netdev
[Top] [All Lists]

Re: [RFC] locking changes for lec.c

To: "chas williams - CONTRACTOR" <chas@xxxxxxxxxxxxxxxx>
Subject: Re: [RFC] locking changes for lec.c
From: "David S. Miller" <davem@xxxxxxxxxxxxx>
Date: Fri, 14 Jan 2005 13:56:12 -0800
Cc: netdev@xxxxxxxxxxx
In-reply-to: <200501061717.j06HHJ5I000508@xxxxxxxxxxxxxxxxxxxxxxx>
References: <200501061717.j06HHJ5I000508@xxxxxxxxxxxxxxxxxxxxxxx>
Sender: netdev-bounce@xxxxxxxxxxx
On Thu, 06 Jan 2005 12:17:20 -0500
"chas williams - CONTRACTOR" <chas@xxxxxxxxxxxxxxxx> wrote:

> after looking at things recently its not clear to me that the combination
> of lec_arp_lock and lec_arp_users was protecting things properly.
> here is a small rewrite that eliminates lec_arp_users completely and
> uses lec_arp_lock to protect the lists in lec_prev: lec_arp_tables,
> lec_arp_empty_ones, et al.

Can HW interrupt paths ever call into this ARP stuff?
If not, probably should just use BH disabled locking
instead of the heavy handed IRQ disabling locks.

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