zephyr: Fix Zephyr heap metadata corruption

Zephyr stores heap metadata just before each allocated chunk. So not only
must the chunk start be aligned to cache line size, but the chunk size
must also be aligned up to cache line size. This is to prevent metadata
corruption in case an invalidate or writeback is called for a chunk -- it
may corrupt metadata for the next chunk if it is located in the same cache
line.

Signed-off-by: Serhiy Katsyuba <serhiy.katsyuba@intel.com>

Author: serhiy-katsyuba-intel

zephyr/lib/alloc.c

@@ -471,10 +471,16 @@ void *rmalloc_align(uint32_t flags, size_t bytes, uint32_t alignment)
 		ptr = (__sparse_force void *)heap_alloc_aligned_cached(heap, alignment, bytes);
 	} else {
 		/*
-		 * XTOS alloc implementation has used dcache alignment,
-		 * so SOF application code is expecting this behaviour.
+		 * Zephyr sys_heap stores metadata at start of each
+		 * heap allocation. To ensure no allocated cached buffer
+		 * overlaps the same cacheline with the metadata chunk,
+		 * align both allocation start and size of allocation
+		 * to cacheline. As cached and non-cached allocations are
+		 * mixed, same rules need to be followed for both type of
+		 * allocations.
 		 */
 		alignment = MAX(PLATFORM_DCACHE_ALIGN, alignment);
+		bytes = ALIGN_UP(bytes, alignment);
 		ptr = heap_alloc_aligned(heap, alignment, bytes);
 	}