[KERNEL] Sub501 Kernel, mods and control apk

[Guest NeoEpsilon]

try

iwconfig --help

just to make sure its not the binary that crashes the system anyway

try

su

try

iwconfig eth0 tx 20dBm

try

iwconfig eth0 tx 10dBm

can't say anything to the clockspeeds as i have now the newer kernel 2.6.32.24-sub501

Oh man...

Stupid mistake... :)

I forgot to su in Terminal Emulator...

Quite surprised that it rebooted the phone instead of throwing a permission error though...

All is good now...

Thanks for the fast reply and help... :)

[Guest woti17]

@sub501

do you think now with the kernel sources, it will be possible to get ethernet over usb working?

not thetering, but reverse thethering: using the pc or netbook that has an internet-connection over eth0 or wlan0, to give

the phone access to the internet over usb cable (not the opposite, using the phones connection for the pc!),

without the android recognizing that it is not connected over wifi or 3g, and disabling market etc ;-)

(a linux kernel should have driver's for rndis, ethernet, cdc?)

[Guest Sub501]

@sub501

do you think now with the kernel sources, it will be possible to get ethernet over usb working?

not thetering, but reverse thethering: using the pc or netbook that has an internet-connection over eth0 or wlan0, to give

the phone access to the internet over usb cable (not the opposite, using the phones connection for the pc!),

without the android recognizing that it is not connected over wifi or 3g, and disabling market etc ;-)

(a linux kernel should have driver's for rndis, ethernet, cdc?)

Yes , I know, played a little with rndis. I can't give you a straight answer right now but why not possible. So we should have an iface then default route set to it ....

Hmm... managed to do something but is a little bit complicated. Keep digging...

Edited October 29, 2010 by Sub501

[Guest woti17]

Yes , I know, played a little with rndis. I can't give you a straight answer right now but why not possible. So we should have an iface then default route set to it ....

Hmm... managed to do something but is a little bit complicated. Keep digging...

maybe this could be interesting

/sys/devices/platform/msm_hsusb/usb_function_switch

from a script i found in an rom i don't know which it was

echo -n "Enabling usb-network..."

echo 4 > /sys/devices/platform/msm_hsusb/usb_function_switch

ifconfig usb0 $IP

i tried it

echo 4 > usb_function_switch ; cat usb_function_switch

and the content of usb_function_switch changed from

ether:disable

usb_mass_storage:enable

adb:enable

diag:disable

modem:disable

serial:disable

projector:disable

to

ether:enable

usb_mass_storage:disable

adb:disable

diag:disable

modem:disable

serial:disable

projector:disable

(0..all disable, 1..massstorage, 2..adb, 3..mass+adb, 4..ether, 5..eth+mass, 6..eth+adb, 7..eth+mass+adb ... seems to be a binary 0,1 pattern)

Edited October 29, 2010 by woti17

[Guest Sub501]

maybe this could be interesting

/sys/devices/platform/msm_hsusb/usb_function_switch

from a script i found in an rom i don't know which it was

echo -n "Enabling usb-network..."

echo 4 > /sys/devices/platform/msm_hsusb/usb_function_switch

ifconfig usb0 $IP

i tried it

echo 4 > usb_function_switch ; cat usb_function_switch

and the content of usb_function_switch changed from

ether:disable

usb_mass_storage:enable

adb:enable

diag:disable

modem:disable

serial:disable

projector:disable

to

ether:enable

usb_mass_storage:disable

adb:disable

diag:disable

modem:disable

serial:disable

projector:disable

(0..all disable, 1..massstorage, 2..adb, 3..mass+adb, 4..ether, 5..eth+mass, 6..eth+adb, 7..eth+mass+adb ... seems to be a binary 0,1 pattern)

did you try to set the default route to usb then in PC (linux) to NAT the packets?

Edited October 30, 2010 by Sub501

[Guest Sub501]

Great kernel Sub501 :(...

BTW, I use your 2.6.32.15 kernel...

Using the default init mods, the CPU was supposed to be set @806Mhz MAX when the battery is <30%, correct..?

Then again, Android System Info and Quadrant reported the CPU @768Mhz... :)

In Android System Info, the reported "available" speed was the following...

245

384

422.4

460.8

537.6

576

614.4

652.8

691.2

729.6

768

844.8

921.6

998.4

1075.2

1113.6

A few steps are missing... :)

Is this the result of adding a new steps..?

I've been wondering is this reported clock the "real" clock, or merely a false report...

For example, the CPU was running @806Mhz and yet reported @768Mhz...

Secondly, how can we know how many mW/dB the Wi-Fi radio is using @stock and after using your mod..?

Thanks for your time and great work...

Yeah 806 is not listed in cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_frequencies .

Have to find out why.

[Guest woti17]

did you try to set the default route to usb then in PC (linux) to NAT the packets?

one step further

on the phone i unchecked wifi and 3g, connected usb cable between pc and phone

then i did in terminal

ifconfig eth0 down

echo 7 > /sys/devices/platform/msm_hsusb/usb_function_switch

ifconfig usb0 192.168.178.73 up

on the linux netbook i did

ifconfig eth0 down

ifconfig wlan0 down

ifconfig usb0 192.168.178.10 up

now i can ping the phone from pc and the pc from the phone

- which makes me happy, because i never got to that point until now

the nat will not work i think, because the pc is already behind a nat

now i will try to make a usb/eth bridge on the pc again, and then find out how to setup dns working on the phone (think its a set property)

Edited October 30, 2010 by woti17

[Guest Sub501]

the nat will not work i think, because the pc is already behind a nat

between you phone and PC is a private class C network different than the PC-ROUTER network. NAT will work, your PC will became router for your phone.

[Guest Sub501]

Final version of Kernel 2.6.32.25 is available in the first post.

- Added smartass governor

- Added VDD patch

- Available freqs: 128000 245000 384000 422400 460800 499200 537600 576000 614400 652800 691200 729600 768000 806400 844800 883200 921600 960000 998400 1036800 1075200 1113600

Edited October 31, 2010 by Sub501

[Guest Sub501]

Great kernel Sub501 :(...

BTW, I use your 2.6.32.15 kernel...

Using the default init mods, the CPU was supposed to be set @806Mhz MAX when the battery is <30%, correct..?

Then again, Android System Info and Quadrant reported the CPU @768Mhz... :)

In Android System Info, the reported "available" speed was the following...

245

384

422.4

460.8

537.6

576

614.4

652.8

691.2

729.6

768

844.8

921.6

998.4

1075.2

1113.6

A few steps are missing... :)

Is this the result of adding a new steps..?

I've been wondering is this reported clock the "real" clock, or merely a false report...

For example, the CPU was running @806Mhz and yet reported @768Mhz...

Secondly, how can we know how many mW/dB the Wi-Fi radio is using @stock and after using your mod..?

Thanks for your time and great work...

Fixed in 20101031 (see post #1). Available freqs: 128000 245000 384000 422400 460800 499200 537600 576000 614400 652800 691200 729600 768000 806400 844800 883200 921600 960000 998400 1036800 1075200 1113600

[Guest Sub501]

sub501 have you look at the smartass govenor i post some links for you yesterday

Available in 20101031. See post #1

[Guest woti17]

Available in 20101031. See post #1

installed.

what do i have to do to try the brainfuck-scheduler (bfs)?

can smartass govenor been mixed up with other govenors or is it: all except smartass mixed _or_ smartass?

sounds kind of funny btw ;-)

Edited October 31, 2010 by woti17

[Guest Sub501]

installed.

what do i have to do to try the brainfuck-scheduler (bfs)?

can smartass govenor been mixed up with other govenors or is it: all except smartass mixed _or_ smartass?

sounds kind of funny btw ;-)

Do you have a patch for BFS?

What do you mean by mixed? You have governors in kernel, only one governor can be active at a time.

[Guest woti17]

Do you have a patch for BFS?

What do you mean by mixed? You have governors in kernel, only one governor can be active at a time.

i meant, i can switch between powersave and ondemand for example for screen off/screen on

but is it useful ie to switch between smartass and powersave, or is it better if you use smartass, to use only smartass because its optimiszed for all circumstances (ie screen on/off, various battery charges, etc)

bfs, i thank the ziggy kernel at xda has bfs, looked at the config

[Guest Sub501]

i meant, i can switch between powersave and ondemand for example for screen off/screen on

but is it useful ie to switch between smartass and powersave, or is it better if you use smartass, to use only smartass because its optimiszed for all circumstances (ie screen on/off, various battery charges, etc)

bfs, i thank the ziggy kernel at xda has bfs, looked at the config

Yeah you can mix this way. /data/etc/powermanagement.conf will let you define whatever gov you want for screen on/off

I don't know which is better, test and post the results. I'm still using ondemand/powersave schema

A good friend of mine pointed http://ck.kolivas.org/ . The BFS patch is there, next kernel update? 1-2 weeks.

Edited November 1, 2010 by Sub501

[Guest TheUntouchable]

Sub501, perhaps this app is interesting for you:

CurrentWidget

http://forum.xda-developers.com/showthread.php?t=797037

[Guest Sub501]

Sub501, perhaps this app is interesting for you:

CurrentWidget

http://forum.xda-developers.com/showthread.php?t=797037

Useful tool, thanks

[Guest Zarch1972]

Sub501, perhaps this app is interesting for you:

CurrentWidget

http://forum.xda-developers.com/showthread.php?t=797037

Its showing 4 to 6mA when screen off for me.

Running sub501 kernel from 29 Oct (2.6.32.15?) - before latest update.

Radio: 32.48.00.32U_5.10.05.30 (this has made a nice subtle difference to my battery over 32.44.00.32U_5.09.05.30_2 from the R9 kitchen)

I assume that is all hunky dory?

Edited November 2, 2010 by Zarch1972

[Guest TheUntouchable]

As you can see i have the same results :)

Phone: HTC Desire

Kernel: Sub501 Kernel 2.6.32.25

Mods: Sub501 Init-mods (Undervolted, Built-in governors, cfq and deadline ioschedulers)

ROM: MoDaCo 22/Oct r9 WIP

Radio: 32.48.00.32U_5.11.05.14

Standby: 4mA - 6mA

Screen on: 267mA - 471mA

[Guest Sub501]

As you can see i have the same results :)

Phone: HTC Desire

Kernel: Sub501 Kernel 2.6.32.25

Mods: Sub501 Init-mods (Undervolted, Built-in governors, cfq and deadline ioschedulers)

ROM: MoDaCo 22/Oct r9 WIP

Radio: 32.48.00.32U_5.11.05.14

Standby: 4mA - 6mA

Screen on: 267mA - 471mA

Try to lower voltages in /data/etc/undervolt.conf :

-25

Save, reboot then cat /proc/undervolt to see the new voltages. Post the results of monitoring.

I have screen bright set to manual ~40%. With screen on, ondemand 245 - 1Ghz:

Screen on: 121 - 210mA (using default undervolt provided in mod update.zip)

Edited November 2, 2010 by Sub501

[Guest Zarch1972]

Try to lower voltages in /data/etc/undervolt.conf :

-25

Save, reboot then cat /proc/undervolt to see the new voltages. Post the results of monitoring.

I have screen bright set to manual ~40%. With screen on, ondemand 245 - 1Ghz:

Screen on: 121 - 210mA (using default undervolt provided in mod update.zip)

Firstly, I must say that using vi in terminal emulator is a right laugh.... good job I use vi every day at work!

AMOLED Screen (original Desire)

This afternoon with the standard sub501 kernel and configs: (no changes at all)

Screen off 4mA - 6mA

max 556 mA (not sure what i was doing at 17:17, but the Current Widget log says 556mA), but regularly in the 300-450 range.

Just now with -25 undervolt in place:

Screen off 4mA - 6mA

Screen on with Angry Birds playing : 303mA

Will continue to monitor and see if I get anything above 303mA with the -25 change in place.

Edited November 3, 2010 by Zarch1972

[Guest woti17]

amoled

kernel 2.6.32.25

powersave 245/ondemand max 576-998 depending on battery-charge

wifi tx power reduced 10mW

rom 2.14.207.1

baseband 32.48.00.32U_5.11.05.14

cifs+configs module

with sub501 unmodified undervolt.conf and additional -25mV, doing nothing special

i get with with screen off: 3mA-6mA

with screen on 10%/wifi on about 98-140mA

with screen on 40%/wifi on about 165-170mA

950 925 19200

950 925 128000

950 925 245000

975 950 384000

975 950 422400

1000 975 460800

1025 1000 499200

1025 1000 537600

1050 1025 576000

1075 1050 614400

1100 1075 652800

1125 1100 691200

1150 1125 729600

1175 1150 768000

1200 1175 806400

1200 1175 844800

1225 1200 883200

1225 1200 921600

1250 1225 960000

1250 1225 998400

Edited November 3, 2010 by woti17

[Guest NeoEpsilon]

I believe we also need to post the display type... :)

AMOLED and SLCD should have different power consumption...

HTC Desire-SLCD

Sub501 Kernel 2.6.32.25

Radio 32.44.00.32U_5.09.05.30_2

My own ROM (2.13.707.1)

Brightness 35% + Wi-Fi ON

Undervolt+Interactive Governor (I only copied and load undervolt.ko, configs.ko, bcm4329.ko)

Screen ON, Idle 152-159mA

Screen OFF, Idle 6-7mA (never lower than 6mA)

Angry Birds, 296-310mA

If I recall correctly, using "original" kernel and Interactive...

I've seen an average of 145mA @Idle, Screen OFF is the same 6mA though...

This is interesting because when using Sub501, I'm undervolting aggresively...

950 925 19200

950 925 128000

950 925 245000

975 925 384000

975 925 422400

1000 950 460800

1025 975 499200

1025 975 537600

1050 975 576000

1075 1000 614400

1100 1025 652800

1125 1050 691200

1150 1075 729600

1175 1075 768000

1200 1125 806400

1200 1125 844800

1225 1175 883200

1225 1175 921600

1250 1175 960000

1250 1175 998400

Any thought..? :)

I haven't got any time to test previous ROM because I deleted my nandroid backup...

Edited November 3, 2010 by NeoEpsilon

[Guest Sub501]

Any thought..? :)

I haven't got any time to test previous ROM because I deleted my nandroid backup...

I don't get it? Is it something not ok or?

[Guest Sub501]

In my opinion the mA will vary just because we also have different applications running in the background. For example during my facebook/twt background sync or k9mail direct push check the consumption is raising. In the end different usages will generate different results.