• After 15+ years, we've made a big change: Android Forums is now Early Bird Club. Learn more here.

Root [KERNEL] TheOC v1.7.20 kernel for MT Isaac & TG based CM7

I'll take a look at that experimental fix. What drew you to that part of the code?

http://androidforums.com/triumph-al...nel-mt-isaac-tg-based-cm7-12.html#post3678427

^^ this post. I was looking into TG's wakelock on the wifi chip and just came upon that area since 'MMC' and 'wake_unlock' were the two clues. It's just a guesstimate. I'd appreciate any pointers on how to set up tracing & debugging.

As for the mkbootimg line, I've never used that pagesize option in any of the flashable zips that I've posted and it seems to be working fine. It's weird that you would need it. It's probably your system that you're using to build it needs it.
Yeah dunno about that either. Running Ubuntu 11.10 on VMware player 4 on Win 7 64-bit. 3 GB RAM (2 GB for Ubuntu).
 
Really itching on trying the 61mhz frequency as my new min, has anyone had any issues going this low? Remember reports that some phones had issues at 122mhz in that they would not wake. Never had any issues with my phone at this frequency though.

I never had problems with 122 Mhz min, it also helps that I'm not using a governor that locks the sleep frequency to the min either (like interactiveX). Besides a good governor should ramp it pretty fast to max if the demand is there.

BTW, I can understand expanding the range of frequencies, but why add additional intermediates like 188 Mhz? The stock one only came with 245,4xx,768,1024. or something simple like that. I don't believe there's much battery savings to squeeze out by going to 188 vs 245 for example. If anything there may be a cost to configuring and tearing down the voltage driver setups.
 
Hey thanks for the 2-way call recording patch. Pretty obscure, but I used that app awhile back to bust a bill collector. Had to use speaker phone and the results were usable, but crappy because it used the Mic.
 
I never had problems with 122 Mhz min, it also helps that I'm not using a governor that locks the sleep frequency to the min either (like interactiveX). Besides a good governor should ramp it pretty fast to max if the demand is there.

BTW, I can understand expanding the range of frequencies, but why add additional intermediates like 188 Mhz? The stock one only came with 245,4xx,768,1024. or something simple like that. I don't believe there's much battery savings to squeeze out by going to 188 vs 245 for example. If anything there may be a cost to configuring and tearing down the voltage driver setups.

Well, when I was testing the changes for the Interactive governor, it never actually stopped at the 188 MHz, basically, the most used clock states that I saw, was Deep Sleep, 61, and 1024 when my min/max range of clock frequencies were set at 61/1708 MHz -- depending on if I have my wifi enabled or not. So I figured it would be ok to leave it so anyone whose phone couldn't handle 122 and still wanted to go lower than 245 could try the 188--every little ounce of juice helps.
 
Hey thanks for the 2-way call recording patch. Pretty obscure, but I used that app awhile back to bust a bill collector. Had to use speaker phone and the results were usable, but crappy because it used the Mic.

Yeah, I was actually kind of impressed with the sound quality from that CallRecorder app that I mentioned in the OP; it was much better than I expected. If only it was free... :)
 
Hey Mantera, check the default v for the bottom 3 clock speeds. Seems to be an incongruity there. No big deal, of course I can change them. I expect you work on this on the wee hours.

Btw.. That thing about having your wife enabled?
Uh, is that an undocumented feature? Uh, is it, uh.... Configurable?

;)

What - uh, never mind....
 
Hey Mantera, check the default v for the bottom 3 clock speeds. Seems to be an incongruity there. No big deal, of course I can change them. I expect you work on this on the wee hours.

Btw.. That thing about having your wife enabled?
Uh, is that an undocumented feature? Uh, is it, uh.... Configurable?

;)

What - uh, never mind....

Whoops! Thanks. Fixed the spelling.

As for the default voltage for the bottom clocks. The stock default is 900 mV. If you had previously used Incredicontrol to set the voltages previously, any voltage that you had previously set will remain but the new clock speeds will show up with the stock voltage which you'll need to adjust.
 
Ok, I gotcha. I use system tuner pro for that, but I get it. Had thought I set back to default first.

Damn. I thought you had enabled a killer new feature. Now I find out it's only wifi. Buzz kill. I was hoping to over clock the wife... Or under clock, as the demand ramped down.
:D
 
LOL, turns out CM's new re-coded interactive governor is broken (it locks to highest cpu freq). Reset stats in CPU spy, then try refresh a few times with screen on & off). Everything else seems to work fine though, including if I switch to ondemand governor & set min to 61 Mhz.
 
LOL!! Wife enabled ;)

Everything working fine here also.
There seems to be an increase in battery drain (7% per hour idle as compared to previous 3% per hour idle) but I'm going to check my background apps, look at autostarts, and re-boot (and possibly re-flash everything)

No issues for me either in underclocking to 61mHz (with no screen off profile utilized to lock that in place to prevent possible waking issues) and performance on this phone is simply amazing!

Thanks to everyone who contributes their work, this phone is phenomenal!! With this phone overclocked and undervolted I love comparing it to other people's phones who are paying $70/month.. the Triumph is a beast considering it's pre-paid.
 
I never had problems with 122 Mhz min, it also helps that I'm not using a governor that locks the sleep frequency to the min either (like interactiveX). Besides a good governor should ramp it pretty fast to max if the demand is there.

I lowered my minimum down to 61 MHz and was about to report that all is fine but I just now looked down at my phone to see the little blue Android on a skateboard. Using ondemand and had 61 at 725, 122 at 750, 184 at 750 and both 245 and 368 at 800. The phone had been idle about 20 minutes before I noticed the blue Android guy so I
 
I never had problems with 122 Mhz min, it also helps that I'm not using a governor that locks the sleep frequency to the min either (like interactiveX). Besides a good governor should ramp it pretty fast to max if the demand is there.

I lowered my minimum down to 61 MHz and was about to report that all is fine but I just now looked down at my phone to see the little blue Android on a skateboard. Using ondemand and had 61 at 725, 122 at 750, 184 at 750 and both 245 and 368 at 800. The phone had been idle about 20 minutes before I noticed the blue Android guy so I
 
LOL, turns out CM's new re-coded interactive governor is broken (it locks to highest cpu freq). Reset stats in CPU spy, then try refresh a few times with screen on & off). Everything else seems to work fine though, including if I switch to ondemand governor & set min to 61 Mhz.

I don't see that... My cpu min/max as I've said before is 61/1708.

Since I saw your post, I changed my governor to Interactive, rebooted my phone, wiped cpu spy stats, and then turned off the screen. My stats:

460 MHz: 0:01:25
Deep Sleep: 0:12:27

No other used clocks.
 
Ah, I see how you got cpufreq_interactive.c compiled, thanks :) I'm going to include this with my next ROM build, along w/ my experimental fix for the "occasional can't go into deep sleep" bug, basically moving line 2490 in /drivers/mmc/host/msm_sdcc.c to outside of the if(mmc) check. It looks like the whole subroutine doesn't do anything otherwise. I haven't seen the bug in about 2 days so far.

Code:
}
wake_unlock(&host->sdio_suspend_wlock);
QUOTE]

I took a look at the function with this line. It doesn't seem like it will hurt anything to move it outside the If statement.

I'll check back in a few days with you to see how your testing is going with this and if it still looks good, I'll drop it into the next release.
 
When I first switched to interactive, it pegged on 1.51. After I saw your post, I rebooted and it seemed to operate correctly.
 
Guys, once again, thank you!
I tried to look into the idea of writing some of my own code. I think I hurt my brain, felt somthing pop. I wish I could just change the voltage and frequency tables on my wife's phone. She's running stock + the OV kernel from pwnyourace, but that doesn't have voltage control. I would use the one by b_random, but his seems to have a battery drain issue.
On top of all of that, my wife doesn't want me messing with her phone

so, shhh :-)
 
I don't see that... My cpu min/max as I've said before is 61/1708.

Since I saw your post, I changed my governor to Interactive, rebooted my phone, wiped cpu spy stats, and then turned off the screen. My stats:

460 MHz: 0:01:25
Deep Sleep: 0:12:27

No other used clocks.

Doesn't that seem odd? only 460 & deep sleep? it should be mostly deep sleep (except if just reset stats), followed by min freq, then max, and also a few in between since it goes through the freq tables briefly.
 
Doesn't that seem odd? only 460 & deep sleep? it should be mostly deep sleep (except if just reset stats), followed by min freq, then max, and also a few in between since it goes through the freq tables briefly.

Not really. That was only a sample of 13 minutes... And that 1 minute at 460, I attribute to the time while I was clearing the stats and turning off my wifi and the screen was on; these were all activities that should not require the cpu to ramp up too much--remember the load limit before ramping up the cpu also got upped with this update for Interactive.

I guess it just needs more data before we can say one way or the other.
 
I don't think it's working properly, prev version, interactiveX, & ondemand governors all idle at the low freq when I just wait at the cpu spy screen, hit refresh every 5 secs and the low freq time goes up by about the same. it shouldn't have any demand for higher when Im just waiting on a static screen. Yours looks like it s idling at 460 mhz when it should be lower. Mine is stuck at 1024 until screen goes off, then deep sleep registers.

I think there are other dependencies in the CM kernel that prevents it from working properly in the MT, which is why it's behaving so oddly on each of our phones. Tried rebooting multiple times and its just stuck at high freq for me.
 
I don't think it's working properly, prev version, interactiveX, & ondemand governors all idle at the low freq when I just wait at the cpu spy screen, hit refresh every 5 secs and the low freq time goes up by about the same. it shouldn't have any demand for higher when Im just waiting on a static screen. Yours looks like it s idling at 460 mhz when it should be lower. Mine is stuck at 1024 until screen goes off, then deep sleep registers.

I think there are other dependencies in the CM kernel that prevents it from working properly in the MT, which is why it's behaving so oddly on each of our phones. Tried rebooting multiple times and its just stuck at high freq for me.

Yeah, after more testing, there definitely looks like there's something not behaving properly. At least it does go into deep sleep... :) I'll have to go through the code again.
 
I have been running the previous v1.5 for about a week with a min of 245, max 1024, smartassv2 governor; no issues. Just to see if I had problems I lowered the min to 122 around noon yesterday. I noticed nothing, but I had no way to monitor it.

Yesterday evening when I dropped v1.6.11 on top, lowered my min to 61, and installed cpuspy.

While watching a movie it seemed to show almost 2 hrs of deep sleep. It was on power while I slept and didn't ever seem to go into deep sleep, but did log almost 9 hrs @61. In 14.5 hours I've had no problems. It spent 42min @1024, 1:17:21 @245, 9hrs @61, 2hrs in deep sleep; the rest of the time is spread across other frequencies less than 20 minutes in each.

I would say that smartassv2 is doing a great job of keeping the cpu as low as possible. I suspect that perhaps "deep sleep" is not possible when charging? That is the only place looks like it could have done better.

I did not receive any calls or texts during this test. I played a couple of games before bed, to test the tilt sensors. But it all seem stable from my brief sample.

Just providing another reference point for the group to help isolate interactions with other tweaks, that might be interfering with this new kernel. I'll report anything new that I find, but I suspect I'll be trying the new reloaded ROM before the weekend.
 
I have been running the previous v1.5 for about a week with a min of 245, max 1024, smartassv2 governor; no issues. Just to see if I had problems I lowered the min to 122 around noon yesterday. I noticed nothing, but I had no way to monitor it.

Yesterday evening when I dropped v1.6.11 on top, lowered my min to 61, and installed cpuspy.

While watching a movie it seemed to show almost 2 hrs of deep sleep. It was on power while I slept and didn't ever seem to go into deep sleep, but did log almost 9 hrs @61. In 14.5 hours I've had no problems. It spent 42min @1024, 1:17:21 @245, 9hrs @61, 2hrs in deep sleep; the rest of the time is spread across other frequencies less than 20 minutes in each.

I would say that smartassv2 is doing a great job of keeping the cpu as low as possible. I suspect that perhaps "deep sleep" is not possible when charging? That is the only place looks like it could have done better.

I did not receive any calls or texts during this test. I played a couple of games before bed, to test the tilt sensors. But it all seem stable from my brief sample.

Just providing another reference point for the group to help isolate interactions with other tweaks, that might be interfering with this new kernel. I'll report anything new that I find, but I suspect I'll be trying the new reloaded ROM before the weekend.

Thanks for the info. Smartassv2 should be fine nothing was changed for that governor between the two versions.

It was only Interactive using the new code from CM that seems to be having an issue.

And if you're charging, I wouldn't expect the phone to go into deep sleep. Pretty much any time that you have you're phone hooked up to something, I wouldn't expect it go into deep sleep.
 
Back
Top Bottom