> Subject: Re: [PATCH V2 20/26] imx: dynamic setting mmcdev and mmcroot > > On Tue, Apr 5, 2022 at 10:56 PM Peng Fan (OSS) <peng....@oss.nxp.com> > wrote: > > > > From: Peng Fan <peng....@nxp.com> > > > > Dynamic setting mmcdev and mmcroot. > > Then when boot linux, we can have correct "root=/dev/mmcblk[x]p2" > > > > Signed-off-by: Peng Fan <peng....@nxp.com> > > --- > > arch/arm/include/asm/mach-imx/sys_proto.h | 2 + > > board/freescale/common/Makefile | 3 ++ > > board/freescale/common/mmc.c | 49 +++++++++++++++++++++++ > > 3 files changed, 54 insertions(+) > > create mode 100644 board/freescale/common/mmc.c > > > > diff --git a/arch/arm/include/asm/mach-imx/sys_proto.h > > b/arch/arm/include/asm/mach-imx/sys_proto.h > > index 0c0c7814fb2..37fd427cc00 100644 > > --- a/arch/arm/include/asm/mach-imx/sys_proto.h > > +++ b/arch/arm/include/asm/mach-imx/sys_proto.h > > @@ -228,6 +228,8 @@ int mxs_reset_block(struct mxs_register_32 *reg); > > int mxs_wait_mask_set(struct mxs_register_32 *reg, u32 mask, u32 > > timeout); int mxs_wait_mask_clr(struct mxs_register_32 *reg, u32 > > mask, u32 timeout); > > > > +void board_late_mmc_env_init(void); > > + > > unsigned long call_imx_sip(unsigned long id, unsigned long reg0, > > unsigned long reg1, unsigned long reg2, > > unsigned long reg3); diff --git > > a/board/freescale/common/Makefile b/board/freescale/common/Makefile > > index f13965daf2e..4df484935f4 100644 > > --- a/board/freescale/common/Makefile > > +++ b/board/freescale/common/Makefile > > @@ -63,6 +63,9 @@ obj-$(CONFIG_ZM7300) += zm7300.o > > obj-$(CONFIG_POWER_PFUZE100) += pfuze.o > > obj-$(CONFIG_DM_PMIC_PFUZE100) += pfuze.o > > obj-$(CONFIG_POWER_MC34VR500) += mc34vr500.o > > +ifneq (,$(filter $(SOC), imx8ulp)) > > +obj-y += mmc.o > > +endif > > > > obj-$(CONFIG_LS102XA_STREAM_ID) += ls102xa_stream_id.o > > > > diff --git a/board/freescale/common/mmc.c > > b/board/freescale/common/mmc.c new file mode 100644 index > > 00000000000..8cd5079f962 > > --- /dev/null > > +++ b/board/freescale/common/mmc.c > > @@ -0,0 +1,49 @@ > > +// SPDX-License-Identifier: GPL-2.0+ > > +/* > > + * Copyright 2016 Freescale Semiconductor, Inc. > > + * Copyright 2018-2022 NXP > > + */ > > + > > +#include <common.h> > > +#include <command.h> > > +#include <asm/arch/sys_proto.h> > > +#include <linux/errno.h> > > +#include <asm/io.h> > > +#include <stdbool.h> > > +#include <mmc.h> > > +#include <env.h> > > + > > +static int check_mmc_autodetect(void) { > > + char *autodetect_str = env_get("mmcautodetect"); > > + > > + if (autodetect_str && !strcmp(autodetect_str, "yes")) > > + return 1; > > + > > + return 0; > > +} > > + > > +/* This should be defined for each board */ __weak int > > +mmc_map_to_kernel_blk(int dev_no) { > > + return dev_no; > > +} > > + > > +void board_late_mmc_env_init(void) > > +{ > > + char cmd[32]; > > + char mmcblk[32]; > > + u32 dev_no = mmc_get_env_dev(); > > + > > + if (!check_mmc_autodetect()) > > + return; > > + > > + env_set_ulong("mmcdev", dev_no); > > + > > + /* Set mmcblk env */ > > + sprintf(mmcblk, "/dev/mmcblk%dp2 rootwait rw", > mmc_map_to_kernel_blk(dev_no)); > > + env_set("mmcroot", mmcblk); > > + > > + sprintf(cmd, "mmc dev %d", dev_no); > > + run_command(cmd, 0); > > +} > > -- > > 2.35.1 > > > > Peng, > > I see Stefano already applied this but I'm not sure I agree with it. > Why should you assume that U-Boot and Linux have the same device mapping? > The kernel device mapping is not guaranteed to be consistent. > Every time I have asked about this I've been told the standard was to use a > boot > script that used 'part' to determine the UUID of the boot device from U-Boot's > perspective then use root=PARTUUID= to match that from the kernel's > perspective. > > For example if using CONFIG_DISTRO_DEFAULTS=y (which I think everyone > should be using) your bootscript would look like this: > part uuid ${devtype} ${devnum}:${distro_bootpart} uuid setenv bootargs > "root=PARTUUID=${uuid} rootwait $bootargs"
For NXP internal release, we still use legacy way, with cmd in config header file, let me check and update to fix this. Thanks for pointing out. Thanks, Peng. > > Best Regards, > > Tim