This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
Re: [PATCH] filename in tapset nfs.proc.open and nfs.proc.release
- From: David Smith <dsmith at redhat dot com>
- To: "Frank Ch. Eigler" <fche at redhat dot com>
- Cc: hai wu <haiwu dot us at gmail dot com>, systemtap at sourceware dot org
- Date: Wed, 17 Dec 2014 09:11:19 -0600
- Subject: Re: [PATCH] filename in tapset nfs.proc.open and nfs.proc.release
- Authentication-results: sourceware.org; auth=none
- References: <CAJ1=nZcYtxopEs-XGj-Cs89dfzA2LF8oScnc_b+JE1zysg1T=Q at mail dot gmail dot com> <548F3F8D dot 8000900 at redhat dot com> <y0m4mswhcu8 dot fsf at fche dot csb>
On 12/15/2014 02:34 PM, Frank Ch. Eigler wrote:
> David Smith <dsmith@redhat.com> writes:
>
>> [...] Thanks for the patch, but I don't think this is really what
>> we want here. Here's why: [...]
>
> The full pathname is almost certainly more useful in the many probes
> where __file_filename() is currently used (I recall receiving some
> complaints about that long ago, just not getting to the bottom), and
> the cost is not -that- high. But point taken, some might not want to
> pay that.
>
> One way to have it both ways is to extend each these probe aliases
> (only some dozen or so) to define a new context variable like
> 'fullpath' sibling of 'file*name'. It would fully resolve the path
> name. (I'd certainly switch argstr over to print that instead of the
> last component.)
I'm OK with providing a 'fullpath' variable (which would get optimized
out if you didn't use it), but I'd need more convincing about the wisdom
of switching argstr over to using it. As you said above "some might not
want to pay that".
--
David Smith
dsmith@redhat.com
Red Hat
http://www.redhat.com
256.217.0141 (direct)
256.837.0057 (fax)