From: Jeff Garzik <jgarzik@pobox.com>
To: Thomas Spatzier <thomas.spatzier@de.ibm.com>
Cc: linux-kernel@vger.kernel.org, Netdev <netdev@oss.sgi.com>
Subject: Re: [patch 4/10] s390: network driver.
Date: Sat, 13 Nov 2004 20:29:13 -0500 [thread overview]
Message-ID: <4196B4E9.40502@pobox.com> (raw)
In-Reply-To: <OF88EC0E9F.DE8FC278-ONC1256F4A.0038D5C0-C1256F4A.00398E11@de.ibm.com>
Thomas Spatzier wrote:
>
>
>
>>You should be using netif_carrier_{on,off} properly, and not drop the
>>packets. When (if) link comes back, you requeue the packets to hardware
>>(or hypervisor or whatever). Your dev->stop() should stop operation and
>>clean up anything left in your send/receive {rings | buffers}.
>>
>
>
> When we do not drop packets, but call netif_stop_queue the write queues
> of all sockets associated to the net device are blocked as soon as they
> get full. This causes problems with programs such as the zebra routing
> daemon. So we have to keep the netif queue running in order to not block
> any programs.
This is very, very wrong. You are essentially creating in in-driver
/dev/null simulator. This does nothing but chew CPU cycles by having
the driver drop packets, rather than allowing the system to work as it
should.
Queues are DESIGNED to fill up under various conditions.
Would not the zebra routing software have the same problems with cable
pull under an e1000 or tg3 gigabit NIC?
> We also had a look at some other drivers and the common behaviour seems to
> be that packets are lost if the network cable is pulled out.
Which drivers, specifically, are these?
The most popular drivers -- e1000, tg3, etc. -- do not do this, for very
good reasons.
Jeff
next parent reply other threads:[~2004-11-14 1:29 UTC|newest]
Thread overview: 67+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <OF88EC0E9F.DE8FC278-ONC1256F4A.0038D5C0-C1256F4A.00398E11@de.ibm.com>
2004-11-14 1:29 ` Jeff Garzik [this message]
2004-11-15 7:52 ` [patch 4/10] s390: network driver Paul Jakma
2004-11-21 8:16 ` Paul Jakma
2004-11-29 15:57 ` Thomas Spatzier
2004-11-29 16:30 ` Paul Jakma
2004-11-29 16:41 ` Thomas Spatzier
2004-11-29 20:27 ` Paul Jakma
2004-11-30 7:22 ` Thomas Spatzier
2004-12-05 6:25 ` Paul Jakma
2004-12-06 11:01 ` Post Network dev questions to netdev Please WAS(Re: " jamal
2004-12-06 11:27 ` jamal
2004-12-06 14:42 ` Hasso Tepper
2004-12-07 1:13 ` Herbert Xu
2004-12-07 2:22 ` jamal
2004-12-10 15:37 ` Paul Jakma
2004-12-14 7:40 ` Thomas Spatzier
2004-12-15 13:50 ` jamal
2004-12-15 15:03 ` Thomas Spatzier
2004-12-19 19:29 ` jamal
2004-12-19 22:29 ` Tommy Christensen
2004-12-19 23:05 ` jamal
2004-12-19 23:46 ` Tommy Christensen
2004-12-20 0:15 ` Jeff Garzik
2004-12-20 14:10 ` jamal
2004-12-20 18:54 ` Jeff Garzik
2004-12-21 0:13 ` Tommy Christensen
2004-12-21 1:19 ` Jeff Garzik
2004-12-22 10:56 ` Thomas Spatzier
2004-12-22 11:07 ` Jeff Garzik
2004-12-22 13:48 ` jamal
2005-01-03 9:10 ` Thomas Spatzier
2005-01-03 15:05 ` jamal
2005-01-04 23:28 ` Jeff Garzik
2005-01-05 3:19 ` jamal
2005-01-05 6:30 ` Paul Jakma
2005-01-05 13:16 ` jamal
2005-01-05 14:29 ` Paul Jakma
2005-01-06 13:55 ` jamal
2005-01-05 15:35 ` Tommy Christensen
2005-01-06 13:58 ` jamal
2005-01-06 15:06 ` Tommy Christensen
2005-01-07 13:32 ` jamal
2005-01-07 15:26 ` Tommy Christensen
2005-01-10 13:18 ` jamal
2005-01-16 23:10 ` jamal
2005-01-17 12:04 ` Hasso Tepper
2005-01-17 22:04 ` Tommy Christensen
2005-01-17 22:13 ` Peter Buckingham
2005-01-17 22:36 ` jamal
2005-01-17 22:53 ` Tommy Christensen
2005-01-17 21:38 ` Tommy Christensen
2005-01-30 23:39 ` Tommy Christensen
2005-01-31 0:09 ` jamal
2005-01-31 0:12 ` jamal
2005-01-31 0:31 ` Tommy Christensen
2005-01-31 3:26 ` jamal
2005-01-31 12:16 ` Tommy Christensen
2005-03-13 17:49 ` Hasso Tepper
2005-01-05 6:26 ` Paul Jakma
2004-12-20 14:16 ` Paul Jakma
2004-12-20 18:56 ` Jeff Garzik
2004-12-26 5:36 ` Herbert Xu
2004-12-19 22:43 ` Jeff Garzik
2004-12-19 23:54 ` Paul Jakma
2004-12-20 14:11 ` jamal
2004-12-07 2:39 ` jamal
2004-12-06 18:44 ` Paul Jakma
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=4196B4E9.40502@pobox.com \
--to=jgarzik@pobox.com \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@oss.sgi.com \
--cc=thomas.spatzier@de.ibm.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).