Re: Inserting delays with systemtap

["Bryn M. Reeves" <bmr at redhat dot com>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 03/07/2012 03:23 PM, Frank Ch. Eigler wrote:
> bmr wrote:
>> Would it be worth having something like this in a tapset?
> 
> Absolutely.  Would you mind writing it up as a patch against some
> new file like .../tapset/guru-delay.stp?

Happy to - I'll try to get something done today or tomorrow.

>> The requirement to use STP_NO_OVERLOAD for long-running delays
>> [...]
> 
> ... is a problem.  It may be possible to automagically disable 
> overload processing if any of these functions are used, by
> including:
> 
> %{ /* guru */ #undef STP_OVERLOAD %}

Ah, thanks - I tried adding -DSTP_NO_OVERLOAD to the shabang line
along with -g (which worked) but I appear to hit one of the safety
regexes and it's rejected:

# head -1 ./cdc-acm-test.stp
#!/usr/bin/stap -DSTP_NO_OVERLOAD -g

# ././cdc-acm-test.stp
ERROR: Safety pattern mismatch for -D parameter ('STP_NO_OVERLOAD -g'
vs. '^[a-z_][a-z_0-9]*(=-?[a-z_0-9]+)?$') rc=1

In some ways I didn't think this was so bad - I guess it's definitely
a feature you want to know what you're doing with!

Cheers,
Bryn.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk9Xfi4ACgkQ6YSQoMYUY95p/QCeOLMfiu2t1nruTJNKTo9GRM/h
GZMAoJPHyjCVOqdT8s6k6pccLXP/4iLg
=o/Pc
-----END PGP SIGNATURE-----