Re: [PATCH bpf-next 0/3] Fixes for TC-BPF series

From: Yaniv Agman <yanivagman@gmail.com>
To: memxor@gmail.com
Cc: andrii@kernel.org, ast@kernel.org, bpf@vger.kernel.org,
	daniel@iogearbox.net, netdev@vger.kernel.org, toke@redhat.com
Subject: Re: [PATCH bpf-next 0/3] Fixes for TC-BPF series
Date: Mon, 14 Jun 2021 12:32:07 +0300	[thread overview]
Message-ID: <CAMy7=ZUXRJni3uUVWkWFu8Dkc5XCVsM54i_iLDwHQ5Y0Z3inJw@mail.gmail.com> (raw)

Hi Kartikeya,

I recently started experimenting with the new tc-bpf API (which is
great, many thanks!) and I wanted to share a potential problem I
found.
I'm using this "Fixes for TC-BPF series" thread to write about it, but
it is not directly related to this patch set.

According to the API summary given in
https://lore.kernel.org/bpf/20210512103451.989420-3-memxor@gmail.com/,
"It is advised that if the qdisc is operated on by many programs,
then the program at least check that there are no other existing
filters before deleting the clsact qdisc."
In the example given, one should:

/* set opts as NULL, as we're not really interested in
* getting any info for a particular filter, but just
* detecting its presence.
*/
r = bpf_tc_query(&hook, NULL);

However, following in this summary, where bpf_tc_query is described,
it is written that the opts argument cannot be NULL.
And indeed, when I tried to use the example above, an error (EINVAL)
was returned (as expected?)

Am I missing something?

Yaniv