p vid, saw it, I had been checking photos and movies, listening to music, used both Osso-mediaplayer and Open Media Player, browsing the net, applied speedpatch (I've had it removed previously) and still rock-solid.
I Think I'm fine with this, OC@900Mhz, Smartreflex on :D EDIT: Seems like I spoke too quick... just had a reboot while launching 'qreddit', and then I reboot again to ensure things would go fine, and then Desktop was unresponsive and laggy, checked dmesg and it was full of Code:
HWRecoveryResetSGX: SGX Hardware Recovery triggered One more reboot fixed it... I don't know what was that, it had never happened to my N900 before... |
Re: [Announce] kernel-power unstable v49
Quote:
/Estel |
Re: [Announce] kernel-power unstable v49
Quote:
i haven't installed any sort of speedpatch, swappolube or something like that but still got a reboot at 900MHz...i am really very confused..but one thing for sure smart reflex do works :P my battery backup has increased to several hours... @freemangordon...thanks a lot pal ! and can u please check the 900 freq or provide some more voltage boost to it...i always see sr_vdd1 voltage stuck at 55 only at 900.. thanku.. |
Re: [Announce] kernel-power unstable v49
seems like there is alot of testers on the variation of settings... i really hope all stable tweaks can be shared somehow... now i hv to go back the threads and find where was the thread i have looked at that i want to try.... anyway.... great job to alll...
am using the kp49 from the link provided... on dsp profile provided... no tweak yet... i see new driver for bq27x00 driver... how do i use it? modprob or ther is a new short command for it? |
Re: [Announce] kernel-power unstable v49
Quote:
EDIT: @strange1712 - this message appears sometimes if device is rebooted instead of powered down. It happens on stock kernel too, not related to KP. On the other hand AFAIK speedpatch has its own frequency profiles, which are not compatible with latest KP because of the frequencies change. Please remove it before continue testing. |
Re: [Announce] kernel-power unstable v49
Quote:
|
Re: [Announce] kernel-power unstable v49
I got the latest (Nov 16) KP 49 from the first post, but I kinda have a problem. My profile is:
Code:
# kernel configuration file generated by /usr/sbin/kernel-config Some light? |
Re: [Announce] kernel-power unstable v49
@fahadj2003 - Ok, if you say so. I've tried that crap once, and removed it in a blink of an eye as it caused my device to be laggy as hell, and to freeze h-d leading to reboot.
Anyway, it could be a coincidence that restart happend after installing "speedpatch", but it could be not. |
Re: [Announce] kernel-power unstable v49
Quote:
|
Re: [Announce] kernel-power unstable v49
Quote:
Or best throw that profile away and use dsp with modified MAX frequency. EDIT: Just wondering what is your battery life with this profile. 6 hours? |
Re: [Announce] kernel-power unstable v49
Quote:
Quote:
jus recorded it at 848x400.. at 600 dsp.. froze at first *dunno why*.. but worked fine later.. # DO NOT EDIT THIS FILE - CREATE A COPY OF IT FOR CUSTOMIZATION # minimum frequency to use MINFREQ=900 # maximum frequency to use MAXFREQ=1100 # list of frequency configurations: each "frequency:volt,dsprate" FREQS="0:25,90 125:25,90 250:33,180 500:43,360 550:49,400 600:55,430 720:55,520 805:55,600 850:55,600 900:55,600 950:55,600 1000:61,600 1100:72,600 1150:72,600" SMARTREFLEX_VDD1=0 SMARTREFLEX_VDD2=0 GOVERNOR=ondemand UP_THRESHOLD=95 SAMPLING_RATE=100000 IGNORE_NICE_LOAD=1 POWERSAVE_BIAS=56 |
Re: [Announce] kernel-power unstable v49
@freemangordon...
oops ! thanks a lot ... u saved my beloved n900 cpu from frying :p okay, now i am using these settings and everything seems to be good till now.. # DO NOT EDIT THIS FILE - CREATE A COPY OF IT FOR CUSTOMIZATION # minimum frequency to use MINFREQ=125 # maximum frequency to use MAXFREQ=850 # list of frequency configurations: each "frequency:volt,dsprate" FREQS="0:30,90 125:30,90 250:38,180 500:48,360 550:54,400 600:60,430 720:60,480 805:60,540 850:60,600 900:60,600 950:60,600 1000:60,600 1100:72,600 1150:72,600" SMARTREFLEX_VDD1=1 SMARTREFLEX_VDD2=1 GOVERNOR=ondemand UP_THRESHOLD=95 SAMPLING_RATE=300000 IGNORE_NICE_LOAD=0 POWERSAVE_BIAS=0 and sr_vdd1 shows this 720:45 805:49 850:52 are these values good or i still need to reduce frequency and dsp..(p.s - no reboots till now with these settings) thanks... |
Re: [Announce] kernel-power unstable v49
Quote:
Nothing at all :D |
Re: [Announce] kernel-power unstable v49
Quote:
The dsp profile, hmmm. When I load it, and then start to capture anything in HD, the preview of camerui2 is laggy, not the video... I mean, as you are recording, you see what is being recorded. <-- This part is laggy. But the video recorded is great. But with that profile, everything is fast, but now I guess that it is frying my N900 right? EDIT: I have changed the 700 and 750 steps to 720. That post was a typo. |
Re: [Announce] kernel-power unstable v49
Quote:
Quote:
Just use dsp with limits 250(or 125)-900, remove all "magical" patches and you will fine with both performance and battery life. |
Re: [Announce] kernel-power unstable v49
Using pali's 20111116 build is generally stable. I had just one exception with the previous 20111112 build: moving all the new KP build packages over from my Windows box via WiFi & CIFS resulted in a reboot. (stock kernel-power-settings-13 'dsp' profile).
Push it to testing already, but i dunno what the _current_ issue with the repository is. |
Re: kernel-power patches/updates
well well i have been off for a long while
now just to make somethings clear : for who are using batterypatch and using KP latest build : edit overclock profile and remove 700 and change 750 to 720 and you will be fine for who are using speedpatch (and doesn't having problems wih it) then continue using it, there are no conflicts betwean it and KP and the only problem than speedpatch may cause is the terminal message still i can't solve that and speedpatch doesn't cause unexpected reboots @maacruz actually it does assign processes to cgroup you can check your self the 1st version of speedpatch didn't do that @freemangordon Quote:
and i don't know why it caused you troubles anyway after installing speedpatch you and some others may think that moving from A to B as everything in /sysparts is moved to /dev/cgroup doesn't improve anything but that is wrong it actually improves multiple things as linus torvalds said on N900 the most imroved thing are multitask and some other programms EX: transmission, nokia panorama ...etc hope i made everything clear :) |
Re: [Announce] kernel-power unstable v49
Ever since I updated to 16/11 kp49, xorg doesn't seem to wanna settle down. Idle CPU mode is at 250mhz 30% load minimum, xorg is 15% minimum.
|
Re: kernel-power patches/updates
Quote:
|
Re: [Announce] kernel-power unstable v49
Quote:
|
Re: [Announce] kernel-power unstable v49
No "sgx-recovery" message in dmesg. And I rebooted the device 3 times already, still the same.
EDIT: Cleared dmesg log and did a power-cycle. Back to normal. |
Re: [Announce] kernel-power unstable v49
Take the battery out for 30 seconds. Then have a look.
I have been away for a while am well horny about the inclusion of compcache&hd recording. Is there a wiki entry that details the instructions required to get all this cool **** up to date? OR Should I just wait for KP49 to hit the repos and take it form there? |
Re: [Announce] kernel-power unstable v49
guys
what is default file which is loading after every reboot? |
Re: [Announce] kernel-power unstable v49
Unfortunately, due to Nokians being slow on releasing code of some LGPL'ed binary, + DSP binary part, that we aren't going to have source code for, manual part of installation will be still required. I don't think, that we're going to re-distribute closed source DSP binary - as freemangordon said, it's probably not even belonging to Nokia. People need to extraxt those thing from harmattan rootfs (fixme?).
/Estel |
Re: kernel-power patches/updates
Quote:
Pali, freemangordon and the rest of the respectable members of this forum, I ask for your forgiveness for going off-topic, this will be the only and last time. Code:
/tmp $ apt-cache showpkg speedpatch (me apt-get -d install speedpatch) (me dpkg -x speedpatch /tmp/speedpatch) Code:
/tmp/speedpatch $ find -type f rc.local -> this just creates the cgroup user directory and sets the cgroup_clean script as release agent evet.d/speedpatch -> just the same cgroup_clean -> this just removes an empty cgroup directory .profile.speedpatch -> this just calls .bashrc as is customary .bashrc.speedpatch -> mmmm how funny. Code:
/tmp/speedpatch $ cat home/user/.bashrc.speedpatch Just like I do in my desktop... to launch cpu hungry processes like kernel builds from the xterm Of course this does nothing for processes not launched from a xterm. Let's see what's said in the "speedpatch" thread. Quote:
It doesn't create any of those cgroups, and of course, it doesn't monitor anything nor moves any process to any cgroup. It just creates a cgroup for each interactive shell, but how many processes are launched from a shell in the N900? None. Processes are lauched by dbus from GUI actions. So, as I said before, all processes remain under the main cgroup, and this "speedpatch" does nothing. Unless you want to compile the kernel in your N900 or something like that :D Ahhh Karam, what a naugty joker you are! |
Re: [Announce] kernel-power unstable v49
Lol you did that maacruz didn'tchu
I actually investigated it since the beginning and have known that but didn't want to post my findings anywhere not to let down the spirits. At least you get a placebo! Where is epitaph BTW? He invented that crap in the first place... [/END-OFFTOPIC] |
All times are GMT. The time now is 10:06. |
vBulletin® Version 3.8.8