Skip to content

RFC user space rt-safe printing and tracing

Jan Kiszka xenomai 1 version In review

Get series with b4
b4 am [email protected]

Versions newest first

  1. v1 RFC Latest2 patches 7 repliesIn review

Patches in v1 No applies recorded yet

#Subject and filesSizeRepliesApplied
1/2real-time print libraryconfigure.in examples/native/Makefile examples/native/rtprint.c +10 more+669 -37 replies
2/2examplary syscall instrumentationconfigure.in include/asm-generic/Makefile.am include/asm-generic/bits/bind.h +9 more+200 -330 replies

Reviews trailers per patch

No Reviewed-by, Acked-by or Tested-by trailers for v1 yet.

Read the v1 thread

Cover letter [Xenomai-core] [RFC][PATCH 0/2] user space rt-safe printing and tracing

Hi,

fresh from the development labs, guaranteed to be incomplete and lightly tested, here comes a patch series that addresses two issues bugging me (and maybe others) for quite some time:

1. How to print from RT (primary) context without stepping down to secondary mode?

2. How to trace Xenomai syscalls without hacking on strace _and_ I-pipe (ptrace for RT-syscalls) _and_ without suffering from its temporal intrusiveness (traced processes are stopped while examining their syscall arguments)?

The following two patches (against trunk) aim at answering both questions. The first one provides a generic rtprint library, the second one exploits it by instrumenting ridiculous subsets of libnative and libpthread_rt. Find details on design and usage will be found in the following mails.

Take this series as food for discussion about its usefulness, code organisation and remaining quirks, and also as an offer to early adopters. While I think that librtprint should be generally acceptable, I would like to gain some comments on my approach for per-process syscall tracing (and its impact on the skin libraries) first. Then, if this tracing is considered viable, helping hands for the boooooring instrumentation work would be *very* welcome!

Jan