I agree this change.

Now work fine this hints.

• add cpu type to ar71xx_gpio_oe_is_high()

I checked code and I think this problem is case of ar71xx_gpio_oe_is_high().

AR9344 data sheet say.

This is log.

• delete new impliment and change logic

Sorry my missing opration.

This is my test environment.

This is dmesg.

Sorry my missing comment.

• delete dtsi file

loos Please show me complete dts file use gpioiic.c.

Now this patch is conflict. I hope some body make this future.

• review refict

OK. I copy only my repository that is I want use.

My understand is sys/gnu/dts have only openwrt(opneled) dts files. And sys/dts have freebsd original dts files. MZK-W04N-XX.dts is my missing add dts file then move to freebsd dts directory. Do you think how to sync openwrt(opneled) dts files ?

In D13648#285288, @mizhka_gmail.com wrote:

Hi,
Small cosmetic changes are required.
But I don't get idea why this marker lookup can't be done by flash_slicer code.
Could you please clarify benefit?
Thank you!

I can't update this review by arc command. I will put new review.

In D18001#384941, @sgalabov wrote:

! In D18001#384940, @yamori813_yahoo.co.jp wrote:
Mediatek soc is some time CPU print name and CHIPID is different. I have one that is CPU printed RT3050F but CHIPID is RT3350. I think Mediatek code need check SYSCTL_CHIPID0_3 and SYSCTL_CHIPID4_7. Sure thing first is CHIPID and second is FDT I believe.

Can you please share a dmesg output from the case where CPU is printed as RT3050F but CHIPID is RT3350 actually?

In D18001#384902, @sgalabov wrote:

Please split Mediatek and Atheros in separate reviews.
In terms of Mediatek, it looks like there's a bit of code duplication. Why don't we just implement the CPU ID detection within mtk_soc_try_early_detect() instead? We already have the correct FDT mappings there, so all we have to do is fill in the cpu_model in there.

This is dmesg on MT7620.

Broadcom soc support is D12737.

• upgrade dts file at openwrt v18.06.1

In D17880#382169, @manu wrote:

In D17880#382015, @yamori813_yahoo.co.jp wrote:

I seem uart_dev_snps.c use ather soc uart. Not only allwinner, And ALLWINNER_32 and ALLWINNER_64 define is not valid.

The ALLWINNER_XX defines are fake ones so one need to choose between the two.
If you have another device that uses snps uart driver feel free to submit some early routines if those ones doesn't work for you.
Also aw_machdep.c isn't used in aarch64 build.

I seem uart_dev_snps.c use ather soc uart. Not only allwinner, And ALLWINNER_32 and ALLWINNER_64 define is not valid.

This code have compile error now. I will check, but now abandon.

I want add RT2880 and RT5350 and MT7620 support in this code. If I will done it, then I will put review again. Thanks.

I want add more support chip. But I don't know when that will be.

This is original README

• add support eeprom in flash

• add more setup

• fixed bug

• fixed led port

This is connect figure.

• forget add kernel address setting

• too mach delete INTRNG
• refict review

Currently We have almost same code in these file.

Sorry my miss operation.

• fix typo by arc
• repare from master file

• impliment of a few u-boot have space before header

• forget delete

• delete kernel option make to default

• check 64K and 128K boundly. add comment and static

I think dts must not have kernel size because of it's change kernel configuration. This code separate to kernel and rootfs from firmware partition.

• rewrite rootfs check logic

• review refict

Current code is very dangerous because of doing sector erase to address 0. 0 is boot sector in mips. Almost system is sector 0 is protected then no effect. I see Juniper not tested this code on AMD type flash.

• fixed typo

If kernel configuration change and kernel size make big then std partition information is missing. In this case boot stop at mountroot. It's inconvenient. Non-FDT slicer geom_map have same function. I think more better this function set by FDT information than kernel option. But I don't know FDT access in geom code.

• use err value at return

This is dts

• remove child

• review refict

• gpioiic is sepalate like gpioled

In D12649#264042, @gonzo wrote:

Custom dts files shouldn't be put to sys/gnu/dts. This directory is periodically synced with upstream tree, so custom changes may be hard to maintain. The right way to new .dts is to put it to boot/fdt/dts/mips. All include operators should just work

• review refict

In D12649#263981, @loos wrote:

I don't like all the duplicate code in gpioiic.c.
Can you explain which bugs are you trying to fix ?
gpioiic seems to works fine with FDT, but if you want to use the new API, maybe you should consider something like what was done with gpioled (and its FDT counterpart).

This patch can boot all my BCM devices. But sill we have hangup problem at boot time.

• add BCM5312 support

• review refict

• review refict

I tested on /dev/input/event0 and mruby binding.

This is hints for Buffalo WZR-HP-G300NH.