OSDev.org

matio wrote:Next up, keyboard support!

I find it hilarious that you have multitasking, but no keyboard support

@earlz: It's really not that incredible; A PS/2 keyboard driver is technically useless for the first bout of the bootstrapping race. The PS/2 keyboard was used as an example for coding drivers in more than one tutorial, so it has become a de facto 'must have' driver for most projects, even though it really can't be used for anything while a kernel initializes itself, and a logically thinking individual would want to initialize all user input devices at a particular stage in the OS (enumeration, or whatever term you coin), when it simply makes more sense.

In fact, if a kernel is properly designed, one would find it best to leave the initialization of the PS/2 keyboard driver until later on. It's not used, and a static method of initialization really goes an limits how you implement locales (whatever that means for your OS) later on.

Initializing a PS/2 controller driver at early boot makes sense, though.

earlz wrote:

matio wrote:Next up, keyboard support!

I find it hilarious that you have multitasking, but no keyboard support

I find it hilarious that you find that hilarious. A keyboard driver isn't exactly a necessity - I managed fine without one for 8 months in my current kernel's development, and I had multitasking 5 months before a keyboard driver.

AlexExtreme wrote:
I find it hilarious that you find that hilarious. A keyboard driver isn't exactly a necessity - I managed fine without one for 8 months in my current kernel's development, and I had multitasking 5 months before a keyboard driver.

I second that. I had GNU Bash running in a previous project before I implemented any kind of keyboard support.

I have keyboard support hacked in for easy testing (copied from some previous old project), but still no proper driver support. Also I doubt whether I'll ever add PS/2 keyboard support, since USB is the way to go.


JAL

jal wrote:I have keyboard support hacked in for easy testing (copied from some previous old project), but still no proper driver support. Also I doubt whether I'll ever add PS/2 keyboard support, since USB is the way to go.


JAL

Aren't many (all?) internal laptop keyboards still PS/2? Also, at least round here, 90% of keyboards in stores are PS/2 (Which is annoying - I have a USB hub attached to my desk which a USB keyboard would go into...)

Owen wrote:Aren't many (all?) internal laptop keyboards still PS/2?

I didn't think so. If so, I'd have to add support for it of course.

Also, at least round here, 90% of keyboards in stores are PS/2 (Which is annoying - I have a USB hub attached to my desk which a USB keyboard would go into...)

Where exactly is "round here"? Here, in the Netherlands, almost all keyboards are USB, and PCs without PS/2 connections are common.


JAL

jal wrote:

Owen wrote:Aren't many (all?) internal laptop keyboards still PS/2?

I didn't think so. If so, I'd have to add support for it of course.

As far as I know, most laptop keyboards are either PS/2 or connected directly to an internal USB hub.

Also, most systems allow you to turn on PS/2 emulation, so a USB keyboard usually ends up being no different to a PS/2 keyboard (to your OS, anyway). This is how I used to test my OS a couple of years ago on real hardware, as I stopped using PS/2 keyboards a long time ago .

pcmattman wrote:Also, most systems allow you to turn on PS/2 emulation, so a USB keyboard usually ends up being no different to a PS/2 keyboard (to your OS, anyway). This is how I used to test my OS a couple of years ago on real hardware, as I stopped using PS/2 keyboards a long time ago :).

Yeah, I know about the emulation, but I'm not too fond of using legacy stuff. Of course I can use the PIC, PIT, PS/2 and such, but I'd rather concentrate on the latest standards.


JAL

But I'd like to test your OS on my laptop, too.

(Which has a PS/2 keyboard)

jal wrote:Where exactly is "round here"? Here, in the Netherlands, almost all keyboards are USB, and PCs without PS/2 connections are common.

United Kingdom (Updated profile). It's possible the situation has changed; I haven't been in a physical computer store for ~1 year

I would say a PS/2 driver should come before a USB keyboard driver because it is trivial to implement(once you have a HID "Stack" for handling such input in your kernel from any kind of driver) and almost every computer out there has PS/2 or PS/2 emulation.

That being said.. I find having a keyboard is very useful in my OS because I don't have to recompile the kernel when I want to test different things in different orders.. I usually implement a very basic shell in my OSs too wit a special "test" command. This is so I can do `test 10` and run test #10 for my OS and thus don't have to recompile to do past tests and such.. and like I said, PS/2 keyboard is sooo trivial to implement(especially if you just do a hacky implementation ignoring keyboard LEDs and such

earlz wrote: I find having a keyboard is very useful in my OS because I don't have to recompile the kernel when I want to test different things in different orders.. I usually implement a very basic shell in my OSs too wit a special "test" command

I personally am doing things through boot time switches passed to the kernel. Be default it displays a progress bar and splash screen, however I can pass switches to the kernel to configure it to test different things and display information via text mode. No recompiling needed: just edit my bootloaders config file to test something.

I personally rather not rewrite the same driver twice. I rather wait until the core system is up and running and I can write them as user mode drivers instead.

Just my 2 cents.

QFT at the post above me (neon). Microkernels and modularity all the way. UDI and modular mode drivers 'till death. Dynamic loading and unloading of OS components to save memory and make a clean internal OS structure is probably a great design decision.

I have yet to implement anything that would actually intensively do that, so I could be wrong of course.

Being in real mode, keyboard support was as easy as calling interrupt 16h.