linux-hardening - RE: [PATCH] crypto/caam: Avoid GCC constprop bug warning

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:   Sat, 29 Oct 2022 11:40:55 +0000
From:   David Laight <David.Laight@...LAB.COM>
To:     'Kees Cook' <keescook@...omium.org>,
        Horia Geantă <horia.geanta@....com>
CC:     Pankaj Gupta <pankaj.gupta@....com>,
        Gaurav Jain <gaurav.jain@....com>,
        Herbert Xu <herbert@...dor.apana.org.au>,
        "David S. Miller" <davem@...emloft.net>,
        "linux-crypto@...r.kernel.org" <linux-crypto@...r.kernel.org>,
        kernel test robot <lkp@...el.com>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "linux-hardening@...r.kernel.org" <linux-hardening@...r.kernel.org>
Subject: RE: [PATCH] crypto/caam: Avoid GCC constprop bug warning

From: Kees Cook
> Sent: 28 October 2022 22:06
> 
> GCC 12 appears to perform constant propagation incompletely(?) and can
> no longer notice that "len" is always 0 when "data" is NULL. Expand the
> check to avoid warnings about memcpy() having a NULL argument:
...
> 
> diff --git a/drivers/crypto/caam/desc_constr.h b/drivers/crypto/caam/desc_constr.h
> index 62ce6421bb3f..ddbba8b00ab7 100644
> --- a/drivers/crypto/caam/desc_constr.h
> +++ b/drivers/crypto/caam/desc_constr.h
> @@ -163,7 +163,7 @@ static inline void append_data(u32 * const desc, const void *data, int len)
>  {
>  	u32 *offset = desc_end(desc);
> 
> -	if (len) /* avoid sparse warning: memcpy with byte count of 0 */
> +	if (data && len) /* avoid sparse warning: memcpy with byte count of 0 */
>  		memcpy(offset, data, len);

I'd guess non-constant zero lengths are unlikely?
So how about:
	/* Avoid calling memcpy() when there is never a buffer */
	if (!__builtin_constant(len) || len)
		memcpy(offset, data, len);

Then the test should never actually end up in the object code.

	David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)