[PATCH] IPv6 MIB:ipv6DefaultRouterTable

[PATCH] IPv6 MIB:ipv6DefaultRouterTable

[Shirley Ma]

[-- Attachment #1: Type: text/plain, Size: 91 bytes --]

This patch is agaist 2.6.1 kernel.

Thanks
Shirley Ma
IBM Linux Technology Center



[-- Attachment #2: linux-2.6.1-ipv6mib7.patch --]
[-- Type: text/x-diff, Size: 1121 bytes --]

diff -urN linux-2.6.1/include/linux/rtnetlink.h linux-2.6.1-ipv6mib7/include/linux/rtnetlink.h
--- linux-2.6.1/include/linux/rtnetlink.h	2004-01-08 22:59:55.000000000 -0800
+++ linux-2.6.1-ipv6mib7/include/linux/rtnetlink.h	2004-01-13 10:41:55.000000000 -0800
@@ -247,7 +247,7 @@
 {
 	__u32	rta_clntref;
 	__u32	rta_lastuse;
-	__s32	rta_expires;
+	__u32	rta_expires;	/* seconds */
 	__u32	rta_error;
 	__u32	rta_used;
 
diff -urN linux-2.6.1/net/ipv6/route.c linux-2.6.1-ipv6mib7/net/ipv6/route.c
--- linux-2.6.1/net/ipv6/route.c	2004-01-08 22:59:48.000000000 -0800
+++ linux-2.6.1-ipv6mib7/net/ipv6/route.c	2004-01-13 10:41:55.000000000 -0800
@@ -1535,8 +1535,8 @@
 		RTA_PUT(skb, RTA_OIF, sizeof(int), &rt->rt6i_dev->ifindex);
 	RTA_PUT(skb, RTA_PRIORITY, 4, &rt->rt6i_metric);
 	ci.rta_lastuse = jiffies_to_clock_t(jiffies - rt->u.dst.lastuse);
-	if (rt->rt6i_expires)
-		ci.rta_expires = jiffies_to_clock_t(rt->rt6i_expires - jiffies);
+	if (rt->rt6i_expires && time_after(rt->rt6i_expires, jiffies))
+		ci.rta_expires = (rt->rt6i_expires - jiffies)/HZ;
 	else
 		ci.rta_expires = 0;
 	ci.rta_used = rt->u.dst.__use;

Re: [PATCH] IPv6 MIB:ipv6DefaultRouterTable

[YOSHIFUJI Hideaki / 吉藤英明]

In article <200401141622.09299.mashirle@us.ibm.com> (at Wed, 14 Jan 2004 16:22:09 -0800), Shirley Ma <mashirle@us.ibm.com> says:

> This patch is agaist 2.6.1 kernel.

Wrong. Do not change user interface.

--yoshfuji

Re: [PATCH] IPv6 MIB:ipv6DefaultRouterTable

[Shirley Ma]

[-- Attachment #1: Type: text/plain, Size: 992 bytes --]





This is a bug fix for ipDefaultRouter Table. According to the new IP MIBs,
the DefaultRouterLifeTime(expire) entry in this table is unsigned 32. It
measures the remaining length of time in seconds. If it's expire, the value
should be 0.

Thanks
Shirley Ma
IBM Linux Technology Center
15300 SW Koll Parkway
Beaverton, OR 97006-6063
Phone: (503) 578-7638
FAX:      (503) 578-3228



YOSHIFUJI Hideaki / 吉藤英明
<yoshfuji@linux-ipv6.org>@cerberus.hongo.wide.ad.jp> on 01/15/2004 01:03:15
AM

Sent by:    "Hideaki YOSHIFUJI" <yoshfuji@cerberus.hongo.wide.ad.jp>


To:    mashirle@us.ltcfwd.linux.ibm.com
cc:    davem@redhat.com, kuznet@ms2.inr.ac.ru, netdev@oss.sgi.com, Shirley
       Ma/Beaverton/IBM@IBMUS
Subject:    Re: [PATCH] IPv6 MIB:ipv6DefaultRouterTable


In article <200401141622.09299.mashirle@us.ibm.com> (at Wed, 14 Jan 2004
16:22:09 -0800), Shirley Ma <mashirle@us.ibm.com> says:

> This patch is agaist 2.6.1 kernel.

Wrong. Do not change user interface.

--yoshfuji


[-- Attachment #2: Type: text/html, Size: 1309 bytes --]

Re: [PATCH] IPv6 MIB:ipv6DefaultRouterTable

[YOSHIFUJI Hideaki / 吉藤英明]

In article <OF87A9D639.F6A8D26B-ON88256E1C.0064AD99@us.ibm.com> (at Thu, 15 Jan 2004 10:26:21 -0800), Shirley Ma <xma@us.ibm.com> says:

> This is a bug fix for ipDefaultRouter Table. According to the new IP MIBs,
> the DefaultRouterLifeTime(expire) entry in this table is unsigned 32. It
> measures the remaining length of time in seconds. If it's expire, the value
> should be 0.

Kernel does not have to provide MIB in as-is format.

--yoshfuji

Re: [PATCH] IPv6 MIB:ipv6DefaultRouterTable

[Shirley Ma]

[-- Attachment #1: Type: text/plain, Size: 1151 bytes --]





Agree.

But if no applications use this value, and nothing is boken. It's not
harmful to change the kernel to provide MIBs in as-is format.

Thanks
Shirley Ma
IBM Linux Technology Center
15300 SW Koll Parkway
Beaverton, OR 97006-6063
Phone: (503) 578-7638
FAX:      (503) 578-3228



YOSHIFUJI Hideaki / 吉藤英明
<yoshfuji@linux-ipv6.org>@cerberus.hongo.wide.ad.jp> on 01/15/2004 10:36:22
AM

Sent by:    "Hideaki YOSHIFUJI" <yoshfuji@cerberus.hongo.wide.ad.jp>


To:    Shirley Ma/Beaverton/IBM@IBMUS
cc:    mashirle@us.ltcfwd.linux.ibm.com, davem@redhat.com,
       kuznet@ms2.inr.ac.ru, netdev@oss.sgi.com, yoshfuji@linux-ipv6.org
Subject:    Re: [PATCH] IPv6 MIB:ipv6DefaultRouterTable


In article <OF87A9D639.F6A8D26B-ON88256E1C.0064AD99@us.ibm.com> (at Thu, 15
Jan 2004 10:26:21 -0800), Shirley Ma <xma@us.ibm.com> says:

> This is a bug fix for ipDefaultRouter Table. According to the new IP
MIBs,
> the DefaultRouterLifeTime(expire) entry in this table is unsigned 32. It
> measures the remaining length of time in seconds. If it's expire, the
value
> should be 0.

Kernel does not have to provide MIB in as-is format.

--yoshfuji


[-- Attachment #2: Type: text/html, Size: 1499 bytes --]

Re: [PATCH] IPv6 MIB:ipv6DefaultRouterTable

[YOSHIFUJI Hideaki / 吉藤英明]

In article <OFC6D64AE0.CEE7DA5F-ON87256E1C.0068A56A@us.ibm.com> (at Thu, 15 Jan 2004 11:09:47 -0800), Shirley Ma <xma@us.ibm.com> says:

> But if no applications use this value, and nothing is boken. It's not
> harmful to change the kernel to provide MIBs in as-is format.

Shirley, nobody can prove that nothing will be broken. :-)

--yoshfuji

Re: [PATCH] IPv6 MIB:ipv6DefaultRouterTable

[Shirley Ma]

[-- Attachment #1: Type: text/plain, Size: 1360 bytes --]






Hi, Yoshifuji,

This rule would mean that any new kernel development *could* break
something somewhere else? ;-)

I don't see any application uses these values for now, so I change them.
It's good for kernel to provide the real MIBs information. If later on,
some applications start to use these values, they can get the correct
numbers without converting. Unless you can give me an broken example. :-)

Thanks
Shirley Ma
IBM Linux Technology Center
15300 SW Koll Parkway
Beaverton, OR 97006-6063
Phone: (503) 578-7638
FAX:      (503) 578-3228



YOSHIFUJI Hideaki / 吉藤英明
<yoshfuji@linux-ipv6.org>@cerberus.hongo.wide.ad.jp> on 01/15/2004 11:19:50
AM

Sent by:    "Hideaki YOSHIFUJI" <yoshfuji@cerberus.hongo.wide.ad.jp>


To:    Shirley Ma/Beaverton/IBM@IBMUS
cc:    mashirle@us.ltcfwd.linux.ibm.com, davem@redhat.com,
       kuznet@ms2.inr.ac.ru, netdev@oss.sgi.com, yoshfuji@linux-ipv6.org
Subject:    Re: [PATCH] IPv6 MIB:ipv6DefaultRouterTable


In article <OFC6D64AE0.CEE7DA5F-ON87256E1C.0068A56A@us.ibm.com> (at Thu, 15
Jan 2004 11:09:47 -0800), Shirley Ma <xma@us.ibm.com> says:

> But if no applications use this value, and nothing is boken. It's not
> harmful to change the kernel to provide MIBs in as-is format.

Shirley, nobody can prove that nothing will be broken. :-)

--yoshfuji


[-- Attachment #2: Type: text/html, Size: 1685 bytes --]

Re: [PATCH] IPv6 MIB:ipv6DefaultRouterTable

[YOSHIFUJI Hideaki / 吉藤英明]

In article <OF6C177E72.62FA8BF9-ON87256E1C.006A8953@us.ibm.com> (at Thu, 15 Jan 2004 11:30:25 -0800), Shirley Ma <xma@us.ibm.com> says:

> I don't see any application uses these values for now, so I change them.
> It's good for kernel to provide the real MIBs information. If later on,
> some applications start to use these values, they can get the correct
> numbers without converting. Unless you can give me an broken example. :-)

ip -6 route ?

--yoshfuji

Re: [PATCH] IPv6 MIB:ipv6DefaultRouterTable

[David S. Miller]

On Fri, 16 Jan 2004 04:19:50 +0900 (JST)
YOSHIFUJI Hideaki / _$B5HF#1QL@ <yoshfuji@linux-ipv6.org> wrote:

> In article <OFC6D64AE0.CEE7DA5F-ON87256E1C.0068A56A@us.ibm.com> (at Thu, 15 Jan 2004 11:09:47 -0800), Shirley Ma <xma@us.ibm.com> says:
> 
> > But if no applications use this value, and nothing is boken. It's not
> > harmful to change the kernel to provide MIBs in as-is format.
> 
> Shirley, nobody can prove that nothing will be broken. :-)

And also, the current situation is OK as long as this value we present now
can be converted into the desired value.  Is this the case?

Re: [PATCH] IPv6 MIB:ipv6DefaultRouterTable

[David S. Miller]

On Thu, 15 Jan 2004 11:30:25 -0800
Shirley Ma <xma@us.ibm.com> wrote:

> I don't see any application uses these values for now, so I change them.
> It's good for kernel to provide the real MIBs information. If later on,
> some applications start to use these values, they can get the correct
> numbers without converting. Unless you can give me an broken example. :-)

Format and layout of procfs files have to stay as they are, this is just the
nature of the game.

We can create new files to present information in new ways and formats,
but I totally prefer using netlink based information gathering over making
new procfs files.

Re: [PATCH] IPv6 MIB:ipv6DefaultRouterTable

[YOSHIFUJI Hideaki / 吉藤英明]

In article <OF6C177E72.62FA8BF9-ON87256E1C.006A8953@us.ibm.com> (at Thu, 15 Jan 2004 11:30:25 -0800), Shirley Ma <xma@us.ibm.com> says:

> This rule would mean that any new kernel development *could* break
> something somewhere else? ;-)

Please try to avoid changing the interface as long as you can.
Please add something rather change it.

In these case, we can avoid changing the interface,
so let's avoid breaking interface.

For ipv6inetNetToMeiaTable case, it is ok to add "jiffies," 
but it wrong to change the value's context.
If you really want to "change" that, please change the name.

Thanks.

-- 
Hideaki YOSHIFUJI @ USAGI Project <yoshfuji@linux-ipv6.org>
GPG FP: 9022 65EB 1ECF 3AD1 0BDF  80D8 4807 F894 E062 0EEA

Re: [PATCH] IPv6 MIB:ipv6DefaultRouterTable

[YOSHIFUJI Hideaki / 吉藤英明]

In article <20040115113700.166ddfb0.davem@redhat.com> (at Thu, 15 Jan 2004 11:37:00 -0800), "David S. Miller" <davem@redhat.com> says:

> And also, the current situation is OK as long as this value we present now
> can be converted into the desired value.  Is this the case?

I think so for the routing table AFAIK.
For neighbours, "jiffies" is required to be exported.

--yoshfuji

Re: [PATCH] IPv6 MIB:ipv6DefaultRouterTable

[YOSHIFUJI Hideaki / 吉藤英明]

In article <20040115113840.4b7eb5d8.davem@redhat.com> (at Thu, 15 Jan 2004 11:38:40 -0800), "David S. Miller" <davem@redhat.com> says:

> We can create new files to present information in new ways and formats,
> but I totally prefer using netlink based information gathering over making
> new procfs files.

Although this discussion is about rtnetlink message,
problem will be solved if we provide jiffies for neigbours 
IMHO.

--yoshfuji @ let me sleep...

Re: [PATCH] IPv6 MIB:ipv6DefaultRouterTable

[Shirley Ma]

[-- Attachment #1: Type: text/plain, Size: 1082 bytes --]





OK, I agree with both of you. Then this patch is not needed. Yoshifuji, you
can sleep in peace. ;-)

Thank you all!
Shirley Ma
IBM Linux Technology Center
15300 SW Koll Parkway
Beaverton, OR 97006-6063
Phone: (503) 578-7638
FAX:      (503) 578-3228



YOSHIFUJI Hideaki / 吉藤英明 <yoshfuji@linux-ipv6.org>@oss.sgi.com on
01/15/2004 11:53:41 AM

Sent by:    netdev-bounce@oss.sgi.com


To:    davem@redhat.com
cc:    Shirley Ma/Beaverton/IBM@IBMUS, mashirle@us.ltcfwd.linux.ibm.com,
       kuznet@ms2.inr.ac.ru, netdev@oss.sgi.com, yoshfuji@linux-ipv6.org
Subject:    Re: [PATCH] IPv6 MIB:ipv6DefaultRouterTable


In article <20040115113840.4b7eb5d8.davem@redhat.com> (at Thu, 15 Jan 2004
11:38:40 -0800), "David S. Miller" <davem@redhat.com> says:

> We can create new files to present information in new ways and formats,
> but I totally prefer using netlink based information gathering over
making
> new procfs files.

Although this discussion is about rtnetlink message,
problem will be solved if we provide jiffies for neigbours
IMHO.

--yoshfuji @ let me sleep...



[-- Attachment #2: Type: text/html, Size: 1435 bytes --]