Quotes of the week [LWN.net]

Many kernel developers believe that userspace is burned into ROM and the only thing they can change is the kernel. That turns out to be incorrect.

-- Avi Kivity

+	if (iommu->cap == (uint64_t)-1 && iommu->ecap == (uint64_t)-1) {
+		/* Promote an attitude of violence to a BIOS engineer today */

-- David Woodhouse

You don't get a consistent filesystem with ext2, either. And if your claim is that several hundred lines of fsck output detailing the filesystem's destruction somehow makes things all better, I suspect most users would disagree with you.

-- Ted Ts'o

I recommend a sledgehammer. If you want to lose your data, you might as well have some fun.

-- Rik van Riel

to post comments

Sledgehammer

Posted Aug 27, 2009 5:18 UTC (Thu) by magfr (subscriber, #16052) [Link] (6 responses)

I would use a pickaxe.

Modern hard drives are pretty sturdy and a sledge hammer is rather blunt so it tend to just bounce off the disk while the pickaxe have a nice effect when hitting the platters.

Sledgehammer

Posted Aug 27, 2009 15:36 UTC (Thu) by PaulMcKenney (✭ supporter ✭, #9624) [Link] (4 responses)

Indeed!!! The last time I attempted to destroy a memory stick, it took three blows with a 16-pound (7-and-a-quarter kilogram) sledgehammer to crack the silicon. And that was with the memory stick resting on a slab of concrete.

Sledgehammer

Posted Aug 27, 2009 16:32 UTC (Thu) by sumC (guest, #1262) [Link] (3 responses)

You should write that as 7,25 kg ;)

Sledgehammer

Posted Aug 27, 2009 17:15 UTC (Thu) by PaulMcKenney (✭ supporter ✭, #9624) [Link]

Hey, I made it across the Atlantic! Getting across the Channel is your problem. :-)

Sledgehammer

Posted Sep 3, 2009 13:48 UTC (Thu) by efexis (guest, #26355) [Link] (1 responses)

Seven thousand two hundred and fifty kilogram?!!

Sledgehammer

Posted Sep 3, 2009 15:18 UTC (Thu) by PaulMcKenney (✭ supporter ✭, #9624) [Link]

The other way to interpret it would be a blow from a 7kg sledgehammer followed by another blow from a 25kg sledgehammer. Which probably would have made the third blow redundant from the viewpoint of the memory stick and impossible from the viewpoint of my muscles and bones. :-)

Sledge hammer, Pickaxe, etc.

Posted Aug 31, 2009 16:31 UTC (Mon) by pr1268 (guest, #24648) [Link]

I would use a pickaxe.

Modern hard drives are pretty sturdy and a sledge hammer is rather blunt so it tend to just bounce off the disk while the pickaxe have a nice effect when hitting the platters.

Sure, but a little plastic explosive stuck directly on the hard drive casing is a more spectacular way to go! Of course, there'd be some collateral damage to deal with.

Disclaimer: I'm only kidding here. Please, don't anybody try this. :)

Quotes of the week

Posted Aug 27, 2009 20:10 UTC (Thu) by quotemstr (subscriber, #45331) [Link] (1 responses)

Avi Kivity is right. The set of things you can do with an ioctl interface (or better yet, a new system call) is a superset of the things that can be done with a filesystem interface, and explicit binary interfaces provide precise kinds of information that text interfaces lack. The discoverability advantage of filesystem interfaces can be addressed by simply writing a filesystem layer on top of the binary interface. As a developer, I'd appreciate a clear and atomic interface to get or set anything I need.

Quotes of the week

Posted Sep 1, 2009 2:36 UTC (Tue) by Kamilion (subscriber, #42576) [Link]

"As a developer, I'd appreciate a clear and atomic interface to get or set anything I need."

Yes, seriously, when are we going to see a dynamic etcfs?
It's the only thing truly holding back a 'local cloud' ramboot cluster.
Sure, there's things like augeas.net and fusefs-sqlfs or sshfs, but there's nothing I've found that can be easily abused into keeping a dynamically populated etc against a flatfile/database/archive without resorting to tmpfs and a tar/cpio hook or rsync at shutdown.

HTTP's everywhere today, GPXE has support for loading boot files over it, nginx is tiny and can remap it's directories to other servers with ease, and the only thing missing is proper mounting of an HTTP directory structure without fuse. People have been messing with this for over 10 years ( http://okmij.org/ftp/HTTP-VFS.html ) and still has yet to occur.

Probably because NFS-root was easier to set up... ;)

It's gotten to the point where HTTP-root should be feasible and so dead simple to deploy home-clusters that children could do it.

I mean, come on, how many people have successfully struggled through installing phpmyadmin? ;)