From: Jesper Dangaard Brouer <jdb@comx.dk>
To: Patrick McHardy <kaber@trash.net>
Cc: Jesper Dangaard Brouer <hawk@diku.dk>,
"netdev@vger.kernel.org" <netdev@vger.kernel.org>
Subject: Re: [PATCH 2/2]: [NET_SCHED]: Making rate table lookups more flexible.
Date: Mon, 03 Sep 2007 16:19:11 +0200 [thread overview]
Message-ID: <1188829151.16405.10.camel@localhost.localdomain> (raw)
In-Reply-To: <46DB2846.2030400@trash.net>
[-- Attachment #1: Type: text/plain, Size: 1296 bytes --]
On Sun, 2007-09-02 at 23:16 +0200, Patrick McHardy wrote:
> Jesper Dangaard Brouer wrote:
> > On Sun, 2 Sep 2007, Patrick McHardy wrote:
> >>
> >
> > Lets focus on the general case, where the functionality actually is
> > needed right away.
> >
> > In the general case:
> >
> > - The rate table needs to be aligned (cell_align=-1).
> > (currently, we miscalculates up to 7 bytes on every lookup)
>
> We will always do that, thats a consequence of storing the
> transmission times for multiples of 8b.
The issue is that we use the lower boundary for calculating the transmit
cost. Thus, a 15 bytes packet only have a transmit cost of 8 bytes.
> > - The existing tc overhead calc can be made more accurate.
> > (by adding overhead before doing the lookup, instead of the
> > current solution where the rate table is modified with its
> > limited resolution)
>
> Please demonstrate this with patches (one for the overhead
> calculation, one for the cell_align thing), then we can
> continue this discussion.
I have attached a patch for the overhead calculation.
I'll look into the "the cell_align thing" tomorrow.
--
Med venlig hilsen / Best regards
Jesper Brouer
ComX Networks A/S
Linux Network developer
Cand. Scient Datalog / MSc.
Author of http://adsl-optimizer.dk
[-- Attachment #2: pkt_sched.h.patch --]
[-- Type: text/x-patch, Size: 721 bytes --]
commit a29d43b78d5ddb20a808d9270c2de358556ba5ee
Author: Jesper Dangaard Brouer <hawk@comx.dk>
Date: Mon Sep 3 14:35:19 2007 +0200
[IPROUTE2]: Update pkt_sched.h
In struct tc_ratespec, replace 'addend', with 'cell_align' and 'overhead'.
Signed-off-by: Jesper Dangaard Brouer <hawk@comx.dk>
diff --git a/include/linux/pkt_sched.h b/include/linux/pkt_sched.h
index 268c515..a127d63 100644
--- a/include/linux/pkt_sched.h
+++ b/include/linux/pkt_sched.h
@@ -78,7 +78,8 @@ struct tc_ratespec
unsigned char cell_log;
unsigned char __reserved;
unsigned short feature;
- short addend;
+ char cell_align;
+ unsigned char overhead;
unsigned short mpu;
__u32 rate;
};
[-- Attachment #3: overhead_to_kernel.patch --]
[-- Type: text/x-patch, Size: 1754 bytes --]
commit 1e89442468dca983e0e3d24b89093370896f2d6a
Author: Jesper Dangaard Brouer <hawk@comx.dk>
Date: Mon Sep 3 15:49:41 2007 +0200
[IPROUTE2]: Overhead calculation is now done in the kernel.
The only current user is HTB. HTB overhead argument is now
passed on the kernel (in the struct tc_ratespec).
Signed-off-by: Jesper Dangaard Brouer <hawk@comx.dk>
diff --git a/tc/q_htb.c b/tc/q_htb.c
index 53e3f78..b579ebe 100644
--- a/tc/q_htb.c
+++ b/tc/q_htb.c
@@ -206,9 +206,11 @@ static int htb_parse_class_opt(struct qdisc_util *qu, int argc, char **argv, str
if (!buffer) buffer = opt.rate.rate / get_hz() + mtu;
if (!cbuffer) cbuffer = opt.ceil.rate / get_hz() + mtu;
-/* encode overhead and mpu, 8 bits each, into lower 16 bits */
- mpu = (unsigned)mpu8 | (unsigned)overhead << 8;
- opt.ceil.mpu = mpu; opt.rate.mpu = mpu;
+ opt.ceil.overhead = overhead;
+ opt.rate.overhead = overhead;
+
+ opt.ceil.mpu = mpu;
+ opt.rate.mpu = mpu;
if ((cell_log = tc_calc_rtable(opt.rate.rate, rtab, cell_log, mtu, mpu)) < 0) {
fprintf(stderr, "htb: failed to calculate rate table.\n");
diff --git a/tc/tc_core.c b/tc/tc_core.c
index 58155fb..1ab0ba0 100644
--- a/tc/tc_core.c
+++ b/tc/tc_core.c
@@ -73,8 +73,6 @@ int tc_calc_rtable(unsigned bps, __u32 *rtab, int cell_log, unsigned mtu,
unsigned mpu)
{
int i;
- unsigned overhead = (mpu >> 8) & 0xFF;
- mpu = mpu & 0xFF;
if (mtu == 0)
mtu = 2047;
@@ -86,8 +84,6 @@ int tc_calc_rtable(unsigned bps, __u32 *rtab, int cell_log, unsigned mtu,
}
for (i=0; i<256; i++) {
unsigned sz = (i<<cell_log);
- if (overhead)
- sz += overhead;
if (sz < mpu)
sz = mpu;
rtab[i] = tc_calc_xmittime(bps, sz);
next prev parent reply other threads:[~2007-09-03 14:19 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-08-31 12:22 [PATCH 2/2]: [NET_SCHED]: Making rate table lookups more flexible Jesper Dangaard Brouer
2007-09-01 7:10 ` Patrick McHardy
2007-09-01 21:56 ` Jesper Dangaard Brouer
2007-09-02 14:35 ` Patrick McHardy
2007-09-02 18:56 ` Jesper Dangaard Brouer
2007-09-02 21:16 ` Patrick McHardy
2007-09-03 14:19 ` Jesper Dangaard Brouer [this message]
2007-09-04 16:25 ` Patrick McHardy
2007-09-05 13:58 ` Jesper Dangaard Brouer
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=1188829151.16405.10.camel@localhost.localdomain \
--to=jdb@comx.dk \
--cc=hawk@diku.dk \
--cc=kaber@trash.net \
--cc=netdev@vger.kernel.org \
/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).