Re: Results of systemtap-20071124 snapshot on s390x

[David Smith <dsmith at redhat dot com>]

David Wilder wrote:
> David Smith wrote:
>> David Wilder wrote:
>>> David Smith wrote:
>>>> David Wilder wrote:
>>>>> FAIL: K_MARKER04 compilation failed
>>>> I've been meaning to ask about this for awhile, so here goes.  What
>>>> kernel is this using (and does this kernel have marker support)?  If
>>>> your kernel has markers, this could fail if the first marker found has
>>>> no arguments.  If your kernel doesn't have marker support, this test
>>>> shouldn't be run.
>>>>
>>> I forgot to put the kernel version in the test report.  I am running
>>> 2.4.24-rc3.  And CONFIG_MARKERS is "y".  I plan to take a look at the
>>> failure in the next day or so.  I will let you know what I find.
>>
>> Since your kernel does have markers, then I'll bet the first marker that
>> the test found doesn't actually have an argument (K_MARKER04 tests
>> whether the first argument can be accessed).
>>
>> Can you post the K_MARKER04 section of systemtap.log?
>>
>> If your problem is that the first marker found doesn't have an argument,
>> I'll need to look at changing the test somehow.
>>
> 
> Pass 1: parsed user script and 38 library script(s) in
> 480usr/0sys/501real ms.^M
> Pass 2: analyzed script: 1 probe(s), 1 function(s), 0 embed(s), 0
> global(s) in 20usr/20sys/42real ms.^M
> Pass 3: translated to C into
> "/tmp/stapsIbnv3/stap_be214f08543be97af4999459e36b103f_387.c" in
> 0usr/0sys/1real ms.^M
> Pass 4: compiled C into "stap_be214f08543be97af4999459e36b103f_387.ko"
> in 1170usr/300sys/1485real ms.^M
> Pass 4: compilation failed.  Try again with more '-v' (verbose) options.^M
> FAIL: K_MARKER04 compilation failed
> Pass 1: parsed user script and 38 library script(s) in
> 480usr/0sys/499real ms.^M
> semantic error: invalid marker argument number: identifier '$arg200' at
> <input>:1:49^M
> semantic error: no match for probe point while resolving probe point
> kernel.mark("core_marker_format")^M
> Pass 2: analyzed script: 0 probe(s), 0 function(s), 0 embed(s), 0
> global(s) in 20usr/30sys/44real ms.^M
> Pass 2: analysis failed.  Try again with more '-v' (verbose) options.^M

Interesting.  The problem isn't that the marker the test used didn't
have an argument.  Otherwise, you would have gotten an error that looked
like the error message you got above (that comes from K_MARKER05). For
instance, here's what I get when I try to compile a probe that uses
'$arg1' when a marker doesn't have any arguments.

  # stap -e 'probe kernel.mark("mm_handle_fault_exit") { print($arg1)}'
  semantic error: invalid marker argument number: identifier '$arg1'
  at <input>:1:51 while resolving probe point
  kernel.mark("mm_handle_fault_exit")
  Pass 2: analysis failed.  Try again with more '-v' (verbose) options.

What's weird is that only K_MARKER04 is failing.  K_MARKER01 and
K_MARKER02 compile correctly, yet K_MARKER04 doesn't compile.  The only
difference between K_MARKER02 and K_MARKER04 is that K_MARKER04 prints
$arg1.

Can you find a marker that has an argument and do a stap command like
this and report the output?

  # stap -vvv -p4 -e 'probe kernel.mark("FOO") { print($arg1) }'

-- 
David Smith
dsmith@redhat.com
Red Hat
http://www.redhat.com
256.217.0141 (direct)
256.837.0057 (fax)