netdev
[Top] [All Lists]

Re: [PATCH]: r8169: Expose hardware stats via ethtool

To: Richard Dawe <rich@xxxxxxxxxxxxxxxxxxxx>
Subject: Re: [PATCH]: r8169: Expose hardware stats via ethtool
From: Jeff Garzik <jgarzik@xxxxxxxxx>
Date: Sun, 27 Feb 2005 17:59:45 -0500
Cc: Francois Romieu <romieu@xxxxxxxxxxxxx>, Jon Mason <jdmason@xxxxxxxxxx>, netdev@xxxxxxxxxxx
In-reply-to: <42224F76.9000602@xxxxxxxxxxxxxxxxxxxx>
References: <42208D83.80803@xxxxxxxxxxxxxxxxxxxx> <200502261132.29261.jdmason@xxxxxxxxxx> <4220B9C6.1010106@xxxxxxxxx> <20050226181213.GA13230@xxxxxxxxxxxxxxxxxxxxxxxxxx> <42224F76.9000602@xxxxxxxxxxxxxxxxxxxx>
Sender: netdev-bounce@xxxxxxxxxxx
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.3) Gecko/20040922
Richard Dawe wrote:
BUG() is a bit exagerated imho.


It seems like a pretty good way of avoiding a buffer overrun to me. E.g.: you copy an extra statistic in rtl8169_get_ethtool_stats(), but forget to update the stats length. Is it not better to crash early, than encounter random behaviour later?


Yeah, that's why the BUG() is present in 8139cp: force an oops rather than continue corrupting memory, if the programmer made an error.

        Jeff



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