In io_write_buf() adjust the calculation of the multiplex message size

so that the loop works even for buffers bigger than 0xffffff. The code
does not produce such big buffers but better make the code correct.
Reported by Blago Dachev (blago (at) dachev (dot) com)
OK benno@

diff --git a/usr.bin/rsync/io.c b/usr.bin/rsync/io.c
index ea35bdb..87019a9 100644 (file)
--- a/usr.bin/rsync/io.c
+++ b/usr.bin/rsync/io.c
@@ -1,4 +1,4 @@
-/*     $OpenBSD: io.c,v 1.20 2021/06/30 13:10:04 claudio Exp $ */
+/*     $OpenBSD: io.c,v 1.21 2021/12/28 11:59:48 claudio Exp $ */
 /*
  * Copyright (c) 2019 Kristaps Dzonsons <kristaps@bsd.lv>
  *
@@ -143,7 +143,7 @@ io_write_buf(struct sess *sess, int fd, const void *buf, size_t sz)
        }
 
        while (sz > 0) {
-               wsz = sz & 0xFFFFFF;
+               wsz = (sz < 0xFFFFFF) ? sz : 0xFFFFFF;
                tag = (7 << 24) + wsz;
                tagbuf = htole32(tag);
                if (!io_write_blocking(fd, &tagbuf, sizeof(tagbuf))) {