Solved: Invalid partition table of otadata size can corrupt ota_0 partition

[this.wiederkehr]

@xykon said in New beta firmware updater 1.13.1.b1:

'otadata' : ["0x190000", "0x1000"]
'ota_0' : ["0x1a0000", "0x180000"]

According to the esp-idf docs otadata should be 0x2000 in size.

Reading the source code in https://github.com/pycom/pycom-esp-idf/blob/master/components/app_update/esp_ota_ops.c I think making this partition only 4k is a serius bug. and will lead to having the first block (4k) of ota_0 data erased when updating the boot partition.

static esp_err_t rewrite_ota_seq(uint32_t seq, uint8_t sec_id, const esp_partition_t *ota_data_partition)
{
    esp_err_t ret;

    if (sec_id == 0 || sec_id == 1) {
        s_ota_select[sec_id].ota_seq = seq;
        s_ota_select[sec_id].crc = ota_select_crc(&s_ota_select[sec_id]);
        ret = esp_partition_erase_range(ota_data_partition, sec_id * SPI_FLASH_SEC_SIZE, SPI_FLASH_SEC_SIZE);
        if (ret != ESP_OK) {
            return ret;
        } else {
            return esp_partition_write(ota_data_partition, SPI_FLASH_SEC_SIZE * sec_id, &s_ota_select[sec_id].ota_seq, sizeof(ota_select));
        }
    } else {
        return ESP_ERR_INVALID_ARG;
    }
}

[this.wiederkehr]

Checking pycom micropython source I see, that you are not useing the esp-idf functions for performing the update .... so no danger of corrupting the ota_0 slot. Still I wonder why you are doing "your one thing" instead of using the esp-idf functions ...

(https://github.com/pycom/pycom-micropython-sigfox/blob/83e2f7f59e28d726a5a1d79377a5ec37f9e0065b/esp32/ftp/updater.c)