[PATCH v2] ARM: at91: add basic support for new SoC lan966x

Mon Sep 20 03:47:14 PDT 2021

> -----Original Message-----
> From: Alexandre Belloni [mailto:alexandre.belloni at bootlin.com]
> Sent: Friday, September 17, 2021 8:03 PM
> To: Nicolas Ferre - M43238 <Nicolas.Ferre at microchip.com>
> Cc: Kavyasree Kotagiri - I30978 <Kavyasree.Kotagiri at microchip.com>; Ludovic
> Desroches - M43218 <Ludovic.Desroches at microchip.com>;
> linux at armlinux.org.uk; linux-arm-kernel at lists.infradead.org; linux-
> kernel at vger.kernel.org; Manohar Puri - I30488
> <Manohar.Puri at microchip.com>; Madhuri Sripada - I34878
> <Madhuri.Sripada at microchip.com>
> Subject: Re: [PATCH v2] ARM: at91: add basic support for new SoC lan966x
> 
> EXTERNAL EMAIL: Do not click links or open attachments unless you know the
> content is safe
> 
> Hello,
> 
> I wanted to reply earlier but I still have a few comments...
> 
> On 17/09/2021 15:41:01+0200, Nicolas Ferre wrote:
> > On 31/08/2021 at 12:21, kavyasree.kotagiri at microchip.com wrote:
> > > From: Kavyasree Kotagiri <Kavyasree.Kotagiri at microchip.com>
> > >
> > > This patch introduces Microchip LAN966X ARMv7 based SoC family
> > > of multiport gigabit AVB/TSN-capable ethernet switches.
> > > It supports two SKUs: 4-port LAN9662 with multiprotocol
> > > processing support and 8-port LAN9668 switch.
> > >
> > > LAN966X includes copper and serial ethernet interfaces,
> > > peripheral interfaces such as PCIe, USB, TWI, SPI, UART, QSPI,
> > > SD/eMMC, Parallel Interface (PI) as well as synchronization
> > > and trigger inputs/outputs.
> > >
> > > Signed-off-by: Kavya Sree Kotagiri <kavyasree.kotagiri at microchip.com>
> >
> > Your name fixed with your permission and as Rob suggested in a previous
> > series.
> >
> > > ---
> > > v1 -> v2:
> > > - Removed lan966x_dt_device_init().
> > >    System boots fine without this function.
> > >
> > >   arch/arm/mach-at91/Kconfig   | 13 +++++++++++++
> > >   arch/arm/mach-at91/Makefile  |  1 +
> >
> > Kconfig and Makefile files updated to avoid merge conflict with SAMA7G5
> that
> > was integrated in 5.15-rc1.
> >
> > >   arch/arm/mach-at91/lan966x.c | 25 +++++++++++++++++++++++++
> >
> >
> > >   3 files changed, 39 insertions(+)
> > >   create mode 100644 arch/arm/mach-at91/lan966x.c
> > >
> > > diff --git a/arch/arm/mach-at91/Kconfig b/arch/arm/mach-at91/Kconfig
> > > index ccd7e80ce943..06cb425af761 100644
> > > --- a/arch/arm/mach-at91/Kconfig
> > > +++ b/arch/arm/mach-at91/Kconfig
> > > @@ -122,6 +122,14 @@ config SOC_SAM9X60
> > >     help
> > >       Select this if you are using Microchip's SAM9X60 SoC
> > > +config SOC_LAN966X
> > > +   bool "ARMv7 based Microchip LAN966X SoC family"
> > > +   depends on ARCH_MULTI_V7
> > > +   select SOC_LAN966
> > > +   select DW_APB_TIMER_OF
> > > +   help
> > > +     This enables support for ARMv7 based Microchip LAN966X SoC
> family.
> > > +
> 
> I don't think SOC_LAN966X is a good name or said differently as it is
> not part specific, I don't get why this isn't merged with SOC_LAN966. I
> would have one or the other.
> 
I am using two different names to represent LAN966 as a new SoC family and 
LAN966X as its SKU's - LAN9662 and LAN9668

> > >   comment "Clocksource driver selection"
> > >   config ATMEL_CLOCKSOURCE_PIT
> > > @@ -188,6 +196,11 @@ config SOC_SAMA5
> > >     select SOC_SAM_V7
> > >     select SRAM if PM
> > > +config SOC_LAN966
> > > +   bool
> > > +   select ARM_GIC
> > > +   select MEMORY
> > > +
> > >   config ATMEL_PM
> > >     bool
> > > diff --git a/arch/arm/mach-at91/Makefile b/arch/arm/mach-
> at91/Makefile
> > > index f565490f1b70..93cfd5b4e6d4 100644
> > > --- a/arch/arm/mach-at91/Makefile
> > > +++ b/arch/arm/mach-at91/Makefile
> > > @@ -6,6 +6,7 @@
> > >   # CPU-specific support
> > >   obj-$(CONFIG_SOC_AT91RM9200)      += at91rm9200.o
> > >   obj-$(CONFIG_SOC_AT91SAM9)        += at91sam9.o
> > > +obj-$(CONFIG_SOC_LAN966X)  += lan966x.o
> > >   obj-$(CONFIG_SOC_SAM9X60) += sam9x60.o
> > >   obj-$(CONFIG_SOC_SAMA5)           += sama5.o
> > >   obj-$(CONFIG_SOC_SAMV7)           += samv7.o
> > > diff --git a/arch/arm/mach-at91/lan966x.c b/arch/arm/mach-
> at91/lan966x.c
> > > new file mode 100644
> > > index 000000000000..154d616569ae
> > > --- /dev/null
> > > +++ b/arch/arm/mach-at91/lan966x.c
> > > @@ -0,0 +1,25 @@
> > > +// SPDX-License-Identifier: GPL-2.0-or-later
> > > +/*
> > > + * Setup code for LAN966X
> > > + *
> > > + * Copyright (C) 2021 Microchip Technology, Inc. and its subsidiaries
> > > + *
> > > + */
> > > +
> > > +#include <linux/of.h>
> > > +#include <linux/of_platform.h>
> > > +
> > > +#include <asm/mach/arch.h>
> > > +#include <asm/system_misc.h>
> > > +
> 
> I'm not sure all those include are actually used.
> 
> > > +#include "generic.h"
> > > +
> > > +static const char *const lan966x_dt_board_compat[] __initconst = {
> > > +   "microchip,lan966x",
> >
> > This compatibility string will need to be documented in
> > Documentation/arm/microchip.rst
> > and
> > Documentation/devicetree/bindings/arm/atmel-at91.yaml (or a similar file,
> as
> > you wish).
> >
> > You'll add it to your DT patches.
> >
> 
> Yes and you'll have to document lan9662 and lan9668. Really, wildcards
> in the compatible strings are usually not a good idea.
> 
> > > +   NULL
> > > +};
> > > +
> > > +DT_MACHINE_START(lan966x_dt, "Microchip LAN966X")
> > > +   /* Maintainer: Microchip */
> > > +   .dt_compat      = lan966x_dt_board_compat,
> > > +MACHINE_END
> > >
> >
> > All the rest looks good to me: no need to resend. It's queued in at91-soc
> > branch targeting 5.16.
> > You'll see it appearing in linux-next in the next few days.
> >
> 
> Well, I'm still wondering one thing, is arch/arm/mach-at91/lan966x.c
> necessary at all to boot the platform?
> 
Yes, it is needed as it introduces new compatible string for new SoC LAN966X.
Could you please let me know if there is a different way to do this?

> --
> Alexandre Belloni, co-owner and COO, Bootlin
> Embedded Linux and Kernel engineering
> https://bootlin.com