* Re: [PATCH 1/3] dt-bindings: can: rockchip: add rk3588v2 CAN-FD compatible
@ 2026-07-02 8:17 Cunhao Lu
2026-07-02 8:49 ` Marc Kleine-Budde
0 siblings, 1 reply; 5+ messages in thread
From: Cunhao Lu @ 2026-07-02 8:17 UTC (permalink / raw)
To: krzk
Cc: Marc Kleine-Budde, heiko, linux-can, mailhol, kernel, robh,
krzk+dt, conor+dt, devicetree, linux-arm-kernel, linux-rockchip,
linux-kernel
Hi Krzysztof,
thanks for the review.
> DT properties have nothing to do here. Are you saying it is fully
> compatible? Then express it properly with fallback.
No, I do not think RK3588v2 is fully compatible with RK3568v2. The RX
FIFO count field uses a different bitfield on RK3588v2, so using
rockchip,rk3568v2-canfd as a fallback would describe a register layout
that does not match the hardware.
Both standard and extended CAN-FD frames worked in my local RK3588v2
testing, but that only confirms CAN-FD functionality on this variant. It
does not make the RK3568v2 register layout a valid fallback for
RK3588v2.
I will reword the commit message in v2 to explain this hardware
difference instead of talking about the DT properties.
> Do not explain what you did. Explain WHY.
>
> Driver is irrelevant here.
Understood. I will remove the driver-focused wording from the binding
patch. The v2 commit message will explain why RK3588v2 needs its own
compatible from the hardware description point of view.
> We do not take anonymous contributions.
I will use my real name in v2.
> So that's just enum with previous entry
I will fold the single-compatible entries into an enum in v2.
> And why is this different from rk3588?
The reason I used rk3588v2 is that this is how the controller identifies
itself at runtime:
rockchip_canfd fea60000.can can0: Rockchip-CANFD rk3588v2 rev2.2 (errata 0x0000) found
rockchip_canfd fea70000.can can1: Rockchip-CANFD rk3588v2 rev2.2 (errata 0x0000) found
I also checked Rockchip's vendor kernel 6.1 sources. The RK3588 DTS uses
the binding name "rockchip,can-2.0" for these controllers, while the
upstream driver has SoC-specific compatibles for the Rockchip CAN-FD IP.
That is why I chose the SoC/IP variant name rockchip,rk3588v2-canfd.
If the preferred binding name is rockchip,rk3588-canfd instead, I can
change the compatible in v2.
Thanks,
Cunhao Lu
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH 1/3] dt-bindings: can: rockchip: add rk3588v2 CAN-FD compatible
2026-07-02 8:17 [PATCH 1/3] dt-bindings: can: rockchip: add rk3588v2 CAN-FD compatible Cunhao Lu
@ 2026-07-02 8:49 ` Marc Kleine-Budde
0 siblings, 0 replies; 5+ messages in thread
From: Marc Kleine-Budde @ 2026-07-02 8:49 UTC (permalink / raw)
To: Cunhao Lu
Cc: krzk, heiko, linux-can, mailhol, kernel, robh, krzk+dt, conor+dt,
devicetree, linux-arm-kernel, linux-rockchip, linux-kernel,
shawn.lin
[-- Attachment #1: Type: text/plain, Size: 1519 bytes --]
On 02.07.2026 16:17:25, Cunhao Lu wrote:
> > And why is this different from rk3588?
>
> The reason I used rk3588v2 is that this is how the controller identifies
> itself at runtime:
There are 2 silicon revision (v2 and v3) of the rk3568 available on the
market. The v2 has several errata in the CAN-FD core, the v3 has some of
them fixed...
> rockchip_canfd fea60000.can can0: Rockchip-CANFD rk3588v2 rev2.2 (errata 0x0000) found
> rockchip_canfd fea70000.can can1: Rockchip-CANFD rk3588v2 rev2.2 (errata 0x0000) found
However, the version register _in_ the CAN-FD was not changed. :/
This is why we have a rockchip,rk3568v2-canfd and a
rockchip,rk3568v3-canfd compatible.
> I also checked Rockchip's vendor kernel 6.1 sources. The RK3588 DTS uses
> the binding name "rockchip,can-2.0" for these controllers, while the
> upstream driver has SoC-specific compatibles for the Rockchip CAN-FD IP.
> That is why I chose the SoC/IP variant name rockchip,rk3588v2-canfd.
>
> If the preferred binding name is rockchip,rk3588-canfd instead, I can
> change the compatible in v2.
Yes, please use rockchip,rk3588-canfd, unless we get information from
rockchips about CAN-FD errata and SoC silicon revisions.
regards,
Marc
--
Pengutronix e.K. | Marc Kleine-Budde |
Embedded Linux | https://www.pengutronix.de |
Vertretung Nürnberg | Phone: +49-5121-206917-129 |
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-9 |
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH 1/3] dt-bindings: can: rockchip: add rk3588v2 CAN-FD compatible
@ 2026-07-02 8:57 Cunhao Lu
0 siblings, 0 replies; 5+ messages in thread
From: Cunhao Lu @ 2026-07-02 8:57 UTC (permalink / raw)
To: Marc Kleine-Budde
Cc: krzk, heiko, linux-can, mailhol, kernel, robh, krzk+dt, conor+dt,
devicetree, linux-arm-kernel, linux-rockchip, linux-kernel
Hi Marc,
thanks for the explanation.
> However, the version register _in_ the CAN-FD was not changed. :/
> This is why we have a rockchip,rk3568v2-canfd and a
> rockchip,rk3568v3-canfd compatible.
>
> Yes, please use rockchip,rk3588-canfd, unless we get information from
> rockchips about CAN-FD errata and SoC silicon revisions.
Understood. I changed the v2 series to use rockchip,rk3588-canfd
instead of rockchip,rk3588v2-canfd.
This includes the binding, the RK3588 DTS nodes, the driver OF match,
and the internal driver model name.
Thanks,
Cunhao Lu
^ permalink raw reply [flat|nested] 5+ messages in thread
[parent not found: <20260701070128.2096267-1-1579567540@qq.com>]
* [PATCH 1/3] dt-bindings: can: rockchip: add rk3588v2 CAN-FD compatible
[not found] <20260701070128.2096267-1-1579567540@qq.com>
@ 2026-07-01 7:01 ` 1579567540
2026-07-02 7:43 ` Krzysztof Kozlowski
0 siblings, 1 reply; 5+ messages in thread
From: 1579567540 @ 2026-07-01 7:01 UTC (permalink / raw)
To: Marc Kleine-Budde, linux-can
Cc: Vincent Mailhol, kernel, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Heiko Stuebner, devicetree, linux-arm-kernel,
linux-rockchip, linux-kernel, luch00
From: luch00 <1579567540@qq.com>
The RK3588 CAN-FD controller uses the same DT properties as the
existing Rockchip CAN-FD blocks, so extend the current schema with a
SoC-specific rockchip,rk3588v2-canfd compatible instead of creating a
new binding file.
Keep RK3588v2 as its own compatible rather than an rk3568v2 fallback.
Driver support uses separate match data and the RX FIFO count field
layout differs from rk3568v2, so a dedicated compatible is the safer
description.
Signed-off-by: luch00 <1579567540@qq.com>
---
.../devicetree/bindings/net/can/rockchip,rk3568v2-canfd.yaml | 1 +
1 file changed, 1 insertion(+)
diff --git a/Documentation/devicetree/bindings/net/can/rockchip,rk3568v2-canfd.yaml b/Documentation/devicetree/bindings/net/can/rockchip,rk3568v2-canfd.yaml
index a077c0330..aa31ec78e 100644
--- a/Documentation/devicetree/bindings/net/can/rockchip,rk3568v2-canfd.yaml
+++ b/Documentation/devicetree/bindings/net/can/rockchip,rk3568v2-canfd.yaml
@@ -17,6 +17,7 @@ properties:
compatible:
oneOf:
- const: rockchip,rk3568v2-canfd
+ - const: rockchip,rk3588v2-canfd
- items:
- const: rockchip,rk3568v3-canfd
- const: rockchip,rk3568v2-canfd
--
2.34.1
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH 1/3] dt-bindings: can: rockchip: add rk3588v2 CAN-FD compatible
2026-07-01 7:01 ` 1579567540
@ 2026-07-02 7:43 ` Krzysztof Kozlowski
0 siblings, 0 replies; 5+ messages in thread
From: Krzysztof Kozlowski @ 2026-07-02 7:43 UTC (permalink / raw)
To: 1579567540
Cc: Marc Kleine-Budde, linux-can, Vincent Mailhol, kernel,
Rob Herring, Krzysztof Kozlowski, Conor Dooley, Heiko Stuebner,
devicetree, linux-arm-kernel, linux-rockchip, linux-kernel
On Wed, Jul 01, 2026 at 03:01:26PM +0800, 1579567540@qq.com wrote:
> From: luch00 <1579567540@qq.com>
>
> The RK3588 CAN-FD controller uses the same DT properties as the
> existing Rockchip CAN-FD blocks, so extend the current schema with a
DT properties have nothing to do here. Are you saying it is fully
compatible? Then express it properly with fallback.
> SoC-specific rockchip,rk3588v2-canfd compatible instead of creating a
> new binding file.
>
> Keep RK3588v2 as its own compatible rather than an rk3568v2 fallback.
Do not explain what you did. Explain WHY.
> Driver support uses separate match data and the RX FIFO count field
Driver is irrelevant here.
> layout differs from rk3568v2, so a dedicated compatible is the safer
> description.
>
> Signed-off-by: luch00 <1579567540@qq.com>
We do not take anonymous contributions.
> ---
> .../devicetree/bindings/net/can/rockchip,rk3568v2-canfd.yaml | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/Documentation/devicetree/bindings/net/can/rockchip,rk3568v2-canfd.yaml b/Documentation/devicetree/bindings/net/can/rockchip,rk3568v2-canfd.yaml
> index a077c0330..aa31ec78e 100644
> --- a/Documentation/devicetree/bindings/net/can/rockchip,rk3568v2-canfd.yaml
> +++ b/Documentation/devicetree/bindings/net/can/rockchip,rk3568v2-canfd.yaml
> @@ -17,6 +17,7 @@ properties:
> compatible:
> oneOf:
> - const: rockchip,rk3568v2-canfd
> + - const: rockchip,rk3588v2-canfd
So that's just enum with previous entry
And why is this different from rk3588?
> - items:
> - const: rockchip,rk3568v3-canfd
> - const: rockchip,rk3568v2-canfd
> --
> 2.34.1
>
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2026-07-02 8:57 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-07-02 8:17 [PATCH 1/3] dt-bindings: can: rockchip: add rk3588v2 CAN-FD compatible Cunhao Lu
2026-07-02 8:49 ` Marc Kleine-Budde
-- strict thread matches above, loose matches on Subject: below --
2026-07-02 8:57 Cunhao Lu
[not found] <20260701070128.2096267-1-1579567540@qq.com>
2026-07-01 7:01 ` 1579567540
2026-07-02 7:43 ` Krzysztof Kozlowski
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®