[Bug translator/18191] New: DWARF-based sizeof(), array_size(), typeof() operators

[Bug translator/18191] New: DWARF-based sizeof(), array_size(), typeof() operators

[dsmith at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=18191

            Bug ID: 18191
           Summary: DWARF-based sizeof(), array_size(), typeof() operators
           Product: systemtap
           Version: unspecified
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: translator
          Assignee: systemtap at sourceware dot org
          Reporter: dsmith at redhat dot com

As mentioned by Stephane Chazelas on the mailing list:

<https://sourceware.org/ml/systemtap/2015-q1/msg00236.html>

It might be nice to have DWARF-based sizeof(), array_size(), typeof()
operators.

-- 
You are receiving this mail because:
You are the assignee for the bug.

[Bug translator/18191] DWARF-based sizeof(), array_size(), typeof() operators

[stephane.chazelas+sourceware at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=18191

Stephane Chazelas <stephane.chazelas+sourceware at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |stephane.chazelas+sourcewar
                   |                            |e at gmail dot com

--- Comment #1 from Stephane Chazelas <stephane.chazelas+sourceware at gmail dot com> ---
[Thanks David]

I don't know about the other ones, but a DWARF-based sizeof() (at least for
kernel symbols) should just be a matter of:

function sizeof:long (addr:long) %{ /* pure */ /* pragma:symbols */
  STAP_RETVALUE = -1;
  _stp_kallsyms_lookup(STAP_ARG_addr, (unsigned long*)&(STAP_RETVALUE), NULL,
NULL, NULL);
%}

See
http://stackoverflow.com/questions/29034267/get-size-of-target-array-in-systemtap
for where the request is coming from.

-- 
You are receiving this mail because:
You are the assignee for the bug.

[Bug translator/18191] DWARF-based sizeof(), array_size(), typeof() operators

[fche at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=18191

Frank Ch. Eigler <fche at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |fche at redhat dot com

--- Comment #2 from Frank Ch. Eigler <fche at redhat dot com> ---
(In reply to Stephane Chazelas from comment #1)

> function sizeof:long (addr:long) %{ /* pure */ /* pragma:symbols */
>   STAP_RETVALUE = -1;
>   _stp_kallsyms_lookup(STAP_ARG_addr, (unsigned long*)&(STAP_RETVALUE),
> NULL, NULL, NULL);
> %}

Note that the above great start is not actually DWARF-based - it's just
symbol-table based.  We might want to do a few things extra:

- rename to symsize (since that's what it does)

- throw an error if the address can't be resolved to a known symbol, instead of
returning a sentinel -1

- having a user-space variant usymsize() that passes 'current' as the last
parameter to the underlying lookup function

-- 
You are receiving this mail because:
You are the assignee for the bug.