[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20080604154124.5b5b5bf7.akpm@linux-foundation.org>
Date: Wed, 4 Jun 2008 15:41:24 -0700
From: Andrew Morton <akpm@...ux-foundation.org>
To: KOSAKI Motohiro <kosaki.motohiro@...fujitsu.com>
Cc: kobayashi.kk@...s.nec.co.jp, kosaki.motohiro@...fujitsu.com,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH 4/4] per-task-delay-accounting: /proc export for memory
reclaim delay
On Wed, 04 Jun 2008 12:07:29 +0900
KOSAKI Motohiro <kosaki.motohiro@...fujitsu.com> wrote:
> Hi
>
> > diff --git a/fs/proc/array.c b/fs/proc/array.c
> > index 9e3b8c3..a3e6e86 100644
> > --- a/fs/proc/array.c
> > +++ b/fs/proc/array.c
> > @@ -498,7 +498,7 @@ static int do_task_stat(struct seq_file *m, struct pid_namespace *ns,
> >
> > seq_printf(m, "%d (%s) %c %d %d %d %d %d %u %lu \
> > %lu %lu %lu %lu %lu %ld %ld %ld %ld %d 0 %llu %lu %ld %lu %lu %lu %lu %lu \
> > -%lu %lu %lu %lu %lu %lu %lu %lu %d %d %u %u %llu %lu %ld\n",
> > +%lu %lu %lu %lu %lu %lu %lu %lu %d %d %u %u %llu %llu %lu %ld\n",
> > pid_nr_ns(pid, ns),
> > tcomm,
> > state,
> > @@ -544,6 +544,7 @@ static int do_task_stat(struct seq_file *m, struct pid_namespace *ns,
> > task->rt_priority,
> > task->policy,
> > (unsigned long long)delayacct_blkio_ticks(task),
> > + (unsigned long long)delayacct_freepages_ticks(task),
> > cputime_to_clock_t(gtime),
> > cputime_to_clock_t(cgtime));
>
> userland program oftern below access.
> thus, this patch break userland compatibility.
>
> ----------------------------------------
> #!/usr/bin/perl
>
> $stat = `cat /proc/$pid/stat`;
> split
> @stat_list = split(/ / , $stat);
> print stat_list[$index];
> ^
> |
> use array index number
>
Any application which breaks if we add more things to /proc/pid/stat
just needs to be fixed.
But we can't add new fields in the middle of the line! We can only add
new items to the end.
But I think it would be good if we just weren't to make this change.
Do we really want to keep reporting the same information in two
separate ways?
If so, a new /proc/pid/delays would be a better way, but it might be a
bit late for doing that now, given that we already put
delayacct_blkio_ticks into /proc/pid/stat.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists