Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   Aleksa Sarai <cyphar@cyphar.com>
To:     "Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com>,
        Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>,
        "David S. Miller" <davem@davemloft.net>,
        Masami Hiramatsu <mhiramat@kernel.org>,
        Jonathan Corbet <corbet@lwn.net>,
        Peter Zijlstra <peterz@infradead.org>,
        Ingo Molnar <mingo@redhat.com>,
        Arnaldo Carvalho de Melo <acme@kernel.org>,
        Alexander Shishkin <alexander.shishkin@linux.intel.com>,
        Jiri Olsa <jolsa@redhat.com>,
        Namhyung Kim <namhyung@kernel.org>,
        Steven Rostedt <rostedt@goodmis.org>,
        Shuah Khan <shuah@kernel.org>,
        Alexei Starovoitov <ast@kernel.org>,
        Daniel Borkmann <daniel@iogearbox.net>
Cc:     Aleksa Sarai <cyphar@cyphar.com>, Aleksa Sarai <asarai@suse.de>,
        Christian Brauner <christian@brauner.io>,
        Brendan Gregg <bgregg@netflix.com>, netdev@vger.kernel.org,
        linux-doc@vger.kernel.org, linux-kernel@vger.kernel.org,
        linux-kselftest@vger.kernel.org
Subject: [PATCH v3 0/2] kretprobe: produce sane stack traces
Date:   Thu,  1 Nov 2018 19:35:49 +1100
Message-Id: <20181101083551.3805-1-cyphar@cyphar.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Historically, kretprobe has always produced unusable stack traces
(kretprobe_trampoline is the only entry in most cases, because of the
funky stack pointer overwriting). This has caused quite a few annoyances
when using tracing to debug problems[1] -- since return values are only
available with kretprobes but stack traces were only usable for kprobes,
users had to probe both and then manually associate them.

This patch series stores the stack trace within kretprobe_instance on
the kprobe entry used to set up the kretprobe. This allows for
DTrace-style stack aggregation between function entry and exit with
tools like BPFtrace -- which would not really be doable if the stack
unwinder understood kretprobe_trampoline.

We also revert commit 76094a2cf46e ("ftrace: distinguish kretprobe'd
functions in trace logs") and any follow-up changes because that code is
no longer necessary now that stack traces are sane. *However* this patch
might be a bit contentious since the original usecase (that ftrace
returns shouldn't show kretprobe_trampoline) is arguably still an
issue. Feel free to drop it if you think it is wrong.

Patch changelog:
 v3:
   * kprobe: fix build on !CONFIG_KPROBES
 v2:
   * documentation: mention kretprobe stack-stashing
   * ftrace: add self-test for fixed kretprobe stacktraces
   * ftrace: remove [unknown/kretprobe'd] handling
   * kprobe: remove needless EXPORT statements
   * kprobe: minor corrections to current_kretprobe_instance (switch
     away from hlist_for_each_entry_safe)
   * kprobe: make maximum stack size 127, which is the ftrace default

Aleksa Sarai (2):
  kretprobe: produce sane stack traces
  trace: remove kretprobed checks

 Documentation/kprobes.txt                     |   6 +-
 include/linux/kprobes.h                       |  27 +++++
 kernel/events/callchain.c                     |   8 +-
 kernel/kprobes.c                              | 101 +++++++++++++++++-
 kernel/trace/trace.c                          |  11 +-
 kernel/trace/trace_output.c                   |  34 +-----
 .../test.d/kprobe/kretprobe_stacktrace.tc     |  25 +++++
 7 files changed, 177 insertions(+), 35 deletions(-)
 create mode 100644 tools/testing/selftests/ftrace/test.d/kprobe/kretprobe_stacktrace.tc

-- 
2.19.1