RE: [PATCH 4/4] arm64: dts: renesas: r9a07g0{43,44,54}: Update usbhs family compatible

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Krzysztof Kozlowski,

> -----Original Message-----
> From: Krzysztof Kozlowski <krzysztof.kozlowski@xxxxxxxxxx>
> Sent: Sunday, March 10, 2024 8:06 AM
> Subject: Re: [PATCH 4/4] arm64: dts: renesas: r9a07g0{43,44,54}: Update usbhs family compatible
> 
> On 10/03/2024 08:57, Biju Das wrote:
> >>>>> old DTB + old kernel will have 16 pipe buffers old DTB + newer
> >>>>> kernel will have 9 pipe buffers.
> >>>>> New DTB + new kernel will have 9 pipe buffer.
> >>>>
> >>>> You missed new DTB and old kernel. This breaks all users of DTS.
> >>>> That's the entire point of your broken generic compatibles which you did not address.
> >>>
> >>> As per my knowledge, there is no user in RZ/G2L is using new DTB and old kernel.
> >>> So, it is safe.
> >>
> >> If there is a user for such change, we could use
> >>
> >> 	compatible = "renesas,usbhs-r9a07g043",
> >> 	             "renesas,rzg2l-usbhs",
> >> 			 "renesas,rza2-usbhs";
> >>
> >> Or
> >>
> >> 	compatible = "renesas,usbhs-r9a07g043",
> >> 			 "renesas,rza2-usbhs";
> >>
> >>
> >> The former consumes less memory compared to the later.
> >>
> >> As later requires, 3 platform structures for rz/g2l, rz/v2l and
> >> rz/gul whereas the former requires just 1.
> >
> > Another way is using RZ/G2L SoC fallback compatible for both RZ/V2L
> > and RZ/Five varients
> >
> > 	compatible = "renesas,usbhs-r9a07g043",
> > 	             "renesas, usbhs-r9a07g044",
> 
> How does it solve anything? Nothing binds to this.

This will solve all the issues.

The usbhs-r9a07g043 IP is derived from usbhs-r9a07g044 that is derived from rza2-usbhs.

> 
> > 			 "renesas,rza2-usbhs";
> >
> > This will fit into all ABI combinations with optimized memory usage in driver related to platform
> structure.
> >
> > old DTB + old kernel will have 16 pipe buffers old DTB + new kernel
> > will have 9 pipe buffers New DTB + old kernel will have 16 pipe
> > buffers New DTB + new kernel will have 9 pipe buffers
> 
> Maybe just drop incorrect generic compatibles since you are breaking users? You did not provide any
> rationale against my arguments.

The above model does not break any users. Can you please give an example which breaks the users?


Cheers,
Biju




[Index of Archives]     [Device Tree Compilter]     [Device Tree Spec]     [Linux Driver Backports]     [Video for Linux]     [Linux USB Devel]     [Linux PCI Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [XFree86]     [Yosemite Backpacking]


  Powered by Linux