Backport upstream CryptoNight slow-hash width/index fixes to resolve cn_lite_slow_hash_v1 mismatch

Author: wrkzdev

include/hash-ops.h

@@ -77,9 +77,9 @@ void cn_slow_hash(
     int light,
     int variant,
     int prehashed,
-    uint64_t page_size,
-    uint64_t scratchpad,
-    uint64_t iterations,
+    uint32_t page_size,
+    uint32_t scratchpad,
+    uint32_t iterations,
     uint64_t mask);
 
 void hash_extra_blake(const void *data, size_t length, char *hash);

src/slow-hash-arm.c

@@ -257,14 +257,14 @@ void cn_slow_hash(
     int light,
     int variant,
     int prehashed,
-    uint64_t page_size,
-    uint64_t scratchpad,
-    uint64_t iterations,
+    uint32_t page_size,
+    uint32_t scratchpad,
+    uint32_t iterations,
     uint64_t mask)
 {
-    uint64_t TOTALBLOCKS = (page_size / AES_BLOCK_SIZE);
-    uint64_t init_rounds = (scratchpad / INIT_SIZE_BYTE);
-    uint64_t aes_rounds = (iterations / 2);
+    uint32_t TOTALBLOCKS = (page_size / AES_BLOCK_SIZE);
+    uint32_t init_rounds = (scratchpad / INIT_SIZE_BYTE);
+    uint32_t aes_rounds = (iterations / 2);
 
     RDATA_ALIGN16 uint8_t expandedKey[240];
 
@@ -500,13 +500,13 @@ void cn_slow_hash(
     int light,
     int variant,
     int prehashed,
-    uint64_t page_size,
-    uint64_t scratchpad,
-    uint64_t iterations,
+    uint32_t page_size,
+    uint32_t scratchpad,
+    uint32_t iterations,
     uint64_t mask)
 {
-    uint64_t init_rounds = (scratchpad / INIT_SIZE_BYTE);
-    uint64_t aes_rounds = (iterations / 2);
+    uint32_t init_rounds = (scratchpad / INIT_SIZE_BYTE);
+    uint32_t aes_rounds = (iterations / 2);
 
     uint8_t text[INIT_SIZE_BYTE];
     uint8_t a[AES_BLOCK_SIZE];

src/slow-hash-portable.c

@@ -115,13 +115,13 @@ void cn_slow_hash(
     int light,
     int variant,
     int prehashed,
-    uint64_t page_size,
-    uint64_t scratchpad,
-    uint64_t iterations,
+    uint32_t page_size,
+    uint32_t scratchpad,
+    uint32_t iterations,
     uint64_t mask)
 {
-    uint64_t init_rounds = (scratchpad / INIT_SIZE_BYTE);
-    uint64_t aes_rounds = (iterations / 2);
+    uint32_t init_rounds = (scratchpad / INIT_SIZE_BYTE);
+    uint32_t aes_rounds = (iterations / 2);
 
     uint8_t text[INIT_SIZE_BYTE];
     uint8_t a[AES_BLOCK_SIZE];
@@ -182,7 +182,7 @@ void cn_slow_hash(
     for (i = 0; i < aes_rounds; i++)
     {
         // Iteration 1
-        j = U64(a)[0] & mask;
+        j = (*(uint32_t *)a) & mask;
         p = &long_state[j];
         aesb_single_round(p, p, a);
         copy_block(c1, p);
@@ -192,7 +192,7 @@ void cn_slow_hash(
         VARIANT1_1(p);
 
         // Iteration 2
-        j = U64(c1)[0] & mask;
+        j = (*(uint32_t *)c1) & mask;
         p = &long_state[j];
         copy_block(c, p);
 

src/slow-hash-x86.c

@@ -408,7 +408,7 @@ BOOL SetLockPagesPrivilege(HANDLE hProcess, BOOL bEnable)
  * the allocated buffer.
  */
 
-void slow_hash_allocate_state(uint64_t page_size)
+void slow_hash_allocate_state(uint32_t page_size)
 {
     if (hp_state != NULL)
     {
@@ -444,7 +444,7 @@ void slow_hash_allocate_state(uint64_t page_size)
  *@brief frees the state allocated by slow_hash_allocate_state
  */
 
-void slow_hash_free_state(uint64_t page_size)
+void slow_hash_free_state(uint32_t page_size)
 {
     if (hp_state == NULL)
     {
@@ -505,14 +505,14 @@ void cn_slow_hash(
     int light,
     int variant,
     int prehashed,
-    uint64_t page_size,
-    uint64_t scratchpad,
-    uint64_t iterations,
+    uint32_t page_size,
+    uint32_t scratchpad,
+    uint32_t iterations,
     uint64_t mask)
 {
-    uint64_t TOTALBLOCKS = (page_size / AES_BLOCK_SIZE);
-    uint64_t init_rounds = (scratchpad / INIT_SIZE_BYTE);
-    uint64_t aes_rounds = (iterations / 2);
+    uint32_t TOTALBLOCKS = (page_size / AES_BLOCK_SIZE);
+    uint32_t init_rounds = (scratchpad / INIT_SIZE_BYTE);
+    uint32_t aes_rounds = (iterations / 2);
 
     RDATA_ALIGN16 uint8_t expandedKey[240]; /* These buffers are aligned to use later with SSE functions */