Skip to content

Releases: Nova-Kernels/kernel_xiaomi_mt6785

NoVA-4.14.329

13 Nov 06:22
Compare
Choose a tag to compare

Changelogs:
Upstreamed Kernel source to 4.14.329 - the latest revision
Updated KernelSU to latest version
Built with Neutron clang 18
[TREEWIDE] Optimise during build time
[TREEWIDE] Suppress logspams and stop debug tracers
[REVERT] DECnet support from kernel ( makes some roms to bootloop )

BACKPORT: mm: fix pageblock heuristic
Disable oom_dump_tasks
Disable syslimiter driver
Disable Debug and tracers from defconfig
Enable mtk_task_turbo
Enable MTK_PERFMGR_TOUCH_BOOST by default
Reduce CPU intensive features which are not used as much

arm64: make it support int128 for more efficiency
blkdev: disable iostats and add_random
batman-adv: some improvements
cpumasks: add big cpu masks
cpufreq: interactive: enable back and minor improvements
cpufreq: schedutil: enforce realtime priority
crypto: significant improvements in algorithm
devfreq: Introduce devfreq boost and adapt it to our gpu
dm-verity: enable high priority workqueues on verify_wq
fs-verity: significant improvements
ged: merge some ALPs
ged_ski: fix gpu_freq always returning to 0
gpio: significant improvements
gpu: sync wakeup on event handoff
gpu-bifrost: Reduce job rescheduling latency
kernel: fix cpufreq mem leaks
kernel: warn when IRQ's affinity is rewritten
kernel: allow wakeup IRQs to cancel ongoing suspend
kernel: introduce SBalance IRQ Balancer
kernel: irq: properly align perf IRQs and disallow userspace to change its affinity
mali: kbase: app adaptive power policy and enhance power policy
mm: do not reserve memory for users / admins
mm / bdi: Introduce SWP_SYNCHRONOUS_IO and fix further related issues
mm / vmstat: use power efficient wq
mtk-fpsgo: merge alps and mark wq as high-priority
net: improvements on ipv4/ipv6 protocols
PM / freezer: reduce timeout to 1 second
PM : Clear wakeups before running PM callbacks and abort suspend while wakelock is present
sched: cass: Introduce Capacity Aware Superset Scheduler and adapt it for us
sched: core: set nr_migrate to 32
sched: core: merge some alps
sched: disable ttwu_queue
sched: deadline: fix stale throttling on de-/boosted tasks
sched: fair: Allow idle cpu groups to pull tasks from overloaded group
sched: fair: fix load_balance and improve the algorithm significantly
sched: fair: some optimizations in the balancer and idle cores logic
sched: idle: improve the idle loop
sched: rt: do not push tasks if there are none
sensors: reduce gyroscope delays
SLMK: improve the algorithm
tcp: allow drivers to tweak TSQ logic
tcp: allow TLP in ECN
tcp: set initial window size to 64k (internet speed improvement)
thermal: core: skpi update disable thermal zones after being suspended
thermal: merge some ALPs
touchscreen: mark display processing as high priority wq
zram: do not allow the default algorithm to be changed
zram: allow overriding zram size from kernel
zstd: merge optimizations from Samsung