move cluster table management from ec3_reader/writer to ec3_pipeline

src/zstd.c

@@ -7,24 +7,28 @@ static enum ec3_status compress(
 	const void *src,
 	size_t len,
 	void *dest,
+	size_t dest_max,
 	size_t *nr_written)
 {
-	*nr_written = ZSTD_compress(dest, len, src, len, 10);
+	*nr_written = ZSTD_compress(dest, dest_max, src, len, 10);
 	return EC3_SUCCESS;
 }
 
 static enum ec3_status decompress(
 	struct ec3_pipeline_stage *stage,
 	const void *src,
+	size_t len,
 	void *dest,
+	size_t dest_max,
 	size_t *nr_read)
 {
+	*nr_read = ZSTD_decompress(dest, dest_max, src, len);
 	return EC3_SUCCESS;
 }
 
 const struct ec3_pipeline_stage_type pipeline_zstd = {
 	.t_id = EC3_PIPELINE_ZSTD,
 	.t_flags = EC3_PIPELINE_F_BUFFERED,
-	.t_data_in = decompress,
-	.t_data_out = compress,
+	.t_cluster_in = decompress,
+	.t_cluster_out = compress,
 };