{
size_t block_len = block_length(sha_type);
+ // Max number of blocks to pass per each call to esp_sha_lock_memory_block()
+ // (keep low enough to avoid starving interrupt handlers, especially if reading
+ // data into SHA via flash cache, but high enough that spinlock overhead is low)
+ const size_t BLOCKS_PER_CHUNK = 100;
+ const size_t MAX_CHUNK_LEN = BLOCKS_PER_CHUNK * block_len;
+
esp_sha_lock_engine(sha_type);
SHA_CTX ctx;
ets_sha_init(&ctx);
- while(ilen > 0) {
- size_t chunk_len = (ilen > block_len) ? block_len : ilen;
+ while (ilen > 0) {
+ size_t chunk_len = (ilen > MAX_CHUNK_LEN) ? MAX_CHUNK_LEN : ilen;
// Wait for idle before entering critical section
- // (to reduce time spent in it), then check after
+ // (to reduce time spent in it), then check again after
esp_sha_wait_idle();
esp_sha_lock_memory_block();
esp_sha_wait_idle();
DPORT_STALL_OTHER_CPU_START();
- {
+ while (chunk_len > 0) {
// This SHA ROM function reads DPORT regs
- ets_sha_update(&ctx, sha_type, input, chunk_len * 8);
+ // (can accept max one SHA block each call)
+ size_t update_len = (chunk_len > block_len) ? block_len : chunk_len;
+ ets_sha_update(&ctx, sha_type, input, update_len * 8);
+
+ input += update_len;
+ chunk_len -= update_len;
+ ilen -= update_len;
}
DPORT_STALL_OTHER_CPU_END();
- input += chunk_len;
- ilen -= chunk_len;
-
if (ilen == 0) {
/* Finish the last block before releasing the memory
block lock, as ets_sha_update() may have written data to
#include <string.h>
#include "esp_types.h"
#include "esp_clk.h"
+#include "esp_log.h"
+#include "esp_timer.h"
+#include "esp_heap_caps.h"
+#include "idf_performance.h"
#include "unity.h"
#include "test_utils.h"
are tested as part of mbedTLS tests. Only esp_sha() is different.
*/
-TEST_CASE("Test esp_sha() function", "[hw_crypto]")
+#define TAG "sha_test"
+
+TEST_CASE("Test esp_sha()", "[hw_crypto]")
+{
+ const size_t BUFFER_SZ = 32 * 1024 + 6; // NB: not an exact multiple of SHA block size
+
+ int64_t begin, end;
+ uint32_t us_sha1, us_sha512;
+ uint8_t sha1_result[20] = { 0 };
+ uint8_t sha512_result[64] = { 0 };
+ void *buffer = heap_caps_malloc(BUFFER_SZ, MALLOC_CAP_8BIT|MALLOC_CAP_INTERNAL);
+ TEST_ASSERT_NOT_NULL(buffer);
+ memset(buffer, 0xEE, BUFFER_SZ);
+
+ const uint8_t sha1_expected[20] = { 0xc7, 0xbb, 0xd3, 0x74, 0xf2, 0xf6, 0x20, 0x86,
+ 0x61, 0xf4, 0x50, 0xd5, 0xf5, 0x18, 0x44, 0xcc,
+ 0x7a, 0xb7, 0xa5, 0x4a };
+ const uint8_t sha512_expected[64] = { 0xc7, 0x7f, 0xda, 0x8c, 0xb3, 0x58, 0x14, 0x8a,
+ 0x52, 0x3b, 0x46, 0x04, 0xc0, 0x85, 0xc5, 0xf0,
+ 0x46, 0x64, 0x14, 0xd5, 0x96, 0x7a, 0xa2, 0x80,
+ 0x20, 0x9c, 0x04, 0x27, 0x7d, 0x3b, 0xf9, 0x1f,
+ 0xb2, 0xa3, 0x45, 0x3c, 0xa1, 0x6a, 0x8d, 0xdd,
+ 0x35, 0x5e, 0x35, 0x57, 0x76, 0x22, 0x74, 0xd8,
+ 0x1e, 0x07, 0xc6, 0xa2, 0x9e, 0x3b, 0x65, 0x75,
+ 0x80, 0x7d, 0xe6, 0x6e, 0x47, 0x61, 0x2c, 0x94 };
+
+ begin = esp_timer_get_time();
+ esp_sha(SHA1, buffer, BUFFER_SZ, sha1_result);
+ end = esp_timer_get_time();
+ TEST_ASSERT_EQUAL_HEX8_ARRAY(sha1_expected, sha1_result, sizeof(sha1_expected));
+ us_sha1 = end - begin;
+ ESP_LOGI(TAG, "esp_sha() 32KB SHA1 in %u us", us_sha1);
+
+ begin = esp_timer_get_time();
+ esp_sha(SHA2_512, buffer, BUFFER_SZ, sha512_result);
+ end = esp_timer_get_time();
+ TEST_ASSERT_EQUAL_HEX8_ARRAY(sha512_expected, sha512_result, sizeof(sha512_expected));
+
+ us_sha512 = end - begin;
+ ESP_LOGI(TAG, "esp_sha() 32KB SHA512 in %u us", us_sha512);
+
+ free(buffer);
+
+ TEST_PERFORMANCE_LESS_THAN(ESP32_TIME_SHA1_32KB, "%dus", us_sha1);
+ TEST_PERFORMANCE_LESS_THAN(ESP32_TIME_SHA512_32KB, "%dus", us_sha512);
+}
+
+TEST_CASE("Test esp_sha() function with long input", "[hw_crypto]")
{
const void* ptr;
spi_flash_mmap_handle_t handle;