2003-05-03 13:59:10

by Vinay K Nallamothu

[permalink] [raw]
Subject: [PATCH 2.{4,5}.x] mod_timer fix for sch_cbq.c

Hi,

sch_cbq.c: trivial {del,add}_timer to mod_timer conversions.

vinay

--- linux-2.5.68/net/sched/sch_cbq.c 2003-03-25 10:08:36.000000000 +0530
+++ linux-2.5.68-nvk/net/sched/sch_cbq.c 2003-05-03 19:29:08.000000000 +0530
@@ -1056,11 +1056,9 @@
sch->stats.overlimits++;
if (q->wd_expires && !netif_queue_stopped(sch->dev)) {
long delay = PSCHED_US2JIFFIE(q->wd_expires);
- del_timer(&q->wd_timer);
if (delay <= 0)
delay = 1;
- q->wd_timer.expires = jiffies + delay;
- add_timer(&q->wd_timer);
+ mod_timer(&q->wd_timer, jiffies + delay);
sch->flags |= TCQ_F_THROTTLED;
}
}


2003-05-03 18:47:46

by David Miller

[permalink] [raw]
Subject: Re: [PATCH 2.{4,5}.x] mod_timer fix for sch_cbq.c

On Sat, 2003-05-03 at 07:16, Vinay K Nallamothu wrote:
> Hi,
>
> sch_cbq.c: trivial {del,add}_timer to mod_timer conversions.

Vinay, I sent you email earlier today saying that you
need to send networking patches to the networking lists
and to the networking maintainers.

Yet, you're still shoveling these patches to lkml.
What is the problem?

--
David S. Miller <[email protected]>

2003-05-04 00:54:44

by jw schultz

[permalink] [raw]
Subject: Re: [PATCH 2.{4,5}.x] mod_timer fix for sch_cbq.c

On Sat, May 03, 2003 at 07:53:32AM -0700, David S. Miller wrote:
> On Sat, 2003-05-03 at 07:16, Vinay K Nallamothu wrote:
> > Hi,
> >
> > sch_cbq.c: trivial {del,add}_timer to mod_timer conversions.
>
> Vinay, I sent you email earlier today saying that you
> need to send networking patches to the networking lists
> and to the networking maintainers.
>
> Yet, you're still shoveling these patches to lkml.
> What is the problem?
>

This looks like a timer fix that happens to be in network
code, not a network fix. Judging from the others he just
posted he is fixing/upgrading timer code.

He only CC'd the kernel list. The bug was sent to
[email protected]. You are askiing him to filter
every piddly fix to identify which group it should go to.

I sypmathize with your position but i think you are setting
up too high a barrier. When multiple poeple repeat the same
mistake it is called a systems problem and you fix the
system.

It strikes me that a filter subscribed to lkml could examine
the postings for lines like:

--- linux-2.5.68/net/sched/sch_cbq.c 2003-03-25 10:08:36.000000000 +0530
+++ linux-2.5.68-nvk/net/sched/sch_cbq.c 2003-05-03 19:29:08.000000000 +0530

and if present recognize "this is for the network list" and
send them there if the networking list isn't already
specified as a destination. Such functionality, once
created, could be expanded for other lists or individuals
that want it.


--
________________________________________________________________
J.W. Schultz Pegasystems Technologies
email address: [email protected]

Remember Cernan and Schmitt