Re: [PATCH net-next v2 2/4] bpf: test_run: Use system page pool for XDP live frame mode

From: John Fastabend <john.fastabend@gmail.com>
To: "Alexander Lobakin" <aleksander.lobakin@intel.com>,
	"Toke Høiland-Jørgensen" <toke@redhat.com>
Cc: Alexei Starovoitov <ast@kernel.org>,
	 Daniel Borkmann <daniel@iogearbox.net>,
	 Andrii Nakryiko <andrii@kernel.org>,
	 Martin KaFai Lau <martin.lau@linux.dev>,
	 Eduard Zingerman <eddyz87@gmail.com>,
	 Song Liu <song@kernel.org>,
	 Yonghong Song <yonghong.song@linux.dev>,
	 John Fastabend <john.fastabend@gmail.com>,
	 KP Singh <kpsingh@kernel.org>,
	 Stanislav Fomichev <sdf@google.com>,
	 Hao Luo <haoluo@google.com>,  Jiri Olsa <jolsa@kernel.org>,
	 "David S. Miller" <davem@davemloft.net>,
	 Jakub Kicinski <kuba@kernel.org>,
	 Jesper Dangaard Brouer <hawk@kernel.org>,
	 Eric Dumazet <edumazet@google.com>,
	 Paolo Abeni <pabeni@redhat.com>,
	 bpf@vger.kernel.org,  netdev@vger.kernel.org
Subject: Re: [PATCH net-next v2 2/4] bpf: test_run: Use system page pool for XDP live frame mode
Date: Wed, 03 Apr 2024 13:39:35 -0700	[thread overview]
Message-ID: <660dbe87d8797_1cf6b2083c@john.notmuch> (raw)
In-Reply-To: <2746b1d9-2e1f-4e66-89ed-19949a189a92@intel.com>

Alexander Lobakin wrote:
> From: Toke Høiland-Jørgensen <toke@redhat.com>
> Date: Tue, 20 Feb 2024 22:03:39 +0100
> 
> > The BPF_TEST_RUN code in XDP live frame mode creates a new page pool
> > each time it is called and uses that to allocate the frames used for the
> > XDP run. This works well if the syscall is used with a high repetitions
> > number, as it allows for efficient page recycling. However, if used with
> > a small number of repetitions, the overhead of creating and tearing down
> > the page pool is significant, and can even lead to system stalls if the
> > syscall is called in a tight loop.
> > 
> > Now that we have a persistent system page pool instance, it becomes
> > pretty straight forward to change the test_run code to use it. The only
> > wrinkle is that we can no longer rely on a custom page init callback
> > from page_pool itself; instead, we change the test_run code to write a
> > random cookie value to the beginning of the page as an indicator that
> > the page has been initialised and can be re-used without copying the
> > initial data again.
> > 
> > The cookie is a random 128-bit value, which means the probability that
> > we will get accidental collisions (which would lead to recycling the
> > wrong page values and reading garbage) is on the order of 2^-128. This
> > is in the "won't happen before the heat death of the universe" range, so
> > this marking is safe for the intended usage.
> > 
> > Reviewed-by: Alexander Lobakin <aleksander.lobakin@intel.com>
> > Tested-by: Alexander Lobakin <aleksander.lobakin@intel.com>
> > Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
> 
> Hey,
> 
> What's the status of this series, now that the window is open?
> 
> Thanks,
> Olek

Hi Toke,

I read the thread from top to bottom so seems someone else notices the
2^128 is unique numbers not the collision probability. Anywaays I'm still
a bit confused, whats the use case here? Maybe I need to understand
what this XDP live frame mode is better?

Could another solution be to avoid calling BPF_TEST_RUN multiple times
in a row? Or perhaps have a BPF_SETUP_RUN that does the config and lets
BPF_TEST_RUN skip the page allocation? Another idea just have the first
run of BPF_TEST_RUN init a page pool and not destroy it.

Thanks,
John