[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <694A8714-F178-4211-A9DC-DA7604B8D5B0@fb.com>
Date: Thu, 18 Jun 2020 20:18:06 +0000
From: Song Liu <songliubraving@...com>
To: Peter Zijlstra <peterz@...radead.org>
CC: linux-kernel <linux-kernel@...r.kernel.org>,
Kan Liang <kan.liang@...ux.intel.com>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
Andi Kleen <ak@...ux.intel.com>,
"Josh Poimboeuf" <jpoimboe@...hat.com>, Tejun Heo <tj@...nel.org>,
Andrii Nakryiko <andriin@...com>,
Francis Ricci <fjricci@...com>
Subject: Re: WARNING with LBR + precise_ip=2 + bpf_get_stackid()
> On Jun 18, 2020, at 11:33 AM, Peter Zijlstra <peterz@...radead.org> wrote:
>
> On Thu, Jun 18, 2020 at 05:25:24PM +0000, Song Liu wrote:
>
[...]
>> 0xfffffe00004d1380 __start_orc_unwind
>> 0xfffffe00004d1388 unwind_next_frame.cold.7
>> 0xfffffe00004d13c8 perf_callchain_kernel
>> 0xfffffe00004d1418 entry_SYSCALL_64
>> 0xfffffe00004d1450 get_perf_callchain
>> 0xfffffe00004d14b0 bpf_get_stack
>
> This, this is where it gets buggerd. This is a BPF problem, not a perf
> problem.
>
> You'll note that if you ask perf for a PERF_SAMPLE_CALLCHAIN we'll do
> that early in setup_pebs_fixed_sample_data().
We did notice perf works fine in similar scenario. I guess need similar
logic in BPF.
Thanks,
Song
Powered by blists - more mailing lists