@akhilnarang it would be great to see some kind of changelog when you release![]()
Merged in linux-stable v4.4.146
Wireguard 20180802 snapshot
Fixed most compile-time warnings with Clang 7
Thanks for your hard work!v1.9 is up! (Sorry forgot to post yesterday itself)
https://github.com/akhilnarang/whyred/releases/tag/Derp-v1.9
Changelog:
Code:Merged in linux-stable v4.4.146 Wireguard 20180802 snapshot Fixed most compile-time warnings with Clang 7
I want update to 1.9, do I have to flash it again or can it go directly?
Does this kernel work on android 9 (pie)?
Very much, yes. It works flawlessly and compatible with whyred P GSI.
I also had this several times in several roms. And also maybe on another kernel. Maybe it has something to do with magisk, if you use it. Maybe remove magisk, clean Flash kernel, boot, then reboot to recovery and Flash magisk backMy minimum frequency on little cores gets reset to 1401mhz even though I set it to 633 mhz, after a minute or two it goes to 1401mhz again. I did a clean flash of 5th August AOSiP and had even on stock kernel, and on this update again. I don't get it, it was okay on previous rom+kernel builds.
It's not get reset. Check freq stats after few days of normal usage. You'll see.My minimum frequency on little cores gets reset to 1401mhz even though I set it to 633 mhz, after a minute or two it goes to 1401mhz again. I did a clean flash of 5th August AOSiP and had even on stock kernel, and on this update again. I don't get it, it was okay on previous rom+kernel builds.
I want update to 1.9, do I have to flash it again or can it go directly?
I have a poor notification led pulse, it's not smooth
I heard many units has this issue, mine is global RN5
I'm currently on LOS, can it be fixed with some kernel modifications?
Thanks for your hard work!
Any chance to include vibration strengh control?
Thanks!
Had this while using latest aosip official (July), flashed aex and aosip beta of 5th August after that, neither of them have the problem, I think it's something related to the official build of July cause I even tried other kernels and the problem remainedMy minimum frequency on little cores gets reset to 1401mhz even though I set it to 633 mhz, after a minute or two it goes to 1401mhz again. I did a clean flash of 5th August AOSiP and had even on stock kernel, and on this update again. I don't get it, it was okay on previous rom+kernel builds.
Changelog:
Merged in linux-stable v4.4.147
Merged in CAF LA.UM.6.2.r1-10300-sdm660.0
Changelog:
Merged in linux-stable v4.4.147
Merged in CAF LA.UM.6.2.r1-10300-sdm660.0
POPP?- Disable PPP
Linux is not realtime. 1000Hz only causes timing mismatches.- 1000hz (scheduler becomes faster to respond to realtime tasks)
No such thing as realtime.- Disable RCU Boost (this causes overhead and jitter for realtime tasks)
You don't even have schedutil...- Schedutil iowait boost
MSM drivers have round dependencies on it.- Disable Ethernet
Fair enough- Disable Audit
You don't even have EAS...- Dynamic Schedtune Boosting
Why?- CPU Boost
Used by perfd.- Remove MSM Performance
You'd rather have a device hung indefinitely over something that- Remove Watchdog
What is this even?- Disable IRQ helper
Fair- Increase KGSL thread priority
Do you even know what you're talking about...- Proper HZ value for userspace
Absolute garbage.- Wakelock blocking
- Adrenoboost
Barely any gain.- Boost adreno tz work queue priority
Care to elaborate?- Disable MASSIVE amounts of logging
Really?- Remove MENU idle governor (we use QCOM)
./drivers/cpuidle/Kconfig:7: select CPU_IDLE_GOV_MENU if (NO_HZ || NO_HZ_IDLE)
$ grep NO_HZ .config
CONFIG_NO_HZ_COMMON=y
CONFIG_NO_HZ_IDLE=y
# CONFIG_NO_HZ_FULL is not set
CONFIG_NO_HZ=y
CONFIG_RCU_FAST_NO_HZ=y
Have you EVER looked at the GitHub repository?- QCACLD-3.0 latest tag
- FW-API latest tag
- QCA-WIFI-HOST-CMN latest tag
- Merged latest CAF tags
Garbage. These drivers don't need priority inheritence and will only break when forced to use it.- Child process priority (better for realtime tasks)
- Use rt_mutex for fbdev
- Use rt_mutex for fingerprint
- Use rt_mutex for ashmem
- Use rt_mutex for ion
Again, have you EVER looked at the GitHub repository or you just paste your handy- Latest stable SDCARDFS
NO.- I2C/SPI overclock (500mhz-->800mhz)
Qualcomm drivers love DebugFS.- Disable DEBUGFS
- Extreme binder modification (rt_mutex, removed debugging, cleaned up)
- Disable debug monitors
- Tune mm BDI ratios
- Use flat memory model
- Disable integrity subsystem
- Disable NFC remnants
- Optimize touchscreen driver for latency and performance
- Disable excessive logging in fingerprint drivers
- Add brightness flooring interface (use true min brightness)
- Performance critical IRQs for fingerprint, mdss, kgsl, touchscreen, etc
- DTS Eagle support
- Purify LMK from MIUI excecptions
- Remove unused compiled drivers
- Display hacks for better latency and sync
- Disable PSTORE
- Improve LED animation (kanged from Onyx)
- Net hacks for better latency
- Spoof verity and bootloader unlock status (green/locked)
- Wireguard VPN (latest)
- Better initial frequency for GPU
- Anxiety I/O scheduler
- Disable stack protector
- Remove some devfreq drivers
- Disable printk by default
- Disable gentle fair sleepers
- Use laptop mode by default (better batt)
- Disable some unused CGROUPs
- Enable NEON accelerated
- Chacha20 random
- Reduced cache pressure
- Optimized nr_requests for latency
- Reduced dirty ratios
- Disable UFS support
- Use fq_codel instead of pfifo_fast
- Optimized readahead (512KB)
- Spoof fast charging to enable quick charge 4 for all models
- Disable CABC (don't auto adjust brightness)
- Analog dimming
- NTFS R/W
- KCAL
- -Ofast
- Fast math computation
Would be nice
@akhilnarang seen that you got poco f1 I'm curious if you could share your plans about whyred support it the near future. Are you planning to "leave" whyred development and focus on poco f1?
Thank you in advance for all your efforts and work put into whyred!