Subject: Re: DMA caused double fault
To: None <tech-kern@NetBSD.org>
From: Izumi Tsutsui <tsutsui@ceres.dti.ne.jp>
List: tech-kern
Date: 04/29/2006 00:10:56
In article <445229F9.1020909@anastigmatix.net>
nblists@anastigmatix.net wrote:

> As a rather subtle point in the usage of 'corrupt' (looking in an
> old Webster unabridged that predates most the word's specific use
> in informatics), there are about 7 senses of the word but they all
> involve the notion that something has transitioned from an original
> right or pure state into an impaired one.

Thank you for explanataion. In that case I can only say:
- uninitialized variables might cause different behavour
  between the first powercycle and second reboot
- passing wrong parameters to the DMA bus-master could cause
  any unexpected behavour because it might write junk data
  to any random addresses, which would not be initialized later
- putting a bunch of assertions or printfs might help on such situation
---
Izumi Tsutsui