linux-kernel - Re: [PATCH] kernel: kallsyms: memory override issue, need check destination buffer length

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

Date:	Thu, 11 Apr 2013 13:38:15 +0930
From:	Rusty Russell <rusty@...tcorp.com.au>
To:	Chen Gang <gang.chen@...anux.com>,
	Stephen Boyd <sboyd@...eaurora.org>
Cc:	Andrew Morton <akpm@...ux-foundation.org>,
	"linux-kernel\@vger.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] kernel: kallsyms: memory override issue, need check destination buffer length

Chen Gang <gang.chen@...anux.com> writes:
>   We don't export any symbols > 128 characters, but if we did then
>   kallsyms_expand_symbol() would overflow the buffer handed to it.
>   So we need check destination buffer length when copying.
>
>   the related test:
>     if we define an EXPORT function which name more than 128.
>     will panic when call kallsyms_lookup_name by init_kprobes on booting.
>     after check the length (provide this patch), it is ok.
>
>   Implementaion:
>     add additional destination buffer length parameter (maxlen)
>     if uncompressed string is too long (>= maxlen), it will be truncated.
>     not check the parameters whether valid, since it is a static function.
>
>
> Signed-off-by: Chen Gang <gang.chen@...anux.com>
> Signed-off-by: Rusty Russell <rusty@...tcorp.com.au>

??!!!  I never signed this off!  I've never seen it before!

Minor comments below:

> -static unsigned int kallsyms_expand_symbol(unsigned int off, char *result)
> +static unsigned int kallsyms_expand_symbol(unsigned int off,
> +					   char *result, int maxlen)

'size_t maxlen' would be more explicit.

>  {
>  	int len, skipped_first = 0;
>  	const u8 *tptr, *data;
> +	char *begin = result;
>   	/* Get the compressed symbol length from the first symbol byte. */
>  	data = &kallsyms_names[off];
> @@ -113,14 +116,16 @@ static unsigned int kallsyms_expand_symbol(unsigned int off, char *result)
>   		while (*tptr) {
>  			if (skipped_first) {
> -				*result = *tptr;
> -				result++;
> +				*result++ = *tptr;
> +				if (result - begin == maxlen - 1)
> +					goto tail;

You could just decrement maxlen instead, and handle maxlen == 0 at the
same time:

                if (maxlen <= 1)
                        goto tail;
                *result = *tptr;
                result++;
                maxlen--;

> @@ -176,7 +181,7 @@ unsigned long kallsyms_lookup_name(const char *name)
>  	unsigned int off;
>   	for (i = 0, off = 0; i < kallsyms_num_syms; i++) {
> -		off = kallsyms_expand_symbol(off, namebuf);
> +		off = kallsyms_expand_symbol(off, namebuf, sizeof(namebuf));
>   		if (strcmp(namebuf, name) == 0)
>  			return kallsyms_addresses[i];

I prefer to use ARRAY_SIZE(namebuf) instead of sizeof(namebuf).  That
way we break compile if the declaration is changed from an array to a
pointer one day.

Same for the others.

Thanks,
Rusty.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/