Does it physically disable the MMC port in hardware, and the OS is forced to obey, or is it merely a signal that causes the OS to release the MMC device?
That being said, I've not looked particularly hard
there are a few places (I'm lead to believe) that sell replacement back covers.. but another popular solution is a bit of a bodge.. using a rubber magnet (think fridge magnet) one of the ones that is flexible and thin like rubber..
These flat rubber magnets can be cut into a suitable small blob and wedged into the hole.. That should be enough to keep the MMC sensor happy, even if it won't keep the kick stand in place..
This site seems to sell a replacement cover for €15 which isn't bad!
I've got a few flexible magents stuck to the fridge at home, can whack one in the post if that'd help!
I'm between workstations at home.. I've just bought a new PC (first time I've bought a desktop since 2002 I think) anyway, I'm getting ready to do some more substantial dev work .. both on Maemo and MeeGo.. but I just need a quiet weekend to get sorted.. fingers crossed that might be this weekend.. especially if the 2nd processor part turns up this week!
I'll do some more kernel patch investigation work then.. and if I spot something, I might suggest it to Titan for incorporation into his overclocking uber kernel.. perhaps a system setting to assume the sensor is open / closed / automatic.. A little desktop widget might be handy too..
Hey there!
Cpitchford, i dont know how much i can beg you to work on a kernel patch to tell the phone to disregard the status of the back cover!
I really need the MMC to work, Nitdroid, Meego, everything awesome needs that damn thing and it wont even read!!!
I would do it myself but i dont have enough experience :/
ALTHOUGH you can alter the gconf and basically manually control whether the phone thinks the cover is on or not. It still didnt pick up my memory card, but the camera shutter was working, so i guess it was semi working?
Despite this, Dmesg still reported that the mmc was missing, so this could say that Dmesg or the phone in general doesnt rely on gconf to determine whether the cover is on or not.
But things like the camera use gconf.
But i think the best/only way to get round this is a kernel patch, i mean i dont know how legendary you would be if you could do it, and id be willing to help in any way i can
p.s. i know the drill Checked the magnet, checked this and that and it is actually a raped sensor on the mobo!
Had same problem. Hardware fault on the cover sensor - NOT missing magnet.
Fixed it with attached.
DISCLAIMER - READ FIRST AND CONSIDER: I am not a programmer; I like to tinker. YMMV.
My phone is running kernel-power49 - the module was compiled accordingly. I have no idea whether it will work with other kernels including the default (couldn't see any difference in the source but I can't guarantee it). Please don't blame me if it bricks your phone! 
After messing about for hours editing and recompiling kernels, I discovered that you only need one module.
****Rename attached file to omap_hsmmc.ko (forum wouldn't let me attach without changing extension)****
Copy the file to /lib/modules/current (as root) on your N900. Also copy to /lib/modules/<name of your kernel> - in my case : /lib/modules/2.6.28.10-power49
NB: I'm not actually sure it needs to go in both but I did just to make sure. Perhaps someone more knowledgeable can advise.... ?
Reboot and VOILA: working mmc slot. 
All I did was edit the source of the module to remove the cover checks. I only successfully completed this a few minutes ago but all seems well. I'll come back in a couple of days with an update.
It's a fairly crude hack and I might go back to it to see if it can be improved but I've spent far too much time recently farting around trying to keep this brilliant but seemingly delicate phone working adequately. I also have the communications error - currently fixed with some folded paper between the sim holder and battery. 
Always greatly admired NOKIA stuff... never again!
