netdev
[Top] [All Lists]

Re: [PATCH 1/7] netpoll: shorten carrier detect timeout

To: Matt Mackall <mpm@xxxxxxxxxxx>
Subject: Re: [PATCH 1/7] netpoll: shorten carrier detect timeout
From: Patrick McHardy <kaber@xxxxxxxxx>
Date: Fri, 11 Mar 2005 05:53:16 +0100
Cc: Jeff Garzik <jgarzik@xxxxxxxxx>, netdev@xxxxxxxxxxx, Jeff Moyer <jmoyer@xxxxxxxxxx>
In-reply-to: <20050311044246.GT3120@xxxxxxxxx>
References: <2.454130102@xxxxxxxxxxx> <422A4A38.4040303@xxxxxxxxx> <20050306002015.GD3120@xxxxxxxxx> <422A564D.4080809@xxxxxxxxx> <20050310230117.GP3120@xxxxxxxxx> <42311FF9.5010007@xxxxxxxxx> <20050311044246.GT3120@xxxxxxxxx>
Sender: netdev-bounce@xxxxxxxxxxx
User-agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.7.5) Gecko/20050106 Debian/1.7.5-1
Matt Mackall wrote:
On Fri, Mar 11, 2005 at 05:35:05AM +0100, Patrick McHardy wrote:

So I'm going to change it from "flaky" to "untrustworthy" and add a
comment.

Why don't you trust an instaneously available carrier? Any
reason to assume there will be false positives?


Because I had reports of people losing all their boot messages until
this logic was added (about a year ago now?). I don't remember which
NICs were implicated, but some apparently report carrier is always
available.

If this problem is not common, I think it would be better to make
this behaviour dependant on a boot parameter instead of forcing
everyone to wait for 4s. Additionally you could have a blacklist
of flaky NICs.

Regards
Patrick

<Prev in Thread] Current Thread [Next in Thread>