Yank555.lu Android Stuff

TylonHH wrote:As far as I know you will always have a delay on non sammy roms. Because Samsung didn't published the mali driver.
You didn't had that delay on CM before?

---------------
gesendet vom Barhocker

yank555 wrote:

Hopesfall wrote:i tried that already, cranked it up to 600 Mhz. still getting delays :/

i flashed CM11 just because of this delay, i was hoping the delay would be completely gone.

Well a few things to consider :

1) stock screen off max is 1400

2) the wakeup lag is inherent to the Exynos platform (don't have it on Qualcomm)

3) governors behave differently, some rev up faster than others

Unfortunately there is no perfect medicine for wakeup lag, once the device is in deep sleep, it will not wake up instantly, and some battery saving additions may contribute to make it worse.

You need to find a balance that is acceptable for you, it will not likely get perfect on the i9300 ...

I've added a "wakeup kick" in my Note 3 kernel, which boosts scaling min for a certain time on wakeup, could help ... I'll try to port it to i9300 when I have spare time, but I don't think it will help, as I'm afraid it will be too late when it'll be activated (cf. point 2 above).

JP.

Hopesfall wrote:

TylonHH wrote:As far as I know you will always have a delay on non sammy roms. Because Samsung didn't published the mali driver.
You didn't had that delay on CM before?

---------------
gesendet vom Barhocker

I was on stock before, sammy roms had even bigger problems with the delay. although i must admit i was hoping CM would resolve it.

yank555 wrote:

Hopesfall wrote:i tried that already, cranked it up to 600 Mhz. still getting delays :/

i flashed CM11 just because of this delay, i was hoping the delay would be completely gone.

Well a few things to consider :

1) stock screen off max is 1400

2) the wakeup lag is inherent to the Exynos platform (don't have it on Qualcomm)

3) governors behave differently, some rev up faster than others

Unfortunately there is no perfect medicine for wakeup lag, once the device is in deep sleep, it will not wake up instantly, and some battery saving additions may contribute to make it worse.

You need to find a balance that is acceptable for you, it will not likely get perfect on the i9300 ...

I've added a "wakeup kick" in my Note 3 kernel, which boosts scaling min for a certain time on wakeup, could help ... I'll try to port it to i9300 when I have spare time, but I don't think it will help, as I'm afraid it will be too late when it'll be activated (cf. point 2 above).

JP.

Thanks for your reply I'll try those definitely.

Tell me, is there a way to passively disable deep sleep altogether? like not letting the processor go into deep sleep but instead keeping it at designated freq when the screen is off? I'm assuming that the first option does exactly that or?

I mean SGS3 is an awesome device even today and it's really sad to see that it has problems with this delay.

yank555 wrote:I believe there are apps that do that, but a wakelock will eat up a steady ~60mA/h ...

The delay is a hardware design issue, so except by impeeding deep sleep, I don't expect any good solution.

JP.

Send from my 'proudly eFused' Note 3 (n9005) running Temasek v75 on Yank555.lu htle v1.0a-beta1 (3.4.89) kernel.

Hopesfall wrote:

yank555 wrote:I believe there are apps that do that, but a wakelock will eat up a steady ~60mA/h ...

The delay is a hardware design issue, so except by impeeding deep sleep, I don't expect any good solution.

JP.

Send from my 'proudly eFused' Note 3 (n9005) running Temasek v75 on Yank555.lu htle v1.0a-beta1 (3.4.89) kernel.

I know there are apps to do it, but they are really unreliable(OS kills them), that's why i asked is there a way to disable it so it remains persistent?
I want to try it out for few days without deep sleep, if the battery can hold for one day i'll gladly disable it because i always charge my phone at night and in most cases the battery is well over 50% when i plug it in after a day's usage.

cornelha wrote:Hey Yank,

Can you please tell me which files need to be included in a repacked zip. I want to keep my configuration after every Omni nightly update. I know how to repack a kernel, just need to make sure about the configuration files needed for your kernel.

yank555 wrote:

cornelha wrote:Hey Yank,

Can you please tell me which files need to be included in a repacked zip. I want to keep my configuration after every Omni nightly update. I know how to repack a kernel, just need to make sure about the configuration files needed for your kernel.

The generated config file is : /system/etc/init.kernel.sh

And of course the boot image and kernel modules need to be in there

JP.

Code: Select all

ui_print(" ");
ui_print("*Yank555 Kernel*");
ui_print(" ");
show_progress(0.100000, 0);
show_progress(0.100000, 20);

ui_print("* Flashing kernel *");

assert(package_extract_file("boot.img", "/tmp/boot.img"),
show_progress(0.100000, 40);
write_raw_image("/tmp/boot.img", "/dev/block/mmcblk0p5"),
show_progress(0.100000, 60);
delete("/tmp/boot.img"));

ui_print("Extracting modules and config..");
show_progress(0.100000, 80);
run_program("/sbin/busybox", "mount", "/system");
assert(package_extract_dir("system", "/system"));
ui_print("Fixing Permissions..");
set_perm_recursive(0,0,0755,0644,"/system");
ui_print("Unmounting File System..");
run_program("/sbin/busybox", "umount", "/system");

show_progress(0.100000, 100);
ui_print(" ");
ui_print(" ");
ui_print("  * Installation Finished Succesful *");

cornelha wrote:I borked it somewhere, could you please have a look at my scipt?

Code: Select all

ui_print(" ");
ui_print("*Yank555 Kernel*");
ui_print(" ");
show_progress(0.100000, 0);
show_progress(0.100000, 20);

ui_print("* Flashing kernel *");

assert(package_extract_file("boot.img", "/tmp/boot.img"),
show_progress(0.100000, 40);
write_raw_image("/tmp/boot.img", "/dev/block/mmcblk0p5"),
show_progress(0.100000, 60);
delete("/tmp/boot.img"));

ui_print("Extracting modules and config..");
show_progress(0.100000, 80);
run_program("/sbin/busybox", "mount", "/system");
assert(package_extract_dir("system", "/system"));
ui_print("Fixing Permissions..");
set_perm_recursive(0,0,0755,0644,"/system");
ui_print("Unmounting File System..");
run_program("/sbin/busybox", "umount", "/system");

show_progress(0.100000, 100);
ui_print(" ");
ui_print(" ");
ui_print("  * Installation Finished Succesful *");

File Structrure is as follows
system/lib/modules/*
system/etc/init.kernel.sh
boot.img

It just doesn't boot. I'm pretty sure it's something very stupid somewhere that I missed.

yank555 wrote:

cornelha wrote:I borked it somewhere, could you please have a look at my scipt?

Code: Select all

ui_print(" ");
ui_print("*Yank555 Kernel*");
ui_print(" ");
show_progress(0.100000, 0);
show_progress(0.100000, 20);

ui_print("* Flashing kernel *");

assert(package_extract_file("boot.img", "/tmp/boot.img"),
show_progress(0.100000, 40);
write_raw_image("/tmp/boot.img", "/dev/block/mmcblk0p5"),
show_progress(0.100000, 60);
delete("/tmp/boot.img"));

ui_print("Extracting modules and config..");
show_progress(0.100000, 80);
run_program("/sbin/busybox", "mount", "/system");
assert(package_extract_dir("system", "/system"));
ui_print("Fixing Permissions..");
set_perm_recursive(0,0,0755,0644,"/system");
ui_print("Unmounting File System..");
run_program("/sbin/busybox", "umount", "/system");

show_progress(0.100000, 100);
ui_print(" ");
ui_print(" ");
ui_print("  * Installation Finished Succesful *");

File Structrure is as follows
system/lib/modules/*
system/etc/init.kernel.sh
boot.img

It just doesn't boot. I'm pretty sure it's something very stupid somewhere that I missed.

I will just assume mmcblk0p5 is the right partition, can't really check right now.

This :

set_perm_recursive(0,0,0755,0644,"/system");

is a bit excessive, you'll remove all execution access of every file (644 = rw-r--r--) in whole /system, so it just can't boot, nothing can be executed

Set permissions only where you changed stuff, or better, only on the files you change

JP.