linux-kernel - Re: [BUGFIX PATCH perf/core ] perf-probe: Fix to show lines of sys

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: <20150812134042.GE31059@kernel.org>
Date:	Wed, 12 Aug 2015 10:40:42 -0300
From:	Arnaldo Carvalho de Melo <acme@...nel.org>
To:	Masami Hiramatsu <masami.hiramatsu.pt@...achi.com>
Cc:	Namhyung Kim <namhyung@...nel.org>, Jiri Olsa <jolsa@...hat.com>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	David Ahern <dsahern@...il.com>,
	Brendan Gregg <brendan.d.gregg@...il.com>
Subject: Re: [BUGFIX PATCH perf/core ] perf-probe: Fix to show lines of sys_
 functions correctly

Em Wed, Aug 12, 2015 at 10:23:55AM -0300, Arnaldo Carvalho de Melo escreveu:
> Em Wed, Aug 12, 2015 at 10:24:07AM +0900, Masami Hiramatsu escreveu:
> > "perf probe --lines sys_poll" shows only the first line of
> > sys_poll, because the SYSCALL_DEFINE macro
> > ----
 
> Thanks! Just try prefixing those ---- lines one space so that git-am
> works on your messages :-)
 
> Now to build and test, will report here the results.

Ok, now -L works:

  [root@zoo ~]# perf probe -L sys_select
  <SyS_select@...me/git/linux/fs/select.c:0>
      0  SYSCALL_DEFINE5(select, int, n, fd_set __user *, inp, fd_set __user *, outp,
                        fd_set __user *, exp, struct timeval __user *, tvp)
         {
      3         struct timespec end_time, *to = NULL;
                struct timeval tv;
                int ret;
         
      7         if (tvp) {
      8                 if (copy_from_user(&tv, tvp, sizeof(tv)))
      9                         return -EFAULT;
         
     11                 to = &end_time;
     12                 if (poll_select_set_timeout(to,
                                        tv.tv_sec + (tv.tv_usec / USEC_PER_SEC),
     14                                 (tv.tv_usec % USEC_PER_SEC) * NSEC_PER_USEC))
     15                         return -EINVAL;
                }
         
     18         ret = core_sys_select(n, inp, outp, exp, to);
     19         ret = poll_select_copy_remaining(&end_time, tvp, 1, ret);
         
     21         return ret;
         }
         
         static long do_pselect(int n, fd_set __user *inp, fd_set __user *outp,

But then adding a probe doesn't work on sys_select, even with -L stating that I
can add a probe at that line, maybe the code you added to make -L work needs to
be shared with the other operations in 'perf probe'?

  [root@zoo ~]# 
  [root@zoo ~]# perf probe 'get_timeval=sys_select:11 tv_sec=tv.tv_sec tv_usec=tv.tv_usec'
  Probe point 'sys_select:11' not found.
    Error: Failed to add events.
  [root@zoo ~]#

  [root@zoo ~]# perf probe -V sys_select
  Available variables at sys_select
        @<SyS_select+0>
                long int        exp
                long int        inp
                long int        n
                long int        outp
                long int        tvp
  [root@zoo ~]# perf probe -V sys_select:11
  Failed to find the address of sys_select:11
    Error: Failed to show vars.
  [root@zoo ~]# 

While this works:

  [root@zoo ~]# perf probe 'vfs_getname=getname_flags:72 pathname=filename:string'
  Added new event:
    probe:vfs_getname    (on getname_flags:72 with pathname=filename:string)

  You can now use it in all perf tools, such as:

	perf record -e probe:vfs_getname -aR sleep 1

- Arnaldo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/