| 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
| ||
|
Date: Thu, 27 Jun 2019 15:07:56 -0700
From: "Jonathan Lemon" <jonathan.lemon@...il.com>
To: "Jesper Dangaard Brouer" <brouer@...hat.com>
Cc: "Willem de Bruijn" <willemdebruijn.kernel@...il.com>,
"Toke Høiland-Jørgensen" <toke@...hat.com>,
"Machulsky, Zorik" <zorik@...zon.com>,
"Jubran, Samih" <sameehj@...zon.com>, davem@...emloft.net,
netdev@...r.kernel.org, "Woodhouse, David" <dwmw@...zon.co.uk>,
"Matushevsky, Alexander" <matua@...zon.com>,
"Bshara, Saeed" <saeedb@...zon.com>,
"Wilson, Matt" <msw@...zon.com>,
"Liguori, Anthony" <aliguori@...zon.com>,
"Bshara, Nafea" <nafea@...zon.com>,
"Tzalik, Guy" <gtzalik@...zon.com>,
"Belgazal, Netanel" <netanel@...zon.com>,
"Saidi, Ali" <alisaidi@...zon.com>,
"Herrenschmidt, Benjamin" <benh@...zon.com>,
"Kiyanovski, Arthur" <akiyano@...zon.com>,
"Daniel Borkmann" <borkmann@...earbox.net>,
"Ilias Apalodimas" <ilias.apalodimas@...aro.org>,
"Alexei Starovoitov" <alexei.starovoitov@...il.com>,
"Jakub Kicinski" <jakub.kicinski@...ronome.com>,
xdp-newbies@...r.kernel.org
Subject: Re: XDP multi-buffer incl. jumbo-frames (Was: [RFC V1 net-next 1/1]
net: ena: implement XDP drop support)
On 26 Jun 2019, at 13:00, Jesper Dangaard Brouer wrote:
> On Wed, 26 Jun 2019 09:42:07 -0700 "Jonathan Lemon"
> <jonathan.lemon@...il.com> wrote:
>
>> If all packets are collected together (like the bulk queue does), and
>> then passed to XDP, this could easily be made backwards compatible.
>> If the XDP program isn't 'multi-frag' aware, then each packet is just
>> passed in individually.
>
> My proposal#1 is XDP only access first-buffer[1], as this simplifies
> things.
>
> (AFAIK) What you are proposing is that all the buffers are passed to
> the XDP prog (in form of a iovec). I need some more details about
> your
> suggestion.
I was thinking this over yesterday - and was probably conflating packets
and buffers a bit. Suppose that for the purposes of this discussion,
we're
talking about a single packet that is split over multiple buffer areas.
Say, on RX, with header split:
buf[0] = header
buf[1] = data
For LRO (hw recv) and jumbo frames (and TSO):
buf[0] = hdr + data
buf[1] = data
buf[n] = data
GRO cases, where individual packets are reassembled by software, aren't
handled here.
> Specifically:
>
> - What is the semantic when a 3 buffer packet is input and XDP prog
> choose to return XDP_DROP for packet #2 ?
>
> - Same situation of packet #2 wants a XDP_TX or redirect?
The collection of buffers represents a single packet, so this isn't
applicable here, right?
However, just thinking about incomplete data words (aka: pullup) gives
me a headache - seems this would complicate the BPF/verifier quite a
bit.
So perhaps just restricting things to the first entry would do for now?
As far as the exact data structure used to hold the buffers, it would
be nice if it had the same layout as a bio_vec, in case someone wanted
to get clever and start transferring things over directly.
--
Jonathan
>> Of course, passing in the equivalent of a iovec requires some form of
>> loop support on the BPF side, doesn't it?
>
> The data structure used for holding these packet buffers/segments also
> needs to be discussed. I would either use an array of bio_vec[2] or
> skb_frag_t (aka skb_frag_struct). The skb_frag_t would be most
> obvious, as we already have to write this when creating an SKB, in
> skb_shared_info area. (Structs listed below signature).
>
> The problem is also that size of these structs (16 bytes) per
> buffer/segment, and we likely need to support 17 segments, as this
> need
> to be compatible with SKBs (size 272 bytes).
>
> My idea here is that we simply use the same memory area, that we have
> to
> store skb_shared_info into. As this allow us to get the SKB setup for
> free, when doing XDP_PASS or when doing SKB alloc after XDP_REDIRECT.
>
>
> [1]
> https://github.com/xdp-project/xdp-project/blob/master/areas/core/xdp-multi-buffer01-design.org#proposal1-xdp-only-access-first-buffer
>
> [2]
> https://lore.kernel.org/netdev/20190501041757.8647-1-willy@infradead.org/
> --
> Best regards,
> Jesper Dangaard Brouer
> MSc.CS, Principal Kernel Engineer at Red Hat
> LinkedIn: http://www.linkedin.com/in/brouer
>
>
> $ pahole -C skb_frag_struct vmlinux
> struct skb_frag_struct {
> struct {
> struct page * p; /* 0 8 */
> } page; /* 0 8 */
> __u32 page_offset; /* 8 4 */
> __u32 size; /* 12 4 */
>
> /* size: 16, cachelines: 1, members: 3 */
> /* last cacheline: 16 bytes */
> };
>
> $ pahole -C bio_vec vmlinux
> struct bio_vec {
> struct page * bv_page; /* 0 8 */
> unsigned int bv_len; /* 8 4 */
> unsigned int bv_offset; /* 12 4 */
>
> /* size: 16, cachelines: 1, members: 3 */
> /* last cacheline: 16 bytes */
> };
>
> $ pahole -C skb_shared_info vmlinux
> struct skb_shared_info {
> __u8 __unused; /* 0 1 */
> __u8 meta_len; /* 1 1 */
> __u8 nr_frags; /* 2 1 */
> __u8 tx_flags; /* 3 1 */
> short unsigned int gso_size; /* 4 2 */
> short unsigned int gso_segs; /* 6 2 */
> struct sk_buff * frag_list; /* 8 8 */
> struct skb_shared_hwtstamps hwtstamps; /* 16 8 */
> unsigned int gso_type; /* 24 4 */
> u32 tskey; /* 28 4 */
> atomic_t dataref; /* 32 0 */
>
> /* XXX 8 bytes hole, try to pack */
>
> void * destructor_arg; /* 40 8 */
> skb_frag_t frags[17]; /* 48 272 */
>
> /* size: 320, cachelines: 5, members: 13 */
> /* sum members: 312, holes: 1, sum holes: 8 */
> };
Powered by blists - more mailing lists