[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <Z1fTbcRiDRPU9IPQ@pop-os.localdomain>
Date: Mon, 9 Dec 2024 21:36:45 -0800
From: Cong Wang <xiyou.wangcong@...il.com>
To: Daniel Borkmann <daniel@...earbox.net>
Cc: netdev@...r.kernel.org, bpf@...r.kernel.org,
Cong Wang <cong.wang@...edance.com>
Subject: Re: [Patch bpf v2 0/4] bpf: a bug fix and test cases for
bpf_skb_change_tail()
On Fri, Dec 06, 2024 at 10:35:28PM +0100, Daniel Borkmann wrote:
> Hi Cong,
>
> On 11/29/24 2:22 AM, Cong Wang wrote:
> > From: Cong Wang <cong.wang@...edance.com>
> >
> > This patchset fixes a bug in bpf_skb_change_tail() helper and adds test
> > cases for it, as requested by Daniel and John.
> >
> > ---
> > v2: added a test case for TC where offsets are positive
> > fixed a typo in 1/4 patch description
> > reduced buffer size in the sockmap test case
>
> I ran the selftest several times but it's repeatedly failing whereas
> without the series bpf tree CI seems fine. The CI fails on tc tests,
> so potentially patch 4 is causing this.
Ah, thanks for catching it.
Previously, the CI job failed due to flaky tests, which are tests that
inconsistently pass or fail. However, this time the failure indicates
a genuine issue.
>
> Switching over to tcx APIs from libbpf might automatically address
> this given the failures seem to be in 'revision unexpected' which is
> likely due to legacy libbpf tc APIs detaching but not deleting the
> underlying qdisc.
Sure, thanks for the hint. I will update this patchset.
Thanks.
Powered by blists - more mailing lists