10:45 <NebuK_work> yoho

10:45 <NebuK_work> maybe you guys know a bit more about elfutils / libdw as needed by systemtap

10:46 <NebuK_work> for some reason on one of my systems, not everything from libdwfl is built, it seems dwfl_linux_kernel_report_offline is missing

12:16 <mjw> NebuK_work, That is weird, do you have more information how that could have happened?

12:16 <mjw> which version, how did you configure, etc.?

12:52 <mjw> BTW. For people playing along at home ^. We discussed on #elfutils a bit. The issue is a "patch" in openwrt that just completely rips out dwfl_linux_kernel_report_offline (but leaves the header intact...)

12:52 <mjw> https://github.com/lede-project/source/blob/77c25c2dd164c28021ca0095a111adc7acc29e62/package/libs/elfutils/patches/101-no-fts.patch

12:52 <mjw> So basically that is an unusable elfutils build...

12:54 <mjw> Note how they just remove linux-kernel-modules.c from the build (but leave the file itself) and rip out random calls to dwfl_linux_kernel_report... So I have no idea how anything works on that setup.

13:13 <fche> funkyh

13:32 <NebuK_work> and i really would’ve liked to include systemtap … it’d ease kernel debugging on the target devices so much

13:33 <fche> you can build systemtap with an elfutils source tree of your own.

13:34 <fche> % configure --with-elfutils=ELFUTILS_SOURCE_TREE ...

13:34 <mjw> but they use "musl" so it won't build out of the box...

13:34 <mjw> for some reason they did submit some of their patches, but not all...

13:35 <mjw> Well, this crazy one we obviously wouldn't have accepted.

13:37 <NebuK_work> hm i guess what i’ll have to do if the stubbing of the missing functions doesn’t work is provide some scripting to build systemtap on the „real“ host, but staprun for the target (staprun ~only~ compiles fine with musl)

13:48 <NebuK_work> sorry, re

13:49 <fche> can you build your own copy of stap from unbroken sources?

13:54 <NebuK_work> outside of the openwrt buildsystem yes

14:11 <NebuK_work> fche: or … did i misunderstand your question?

14:11 <fche> nope, that was it -

14:12 <fche> so in theory one can still use the tool, even if the distro copy is broken

14:13 <NebuK_work> yeah, it’s unproblematic on the OS-host-side (not openwrt-host-side)

17:58 <jistone> NebuK_work, I guess you left #elfutils, so I'll copy what I just found:

17:58 <jistone> <jistone> http://wiki.musl-libc.org/wiki/FAQ#Q:_why_is_fts.h_not_included_.3F

17:58 <jistone> <jistone> references glibc 15838 for want of fts64, which was closed dupe of 11460, which mjw fixed!

17:58 <jistone> <jistone> I also found https://github.com/pullmoll/musl-fts

18:00 <fche> heh

18:00 <fche> way to go mark

18:11 <irker831> systemtap: fche systemtap.git:refs/heads/master * release-3.0-301-gd1025fe / NEWS man/stapprobes.3stap: NEWS: mention that .return { $var } should be replaced with @entry($var) http://tinyurl.com/gsyt86f

19:06 <irker831> systemtap: mcermak systemtap.git:refs/heads/master * release-3.0-302-gb8d6168 / configure.ac initscript/99stap/check.in initscript/99stap/install.in initscript/99stap/start-staprun.sh initscript/systemtap.in systemtap.spec: Fix the boot time probing feature for RHEL6 and Fedora19+ systems. http://tinyurl.com/zooaxrx

19:07 <irker831> systemtap: mcermak systemtap.git:refs/heads/master * release-3.0-303-g9ed4785 / Makefile.in aclocal.m4 configure doc/Makefile.in doc/SystemTap_Tapset_Reference/Makefile.in doc/beginners/Makefile.in java/Makefile.in man/Makefile.in man/cs/Makefile.in stapdyn/Makefile.in staprun/Makefile.in: Autoreconf following previous commit b8d6168f1f8d8a303b0d. http://tinyurl.com/j8t6yrd