linux-kernel - Re: [PATCH RESEND bpf-next 10/15] ice: Implement VLAN tag hint

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]

Message-ID: <ZGI2oDcWX+o9Ea0T@lincoln>
Date:   Mon, 15 May 2023 15:41:52 +0200
From:   Larysa Zaremba <larysa.zaremba@...el.com>
To:     Stanislav Fomichev <sdf@...gle.com>
CC:     <bpf@...r.kernel.org>, Alexei Starovoitov <ast@...nel.org>,
        "Daniel Borkmann" <daniel@...earbox.net>,
        Andrii Nakryiko <andrii@...nel.org>,
        "Jakub Kicinski" <kuba@...nel.org>,
        Martin KaFai Lau <martin.lau@...ux.dev>,
        Song Liu <song@...nel.org>, Yonghong Song <yhs@...com>,
        John Fastabend <john.fastabend@...il.com>,
        KP Singh <kpsingh@...nel.org>, Jiri Olsa <jolsa@...nel.org>,
        Jesse Brandeburg <jesse.brandeburg@...el.com>,
        "Tony Nguyen" <anthony.l.nguyen@...el.com>,
        Anatoly Burakov <anatoly.burakov@...el.com>,
        Jesper Dangaard Brouer <brouer@...hat.com>,
        Alexander Lobakin <alexandr.lobakin@...el.com>,
        Magnus Karlsson <magnus.karlsson@...il.com>,
        Maryam Tahhan <mtahhan@...hat.com>,
        <xdp-hints@...-project.net>, <netdev@...r.kernel.org>,
        <intel-wired-lan@...ts.osuosl.org>, <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH RESEND bpf-next 10/15] ice: Implement VLAN tag hint

On Fri, May 12, 2023 at 11:31:21AM -0700, Stanislav Fomichev wrote:
> On 05/12, Larysa Zaremba wrote:
> > Implement .xmo_rx_vlan_tag callback to allow XDP code to read
> > packet's VLAN tag.
> > 
> > Signed-off-by: Larysa Zaremba <larysa.zaremba@...el.com>
> > ---
> >  drivers/net/ethernet/intel/ice/ice_txrx_lib.c | 44 +++++++++++++++++++
> >  1 file changed, 44 insertions(+)
> > 
> > diff --git a/drivers/net/ethernet/intel/ice/ice_txrx_lib.c b/drivers/net/ethernet/intel/ice/ice_txrx_lib.c
> > index 1caa73644e7b..39547feb6106 100644
> > --- a/drivers/net/ethernet/intel/ice/ice_txrx_lib.c
> > +++ b/drivers/net/ethernet/intel/ice/ice_txrx_lib.c
> > @@ -627,7 +627,51 @@ static int ice_xdp_rx_hash(const struct xdp_md *ctx, u32 *hash,
> >  	return 0;
> >  }
> >  
> > +/**
> > + * ice_xdp_rx_ctag - VLAN tag XDP hint handler
> > + * @ctx: XDP buff pointer
> > + * @vlan_tag: destination address
> > + *
> > + * Copy VLAN tag (if was stripped) to the destination address.
> > + */
> > +static int ice_xdp_rx_ctag(const struct xdp_md *ctx, u16 *vlan_tag)
> > +{
> > +	const struct ice_xdp_buff *xdp_ext = (void *)ctx;
> > +	netdev_features_t features;
> > +
> 
> [..]
> 
> > +	features = xdp_ext->rx_ring->netdev->features;
> > +
> > +	if (!(features & NETIF_F_HW_VLAN_CTAG_RX))
> > +		return -EINVAL;
> 
> Passing-by comment: why do we need to check features?
> ice_get_vlan_tag_from_rx_desc seems to be checking a bunch of
> fields in the descriptors, so that should be enough?

Unfortunately, it is not enough, because it only checks, if there is a valid 
value in the descriptor, without distinguishing c-tag from s-tag. In this
hardware, c-tag and s-tag are mutually exclusive, so they can occupy same 
descriptor fields. Checking netdev features is just the easiest way to tell them 
apart.

I guess, storing this information in in the ring structure would be more 
efficient than checking netdev features. I know Piotr Raczynski indends to 
review this series, so maybe he would provide some additional 
feedback/suggestions.

> 
> > +
> > +	*vlan_tag = ice_get_vlan_tag_from_rx_desc(xdp_ext->eop_desc);
> 
> Should we also do the following:
> 
> if (!*vlan_tag)
> 	return -ENODATA;
> 
> ?

Oh, returning VLAN tag with zero value really made sense to me at the beginning,
but after playing with different kinds of packets, I think returning error makes 
more sense. Will change.

> 
> > +	return 0;
> > +}
> > +
> > +/**
> > + * ice_xdp_rx_stag - VLAN s-tag XDP hint handler
> > + * @ctx: XDP buff pointer
> > + * @vlan_tag: destination address
> > + *
> > + * Copy VLAN s-tag (if was stripped) to the destination address.
> > + */
> > +static int ice_xdp_rx_stag(const struct xdp_md *ctx, u16 *vlan_tag)
> > +{
> > +	const struct ice_xdp_buff *xdp_ext = (void *)ctx;
> > +	netdev_features_t features;
> > +
> > +	features = xdp_ext->rx_ring->netdev->features;
> > +
> > +	if (!(features & NETIF_F_HW_VLAN_STAG_RX))
> > +		return -EINVAL;
> > +
> > +	*vlan_tag = ice_get_vlan_tag_from_rx_desc(xdp_ext->eop_desc);
> > +	return 0;
> > +}
> > +
> >  const struct xdp_metadata_ops ice_xdp_md_ops = {
> >  	.xmo_rx_timestamp		= ice_xdp_rx_hw_ts,
> >  	.xmo_rx_hash			= ice_xdp_rx_hash,
> > +	.xmo_rx_ctag			= ice_xdp_rx_ctag,
> > +	.xmo_rx_stag			= ice_xdp_rx_stag,
> >  };
> > -- 
> > 2.35.3
> >