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]
Date:	Wed, 30 Sep 2015 10:18:28 -0300
From:	Arnaldo Carvalho de Melo <acme@...nel.org>
To:	平松雅巳 / HIRAMATU,MASAMI 
	<masami.hiramatsu.pt@...achi.com>
Cc:	David Ahern <dsahern@...il.com>, Jiri Olsa <jolsa@...nel.org>,
	Wang Nan <wangnan0@...wei.com>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [perf probe] How to ask what variables can be collected at
 function return?

Em Wed, Sep 30, 2015 at 08:05:09AM +0000, 平松雅巳 / HIRAMATU,MASAMI escreveu:
> From: Arnaldo Carvalho de Melo [mailto:acme@...nel.org]
> >
> >How to figure out which variables can I collect at function return?
> >None?
> 
> At first, it seems the below pattern failed to get the probe point itself,
> because you're specifying the return point of an inlined function, which
> is nowhere.

Would be really nice to have this warning emitted to the user, as well
as bits and pieces of the following explanations.

- Arnaldo
 
> If you give a function which has actual instance, perf-probe may show
> the variables which will be accessable in the entry of the function as below.
> 
> $ sudo ./perf probe -V "vfs_read%return"
> Available variables at vfs_read%return
>         @<vfs_read+0>
>                 char*   buf
>                 loff_t* pos
>                 size_t  count
>                 struct file*    file
> 
> However, this may not work on the arguments passed by registers,
> since those may be already broken by the operations in the function.
> 
> So, to get correct accessible variables at return, you must give an actual
> line number in the function currently.
> 
> $ sudo ./perf probe -L vfs_read
> <vfs_read@...me/mhiramat/ksrc/linux-3/fs/read_write.c:0>
>       0  ssize_t vfs_read(struct file *file, char __user *buf, size_t count, lof
>       1  {
>                 ssize_t ret;
> 
>       4         if (!(file->f_mode & FMODE_READ))
>       5                 return -EBADF;
>       6         if (!(file->f_mode & FMODE_CAN_READ))
>       7                 return -EINVAL;
>       8         if (unlikely(!access_ok(VERIFY_WRITE, buf, count)))
>       9                 return -EFAULT;
> 
>      11         ret = rw_verify_area(READ, file, pos, count);
>      12         if (ret >= 0) {
>                         count = ret;
>      14                 ret = __vfs_read(file, buf, count, pos);
>      15                 if (ret > 0) {
>      16                         fsnotify_access(file);
>      17                         add_rchar(current, ret);
>                         }
>      19                 inc_syscr(current);
>                 }
> 
>                 return ret;
>      23  }
> 
> $ sudo ./perf probe -V vfs_read:23
> Available variables at vfs_read:23
>         @<vfs_read+214>
>                 char*   buf
>                 loff_t* pos
>                 size_t  count
>                 struct file*    file
> 
> 
> >
> >I think we need a better error message :-)
> >
> >[root@zoo ~]# perf probe -V getname_flags%return Return probe must be on the head of a real function.
> 
> It seems that perf probe said " Return probe must be on the head of a real function."
> So this means it can not find the given probe point because the %return is not 
> put on the entry of the function which has actual instance.
> 
> Thanks.
> 
> 
> >Debuginfo analysis failed.
> >  Error: Failed to show vars.
> >[root@zoo ~]#
> >
> >
> >- 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/

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ