From: Patrick McHardy <kaber@trash.net>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: netdev@vger.kernel.org, bugzilla-daemon@bugzilla.kernel.org,
bugme-daemon@bugzilla.kernel.org, stas@crypt.org.ru
Subject: Re: [Bugme-new] [Bug 14875] New: iproute2: problems with "tc filter replace" and u32 hashing filters
Date: Tue, 05 Jan 2010 05:46:56 +0100 [thread overview]
Message-ID: <4B42C440.3070906@trash.net> (raw)
In-Reply-To: <20100104123314.a8e1de57.akpm@linux-foundation.org>
Andrew Morton wrote:
> (switched to email. Please respond via emailed reply-to-all, not via the
> bugzilla web interface).
>
>> http://bugzilla.kernel.org/show_bug.cgi?id=14875
>>
>> I'm using u32 hashing filters and have some issues with "tc filter replace"
>> command.
>>
>> Issue 1: "tc filter replace" command does not replace filters inside u32 hash
>> tables and works like "tc filter add" command.
>>
>> Consider the following scenario:
>>
>> ...
>> 2. Add filter for IP-address 10.0.0.1
>>
>> tc class add dev eth1 parent 1: classid 1:3 htb rate 256kibit ceil 256kibit
>> tc qdisc add dev eth1 parent 1:3 handle 3:0 pfifo limit 50
>> tc filter add dev eth1 parent 1: pref 20 u32 ht 100:1: \
>> match ip src 10.0.0.1 flowid 1:3
>>
>> 3. Try to replace filter for 10.0.0.1 with a new one for 10.0.0.2
>>
>> tc filter replace dev eth1 parent 1: pref 20 u32 ht 100:1: \
>> match ip src 10.0.0.2 flowid 1:3
>>
>> I expect that filter for 10.0.0.1 in hash table 100:1: have been replaced by
>> new
>> rule for 10.0.0.2. But "tc -p filter show dev eth1" outputs two filters for
>> both 10.0.0.1 and 10.0.0.2:
>>
>> filter parent 1: protocol ip pref 10 u32 fh 100:1:800 order 2048 key ht 100
>> bkt 1 flowid 1:3
>> match IP src 10.0.0.1/32
>> filter parent 1: protocol ip pref 10 u32 fh 100:1:801 order 2049 key ht 100
>> bkt 1 flowid 1:3
>> match IP dst 10.0.0.2/32
>>
>> It means that "tc filter replace" command did not delete the filter 100:1:800,
>> but attached a new one with handle 100:1:801, just like the "tc filter add"
>> command. I think it is a wrong behaviour for "replace" command.
You need to specify a handle for the filters to get replaced.
tc filter add dev eth1 parent 1: pref 20 handle 100:1 u32 ht 100:1: \
match ip src 10.0.0.1 flowid 1:3
tc filter replace dev eth1 parent 1: pref 20 handle 100:1 u32 ht 100:1: \
match ip src 10.0.0.2 flowid 1:3
works fine.
>> Issue 2: It seems that tc does not provide any syntax to replace a single
>> filter
>> inside the hash table. The command with explicit handle number
>>
>> tc filter replace dev eth1 parent 1: pref 20 u32 ht 100:1:800 \
>> match ip dst 10.0.0.3 flowid 1:3
>>
>> gives the error message: "ht" must be a hash table.
>>
>> The similar command with "handle 100:1:800" prints "What is "handle"?" and
>> usage information.
Handles consist of a major and minor number, not three numbers,
see above.
next prev parent reply other threads:[~2010-01-05 4:47 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <bug-14875-10286@http.bugzilla.kernel.org/>
2010-01-04 20:33 ` [Bugme-new] [Bug 14875] New: iproute2: problems with "tc filter replace" and u32 hashing filters Andrew Morton
2010-01-05 4:46 ` Patrick McHardy [this message]
2010-01-05 16:00 ` Stanislav Kruchinin
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=4B42C440.3070906@trash.net \
--to=kaber@trash.net \
--cc=akpm@linux-foundation.org \
--cc=bugme-daemon@bugzilla.kernel.org \
--cc=bugzilla-daemon@bugzilla.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=stas@crypt.org.ru \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.