linux-kernel - Re: namei_vfat.c array subscript is above array bounds

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Sat, 08 Jul 2017 16:42:25 +0900
From:   OGAWA Hirofumi <hirofumi@...l.parknet.co.jp>
To:     Meelis Roos <mroos@...ux.ee>
Cc:     Linux Kernel list <linux-kernel@...r.kernel.org>
Subject: Re: namei_vfat.c array subscript is above array bounds

Meelis Roos <mroos@...ux.ee> writes:

>> Source is same with above?
>
> Yes, same checksums.

> CONFIG_X86_USE_3DNOW=y
> CONFIG_UBSAN=y
> CONFIG_UBSAN_SANITIZE_ALL=y

OK. Succeed to reproduce. gcc's ubsan (CONFIG_UBSAN_SANITIZE_ALL=y) with
CONFIG_X86_USE_3DNOW=y outputs warnings of that.

But warnings are strange. The target of source is,

vfat_create_shortname:

	extlen = 0;
	if (ext_start) {
		for (p = ext, ip = ext_start; extlen < 3 && ip < end; ip++) {
			chl = to_shortname_char(nls, charbuf, sizeof(charbuf),
						ip, &ext_info);
			if (chl == 0)
				continue;

			if ((extlen + chl) > 3) {
				is_shortname = 0;
				break;
			}
			for (chi = 0; chi < chl; chi++) {
				*p++ = charbuf[chi];
				extlen++;
			}
			if (extlen >= 3) {
				if (ip + 1 != end)
					is_shortname = 0;
				break;
			}
		}
	}

[...]

	memcpy(name_res + 8, ext, extlen);    <= here

name_res == name_res[11], but extlen never be bigger than 3 (if I'm not
missing something). And extlen is not constant, but gcc outputs the
warnings on __constant_memcpy3d().

#define memcpy(t, f, n)				\
	(__builtin_constant_p((n))		\
	 ? __constant_memcpy3d((t), (f), (n))	\
	 : __memcpy3d((t), (f), (n)))

And changing

	memcpy(name_res + 8, ext, extlen);

to

        __builtin_memcpy(name_res + 8, ext, extlen);

doesn't output warnings.

So, I'm not sure though, it looks like the bug of ubsan(?). (BTW, for
now, you can set CONFIG_UBSAN_SANITIZE_ALL=n to disable ubsan.)

Thanks.
-- 
OGAWA Hirofumi <hirofumi@...l.parknet.co.jp>