Newsletter #26 - June 2024 (GarlicOS & Duo updates)

The newsletters are back!

Those who have been with me for a while know that I'm the kind of person who prefers to work in silence, come out of hiding every once in a while to drop a release, and then go straight back to working in silence some more.

I guess you could say I've never been much for idle talk.

This approach allows me to focus on the work, get things done, and be as efficient as a one-man army can be, but it also causes rather long periods of radio silence that might be misinterpreted as something else.

I used to do frequent newsletters containing technological write-ups of what I'd been working on behind closed doors, which helped bridge these gaps. However, with ever-increasing workloads and the need to make time for more and more complex projects, these were among the first things that had to be cut down for what I assumed would be the greater good.

But I realize that people want to be kept in the loop—I know I would—which brings us back to the topic at hand: the newsletters are back!

How's GarlicOS coming along?

I know that a lot of you are here for GarlicOS, so let's take a closer look at what's being worked on and what's been grinding my gears for the last few weeks.

Going back a few weeks, I dropped an experimental update for GarlicOS that brought Anbernic's new Micro-Handheld, the RG28XX, into the fold.

This acted as something of a framework test to see how hard, or easy, it has become to port GarlicOS to newer devices.

The goal is, after all, to make Retro-OS development easy to understand and achievable by everybody regardless of the device's hardware platform or the person's technological background.

In that regard, things looked quite rosy, with less than eight hours needed to port everything from start to finish, not including the soldering and sanding work needed to get the UART wired up and looking pretty, that is.

With that part cleared, I decided it was time to change focus towards unlocking as many advanced hardware features as possible, ranging from WiFi support to hardware acceleration and all the prerequisites that go hand-in-hand with these.

A week or so later we had working hardware acceleration in my internal testing builds.

Once again, things looked rosy, except for the fact that performance was still subpar at best when compared to other Retro-OSes out there.

Which brings us to the final stretch of work-in-progress GarlicOS development.

I've been busy building, testing, and debugging several hundred GarlicOS builds over the last few weeks, trying to pinpoint the cause of these performance issues, but so far remain unsuccessful in isolating a definite cause.

I initially suspected that these stemmed from the previous lack of hardware acceleration, but now that we have that part sorted, it's become crystal clear that these issues stem from somewhere else.

Not wanting to deliver something underbaked, I decided I'd take the time I need to figure out just what is causing these issues, which I hope explains the recent delay.

I'll make sure to keep you all posted going forward.

How's Duo coming along?

It's no secret that I'm a big fan of game streaming.

Viewing it through the lens of handheld gaming, which is my primary way of consuming games, there's always been this divide between running games locally with reduced graphical fidelity, accelerating them via the use of eGPUs, or streaming them.

And while I prefer to play things locally, if that is an option, I've always found eGPUs to be too clunky and expensive for my liking.

Given my extensive use of game streaming, it didn't take me long to pinpoint its weakest links: the barrier of entry, setup requirements—both physical and virtual—and, most importantly, the negative side-effects it imposes on the host system by occupying its console session during use.

One by one, I tackled these, implemented solutions for them, and built what is now dubbed Duo, a spiritual successor to Castor and my take on what I'd consider a wireless and affordable eGPU for handheld use.

As part of this ongoing journey, Duo has seen several improvements over the last few weeks, with the biggest one being a full rewrite of its monitor driver in version 1.4.6, moving it from process library injection to its own UMDF driver wrapper library.

This rewrite allowed us to rid ourselves of one of Duo's biggest weak points: the need for Memory Integrity to be disabled for it to function properly.

Another change was the recent replacement of its RdpWrap dependency with TermWrap, which made the whole system a lot more resilient to Windows Updates.

I hope that by carefully rewriting certain parts of Duo, we'll eventually reach a point where antivirus false positives and Windows Updates are no longer cause for concern.

Time for a summer break

As some of you know, my start into 2024 wasn't the best.

I contracted pneumonia sometime late last year, and it took me a considerable amount of time to recover from it fully.

Things are back to normal now, the projects are moving again, and overall life's been good since my recovery, but these troubling months did leave scars.

They made me realize that there's more to life than work, and to celebrate that, I've decided that I'll be granting myself a much-needed 2-week break sometime this summer to detox from work & tech—something I haven't done in the last 3 years or so.

I haven't decided on the exact date yet, but I'll let you all know ahead of time.

With all of this said, there's only one thing left to do, and that is to thank you all for your continued support!

I wouldn't be able to work on the things I do without all of you, and no day goes by where I'm not thankful for the opportunities I've been given by all of you.

I wish all of you a wonderful summer break!

- Black-Seraph