[PATCH v3] Fix flashing of eMMC user area with Fastboot

[Sean Anderson]

On 5/14/21 6:10 PM, Oleh Kravchenko wrote:
 > Hello Sean,
 > Could you please clarify what you have mean?
 >
 > I think you pointing to this?
 >> fastboot_raw_partition_<raw partition name>=<offset> <size> [mmcpart <num>]
 >
 > Because I don't have idea how aliases will help to flash.

No, specifically I was requesting that you try "fastboot flash 0.1:0
foo.img" (or 1.1:0 depending on your mmc). And please send me the
failing output from the U-Boot side.

(and please address the comments inline below as well)

--Sean

 >
 > 15.05.21 00:45, Sean Anderson пише:
 >> On 5/14/21 5:26 PM, Oleh Kravchenko wrote:
 >>> Hello guys,
 >>> Could you please review and merge this patch?
 >>
 >> Did you have any luck with the second suggestion [1] I made for your
 >> original patch?
 >>
 >> [1] https://lists.denx.de/pipermail/u-boot/2021-May/449778.html
 >>
 >> --Sean
 >>
 >>>
 >>> PR successfully passed CI:
 >>> https://github.com/u-boot/u-boot/pull/75
 >>>
 >>> 15.05.21 00:15, Oleh Kravchenko пише:
 >>>> 'gpt' and 'mmc0' fastboot partitions have been treated as the same device,
 >>>> but it is wrong.
 >>>>
 >>>> Signed-off-by: Oleh Kravchenko <oleg at kaa.org.ua>
 >>>> Cc: Pantelis Antoniou <panto at antoniou-consulting.com>
 >>>> Cc: Marek Vasut <marex at denx.de>
 >>>> ---
 >>>> Changes for v2:
 >>>>       - code cleanup;
 >>>> Changes for v3:
 >>>>       - QA passed at https://github.com/u-boot/u-boot/pull/75;
 >>>>
 >>>>     drivers/fastboot/fb_mmc.c | 25 ++++++++++++++++++++-----
 >>>>     1 file changed, 20 insertions(+), 5 deletions(-)
 >>>>
 >>>> diff --git a/drivers/fastboot/fb_mmc.c b/drivers/fastboot/fb_mmc.c
 >>>> index 2f3837e559..647d3f6c1b 100644
 >>>> --- a/drivers/fastboot/fb_mmc.c
 >>>> +++ b/drivers/fastboot/fb_mmc.c
 >>>> @@ -532,12 +532,7 @@ void fastboot_mmc_flash_write(const char *cmd, void *download_buffer,
 >>>>     #endif
 >>>>
 >>>>     #if CONFIG_IS_ENABLED(EFI_PARTITION)
 >>>> -#ifndef CONFIG_FASTBOOT_MMC_USER_SUPPORT
 >>>>         if (strcmp(cmd, CONFIG_FASTBOOT_GPT_NAME) == 0) {
 >>>> -#else
 >>>> -    if (strcmp(cmd, CONFIG_FASTBOOT_GPT_NAME) == 0 ||
 >>>> -        strcmp(cmd, CONFIG_FASTBOOT_MMC_USER_NAME) == 0) {
 >>>> -#endif
 >>>>             dev_desc = fastboot_mmc_get_dev(response);
 >>>>             if (!dev_desc)
 >>>>                 return;
 >>>> @@ -599,9 +594,29 @@ void fastboot_mmc_flash_write(const char *cmd, void *download_buffer,
 >>>>         }
 >>>>     #endif
 >>>>
 >>>> +#if CONFIG_IS_ENABLED(FASTBOOT_MMC_USER_SUPPORT)
 >>
 >> Is it possible to use
 >>
 >> if (CONFIG_IS_ENABLED(...)) { ... }
 >>
 >> here?
 >>
 >>>> +    if (strcmp(cmd, CONFIG_FASTBOOT_MMC_USER_NAME) == 0) {
 >>>> +        dev_desc = fastboot_mmc_get_dev(response);
 >>>> +        if (!dev_desc)
 >>>> +            return;
 >>>> +
 >>>> +        memset(&info, 0, sizeof(info));
 >>>> +        info.start    = 0;
 >>>> +        info.size    = dev_desc->lba;
 >>>> +        info.blksz    = dev_desc->blksz;
 >>>> +        strlcpy((char *)&info.name, cmd, sizeof(info.name));
 >>>> +
 >>>> +        goto write_image;
 >>
 >> Why do we need to skip get_part_info?
 >>
 >>>> +    }
 >>>> +#endif
 >>>> +
 >>>>         if (fastboot_mmc_get_part_info(cmd, &dev_desc, &info, response) < 0)
 >>>>             return;
 >>>>
 >>>> +#if CONFIG_IS_ENABLED(FASTBOOT_MMC_USER_SUPPORT)
 >>
 >> Is conditionally defining this label necessary?
 >>
 >> --Sean
 >>
 >>>> +write_image:
 >>>> +#endif
 >>>> +
 >>>>         if (is_sparse_image(download_buffer)) {
 >>>>             struct fb_mmc_sparse sparse_priv;
 >>>>             struct sparse_storage sparse;
 >>>>
 >>>
 >