Re: [PATCH] instruction tracing probe

[Dave Nomura <dcnltc at us dot ibm dot com>]

I'll check out the man pages.  I forgot to look for them.

I wasn't sure how to make a test case that could be run in a suite due 
to synchorinization problems:
1) user program must be started first so that it's pid can be known to 
the systemtap session.
I could use the pathname variant for taskfinder probes, but the user 
probes that I need to turn  the probe on/off still don't support  
pathnames.
2) the user program has to  be delayed long enough to let the systemtap 
session to start up and do some itrace initialiation.

if you have any suggestions about how to work around these limitations 
in creating a test case or maybe an example test case that already deals 
with these issues I could try and put something together.

As for checking the code in, I don't think I have write permission.  
What is the process for getting the patch committed?

David Smith wrote:
> Dave Nomura wrote:
>   
> > This patch set implements user space instruction tracing by the
> > following itrace probe:
> >     
> ...
>
> Based on a quick reading of the code, it looks reasonable to me.  It is
> missing a few things (that shouldn't keep you from checking it in) like
> testcases and additions to the man pages.
>
> We probably should also look at places where we could share more code
> between the itrace and utrace probes (such as when validating pathnames).
>
>   


--
Dave Nomura
LTC Linux Power Toolchain