PERFORCE change 122892 for review

[Ulf Lilleengen]

On tor, jul 05, 2007 at 12:01:40 +0000, Sonja Milicic wrote:
> http://perforce.freebsd.org/chv.cgi?CH=122892
> 
> Change 122892 by smilicic at tanarri_marilith on 2007/07/05 00:01:06
> 
> 	fixed a bug with log provider creation and added more details to dumpconf
> 
> Affected files ...
> 
> .. //depot/projects/soc2007/smilicic_glog/sys/geom/log/glog.c#4 edit
> Differences ...
> 
> ==== //depot/projects/soc2007/smilicic_glog/sys/geom/log/glog.c#4 (text+ko) ====
* SNIP *
>  g_log_worker_sleep(struct g_log_softc *sc)
>  {
> -	if (g_log_no_events(&sc->sc_events))
> -		tsleep(sc, PRIBIO, "glogidle", hz); 
> +	if (g_log_no_events(&sc->sc_events)){
> +		G_LOG_DEBUG(0, "putting worker to sleep");
> +		tsleep(sc, PRIBIO, "glogidle", hz);
> +	}
>  }
>  
>  /*worker thread*/
> @@ -380,6 +419,7 @@
>  		panic("No softc!");
You should consider to use KASSERT for this, as you have done en log_post_event.
>  	
>  	while (1){
> +                G_LOG_DEBUG(0,"working...");
>  		ev = g_log_get_event(&sc->sc_events);
> +                g_log_stop(sc->sc_geom_log, *force);
> +        else 
> +                panic("Softc is null in ctl_destroy!");
Also a candidate.

Also, I noticed that a lot of places you mix whitespaces and tabs. It would be
nice if you just used one of them. Anyway, thought I'd just see how the other
GEOM student is doing :) Your work looks good so far!

-- 
Ulf Lilleengen