Fix unaligned data accesses on ARM

Some unaligned data access patterns are not entirely OK on Linux/ARM.
Depending on kernel configuration, it may cause bus errors or "only"
lead to massive slowdowns as the kernel works around faulting
instructions and perhaps even emits log messages. (This option does
not seem to exist for 32 bit code running on an ARM64 kernel, though.)

Most of this patch which was originally developed by Steve Langasek
has been part of the Debian package for a while.

libyara/arena.c

@@ -435,6 +435,12 @@ void* yr_arena_ref_to_ptr(
   if (YR_ARENA_IS_NULL_REF(*ref))
     return NULL;
 
+#if defined(__arm__)
+  YR_ARENA_REF tmp_ref;
+  memcpy(&tmp_ref, ref, sizeof(YR_ARENA_REF));
+  ref = &tmp_ref;
+#endif
+
   return yr_arena_get_ptr(arena, ref->buffer_id, ref->offset);
 }
 

libyara/exec.c

@@ -540,7 +540,7 @@ int yr_execute_code(
         break;
 
       case OP_PUSH:
-        r1.i = *(uint64_t*)(ip);
+        memcpy(&r1.i, ip, sizeof(uint64_t));
         ip += sizeof(uint64_t);
         push(r1);
         break;
@@ -550,7 +550,7 @@ int yr_execute_code(
         break;
 
       case OP_CLEAR_M:
-        r1.i = *(uint64_t*)(ip);
+        memcpy(&r1.i, ip, sizeof(uint64_t));
         ip += sizeof(uint64_t);
         #if YR_PARANOID_EXEC
         ensure_within_mem(r1.i);
@@ -559,7 +559,7 @@ int yr_execute_code(
         break;
 
       case OP_ADD_M:
-        r1.i = *(uint64_t*)(ip);
+        memcpy(&r1.i, ip, sizeof(uint64_t));
         ip += sizeof(uint64_t);
         #if YR_PARANOID_EXEC
         ensure_within_mem(r1.i);
@@ -570,7 +570,7 @@ int yr_execute_code(
         break;
 
       case OP_INCR_M:
-        r1.i = *(uint64_t*)(ip);
+        memcpy(&r1.i, ip, sizeof(uint64_t));
         ip += sizeof(uint64_t);
         #if YR_PARANOID_EXEC
         ensure_within_mem(r1.i);
@@ -579,7 +579,7 @@ int yr_execute_code(
         break;
 
       case OP_PUSH_M:
-        r1.i = *(uint64_t*)(ip);
+        memcpy(&r1.i, ip, sizeof(uint64_t));
         ip += sizeof(uint64_t);
         #if YR_PARANOID_EXEC
         ensure_within_mem(r1.i);
@@ -589,7 +589,7 @@ int yr_execute_code(
         break;
 
       case OP_POP_M:
-        r1.i = *(uint64_t*)(ip);
+        memcpy(&r1.i, ip, sizeof(uint64_t));
         ip += sizeof(uint64_t);
         #if YR_PARANOID_EXEC
         ensure_within_mem(r1.i);
@@ -599,7 +599,7 @@ int yr_execute_code(
         break;
 
       case OP_SET_M:
-        r1.i = *(uint64_t*)(ip);
+        memcpy(&r1.i, ip, sizeof(uint64_t));
         ip += sizeof(uint64_t);
         #if YR_PARANOID_EXEC
         ensure_within_mem(r1.i);
@@ -611,7 +611,7 @@ int yr_execute_code(
         break;
 
       case OP_SWAPUNDEF:
-        r1.i = *(uint64_t*)(ip);
+        memcpy(&r1.i, ip, sizeof(uint64_t));
         ip += sizeof(uint64_t);
         #if YR_PARANOID_EXEC
         ensure_within_mem(r1.i);
@@ -802,7 +802,7 @@ int yr_execute_code(
         break;
 
       case OP_PUSH_RULE:
-        r1.i = *(uint64_t*)(ip);
+        memcpy(&r1.i, ip, sizeof(uint64_t));
         ip += sizeof(uint64_t);
 
         rule = &context->rules->rules_list_head[r1.i];
@@ -841,7 +841,7 @@ int yr_execute_code(
       case OP_MATCH_RULE:
         pop(r1);
 
-        r2.i = *(uint64_t*)(ip);
+        memcpy(&r2.i, ip, sizeof(uint64_t));
         ip += sizeof(uint64_t);
 
         rule = &context->rules->rules_list_head[r2.i];
@@ -1316,7 +1316,7 @@ int yr_execute_code(
         break;
 
       case OP_IMPORT:
-        r1.i = *(uint64_t*)(ip);
+        memcpy(&r1.i, ip, sizeof(uint64_t));
         ip += sizeof(uint64_t);
 
         result = yr_modules_load((char*) r1.p, context);
@@ -1356,8 +1356,7 @@ int yr_execute_code(
         break;
 
       case OP_INT_TO_DBL:
-
-        r1.i = *(uint64_t*)(ip);
+        memcpy(&r1.i, ip, sizeof(uint64_t));
         ip += sizeof(uint64_t);
 
         #if YR_PARANOID_EXEC

libyara/notebook.c

@@ -146,6 +146,15 @@ void* yr_notebook_alloc(
 
   void *ptr = notebook->page_list_head->data + notebook->page_list_head->used;
 
+#if defined(__arm__)
+  uintptr_t misalignment = (uintptr_t)ptr & 3;
+  if (misalignment)
+  {
+    size += 4-misalignment;
+    ptr += 4-misalignment;
+  }
+#endif
+
   notebook->page_list_head->used += size;
 
   return ptr;