On 990906, at 13:04:34, [EMAIL PROTECTED] wrote:
> David, I hope you haven't plunged blindly ahead on my say-so and
> destroyed the data on your hd. Recent versions of the IDE driver are
> fairly good at detecting and evading buggy hardware, but if you have a
> VLB system with a cmd640 or rz1000 IDE chip, it is BOOBY-TRAPPED, and
> will make garbage out of your data at the slightest provokation. The
> IDE driver can't detect those VLB chips. If you tell it you have one,
> (if you do), you should be right. I recommend you read
> /usr/src/linux/Dpcumentation/ide.txt if you are going to play with
> hdparm, as well as the hdparm man page.
>
> I did destroy a fs with cmd640vlb, just by using both channels without
> telling the IDE driver it was there. I have the PCI version now, and
> hdparm won't _let_ me set umask. Fortunately, I have a modem with FIFO.
Well, I'm trying to be cautious, but I usually obtain good judgement the
hard way: "good judgement come from experience, experience comes from
bad judgement". So far, I've backed up the windows partition (I don't
know yet how to use the IDE tape under linux), and I'm just reading the
hdparm man page and querying the disk.
Anyway, the system has a BCM motherboard with Intel 440ZX chipset and
Intel PCI IDE controller chip, a Quantm disk, and an ActionTec PCI modem
(supposedly with a FIFO). Given this hardware, does it still make sense
to alter the disk parameters to solve my receive packet framing errors?
--
David Ellement
