From 5562545b5562f6d12a4ef991fae158bf4ccf92b6 Mon Sep 17 00:00:00 2001 From: Joerg Sonnenberger Date: Sat, 9 Sep 2017 17:47:32 +0200 Subject: [PATCH] Avoid a read off-by-one error for UTF16 names in RAR archives. Reported-By: OSS-Fuzz issue 573 CVE: CVE-2017-14502 Upstream-Status: Backport Signed-off-by: Zhixiong Chi --- libarchive/archive_read_support_format_rar.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/libarchive/archive_read_support_format_rar.c b/libarchive/archive_read_support_format_rar.c index cbb14c3..751de69 100644 --- a/libarchive/archive_read_support_format_rar.c +++ b/libarchive/archive_read_support_format_rar.c @@ -1496,7 +1496,11 @@ read_header(struct archive_read *a, struct archive_entry *entry, return (ARCHIVE_FATAL); } filename[filename_size++] = '\0'; - filename[filename_size++] = '\0'; + /* + * Do not increment filename_size here as the computations below + * add the space for the terminating NUL explicitly. + */ + filename[filename_size] = '\0'; /* Decoded unicode form is UTF-16BE, so we have to update a string * conversion object for it. */ -- 1.9.1