mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Jonas Karlman <jonas@kwiboo.se>
To: Cunhao Lu <1579567540@qq.com>
Cc: Stephen Boyd <sboyd@kernel.org>,
	Brian Masney <bmasney+clk@redhat.com>,
	Jerome Brunet <jbrunet+clk@baylibre.com>,
	Rob Herring <robh@kernel.org>,
	Krzysztof Kozlowski <krzk+dt@kernel.org>,
	Conor Dooley <conor+dt@kernel.org>,
	Heiko Stuebner <heiko@sntech.de>,
	Philipp Zabel <p.zabel@pengutronix.de>,
	Elaine Zhang <zhangqing@rock-chips.com>,
	Richard Cochran <richardcochran@gmail.com>,
	linux-clk <linux-clk@vger.kernel.org>,
	devicetree <devicetree@vger.kernel.org>,
	linux-arm-kernel <linux-arm-kernel@lists.infradead.org>,
	linux-rockchip <linux-rockchip@lists.infradead.org>,
	linux-kernel <linux-kernel@vger.kernel.org>,
	netdev <netdev@vger.kernel.org>,
	Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
Subject: Re: [PATCH v6 1/4] dt-bindings: clock: rockchip: Add RK3572 CRU
Date: Fri, 20 Nov 2026 11:37:18 +0100	[thread overview]
Message-ID: <38fbfd13-526c-47cd-b723-2ac4e77682ae@kwiboo.se> (raw)
In-Reply-To: <tencent_BB5901AB88D47E874D0E348B32A609A92308@qq.com>

Hi Cunhao Lu,

On 11/20/2026 10:40 AM, Cunhao Lu wrote:
> Hi Jonas,
> 
> Thank you for your review.
> 
>> [...]
>>> +#define PCLK_DDR_MON 527
>>> +
>>> +/* Secure clocks */
>>> +#define ACLK_CCI 600
>>> +#define SCLK_DDR 601
>>> +#define CLK_STIMER0_ROOT 602
>>> +#define PCLK_SECURE_S 603
>>> +#define HCLK_SECURE_S 604
>>> +#define ACLK_SECURE_S 605
>>> +#define CLK_PKA_CRYPTO_S 606
>>> +#define PCLK_KLAD 607
>>> +#define HCLK_CRYPTO_S 608
>>> +#define HCLK_KLAD 609
>>> +#define ACLK_CRYPTO_S 610
>>> +#define PCLK_OTPC_S 611
>>> +#define CLK_OTPC_S 612
>>> +#define PCLK_WDT_S 613
>>> +#define TCLK_WDT_S 614
>>> +#define HCLK_TRNG_S 615
>>
>> Are these secure clocks meant to be used as SCMI clock ids? Why a gap from
>> 527 to 600? If I am not mistaken SCMI ids should similar to Linux clock ids
>> start at 0 and be a consecutive series.
> 
> Yes, these are SCMI clock IDs. The CRU IDs end at 527, but the SCMI
> IDs are a separate set. The vendor header [1] uses 600-615 for this
> group, matching the IDs in the released RK3572 BL31 firmware.
> 
> I haven't found the RK3572 BL31 source on GitHub, so I can't rebuild
> the firmware. That also means I can't change the clock IDs it exposes
> via SCMI. If I numbered this group from zero in the binding, Linux
> would request ID 1 for the DDR clock, but the firmware on existing
> boards expects 601. With those IDs, I wouldn't be able to read the
> correct clock data on my board.
> 
> I'll add the SCMI_ prefix in the next version and keep the firmware IDs
> unchanged.

Great, I think the SCMI standard may dictate that IDs should start at 0,
and I expect the vendor firmware possible would fail SCMI validation.

Not sure there is any point to use different IDs for vendor and mainline
firmware. Hopefully we can nudge Rockchip to closer follow specification
in firmware for next SoC.

> 
>> [...]
>>> +/* CRU-->SOFTRST_CON01 */
>>> +#define SRST_ARESETN_TOP_BIU                     0
>>> +#define SRST_PRESETN_TOP_BIU                     1
>>> +#define SRST_ARESETN_TOP_MID_BIU                 2
>>> +#define SRST_HRESETN_TOP_BIU                     3
>>
>> Please drop the unnecessary/redundant RESETN in all constants. The SRST
>> prefix is already a short for Software Reset.
> 
> I'll fix the names in the next version.
> 
>> Please see [1] for the header I prepared for U-Boot mainline RK3572 support,
>> it should have all these symbols cleaned up, to match naming patterns used
>> for all currently supported Rockchip SoCs.
>>
>> [1] https://git.u-boot-project.org/u-boot/contributors/kwiboo/u-boot/-/blob/rk3562/include/dt-bindings/reset/rockchip%2Crk3572-cru.h
> 
> Thanks, I'll follow the naming convention in your header in the next
> revision. I did notice two possible issues with the reset IDs:
> 
> Your header skips 421 between SRST_LINKSYM_HDMITXPHY0 (420) and
> SRST_P_PHPPHY_CRU (422). I've fixed this gap in the current patch.

Looks like I just applied renaming and not re-numbering, so this gap
originates from vendor source header. Good catch :-)

> 
> Your header also assigns ID 453 to both SRST_H_PMU1_BIU and
> SRST_H_PMU1_FW, but the lookup table [2] maps them to bits 9 and 10
> of PMU1 SOFTRST_CON03, respectively. Could you confirm whether these
> two reset lines are meant to share an ID?

This duplicate seem to originate from the reset header in vendor U-Boot
v5 branch [3], that I used for my mainline U-Boot work on RK3572. That
and the fact that I did not re-number the indices. So I am expecting the
vendor Linux 6.12 version is more correct (does not contain the vendor
U-Boot duplication of 453).

[3] https://github.com/rockchip-linux/u-boot/blob/v5/dts/upstream/include/dt-bindings/reset/rockchip%2Crk3572-cru.h

Regards,
Jonas

> 
> [1] https://github.com/rockchip-linux/kernel/blob/470f9dccbdc42e7b8a824d0a5c5640a10e9457d2/include/dt-bindings/clock/rockchip,rk3572-cru.h#L542-L560
> [2] https://git.u-boot-project.org/u-boot/contributors/kwiboo/u-boot/-/blob/85afc941266af08ab30a897dde794f952321025a/drivers/reset/rst-rk3572.c#L605-606
> 
> Best regards,
> Cunhao Lu
> 
> pw-bot: cr


  reply	other threads:[~2026-11-20 10:37 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-17  7:06 [PATCH v6 0/4] clk: rockchip: Add RK3572 clock and reset support Cunhao Lu via B4 Relay
2026-11-17  7:06 ` [PATCH v6 1/4] dt-bindings: clock: rockchip: Add RK3572 CRU Cunhao Lu via B4 Relay
2026-11-19 11:59   ` Jonas Karlman
2026-11-20  9:40     ` Cunhao Lu
2026-11-20 10:37       ` Jonas Karlman [this message]
2026-11-17  7:06 ` [PATCH v6 2/4] clk: rockchip: Add 24-bit fractional divider support Cunhao Lu via B4 Relay
2026-11-17  7:06 ` [PATCH v6 3/4] clk: rockchip: Add composite clock macro with mux table Cunhao Lu via B4 Relay
2026-11-17  7:06 ` [PATCH v6 4/4] clk: rockchip: Add RK3572 clock and reset support Cunhao Lu via B4 Relay

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=38fbfd13-526c-47cd-b723-2ac4e77682ae@kwiboo.se \
    --to=jonas@kwiboo.se \
    --cc=1579567540@qq.com \
    --cc=bmasney+clk@redhat.com \
    --cc=conor+dt@kernel.org \
    --cc=devicetree@vger.kernel.org \
    --cc=heiko@sntech.de \
    --cc=jbrunet+clk@baylibre.com \
    --cc=krzk+dt@kernel.org \
    --cc=krzysztof.kozlowski@oss.qualcomm.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-clk@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-rockchip@lists.infradead.org \
    --cc=netdev@vger.kernel.org \
    --cc=p.zabel@pengutronix.de \
    --cc=richardcochran@gmail.com \
    --cc=robh@kernel.org \
    --cc=sboyd@kernel.org \
    --cc=zhangqing@rock-chips.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®