re: missing SDT_PROVIDER_DEFINE(sdt)

[matthew green <mrg%eterna.com.au@localhost>]

Taylor R Campbell writes:
>    Date: Fri, 1 Apr 2016 12:29:19 +0900
>    From: Kengo NAKAHARA <k-nakahara%iij.ad.jp@localhost>
> 
>    There is SDT_PROVIDER_DECLARE(sdt) in sdt.h.
>        http://nxr.netbsd.org/xref/src/sys/sys/sdt.h#416
>    However, there is no SDT_PROVIDER_DEFINE(sdt). I think this is
>    inconsistent.
> 
>    Currently, "sdt" provider is not used yet, so this inconsistency doesn't
>    cause problems. "sdt" provider will be used by ipf if it enable sdt.
>    That is, when it enabled the following macros.
>        http://nxr.netbsd.org/xref/src/sys/external/bsd/ipf/netinet/ip_compa=
> t.h#2831
> 
>    I think the following patch is required to prevent future problems such
>    as ipf enabled sdt.
> 
> I'm actually a little puzzled by all that.  Why doesn't ipf just
> define its own provider?  ipf appears to be the only use of the sdt
> provider.  I'd sooner get rid of DTRACE_PROBE* altogether and just
> adapt ipf to use its own provider.  But I have no objection if you'd
> rather just add an SDT_PROVIDER_DEFINE(sdt).

i assume this was because ipf was ported to solaris and solaris
provides a (generic?) "sdt" provider, not that it has been ported
to netbsd dtrace in any way.



.mrg.