lkml.org 
[lkml]   [2000]   [Jul]   [21]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: The big IDE fight in a different light

Sorry Scoot,

NO kernels are safe with an unfiltered HDIO_DRIVE_CMD active in the IOCTL.

I have to finish 2.4 and fix 2.2 and 2.0........

The only safe thing to do in 2.2 and 2.0 is to disable it or use 2.4 plus
the patch. Since most do not want the patch and put there heads in the
sand and their butts in the air and I am trying not to kick them to hard.

You make the best point and that is why I am having to become so ugly to
win the issue.

Respectfully,

On Fri, 21 Jul 2000, Scott Long wrote:

> Ok. I don't work on the kernel, and I never post on this list. I do read
> it however, and I've been a little unnerved by the enormous argument
> between Andre and others over the past two days.
>
> I've found that the best way to break up a fight isn't to try to
> convince one side that the other side is correct. Instead, I try to
> distract the combatants with a different issue. And this is my attempt
> to do that.
>
> I think the kernel people need to remember one very important thing: you
> are no longer working for yourselves, as a hobby. Your ideals about what
> Linux "should" be are no longer as relevant as they once were. You are
> working for MILLIONS of users who put their faith in you and trust in
> you. And it is their opinions that matter ultimately, not yours. At
> least, if you want to continue on this path of world domination.
>
> The discussion (or rather shit-flinging contest) seems to be centered
> around the ability to physically fry an IDE disk. First of all, if this
> is possible in 2.2, is it possible in 2.0? I will back up to a 2.0
> kernel if not. If it is only possible in 2.4, then I WILL NOT UPGRADE TO
> 2.4. And no one in this office will, either. If all versions of Linux
> are susceptible, maybe we'll look for a different OS where it is more
> difficult to accomplish this damage. We cannot afford data destruction,
> but even more importantly we CERTAINLY cannot afford having our disks
> physically destroyed.
>
> I understand the argument that if you are root, you can do anything you
> want. But remember also that most hacker kids do not understand the full
> potential of their root access. Most hacker kids cannot write ioctl()'s
> that will do the damage you describe, MUCH LESS do the necessary
> bit-banging. I believe it was an unfortunate decision on Andre's part to
> publish these destructive sequences. I realize that it was his way to
> try to coax the kernel people into accepting his argument, by giving the
> hackers what they need to do this damage.
>
> You may think my arguments make little logical sense. THAT DOESN'T
> MATTER. It matters what I believe, not what you believe. I am the one
> using the kernel in the real world, along with the millions of other
> users who have about the same level of technical knowledge (perhaps less
> -- I am, at least, a programmer and understand your discussion). If they
> believe their OS could damage their disks, then they will switch to a
> different OS REGARDLESS of stupid technical issues and technical points
> made by people who are among the 0.01% of people who understand it. It
> is unfortunate but as Linux becomes accepted into the mainstream, user
> opinion will be dominated not by fact but by common schema. The
> "correctness" of these schema is irrelevant. You, the kernel developers,
> will have to go with the flow or risk being dumped. Users will sleep
> better at night knowing that this sort of disk damage has been made more
> difficult to accomplish. And that will be a selling point (not in a
> literal sense obviously) for Linux.
>
> So don't try to convince me that *I* am wrong by giving me technical
> spew. It won't work. And it won't work for the many other users out
> there. Make your decisions based on what the USERS will say, not on
> technicalities. It's your ass on the line. We can always go back to
> Microsoft if we want.
>
> And so in conclusion, I, a "mere user," want this safeguard in the
> kernel. If it's not there, I won't use the system. Period. BeOS is nice
> too. A simple API, good threading, good SMP. Just what I need.
>
> And I think other users feel the same. Keep that in mind.
>
> Regards,
> Scott
>

Andre Hedrick
The Linux ATA/IDE guy



-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:57    [W:0.095 / U:1.296 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site