 So we're here at the Lenaro Connect here in Vancouver and who are you? I'm Todd Chos, I'm from the Android kernel team at Google. And who are you? I'm Bero, I'm from the Android team at Lenaro. So you've been working on Android for a while, right? Yeah, it's been about seven years. So how's it been? It's really good. A lot of things have improved since I got started, and especially things are getting much more open, there's no longer so many hidden components, and especially also on the kernel side we are getting much closer to being able to run Android on a mainline kernel. Not a hundred percent there yet, but both his team and our team are trying to make sure it's going to happen. So did you do a lot of stuff in that direction? Yeah, so on the Android kernel team we're trying to get a lot better about working closer to mainline, and a lot of the work that our team and the team at Lenaro have been doing have been enabling different kinds of test boards and devices that can run mainline or close to it. Their team has been helping us in getting a bunch of the Android specific things that used to be kept out of tree, into the Linux tree, all of this towards trying to run much more recent kernels on Android devices. One of the big areas that we'd really love to do is phone form factors, and that's been especially hard because in general the SOC vendors have to do a lot of work. They had a lot of code to support the SOCs, and that's done in-house, and in very few of those cases does that go upstream, enabling running phone form factors with mainline kernels. But we're hopeful that this year with Lenaro's help we may have at least one, maybe several. And last year there was a keynote about the project treble. So how's it been going for that? It's been going very well, so I'd say it's been a big success both in terms of modularizing Android, a little less spaghetti code, which the point of that was to enable faster upgrades. When we announced Pi this year we had several partners who were right with us on stage with devices that were shipping on Pi, and that was largely thanks to treble. And in the kernel area it's also resulted in better processes and disciplines around taking patches from the long-term stable, and I think it's resulted in more solid kernels. And is the secure support a big deal? It is a big deal because without that, since devices typically ship when the kernels are already two years old, a two-year support model just didn't work because those critical security and bug fixes that normally would come down from LTS, they stop when it goes out of support. And so having a six-year support really enables us to have much better devices. So you've been yourself, for example, a lot involved with Lenaro, connecting talking with them? Yes, so you're very involved with the teams doing testing of kernels and that's part of why this LTS merging is working well because of the testing that Lenaro's doing. We also have folks like Barrow working on AOSP and making it more performant and we've got other teams within Lenaro helping us take patches and bring them upstream, so Lenaro's been key to all of this. So for many years I've been doing videos with you, and at one point it felt like the Lenaro Android had the speed up because there was some acceleration kind of happening. Is this stuff that Google took in and kind of, you know, what do you mean? Is it all open source? So I don't know the specific ones, but there have been plenty of cases where Lenaro's helped us with performance. I think one area that stands out is the work that they help with in art and the JIT compilers for Android apps. There's plenty of other examples of performance help that Lenaro's given. What are you doing with the art, for example, and that kind of stuff? Yeah, so the particular speed ups you're talking about is probably what we did with Demoing four or five years back in Hong Kong, which was essentially the Android rebuild with newer tool chains. The tool chains in AOSP have always been a little bit behind, especially when it was still using GCC. And we used to make a big effort here, moving it to newer GCC releases that has all gone upstream. And nowadays, of course, Android has moved to using Clang as the primary compiler, and we are still trying to help out there. We have a system setup that runs nightly builds of the LLVM in Clang tool chain and then builds AOSP with that and notifies us if anything goes wrong. So we immediately upstream any changes needed to make newer versions of Clang work. And we also find a lot of bugs in the compilers that can be fixed before release by doing that as a big compiler test. So are you involved with the LKTS, what's called the Lenaro testing? KFT. You're involved with that? Yes, somewhat. I'm mostly working on the user land side of things. My primary work is on the tool chain things we just talked about. But I'm somewhat involved definitely looking at the test results all the time and sometimes helping figure out where regression came from. Just making sure that we don't send total crap patches upstream. And the LKFT, so you've been using that? Absolutely. So LKFT has become critical to testing the kernels both at the LTS level where the long-term stable releases when Greg H puts them together. LKFT has been used to test that and the RCs. And then when those merge down to common, LKFT is used there to verify it. It's been critical to verifying the kernels. And you mentioned Greg many times. Greg Korhachman, he's an important guy that you have to work with. Yeah, so he's one of the key kernel maintainers and he also works closely with us. So he actually does the merges from the LTS kernel to the common kernels. We meet with him regularly, work with him. And then Lenaro has been working with him closely on LTS just to find the right patches to help look through other projects and find where the bug fixes and potential security fixes are. And those are used by Greg in creating the LTS releases. So Lenaro has been very involved in that as well. Have you been involved with the Android since the beginning? No, I've been involved in Android for almost four years. And Android is so cool. It's awesome. It's the best, right? Because it's also the platform that made Linux the most popular, right? I mean, it's the most popular Linux thing out there. Is it one of those, at least? So it's a big deal that you make things the right way. It is certainly a very large Linux distribution. I don't want to overclaim, but definitely it's important to the Linux community. And it's important that you do security the right way. So is it possible that I can ask, how secure is Android? Is it like perfect? Or there's some issues? Of course it's not perfect like any large software project, but we're working to make it better all the time. Because it's kind of the architecture of Android is siloed and there's a lot of things happening through the Play Store, security stuff, right? It scans all your apps and you don't get malicious apps. That's important. There's a lot of work being done on a lot of fronts to make it secure. And it is quite secure. And just also it's a huge work to make it work. So Google is making it easier for all the manufacturers and all the chip vendors always becomes easier than easier? I'm not sure that the vendors would agree with that, that we're making it easier and easier, but we have good work in relationship. And I think it works pretty well. Cool. Thanks a lot. Thank you.