13:25 <fche> tux3, :)

13:25 <fche> thanks mjw for explaining

13:31 <mjw> sorry, missed it on irc. Hope the bugzilla entry was useful

13:57 <fche> yup

14:54 <irker038> systemtap: smakarov systemtap.git:master * release-4.3-76-g21109b093 / runtime/linux/uprobes/Makefile runtime/linux/uprobes/uprobes.txt testsuite/systemtap.base/stap_hwbkpt_kmod.Makefile testsuite/systemtap.base/stap_kmodule.Makefile testsuite/systemtap.context/makefile1 testsuite/systemtap.context/makefile2 testsuite/systemtap.printf/memory1_module.Makefile: tentative fix: update hand-written Makefiles to use KBUILD_EXTMOD=

15:33 <irker038> systemtap: fche systemtap.git:master * release-4.3-77-gdf1fc1ddb / man/stapprobes.3stap: man stapprobes.3stap: add @var("var","module") docs

20:02 <agentzh> fche: what do you think of this patch? https://gist.github.com/agentzh/2eb773f23ed65122c8fc07baf32cbe27

20:02 <agentzh> for PR26673: https://sourceware.org/bugzilla/show_bug.cgi?id=26673

20:03 <fche> moved into the context structure, sounds good.

20:03 <agentzh> k, so good to commit?

20:03 <fche> yup

20:03 <fche> assuming it works :

20:03 <agentzh> cool, thanks, will do.

20:03 <fche> :)

20:03 <agentzh> yep, it works.

20:04 <agentzh> i'll attach some tests to it.

20:38 <agentzh> fche: added a test case to cover that kernel bt fix: https://gist.github.com/agentzh/43abe9f60a5754584f5a8c84c8e387c7

20:38 <agentzh> this is the v2 patch.

20:39 <agentzh> passing on my side.

20:39 <fche> nice

20:58 <irker038> systemtap: yichun systemtap.git:master * release-4.3-78-g3674b8d35 / runtime/common_probe_context.h runtime/stack.c testsuite/systemtap.context/max-kernel-bt.exp testsuite/systemtap.context/max-kernel-bt_1.stp: PR26673: ko gcc compilation error would happen when -DMAXBACKTRACE=N and N > 60

21:04 <agentzh> fche: what do you think of this doc patch? https://gist.github.com/agentzh/099dd1e8c57cde2601e4f83d2c60a598

21:06 <agentzh> where did you add? a quick grep didn't reveal anything.

21:06 <fche> man stapprobes

21:07 <agentzh> ah okay. i only grepped doc/.

21:07 <agentzh> that's why

21:08 <agentzh> will commit the doc/ changes.

21:09 <irker038> systemtap: yichun systemtap.git:master * release-4.3-79-g46cd7e5d0 / doc/SystemTap_Beginners_Guide/en-US/ScriptConstructs.xml doc/langref.tex: Doc: documented the @var(NAME, EXE) notation in the beginners guide and langref docs

21:10 <agentzh> fche: i wonder if you'll also welcome a non-dwarf kernel backtrace dumper.

21:10 <agentzh> similar to the dump_stack() kernel C func but not to dmesg.

21:11 <fche> um, don't we already call into the kernel backtracer as a fallback?

21:11 <agentzh> it didn't give meaningful backtraces.

21:11 <agentzh> far from dump_stack().

21:12 <agentzh> am i missing anything here?

21:14 <agentzh> fche: i filed a PR for it in the past: https://sourceware.org/bugzilla/show_bug.cgi?id=25853

21:16 <fche> hm

21:16 <fche> the warning message gives the very tip you need, doesn't it?

21:16 <fche> stap -d kernel ?

21:16 <fche> I believe those "(inexact)" lines are coming from the kernel backtracer

21:16 <fche> but aren't mapped to symbol names for some reason

21:18 <agentzh> fche: tried -d kernel but still no improvement.

21:18 <agentzh> so i thought it was irrelevant.

21:18 <agentzh> i can only get better backtrace with kernel dwarf available.

21:19 <agentzh> but dump_stack() works beautifully.

21:19 <agentzh> in the same system.

21:20 <fche> well anyway the code is already there to back down to the kernel, IIRC

21:20 <fche> tuning it to make it more complete would be great

21:20 <agentzh> okay, will do.

21:20 <agentzh> my colleage kerneltoast will look into it. he's not around atm.

21:21 <agentzh> *colleague

21:23 <irker038> systemtap: alizhang systemtap.git:azhang/pr13838 * release-4.3-85-g0d1e23cb5 / runtime/softfloat.c runtime/softfloat.h runtime/softfloat/internals.h runtime/softfloat/primitives.h: add more helper functions declaratio and definition so that program can compile