From: Thomas Graf <tgraf@suug.ch>
To: jamal <hadi@cyberus.ca>
Cc: "David S. Miller" <davem@davemloft.net>, netdev@oss.sgi.com
Subject: Re: [PATCH 7/6] cls_fw: CONFIG_NET_CLS_IND is not dependant on CONFIG_NET_CLS_ACT
Date: Fri, 29 Oct 2004 14:53:56 +0200 [thread overview]
Message-ID: <20041029125356.GM12289@postel.suug.ch> (raw)
In-Reply-To: <1099053306.1025.95.camel@jzny.localdomain>
> > Regarding the generic statistics for actions:
> >
> > I thought about dumping the most common stats basic,queue,rate_est
> > in tcf_action_copy_stats and let the action module dump additional
> > stats in its get_stats implementation.
>
> I think the action code seems clean to me as is.
> Here are my thoughts:
> Most of them just need queue stats (actually at the moment all of them)
> so in tca_gen, struct tc_stats stats needs replacement to make sure they
> only have a queue stat because it is generic.
What about the rate estimator and basic stats in pkt_act.h? tca_gen
requires at least basic,queue and rate_est for me to compile.
> All actions call
> qdisc_copy_stats calls in their stats dumpers. Replace that call with
> the magic you have in the new stats dumping scheme.
That's a lot of duplicated code which could be avoided easly. It's
only a cosmetic thing but I don't see any reason for duplicating code :->
So this would be my way of doing it:
tcf_action_copy_stats {
start_copy()
copy_basic_stats()
copy_queue_stats()
copy_rate_est()
call get_stats() and let the action module dump additional stats
finish_copy
}
I'd also suggest to make a new TLV type TCA_ACT_STATS and not reuse
TCA_STATS.
Another issue... do we want the compatibility stuff and provide the
old tc_stats? I'd say no but it's your call.
> PS:- I hope this is after the bk snapshot with current cleanups you have
> is tested?
Sure. Always planning ahead.
next prev parent reply other threads:[~2004-10-29 12:53 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2004-10-29 0:21 [PATCHSET 0/6] PKT_SCHED: Generic classifier routines / cls_fw cleanup Thomas Graf
2004-10-29 0:22 ` [PATCH 1/6] PKT_SCHED: Add generic classifier routines Thomas Graf
2004-10-29 10:23 ` [RESEND " Thomas Graf
2004-10-29 0:23 ` [PATCH 2/6] cls_fw: Cleanup fw_classify Thomas Graf
2004-10-29 0:24 ` [PATCH 3/6] cls_fw: Use generic routines to configure action/policer Thomas Graf
2004-10-29 0:24 ` [PATCH 4/6] cls_fw: Use generic routines to dump action/policer Thomas Graf
2004-10-29 0:25 ` [PATCH 5/6] cls_fw: Whitespace/ifdef fixes Thomas Graf
2004-10-29 0:26 ` [PATCH 6/6] PKT_SCHED: break is not enough to stop walking Thomas Graf
2004-10-29 10:26 ` [PATCH 7/6] cls_fw: CONFIG_NET_CLS_IND is not dependant on CONFIG_NET_CLS_ACT Thomas Graf
2004-10-29 11:39 ` jamal
2004-10-29 11:53 ` Thomas Graf
2004-10-29 12:35 ` jamal
2004-10-29 12:53 ` Thomas Graf [this message]
2004-10-29 13:12 ` jamal
2004-10-29 14:24 ` Thomas Graf
2004-10-29 14:53 ` jamal
2004-11-02 0:39 ` David S. Miller
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20041029125356.GM12289@postel.suug.ch \
--to=tgraf@suug.ch \
--cc=davem@davemloft.net \
--cc=hadi@cyberus.ca \
--cc=netdev@oss.sgi.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).