[Bug testsuite/18971] New: process_by_pid.exp issues

["mcermak at redhat dot com" <sourceware-bugzilla at sourceware dot org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=18971

            Bug ID: 18971
           Summary: process_by_pid.exp issues
           Product: systemtap
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: testsuite
          Assignee: systemtap at sourceware dot org
          Reporter: mcermak at redhat dot com
  Target Milestone: ---

Created attachment 8612
  --> https://sourceware.org/bugzilla/attachment.cgi?id=8612&action=edit
proposed partial patch

The process_by_pid.exp fails thusly:

=======
Running ./systemtap.base/process_by_pid.exp ... 
Executing on host: gcc ./systemtap.base/process_by_pid.c  -g -isystem.
-isystem/root/mcermak-systemtap/systemtap-build/include  -lm   -o
./process_by_pid    (timeout = 300)
spawn -ignore SIGHUP gcc ./systemtap.base/process_by_pid.c -g -isystem.
-isystem/root/mcermak-systemtap/systemtap-build/include -lm -o
./process_by_pid^M
starting the stap script
executing: stap -v ./systemtap.base/process_by_pid.stp 30007 process_by_pid
spawn stap -v ./systemtap.base/process_by_pid.stp 30007 process_by_pid^M
Pass 1: parsed user script and 110 library script(s) using
111128virt/31924res/3016shr/29492data kb, in 580usr/20sys/598real ms.^M
Pass 2: analyzed script: 22 probe(s), 3 function(s), 3 embed(s), 1 global(s)
using 112620virt/34804res/4168shr/30984data kb, in 990usr/510sys/1391real ms.^M
Pass 3: translated to C into
"/tmp/stapvzdqbe/stap_e2f09c9e92c337751f6b1c086dc5582b_11442_src.c" using
112620virt/35172res/4520shr/30984data kb, in 20usr/120sys/138real ms.^M
Pass 4: compiled C into "stap_e2f09c9e92c337751f6b1c086dc5582b_11442.ko" in
3070usr/590sys/3425real ms.^M
Pass 5: starting run.^M
systemtap starting probe^M
PASS: process_by_pid startup
PASS: process_by_pid load generation
Executing: kill -INT -30023
systemtap ending probe^M
pass^M
PASS: process_by_pid shutdown and output
Pass 5: run completed in 10usr/90sys/20839real ms.^M
Executing: kill -INT -30023
starting the stap script
executing: stap -v --runtime=dyninst ./systemtap.base/process_by_pid.stp 30007
spawn stap -v --runtime=dyninst ./systemtap.base/process_by_pid.stp 30007^M
Pass 1: parsed user script and 40 library script(s) using
82904virt/4392res/2980shr/1268data kb, in 20usr/0sys/23real ms.^M
Pass 2: analyzed script: 22 probe(s), 3 function(s), 1 embed(s), 1 global(s)
using 83160virt/5452res/3804shr/1524data kb, in 10usr/0sys/6real ms.^M
Pass 3: translated to C into
"/tmp/stapqGQlb1/stap_052540e811cde79b472386b9861faa00_8992_src.c" using
83268virt/6060res/4264shr/1632data kb, in 20usr/140sys/180real ms.^M
Pass 4: compiled C into "stap_052540e811cde79b472386b9861faa00_8992.so" in
1310usr/40sys/1360real ms.^M
Pass 5: starting run.^M
stapdyn: WARNING: process probes require a target (-c or -x)^M
systemtap starting probe^M
FAIL: process_by_pid (dyninst) startup (timeout)
Executing: kill -INT -30463
Executing: kill -INT -30463
Executing: kill -INT 30005
Executing: kill -KILL 30005
kill: kill: sending signal to 30005 failed: No such process
Executing: kill -INT 30007
Executing: kill -KILL 30007
kill: kill: sending signal to 30007 failed: No such process
Executing: kill -INT 30009
Executing: kill -KILL 30009
kill: kill: sending signal to 30009 failed: No such process
testcase ./systemtap.base/process_by_pid.exp completed in 215 seconds

                === systemtap Summary === 

# of expected passes            3
# of unexpected failures        1
runtest completed at Wed Sep 16 09:43:42 2015
=======

It's hard to see what expect does for dyninst runtime, since in this case
RUNTESTFLAGS='-d' doesn't work. To quickly see what expect reads following can
be used:

=======
-re {^.+\r\n} {
  parray expect_out
  exp_continue
}
=======

1) TC fails since in dyninst mode -c or -x is needed (stapdyn: WARNING: process
probes require a target (-c or -x)) (PR17352).

2) TC fails since stap_run() gets unexpected warnings similar to "stapdyn:
WARNING: Couldn't instrument the function containing" ...

3) TC fails since plt probes aren't being hit in dyninst mode (PR17359).

Attached patch might solve 1) and 2). For 3), the fail might possibly be marked
as KFAIL?

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