[Kernel/Recovery][P75XX] A1 Kernel V1.9/TWRP2.6.3.0

kashortiexda said:

+1 ^^ battery charging symbol, though not plugged in

Sent from my GT-P7500 using Tapatalk 2

Click to expand...

Click to collapse

jaswinky said:

The boot problem is difficult to diagnose, it may depend, apart of the voltages, of how cold or hot is the CPU and the system, the ambient temperature, etc... That could explain why some people have the issue, and some people don't. The fact that 1.5.2 and more even 1.5.3 has fix the problem to most people, could mean that maybe there still is or are others possible voltages adjusts needed to do...

AAccount, Do yo mean 1.5.4 set ACmod by default and we can, as users, desactive it? how? Could you post last source code to try see if more voltage could be fixed?

Thanks!

Click to expand...

Click to collapse

jaswinky said:

Great! this way to manage USB charging is cleaner...

What do you mean with your tablet? Is it ok??

Click to expand...

Click to collapse

jaswinky said:

I don't really think it's even worth to overclock to 1.6GHz. I don't know any P75xx kernel to 1,6GHz in HC, so this whould be for any reason... The experiences I have been having with overclock is that you can do it safetly with good brands and devices, because the design is oversized to sure the people will not have any problem. Cheaper devices could not be too oversized. Example: my SGS2, with siyah kernel, I can put from 1,2GHz to 1,6GHz (33%), that's all. I tried 1,4Gz with no problem (I have clocked it to 1,2GHz because I don't need yet 1,4 or 1,6, so that I save battery); our tab, to 1,4GHz (40%) with no problem. However, I also have a LG Optimus Black overcloced from 1GHz to 1,1GHz, because 1,2GHz had a problem with apk installations from market . You cannot predict the effects of a bad overclocking... this booting bug is the same thing... maybe the voltage differences are in the limit and small difference in temperature (note that in those small voltages, the amperage is more sensible to small temperature changes because the impedances) may cause a insufficient suply...

Click to expand...

Click to collapse

falcon67 said:

Is there possibility to make 4.0.4 kernel for t mobile SGH-T859 galaxy tab 10.1. I was able to install aneroid 3.2 GT-P7500 zip rom on t mobile galaxy tab without no issues. With android 4.0.4 zip rom flash , t mobile tab 10.1 does not boot up at all.


Sent from my GT-P7500 using xda app-developers app

Click to expand...

Click to collapse

jaswinky said:

Consider that it was only my opinion about it... I think now, it is more important to improve more and more 1,4GHz than start with a 1,6GHz project some inestable... Maybe in some moths when all tablets will be quad at 1,5GHz, and the apks will be made for them, it would a great idea to start or continue this project...

That what happen for me with my SGS2, now I don't need 1,4GHz or 1,6GHz, but I know in some monthes when most apk will be design for SGS3, then these apks will go perfect for me in my SGS2 at 1,6GHz... so I prefer to wait for that...

Click to expand...

Click to collapse

steveman29 said:

Hello again i know we asked for many features and you added all of them. Could you please also enable swap so we can add a swap file with swapper? It would help a lot because of the low memory. Thank you in advance!

Sent from my GT-P7510 using Tapatalk 2

Click to expand...

Click to collapse

root@android:/sdcard # mkswap swap.disk
Setting up swapspace version 1, size = 33550336 bytes
UUID=94658bfe-a265-4aed-afaf-d2de1fc485f2
root@android:/sdcard # swapon swap.disk
swapon: swap.disk: Invalid argument
1|root@android:/sdcard # ls /data/local
tmp
root@android:/sdcard # mv swap.disk /data/local
failed on 'swap.disk' - Cross-device link
255|root@android:/sdcard # cp swap.disk /data/local
root@android:/sdcard # cd /data/local
root@android:/data/local # ls
swap.disk
tmp
root@android:/data/local # mkswap swap.disk
Setting up swapspace version 1, size = 33550336 bytes
UUID=bdf56bf0-7311-4d4c-a852-430cc6753198
root@android:/data/local # swapon swap.disk
root@android:/data/local # free 
             total         used         free       shared      buffers
Mem:        738984       726600        12384            0        97964
-/+ buffers:             628636       110348
Swap:        32764            0        32764
root@android:/data/local #

jaswinky said:

I think where could be the reason why still happen the freeze boot issue a few people (I read one here and one in htcmania):

Sorry, I don`t know how to insert code as you do...

Again the init function where initial voltages are assigned:

Code:

void __init tegra_soc_init_dvfs(void)
{
	...
	BUG_ON(speedo_id >= ARRAY_SIZE(cpu_speedo_nominal_millivolts));
	tegra2_dvfs_rail_vdd_cpu.nominal_millivolts =
		cpu_speedo_nominal_millivolts[speedo_id];
	BUG_ON(speedo_id >= ARRAY_SIZE(core_speedo_nominal_millivolts));
	tegra2_dvfs_rail_vdd_core.nominal_millivolts =
		core_speedo_nominal_millivolts[speedo_id];
	tegra2_dvfs_rail_vdd_aon.nominal_millivolts =
		core_speedo_nominal_millivolts[speedo_id];
...
}

Here is assigned the nominal voltages from the array due to the speedo_id. Ok, the problem is that tegra2_dvfs_rail_vdd_cpu.nominal_millivolts, tegra2_dvfs_rail_vdd_core.nominal_millivolts and tegra2_dvfs_rail_vdd_aon.nominal_millivolts were created with a default value at begining of the file. This mains that in any instant, even with a reduced probability, it could coexist a cpu nominal value assigned from speedo_id array and a core nominal default value. If so, see this:

Code:

static const int cpu_speedo_nominal_millivolts[] =
/* spedo_id  0,    1,    2 */
	{ 1100, [B][COLOR="Red"]1225[/COLOR][/B], 1125 };

static const int core_speedo_nominal_millivolts[] =
/* spedo_id  0,    1,    2 */
	{ 1225, 1275, 1300 };
...
static struct dvfs_rail tegra2_dvfs_rail_vdd_cpu = {
	.reg_id = "vdd_cpu",
	.max_millivolts = 1250,
	.min_millivolts = 900,
	.nominal_millivolts = 1175,
};

static struct dvfs_rail tegra2_dvfs_rail_vdd_core = {
	.reg_id = "vdd_core",
	.max_millivolts = 1300,
	.min_millivolts = 950,
	.nominal_millivolts = [B][COLOR="red"]1225[/COLOR][/B],
	.step = 150, /* step vdd_core by 150 mV to allow vdd_aon to follow */
};

Nominal values in bold and red could coexist... It is difficult, I know, but maybe because this reason only one or two people are still reporting the problem...

Maybe decrease first bold value from 1225 to 1175 as cpu default nominal value?, or increase second bold value to 1275 as nominal core array value?

Click to expand...

Click to collapse

decatf said:

Can I make a request that the first post also detail what voltage changes were made from the stock settings? That's pretty important when it comes to overclocking.

Click to expand...

Click to collapse