linuxmafia.com 2013/2013.06.linuxmafia.com/linuxmafia.com/pub/linux/security/f00f/collinstmp.txt

Text File | 1997-11-20 | 6.9 KB | 188 lines

[Added comp.sys.intel.  Followups set.]

Jonas Munsin <[e-mail]> wrote:
: If you check www.slashdot.org you will see that the bug was raported
: at Fri Nov 07 at 2:56PM (actually it might have been reported even
: earlier, as that might be the date of the correction (which was a false
: correction)). And it that first post, #linuxos on efnet given credit.

_Not_ Friday the 7th.  Thursday, 6 Nov 97, 9:57pm USA Central Standard Time.
Full text follows (snipping a few extraneous header lines):

---<snip>---

From: [e-mail]
Newsgroups: comp.os.linux.advocacy
Subject: This code will lock up any P5 machine, even usermode Linux!  (F0 0F C7 C8)
Date: Thu, 06 Nov 1997 21:57:33 -0800
Organization: The University of Texas at Austin, Austin, Texas
Message-ID: <[e-mail]>
NNTP-Posting-Host: dial-102-5.ots.utexas.edu
X-Mailer: Mozilla 3.0 (Win95; I)

Hi,

        Check this out.  If you execute F0 0F C7 C8 on a P5 it will lock the
machine up.  This is true for any operating system including usermode
Linux.  It's pretty cool.  Basically, the opcodes are an invalid form of
cmpxchg8b eax with a lock prefix.  Has anyone seen this before?  The
problem doesn't show itself for the Pentium Pro or Pentium 2.

---<snip>---

You will note that several news reports, such as Alexander Wolfe's in
_EE Times_, erroneously claim the story broke on comp.sys.intel.
On the other hand, C|Net (www.news.com) rather lamely claims in every
story on the subject -- in defiance of the facts -- to have "first reported"
the bug.  Lastly, a number of stories have gotten wrong the intial post's
date.

Considering how heavily these organisations obviously lean on netnews,
they seem to take little care to report on it correctly.  I wonder if
they'll see fit to mention that Linux patched around the bug in both
the development and release kernels, within seven days (and still not
word one from Micro$oft)?




[comp.sys.intel re-added.  Followups set.]

In comp.os.linux.hardware, [e-mail] (Robert Collins) wrote
as follows:

Rick Moen <[e-mail]> wrote:

>> You will note that several news reports, such as Alexander Wolfe's in
>> _EE Times_, erroneously claim that the story broke on comp.sys.intel.

> So what.

So, the fact that it broke on a _Linux_ newsgroup, and that key technical
facts emerged there _before_ they broke on comp.sys.intel, is newsworthy,
and has been omitted through shoddy reporting.

> Is the actual newsgroup name the piece that's newsworthy?

Here, it is a significant part of what's newsworthy, yes.  For reasons
noted.

>> On the other hand, C|Net (www.news.com) rather lamely claims in every
>> story on the subject -- in defiance of the facts -- to have "first
>> reported" the bug.

> Context is everything.

Isn't it, though?  Reading the C|Net reports leaves you the impression of
enterprising reporters pulling the story out of thin air, rather than
cribbing it from a Linux newsgroup.  C|Net's omission of its source was,
of course, my point.

> From the context of the C/Net article, I knew they meant that they were
> the first *NEWS ORGANIZATION* to report the bug.

...and being Robert R. Collins, _you_ knew where they got it from.  Most
of the other readers did not -- and certainly were _very_ likely to
be misled by that wording.  Which was my point.  Which you're ignoring.

> I don't consider comp.os.linux.advocacy a news organization.  Therefore
> I don't have any problem with the C/Net article.

That is a non-sequitur.  Further, in this case, comp.os.linux.advocacy
in fact functioned in a manner indistinguishable _from_ a news organisation.
And went uncredited by C|Net.

>> Lastly, a number of stories have gotten wrong the intial post's
>> date.

> Who cares?  Is that really significant?

If you care about accuracy, yes.  If they get such fundamental things
wrong in cases where the real data are eminently and easily available,
doesn't that make you wonder what else they're going to bollix where
you cannot similarly check up on them?

> And what was the initial posting date?

Date: Thu, 06 Nov 1997 21:57:33 -0800
(As previously mentioned, USA Central Standard Time.)

> Can you make one blanket statement for all international time zones?

Now, you're getting extremely silly.  (And where do you imagine C|Net
writes from, Djakarta?)

>> Considering how heavily these organisations obviously lean on netnews,
>> they seem to take little care to report on it correctly.  I wonder if
>> they'll see fit to mention that Linux patched around the bug in both
>> the development and release kernels, within seven days (and still not
>> word one from Micro$oft)?

> Again, who cares?

Wait, let me guess...

Um, I know!  People who are bothered if everything outside the usual
Redmond-hypnotised blinkers simply _fails_ to be reported, no matter
how significant?

Or maybe not.  Who's to say?  Maybe nobody _does_ care.  Maybe news is
only news if it comes from major purchasers of advertising space (and
never if it involves freeware).

"We were always at war with Eastasia...."  ;->

> Are we fighting with significance issues here?

_We_?  Maybe not.  However, you appear to have summarily ignored my
points.  To each his own.

> The newsworthiness of the story isn't the date it first appeared on a
> usenet newsgroup; it isn't which newsgroup it first appeared on; and
> it isn't whether or not Linux has a patch.

We have a difference of opinion, apparently, as these appear to me to
be significant facts.  Additionally, the fact that Linux is (or was,
when last I checked) one of only two Intel-based OSes to have a fix,
and the _only_ one to issue public source code for a fix, seems to
me to be not just significant but a _key_ fact.  Likewise omitted.

> The bug itself is the newsworthiness.  All of those other details
> are just filler for the story.

Heh.

Go work as editor for an industry paper.  Tell all your reporters
"It doesn't matter if you get the supporting facts all wrong.  All
that matters is to identify the overall issue."  You'll get fired.

I _do_ hope so, anyway.

> As for Micro$oft?  They have made a "word" on the subject.

Since you're nitpicking, they in fact did not, when I posted that.

>  Their "word" appears at the Intel web site.

[Bronx cheer]  Something to the effect of "We're working closely
with Intel blah blah blah", if I recall correctly, right?"

Thanks, Robert.  I haven't laughed that hard all week!

> Maybe you should get your own story straight before you worry
> about other people.

Nothwithstanding the fact that those people are paid full-time as
_professional reporters on technology_, and I'm not, it seems to
me that I'm doing a whole lot better, just as an amateur commentator
firing off instant Usenet posts when I'm half asleep.

I do admire that rhetorical fluorish, though.  Thanks for attempting
to play.

--
Cheers,                    Long ago, there lived a creature with a
Rick Moen                  voice like a vacuum cleaner.  We know little
rick (at) hugin.imat.com   about it, but we do know that it ate cats.