Re: [PATCH bpf-next] xsk: support AF_PACKET

["Toke Høiland-Jørgensen" <toke@redhat.com>]

Daniel Borkmann <daniel@iogearbox.net> writes:

> On 5/28/21 12:54 PM, Toke Høiland-Jørgensen wrote:
>> Daniel Borkmann <daniel@iogearbox.net> writes:
>>> On 5/28/21 12:00 PM, Magnus Karlsson wrote:
>>>> On Fri, May 28, 2021 at 11:52 AM Jesper Dangaard Brouer
>>>> <brouer@redhat.com> wrote:
>>>>> On Fri, 28 May 2021 17:02:01 +0800
>>>>> Xuan Zhuo <xuanzhuo@linux.alibaba.com> wrote:
>>>>>> On Fri, 28 May 2021 10:55:58 +0200, Toke Høiland-Jørgensen <toke@redhat.com> wrote:
>>>>>>> Xuan Zhuo <xuanzhuo@linux.alibaba.com> writes:
>>>>>>>
>>>>>>>> In xsk mode, users cannot use AF_PACKET(tcpdump) to observe the current
>>>>>>>> rx/tx data packets. This feature is very important in many cases. So
>>>>>>>> this patch allows AF_PACKET to obtain xsk packages.
>>>>>>>
>>>>>>> You can use xdpdump to dump the packets from the XDP program before it
>>>>>>> gets redirected into the XSK:
>>>>>>> https://github.com/xdp-project/xdp-tools/tree/master/xdp-dump
>>>>>>
>>>>>> Wow, this is a good idea.
>>>>>
>>>>> Yes, it is rather cool (credit to Eelco).  Notice the extra info you
>>>>> can capture from 'exit', like XDP return codes, if_index, rx_queue.
>>>>>
>>>>> The tool uses the perf ring-buffer to send/copy data to userspace.
>>>>> This is actually surprisingly fast, but I still think AF_XDP will be
>>>>> faster (but it usually 'steals' the packet).
>>>>>
>>>>> Another (crazy?) idea is to extend this (and xdpdump), is to leverage
>>>>> Hangbin's recent XDP_REDIRECT extension e624d4ed4aa8 ("xdp: Extend
>>>>> xdp_redirect_map with broadcast support").  We now have a
>>>>> xdp_redirect_map flag BPF_F_BROADCAST, what if we create a
>>>>> BPF_F_CLONE_PASS flag?
>>>>>
>>>>> The semantic meaning of BPF_F_CLONE_PASS flag is to copy/clone the
>>>>> packet for the specified map target index (e.g AF_XDP map), but
>>>>> afterwards it does like veth/cpumap and creates an SKB from the
>>>>> xdp_frame (see __xdp_build_skb_from_frame()) and send to netstack.
>>>>> (Feel free to kick me if this doesn't make any sense)
>>>>
>>>> This would be a smooth way to implement clone support for AF_XDP. If
>>>> we had this and someone added AF_XDP support to libpcap, we could both
>>>> capture AF_XDP traffic with tcpdump (using this clone functionality in
>>>> the XDP program) and speed up tcpdump for dumping traffic destined for
>>>> regular sockets. Would that solve your use case Xuan? Note that I have
>>>> not looked into the BPF_F_CLONE_PASS code, so do not know at this
>>>> point what it would take to support this for XSKMAPs.
>>>
>>> Recently also ended up with something similar for our XDP LB to record pcaps [0] ;)
>>> My question is.. tcpdump doesn't really care where the packet data comes from,
>>> so why not extending libpcap's Linux-related internals to either capture from
>>> perf RB or BPF ringbuf rather than AF_PACKET sockets? Cloning is slow, and if
>>> you need to end up creating an skb which is then cloned once again inside AF_PACKET
>>> it's even worse. Just relying and reading out, say, perf RB you don't need any
>>> clones at all.
>> 
>> We discussed this when creating xdpdump and decided to keep it as a
>> separate tool for the time being. I forget the details of the
>> discussion, maybe Eelco remembers.
>> 
>> Anyway, xdpdump does have a "pipe pcap to stdout" feature so you can do
>> `xdpdump | tcpdump` and get the interactive output; and it will also
>> save pcap information to disk, of course (using pcap-ng so it can also
>> save metadata like XDP program name and return code).
>
> Right, and this should yield a significantly better performance compared to
> cloning & pushing traffic into AF_PACKET. I presume not many folks are aware
> of xdpdump (yet) which is probably why such patch was created here..

What, are you implying we haven't achieved world domination yet?
Inconceivable! ;)

> a native libpcap implementation could solve that aspect fwiw and
> additionally hook at the same points as AF_PACKET via BPF but without
> the hassle/overhead of things like dev_queue_xmit_nit() in fast path.
> (Maybe another option could be to have a drop-in replacement
> libpcap.so for tcpdump using it transparently.)

I do believe that Michael was open to adding something like this to
tcpdump/libpcap when I last talked to him about it; and I'm certainly
not opposed to it either! Hooking up tcpdump like this may be a bit of a
firehose, though, so it would be nice to be able to carry over the
kernel-side filtering as well. I suppose it should be possible to write
an eBPF bytecode generator that does a bit of setup and then just
translates the cBPF packet filtering ops, no? This would be cool to have
in any case; IIRC Cloudflare did something like that but took a detour
through C code generation?

-Toke