Re: [PATCH] cputime_t patches broke RLIMIT_CPU

[Roland McGrath]

> So would it not be nicer to just keep everything in seconds instead? 

Yes, that's how it was done before.  The patch I just posted was intended
to fix the apparent typo without getting any deeper.  Below is an untested
alternate patch to restore the old behavior under the new macro regime.

> Alternatively, seriously fix "secs_to_cputime()" to do the proper thing?
> Or did I miss a patch and you already did that?

I wasn't really on a cputime-cleanup rampage, just fixing the obvious bugs
I saw.  I hope Martin would like to do some cleanup.


Thanks,
Roland


--- linux-2.6/kernel/sched.c
+++ linux-2.6/kernel/sched.c
@@ -2299,17 +2325,17 @@ static void account_it_prof(struct task_
 static void check_rlimit(struct task_struct *p, cputime_t cputime)
 {
 	cputime_t total, tmp;
+	unsigned long secs;
 
 	total = cputime_add(p->utime, p->stime);
-	tmp = jiffies_to_cputime(p->signal->rlim[RLIMIT_CPU].rlim_cur);
-	if (unlikely(cputime_gt(total, tmp))) {
+	secs = cputime_to_secs(total);
+	if (unlikely(secs >= p->signal->rlim[RLIMIT_CPU].rlim_cur)) {
 		/* Send SIGXCPU every second. */
 		tmp = cputime_sub(total, cputime);
-		if (cputime_to_secs(tmp) < cputime_to_secs(total))
+		if (cputime_to_secs(tmp) < secs)
 			send_sig(SIGXCPU, p, 1);
 		/* and SIGKILL when we go over max.. */
-		tmp = jiffies_to_cputime(p->signal->rlim[RLIMIT_CPU].rlim_max);
-		if (cputime_gt(total, tmp))
+		if (secs >= p->signal->rlim[RLIMIT_CPU].rlim_max)
 			send_sig(SIGKILL, p, 1);
 	}
 }
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/