Files
esp-idf/components/bootloader_support/src/bootloader_flash.c

623 lines
19 KiB
C
Raw Normal View History

// Copyright 2015-2016 Espressif Systems (Shanghai) PTE LTD
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#include <stddef.h>
#include <bootloader_flash_priv.h>
#include <esp_log.h>
#include <esp_flash_encrypt.h>
bootloader: fix the WRSR format for ISSI flash chips 1. The 2nd bootloader always call `rom_spiflash_unlock()`, but never help to clear the WEL bit when exit. This may cause system unstability. This commit helps to clear WEL when flash configuration is done. **RISK:** When the app starts, it didn't have to clear the WEL before it actually write/erase. But now the very first write/erase operation should be done after a WEL clear. Though the risk is little (all the following write/erase also need to clear the WEL), we still have to test this carefully, especially for those functions used by the OTA. 2. The `rom_spiflash_unlock()` function in the patch of ESP32 may (1) trigger the QPI, (2) clear the QE or (3) fail to unlock the ISSI chips. Status register bitmap of ISSI chip and GD chip: | SR | ISSI | GD25LQ32C | | -- | ---- | --------- | | 0 | WIP | WIP | | 1 | WEL | WEL | | 2 | BP0 | BP0 | | 3 | BP1 | BP1 | | 4 | BP2 | BP2 | | 5 | BP3 | BP3 | | 6 | QE | BP4 | | 7 | SRWD | SRP0 | | 8 | | SRP1 | | 9 | | QE | | 10 | | SUS2 | | 11 | | LB1 | | 12 | | LB2 | | 13 | | LB3 | | 14 | | CMP | | 15 | | SUS1 | QE bit of other chips are at the bit 9 of the status register (i.e. bit 1 of SR2), which should be read by RDSR2 command. However, the RDSR2 (35H, Read Status 2) command for chip of other vendors happens to be the QIOEN (Enter QPI mode) command of ISSI chips. When the `rom_spiflash_unlock()` function trys to read SR2, it may trigger the QPI of ISSI chips. Moreover, when `rom_spiflash_unlock()` try to clear the BP4 bit in the status register, QE (bit 6) of ISSI chip may be cleared by accident. Or if the ISSI chip doesn't accept WRSR command with argument of two bytes (since it only have status register of one byte), it may fail to clear the other protect bits (BP0~BP3) as expected. This commit makes the `rom_spiflash_unlock()` check whether the vendor is issi. if so, `rom_spiflash_unlock()` only send RDSR to read the status register, send WRSR with only 1 byte argument, and also avoid clearing the QE bit (bit 6). 3. `rom_spiflash_unlock()` always send WRSR command to clear protection bits even when there is no protection bit active. And the execution of clearing status registers, which takes about 700us, will also happen even when there's no bits cleared. This commit skips the clearing of status register if there is no protection bits active. Also move the execute_flash_command to be a bootloader API; move implementation of spi_flash_wrap_set to the bootloader
2020-03-12 18:20:31 +08:00
#include "sdkconfig.h"
#include "soc/soc_caps.h"
#if CONFIG_IDF_TARGET_ESP32
# include "soc/spi_struct.h"
# include "soc/spi_reg.h"
/* SPI flash controller */
# define SPIFLASH SPI1
#else
# include "soc/spi_mem_struct.h"
# include "soc/spi_mem_reg.h"
/* SPI flash controller */
# define SPIFLASH SPIMEM1
#endif
#if CONFIG_IDF_TARGET_ESP32S2BETA
#include "esp32/rom/ets_sys.h"
#include "esp32s2beta/rom/spi_flash.h"
#include "esp32s2/rom/ets_sys.h"
#endif
bootloader: fix the WRSR format for ISSI flash chips 1. The 2nd bootloader always call `rom_spiflash_unlock()`, but never help to clear the WEL bit when exit. This may cause system unstability. This commit helps to clear WEL when flash configuration is done. **RISK:** When the app starts, it didn't have to clear the WEL before it actually write/erase. But now the very first write/erase operation should be done after a WEL clear. Though the risk is little (all the following write/erase also need to clear the WEL), we still have to test this carefully, especially for those functions used by the OTA. 2. The `rom_spiflash_unlock()` function in the patch of ESP32 may (1) trigger the QPI, (2) clear the QE or (3) fail to unlock the ISSI chips. Status register bitmap of ISSI chip and GD chip: | SR | ISSI | GD25LQ32C | | -- | ---- | --------- | | 0 | WIP | WIP | | 1 | WEL | WEL | | 2 | BP0 | BP0 | | 3 | BP1 | BP1 | | 4 | BP2 | BP2 | | 5 | BP3 | BP3 | | 6 | QE | BP4 | | 7 | SRWD | SRP0 | | 8 | | SRP1 | | 9 | | QE | | 10 | | SUS2 | | 11 | | LB1 | | 12 | | LB2 | | 13 | | LB3 | | 14 | | CMP | | 15 | | SUS1 | QE bit of other chips are at the bit 9 of the status register (i.e. bit 1 of SR2), which should be read by RDSR2 command. However, the RDSR2 (35H, Read Status 2) command for chip of other vendors happens to be the QIOEN (Enter QPI mode) command of ISSI chips. When the `rom_spiflash_unlock()` function trys to read SR2, it may trigger the QPI of ISSI chips. Moreover, when `rom_spiflash_unlock()` try to clear the BP4 bit in the status register, QE (bit 6) of ISSI chip may be cleared by accident. Or if the ISSI chip doesn't accept WRSR command with argument of two bytes (since it only have status register of one byte), it may fail to clear the other protect bits (BP0~BP3) as expected. This commit makes the `rom_spiflash_unlock()` check whether the vendor is issi. if so, `rom_spiflash_unlock()` only send RDSR to read the status register, send WRSR with only 1 byte argument, and also avoid clearing the QE bit (bit 6). 3. `rom_spiflash_unlock()` always send WRSR command to clear protection bits even when there is no protection bit active. And the execution of clearing status registers, which takes about 700us, will also happen even when there's no bits cleared. This commit skips the clearing of status register if there is no protection bits active. Also move the execute_flash_command to be a bootloader API; move implementation of spi_flash_wrap_set to the bootloader
2020-03-12 18:20:31 +08:00
#ifndef BOOTLOADER_BUILD
/* Normal app version maps to esp_spi_flash.h operations...
*/
static const char *TAG = "bootloader_mmap";
static spi_flash_mmap_handle_t map;
uint32_t bootloader_mmap_get_free_pages(void)
{
return spi_flash_mmap_get_free_pages(SPI_FLASH_MMAP_DATA);
}
const void *bootloader_mmap(uint32_t src_addr, uint32_t size)
{
if (map) {
ESP_LOGE(TAG, "tried to bootloader_mmap twice");
return NULL; /* existing mapping in use... */
}
const void *result = NULL;
uint32_t src_page = src_addr & ~(SPI_FLASH_MMU_PAGE_SIZE - 1);
size += (src_addr - src_page);
esp_err_t err = spi_flash_mmap(src_page, size, SPI_FLASH_MMAP_DATA, &result, &map);
if (err != ESP_OK) {
ESP_LOGE(TAG, "spi_flash_mmap failed: 0x%x", err);
return NULL;
}
return (void *)((intptr_t)result + (src_addr - src_page));
}
void bootloader_munmap(const void *mapping)
{
if (mapping && map) {
spi_flash_munmap(map);
}
map = 0;
}
esp_err_t bootloader_flash_read(size_t src, void *dest, size_t size, bool allow_decrypt)
{
if (allow_decrypt && esp_flash_encryption_enabled()) {
return spi_flash_read_encrypted(src, dest, size);
} else {
return spi_flash_read(src, dest, size);
}
}
esp_err_t bootloader_flash_write(size_t dest_addr, void *src, size_t size, bool write_encrypted)
{
if (write_encrypted) {
#if CONFIG_IDF_TARGET_ESP32
return spi_flash_write_encrypted(dest_addr, src, size);
#elif CONFIG_IDF_TARGET_ESP32S2BETA
return SPI_Encrypt_Write(dest_addr, src, size);
#endif
} else {
return spi_flash_write(dest_addr, src, size);
}
}
esp_err_t bootloader_flash_erase_sector(size_t sector)
{
return spi_flash_erase_sector(sector);
}
esp_err_t bootloader_flash_erase_range(uint32_t start_addr, uint32_t size)
{
return spi_flash_erase_range(start_addr, size);
}
#else //BOOTLOADER_BUILD
/* Bootloader version, uses ROM functions only */
#include <soc/dport_reg.h>
#if CONFIG_IDF_TARGET_ESP32
#include <esp32/rom/spi_flash.h>
#include <esp32/rom/cache.h>
#elif CONFIG_IDF_TARGET_ESP32S2BETA
#include <esp32s2beta/rom/spi_flash.h>
#include <esp32s2beta/rom/cache.h>
#endif
static const char *TAG = "bootloader_flash";
#if CONFIG_IDF_TARGET_ESP32
/* Use first 50 blocks in MMU for bootloader_mmap,
50th block for bootloader_flash_read
*/
#define MMU_BLOCK0_VADDR SOC_DROM_LOW
#define MMU_SIZE (0x320000)
#define MMU_BLOCK50_VADDR (MMU_BLOCK0_VADDR + MMU_SIZE)
#define FLASH_READ_VADDR MMU_BLOCK50_VADDR
#elif CONFIG_IDF_TARGET_ESP32S2BETA
/* Use first 63 blocks in MMU for bootloader_mmap,
63th block for bootloader_flash_read
*/
#define MMU_BLOCK0_VADDR SOC_DROM_LOW
#define MMU_SIZE (0x3f0000)
#define MMU_BLOCK63_VADDR (MMU_BLOCK0_VADDR + MMU_SIZE)
#define FLASH_READ_VADDR MMU_BLOCK63_VADDR
#endif
#define MMU_FREE_PAGES (MMU_SIZE / FLASH_BLOCK_SIZE)
static bool mapped;
// Current bootloader mapping (ab)used for bootloader_read()
static uint32_t current_read_mapping = UINT32_MAX;
uint32_t bootloader_mmap_get_free_pages(void)
{
/**
* Allow mapping up to 50 of the 51 available MMU blocks (last one used for reads)
* Since, bootloader_mmap function below assumes it to be 0x320000 (50 pages), we can safely do this.
*/
return MMU_FREE_PAGES;
}
const void *bootloader_mmap(uint32_t src_addr, uint32_t size)
{
if (mapped) {
ESP_LOGE(TAG, "tried to bootloader_mmap twice");
return NULL; /* can't map twice */
}
if (size > MMU_SIZE) {
ESP_LOGE(TAG, "bootloader_mmap excess size %x", size);
return NULL;
}
uint32_t src_addr_aligned = src_addr & MMU_FLASH_MASK;
uint32_t count = bootloader_cache_pages_to_map(size, src_addr);
#if CONFIG_IDF_TARGET_ESP32
Cache_Read_Disable(0);
Cache_Flush(0);
#elif CONFIG_IDF_TARGET_ESP32S2BETA
uint32_t autoload = Cache_Suspend_ICache();
Cache_Invalidate_ICache_All();
#endif
ESP_LOGD(TAG, "mmu set paddr=%08x count=%d size=%x src_addr=%x src_addr_aligned=%x",
src_addr & MMU_FLASH_MASK, count, size, src_addr, src_addr_aligned );
#if CONFIG_IDF_TARGET_ESP32
2017-04-05 17:26:37 +08:00
int e = cache_flash_mmu_set(0, 0, MMU_BLOCK0_VADDR, src_addr_aligned, 64, count);
#elif CONFIG_IDF_TARGET_ESP32S2BETA
int e = Cache_Ibus_MMU_Set(DPORT_MMU_ACCESS_FLASH, MMU_BLOCK0_VADDR, src_addr_aligned, 64, count, 0);
#endif
2017-04-05 17:26:37 +08:00
if (e != 0) {
ESP_LOGE(TAG, "cache_flash_mmu_set failed: %d\n", e);
#if CONFIG_IDF_TARGET_ESP32
2017-04-05 17:26:37 +08:00
Cache_Read_Enable(0);
#elif CONFIG_IDF_TARGET_ESP32S2BETA
Cache_Resume_ICache(autoload);
#endif
2017-04-05 17:26:37 +08:00
return NULL;
}
#if CONFIG_IDF_TARGET_ESP32
2017-04-05 17:26:37 +08:00
Cache_Read_Enable(0);
#elif CONFIG_IDF_TARGET_ESP32S2BETA
Cache_Resume_ICache(autoload);
#endif
mapped = true;
return (void *)(MMU_BLOCK0_VADDR + (src_addr - src_addr_aligned));
}
void bootloader_munmap(const void *mapping)
{
if (mapped) {
#if CONFIG_IDF_TARGET_ESP32
/* Full MMU reset */
Cache_Read_Disable(0);
Cache_Flush(0);
mmu_init(0);
#elif CONFIG_IDF_TARGET_ESP32S2BETA
//TODO, save the autoload value.
Cache_Suspend_ICache();
Cache_Invalidate_ICache_All();
Cache_MMU_Init();
#endif
mapped = false;
current_read_mapping = UINT32_MAX;
}
}
static esp_err_t spi_to_esp_err(esp_rom_spiflash_result_t r)
{
switch (r) {
case ESP_ROM_SPIFLASH_RESULT_OK:
return ESP_OK;
case ESP_ROM_SPIFLASH_RESULT_ERR:
return ESP_ERR_FLASH_OP_FAIL;
case ESP_ROM_SPIFLASH_RESULT_TIMEOUT:
return ESP_ERR_FLASH_OP_TIMEOUT;
default:
return ESP_FAIL;
}
}
static esp_err_t bootloader_flash_read_no_decrypt(size_t src_addr, void *dest, size_t size)
{
#if CONFIG_IDF_TARGET_ESP32
Cache_Read_Disable(0);
Cache_Flush(0);
#elif CONFIG_IDF_TARGET_ESP32S2BETA
uint32_t autoload = Cache_Suspend_ICache();
#endif
esp_rom_spiflash_result_t r = esp_rom_spiflash_read(src_addr, dest, size);
#if CONFIG_IDF_TARGET_ESP32
Cache_Read_Enable(0);
#elif CONFIG_IDF_TARGET_ESP32S2BETA
Cache_Resume_ICache(autoload);
#endif
return spi_to_esp_err(r);
}
static esp_err_t bootloader_flash_read_allow_decrypt(size_t src_addr, void *dest, size_t size)
{
uint32_t *dest_words = (uint32_t *)dest;
for (int word = 0; word < size / 4; word++) {
uint32_t word_src = src_addr + word * 4; /* Read this offset from flash */
uint32_t map_at = word_src & MMU_FLASH_MASK; /* Map this 64KB block from flash */
uint32_t *map_ptr;
if (map_at != current_read_mapping) {
/* Move the 64KB mmu mapping window to fit map_at */
#if CONFIG_IDF_TARGET_ESP32
Cache_Read_Disable(0);
Cache_Flush(0);
#elif CONFIG_IDF_TARGET_ESP32S2BETA
uint32_t autoload = Cache_Suspend_ICache();
Cache_Invalidate_ICache_All();
#endif
ESP_LOGD(TAG, "mmu set block paddr=0x%08x (was 0x%08x)", map_at, current_read_mapping);
#if CONFIG_IDF_TARGET_ESP32
int e = cache_flash_mmu_set(0, 0, FLASH_READ_VADDR, map_at, 64, 1);
#elif CONFIG_IDF_TARGET_ESP32S2BETA
int e = Cache_Ibus_MMU_Set(DPORT_MMU_ACCESS_FLASH, FLASH_READ_VADDR, map_at, 64, 1, 0);
#endif
if (e != 0) {
ESP_LOGE(TAG, "cache_flash_mmu_set failed: %d\n", e);
#if CONFIG_IDF_TARGET_ESP32
Cache_Read_Enable(0);
#elif CONFIG_IDF_TARGET_ESP32S2BETA
Cache_Resume_ICache(autoload);
#endif
return ESP_FAIL;
}
current_read_mapping = map_at;
#if CONFIG_IDF_TARGET_ESP32
Cache_Read_Enable(0);
#elif CONFIG_IDF_TARGET_ESP32S2BETA
Cache_Resume_ICache(autoload);
#endif
}
map_ptr = (uint32_t *)(FLASH_READ_VADDR + (word_src - map_at));
dest_words[word] = *map_ptr;
}
return ESP_OK;
}
esp_err_t bootloader_flash_read(size_t src_addr, void *dest, size_t size, bool allow_decrypt)
{
if (src_addr & 3) {
ESP_LOGE(TAG, "bootloader_flash_read src_addr 0x%x not 4-byte aligned", src_addr);
return ESP_FAIL;
}
if (size & 3) {
ESP_LOGE(TAG, "bootloader_flash_read size 0x%x not 4-byte aligned", size);
return ESP_FAIL;
}
if ((intptr_t)dest & 3) {
ESP_LOGE(TAG, "bootloader_flash_read dest 0x%x not 4-byte aligned", (intptr_t)dest);
return ESP_FAIL;
}
if (allow_decrypt) {
return bootloader_flash_read_allow_decrypt(src_addr, dest, size);
} else {
return bootloader_flash_read_no_decrypt(src_addr, dest, size);
}
}
esp_err_t bootloader_flash_write(size_t dest_addr, void *src, size_t size, bool write_encrypted)
{
esp_err_t err;
size_t alignment = write_encrypted ? 32 : 4;
if ((dest_addr % alignment) != 0) {
ESP_LOGE(TAG, "bootloader_flash_write dest_addr 0x%x not %d-byte aligned", dest_addr, alignment);
return ESP_FAIL;
}
if ((size % alignment) != 0) {
ESP_LOGE(TAG, "bootloader_flash_write size 0x%x not %d-byte aligned", size, alignment);
return ESP_FAIL;
}
if (((intptr_t)src % 4) != 0) {
ESP_LOGE(TAG, "bootloader_flash_write src 0x%x not 4 byte aligned", (intptr_t)src);
return ESP_FAIL;
}
err = spi_to_esp_err(esp_rom_spiflash_unlock());
if (err != ESP_OK) {
return err;
}
if (write_encrypted) {
#if CONFIG_IDF_TARGET_ESP32
return spi_to_esp_err(esp_rom_spiflash_write_encrypted(dest_addr, src, size));
#elif CONFIG_IDF_TARGET_ESP32S2BETA
// TODO: use the same ROM AP here
return spi_to_esp_err(SPI_Encrypt_Write(dest_addr, src, size));
#endif
} else {
return spi_to_esp_err(esp_rom_spiflash_write(dest_addr, src, size));
}
}
esp_err_t bootloader_flash_erase_sector(size_t sector)
{
return spi_to_esp_err(esp_rom_spiflash_erase_sector(sector));
}
esp_err_t bootloader_flash_erase_range(uint32_t start_addr, uint32_t size)
{
if (start_addr % FLASH_SECTOR_SIZE != 0) {
return ESP_ERR_INVALID_ARG;
}
if (size % FLASH_SECTOR_SIZE != 0) {
return ESP_ERR_INVALID_SIZE;
}
size_t start = start_addr / FLASH_SECTOR_SIZE;
size_t end = start + size / FLASH_SECTOR_SIZE;
const size_t sectors_per_block = FLASH_BLOCK_SIZE / FLASH_SECTOR_SIZE;
esp_rom_spiflash_result_t rc = ESP_ROM_SPIFLASH_RESULT_OK;
for (size_t sector = start; sector != end && rc == ESP_ROM_SPIFLASH_RESULT_OK; ) {
if (sector % sectors_per_block == 0 && end - sector >= sectors_per_block) {
rc = esp_rom_spiflash_erase_block(sector / sectors_per_block);
sector += sectors_per_block;
} else {
rc = esp_rom_spiflash_erase_sector(sector);
++sector;
}
}
return spi_to_esp_err(rc);
}
bootloader: fix the WRSR format for ISSI flash chips 1. The 2nd bootloader always call `rom_spiflash_unlock()`, but never help to clear the WEL bit when exit. This may cause system unstability. This commit helps to clear WEL when flash configuration is done. **RISK:** When the app starts, it didn't have to clear the WEL before it actually write/erase. But now the very first write/erase operation should be done after a WEL clear. Though the risk is little (all the following write/erase also need to clear the WEL), we still have to test this carefully, especially for those functions used by the OTA. 2. The `rom_spiflash_unlock()` function in the patch of ESP32 may (1) trigger the QPI, (2) clear the QE or (3) fail to unlock the ISSI chips. Status register bitmap of ISSI chip and GD chip: | SR | ISSI | GD25LQ32C | | -- | ---- | --------- | | 0 | WIP | WIP | | 1 | WEL | WEL | | 2 | BP0 | BP0 | | 3 | BP1 | BP1 | | 4 | BP2 | BP2 | | 5 | BP3 | BP3 | | 6 | QE | BP4 | | 7 | SRWD | SRP0 | | 8 | | SRP1 | | 9 | | QE | | 10 | | SUS2 | | 11 | | LB1 | | 12 | | LB2 | | 13 | | LB3 | | 14 | | CMP | | 15 | | SUS1 | QE bit of other chips are at the bit 9 of the status register (i.e. bit 1 of SR2), which should be read by RDSR2 command. However, the RDSR2 (35H, Read Status 2) command for chip of other vendors happens to be the QIOEN (Enter QPI mode) command of ISSI chips. When the `rom_spiflash_unlock()` function trys to read SR2, it may trigger the QPI of ISSI chips. Moreover, when `rom_spiflash_unlock()` try to clear the BP4 bit in the status register, QE (bit 6) of ISSI chip may be cleared by accident. Or if the ISSI chip doesn't accept WRSR command with argument of two bytes (since it only have status register of one byte), it may fail to clear the other protect bits (BP0~BP3) as expected. This commit makes the `rom_spiflash_unlock()` check whether the vendor is issi. if so, `rom_spiflash_unlock()` only send RDSR to read the status register, send WRSR with only 1 byte argument, and also avoid clearing the QE bit (bit 6). 3. `rom_spiflash_unlock()` always send WRSR command to clear protection bits even when there is no protection bit active. And the execution of clearing status registers, which takes about 700us, will also happen even when there's no bits cleared. This commit skips the clearing of status register if there is no protection bits active. Also move the execute_flash_command to be a bootloader API; move implementation of spi_flash_wrap_set to the bootloader
2020-03-12 18:20:31 +08:00
#endif // BOOTLOADER_BUILD
bootloader: fix the WRSR format for ISSI flash chips 1. The 2nd bootloader always call `rom_spiflash_unlock()`, but never help to clear the WEL bit when exit. This may cause system unstability. This commit helps to clear WEL when flash configuration is done. **RISK:** When the app starts, it didn't have to clear the WEL before it actually write/erase. But now the very first write/erase operation should be done after a WEL clear. Though the risk is little (all the following write/erase also need to clear the WEL), we still have to test this carefully, especially for those functions used by the OTA. 2. The `rom_spiflash_unlock()` function in the patch of ESP32 may (1) trigger the QPI, (2) clear the QE or (3) fail to unlock the ISSI chips. Status register bitmap of ISSI chip and GD chip: | SR | ISSI | GD25LQ32C | | -- | ---- | --------- | | 0 | WIP | WIP | | 1 | WEL | WEL | | 2 | BP0 | BP0 | | 3 | BP1 | BP1 | | 4 | BP2 | BP2 | | 5 | BP3 | BP3 | | 6 | QE | BP4 | | 7 | SRWD | SRP0 | | 8 | | SRP1 | | 9 | | QE | | 10 | | SUS2 | | 11 | | LB1 | | 12 | | LB2 | | 13 | | LB3 | | 14 | | CMP | | 15 | | SUS1 | QE bit of other chips are at the bit 9 of the status register (i.e. bit 1 of SR2), which should be read by RDSR2 command. However, the RDSR2 (35H, Read Status 2) command for chip of other vendors happens to be the QIOEN (Enter QPI mode) command of ISSI chips. When the `rom_spiflash_unlock()` function trys to read SR2, it may trigger the QPI of ISSI chips. Moreover, when `rom_spiflash_unlock()` try to clear the BP4 bit in the status register, QE (bit 6) of ISSI chip may be cleared by accident. Or if the ISSI chip doesn't accept WRSR command with argument of two bytes (since it only have status register of one byte), it may fail to clear the other protect bits (BP0~BP3) as expected. This commit makes the `rom_spiflash_unlock()` check whether the vendor is issi. if so, `rom_spiflash_unlock()` only send RDSR to read the status register, send WRSR with only 1 byte argument, and also avoid clearing the QE bit (bit 6). 3. `rom_spiflash_unlock()` always send WRSR command to clear protection bits even when there is no protection bit active. And the execution of clearing status registers, which takes about 700us, will also happen even when there's no bits cleared. This commit skips the clearing of status register if there is no protection bits active. Also move the execute_flash_command to be a bootloader API; move implementation of spi_flash_wrap_set to the bootloader
2020-03-12 18:20:31 +08:00
/* dummy_len_plus values defined in ROM for SPI flash configuration */
bootloader: fix the WRSR format for ISSI flash chips 1. The 2nd bootloader always call `rom_spiflash_unlock()`, but never help to clear the WEL bit when exit. This may cause system unstability. This commit helps to clear WEL when flash configuration is done. **RISK:** When the app starts, it didn't have to clear the WEL before it actually write/erase. But now the very first write/erase operation should be done after a WEL clear. Though the risk is little (all the following write/erase also need to clear the WEL), we still have to test this carefully, especially for those functions used by the OTA. 2. The `rom_spiflash_unlock()` function in the patch of ESP32 may (1) trigger the QPI, (2) clear the QE or (3) fail to unlock the ISSI chips. Status register bitmap of ISSI chip and GD chip: | SR | ISSI | GD25LQ32C | | -- | ---- | --------- | | 0 | WIP | WIP | | 1 | WEL | WEL | | 2 | BP0 | BP0 | | 3 | BP1 | BP1 | | 4 | BP2 | BP2 | | 5 | BP3 | BP3 | | 6 | QE | BP4 | | 7 | SRWD | SRP0 | | 8 | | SRP1 | | 9 | | QE | | 10 | | SUS2 | | 11 | | LB1 | | 12 | | LB2 | | 13 | | LB3 | | 14 | | CMP | | 15 | | SUS1 | QE bit of other chips are at the bit 9 of the status register (i.e. bit 1 of SR2), which should be read by RDSR2 command. However, the RDSR2 (35H, Read Status 2) command for chip of other vendors happens to be the QIOEN (Enter QPI mode) command of ISSI chips. When the `rom_spiflash_unlock()` function trys to read SR2, it may trigger the QPI of ISSI chips. Moreover, when `rom_spiflash_unlock()` try to clear the BP4 bit in the status register, QE (bit 6) of ISSI chip may be cleared by accident. Or if the ISSI chip doesn't accept WRSR command with argument of two bytes (since it only have status register of one byte), it may fail to clear the other protect bits (BP0~BP3) as expected. This commit makes the `rom_spiflash_unlock()` check whether the vendor is issi. if so, `rom_spiflash_unlock()` only send RDSR to read the status register, send WRSR with only 1 byte argument, and also avoid clearing the QE bit (bit 6). 3. `rom_spiflash_unlock()` always send WRSR command to clear protection bits even when there is no protection bit active. And the execution of clearing status registers, which takes about 700us, will also happen even when there's no bits cleared. This commit skips the clearing of status register if there is no protection bits active. Also move the execute_flash_command to be a bootloader API; move implementation of spi_flash_wrap_set to the bootloader
2020-03-12 18:20:31 +08:00
extern uint8_t g_rom_spiflash_dummy_len_plus[];
IRAM_ATTR static uint32_t bootloader_flash_execute_command_common(
uint8_t command,
uint32_t addr_len, uint32_t address,
uint8_t dummy_len,
uint8_t mosi_len, uint32_t mosi_data,
uint8_t miso_len)
bootloader: fix the WRSR format for ISSI flash chips 1. The 2nd bootloader always call `rom_spiflash_unlock()`, but never help to clear the WEL bit when exit. This may cause system unstability. This commit helps to clear WEL when flash configuration is done. **RISK:** When the app starts, it didn't have to clear the WEL before it actually write/erase. But now the very first write/erase operation should be done after a WEL clear. Though the risk is little (all the following write/erase also need to clear the WEL), we still have to test this carefully, especially for those functions used by the OTA. 2. The `rom_spiflash_unlock()` function in the patch of ESP32 may (1) trigger the QPI, (2) clear the QE or (3) fail to unlock the ISSI chips. Status register bitmap of ISSI chip and GD chip: | SR | ISSI | GD25LQ32C | | -- | ---- | --------- | | 0 | WIP | WIP | | 1 | WEL | WEL | | 2 | BP0 | BP0 | | 3 | BP1 | BP1 | | 4 | BP2 | BP2 | | 5 | BP3 | BP3 | | 6 | QE | BP4 | | 7 | SRWD | SRP0 | | 8 | | SRP1 | | 9 | | QE | | 10 | | SUS2 | | 11 | | LB1 | | 12 | | LB2 | | 13 | | LB3 | | 14 | | CMP | | 15 | | SUS1 | QE bit of other chips are at the bit 9 of the status register (i.e. bit 1 of SR2), which should be read by RDSR2 command. However, the RDSR2 (35H, Read Status 2) command for chip of other vendors happens to be the QIOEN (Enter QPI mode) command of ISSI chips. When the `rom_spiflash_unlock()` function trys to read SR2, it may trigger the QPI of ISSI chips. Moreover, when `rom_spiflash_unlock()` try to clear the BP4 bit in the status register, QE (bit 6) of ISSI chip may be cleared by accident. Or if the ISSI chip doesn't accept WRSR command with argument of two bytes (since it only have status register of one byte), it may fail to clear the other protect bits (BP0~BP3) as expected. This commit makes the `rom_spiflash_unlock()` check whether the vendor is issi. if so, `rom_spiflash_unlock()` only send RDSR to read the status register, send WRSR with only 1 byte argument, and also avoid clearing the QE bit (bit 6). 3. `rom_spiflash_unlock()` always send WRSR command to clear protection bits even when there is no protection bit active. And the execution of clearing status registers, which takes about 700us, will also happen even when there's no bits cleared. This commit skips the clearing of status register if there is no protection bits active. Also move the execute_flash_command to be a bootloader API; move implementation of spi_flash_wrap_set to the bootloader
2020-03-12 18:20:31 +08:00
{
assert(mosi_len <= 32);
assert(miso_len <= 32);
bootloader: fix the WRSR format for ISSI flash chips 1. The 2nd bootloader always call `rom_spiflash_unlock()`, but never help to clear the WEL bit when exit. This may cause system unstability. This commit helps to clear WEL when flash configuration is done. **RISK:** When the app starts, it didn't have to clear the WEL before it actually write/erase. But now the very first write/erase operation should be done after a WEL clear. Though the risk is little (all the following write/erase also need to clear the WEL), we still have to test this carefully, especially for those functions used by the OTA. 2. The `rom_spiflash_unlock()` function in the patch of ESP32 may (1) trigger the QPI, (2) clear the QE or (3) fail to unlock the ISSI chips. Status register bitmap of ISSI chip and GD chip: | SR | ISSI | GD25LQ32C | | -- | ---- | --------- | | 0 | WIP | WIP | | 1 | WEL | WEL | | 2 | BP0 | BP0 | | 3 | BP1 | BP1 | | 4 | BP2 | BP2 | | 5 | BP3 | BP3 | | 6 | QE | BP4 | | 7 | SRWD | SRP0 | | 8 | | SRP1 | | 9 | | QE | | 10 | | SUS2 | | 11 | | LB1 | | 12 | | LB2 | | 13 | | LB3 | | 14 | | CMP | | 15 | | SUS1 | QE bit of other chips are at the bit 9 of the status register (i.e. bit 1 of SR2), which should be read by RDSR2 command. However, the RDSR2 (35H, Read Status 2) command for chip of other vendors happens to be the QIOEN (Enter QPI mode) command of ISSI chips. When the `rom_spiflash_unlock()` function trys to read SR2, it may trigger the QPI of ISSI chips. Moreover, when `rom_spiflash_unlock()` try to clear the BP4 bit in the status register, QE (bit 6) of ISSI chip may be cleared by accident. Or if the ISSI chip doesn't accept WRSR command with argument of two bytes (since it only have status register of one byte), it may fail to clear the other protect bits (BP0~BP3) as expected. This commit makes the `rom_spiflash_unlock()` check whether the vendor is issi. if so, `rom_spiflash_unlock()` only send RDSR to read the status register, send WRSR with only 1 byte argument, and also avoid clearing the QE bit (bit 6). 3. `rom_spiflash_unlock()` always send WRSR command to clear protection bits even when there is no protection bit active. And the execution of clearing status registers, which takes about 700us, will also happen even when there's no bits cleared. This commit skips the clearing of status register if there is no protection bits active. Also move the execute_flash_command to be a bootloader API; move implementation of spi_flash_wrap_set to the bootloader
2020-03-12 18:20:31 +08:00
uint32_t old_ctrl_reg = SPIFLASH.ctrl.val;
#if CONFIG_IDF_TARGET_ESP32
SPIFLASH.ctrl.val = SPI_WP_REG_M; // keep WP high while idle, otherwise leave DIO mode
#else
bootloader: fix the WRSR format for ISSI flash chips 1. The 2nd bootloader always call `rom_spiflash_unlock()`, but never help to clear the WEL bit when exit. This may cause system unstability. This commit helps to clear WEL when flash configuration is done. **RISK:** When the app starts, it didn't have to clear the WEL before it actually write/erase. But now the very first write/erase operation should be done after a WEL clear. Though the risk is little (all the following write/erase also need to clear the WEL), we still have to test this carefully, especially for those functions used by the OTA. 2. The `rom_spiflash_unlock()` function in the patch of ESP32 may (1) trigger the QPI, (2) clear the QE or (3) fail to unlock the ISSI chips. Status register bitmap of ISSI chip and GD chip: | SR | ISSI | GD25LQ32C | | -- | ---- | --------- | | 0 | WIP | WIP | | 1 | WEL | WEL | | 2 | BP0 | BP0 | | 3 | BP1 | BP1 | | 4 | BP2 | BP2 | | 5 | BP3 | BP3 | | 6 | QE | BP4 | | 7 | SRWD | SRP0 | | 8 | | SRP1 | | 9 | | QE | | 10 | | SUS2 | | 11 | | LB1 | | 12 | | LB2 | | 13 | | LB3 | | 14 | | CMP | | 15 | | SUS1 | QE bit of other chips are at the bit 9 of the status register (i.e. bit 1 of SR2), which should be read by RDSR2 command. However, the RDSR2 (35H, Read Status 2) command for chip of other vendors happens to be the QIOEN (Enter QPI mode) command of ISSI chips. When the `rom_spiflash_unlock()` function trys to read SR2, it may trigger the QPI of ISSI chips. Moreover, when `rom_spiflash_unlock()` try to clear the BP4 bit in the status register, QE (bit 6) of ISSI chip may be cleared by accident. Or if the ISSI chip doesn't accept WRSR command with argument of two bytes (since it only have status register of one byte), it may fail to clear the other protect bits (BP0~BP3) as expected. This commit makes the `rom_spiflash_unlock()` check whether the vendor is issi. if so, `rom_spiflash_unlock()` only send RDSR to read the status register, send WRSR with only 1 byte argument, and also avoid clearing the QE bit (bit 6). 3. `rom_spiflash_unlock()` always send WRSR command to clear protection bits even when there is no protection bit active. And the execution of clearing status registers, which takes about 700us, will also happen even when there's no bits cleared. This commit skips the clearing of status register if there is no protection bits active. Also move the execute_flash_command to be a bootloader API; move implementation of spi_flash_wrap_set to the bootloader
2020-03-12 18:20:31 +08:00
SPIFLASH.ctrl.val = SPI_MEM_WP_REG_M; // keep WP high while idle, otherwise leave DIO mode
#endif
//command phase
bootloader: fix the WRSR format for ISSI flash chips 1. The 2nd bootloader always call `rom_spiflash_unlock()`, but never help to clear the WEL bit when exit. This may cause system unstability. This commit helps to clear WEL when flash configuration is done. **RISK:** When the app starts, it didn't have to clear the WEL before it actually write/erase. But now the very first write/erase operation should be done after a WEL clear. Though the risk is little (all the following write/erase also need to clear the WEL), we still have to test this carefully, especially for those functions used by the OTA. 2. The `rom_spiflash_unlock()` function in the patch of ESP32 may (1) trigger the QPI, (2) clear the QE or (3) fail to unlock the ISSI chips. Status register bitmap of ISSI chip and GD chip: | SR | ISSI | GD25LQ32C | | -- | ---- | --------- | | 0 | WIP | WIP | | 1 | WEL | WEL | | 2 | BP0 | BP0 | | 3 | BP1 | BP1 | | 4 | BP2 | BP2 | | 5 | BP3 | BP3 | | 6 | QE | BP4 | | 7 | SRWD | SRP0 | | 8 | | SRP1 | | 9 | | QE | | 10 | | SUS2 | | 11 | | LB1 | | 12 | | LB2 | | 13 | | LB3 | | 14 | | CMP | | 15 | | SUS1 | QE bit of other chips are at the bit 9 of the status register (i.e. bit 1 of SR2), which should be read by RDSR2 command. However, the RDSR2 (35H, Read Status 2) command for chip of other vendors happens to be the QIOEN (Enter QPI mode) command of ISSI chips. When the `rom_spiflash_unlock()` function trys to read SR2, it may trigger the QPI of ISSI chips. Moreover, when `rom_spiflash_unlock()` try to clear the BP4 bit in the status register, QE (bit 6) of ISSI chip may be cleared by accident. Or if the ISSI chip doesn't accept WRSR command with argument of two bytes (since it only have status register of one byte), it may fail to clear the other protect bits (BP0~BP3) as expected. This commit makes the `rom_spiflash_unlock()` check whether the vendor is issi. if so, `rom_spiflash_unlock()` only send RDSR to read the status register, send WRSR with only 1 byte argument, and also avoid clearing the QE bit (bit 6). 3. `rom_spiflash_unlock()` always send WRSR command to clear protection bits even when there is no protection bit active. And the execution of clearing status registers, which takes about 700us, will also happen even when there's no bits cleared. This commit skips the clearing of status register if there is no protection bits active. Also move the execute_flash_command to be a bootloader API; move implementation of spi_flash_wrap_set to the bootloader
2020-03-12 18:20:31 +08:00
SPIFLASH.user.usr_command = 1;
SPIFLASH.user2.usr_command_bitlen = 7;
SPIFLASH.user2.usr_command_value = command;
//addr phase
SPIFLASH.user.usr_addr = addr_len > 0;
SPIFLASH.user1.usr_addr_bitlen = addr_len - 1;
bootloader: fix the WRSR format for ISSI flash chips 1. The 2nd bootloader always call `rom_spiflash_unlock()`, but never help to clear the WEL bit when exit. This may cause system unstability. This commit helps to clear WEL when flash configuration is done. **RISK:** When the app starts, it didn't have to clear the WEL before it actually write/erase. But now the very first write/erase operation should be done after a WEL clear. Though the risk is little (all the following write/erase also need to clear the WEL), we still have to test this carefully, especially for those functions used by the OTA. 2. The `rom_spiflash_unlock()` function in the patch of ESP32 may (1) trigger the QPI, (2) clear the QE or (3) fail to unlock the ISSI chips. Status register bitmap of ISSI chip and GD chip: | SR | ISSI | GD25LQ32C | | -- | ---- | --------- | | 0 | WIP | WIP | | 1 | WEL | WEL | | 2 | BP0 | BP0 | | 3 | BP1 | BP1 | | 4 | BP2 | BP2 | | 5 | BP3 | BP3 | | 6 | QE | BP4 | | 7 | SRWD | SRP0 | | 8 | | SRP1 | | 9 | | QE | | 10 | | SUS2 | | 11 | | LB1 | | 12 | | LB2 | | 13 | | LB3 | | 14 | | CMP | | 15 | | SUS1 | QE bit of other chips are at the bit 9 of the status register (i.e. bit 1 of SR2), which should be read by RDSR2 command. However, the RDSR2 (35H, Read Status 2) command for chip of other vendors happens to be the QIOEN (Enter QPI mode) command of ISSI chips. When the `rom_spiflash_unlock()` function trys to read SR2, it may trigger the QPI of ISSI chips. Moreover, when `rom_spiflash_unlock()` try to clear the BP4 bit in the status register, QE (bit 6) of ISSI chip may be cleared by accident. Or if the ISSI chip doesn't accept WRSR command with argument of two bytes (since it only have status register of one byte), it may fail to clear the other protect bits (BP0~BP3) as expected. This commit makes the `rom_spiflash_unlock()` check whether the vendor is issi. if so, `rom_spiflash_unlock()` only send RDSR to read the status register, send WRSR with only 1 byte argument, and also avoid clearing the QE bit (bit 6). 3. `rom_spiflash_unlock()` always send WRSR command to clear protection bits even when there is no protection bit active. And the execution of clearing status registers, which takes about 700us, will also happen even when there's no bits cleared. This commit skips the clearing of status register if there is no protection bits active. Also move the execute_flash_command to be a bootloader API; move implementation of spi_flash_wrap_set to the bootloader
2020-03-12 18:20:31 +08:00
#if CONFIG_IDF_TARGET_ESP32
SPIFLASH.addr = (addr_len > 0)? (address << (32-addr_len)) : 0;
#else
SPIFLASH.addr = address;
bootloader: fix the WRSR format for ISSI flash chips 1. The 2nd bootloader always call `rom_spiflash_unlock()`, but never help to clear the WEL bit when exit. This may cause system unstability. This commit helps to clear WEL when flash configuration is done. **RISK:** When the app starts, it didn't have to clear the WEL before it actually write/erase. But now the very first write/erase operation should be done after a WEL clear. Though the risk is little (all the following write/erase also need to clear the WEL), we still have to test this carefully, especially for those functions used by the OTA. 2. The `rom_spiflash_unlock()` function in the patch of ESP32 may (1) trigger the QPI, (2) clear the QE or (3) fail to unlock the ISSI chips. Status register bitmap of ISSI chip and GD chip: | SR | ISSI | GD25LQ32C | | -- | ---- | --------- | | 0 | WIP | WIP | | 1 | WEL | WEL | | 2 | BP0 | BP0 | | 3 | BP1 | BP1 | | 4 | BP2 | BP2 | | 5 | BP3 | BP3 | | 6 | QE | BP4 | | 7 | SRWD | SRP0 | | 8 | | SRP1 | | 9 | | QE | | 10 | | SUS2 | | 11 | | LB1 | | 12 | | LB2 | | 13 | | LB3 | | 14 | | CMP | | 15 | | SUS1 | QE bit of other chips are at the bit 9 of the status register (i.e. bit 1 of SR2), which should be read by RDSR2 command. However, the RDSR2 (35H, Read Status 2) command for chip of other vendors happens to be the QIOEN (Enter QPI mode) command of ISSI chips. When the `rom_spiflash_unlock()` function trys to read SR2, it may trigger the QPI of ISSI chips. Moreover, when `rom_spiflash_unlock()` try to clear the BP4 bit in the status register, QE (bit 6) of ISSI chip may be cleared by accident. Or if the ISSI chip doesn't accept WRSR command with argument of two bytes (since it only have status register of one byte), it may fail to clear the other protect bits (BP0~BP3) as expected. This commit makes the `rom_spiflash_unlock()` check whether the vendor is issi. if so, `rom_spiflash_unlock()` only send RDSR to read the status register, send WRSR with only 1 byte argument, and also avoid clearing the QE bit (bit 6). 3. `rom_spiflash_unlock()` always send WRSR command to clear protection bits even when there is no protection bit active. And the execution of clearing status registers, which takes about 700us, will also happen even when there's no bits cleared. This commit skips the clearing of status register if there is no protection bits active. Also move the execute_flash_command to be a bootloader API; move implementation of spi_flash_wrap_set to the bootloader
2020-03-12 18:20:31 +08:00
#endif
//dummy phase
if (miso_len > 0) {
uint32_t total_dummy = dummy_len + g_rom_spiflash_dummy_len_plus[1];
SPIFLASH.user.usr_dummy = total_dummy > 0;
SPIFLASH.user1.usr_dummy_cyclelen = total_dummy - 1;
} else {
SPIFLASH.user.usr_dummy = 0;
SPIFLASH.user1.usr_dummy_cyclelen = 0;
}
//output data
bootloader: fix the WRSR format for ISSI flash chips 1. The 2nd bootloader always call `rom_spiflash_unlock()`, but never help to clear the WEL bit when exit. This may cause system unstability. This commit helps to clear WEL when flash configuration is done. **RISK:** When the app starts, it didn't have to clear the WEL before it actually write/erase. But now the very first write/erase operation should be done after a WEL clear. Though the risk is little (all the following write/erase also need to clear the WEL), we still have to test this carefully, especially for those functions used by the OTA. 2. The `rom_spiflash_unlock()` function in the patch of ESP32 may (1) trigger the QPI, (2) clear the QE or (3) fail to unlock the ISSI chips. Status register bitmap of ISSI chip and GD chip: | SR | ISSI | GD25LQ32C | | -- | ---- | --------- | | 0 | WIP | WIP | | 1 | WEL | WEL | | 2 | BP0 | BP0 | | 3 | BP1 | BP1 | | 4 | BP2 | BP2 | | 5 | BP3 | BP3 | | 6 | QE | BP4 | | 7 | SRWD | SRP0 | | 8 | | SRP1 | | 9 | | QE | | 10 | | SUS2 | | 11 | | LB1 | | 12 | | LB2 | | 13 | | LB3 | | 14 | | CMP | | 15 | | SUS1 | QE bit of other chips are at the bit 9 of the status register (i.e. bit 1 of SR2), which should be read by RDSR2 command. However, the RDSR2 (35H, Read Status 2) command for chip of other vendors happens to be the QIOEN (Enter QPI mode) command of ISSI chips. When the `rom_spiflash_unlock()` function trys to read SR2, it may trigger the QPI of ISSI chips. Moreover, when `rom_spiflash_unlock()` try to clear the BP4 bit in the status register, QE (bit 6) of ISSI chip may be cleared by accident. Or if the ISSI chip doesn't accept WRSR command with argument of two bytes (since it only have status register of one byte), it may fail to clear the other protect bits (BP0~BP3) as expected. This commit makes the `rom_spiflash_unlock()` check whether the vendor is issi. if so, `rom_spiflash_unlock()` only send RDSR to read the status register, send WRSR with only 1 byte argument, and also avoid clearing the QE bit (bit 6). 3. `rom_spiflash_unlock()` always send WRSR command to clear protection bits even when there is no protection bit active. And the execution of clearing status registers, which takes about 700us, will also happen even when there's no bits cleared. This commit skips the clearing of status register if there is no protection bits active. Also move the execute_flash_command to be a bootloader API; move implementation of spi_flash_wrap_set to the bootloader
2020-03-12 18:20:31 +08:00
SPIFLASH.user.usr_mosi = mosi_len > 0;
#if CONFIG_IDF_TARGET_ESP32
SPIFLASH.mosi_dlen.usr_mosi_dbitlen = mosi_len ? (mosi_len - 1) : 0;
#else
bootloader: fix the WRSR format for ISSI flash chips 1. The 2nd bootloader always call `rom_spiflash_unlock()`, but never help to clear the WEL bit when exit. This may cause system unstability. This commit helps to clear WEL when flash configuration is done. **RISK:** When the app starts, it didn't have to clear the WEL before it actually write/erase. But now the very first write/erase operation should be done after a WEL clear. Though the risk is little (all the following write/erase also need to clear the WEL), we still have to test this carefully, especially for those functions used by the OTA. 2. The `rom_spiflash_unlock()` function in the patch of ESP32 may (1) trigger the QPI, (2) clear the QE or (3) fail to unlock the ISSI chips. Status register bitmap of ISSI chip and GD chip: | SR | ISSI | GD25LQ32C | | -- | ---- | --------- | | 0 | WIP | WIP | | 1 | WEL | WEL | | 2 | BP0 | BP0 | | 3 | BP1 | BP1 | | 4 | BP2 | BP2 | | 5 | BP3 | BP3 | | 6 | QE | BP4 | | 7 | SRWD | SRP0 | | 8 | | SRP1 | | 9 | | QE | | 10 | | SUS2 | | 11 | | LB1 | | 12 | | LB2 | | 13 | | LB3 | | 14 | | CMP | | 15 | | SUS1 | QE bit of other chips are at the bit 9 of the status register (i.e. bit 1 of SR2), which should be read by RDSR2 command. However, the RDSR2 (35H, Read Status 2) command for chip of other vendors happens to be the QIOEN (Enter QPI mode) command of ISSI chips. When the `rom_spiflash_unlock()` function trys to read SR2, it may trigger the QPI of ISSI chips. Moreover, when `rom_spiflash_unlock()` try to clear the BP4 bit in the status register, QE (bit 6) of ISSI chip may be cleared by accident. Or if the ISSI chip doesn't accept WRSR command with argument of two bytes (since it only have status register of one byte), it may fail to clear the other protect bits (BP0~BP3) as expected. This commit makes the `rom_spiflash_unlock()` check whether the vendor is issi. if so, `rom_spiflash_unlock()` only send RDSR to read the status register, send WRSR with only 1 byte argument, and also avoid clearing the QE bit (bit 6). 3. `rom_spiflash_unlock()` always send WRSR command to clear protection bits even when there is no protection bit active. And the execution of clearing status registers, which takes about 700us, will also happen even when there's no bits cleared. This commit skips the clearing of status register if there is no protection bits active. Also move the execute_flash_command to be a bootloader API; move implementation of spi_flash_wrap_set to the bootloader
2020-03-12 18:20:31 +08:00
SPIFLASH.mosi_dlen.usr_mosi_bit_len = mosi_len ? (mosi_len - 1) : 0;
#endif
bootloader: fix the WRSR format for ISSI flash chips 1. The 2nd bootloader always call `rom_spiflash_unlock()`, but never help to clear the WEL bit when exit. This may cause system unstability. This commit helps to clear WEL when flash configuration is done. **RISK:** When the app starts, it didn't have to clear the WEL before it actually write/erase. But now the very first write/erase operation should be done after a WEL clear. Though the risk is little (all the following write/erase also need to clear the WEL), we still have to test this carefully, especially for those functions used by the OTA. 2. The `rom_spiflash_unlock()` function in the patch of ESP32 may (1) trigger the QPI, (2) clear the QE or (3) fail to unlock the ISSI chips. Status register bitmap of ISSI chip and GD chip: | SR | ISSI | GD25LQ32C | | -- | ---- | --------- | | 0 | WIP | WIP | | 1 | WEL | WEL | | 2 | BP0 | BP0 | | 3 | BP1 | BP1 | | 4 | BP2 | BP2 | | 5 | BP3 | BP3 | | 6 | QE | BP4 | | 7 | SRWD | SRP0 | | 8 | | SRP1 | | 9 | | QE | | 10 | | SUS2 | | 11 | | LB1 | | 12 | | LB2 | | 13 | | LB3 | | 14 | | CMP | | 15 | | SUS1 | QE bit of other chips are at the bit 9 of the status register (i.e. bit 1 of SR2), which should be read by RDSR2 command. However, the RDSR2 (35H, Read Status 2) command for chip of other vendors happens to be the QIOEN (Enter QPI mode) command of ISSI chips. When the `rom_spiflash_unlock()` function trys to read SR2, it may trigger the QPI of ISSI chips. Moreover, when `rom_spiflash_unlock()` try to clear the BP4 bit in the status register, QE (bit 6) of ISSI chip may be cleared by accident. Or if the ISSI chip doesn't accept WRSR command with argument of two bytes (since it only have status register of one byte), it may fail to clear the other protect bits (BP0~BP3) as expected. This commit makes the `rom_spiflash_unlock()` check whether the vendor is issi. if so, `rom_spiflash_unlock()` only send RDSR to read the status register, send WRSR with only 1 byte argument, and also avoid clearing the QE bit (bit 6). 3. `rom_spiflash_unlock()` always send WRSR command to clear protection bits even when there is no protection bit active. And the execution of clearing status registers, which takes about 700us, will also happen even when there's no bits cleared. This commit skips the clearing of status register if there is no protection bits active. Also move the execute_flash_command to be a bootloader API; move implementation of spi_flash_wrap_set to the bootloader
2020-03-12 18:20:31 +08:00
SPIFLASH.data_buf[0] = mosi_data;
//input data
SPIFLASH.user.usr_miso = miso_len > 0;
#if CONFIG_IDF_TARGET_ESP32
SPIFLASH.miso_dlen.usr_miso_dbitlen = miso_len ? (miso_len - 1) : 0;
#else
SPIFLASH.miso_dlen.usr_miso_bit_len = miso_len ? (miso_len - 1) : 0;
#endif
bootloader: fix the WRSR format for ISSI flash chips 1. The 2nd bootloader always call `rom_spiflash_unlock()`, but never help to clear the WEL bit when exit. This may cause system unstability. This commit helps to clear WEL when flash configuration is done. **RISK:** When the app starts, it didn't have to clear the WEL before it actually write/erase. But now the very first write/erase operation should be done after a WEL clear. Though the risk is little (all the following write/erase also need to clear the WEL), we still have to test this carefully, especially for those functions used by the OTA. 2. The `rom_spiflash_unlock()` function in the patch of ESP32 may (1) trigger the QPI, (2) clear the QE or (3) fail to unlock the ISSI chips. Status register bitmap of ISSI chip and GD chip: | SR | ISSI | GD25LQ32C | | -- | ---- | --------- | | 0 | WIP | WIP | | 1 | WEL | WEL | | 2 | BP0 | BP0 | | 3 | BP1 | BP1 | | 4 | BP2 | BP2 | | 5 | BP3 | BP3 | | 6 | QE | BP4 | | 7 | SRWD | SRP0 | | 8 | | SRP1 | | 9 | | QE | | 10 | | SUS2 | | 11 | | LB1 | | 12 | | LB2 | | 13 | | LB3 | | 14 | | CMP | | 15 | | SUS1 | QE bit of other chips are at the bit 9 of the status register (i.e. bit 1 of SR2), which should be read by RDSR2 command. However, the RDSR2 (35H, Read Status 2) command for chip of other vendors happens to be the QIOEN (Enter QPI mode) command of ISSI chips. When the `rom_spiflash_unlock()` function trys to read SR2, it may trigger the QPI of ISSI chips. Moreover, when `rom_spiflash_unlock()` try to clear the BP4 bit in the status register, QE (bit 6) of ISSI chip may be cleared by accident. Or if the ISSI chip doesn't accept WRSR command with argument of two bytes (since it only have status register of one byte), it may fail to clear the other protect bits (BP0~BP3) as expected. This commit makes the `rom_spiflash_unlock()` check whether the vendor is issi. if so, `rom_spiflash_unlock()` only send RDSR to read the status register, send WRSR with only 1 byte argument, and also avoid clearing the QE bit (bit 6). 3. `rom_spiflash_unlock()` always send WRSR command to clear protection bits even when there is no protection bit active. And the execution of clearing status registers, which takes about 700us, will also happen even when there's no bits cleared. This commit skips the clearing of status register if there is no protection bits active. Also move the execute_flash_command to be a bootloader API; move implementation of spi_flash_wrap_set to the bootloader
2020-03-12 18:20:31 +08:00
SPIFLASH.cmd.usr = 1;
while (SPIFLASH.cmd.usr != 0) {
}
SPIFLASH.ctrl.val = old_ctrl_reg;
uint32_t ret = SPIFLASH.data_buf[0];
if (miso_len < 32) {
//set unused bits to 0
ret &= ~(UINT32_MAX << miso_len);
}
return ret;
}
uint32_t bootloader_execute_flash_command(uint8_t command, uint32_t mosi_data, uint8_t mosi_len, uint8_t miso_len)
{
const uint8_t addr_len = 0;
const uint8_t address = 0;
const uint8_t dummy_len = 0;
return bootloader_flash_execute_command_common(command, addr_len, address,
dummy_len, mosi_len, mosi_data, miso_len);
}
// cmd(0x5A) + 24bit address + 8 cycles dummy
uint32_t IRAM_ATTR bootloader_flash_read_sfdp(uint32_t sfdp_addr, unsigned int miso_byte_num)
{
assert(miso_byte_num <= 4);
const uint8_t command = CMD_RDSFDP;
const uint8_t addr_len = 24;
const uint8_t dummy_len = 8;
const uint8_t mosi_len = 0;
const uint32_t mosi_data = 0;
const uint8_t miso_len = miso_byte_num * 8;
return bootloader_flash_execute_command_common(command, addr_len, sfdp_addr,
dummy_len, mosi_len, mosi_data, miso_len);
bootloader: fix the WRSR format for ISSI flash chips 1. The 2nd bootloader always call `rom_spiflash_unlock()`, but never help to clear the WEL bit when exit. This may cause system unstability. This commit helps to clear WEL when flash configuration is done. **RISK:** When the app starts, it didn't have to clear the WEL before it actually write/erase. But now the very first write/erase operation should be done after a WEL clear. Though the risk is little (all the following write/erase also need to clear the WEL), we still have to test this carefully, especially for those functions used by the OTA. 2. The `rom_spiflash_unlock()` function in the patch of ESP32 may (1) trigger the QPI, (2) clear the QE or (3) fail to unlock the ISSI chips. Status register bitmap of ISSI chip and GD chip: | SR | ISSI | GD25LQ32C | | -- | ---- | --------- | | 0 | WIP | WIP | | 1 | WEL | WEL | | 2 | BP0 | BP0 | | 3 | BP1 | BP1 | | 4 | BP2 | BP2 | | 5 | BP3 | BP3 | | 6 | QE | BP4 | | 7 | SRWD | SRP0 | | 8 | | SRP1 | | 9 | | QE | | 10 | | SUS2 | | 11 | | LB1 | | 12 | | LB2 | | 13 | | LB3 | | 14 | | CMP | | 15 | | SUS1 | QE bit of other chips are at the bit 9 of the status register (i.e. bit 1 of SR2), which should be read by RDSR2 command. However, the RDSR2 (35H, Read Status 2) command for chip of other vendors happens to be the QIOEN (Enter QPI mode) command of ISSI chips. When the `rom_spiflash_unlock()` function trys to read SR2, it may trigger the QPI of ISSI chips. Moreover, when `rom_spiflash_unlock()` try to clear the BP4 bit in the status register, QE (bit 6) of ISSI chip may be cleared by accident. Or if the ISSI chip doesn't accept WRSR command with argument of two bytes (since it only have status register of one byte), it may fail to clear the other protect bits (BP0~BP3) as expected. This commit makes the `rom_spiflash_unlock()` check whether the vendor is issi. if so, `rom_spiflash_unlock()` only send RDSR to read the status register, send WRSR with only 1 byte argument, and also avoid clearing the QE bit (bit 6). 3. `rom_spiflash_unlock()` always send WRSR command to clear protection bits even when there is no protection bit active. And the execution of clearing status registers, which takes about 700us, will also happen even when there's no bits cleared. This commit skips the clearing of status register if there is no protection bits active. Also move the execute_flash_command to be a bootloader API; move implementation of spi_flash_wrap_set to the bootloader
2020-03-12 18:20:31 +08:00
}
void bootloader_enable_wp(void)
{
bootloader_execute_flash_command(CMD_WRDI, 0, 0, 0); /* Exit OTP mode */
}
uint32_t IRAM_ATTR bootloader_read_flash_id(void)
{
uint32_t id = bootloader_execute_flash_command(CMD_RDID, 0, 0, 24);
id = ((id & 0xff) << 16) | ((id >> 16) & 0xff) | (id & 0xff00);
return id;
}
bootloader: fix the WRSR format for ISSI flash chips 1. The 2nd bootloader always call `rom_spiflash_unlock()`, but never help to clear the WEL bit when exit. This may cause system unstability. This commit helps to clear WEL when flash configuration is done. **RISK:** When the app starts, it didn't have to clear the WEL before it actually write/erase. But now the very first write/erase operation should be done after a WEL clear. Though the risk is little (all the following write/erase also need to clear the WEL), we still have to test this carefully, especially for those functions used by the OTA. 2. The `rom_spiflash_unlock()` function in the patch of ESP32 may (1) trigger the QPI, (2) clear the QE or (3) fail to unlock the ISSI chips. Status register bitmap of ISSI chip and GD chip: | SR | ISSI | GD25LQ32C | | -- | ---- | --------- | | 0 | WIP | WIP | | 1 | WEL | WEL | | 2 | BP0 | BP0 | | 3 | BP1 | BP1 | | 4 | BP2 | BP2 | | 5 | BP3 | BP3 | | 6 | QE | BP4 | | 7 | SRWD | SRP0 | | 8 | | SRP1 | | 9 | | QE | | 10 | | SUS2 | | 11 | | LB1 | | 12 | | LB2 | | 13 | | LB3 | | 14 | | CMP | | 15 | | SUS1 | QE bit of other chips are at the bit 9 of the status register (i.e. bit 1 of SR2), which should be read by RDSR2 command. However, the RDSR2 (35H, Read Status 2) command for chip of other vendors happens to be the QIOEN (Enter QPI mode) command of ISSI chips. When the `rom_spiflash_unlock()` function trys to read SR2, it may trigger the QPI of ISSI chips. Moreover, when `rom_spiflash_unlock()` try to clear the BP4 bit in the status register, QE (bit 6) of ISSI chip may be cleared by accident. Or if the ISSI chip doesn't accept WRSR command with argument of two bytes (since it only have status register of one byte), it may fail to clear the other protect bits (BP0~BP3) as expected. This commit makes the `rom_spiflash_unlock()` check whether the vendor is issi. if so, `rom_spiflash_unlock()` only send RDSR to read the status register, send WRSR with only 1 byte argument, and also avoid clearing the QE bit (bit 6). 3. `rom_spiflash_unlock()` always send WRSR command to clear protection bits even when there is no protection bit active. And the execution of clearing status registers, which takes about 700us, will also happen even when there's no bits cleared. This commit skips the clearing of status register if there is no protection bits active. Also move the execute_flash_command to be a bootloader API; move implementation of spi_flash_wrap_set to the bootloader
2020-03-12 18:20:31 +08:00
#if SOC_CACHE_SUPPORT_WRAP
esp_err_t bootloader_flash_wrap_set(spi_flash_wrap_mode_t mode)
{
uint32_t reg_bkp_ctrl = SPIFLASH.ctrl.val;
uint32_t reg_bkp_usr = SPIFLASH.user.val;
SPIFLASH.user.fwrite_dio = 0;
SPIFLASH.user.fwrite_dual = 0;
SPIFLASH.user.fwrite_qio = 1;
SPIFLASH.user.fwrite_quad = 0;
SPIFLASH.ctrl.fcmd_dual = 0;
SPIFLASH.ctrl.fcmd_quad = 0;
SPIFLASH.user.usr_dummy = 0;
SPIFLASH.user.usr_addr = 1;
SPIFLASH.user.usr_command = 1;
SPIFLASH.user2.usr_command_bitlen = 7;
SPIFLASH.user2.usr_command_value = CMD_WRAP;
SPIFLASH.user1.usr_addr_bitlen = 23;
SPIFLASH.addr = 0;
SPIFLASH.user.usr_miso = 0;
SPIFLASH.user.usr_mosi = 1;
SPIFLASH.mosi_dlen.usr_mosi_bit_len = 7;
SPIFLASH.data_buf[0] = (uint32_t) mode << 4;;
SPIFLASH.cmd.usr = 1;
while(SPIFLASH.cmd.usr != 0)
{ }
SPIFLASH.ctrl.val = reg_bkp_ctrl;
SPIFLASH.user.val = reg_bkp_usr;
return ESP_OK;
}
#endif //SOC_CACHE_SUPPORT_WRAP
/*******************************************************************************
* XMC startup flow
******************************************************************************/
#define XMC_SUPPORT CONFIG_BOOTLOADER_FLASH_XMC_SUPPORT
#define XMC_VENDOR_ID 0x20
#if BOOTLOADER_BUILD
#define BOOTLOADER_FLASH_LOG(level, ...) ESP_LOG##level(TAG, ##__VA_ARGS__)
#else
static DRAM_ATTR char bootloader_flash_tag[] = "bootloader_flash";
#define BOOTLOADER_FLASH_LOG(level, ...) ESP_EARLY_LOG##level(bootloader_flash_tag, ##__VA_ARGS__)
#endif
#if XMC_SUPPORT
//strictly check the model
static IRAM_ATTR bool is_xmc_chip_strict(uint32_t rdid)
{
uint32_t vendor_id = BYTESHIFT(rdid, 2);
uint32_t mfid = BYTESHIFT(rdid, 1);
uint32_t cpid = BYTESHIFT(rdid, 0);
if (vendor_id != XMC_VENDOR_ID) {
return false;
}
bool matched = false;
if (mfid == 0x40) {
if (cpid >= 0x13 && cpid <= 0x20) {
matched = true;
}
} else if (mfid == 0x41) {
if (cpid >= 0x17 && cpid <= 0x20) {
matched = true;
}
} else if (mfid == 0x50) {
if (cpid >= 0x15 && cpid <= 0x16) {
matched = true;
}
}
return matched;
}
esp_err_t IRAM_ATTR bootloader_flash_xmc_startup(void)
{
// If the RDID value is a valid XMC one, may skip the flow
const bool fast_check = true;
if (fast_check && is_xmc_chip_strict(g_rom_flashchip.device_id)) {
BOOTLOADER_FLASH_LOG(D, "XMC chip detected by RDID (%08X), skip.", g_rom_flashchip.device_id);
return ESP_OK;
}
// Check the Manufacturer ID in SFDP registers (JEDEC standard). If not XMC chip, no need to run the flow
const int sfdp_mfid_addr = 0x10;
uint8_t mf_id = (bootloader_flash_read_sfdp(sfdp_mfid_addr, 1) & 0xff);
if (mf_id != XMC_VENDOR_ID) {
BOOTLOADER_FLASH_LOG(D, "non-XMC chip detected by SFDP Read (%02X), skip.", mf_id);
return ESP_OK;
}
BOOTLOADER_FLASH_LOG(I, "XM25QHxxC startup flow");
// Enter DPD
bootloader_execute_flash_command(0xB9, 0, 0, 0);
// Enter UDPD
bootloader_execute_flash_command(0x79, 0, 0, 0);
// Exit UDPD
bootloader_execute_flash_command(0xFF, 0, 0, 0);
// Delay tXUDPD
ets_delay_us(2000);
// Release Power-down
bootloader_execute_flash_command(0xAB, 0, 0, 0);
ets_delay_us(20);
// Read flash ID and check again
g_rom_flashchip.device_id = bootloader_read_flash_id();
if (!is_xmc_chip_strict(g_rom_flashchip.device_id)) {
BOOTLOADER_FLASH_LOG(E, "XMC flash startup fail");
return ESP_FAIL;
}
return ESP_OK;
}
#else
//only compare the vendor id
static IRAM_ATTR bool is_xmc_chip(uint32_t rdid)
{
uint32_t vendor_id = (rdid >> 16) & 0xFF;
return (vendor_id == XMC_VENDOR_ID);
}
esp_err_t IRAM_ATTR bootloader_flash_xmc_startup(void)
{
if (is_xmc_chip(g_rom_flashchip.device_id)) {
BOOTLOADER_FLASH_LOG(E, "XMC chip detected (%08X) while support disabled.", g_rom_flashchip.device_id);
return ESP_FAIL;
}
return ESP_OK;
}
#endif //XMC_SUPPORT