cvs commit: src/usr.bin/pkill pkill.1 pkill.c

Garance A Drosehn gad at FreeBSD.org
Wed Aug 24 22:14:19 GMT 2005


At 10:56 PM +0200 8/24/05, Pawel Jakub Dawidek wrote:
>On Wed, Aug 24, 2005, John Baldwin wrote:
>+> On Wednesday 24 August 2005, Pawel Jakub Dawidek wrote:
>+> >   Log:
>+> >   Modify '-F' option to work nicely with pidfile(3) - a pidfile
>+> >   given as an argument has to be locked.
>+> >
>+>
>+> I guess the only thing that makes me uncomfortable about this
>+> change is that it means pkill won't work with 3rd party program's
>+> pid files since they are probably not using pidfile(3).  Maybe
>+> have some way of turning off the flock requirement?

I must admit that I did not quite understand what you had meant in
your earlier messages about changes to `pkill' to match your new
pid-handling library routines.  I agree with John, this change
doesn't seem quite right to me.

>Some daemons, even if they don't use pidfile(3), they do use flock(2)
>to lock the pidfile (eg. cron(8)).  I added this option to pkill(1)
>in the past, but I wasn't very happy with it until today...
>
>On the other hand, I added it to avoid: kill `cat /var/run/daemon.pid`...
>
>Hmm, what do you think about reverting '-F' behaviour and make it
>use flock(2) only when '-L' is also given?

Something like that would be good, I think.  The previous change
was useful as it was, even though I agree it is even more useful
for pkill to pay attention to file-locking when it can.

I'd slightly prefer something other than adding another option,
but I can't think of anything which is particularly better.  The
only other suggestion might be to have "daemon.lpid" filenames
instead of "daemon.pid", but that also seems icky to me.

-- 
Garance Alistair Drosehn     =      gad at gilead.netel.rpi.edu
Senior Systems Programmer               or   gad at FreeBSD.org
Rensselaer Polytechnic Institute;             Troy, NY;  USA


More information about the cvs-src mailing list