[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CALC8CXdNYGy2itPVkmuS3grfCof9ggHTRVyps85_9X2WwREpXQ@mail.gmail.com>
Date: Thu, 2 Oct 2025 12:18:11 -0400
From: ChaosEsque Team <chaosesqueteam@...il.com>
To: Menglong Dong <menglong8.dong@...il.com>
Cc: andrii@...nel.org, eddyz87@...il.com, mykolal@...com, ast@...nel.org,
daniel@...earbox.net, martin.lau@...ux.dev, song@...nel.org,
yonghong.song@...ux.dev, john.fastabend@...il.com, kpsingh@...nel.org,
sdf@...ichev.me, haoluo@...gle.com, jolsa@...nel.org, shuah@...nel.org,
yikai.lin@...o.com, memxor@...il.com, bpf@...r.kernel.org,
linux-kselftest@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH bpf-next v2 2/3] selftests/bpf: skip recursive functions
for kprobe_multi
Indeed, it is so long. Sir Dong, of Meng.
On Tue, Aug 26, 2025 at 4:05 AM Menglong Dong <menglong8.dong@...il.com> wrote:
>
> Some functions is recursive for the kprobe_multi and impact the benchmark
> results. So just skip them.
>
> Signed-off-by: Menglong Dong <dongml2@...natelecom.cn>
> ---
> v2:
> - introduce trace_blacklist instead of copy-pasting strcmp
> ---
> tools/testing/selftests/bpf/trace_helpers.c | 19 +++++++++++++++++++
> 1 file changed, 19 insertions(+)
>
> diff --git a/tools/testing/selftests/bpf/trace_helpers.c b/tools/testing/selftests/bpf/trace_helpers.c
> index d24baf244d1f..9577979bd84d 100644
> --- a/tools/testing/selftests/bpf/trace_helpers.c
> +++ b/tools/testing/selftests/bpf/trace_helpers.c
> @@ -19,6 +19,7 @@
> #include <gelf.h>
> #include "bpf/hashmap.h"
> #include "bpf/libbpf_internal.h"
> +#include "bpf_util.h"
>
> #define TRACEFS_PIPE "/sys/kernel/tracing/trace_pipe"
> #define DEBUGFS_PIPE "/sys/kernel/debug/tracing/trace_pipe"
> @@ -540,8 +541,20 @@ static bool is_invalid_entry(char *buf, bool kernel)
> return false;
> }
>
> +static const char * const trace_blacklist[] = {
> + "migrate_disable",
> + "migrate_enable",
> + "rcu_read_unlock_strict",
> + "preempt_count_add",
> + "preempt_count_sub",
> + "__rcu_read_lock",
> + "__rcu_read_unlock",
> +};
> +
> static bool skip_entry(char *name)
> {
> + int i;
> +
> /*
> * We attach to almost all kernel functions and some of them
> * will cause 'suspicious RCU usage' when fprobe is attached
> @@ -559,6 +572,12 @@ static bool skip_entry(char *name)
> if (!strncmp(name, "__ftrace_invalid_address__",
> sizeof("__ftrace_invalid_address__") - 1))
> return true;
> +
> + for (i = 0; i < ARRAY_SIZE(trace_blacklist); i++) {
> + if (!strcmp(name, trace_blacklist[i]))
> + return true;
> + }
> +
> return false;
> }
>
> --
> 2.51.0
>
>
Powered by blists - more mailing lists