This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
Re: binary tracing
- From: Tom Zanussi <zanussi at us dot ibm dot com>
- To: Jian Gui <guijian at cn dot ibm dot com>
- Cc: Tom Zanussi <zanussi at us dot ibm dot com>, "Martin Hunt" <hunt at redhat dot com>, systemtap at sources dot redhat dot com
- Date: Fri, 7 Apr 2006 12:08:10 -0500
- Subject: Re: binary tracing
- References: <17461.12179.509628.176681@tut.ibm.com> <OFCAFEA37D.869B6B23-ON48257149.000689F1-48257149.000BBDFF@cn.ibm.com>
Jian Gui writes:
> > The attached tarball contains a patch to systemtap - since it doesn't
> > use bin_write(), it doesn't include bin_write.c, and it doesn't
> > actually make any changes to the runtime other than to comment out a
> > vlog print that was getting in the way of parsing the data in stpd.
> >
> You mentioned to let the translator generate the corresponding tracing
> code and not to use bin_write(). Are you considering to totally remove
> bin_write() interface from runtime? or keep it but add the ability to
> do binary tracing at script level?
No, I wasn't suggesting removing bin_write() - in fact I think Martin
is already rolling it in to the runtime.
I was considering this as an optimization - if there are only binary
specifiers in the printf() statement, the translator might decide to
generate code to write the data into the buffer directly instead of
calling bin_write() to do it.
Tom