Re: [PATCH] tun

netdev

Re: [PATCH] tun

To:	Stephen Hemminger <shemminger@xxxxxxxx>
Subject:	Re: [PATCH] tun
From:	"David S. Miller" <davem@xxxxxxxxxx>
Date:	Mon, 15 Mar 2004 12:47:51 -0800
Cc:	maxk@xxxxxxxxxxxx, vtun@xxxxxxxxxxxxxxxxxxxxxxxx, netdev@xxxxxxxxxxx
In-reply-to:	<20040315102810.2111a5d7@dell_ss3.pdx.osdl.net>
References:	<20040315102810.2111a5d7@dell_ss3.pdx.osdl.net>
Sender:	netdev-bounce@xxxxxxxxxxx

On Mon, 15 Mar 2004 10:28:10 -0800
Stephen Hemminger <shemminger@xxxxxxxx> wrote:

> Tun verifies the user address before doing memcpy_from/toiovec.  There are
> two problems with this strategy.  First, it is possible on an SMP machine
> to construct cases where a page is unmmapped between the verify and the copy,
> and it is also faster to do the verification only once.

Yes, those verify_area calls are superfluous and racy.

Applied, thanks Stephen.

[More with this subject...]

<Prev in Thread]	Current Thread	[Next in Thread>
[PATCH] tun, Stephen Hemminger Re: [PATCH] tun, David S. Miller <=

Previous by Date:	Re: [PATCH] tun - don't obscure error return from misc_register, David S. Miller
Next by Date:	Re: [PATCH] (repost) TCP westwood bw_sample, David S. Miller
Previous by Thread:	[PATCH] tun, Stephen Hemminger
Next by Thread:	[PATCH] IPv4 ARP hash algorithm has poor distribution, Tim Gardner
Indexes:	[Date] [Thread] [Top] [All Lists]