OpenVoiceOS/buildroot-external/board/raspberrypi/kernel.config

CONFIG_SND_SOC_WM8960=m
CONFIG_SND_SOC_TAS5086=m

# CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE is not set
CONFIG_CPU_FREQ_DEFAULT_GOV_ONDEMAND=y

CONFIG_PREEMPT_VOLUNTARY=y
# CONFIG_PREEMPT is not set

CONFIG_ZRAM=y
CONFIG_ZSMALLOC=y
Changes to sound architecture, prepare system no longer needed 2021-02-05 16:10:50 +01:00			`CONFIG_SND_SOC_WM8960=m`
Add CONFIG_SND_SOC_TAS5086=m to kernel config 2021-06-07 11:11:13 +02:00			`CONFIG_SND_SOC_TAS5086=m`

Changes to sound architecture, prepare system no longer needed 2021-02-05 16:10:50 +01:00			`# CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE is not set`
			`CONFIG_CPU_FREQ_DEFAULT_GOV_ONDEMAND=y`
Add and configure 20% of memory as zram swap with a low agressive swappiness 2021-03-01 21:58:30 +01:00
Switch to a voluntary kernel preemption. The respeaker mic hats and possibly (many) others are using the x-powers ac108 adc chip however the driver of this chip is not compatible with 64bit PREEMPT kernels crashing the drivers with "BUG: scheduling while atomic" spinlock issues. PREEMPT is used for low latency desktops PREEMPT_NONE is used for servers As we are not a (low-latency needed) desktop however not a server as well, the middle ground actually works perfectly fine with the benefit of having a more stable respeaker driver. This accounts for ALL different respeaker HATs except the 2-mic as that one is based on the WM8960 chip which is actually already supported mainline if you want anyway. 2022-02-09 14:15:41 +01:00			`CONFIG_PREEMPT_VOLUNTARY=y`
			`# CONFIG_PREEMPT is not set`

Add and configure 20% of memory as zram swap with a low agressive swappiness 2021-03-01 21:58:30 +01:00			`CONFIG_ZRAM=y`
			`CONFIG_ZSMALLOC=y`