|
|
Log in / Subscribe / Register

Quotes of the week

[Posted June 20, 2012 by corbet]

Sorry, it can't always be constructive, but I'll try my best. I'll also try to not cast aversions about your cat, but if you taunt me, all bets are off.
— Greg Kroah-Hartman

Hooks and notifiers are a form of "COME FROM" programming, and they make it very hard to reason about the code. The only way that that can be reasonably mitigated is by having the exact semantics of a hook or notifier -- the preconditions, postconditions, and other invariants -- carefully documented. Experience has shown that in practice that happens somewhere between rarely and never.
— H. Peter Anvin
to post comments

Quotes of the week

Posted Jun 21, 2012 1:51 UTC (Thu) by jackb (guest, #41909) [Link] (16 responses)

I spent all week waiting for this to come out just to see if a particular quote from Linus Torvalds would make it onto this page.

Quotes of the week

Posted Jun 21, 2012 2:08 UTC (Thu) by corbet (editor, #1) [Link] (8 responses)

When a quote is as widely sensationalized as that one was, it doesn't really seem necessary to copy it onto the kernel page as well. I suppose I could have quoted from this related message, but it kind of slipped my mind, sorry.

Quotes of the week

Posted Jun 21, 2012 6:01 UTC (Thu) by jackb (guest, #41909) [Link]

Perhaps it would be best addressed in the future, depending on if there are observable ramifications or not.

Quotes of the week

Posted Jun 21, 2012 7:13 UTC (Thu) by man_ls (guest, #15091) [Link] (3 responses)

We do not all read the lkml, and we do not all follow on tech news. Perhaps I am the only that has not heard about that quote, but it would be nice to have it here if only for future reference.

Quotes of the week

Posted Jun 22, 2012 16:53 UTC (Fri) by giraffedata (guest, #1954) [Link] (2 responses)

I too had no idea what was being referred to, but with clues in this thread, I did a web search and came up with this. The quote is only 3 words, but also comes with a visual gesture.

http://www.wired.com/wiredenterprise/2012/06/torvalds-nvidia-linux/

Quotes of the week

Posted Jun 22, 2012 17:10 UTC (Fri) by man_ls (guest, #15091) [Link] (1 responses)

Thanks, I had managed to miss it in a cursory web search. It is really not qotw material...

Quotes of the week

Posted Jun 22, 2012 23:49 UTC (Fri) by nix (subscriber, #2304) [Link]

It *was* amusing a few minutes later, when the last questioner said that he'd gone to work for <voice type="small">nvidia</voice>...

Quotes of the week

Posted Jun 21, 2012 14:46 UTC (Thu) by clump (subscriber, #27801) [Link]

Thank you for not posting the quote. I'd rather see more respectful discourse. Thank you instead for posting the link in this thread. I wasn't previously aware of Stephen Warren's and am happy to have read it.

Quotes of the week

Posted Jun 26, 2012 15:13 UTC (Tue) by ortalo (guest, #4654) [Link] (1 responses)

IMHO, all this was settled once upon a time in the past when binary-only modules were allowed to load into the kernel. (NB: At that time, it seemed to me that allowing this attitude from hardware manufacturers, especially in the GPU market, was satisfactory for Linus. Personnally, I am glad to see his posture has evolved.)

It does not mean all hope is lost. Maybe now, 15 years later, it's time to re-orient the strategy; either by:
1) changing the policy (boring PR stuff) and finding a technical way of preventing binary-only modules from loading (probably much more funny btw);
2) playing tricks with the API exported to these modules in order to increase significantly the burden of maintaining them behind closed doors.

Of course, any of these options may be even more provocative than the initial demonstration...

PS: I erroneously posted this comment at the root instead of as a reply in the good thread so this is a repost. Is it possible to delete the other one? (ID: 503553)

Quotes of the week

Posted Jun 26, 2012 19:09 UTC (Tue) by dlang (guest, #313) [Link]

If the binary module is not a derivative of the kernel, there are no grounds for saying that it can't be used. That's a very different statement from saying that this is a good thing, or even acceptable.

It's very possible to say that NVIDIA has the right to do what they are doing, but at the same time say that they are one of the worst companies to deal with.

Quotes of the week

Posted Jun 21, 2012 15:10 UTC (Thu) by jwarnica (subscriber, #27492) [Link]

And thank you for correcting that. With the link you gave, I'm up to date.

Quotes of the week

Posted Jun 21, 2012 20:24 UTC (Thu) by flewellyn (subscriber, #5047) [Link] (5 responses)

To what do you refer?

Quotes of the week

Posted Jun 22, 2012 2:25 UTC (Fri) by daniel (guest, #3181) [Link] (4 responses)

I believe the reference is to "constructive" encouragement by Linus for nVidia to start doing the right thing.

Quotes of the week

Posted Jun 22, 2012 7:27 UTC (Fri) by Fowl (subscriber, #65667) [Link] (2 responses)

I think the sad thing is that everyone seems to be thinking that Linux was referring to their binary graphics, but if you watch a bit of the context it's more likely that he was referring to their SoC efforts.

Quotes of the week

Posted Jun 22, 2012 18:56 UTC (Fri) by daniel (guest, #3181) [Link]

<div class="FormattedComment">
I'm pretty sure he was referring to the SoC graphics. Which is where NVidia clearly can't rely on the argument that the Linux segment is not big enough to warrant special attention.<br>
</div>

Quotes of the week

Posted Jun 22, 2012 20:32 UTC (Fri) by leoc (guest, #39773) [Link]

Probably because the question asked was specifically about their "Optimus" video card switching technology being such a pain in the ass, resulting in a lot of otherwise unnecessary work to reverse engineer something that works. Then they issued a condescending response talking about their valiant work updating their README files to point to it.

Quotes of the week

Posted Jun 23, 2012 0:54 UTC (Sat) by bronson (subscriber, #4806) [Link]

Yes, it must refer to this advice from a prominent Kernel developer (warning: it's Linus but not entirely SFW): http://static02.mediaite.com/geekosystem/uploads/2012/06/...

Cast aversions?

Posted Jun 21, 2012 13:59 UTC (Thu) by dskoll (subscriber, #1630) [Link] (1 responses)

I am asperse to casting aversions.

Cast aversions?

Posted Jun 22, 2012 2:29 UTC (Fri) by daniel (guest, #3181) [Link]

I normally only cast aversion on hostile monsters.

Quotes of the week

Posted Jun 22, 2012 23:39 UTC (Fri) by jtc (guest, #6246) [Link]

When I click on the "Article" link (http://mid.gmane.org/%3C4FDD07CF.8090605%40zytor.com%3E) from the H Peter Anvin quote (http://lwn.net/Articles/502634/), I get a page saying: "No such article". And when I google for H peter anvin and the subject of the thread (" [Ksummit-2012-discuss] [ATTEND] Wither the baseline; EFI breakout session"), the only relevant URL I find is the same LWN one.

Is this because I've been watching X-files reruns lately, or is there some other explanation?

Quotes of the week

Posted Jun 26, 2012 15:08 UTC (Tue) by ortalo (guest, #4654) [Link]

IMHO, all this was settled once upon a time in the past when binary-only modules were allowed to load into the kernel. (NB: At that time, it seemed to me that allowing this attitude from hardware manufacturers, especially in the GPU market, was satisfactory for Linus. Personnally, I am glad to see his posture has evolved.)

It does not mean all hope is lost. Maybe now, 15 years later, it's time to re-orient the strategy; either by:
1) changing the policy (boring PR stuff) and finding a technical way of preventing binary-only modules from loading (probably much more funny btw);
2) playing tricks with the API exported to these modules in order to increase significantly the burden of maintaining them behind closed doors.

Of course, any of these options may be even more provocative than the initial demonstration...


Copyright © 2012, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds