kernel - clarify code documentation in vfs_cluster

* Clarify the operation of minreq/maxreq a little better in
  cluster_read().

diff --git a/sys/kern/vfs_cluster.c b/sys/kern/vfs_cluster.c
index 41eda14..7eff091 100644 (file)
--- a/sys/kern/vfs_cluster.c
+++ b/sys/kern/vfs_cluster.c
@@ -122,13 +122,16 @@ cluster_read(struct vnode *vp, off_t filesize, off_t loffset,
         * Then we limit maxreq to max_readahead to ensure it is a reasonable
         * value.
         *
-        * Finally we must ensure that loffset + maxreq does not cross the
+        * Finally we must ensure that (loffset + maxreq) does not cross the
         * boundary (filesize) for the current blocksize.  If we allowed it
         * to cross we could end up with buffers past the boundary with the
         * wrong block size (HAMMER large-data areas use mixed block sizes).
+        * minreq is also absolutely limited to filesize.
         */
        if (maxreq < minreq)
                maxreq = minreq;
+       /* minreq not used beyond this point */
+
        if (maxreq > max_readahead) {
                maxreq = max_readahead;
                if (maxreq > 16 * 1024 * 1024)