Development [Kernel][24.09.2023][Android 13.0.0]Kirisakura 1.1.5_Raviantah for Pixel 6/Pro aka "RAVIOLE"

[Freak07]

by magisk i meant using FKM/EXKM, sorry a little mishap. Thank you anyway.

All good, just wanted to be precise so nobody else gets confused.

Those should work fine as well or use the excellent kernel flasher app from @capntrips , it's open source and available here on XDA.

Let's you easily backup all relevant partitions.

 

[Freak07]

Running 1.0.5_1

Thanks for the log, something is changing your ZRAM size from 3 to 4 GB.
Or rather it sets the size from 3GB to 0GB, then from 0GB to 4GB.

Do you use any mods, tweaks, optimizers or set that yourself in EXKM? (just because I see a selinux denial from exkm, just before the ZRAM stuff happens)

Please try without changing ZRAM size. The problem is EXKM does that very very late during boot, which might cause the issue you experienced.
I came across several devices that experience crashes when ZRAM size is changed that way after ZRAM was set up initially by the system and pages were swapped out already. It shouldn´t be done that way really.

 

[Freak07]

Thanks for the log, something is changing your ZRAM size from 3 to 4 GB.
Or rather it sets the size from 3GB to 0GB, then from 0GB to 4GB.

Do you use any mods, tweaks, optimizers or set that yourself in EXKM? (just because I see a selinux denial from exkm, just before the ZRAM stuff happens)

Please try without changing ZRAM size. The problem is EXKM does that very very late during boot, which might cause the issue you experienced.
I came across several devices that experience crashes when ZRAM size is changed that way after ZRAM was set up initially by the system and pages were swapped out already. It shouldn´t be done that way really.

okay, I checked all the logs previously shared.

All of those that crash right after boot since I said the trace of the crash changed with this issue have ZRAM size changed. the other logs don´t reach that far back, but I´m suspecting something similar.
Shame I didn´t catch that previously.


So to everyone, please do not use anything that changes ZRAM size or alters ZRAM behaviour (kernel managers, magic kernel tweak scripts, other mods, mystic kernel improve apps etc), flash the latest test shared here and post the relevant logs mentioned in case of a crash:

[Kernel][24.09.2023][Android 13.0.0]Kirisakura 1.1.5_Raviantah for Pixel 6/Pro aka "RAVIOLE"

Kirisakura-Kernel for the Pixel 6/Pro Hello everyone, To keep it short: Here is Kirisakura - Kernel for the Google Pixel 6 Pro aka Raven and the Pixel 6 aka Oriole, together Raviole. I would appreciate if everybody that flashes the kernel...

forum.xda-developers.com

 

[michaely123]

Thanks for the log, something is changing your ZRAM size from 3 to 4 GB.
Or rather it sets the size from 3GB to 0GB, then from 0GB to 4GB.

Do you use any mods, tweaks, optimizers or set that yourself in EXKM? (just because I see a selinux denial from exkm, just before the ZRAM stuff happens)

Please try without changing ZRAM size. The problem is EXKM does that very very late during boot, which might cause the issue you experienced.
I came across several devices that experience crashes when ZRAM size is changed that way after ZRAM was set up initially by the system and pages were swapped out already. It shouldn´t be done that way really.

I just put the ZRAM to stock. I guess I set that up a long time ago with EXKM, but I still ended up crashing. Thanks again @Freak07

 

[digger16309]

I always flash with EXKM, but then disable it to prevent its old settings getting applied. But I think I haven't re-booted after that in the past. The ZRAM is set at 3GB but it could be something else in EX that has triggered it.

 

[Justarandomguy]

I don't want to talk too soon but with last update phone did not crash a single time, usually it crashed at least once a day, or after a reboot it was almost 50/50 chance it will crash in a minute, using it since updated, rebooted few times (low battery, going to sleep and so, nothing kernel related) and 0 crashes still. Didn't update or change anything, so it's all kernel related, as always, thanks for your amazing job, you made this device way more battery efficient and so smooooooth

 

[falsehhhhh]

No reboot

 

[Freak07]

I just put the ZRAM to stock. I guess I set that up a long time ago with EXKM, but I still ended up crashing. Thanks again @Freak07

thanks again for the log! please check the next post.

 

[Freak07]

Test Release 1.0.5_3

Hey everyone,

so it seems the probability of the crash is greatly reduced and since the last test the trace changed again, which is good, because it means I´m on the right track.
Additionally to that it seems most users that were affected by the crash changed their ZRAM size via EXKM or any other kernel app. I think this was probably the main cause for the issue to get triggered in the first place.
Thanks to @michaely123 ´s log that shed some further light onto it, here´s the next test build.

Thanks to everyone for flashing all those tests, keeping up with me, providing logs and feedback! all of that is valuable to me, since I still can´t reproduce the crash on my end.

So I´d like to ask, just like the last time:
I´d appreciate if as much users as possible could flash and run this. (especially @michaely123 )
I´d appreciate if those that suffered regular crashes before could flash as well.
In case you get a random reboot please zip up and share the contents of sys/fs/pstore once you phone booted up again just like so many did before, which is great for me to see!



Thanks to everyone flashing those tests and providing logs in case of a crash, you guys absolutely rock!

 

[falsehhhhh]

1.0.5.1 Need to upgrade to 1.0.5.3 ？

 

[Freak07]

1.0.5.1 Need to upgrade to 1.0.5.3 ？

Probably better yeah

 

[michaely123]

Running 1.0.5_3. I'm not sure if this log will help, as the phone rebooted 3 hours ago.

*Update*: Second file just happened right now.

 

[digger16309]

No reboots on the last two test releases and mine always occurred within the first 10 minutes of the first boot and none thereafter. I'm convinced mine were an issue with EXKM settings applied at boot that caused a crash.

 

[cyberbbc]

my phone heats up all the time and loses charge pretty quickly, unlike the stock core. I don't quite understand why, it was much better on the early version of the kernel.
such a temperature while I was typing ....

 

[Freak07]

my phone heats up all the time and loses charge pretty quickly, unlike the stock core. I don't quite understand why, it was much better on the early version of the kernel.
such a temperature while I was typing ....
View attachment 5918529

It´s not related to the kernel.

It´s probably an app or a service always running in background.
You can check for this via top command in terminal. Please search the thread, you´ll find the instructions.

Or your ambient temperature is 10°C higher, because it might be summer where you live now.

 

[Rachid.Ala]

my phone heats up all the time and loses charge pretty quickly, unlike the stock core. I don't quite understand why, it was much better on the early version of the kernel.
such a temperature while I was typing ....
View attachment 5918529

Some app might be preventing your phone to deep sleep.
Check the wakelocks with betterbatterystats app or similar. You'll find detailed instructions if you search betterbatterystats thread

 

[Freak07]

This kernel will derp your installation for any other kernels including the original boot image. Conservative requires a higher min CPU locked frequency than 300mhz so it is necessary as schedutil is a jumpy sob. Sometimes, we'll find schedutil is tuned to only select specific frequencies like maybe 3 because the other frequencies are less efficient. I encourage everyone to experiment with plain schedutil as well. Temperatures can be high as well when kernel boosts to 1.8Ghz the max boost freq. Basically hibernate any programs and check task manager for any processes that stick out, then check to see if load looks like nothing is happening maybe your only consuming 180ma.

This is complete nonsense.

The steps to go back to stock are described on the FAQ.

Changing the CPU governor to any other than the default will introduce a lot of breakage in the framework and essentially cripple the phone.

 

[ada12]

kernel may derp next installation for other kernels. As you may experience a freeze on your next kernel. Just reflash everything and wipe data factory reset. Conservative should use a higher min CPU locked frequency than 300mhz. Schedutil is a jumpy sob. Sometimes, we'll find schedutil is tuned to only select specific frequencies like maybe only 3 because the other frequencies are less efficient. I encourage everyone to experiment with schedutil as well as conservative. Temperatures can be high as well when kernel boosts to 1.8Ghz the max boost freq. Basically hibernate any programs and check task manager for any processes that stick out, then check to see if load looks like nothing is happening maybe your only consuming 180ma.

using conservative governor is very bad idea ! its kill literally Energy Aware Scheduling and you introduce many issue ...

also change to schedutil is even no recommanded google have improved scheduler , governor is called "sched_pixel" in settings see https://github.com/freak07/Kirisaku...1ab44e8336/drivers/soc/google/vh/kernel/sched

 

[digger16309]

Do not let a developer convince you their kernel has no relationship on your temps/performance.

Do you even have this device or use this kernel?

Thousands of us have no heat issues with this kernel for over 18 months. If we did, the thread would be flooded with such complaints. Of course it's a non-kernel user issue.

 

[ada12]

Conservative works by keeping the frequencies as low as possible. More battery, right? Schedutil also does this in milliseconds, yet it be too aggressive on the selections for me.

No conservative gov sets the CPU frequency depending on the current usage.

schedutil (* pixel_sched ) is lot more adapted for multi socs like Tensor .

If you have issue don't spam dev please...