Re: [PATCH v6 0/2] soc: loongson2_pm: add power management support

From: "Arnd Bergmann" <arnd@arndb.de>
To: "Yinbo Zhu" <zhuyinbo@loongson.cn>,
	"Rob Herring" <robh+dt@kernel.org>,
	"Krzysztof Kozlowski" <krzysztof.kozlowski+dt@linaro.org>,
	"Conor Dooley" <conor+dt@kernel.org>,
	linux-pm@vger.kernel.org, devicetree@vger.kernel.org,
	linux-kernel@vger.kernel.org, soc@kernel.org,
	"Ulf Hansson" <ulf.hansson@linaro.org>
Cc: "Jianmin Lv" <lvjianmin@loongson.cn>,
	wanghongliang@loongson.cn, "Liu Peibao" <liupeibao@loongson.cn>,
	loongson-kernel@lists.loongnix.cn, loongarch@lists.linux.dev,
	"Liu Yun" <liuyun@loongson.cn>
Subject: Re: [PATCH v6 0/2] soc: loongson2_pm: add power management support
Date: Mon, 21 Aug 2023 21:03:48 -0400	[thread overview]
Message-ID: <556a5e72-4001-4d28-affb-831f5c8f9e26@app.fastmail.com> (raw)
In-Reply-To: <20230803063703.5659-1-zhuyinbo@loongson.cn>

On Thu, Aug 3, 2023, at 02:37, Yinbo Zhu wrote:
> Loongson-2 platform support Power Management Controller (ACPI) and this
> series patch was to add PM driver that base on dts and PM binding support.
>
> Change in v6:
> 		1. The patch "[PATCH v3 1/3] loongarch: export some arch-specific
> 		   pm interfaces" had been merged into mainline tree in v6.5-rc1
> 		   thus this v6 series patch need drop it and need depend on it
> 		   and it's patch link was:
> https://lore.kernel.org/all/20230615091757.24686-2-zhuyinbo@loongson.cn/
> 		2. Adding Ulf Hansson to Cc.
> 		3. Adding soc@kernel.org to Cc.
> 		4. Keep indented with one tab +2 spaces in Kconfig help text.

I talked to WANG Xuerui on IRC, and he was able to clarify some of the
missing bits of information for me, after which I merged both patches,
even though my concerns are not fully addressed:

- I still think that branching into ROM code from the kernel is a mistake
  and we should have never allowed that as an ad-hoc interface in the ACPI
  variant to start with. It's hard to change that now though, and having
  a DT interface to access the same entry point does not really make it
  worse. This might need a redesign for future firmware though, to have
  a proper runtime interface

- The bigger problem I still see is the DT-enabled boot with PMon without
  the UEFI firmware. This does not impact the DT binding, but I would
  consider all non-UEFI booting firmware images broken and not supported
  by the kernel, as we originally discussed when merging the kernel.
  These should still be fixable by upgrading PMon to a UEFI-enabled
  version.

     Arnd