linux-kernel - Re: [PATCH RFC bpf-next 1/3] bpf: report probe fault to BPF stderr

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAADnVQJAdAxEOWT6avzwq6ZrXhEdovhx3yibgA6T8wnMEnnAjg@mail.gmail.com>
Date: Wed, 1 Oct 2025 19:03:46 -0700
From: Alexei Starovoitov <alexei.starovoitov@...il.com>
To: Menglong Dong <menglong8.dong@...il.com>
Cc: Alexei Starovoitov <ast@...nel.org>, bpf <bpf@...r.kernel.org>, 
	LKML <linux-kernel@...r.kernel.org>, 
	linux-trace-kernel <linux-trace-kernel@...r.kernel.org>, jiang.biao@...ux.dev
Subject: Re: [PATCH RFC bpf-next 1/3] bpf: report probe fault to BPF stderr

On Fri, Sep 26, 2025 at 11:12 PM Menglong Dong <menglong8.dong@...il.com> wrote:
>
> Introduce the function bpf_prog_report_probe_violation(), which is used
> to report the memory probe fault to the user by the BPF stderr.
>
> Signed-off-by: Menglong Dong <menglong.dong@...ux.dev>
> ---
>  include/linux/bpf.h      |  1 +
>  kernel/trace/bpf_trace.c | 18 ++++++++++++++++++
>  2 files changed, 19 insertions(+)
>
> diff --git a/include/linux/bpf.h b/include/linux/bpf.h
> index 6338e54a9b1f..a31c5ce56c32 100644
> --- a/include/linux/bpf.h
> +++ b/include/linux/bpf.h
> @@ -2902,6 +2902,7 @@ void bpf_dynptr_init(struct bpf_dynptr_kern *ptr, void *data,
>  void bpf_dynptr_set_null(struct bpf_dynptr_kern *ptr);
>  void bpf_dynptr_set_rdonly(struct bpf_dynptr_kern *ptr);
>  void bpf_prog_report_arena_violation(bool write, unsigned long addr, unsigned long fault_ip);
> +void bpf_prog_report_probe_violation(bool write, unsigned long fault_ip);
>
>  #else /* !CONFIG_BPF_SYSCALL */
>  static inline struct bpf_prog *bpf_prog_get(u32 ufd)
> diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
> index 8f23f5273bab..9bd03a9f53db 100644
> --- a/kernel/trace/bpf_trace.c
> +++ b/kernel/trace/bpf_trace.c
> @@ -2055,6 +2055,24 @@ void bpf_put_raw_tracepoint(struct bpf_raw_event_map *btp)
>         module_put(mod);
>  }
>
> +void bpf_prog_report_probe_violation(bool write, unsigned long fault_ip)
> +{
> +       struct bpf_stream_stage ss;
> +       struct bpf_prog *prog;
> +
> +       rcu_read_lock();
> +       prog = bpf_prog_ksym_find(fault_ip);
> +       rcu_read_unlock();
> +       if (!prog)
> +               return;
> +
> +       bpf_stream_stage(ss, prog, BPF_STDERR, ({
> +               bpf_stream_printk(ss, "ERROR: Probe %s access faule, insn=0x%lx\n",
> +                                 write ? "WRITE" : "READ", fault_ip);
> +               bpf_stream_dump_stack(ss);
> +       }));

Interesting idea, but the above message is not helpful.
Users cannot decipher a fault_ip within a bpf prog.
It's just a random number.

But stepping back... just faults are common in tracing.
If we start printing them we will just fill the stream to the max,
but users won't know that the message is there, since no one
expects it. arena and lock errors are rare and arena faults
were specifically requested by folks who develop progs that use arena.
This one is different. These faults have been around for a long time
and I don't recall people asking for more verbosity.
We can add them with an extra flag specified at prog load time,
but even then. Doesn't feel that useful.