silver0480

@aprayoga @gprovost Tested self building whole image (dev) , with kernel opt "KASAN" . [/config/kernel/linux-rockchip64-dev.config] - # CONFIG_KASAN is not set + CONFIG_KASAN=y on this image , "sysctl -w vm.drop_cahches=3" PASSED, No kernel panic with stack overflow. # with KASAN , probably kernel stack size is incleased to 16KB -> 32KB # https://lore.kernel.org/patchwork/patch/1028055/ So , can you change KASAN kernel configuration for stable operation ? P.S. but , KASAN catched BUG on boot

@aprayoga Tested vdd via new system ( using Nightly build ). My issue ( karnel panic by "sysctl -w vm.drop_caches=3" ) isn't solved. ( https://forum.armbian.com/topic/16664-system-crashes-after-sync-command/?do=findComment&comment=117238 )

I want more small & lightweight chassis , like helios4.

@Chiefahol @clostro @snakekick Check dmsg (via serial console) at the moment of RED-Blinking. If kernel msg contains " [nnnnn.nnnnn] Unable to handle kernel paging request at virtual address " Try CPU speed limitation. But in my case , NOTHING solved with CPU speed. kernel msg contains "[nnnnn.nnnnn] Insufficient stack space to handle exception!" Then I tried Roll back to kernel 5.8.14 ( via armbian-config ) , got stable operation. on my helios64 , Any 5.9.x / 5.10.x kernels causes same critical issue ( including 5.10.9 with armbi

@gprovost @aprayoga Some extend test. same method. [environment] armbian focal 20.11.6 data dir : 18TB HDD btrfs + 16TB HDD btrfs (both single volume) [method] # sync ; sync ; sync ; sysctl -w vm.drop_caches=3 # ls -CFR [datadir] > [tmp dir] ( about 600,000 files ) # ls -1lFR [datadir] > [tmp dir] ( about 600,000 files ) # zip -j [tmpdir] filelist.zip # sync ; sync ; sync # sysctl -w vm.drop_caches=3 >>>>>>> Kernel Panic (blinking Red LED) [Result] Crashed : 5.10.

@gprovost tested some other kernels (via armbian-config). Looks stable with legacy kernel. Crashed : 5.10.0-rc7-rockchip64 Crashed : 5.9.14-rockchip64 (default kernel) Crashed : 5.9.10-rockchip64 PASSED : 4.4.213-rk3399 I will operate with a legacy kernel for the time being, but LAN LED can't blink with legacy kernel .......... so I want to use 5.x.

@gprovost Thanks. Already tested , but nothing gets better. Tested (Crash) : fix cpufreq Max = 1200000 Min = 1200000 governor = performance Tested (Crash) : fix cpufreq Max = 1008000 Min = 1008000 governor = powersave Tested (Crash) : All ZRAM disabled ( /etc/default/armbian-zram-config ) I took kernel via serial console. (dmesg -E) "sysctl -w vm.drop_caches=3" , cause kernel panic Immediately. root@helios64:~# [50657.866770] Insufficient stack space to handle exception! [50657.866777] ESR: 0x96000047 -- DABT (current EL) [50657.867667]

Hi. My helios64 always crashes with this method : [environment] 1TB m.2 sata ssd for system 16TB seagate exos on data drive 18TB WD Ultrastar DC550 on data drive Linux helios64 5.9.14-rockchip64 #20.11.4 SMP PREEMPT Tue Dec 15 08:52:20 CET 2020 aarch64 aarch64 aarch64 GNU/Linux about 620,000 files on data drive. [method1] #sync #sysctl -w vm.drop.caches = 3 #ls -CFR [datadir] > [tmp dir] #ls -1lFR [datadir] > [tmp dir] #zip -j [tmpdir] filelist.zip #sync >>>>&g