> From: Simon Glass <s...@chromium.org> > Date: Sat, 22 Jan 2022 10:17:08 -0700 > > Hi Mark, > > On Sat, 22 Jan 2022 at 07:45, Mark Kettenis <mark.kette...@xs4all.nl> wrote: > > > > > From: Simon Glass <s...@chromium.org> > > > Date: Fri, 21 Jan 2022 18:40:24 -0700 > > > > > > Hi Mark, > > > > > > On Fri, 14 Jan 2022 at 04:05, Mark Kettenis <kette...@openbsd.org> wrote: > > > > > > > > Add a driver for the NVMe storage controller integrated on > > > > Apple SoCs. This NVMe controller isn't PCI based and deviates > > > > from the NVMe standard in its implementation of the command > > > > submission queue and the integration of an NVMMU that needs > > > > to be managed. This commit tweaks the core NVMe code to > > > > support the linear command submission queue implemented by > > > > this controller. But setting up the submission queue and > > > > managing the NVMMU controller is handled by implementing > > > > the driver ops that were added in an earlier commit. > > > > > > > > Signed-off-by: Mark Kettenis <kette...@openbsd.org> > > > > Tested-on: firefly-rk3399 > > > > Tested-by: Mark Kettenis <kette...@openbsd.org> > > > > --- > > > > configs/apple_m1_defconfig | 1 + > > > > drivers/nvme/Kconfig | 11 ++ > > > > drivers/nvme/Makefile | 1 + > > > > drivers/nvme/nvme_apple.c | 233 +++++++++++++++++++++++++++++++++++++ > > > > 4 files changed, 246 insertions(+) > > > > create mode 100644 drivers/nvme/nvme_apple.c > > > > > > Tested on: Macbook Air M1 > > > Tested-by: Simon Glass <s...@chromium.org> > > > > > > > > > > > diff --git a/configs/apple_m1_defconfig b/configs/apple_m1_defconfig > > > > index cb235e4e7d..1528217b17 100644 > > > > --- a/configs/apple_m1_defconfig > > > > +++ b/configs/apple_m1_defconfig > > > > @@ -11,6 +11,7 @@ CONFIG_DISPLAY_BOARDINFO_LATE=y > > > > # CONFIG_NET is not set > > > > # CONFIG_MMC is not set > > > > CONFIG_DEBUG_UART_ANNOUNCE=y > > > > +CONFIG_NVME_APPLE=y > > > > CONFIG_USB_XHCI_HCD=y > > > > CONFIG_USB_XHCI_DWC3=y > > > > CONFIG_USB_KEYBOARD=y > > > > diff --git a/drivers/nvme/Kconfig b/drivers/nvme/Kconfig > > > > index 78da444c8b..0cb465160b 100644 > > > > --- a/drivers/nvme/Kconfig > > > > +++ b/drivers/nvme/Kconfig > > > > @@ -10,6 +10,17 @@ config NVME > > > > This option enables support for NVM Express devices. > > > > It supports basic functions of NVMe (read/write). > > > > > > > > +config NVME_APPLE > > > > + bool "Apple NVMe controller support" > > > > + select NVME > > > > + help > > > > + This option enables support for the NVMe storage > > > > + controller integrated on Apple SoCs. This controller > > > > + isn't PCI-based based and deviates from the NVMe > > > > + standard implementation in its implementation of > > > > + the command submission queue and the integration > > > > + of an NVMMU that needs to be managed. > > > > + > > > > config NVME_PCI > > > > bool "NVM Express PCI device support" > > > > depends on PCI > > > > diff --git a/drivers/nvme/Makefile b/drivers/nvme/Makefile > > > > index fad9724e17..fa7b619446 100644 > > > > --- a/drivers/nvme/Makefile > > > > +++ b/drivers/nvme/Makefile > > > > @@ -3,4 +3,5 @@ > > > > # Copyright (C) 2017, Bin Meng <bmeng...@gmail.com> > > > > > > > > obj-y += nvme-uclass.o nvme.o nvme_show.o > > > > +obj-$(CONFIG_NVME_APPLE) += nvme_apple.o > > > > obj-$(CONFIG_NVME_PCI) += nvme_pci.o > > > > diff --git a/drivers/nvme/nvme_apple.c b/drivers/nvme/nvme_apple.c > > > > new file mode 100644 > > > > index 0000000000..b0dc8492f0 > > > > --- /dev/null > > > > +++ b/drivers/nvme/nvme_apple.c > > > > @@ -0,0 +1,233 @@ > > > > +// SPDX-License-Identifier: GPL-2.0+ > > > > +/* > > > > + * (C) Copyright 2021 Mark Kettenis <kette...@openbsd.org> > > > > + */ > > > > + > > > > +#include <common.h> > > > > +#include <dm.h> > > > > +#include <mailbox.h> > > > > +#include <mapmem.h> > > > > +#include "nvme.h" > > > > +#include <reset.h> > > > > + > > > > +#include <asm/io.h> > > > > +#include <asm/arch-apple/rtkit.h> > > > > > > asm/arch/ should work > > > > It doesn't. For some reason I end up with > > > > arch/arm/include/asm/arch -> arch-m1 > > > > But this rtkit stuff is expected to be generic and apply to all Apple > > SoCs, not just the M1. > > Hmm so you plan to have an arch/m2 as well? Can we not put everything > in arch/apple then? We don't normally have arch directories for > different generations unless they are completely different.
No, I think a single arch/arch-apple is better. That isolates us from whatever silly marketing names Apple comes up with for their SoCs. The goal is to have a single U-Boot binary that supports multiple generations of the SoC. That means I have to set SYS_SOC to "apple" instead of "m1" and move the existing uart.h file into arch/arm/include/arch/arch-apple/ isn't it? I can add a diff that does this to the series.