netdev - Re: [PATCH RFC v1 03/15] bnxt: add XDP frame size to driver

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: <20200320192201.GA21532@C02YVCJELVCG.greyhouse.net>
Date:   Fri, 20 Mar 2020 15:22:01 -0400
From:   Andy Gospodarek <andy@...yhouse.net>
To:     Jesper Dangaard Brouer <brouer@...hat.com>
Cc:     sameehj@...zon.com, Michael Chan <michael.chan@...adcom.com>,
        Andy Gospodarek <andrew.gospodarek@...adcom.com>,
        netdev@...r.kernel.org, bpf@...r.kernel.org, zorik@...zon.com,
        akiyano@...zon.com, gtzalik@...zon.com,
        Toke Høiland-Jørgensen <toke@...e.dk>,
        Daniel Borkmann <borkmann@...earbox.net>,
        Alexei Starovoitov <alexei.starovoitov@...il.com>,
        John Fastabend <john.fastabend@...il.com>,
        Alexander Duyck <alexander.duyck@...il.com>,
        Jeff Kirsher <jeffrey.t.kirsher@...el.com>,
        David Ahern <dsahern@...il.com>,
        Willem de Bruijn <willemdebruijn.kernel@...il.com>,
        Ilias Apalodimas <ilias.apalodimas@...aro.org>,
        Lorenzo Bianconi <lorenzo@...nel.org>
Subject: Re: [PATCH RFC v1 03/15] bnxt: add XDP frame size to driver

On Tue, Mar 17, 2020 at 06:29:22PM +0100, Jesper Dangaard Brouer wrote:
> This driver uses full PAGE_SIZE pages when XDP is enabled.
> 
> Cc: Michael Chan <michael.chan@...adcom.com>
> Cc: Andy Gospodarek <andrew.gospodarek@...adcom.com>
> Signed-off-by: Jesper Dangaard Brouer <brouer@...hat.com>
> ---
>  drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c |    1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c
> index c6f6f2033880..5e3b4a3b69ea 100644
> --- a/drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c
> +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c
> @@ -138,6 +138,7 @@ bool bnxt_rx_xdp(struct bnxt *bp, struct bnxt_rx_ring_info *rxr, u16 cons,
>  	xdp_set_data_meta_invalid(&xdp);
>  	xdp.data_end = *data_ptr + *len;
>  	xdp.rxq = &rxr->xdp_rxq;
> +	xdp.frame_sz = PAGE_SIZE; /* BNXT_RX_PAGE_MODE(bp) when XDP enabled */

So if we want this to be the true size that the packet inside the the DMA
buffer could grow to, I _think_ this would need to be:

	xdp.frame_sz = PAGE_SIZE - XDP_PACKET_HEADROOM;

but I also noted that in patch 8 of the series that there is a check
against data_end - data_hard_start, so functionally your original patch
appears to be correct.

If the intent was just to capture the size of the [DMA] buffer available
for the datagram, maybe calling this new field 'buf_sz' or similar would
be nice as it does not convey anything about the on-wire size like
'frame_sz' does.


>  	orig_data = xdp.data;
>  
>  	rcu_read_lock();
> 
>