Adding support for Ubiquiti UniFi 6 Plus (U6+)

2024-05-16 16:22| 来源: 网络整理| 查看: 265

Nice!

Strange that it crashes immediately on enabling the uart console. I don't know if it is relalted, but I notice that there seems to be some inconsistency in the definittion of the ramoops device:

elbertmai: [ 0.006618] ramoops: using module parameters [ 0.012243] printk: console [ramoops-1] enabled [ 0.012380] pstore: Registered ramoops as persistent store backend [ 0.012430] ramoops: using 0x100000@0x42fe0000, ecc: 16 ... [ 0.013669] ramoops: already initialized [ 0.013760] ramoops: probe of 42ff0000.ramoops failed with error -22

And this is a bit suspicious since ramoops is registered as console until the uart is running.

Note that the bootloader could be adding device nodes here. Maybe that creates some conflict?

EDIT: I see from the downloaded device-tree you posted earlier that we have this, which is consistent with your boot log:

ramoops@42fe0000 { reg = ; compatible = "ramoops"; console-size = ; ftrace-size = ; pmsg-size = ; record-size = ; };

Maybe you need to disable the ramoops node in arch/arm64/boot/dts/mediatek/mt7981.dtsi ?

(there's currently no way to do that, but that should be added if this device requires it)

EDIT2: Trying to figure out where we got that default ramoops@42ff0000 device from. We have the same for MT7986 as an OpenWrt spcific patch: target/linux/mediatek/patches-5.15/194-dts-mt7968a-add-ramoops.patch

That patch was added by commit 7a0ec001ff79 ("mediatek: sync MT7986 device trees with upstream") without any further justification. But there is no such node upstream: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arch/arm64/boot/dts/mediatek/mt7986a.dtsi#n67

I wonder if this isn't something that should be device specific, and possibly left for the bootloader to fill in on devices doing that.

Any comments @daniel ?

【本文地址】

公司简介

联系我们