ipc: Fix incorrect dcache invalidate

There are few problems with the invalidate call here:

struct sof_ipc_cmd_hdr belongs to IPC3. There is no 'size' field in IPC4
payload at that offset. comp_data is zeroed on init and on IPC4 it is only
used to send responses by get_large_config. Probably get_large_config is
rarely used as, otherwise, that would result in invalidates of dcache with
weird sizes and problems at runtime.

I do not see any point to invalidate dcache here. It's definitely not
needed for IPC4. Not an expert in IPC3, however, it seems IPC3 does not
need such code either.

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

Author: serhiy-katsyuba-intel

src/ipc/ipc-common.c

@@ -55,10 +55,6 @@ int ipc_process_on_core(uint32_t core, bool blocking)
 		return -EACCES;
 	}
 
-	/* The other core will write there its response */
-	dcache_invalidate_region((__sparse_force void __sparse_cache *)MAILBOX_HOSTBOX_BASE,
-				 ((struct sof_ipc_cmd_hdr *)ipc->comp_data)->size);
-
 	/*
 	 * If the primary core is waiting for secondary cores to complete, it
 	 * will also reply to the host