Hi,

On Sat, Aug 28, 2021 at 04:15:05PM +0800, Chuanhong Guo wrote:
> ...
> +&snand {
> +     mediatek,quad-spi;
> +     pinctrl-names = "default";
> +     pinctrl-0 = <&serial_nand_pins>;
> +     status = "okay";
> +     mediatek,bmt-v2;
> +
> +     partitions {
> +             compatible = "fixed-partitions";
> +             #address-cells = <1>;
> +             #size-cells = <1>;
> +
> +             partition@0 {
> +                     label = "Preloader";
> +                     reg = <0x0 0x80000>;
> +                     read-only;
> +             };
> +
> +             partition@80000 {
> +                     label = "ATF";
> +                     reg = <0x80000 0x40000>;
> +                     read-only;
> +             };
> +
> +             partition@c0000 {
> +                     label = "u-boot";
> +                     reg = <0xc0000 0x80000>;
> +                     read-only;
> +             };
> +
> +             partition@140000 {
> +                     label = "u-boot-env";
> +                     reg = <0x140000 0x80000>;
> +                     read-only;
> +             };
> +
> +             factory: partition@1c0000 {
> +                     label = "factory";
> +                     reg = <0x1c0000 0x40000>;
> +                     read-only;
> +             };
> +
> +             partition@200000 {
> +                     label = "ubi";
> +                     reg = <0x200000 0x6400000>;
> +             };

Using UBI on top of BMT will result in UBI becoming entirely unusable
in case of a single bad-block being added/detected by BMT:
As BMT then just excludes the block and all follow-up addresses shift
to the left by one erase block that shifting will result in all
subsequent erase blocks being corrupted from UBI's point of view.

I understand that it's not easy to replace bootloader to get rid of
BMT on every single board. As an acceptable compromise, Felix suggested
to further patch the NAND driver so it would allow excluding areas from
BMT treatment (ie. to not have it mess-up UBI).


Cheers


Daniel

_______________________________________________
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Reply via email to