On Friday 21 April 2023 07:56:50 Fabio Estevam wrote:
+uint mmc_get_env_part(struct mmc *mmc) +{
- uint part;
- if (mmc->part_config == MMCPART_NOAVAILABLE) {
part = 0;- } else {
switch (EXT_CSD_EXTRACT_BOOT_PART(mmc->part_config)) {case 0: /* Booting from this eMMC device is disabled */printf("Error - Booting from this eMMC device is disabled\n");printf("Hint: Use 'mmc partconf' command to choose boot partition\n");return -ENODEV;
This does not look to be correct. Return type of the function is uint but here you are trying to return negative number.
I think that there is some layering or API issue. Caller of this function probably does not expect any failure and use returned value as partition number. But you are reading partition number from the source which may return failure (as the source does not have to contain it).
I'm not sure what is the correct way how to handle these kind of error. I hope that this is something which you would know.
Which partition want to choose? Same from which was u-boot/spl loaded at runtime? If yes then maybe this my patch series for mvebu may be interested for you: https://lore.kernel.org/u-boot/20230413205750.10641-1-pali@kernel.org/t/#u
I added there ability to store emmc partition from which was bootloader loaded into access bits of mmc->part_config variable and then via macro EXT_CSD_EXTRACT_PARTITION_ACCESS() it can be extracted. See function spl_mmc_emmc_boot_partition() in that mvebu patch series.
But I'm not sure how reliable access bits of mmc->part_config are on other platforms. As currently only mvebu in that patch series is doing to use it.
case 1: /* Boot partition 1 is used for booting */part = 1;break;case 2: /* Boot partition 2 is used for booting */part = 2;break;case 7: /* User area is used for booting */part = 0;break;default: /* Other values are reserved / unsupported */printf("Error - This eMMC device has configured Reserved boot option\n");printf("Hint: Use 'mmc partconf' command to choose boot partition\n");return -ENODEV;}- }
- return part;
+}