libidn/0002-Check-codepoint-validity-in-punycode_decode-and-puny.patch
2019-09-30 10:56:38 -04:00

40 lines
1.3 KiB
Diff

From bd4d04075fa126552b31cd11aaa50dad72119e6a Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Tim=20R=C3=BChsen?= <tim.ruehsen@gmx.de>
Date: Fri, 6 Jul 2018 13:05:56 +0200
Subject: [PATCH 2/3] Check codepoint validity in punycode_decode() and
punycode_decode()
These functions were able to generate invalid unicode values resp.
invalid punycode. This is undocumented/unexpected behavior that can
lead to security vulns.
Reported-by: Mike Schiffman (Farsight Security, Inc.)
---
lib/punycode.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/lib/punycode.c b/lib/punycode.c
index d475b6d..f7c63e6 100644
--- a/lib/punycode.c
+++ b/lib/punycode.c
@@ -228,6 +228,8 @@ punycode_encode (size_t input_length,
output[out++] = case_flags ?
encode_basic (input[j], case_flags[j]) : (char) input[j];
}
+ else if (input[j] > 0x10FFFF)
+ return punycode_bad_input;
/* else if (input[j] < n) return punycode_bad_input; */
/* (not needed for Punycode with unsigned code points) */
}
@@ -418,6 +420,8 @@ punycode_decode (size_t input_length,
if (i / (out + 1) > maxint - n)
return punycode_overflow;
n += i / (out + 1);
+ if (n > 0x10FFFF)
+ return punycode_bad_input;
i %= (out + 1);
/* Insert n at position i of the output: */
--
1.8.3.1