01:38 <tonyj> fche: around?

01:44 <fche> hey tony

01:44 <tonyj> why on power64le (stap v3.0) am I getting a bunch of i915 stuff from systemtap.base/tracepoints.stp ?

01:45 <tonyj> hey!

01:45 <tonyj> I end up with /tmp/stapFymGED/stap_21182_aux_82.c which contains #include </lib/modules/4.4.16-56-default/source/drivers/gpu/drm/i915/i915_trace.h>

01:47 <tonyj> this ends up including include/linux/intel-iommu.h but only x86 provides a defn of clflush_cache_range for __iommu_flush_cache()

01:48 <fche> probably some kernel tracepoints that are not declared completely in include/trace/FOO.h files, so stap has to add some heuristic #include's

01:49 <fche> dsmith tweaked possibly-related code back in commit 18b1399443

01:50 <tonyj> this isn't a self built kernel

01:51 <fche> understood; that patch touched i915 header selection though

01:51 <fche> so what exactly is the problem with __iommu_flush_cache?

01:51 <tonyj> nothing defines it on power

01:51 <tonyj> or rather nothing defines clflush_cache_range()

01:54 <tonyj> oic the change, i915 isn't even config enabled on ppc

01:55 <fche> still curious about the actual user-visible problem related to this

01:57 <tonyj> the test fails

01:57 <tonyj> 1 sec

01:59 <tonyj> http://pastebin.com/2E0VtWCY

01:59 <fche> ok but that test case is not really user-visible; are all tracepoints affected badly ?

01:59 <fche> or any?

01:59 <tonyj> we have a bug that the kernel panics on power, so I was trying to run that testsuite

02:02 <tonyj> how the customer isn't seeing it I'm not grokking, they're getting:

02:02 <tonyj> Running ./systemtap.base/tracepoints.exp ...

02:02 <tonyj> Oops: Exception in kernel mode, sig: 5 [#1]

02:02 <tonyj> SMP NR_CPUS=2048 NUMA pSeries

02:03 <fche> a crash caused -by- the testsuite run? interesting

02:03 <tonyj> fche: yes but I can't even get it to compile

02:03 <fche> ISTR something vaguely like that in the s390 system (an s390 kernel bug really) that we papered over in stap by putting some tracepoints into a blacklist

02:04 <fche> ok ... pretty sure I won't have any good advice tonight :)

02:04 <fche> but

02:04 <fche> pls send your stap invocation, stap -vvv output, ... to the mailing list

02:04 <tonyj> right now I just want to see if I can run the testsuite :)

02:05 <tonyj> I will sent it tomorrow, late now.

02:05 <tonyj> thanks!

02:05 <fche> righto

05:24 <irker927> systemtap: mcermak systemtap.git:refs/heads/master * release-3.0-201-g73e7d9e / testsuite/Makefile.am testsuite/Makefile.in: Fix testsuite's list-unfinished makefile target http://tinyurl.com/za3bm74

12:15 <davidlt> hi

12:15 <davidlt> I am trying to run systemtap on centos 7.2, but keep getting: semantic error: missing x86_64 kernel/module debuginfo [man warning::debuginfo] under '/lib/modules/3.10.0-327.28.2.el7.x86_64/build'

12:15 <davidlt> urname -r: 3.10.0-327.28.2.el7.x86_64

12:16 <davidlt> Ah, somehow centos installed wrong version of kernel debug-info

12:32 <fche> run `stap-prep` if needed

15:16 <irker927> systemtap: flu systemtap.git:refs/heads/master * release-3.0-202-g05f925e / parse.cxx staptree.cxx staptree.h tapsets.cxx: move auto_path expansion to tapset resolution phase http://tinyurl.com/zwvxbfx

15:16 <irker927> systemtap: flu systemtap.git:refs/heads/master * release-3.0-203-gb9cc292 / main.cxx testsuite/systemtap.base/auto_path.c testsuite/systemtap.base/auto_path.exp testsuite/systemtap.base/tapset/PATH/tmp/stap_auto_path/a/ap1.stp testsuite/systemtap.base/tapset/PATH/tmp/stap_auto_path/a/b/ap2.stp testsuite/systemtap.base/tapset/PATH/tmp/stap_auto_path/b/ap3.stp: some auto_path test case http://tinyurl.com/ztgrhs4

15:24 <fche> hey felix, thanks, looking it over now

15:25 <irker927> systemtap: fche systemtap.git:refs/heads/master * release-3.0-204-g641e6d9 / runtime/stp_string.c runtime/stp_string.h runtime/vma.c: PR20433: simplify strcpy NULL-tolerance with inlined ?: "" http://tinyurl.com/jzylvvz

15:25 <irker927> systemtap: fche systemtap.git:refs/heads/master * release-3.0-205-g7017ba2 / session.cxx: stap startup message: tweak invitation to use '-i' interactive mode http://tinyurl.com/jb6aboq

16:02 <irker927> systemtap: flu systemtap.git:refs/heads/master * release-3.0-206-gc68fbc6 / parse.cxx staptree.cxx staptree.h tapsets.cxx: store auto_path string in probe point http://tinyurl.com/zxg69tt

19:09 <irker927> systemtap: csanting systemtap.git:refs/heads/master * release-3.0-207-g9a1b87b / man/stap.1.in runtime/vsprintf.c testsuite/systemtap.printf/print_user_buffer.c testsuite/systemtap.printf/print_user_buffer.exp testsuite/systemtap.printf/print_user_buffer.stp: PR15932: printf %m/%M user variant http://tinyurl.com/zl553g7

20:44 <ericlee_> Hi, I'm running a demo code here and always got following error:

20:44 <ericlee_> semantic error: unresolved target-symbol expression: identifier '$optname' at /usr/share/systemtap/tapset/linux/tcp.stp:614:32 source: optstr = tcp_ipv6_sockopt_str($optname) ^

20:44 <ericlee_> code: https://paste.fedoraproject.org/407151/34654147/

20:44 <ericlee_> any tgought?

21:01 <jistone> ericlee_, what kernel and systemtap versions do you have?

21:03 <ericlee_> jistone: kernel: 2.6.32-504.30.3.el6 stap: 2.7-2

21:04 <jistone> ok, let me try my VM

21:04 <ericlee_> jistone: thanks

21:07 <jistone> hmm, it's working fine here, but my kernel and stap are newer

21:08 <jistone> kernel-2.6.32-642.el6.x86_64 systemtap-2.9-4.el6.x86_64

21:08 <ericlee_> jistone: right, I run stap-prep and downloaded debuginfo and works! thanls

21:09 <jistone> oh, yeah, that would help!

21:09 <jistone> :)

21:09 <jistone> glad you figured it out, no help from me :)