linux-kernel - Re: [PATCH v5 16/19] tracing: probeevent: Add array type support

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:   Thu, 15 Mar 2018 16:18:40 +0900
From:   Masami Hiramatsu <mhiramat@...nel.org>
To:     Ravi Bangoria <ravi.bangoria@...ux.vnet.ibm.com>
Cc:     Steven Rostedt <rostedt@...dmis.org>, linux-kernel@...r.kernel.org,
        Ingo Molnar <mingo@...hat.com>,
        Namhyung Kim <namhyung@...nel.org>,
        Tom Zanussi <tom.zanussi@...ux.intel.com>,
        Arnaldo Carvalho de Melo <acme@...nel.org>,
        linux-trace-users@...r.kernel.org, linux-kselftest@...r.kernel.org,
        shuah@...nel.org
Subject: Re: [PATCH v5 16/19] tracing: probeevent: Add array type support

On Thu, 15 Mar 2018 11:18:19 +0530
Ravi Bangoria <ravi.bangoria@...ux.vnet.ibm.com> wrote:

> Hi Masami,
> 
> On 03/08/2018 02:20 PM, Masami Hiramatsu wrote:
> > Add array type support for probe events.
> > This allows user to get arraied types from memory address.
> > The array type syntax is
> >
> > 	TYPE[N]
> >
> > Where TYPE is one of types (u8/16/32/64,s8/16/32/64,
> > x8/16/32/64, symbol, string) and N is a fixed value less
> > than 64.
> >
> > The string array type is a bit different from other types. For
> > other base types, <base-type>[1] is equal to <base-type>
> > (e.g. +0(%di):x32[1] is same as +0(%di):x32.) But string[1] is not
> > equal to string. The string type itself represents "char array",
> > but string array type represents "char * array". So, for example,
> > +0(%di):string[1] is equal to +0(+0(%di)):string.
> 
> I was trying to test this patch. But I'm not getting proper data.
> Here is what I'm doing...
> 
> $ cat hello.c
> #include <stdio.h>
> 
> void foo1(int a[])
> {
>     printf("%d\n", a[2]);
> }
> 
> void foo(int a[])
> {
>     printf("%d\n", a[1]);
>     foo1(a);
> }
> 
> void main()
> {
>     int a[3] = {4, 5, 6};
>     printf("%d\n", a[0]);
>     foo(a);
> }
> 
> $ gcc -g hello.c -o hello
> $ sudo ./perf probe -x ~/hello foo1 'a=a:x32[3]'
> $ sudo cat /sys/kernel/debug/tracing/uprobe_events
> p:probe_hello/foo1 /home/ravi/hello:0x00000000000005fc a=+96(%gpr31):x32[3]
> 
> $ sudo ./perf record -e probe_hello:foo1 ~/hello
> $ sudo ./perf script
> hello  6913 [038]  2857.704470: probe_hello:foo1: (100005fc) a={0xd69e4400,0x7fff,0x0}
> 
> 
> I don't see proper values of the 'a'? Anything wrong with my perf commands :) ?

Ah, I guess since it is a pointer. in main() function, "a" is an array,
but in foo1(), "a" is a pointer to the array.
>From the viewpoint of C source code, both pointer and array is same
expression, but actually it is not same from memory point of view.
perf probe has to be updated to enable it too.

Could you please try to do as below?

$ echo "p:probe_hello/foo1 /home/ravi/hello:0x00000000000005fc a=+0(+96(%gpr31)):x32[3]" | sudo tee /sys/kernel/debug/tracing/uprobe_events

and trace it again.

-- 
Masami Hiramatsu <mhiramat@...nel.org>