Linux rolls out the welcome mat for Microsoft's Copilot key Great news, Linux fans! Support for the Copilot key is coming in the 6.14 kernel. What do you think it should do? A year ago, Microsoft decided that what Windows PC keyboards really needed was a Copilot key. It had been years since the IT mega-corp added a key, but with Microsoft's AI obsession, the addition of one to invoke …
If you hit the Copilot key it should gather up all your /home/* files (including . stuff like .ssh/) and throw them up on publicly accessible dropbox then post a link to twitter.
If you hit the key 5 times in a row, then for accessibility it should just set your root password to 'password', take down your firewall, and (again) post your IP to twitter.
The good people there will surely help you summarize your data.
And hey, still better than Recall.
Actually yes - I was a telecomms software tester, and it was on a test laptop c.1999 - I wanted to see what would _really_ happen.
It was fun to see the system still going without most of the root filesystem, then I started killing processes manually one by one, but until you actually kill a process, it keeps going even without a file system "underneath" it.
It did feel like torture to the machine though ;)
Ah, once again my anecdotage gland is triggered.
Some years ago, in my former life at a large tech company, I had need of being able to remotely trigger a BSOD on a Windows machine - for demo purposes, I hasten to add.
Remote, meaning that the debug trick involving a registry edit and the Scroll Lock key or whatever it is wouldn't cut it.
So I wrote a little application that would, running as Administrator, traverse the process list and kill -9 everything it could, which invariably brought the OS down in very short order - mission accomplished.
Our sales teams loved the demo, because they could show customers a system crash on demand, and our product's ability to recover from it.
Except that someone in Redmond clearly didn't like it, because a few months later, I got a polite request from Very Senior Management to stop offering that demo. Apparently, in their market segment share battle with Chromebooks and Macs, Microsoft didn't appreciate potential customers being reminded that Windows does actually crash (and to be faaaaiiiirrrr, it doesn't do so as often now as, say, Windows 95 or even 2K).
So IIRC I changed the demo so that it triggered a fake-but-convincing "Ransomware"* alert on the screen, allowing the sales folks to use that as a hook for our product.
(* I first typed "Randomware", which would be something else entirely. A Randomware infection presumably and unpredictably puts a picture of a shoe, or a halibut, or the face of Pope Pius XX on your screen?)
I don't know if this is true for all unixes, but IIRC "rm" just removes a directory's reference to a file.
The underlying file continues to exist until (at least) there are no processes that have it open, and there are no other directories / files that have hard links to it's inode.
Or something like that.
This is true, files on Unix and Linux are removed when their reference count reaches zero.
Opening the files increases this by one. :)
If you're really bord you can do:
$ dd if=/dev/urandom of=somebigfile &
$ rm somebigfile
Your HD will then continue to fill, even though you won't be able to find 'somebigfile', until DD crashes and then all the space will magically re-appear.
When a file is opened, It is actually the link count in the in-core cached copy of the inode, not the copy stored on disk that is incremented. The on-disk copy remains unchanged.
If a file is open, there is always a cached copy of the inode that describes the file, stored in memory all the time the file is open at least once. While the file is open, the link count of the in-core copy and on-disk copy of the inode differ, but because the system always looks at the in-core copy if there is one, it always sees the incremented count.
Deleting the file from the filesystem decrements both the in-core and on-disk link counts.
Interestingly, because the paging system of most modern UNIX like systems does not page out text segment pages to the paging disk, but relies on the copy that is on disk that the program was initially loaded from, just having a running copy of a program also loads the inode into the cache, and increments the link count, to keep the program's text on disk even if the file is removed from the filesystem.
But if the system crashes, the link count on disk will be correct, as the process(s) keeping the file open will disappear, as will the incremented link count in memory.
Going back to the days where a system crash more often than not required a full fsck of any mounted filesystem that had open files on them because the link count is wrong is not something I want to go back to!
Before on demand paging (when UNIX systems swapped complete processes out of memory to the swap space, rather than just LRU pages), if you attempted to remove a file corresponding to a running process, it wouldn't let you, although I can't remember the exact mechanism.
I ran "rm -fr /" intentionally on the Hardware Management Consoles (which run Linux, and had been jail-broken to allow root access) of four IBM Power 7 775 supercomputers (two of them still listed in the Top 500) when they were decommissioned.
Like everyone else here, I was astounded by just how long the systems responded to the console after running it. Everything need to respond to pressing enter must have been still cached in memory.
Nothing needing something from the filesystem worked any more (including shutdown/halt) but the power button still worked!
> I ran "rm -fr /" .... when they were decommissioned.
I hope they ran encrypted file system. A decomission deletion should use dd if=/dev/urandom for the purposes of data nuke. Is that ebay offer still listed?
Well, shredded anyway. The disks were handed to the organisation's own IT security team, and they arranged correct disposal according to their rules.
The correct procedure that would have allowed disk re-use in that organisation at the time was a multipass wipe (one pass was not deemed good enough) of the disk with different patterns using something like "blanco" or sometimes just "shred" against the device, but it was deemed cheaper to just get the disks destroyed.
Anyway, the HMC's don't actually store anything of particular value, other than some IP addresses, some fault diagnostic information, and some ID's that were unique to the cluster being destroyed.
"You obviously have :-o"
I have ... I've posted this before, but here it is again just for gits and shiggles.
Just over a billion years ago as the Internet measures time (call it roughly 1984), I received a brand new Sun 2/160. It was a dual pedestal beast, with all of 8 Megs of RAM and a pair of 380 Meg CDC SMD drives. Roughly 65 grand worth.
I decanted it from the boxes-on-pallets, plugged all the cables in, and fired the thing up. Into a beautiful new GUI on the Sony Trinitron monitor, just as advertised. Logged in as root, on purpose as there were no other accounts as yet(!!), using the default password(!!!!) ... and poked around. All was well, near as I could tell. It even found and connected to the corporate network all by itself.
The plan was to repartition the disks to better suit our needs and then reinstall the OS. So I made absolutely certain I had the correct tapes, and did the one thing I had never done as a sysadmin ... closed the GUI, and from the # prompt ran rm -rf / intentionally. I was curious to see how long it would take to lose it's tiny little mind. It trundled away to itself for a few minutes, but seemingly was still working fine, enough of vmunix and the shell were in RAM and the swap partition to keep doing simple stuff. If necessary, as long as nobody turned it off, I think I could have restored it into a full, working system. I was quite surprised, but that wasn't really what I was there for ...
So I shut her down, went and got a cuppa coffee, reached for the first tape and went to fire up the machine ... only to discover it didn't ship with a tape drive, despite one being listed on the packing list. It had a lovely bezel that LOOKED like it might be a tape drive, but the space behind it was empty. Oops. So there I was, 8AM and stuck with 65K worth of dead Sun hardware that I was supposed to demo for the Brass at 4PM.
Fortunately the 1980s Sun had Clues about customer service. One call, and their field service rep had the SCSI tape drive, the requisite cables, and a couple of VMEbus cards, (E)EPROMS and spare OS tapes "just in case" on my desk in under forty minutes. She even hung out and made certain that the system worked properly after we took it apart to install the bits that needed installing, and then partitioned it and re-installed the OS.
I made the 4 o'clock deadline ... and bought the Rep the first of many well deserved dinners.
Silly Con Valley was a very small place back then ... Sometimes I really miss it.
One would hope so, however a quote from the previous Register article that was linked to in today's piece...
" It might replace the right CTRL key unless the keyboard is large enough to accommodate both. In other cases, it might be combined with the Menu key – accessing the Menu functionality would require an Fn key or something similar."
The obvious answer to keyboards that physically don't meet your needs is to not purchase one. If your work chooses to lumber you with such a poor choice, lobby your boss for an improvement.
Personally, I haven't used anything but one variation or another of the Model M for serious keyboarding since '88ish ... and even then, I remap it (caps lock becomes left control, for example ... why the fuck is there a caps lock key on keyboards made since 1975 or so, anyway?)
>” why the fuck is there a caps lock key on keyboards made since 1975 or so, anyway?”
It might surprise you there are also a Num lock key and a Scroll lock key.
Personally, I don’t see the point of a CoPilot key, The functionality can be implemented as per the design school of Win8 which came up with the Charms menu, which required users to know where to move the cursor off the screen and the velocity with which to do it, to bring the menu up.
There are useful reasons for the scroll lock and num lock keys. There is no useful reason to type in all caps.
The extra key under discussion is NOT "the copilot key", rather it is an extra key that Redmond plans to use to launch copilot. Folks running more enlightened OSes can use the new key for whatever they want. None of us are beholden to Redmond.
That was then. This is now. There has been no real need for ALL CAPS for around half a century, outside a few very specialized circumstances that probably 8 9s of computer users will never even see, much less need to use for anything. So why does every keyboard have to have a key that's completely unneeded? (Other than the re-map it to the left control key, of course).
Come on Jake, you're experienced enough to realise that you not having a use for something doesn't mean that nobody else does.
Caps lock is useful when entering a whole bunch of asset tags/serial "numbers"/SKUs/whatever that by convention will be in upper case. Trying to hold shift for each letter but not number when they're mixed like that is a recipe for disaster. Yes, some software could be configured to display that field in upper case, but that's not a universal solution. Having physical things in upper case but the electronic reference in lower case makes it a lot harder to cross-reference/check the entries.
I haven't seen anybody regularly hand-enter SKUs etc. for maybe three decades, certainly not since the technology became cheap enough to give scanners away as party favo(u)rs (remember the CueCat?). In that world, it's all barcodes and the like these days. Faster, and far less chance for error.
Yes, software automagically converts lower-case to upper, if needed. Computers are kind of good at that. Humans, not so much (as you pointed out).
Yes, I'm sure there are more than a few hold-outs using outdated methods, but off-hand I can't think of any.
Plus, it was included originally on typewriters for a reason. Given typewriters were used by typists when no one had even heard of computers, we can be sure it was put there to help typists. Perhaps you need to visit user groups/departments where people are still performing “typist” activities.
I hate my caps lock key so much I have it disabled with a registry setting (at work; home I typo more carefully).
Occasionally it would be handy to have it work - sometimes RDP gets socially confused and the target system thinks capslock is enabled (it can't be, but it thinks it is) but on the whole I enjoy not having to worry aboout accidentally hitting the useless thing and therefore SHOUTING UNTIL I notice.
Don't get me started on those users who seem to thing Caps Lock is there to get capital letters as opposed to using Shift.... why use one key press when two will do? grrr.
Icon because I have to be, at work anyway.
I've just been given a new keyboard at work (the old one had a luxuriant ecosystem flourishing between the keys) which has an extra Fn key between Ctrl and the Windows key (heavily remapped, obviously). I blame Mac. I now have to relearn two decades of muscle memory (but not lose the old muscle memory because I still need it for my home machine).
> Why all the animosity?
This is not animosity towards Linux or Linux users, this is people just asking a joke that has "If modern Microsoft actually did control the behavior of the Copilot key under Linux, how would they use that to f@#$ up Linux as badly as they've f@!#$ed up Windows 11 while trying to be 'helpful'?" as the question.
As to why all the animosity towards Microsoft, it's because Recall and Copilot even exist and they've been relentlessly making Windows 11 worse since it was introduced. They are pure sociopathy that only exist for Microsoft's advantage. So yes, the 'Copilot key' would, in the joke, do something just as sociopathic on the Linux side. This is basically "Ha ha! *sob*" from people who have to use Windows 11.
Like Compose
Though since Caps Lock is really only for mechanical typewriters (was really shift lock), I map Caps Lock to Compose
I suppose on a US keyboard missing the |\ beside Z and having one less key, you could map it to whatever is missing on your keyboard map. Black nail varnish and a grout pen helps to change the labels. It for white keys, tippex and a black marker.a
... but a buddy of mine made one with a blank keycap. His esc key is the "Coke" logo in red & white. Used Testors model paint, thinned down and airbrushed. The logo started as a stencil that he cut out of heat-shrink plastic on his Wife's hobby stencil cutter. Then he shrunk it down to fit before applying the paint (three or four very light coats). It looks pretty good, but he admits it was fiddly to create, and he had a lot of false starts.
I don't think I've ever used the windows key - despite years of using Windows at work, except those odd occasions when curious just to see what happens. In those instances it's more like a big red button just taunting you.
I sometimes used the menu key that seemed fashionable 25 years ago, but now I think about it, that might have been on Linux systems to bring up some menu, probably an old version of kde.
They're the super key(s).
xev quite nicely outputs what each key is if you have any doubts;
state 0x0, keycode 133 (keysym 0xffeb, Super_L)
Unfortunately, the super key(s) are marked with a demonic symbol on many keyboards, but placing a GNU sticker on such key resolves such issue.
On GNU systems it's useful to bind it to an application launcher (super + type program + return) or as the mod key for i3, as that leaves the Ctrl and Meta keys free for the Church of Emacs.
Never really occurred to me that they did that, but then I avoid M$ products more or less instinctively although I do have Win10 on the machine I am currently using. Looked up their range and discovered that the surface keyboard with copilot key is on offer for £400+.
I'll stick to my DAS KEYBOARD - standard QUERTY layout with sound knob and sleep button as the only non-traditional controls. I had one an age ago and then disappeared down the path of gaming keyboards etc. Unfortunately that is the road to madness when you are trying to drive several PCs with different OS's from one another. The Das Keyboard is solid German engineering and I'm sure could be used to beat a burglar to death if called upon to do so. Which is comforting.
I couldn't find a keyboard at the £400 price point you claimed. It's £129.99 including the VAT directly from Microsoft: https://www.microsoft.com/en-gb/d/surface-keyboard/93d9l5tctr0b?icid=Surface-Cat-surfaceaccessories-CTA1-CC1-Surface-Keyboard-FY25Fall-Launch
Unless you've got a link that says otherwise then that seems a lot like FUD to me.
I am deeply attached to my trusty Microsoft Mouse 5000; the poor thing is looking a bit worn now after what must be 15? years of use, but I haven’t found any replacement I like as much. And NOS Mouse 5000s go for silly money on eBay, so it’s clearly not just me!
As a owner of one ms keyboard and user of some in the past. Yes, they are boring but (the old hardware before this copilot refresh) good, the old wave keyboards were actualy really nice to use for extended typing.
As for ms phones, I had a lumia 530, from the time nokia was ms, and the phones were nice. The OS (win 8, 7 was crap and 10 was okayish but too heavy) was also much better than it's normally given credit to it. Yes it was "windows" but the ux was actually thought out (and easy to use single handed), in general consistent, which is something that android lacks even today. What it was lacking was apps, which may or may have not been an issue depending on the particular social media poisons one took.
I have a Microsoft Internet Keyboard Pro as my primary keyboard at home; after 25 years of daily use it's still - knock on wood! - going strong and I haven't found anything else I like as much.
An updated version of this that replaced the unpowered USB1.1 hub with modern USB 3 or C ports would be great.
Long ago (as in looong ago, when M$ Windows was just a graphical shell over DOS), they dabbled in hardware. Eventually they settled into mice and keyboards, but for a time they tried an add-in card. Graphics, iirc. As explained in an anonymously circulated year-end recap ("Not Necessarily the ${some_name} Street News"), the card came in two versions: Smoking and Non-Smoking.
They weren't the only software company to dip a toe into the hardware waters. Aldus, where I worked at the time, bought a local company that was developing a graphics card. I don't think it ever made it to market. Aldus also developed their own word processor, about the time that the first graphical variants of M$ Word arrived. The Aldus program never saw the light of day: M$ called and said to cut it out.
Decades ago, Microsoft made one reasonable model of KB, despite it having rubber dome switches beneath the keycaps.
I picked up a Hotblack Desiato model (black body, black keys, black keylabels, etc.) of the DAS, used, for $US10.00. I just need to discover and hand-label the "floating" keys (WinLogo, Fn, Menu, etc.) which manufacturers tend to place where-ever their hearts fancy.
Hello:
> ... what the heck should it do?
I'll be as clear as possible:
It should fuck the hell off.
By this I mean that there should be an option in every distribution's installer to not do anything but what it did before some dickhead thought that this was a good idea.
eg:
Select keyboard layout -> option 1: no microsoft shit
And then continue with the installation as usual.
My IBM ca. 1995 model 'M' does not and will not ever have any pilot or whatever that crap is called.
.
emacs always needs more keyboard combinations to memorize.
I use my "windows" key as a "super" key for emacs. I also have a "menu" key on the right for the window manager. For example, menu-F2 minimizes a window.
Then my other apps can use the function keys and alt however they want.
to remove the SysReq key from some of their keyboards, including a rather nice wireless mechanical one for office use. Something that you don't expect, so don't discover until too late.
So, just let us remap any key - Copilot included - to serve that magic low level function that gets us gracefully out of jail, when needed.
Indeed, however I have never come across a keyboard without the SysReq key (by whatever name) - until the one in question - so have taken for granted it would be there. Perhaps I have been living a very protected life; alternatively the SysReq function is now deemed too 'low level' for the great unwashed of the buying public, and is therefore surplus to profit requirements.
Not all manufacturers use the same set of key codes for non-standard keys. Two manufacturers might Have a similar innovation and assign a key for take a screenshot or change volume to the keyboard, but without the driver mapping the scan code correctly to the command, it’s not going to happen. The manufacturers will just pick any given scan code so there’s no way to predict that in advance.
> F23!? When did that creep in?
At least two decades before you were made: IBM 1379000 . Other variants are 1390572 1394100, different plug.
There are a lot of others, but that one is best known. New keyboards with F13-F24 are not uncommon either, used with Point-Of-Sale machines. But a dying breed too with touch and barcode/QR scanners.
Why? Because that key will be available on keyboards, of course. Why not use it for something useful?
Note that on a Linux system, it won't run anything Redmond-related .... unless the user asks it to.
Honestly, do you lot think that the label on a keycap defines what it does when pressed? Do you think that somehow MS is going to hide their crappy AI code in the keyboard? And more to the point, do you think that somehow the crappy MS AI code will run on Linux simply by pressing the key?
At this point best option is to use BSD oses. Why ? Because corporates won't commit or interfere with development because any of their competitors can take code so they won't be motivated to. So only commiter wants it to be truly open source will commit. We all know how much GPL have saved linux from corporate contamination.
... All this does is ensure that Linux is aware of the key.
Nowhere does it attach AI-anything to Linux, much less "integrate AI into the OS".
Under Linux, it's just another key that does absolutely nothing until the user tells it to.
And as a by the way, the Linux Kernel will NEVER contain AI code from anywhere, much less from Redmond. That's userspace shit.
I like the idea of one commentard ... have it call up The Doctor in EMACS. Sadly, Manglement wouldn't get the joke.
The kernel, Linux was always aware of the key in atkbd - the F23 scancode just wasn't mapped to anything, as previously nobody used it.
Of course such driver is totally useless by itself - it needs to send the keycodes to other software for anything to happen.
Hasn't nvidia already implemented some parts of Artificial stupIdity in their proprietary peripheral software and considering that such software is arguably an extension to Linux, meaning that Linux effectively already contains such code?
It's rather amazing that people don't comprehend the massive confusion caused by referring to the systemd OS as "Linux" and the kernel, Linux as "Linux kernel" (a naming scheme originating from OS's where the kernel is not given a name as it was developed in the same project and was put under a directory called "kernel" or whatever (and of course there is more confusion caused those who do know better who love calling kernels that don't operate by themselves; "Operating Systems")).
Well, I didn't see any reference to init systems at GNU Linux-libre, I can only assume it is not systemd ;)
Your point about naming is only still an issue to people who don't realize that we have at least (and have done for a long time):
* GNU/Linux
* Android/Linux
* Busybox/Musl/Linux etc.
* "Whatever runs on supercomputers"/Linux
It's wrong.
>only commiter wants it to be truly open source will commit.
The reason "open source" was defined was that corporates don't like hearing about free software and the corresponding discussion about what's right and wrong, thus a movement to end such discussions and to prevent people from even realizing that free software exists, in exchange for funding for an efficient development model was started.
If you want to ensure that your software remains free, no matter where it is used, the license to use is the AGPLv3-or-later (many corporates won't even use such licensed software privately, even though it doesn't require doing anything for such use case) or if you want to defend freedom a bit less, the GPLv3-or-later.
>all know how much GPL have saved linux from corporate contamination.
The reasons why that is the case is due to how the license is GPLv2-only and not GPLv2-or-later and is almost never actually enforced for freedom.
It is rather mainly enforced against freedom for corporate interests (approved kind of proprietary software derivative work from a corporate? That's alright, we'll distribute it and maintain it. Free software that respects the users freedom, under a free software license (GPLv3-or-later) written by non-corporates? We'll enforce our license.).
If you want to avoid corporate contamination, the only options are GNU/Linux-libre or GNU/Hurd (the only non-toy 100% free OS's).
One odd thing about this: Windows maps scancode 0x6E to F23, and so this patch is doing the same for Linux. But if you have a genuine IBM 122-key Host Connected keyboard, F23 only sends 0x6E if you've switched to the rarely-used scancode set 3. At power-up it sends 0x6A.
My best guess is that the Windows developer who originally mapped 0x6E to F23 was using a Key Tronic 122-key keyboard, which does send 0x6E in both modes.
Yes, Clippypilot is shite. I'm not sure what some commentators are getting worked up about, though - this is just an addition in keymapping.
Given the hardware requirements M$ stipulate for a laptop etc to get given a copilot key make them have stupid price tags, I'd be surprised if one ever falls into the "buy that as a daily driver or tinkering machine" category for just about anyone on here...although if I did end up with one, I like the idea of a Pizza key though.... just like a Netpliance from the 1990s:
https://www.youtube.com/watch?v=USQxZc9nmtE
My terminal is attached to a serial port (the S in USB) and gets a login prompt whenever the computer boots.
Or simply use <ctrl><alt>F1 to pop into a handy virtual console.
F2, F3 etc. might also have virtual consoles running on them. Consult /etc/inittab for how many. If you have 3 virtual consoles, F4 will be the GUI you boot into. If you have 4, then F5 will be the GUI. Etc. (This assumes you're not running a particularly brain-dead "consumer oriented" distro ... ).
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
