bpf.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Dan Carpenter <dan.carpenter@oracle.com>
To: ast@kernel.org
Cc: bpf@vger.kernel.org
Subject: [bug report] bpf: Support attaching tracing BPF program to other BPF programs
Date: Wed, 11 Mar 2020 15:10:01 +0300	[thread overview]
Message-ID: <20200311121001.GA20472@mwanda> (raw)

Hello Alexei Starovoitov,

This is a semi-automatic email about new static checker warnings.

The patch 5b92a28aae4d: "bpf: Support attaching tracing BPF program
to other BPF programs" from Nov 14, 2019, leads to the following
Smatch complaint:

    kernel/bpf/btf.c:3744 btf_ctx_access()
    error: we previously assumed 't' could be null (see line 3706)

kernel/bpf/btf.c
  3705		/* if (t == NULL) Fall back to default BPF prog with 5 u64 arguments */
  3706		nr_args = t ? btf_type_vlen(t) : 5;
                          ^
Check

  3707		if (prog->aux->attach_btf_trace) {
  3708			/* skip first 'void *__data' argument in btf_trace_##name typedef */
  3709			args++;
  3710			nr_args--;
  3711		}
  3712	
  3713		if (arg == nr_args) {
  3714			if (prog->expected_attach_type == BPF_TRACE_FEXIT) {
  3715				if (!t)
  3716					return true;
  3717				t = btf_type_by_id(btf, t->type);
  3718			} else if (prog->expected_attach_type == BPF_MODIFY_RETURN) {
  3719				/* For now the BPF_MODIFY_RETURN can only be attached to
  3720				 * functions that return an int.
  3721				 */
  3722				if (!t)
  3723					return false;
  3724	
  3725				t = btf_type_skip_modifiers(btf, t->type, NULL);
  3726				if (!btf_type_is_int(t)) {
  3727					bpf_log(log,
  3728						"ret type %s not allowed for fmod_ret\n",
  3729						btf_kind_str[BTF_INFO_KIND(t->info)]);
  3730					return false;
  3731				}
  3732			}

Smatch is complaining that maybe t is NULL and prog->expected_attach_type
is neither BPF_TRACE_FEXIT nor BPF_MODIFY_RETURN.

  3733		} else if (arg >= nr_args) {
  3734			bpf_log(log, "func '%s' doesn't have %d-th argument\n",
  3735				tname, arg + 1);
  3736			return false;
  3737		} else {
  3738			if (!t)
  3739				/* Default prog with 5 args */
  3740				return true;
  3741			t = btf_type_by_id(btf, args[arg].type);
  3742		}
  3743		/* skip modifiers */
  3744		while (btf_type_is_modifier(t))
                                            ^
Dereference

  3745			t = btf_type_by_id(btf, t->type);
  3746		if (btf_type_is_int(t) || btf_type_is_enum(t))

regards,
dan carpenter

                 reply	other threads:[~2020-03-11 12:10 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20200311121001.GA20472@mwanda \
    --to=dan.carpenter@oracle.com \
    --cc=ast@kernel.org \
    --cc=bpf@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).