Date: Wed, 22 Apr 2020 06:03:11 +0300 From: Yuri Pankov <ypankov@fastmail.com> To: Benjamin Kaduk <kaduk@mit.edu> Cc: freebsd-hackers@freebsd.org Subject: Re: "most interesting" process in w(1) Message-ID: <a4199844-36a6-6274-dfe3-84bddad45652@fastmail.com> In-Reply-To: <20200422024436.GA27494@kduck.mit.edu> References: <b43e40b2-c2a0-cf55-a932-ab2e0cefdcdb@fastmail.com> <20200422024436.GA27494@kduck.mit.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
Benjamin Kaduk wrote:
> On Wed, Apr 22, 2020 at 12:09:48AM +0300, Yuri Pankov wrote:
>> Looking at how w(1) finds "most interesting" process for terminal, I
>> noticed the following code which looks strange to me (usr.bin/w/w.c,
>> line 360 in HEAD)
>>
>> for (ep = ehead; ep != NULL; ep = ep->next) {
>> if (ep->tdev == kp->ki_tdev) {
>> /*
>> * proc is associated with this terminal
>> */
>> if (ep->kp == NULL && kp->ki_pgid == kp->ki_tpgid) {
>> /*
>> * Proc is 'most interesting'
>> */
>> if (proc_compare(ep->kp, kp))
>> ep->kp = kp;
>> ...
>> }
>> }
>> }
>>
>> Given the (ep->kp == NULL) check, proc_compare() becomes no-op, it will
>> always select kp, and that's the only place we ever set ep->kp, so the
>> first matching process is always "most interesting". If that's really
>> what we want, we could do without the proc_compare() call. What am I
>> missing here?
>
> I suspect it makes more sense as a "!=" than a "==" (with ep->kp = kp;
> always occurring for the "== NULL" case).
(ep->kp == NULL) check seems to be unrelated addition in r29310 (!), the
code around it gives no hints on why it was added.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?a4199844-36a6-6274-dfe3-84bddad45652>