linux-kernel - Re: [PATCH] crypto: xts: use memmove to avoid overlapped memory copy

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:   Fri, 17 Jul 2020 08:59:54 +0300
From:   Ard Biesheuvel <ardb@...nel.org>
To:     Herbert Xu <herbert@...dor.apana.org.au>
Cc:     Colin King <colin.king@...onical.com>,
        "David S . Miller" <davem@...emloft.net>,
        Linux Crypto Mailing List <linux-crypto@...r.kernel.org>,
        kernel-janitors@...r.kernel.org,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] crypto: xts: use memmove to avoid overlapped memory copy

On Fri, 17 Jul 2020 at 08:21, Herbert Xu <herbert@...dor.apana.org.au> wrote:
>
> On Thu, Jul 16, 2020 at 06:56:30PM +0300, Ard Biesheuvel wrote:
> > On Thu, 16 Jul 2020 at 18:29, Colin King <colin.king@...onical.com> wrote:
> > >
> > > From: Colin Ian King <colin.king@...onical.com>
> > >
> > > There is a memcpy that performs a potential overlapped memory copy
> > > from source b to destination b + 1.  Fix this by using the safer
> > > memmove instead.
> > >
> > > Addresses-Coverity: ("Overlapping buffer in memory copy")
> > > Fixes: 8083b1bf8163 ("crypto: xts - add support for ciphertext stealing")
> > > Signed-off-by: Colin Ian King <colin.king@...onical.com>
> > > ---
> > >  crypto/xts.c | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/crypto/xts.c b/crypto/xts.c
> > > index 3565f3b863a6..fa3e6e7b7043 100644
> > > --- a/crypto/xts.c
> > > +++ b/crypto/xts.c
> > > @@ -169,7 +169,7 @@ static int cts_final(struct skcipher_request *req,
> > >                                       offset - XTS_BLOCK_SIZE);
> > >
> > >         scatterwalk_map_and_copy(b, rctx->tail, 0, XTS_BLOCK_SIZE, 0);
> > > -       memcpy(b + 1, b, tail);
> > > +       memmove(b + 1, b, tail);
> >
> > This is a false positive: tail is guaranteed to be smaller than
> > sizeof(*b), so memmove() is unnecessary here.
> >
> > If changing to memcpy(&b[1], &b[0], tail) makes the warning go away, i
> > am fine with it, but otherwise we should just leave it as is.
>
> How about a comment perhaps?
>

Or change it to b[1] = b[0] (assuming the compiler allows struct
assignment in that way). This will always copy XTS_BLOCK_SIZE bytes,
but we have sufficient space, and it is probably more efficient  too
in most cases.


> Cheers,
> --
> Email: Herbert Xu <herbert@...dor.apana.org.au>
> Home Page: http://gondor.apana.org.au/~herbert/
> PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt