Not only is this an insanely cool project, the writeup is great. I was hooked the whole way through. I particularly love this part:
> At this point, the system was trying to find a framebuffer driver so that the Mac OS X GUI could be shown. As indicated in the logs, WindowServer was not happy - to fix this, I’d need to write my own framebuffer driver.
I'm surprised by how well abstracted MacOS is (was). The I/O Kit abstraction layers seemed to actually do what they said. A little kudos to the NeXT developers for that.
I felt similarly. The learning curve was a tad steep, especially since I had never written a driver before, but once I figured out how to structure things and saw the system come alive, I grew to appreciate the approach IOKit takes.
With that said, I haven't developed drivers for any other platforms, so I really can't say if the abstraction is good compared to what's used by modern systems.
IOKit was actually built from the ground up for OS X! NeXT had a different driver model called DriverKit. I've never coded against either, but my understanding was they're pretty different beasts. (I could be wrong)
That said, indeed, the abstraction layer here is delightful! I know that some NetBSD devs managed to get PPC Darwin running under a Mach/IOKit compatibility layer back in the day, up to running Xquartz on NetBSD! With NetBSD translating IOKit calls. :-)
There’s a great video of a NeXT-era Steve Jobs keynote floating around—I think the one where he announces the x86 port as NeXT was transitioning to a software-only company—where he specifically calls out DriverKit and how great it is.
Steve was not a developer but he made it his business to care about what they cared about.
Yeah - even from the start, I remember NeXT marketing was spending a disproportionate amount of their time selling NeXT’s “object technology”, AppKit and Interface Builder, DPS as an advanced graphics model. It was good hunch from Steve, given how how modern NeXTSTEP feels in retrospect.
For some reason, though, it means that people overlook how NeXT’s hardware was _very_ far from fast. You weren’t going to get SGI level oomph from m68k and MO disks.
Yes, you're right! I'm just dolt who's never checked what a .kext on OS X actually is.
I had been under the impression that DriverKit drivers were quite a different beast, but they're really not. Here's the layout of a NS ".config" bundle:
The driver itself is a Mach-O MH_OBJECT image, flagged with MH_NOUNDEFS. (except for the _reloc images, which are MH_PRELOAD. No clue how these two files relate/interact!)
OS X added a dedicated image type (MH_KEXT_BUNDLE) and they cleaned up a bit, standardized on plists instead of the "INI-esque" .table files, but yeah, basically the same.
IOKit was almost done in Java; C++ was the engineering plan to stop that from happening.
Remember: there was a short window of time where everyone thought Java was the future and Java support was featured heavily in some of the early OS X announcements.
Also DriverKit's Objective-C model was not the same as userspace. As I recall the compiler resolved all message sends at compile time. It was much less dynamic.
Mostly because they thought Objective-C wasn't going to land well with the Object Pascal / C++ communities, given those were the languages on Mac OS previously.
To note that Android Things did indeed use Java for writing drivers, and on Android since Project Treble, and the new userspace driver model since Android 8, that drivers are a mix of C++, Rust and some Java, all talking via Android IPC with the kernel.
Yes, also the same reason why Java was originally introduced, Apple was afraid that the developer community educated in Object Pascal / C++, wasn't keen into learning Objective-C.
When those fears proved not true, and devs were actually welcoming Objective-C, it was when they dropped Java and the whole Java/Objective-C runtime interop.
And there are enough parallels to Linux's stack, I'm thinking about looking through the Linux on Wii project more and comparing how it handles fb issues in comparison. I loved reading this whole post, crazy how many OSes have now been run on the humble Wii!
I’d say it’s more about how much explanation is needed. There are cool abstractions that require explanation because they aren’t intuitive at first, and then it clicks. But usually if I find endless explanations of why indirection is better because it aligns with someone’s conceptual model, that’s to me a bad abstraction. Not because it’s leaky, but because it resists understanding.
> At this point, the system was trying to find a framebuffer driver so that the Mac OS X GUI could be shown. As indicated in the logs, WindowServer was not happy - to fix this, I’d need to write my own framebuffer driver.
I'm surprised by how well abstracted MacOS is (was). The I/O Kit abstraction layers seemed to actually do what they said. A little kudos to the NeXT developers for that.