Does the Meiwes case bode ill for S/M?

By all accounts, Marwin Meiwes and his breakfast, er, victim are probably well-known to non-US readers. Although the Meiwes case has been downplayed in the US, the actual case is somewhat spectacular. Miewes advertised on the Internet his "search for a young man 18-30 for slaughtering." Bernd Brandes showed up on Meiwes door and offered himself up. There's absolutely no doubt that Brandes and Meiwes were very serious about their agreement: Meiwes would kill and eat Brandes.

German prosecutors want to charge Meiwes with murder. But they can't. Brandes clearly and without apparent duress repeatedly stated his desired to be killed and eaten. Under German law, the most they can charge him with is "unauthorized consensual euthanasia," which carries a maximum penalty of five years in prison. The same is true of cannibalism; the most prosecutors can come up with is "disturbing the peace of the dead," which is not a felony.

A lawyer at the U.N., Noah Leavitt, has proposed a different tack: charge Meiwes with torture. As Leavitt notes:

These international and European law sources appropriately recognize that, when the crime is torture, the victim's consent is almost completely irrelevant. As has been exhaustibly documented in studies from around the world, torture victims will "consent" to almost anything if faced with enough pain. As a result, a victim's consent to torture can never be trusted.

This represents a chilling new possibility for prosecutors who want to appear moral and "doing something": since consent is never a defense in cases of torture, if a prosecutor can convince a judge that S/M constitutes torture, S/M will become de-facto illegal in that district. And it only takes a few precedents to make such an assumption the law of the land.

Linux 2.6.1 on a Thinkpad 600E

As always, I've been trying to upgrade my Linux box and keep it current with the most recent deployments. Now that the 2.6.1 kernel is out I've been working out how to install it on Lain, which is a thinkpad 600E.

My first try through was a complete bust. I didn't understand the new module model well enough, but after re-reading the instructions I managed to get it to work; I could reboot to 2.4 and run in 2.6 without much problem. The sound is different in 2.6; instead of a hacked cs4232 card, it actually recognized the cs4600 card and uses it. Unfortunately, this doesn't do me much good as I still can't get sound working.

But the real nightmare was that APM didn't work. That's a killer on a laptop, naturally, where having power management is essential, especially if you want the laptop to suspend correctly. I finally found the culprit. The relevant discussion regarding the Thinkpad 600 line can be found here. Note that if the patch isn't in the correct places in 2.6.2 you may have to hack this patch in by hand. It's completely cargo-cult; I don't understand quite what it does in the place that it's in, but I understand the effect: it pokes the port to turn off the floppy driver in the BIOS. A closer look at the code shows that this is where are block devices are instantiated in the kernel and hardware responsibility for them is surrendered.

--- linux/drivers/block/ll_rw_blk.c_ORIG        Mon Jan  5 20:51:28 2004
+++ linux/drivers/block/ll_rw_blk.c     Mon Jan  5 20:35:23 2004
@@ -2690,6 +2690,9 @@
        for (i = 0; i < ARRAY_SIZE(congestion_wqh); i++)
                init_waitqueue_head(&congestion_wqh[i]);
  
+#if defined(__i386__)  /* Do we even need this? */
+       outb_p(0xc, 0x3f2);
+#endif
        return 0;
 }

I had hoped that would be the end of it, but no. Of course not. I was given another rude shock when, after popping out the network card, ifconfig locked up trying to shut down the ethernet port. This was not the case on the 2.4.21 kernel.

I had hoped to have the 2.6.1 kernel up and running quickly. It promises a great many improvements. But not today. I've gone back to 2.4.21 and will probably be there for a few more weeks.