Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 22 Nov 2006 21:42:04 +0000
From:      John Birrell <jb@what-creek.com>
To:        "Devon H. O'Dell" <devon.odell@gmail.com>
Cc:        freebsd-current@freebsd.org, "Cai, Quanqing" <caiquanqing@gmail.com>, Kris Kennaway <kris@obsecurity.org>
Subject:   Re: KDTRACE is gone?
Message-ID:  <20061122214203.GA48004@what-creek.com>
In-Reply-To: <9ab217670611221016q5bd1bf84v4ef878391eb2a67a@mail.gmail.com>
References:  <2b22951e0611212109t69b01400q5eb0ba15b028ce68@mail.gmail.com> <20061122051359.GA42639@what-creek.com> <20061122165735.GA37930@xor.obsecurity.org> <1164216949.18579.19.camel@buffy.york.ac.uk> <9ab217670611221016q5bd1bf84v4ef878391eb2a67a@mail.gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, Nov 22, 2006 at 01:16:17PM -0500, Devon H. O'Dell wrote:
> Yeah, it could be done. However, the DTrace provider (providing BEGIN,
> END, and ERROR, and code that allows for other providers to hook in)
> is > 13,000 lines of code and comments, so it'd be a very non-trivial
> task.

The DTrace provider isn't actually relevent to this discussion. That
is find staying as a kernel module and CDDL'd.

The issue is the hooks that the DTrace modules register with. These
are very small bits of BSD licensed code that I write, and which
get added inline to existing BSD licensed sources like trap.c. These
little bits of code tend to require calling the DTrace kernel API or
methods and techniques defined by CDDL headers and not documented
anywhere else.

I have been very careful not to copy CDDL stuff into FreeBSD headers
for fear of causing FreeBSD legal problems in future like the
claims being made in the SCO vs IBM (Linux) litigation. Remember that
the OpenSolaris source is based on System V and Sun Microsystems have
extended it to include DTrace.

I need to find an alternative to the KDTRACE implementation -- that's
why I removed it for now.

I want to come up with a functional equivalent to the Solaris DTrace
implementation. Having it as an option (amongst a heap of other 
options) that a user /might/ choose to compile in, isn't a good
design IMO.

--
John Birrell



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20061122214203.GA48004>