home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!uknet!root44!hrc63!mrcu!ajh
- From: ajh@uk.co.gec-mrc (Andrew Haylett)
- Newsgroups: comp.os.linux
- Subject: Would a trace(1) facility be useful?
- Message-ID: <1866@oberon>
- Date: 30 Jul 92 09:25:24 GMT
- Organization: GEC-Marconi Research Centre, Great Baddow, UK
- Lines: 17
- X-Newsreader: Tin 1.1 PL4
-
- I've made substantial use in the past of Sun's trace(1) facility for tracing
- systems calls and arguments from user code. Admittedly, it is more essential in
- a binary-only system, but such a facility could still be useful and interesting
- for Linux.
-
- To implement this would probably be fairly tricky, and I'm not up to it. We
- have a fairly complete ptrace(2), but I think it would need to support the
- PTRACE_SYSCALL function to trap on system calls. The PD 'strace' package by
- Paul Kranenburg could form a starting point for a trace implementation, though
- it was written for Suns and would probably have to be disembowelled. There may
- be simpler code lying around.
-
- And whether any of this is POSIX, or just a Sun-ism, I don't know at all. Any
- comments?
- --
- Andrew Haylett | Inet: ajh@gec-mrc.co.uk | Fax: +44 245 75244
- GEC-Marconi Research | Tel: +44 245 73331 x.3283 | Telex: 995016 GECRES G
-