[U-Boot] [PATCH 4/5] cmd: mvebu: bubt: Add support for legacy Marvell SoCs

[Jon Nettleton]

On Wed, Aug 29, 2018 at 3:36 PM <kostap at marvell.com> wrote:
>
> From: Konstantin Porotchkin <kostap at marvell.com>
>
> Add support for image load and basic verification in bubt
> for legacy Marvell SoCs (A38x, A39x, ...)
>
> Signed-off-by: Konstantin Porotchkin <kostap at marvell.com>
> Cc: Igal Liberman <igall at marvell.com>
> Cc: Stefan Roese <sr at denx.de>
> ---
>  cmd/mvebu/bubt.c | 75 +++++++++++++++++++++++++++++++++++++++++++++++++++++---
>  1 file changed, 72 insertions(+), 3 deletions(-)
>
> diff --git a/cmd/mvebu/bubt.c b/cmd/mvebu/bubt.c
> index b4d371f..e10d079 100644
> --- a/cmd/mvebu/bubt.c
> +++ b/cmd/mvebu/bubt.c
> @@ -83,6 +83,33 @@ struct mvebu_image_info {
>         u32     encrypt_start_offset;
>         u32     encrypt_size;
>  };
> +
> +#else /* Older Armada SoCs - A38x, A39x, ... */
> +
> +#define        LEGACY_HDR_VERSION      1
> +
> +struct legacy_image_header {
> +/*     type    name                    byte order */
> +       u8      block_id;               /*   0   */
> +       u8      flags;                  /*   1   */
> +       u16     nand_pge_size;          /*  2-3  */
> +       u32     block_size;             /*  4-7  */
> +       u8      version;                /*   8   */
> +       u8      hdr_size_msb;           /*   9   */
> +       u16     hdr_size_lsb;           /* 10-11 */
> +       u32     source_addr;            /* 12-15 */
> +       u32     destination_addr;       /* 16-19 */
> +       u32     execution_addr;         /* 20-23 */
> +       u8      options;                /*  24   */
> +       u8      nand_block_size;        /*  25   */
> +       u8      nand_technology;        /*  26   */
> +       u8      rsvd4;                  /*  27   */
> +       u16     rsvd2;                  /* 28-29 */
> +       u8      ext;                    /*  30   */
> +       u8      checksum;               /*  31   */
> +
> +};
> +
>  #endif /* CONFIG_ARMADA_XXX */
>
>  struct bubt_dev {
> @@ -618,11 +645,53 @@ static int check_image_header(void)
>         return 0;
>  }
>
> -#else /* Not ARMADA? */
> +#else /* Legacy SoCs */
> +u8 do_checksum8(u8 *start, u32 len)
> +{
> +       u8 sum = 0;
> +       u8 *startp = start;
> +
> +       do {
> +               sum += *startp;
> +               startp++;
> +       } while (--len);
> +
> +       return sum;
> +}
> +
>  static int check_image_header(void)
>  {
> -       printf("bubt cmd does not support this SoC device or family!\n");
> -       return -ENOEXEC;
> +       struct legacy_image_header *hdr =
> +                       (struct legacy_image_header *)get_load_addr();
> +       u32 header_len = hdr->hdr_size_lsb + (hdr->hdr_size_msb << 16);
> +       u8 checksum;
> +       u8 checksum_ref = hdr->checksum;
> +
> +       /*
> +        * For now compare checksum, and header version. Later we can
> +        * verify more stuff on the header like interface type, etc
> +        */
> +       if (hdr->version != LEGACY_HDR_VERSION) {
> +               printf("ERROR: Bad HDR Version 0x%x != 0x%x\n",
> +                      hdr->version, LEGACY_HDR_VERSION);
> +               return -ENOEXEC;
> +       }
> +
> +       /* The checksum value is discarded from checksum calculation */
> +       hdr->checksum = 0;
> +
> +       checksum = do_checksum8((u8 *)hdr, header_len);
> +       if (checksum != checksum_ref) {
> +               printf("Error: Bad Image checksum. 0x%x != 0x%x\n",
> +                      checksum, checksum_ref);
> +               return -ENOEXEC;
> +       }
> +
> +       /* Restore the checksum before writing */
> +       hdr->checksum = checksum_ref;
> +       printf("Image checksum...OK!\n");
> +
> +       return 0;
>  }
>  #endif
>

Konstantin,

I already have a more complete integration for the legacy SOCs into
the bubt command.  It incorporates much of the same code that is used
to generate the images in kwimage.c.  We haven't submitted it to
mainline yet as we were waiting for a few more patchsets to land.

The features I have incorporated are supporting all possible boot
devices, image flashing from USB devices, and most importantly image
verification before flashing, so you can't brick your device by
flashing an image for eMMC onto the SPI device.

If you could take a look and comment I would be happy to work with you
to get this support into mainline.

Thanks,
Jon