From: Florian Westphal
> Sent: 08 January 2016 16:24
> To: David Laight
> David Laight <David.Laight@xxxxxxxxxx> wrote:
> > From: Pablo Neira Ayuso
> > > Sent: 08 January 2016 14:02
> > > From: Florian Westphal <fw@xxxxxxxxx>
> > >
> > > Needed to convert the (64bit) conntrack counters to BE ordering.
> > >
> > ...
> > > switch (priv->size) {
> > > + case 8: {
> > > + u64 src64;
> > > +
> > > + switch (priv->op) {
> > > + case NFT_BYTEORDER_NTOH:
> > > + for (i = 0; i < priv->len / 8; i++) {
> > > + src64 = get_unaligned_be64(&src[i]);
> > > + src64 = be64_to_cpu((__force __be64)src64);
> > > + put_unaligned_be64(src64, &dst[i]);
> > > + }
> > > + break;
> > > + case NFT_BYTEORDER_HTON:
> > > + for (i = 0; i < priv->len / 8; i++) {
> > > + src64 = get_unaligned_be64(&src[i]);
> > > + src64 = (__force u64)cpu_to_be64(src64);
> > > + put_unaligned_be64(src64, &dst[i]);
> > > + }
> > > + break;
> > > + }
> > > + break;
> >
> > That is horrid.
>
> Yes, sorry for this, however ...
>
> > On a little-endian system you are byteswapping the data 3 times.
> > Image the code on a cpu that doesn't support misaligned transfers
> > and doesn't have a byteswap instruction.
>
> diff --git a/net/netfilter/nft_byteorder.c b/net/netfilter/nft_byteorder.c
> --- a/net/netfilter/nft_byteorder.c
> +++ b/net/netfilter/nft_byteorder.c
> @@ -46,16 +46,16 @@ static void nft_byteorder_eval(const struct nft_expr *expr,
> switch (priv->op) {
> case NFT_BYTEORDER_NTOH:
> for (i = 0; i < priv->len / 8; i++) {
> - src64 = get_unaligned_be64(&src[i]);
> + src64 = get_unaligned((u64 *)&src[i]);
> src64 = be64_to_cpu((__force __be64)src64);
> - put_unaligned_be64(src64, &dst[i]);
> + put_unaligned(src64, (u64 *)&dst[i]);
Why not just:
src64 = get_unaligned_be64(&src[i]);
put_unaligned(src64, (u64 *)&dst[i]);
...
> Results in identical object code.
How many architectures have you looked at?
Try sparc64, ppc64 and arm64, especially le variants.
(le sparc would be bad - but doesn't exist).
Also older compilers that don't have builtin support for misaligned
memory accesses.
Actually there is (probably) a generic problem with:
put_unaligned(src64, (u64 *)&dst[i]);
The C language rules for casts effectively forbid pointers to
misaligned structures (the compiler is allowed to assume they
are aligned however hard you try to persuade it otherwise).
I suspect that means that the (u64 *) cast is allowed to mask
off the bottom address bits.
David
--
To unsubscribe from this list: send the line "unsubscribe netfilter-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
