[src/trunk]: src/sys/kern This had a similar problem to that reported in PR l...

[kre <kre%NetBSD.org@localhost>]

details:   https://anonhg.NetBSD.org/src/rev/600ddfad63f7
branches:  trunk
changeset: 997542:600ddfad63f7
user:      kre <kre%NetBSD.org@localhost>
date:      Tue Mar 12 00:25:44 2019 +0000

description:
This had a similar problem to that reported in PR lib/54053
for the userland (libc) version of humanize_number(),
except in a much more limited, and less harmful, way ...

If the value to be printed was > ~0/10 (eg ~0 itself) and
the buffer given was big enough to hold it, "umax" would overflow,
and the expected result was not produced (here at least the effect was
simply to scale the result for huge values by one unnecessary step,
rather than maximally scale all values usually generating 0E, but
incorrect nonetheless.)

diffstat:

 sys/kern/subr_humanize.c |  21 +++++++++++++++------
 1 files changed, 15 insertions(+), 6 deletions(-)

diffs (54 lines):

diff -r 83c943bccc3c -r 600ddfad63f7 sys/kern/subr_humanize.c
--- a/sys/kern/subr_humanize.c  Mon Mar 11 20:38:27 2019 +0000
+++ b/sys/kern/subr_humanize.c  Tue Mar 12 00:25:44 2019 +0000
@@ -1,4 +1,4 @@
-/*     $NetBSD: subr_humanize.c,v 1.1 2009/10/02 15:48:41 pooka Exp $  */
+/*     $NetBSD: subr_humanize.c,v 1.2 2019/03/12 00:25:44 kre Exp $    */
 
 /*-
  * Copyright (c) 1997, 1998, 1999, 2002, 2007, 2008 The NetBSD Foundation, Inc.
@@ -30,7 +30,7 @@
  */
 
 #include <sys/cdefs.h>
-__KERNEL_RCSID(0, "$NetBSD: subr_humanize.c,v 1.1 2009/10/02 15:48:41 pooka Exp $");
+__KERNEL_RCSID(0, "$NetBSD: subr_humanize.c,v 1.2 2019/03/12 00:25:44 kre Exp $");
 
 #include <sys/types.h>
 #include <sys/systm.h>
@@ -55,7 +55,7 @@
                /* prefixes are: (none), kilo, Mega, Giga, Tera, Peta, Exa */
        const char *prefixes;
        int             r;
-       uint64_t        umax;
+       uint64_t        umax, b10;
        size_t          i, suffixlen;
 
        if (buf == NULL || suffix == NULL)
@@ -77,14 +77,23 @@
                prefixes = " kMGTPE"; /* SI for decimal multiplies */
 
        umax = 1;
+       b10 = bytes/10;
        for (i = 0; i < len - suffixlen - 3; i++) {
+               if (umax > b10) {
+                       /*
+                        * there is space for the unscaled number
+                        * but bytes might be ~0 - there is no bigger
+                        * value available for umax, so we must skip
+                        * the "bytes >= umax" test below
+                        */
+                       i = 0;
+                       goto nodiv;
+               }
                umax *= 10;
-               if (umax > bytes)
-                       break;
        }
        for (i = 0; bytes >= umax && prefixes[i + 1]; i++)
                bytes /= divisor;
-
+ nodiv:
        r = snprintf(buf, len, "%qu%s%c%s", (unsigned long long)bytes,
            i == 0 ? "" : " ", prefixes[i], suffix);