| Message ID | 1509974198-14831-2-git-send-email-jorge.ramirez-ortiz@linaro.org |
|---|---|
| State | Accepted |
| Commit | 9a16153834d09372a5768ba499e5411f12801bda |
| Headers | show |
| Series | [1/2] env: Save environment at the end of an MMC partition | expand |
On Mon, Nov 06, 2017 at 02:16:38PM +0100, Jorge Ramirez-Ortiz wrote: > Save the environment data at the end of the boot partition on emmc > > Signed-off-by: Jorge Ramirez-Ortiz <jorge.ramirez-ortiz@linaro.org> > --- > arch/arm/dts/dragonboard410c-uboot.dtsi | 7 +++++++ > configs/dragonboard410c_defconfig | 2 ++ > include/configs/dragonboard410c.h | 1 + > 3 files changed, 10 insertions(+) Implementation this seems fine. Is there a reason to do this vs ENV_IS_IN_FAT? Or ENV_IS_IN_EXT ? And have it be in a filesystem instead? That usually seems to be best for eval/dev/etc style boards. -- Tom
On 11/07/2017 11:50 PM, Tom Rini wrote: > On Mon, Nov 06, 2017 at 02:16:38PM +0100, Jorge Ramirez-Ortiz wrote: > >> Save the environment data at the end of the boot partition on emmc >> >> Signed-off-by: Jorge Ramirez-Ortiz <jorge.ramirez-ortiz@linaro.org> >> --- >> arch/arm/dts/dragonboard410c-uboot.dtsi | 7 +++++++ >> configs/dragonboard410c_defconfig | 2 ++ >> include/configs/dragonboard410c.h | 1 + >> 3 files changed, 10 insertions(+) > Implementation this seems fine. Is there a reason to do this vs > ENV_IS_IN_FAT? Or ENV_IS_IN_EXT ? And have it be in a filesystem > instead? That usually seems to be best for eval/dev/etc style boards. I think having it on a partition offset (a partition that typically just stores the uboot image) can be very convenient when having the filesystem on an NFS share. It means that the user can completely ignore flashing and maintaining the filesystem until he is ready. The only issue I see is that it is up to the user - when he reprograms a new image or modifies the size of the environemnt- to make sure the environment and whatever image is in the partition dont overlap each other but I think that should be acceptable? >
On Wed, Nov 08, 2017 at 08:49:09AM +0100, Jorge Ramirez wrote: > On 11/07/2017 11:50 PM, Tom Rini wrote: > >On Mon, Nov 06, 2017 at 02:16:38PM +0100, Jorge Ramirez-Ortiz wrote: > > > >>Save the environment data at the end of the boot partition on emmc > >> > >>Signed-off-by: Jorge Ramirez-Ortiz <jorge.ramirez-ortiz@linaro.org> > >>--- > >> arch/arm/dts/dragonboard410c-uboot.dtsi | 7 +++++++ > >> configs/dragonboard410c_defconfig | 2 ++ > >> include/configs/dragonboard410c.h | 1 + > >> 3 files changed, 10 insertions(+) > >Implementation this seems fine. Is there a reason to do this vs > >ENV_IS_IN_FAT? Or ENV_IS_IN_EXT ? And have it be in a filesystem > >instead? That usually seems to be best for eval/dev/etc style boards. > > I think having it on a partition offset (a partition that typically just > stores the uboot image) can be very convenient when having the filesystem on > an NFS share. > > It means that the user can completely ignore flashing and maintaining the > filesystem until he is ready. The only issue I see is that it is up to the > user - when he reprograms a new image or modifies the size of the > environemnt- to make sure the environment and whatever image is in the > partition dont overlap each other but I think that should be acceptable? Well, if you think environment in flash is best, that's fine. I just want to make sure the alternatives are known as I keep surprising people when I note that the env can be a file in FAT or similar. Thanks! -- Tom
On 11/09/2017 02:58 AM, Tom Rini wrote: > On Wed, Nov 08, 2017 at 08:49:09AM +0100, Jorge Ramirez wrote: >> On 11/07/2017 11:50 PM, Tom Rini wrote: >>> On Mon, Nov 06, 2017 at 02:16:38PM +0100, Jorge Ramirez-Ortiz wrote: >>> >>>> Save the environment data at the end of the boot partition on emmc >>>> >>>> Signed-off-by: Jorge Ramirez-Ortiz <jorge.ramirez-ortiz@linaro.org> >>>> --- >>>> arch/arm/dts/dragonboard410c-uboot.dtsi | 7 +++++++ >>>> configs/dragonboard410c_defconfig | 2 ++ >>>> include/configs/dragonboard410c.h | 1 + >>>> 3 files changed, 10 insertions(+) >>> Implementation this seems fine. Is there a reason to do this vs >>> ENV_IS_IN_FAT? Or ENV_IS_IN_EXT ? And have it be in a filesystem >>> instead? That usually seems to be best for eval/dev/etc style boards. >> I think having it on a partition offset (a partition that typically just >> stores the uboot image) can be very convenient when having the filesystem on >> an NFS share. >> >> It means that the user can completely ignore flashing and maintaining the >> filesystem until he is ready. The only issue I see is that it is up to the >> user - when he reprograms a new image or modifies the size of the >> environemnt- to make sure the environment and whatever image is in the >> partition dont overlap each other but I think that should be acceptable? > Well, if you think environment in flash is best, that's fine. I just > want to make sure the alternatives are known as I keep surprising people > when I note that the env can be a file in FAT or similar. Thanks! cool. yes please, I think - in my experience with poplar and now the db410 this is a useful configuration to have >
On 11/06/2017 02:16 PM, Jorge Ramirez-Ortiz wrote:
> Save the environment data at the end of the boot partition on emmc
any progress on this patchset?
the db410c emmc support remains broken on the dragonboard 410c.
TIA
Jorge
On Mon, Nov 06, 2017 at 02:16:38PM +0100, Jorge Ramirez-Ortiz wrote: > Save the environment data at the end of the boot partition on emmc > > Signed-off-by: Jorge Ramirez-Ortiz <jorge.ramirez-ortiz@linaro.org> Applied to u-boot/master, thanks! -- Tom
diff --git a/arch/arm/dts/dragonboard410c-uboot.dtsi b/arch/arm/dts/dragonboard410c-uboot.dtsi index cc2c175..c3475c4 100644 --- a/arch/arm/dts/dragonboard410c-uboot.dtsi +++ b/arch/arm/dts/dragonboard410c-uboot.dtsi @@ -6,6 +6,13 @@ * SPDX-License-Identifier: GPL-2.0+ */ +/ { + config { + u-boot,mmc-env-partition = "boot"; + }; +}; + + &pm8916_gpios { usb_hub_reset_pm { gpios = <&pm8916_gpios 2 0>; diff --git a/configs/dragonboard410c_defconfig b/configs/dragonboard410c_defconfig index a85aa74..4c04544 100644 --- a/configs/dragonboard410c_defconfig +++ b/configs/dragonboard410c_defconfig @@ -43,3 +43,5 @@ CONFIG_USB_ETHER_ASIX88179=y CONFIG_USB_ETHER_MCS7830=y CONFIG_USB_ETHER_SMSC95XX=y CONFIG_OF_LIBFDT_OVERLAY=y +CONFIG_OF_CONTROL=y +CONFIG_ENV_IS_IN_MMC=y diff --git a/include/configs/dragonboard410c.h b/include/configs/dragonboard410c.h index 6b5d59d..d2447b2 100644 --- a/include/configs/dragonboard410c.h +++ b/include/configs/dragonboard410c.h @@ -100,6 +100,7 @@ REFLASH(dragonboard/u-boot.img, 8)\ BOOTENV #define CONFIG_ENV_SIZE 0x2000 +#define CONFIG_SYS_MMC_ENV_DEV 0 /* mmc0 = emmc, mmc1 = sd */ #define CONFIG_ENV_VARS_UBOOT_CONFIG /* Size of malloc() pool */
Save the environment data at the end of the boot partition on emmc Signed-off-by: Jorge Ramirez-Ortiz <jorge.ramirez-ortiz@linaro.org> --- arch/arm/dts/dragonboard410c-uboot.dtsi | 7 +++++++ configs/dragonboard410c_defconfig | 2 ++ include/configs/dragonboard410c.h | 1 + 3 files changed, 10 insertions(+)