 Great, so welcome to 2023 Linux storage file system memory management, BPF, basically the entire internal at this point. Yeah, three minutes. I'm very sorry. First off, I'd like to thank the wonderful program committee. You guys can stand up. Got Martin Javier for storage. Daniel and Martin, there you are, Martin. For BPF, me and Amir for file system, but really just Amir because I'm lazy. McCull and Dan. Where's Dan? Dan Williams is around here somewhere. So big shout out to these guys. They did a fantastic job organizing the schedule and everything and keeping this whole thing running smoothly. We'd also also like to thank our sponsors, Diamond, Meta, Platinum, Samsung, Gold, IBM, NetApp, NVIDIA Seagate, Silver, EBBF Foundation, and Intel. Without their funding, we wouldn't have such a nice place to be. So thank you to our sponsors. Just like last year, we are hybrid virtual in person. So clearly there's a lot more of us in person this year, but there is still a significant number of people that are virtual as well. This is going over Zoom and the other breakout sessions, the camera for the guys on Zoom, the camera will pan to the audience so you can see who's talking. This is always important, but especially important in these cases. If you're going to speak, microphones. And we'll pass these out and we'll chase you down. If you want to speak, just raise your hands. We'll get a microphone to you before you start talking. I feel like there was something else and I should have written it down. If you're virtual and want to speak, raise your hands. The people in the back are keeping track of the Zoom stuff and that'll let you know. For me, I also will log into the Zoom thing when I'm in a room. And so I'll keep track of questions and stuff because sometimes people just want to ask a question. Don't really want to be put up on the screen or whatever. I kind of leave it up to you guys with the tracks, but there are Zoom links on the schedule so everybody can log in. That's what I plan to do. Does that sound good? I try to get a list of virtual attendees. I just have a list of all of the attendees. So if you guys are missing key contributors that you think might be virtual, either reach out to them directly or log into the Zoom and check and see where they're at. Perfect. So that's Zoom. The schedule is on the... If you get a Linux foundation, you get the schedule. You can go to the Google Docs and you can see it there. We will be updating this. I know I have to add the sFrames thing and I think it's on the schedule or everything. We're good? Excellent. Perfect. All right. We will have room for boffs. We'll get a thing up here to suggest for boffs or not boffs. Lightning talks, thank you. Sorry. I'm very tired. But yeah, if you guys have any questions, have any issues, feel free to hit any one of us there on the program committee. It's been a pleasure doing this and you'll notice there's a lot more people here. That's another thing. We increased the attendance to 150 people this year. The goal here is to allow for more junior people. I'm honestly getting tired of seeing the same faces all the time. So it's kind of nice to see the new people and make sure to get them included. Like obviously those of us who've been here a while are going to talk a lot. But if you're new here, feel free to jump in and if you want to just watch, that's also cool too. So this is kind of like trying to help grow the community and kind of like add more support for everybody. As usual, I'm way ahead of schedule. We're going to talk about like the last years of development, like the last year worth of development and stuff. We're going to just have the track leaders kind of come up and do their spiel about what's happened. I don't know what's happened because I don't pay attention. Who wants to go first? Y'all storage? You want to go first? You and Martin? Well, I don't pay attention either. So absolutely nothing has happened the last year in storage. And I can say that conclusively because I had to look at the diffs this morning when Joseph said, can you say what's happened? So we haven't had any major, major developments or new technologies added. It's been like a slow and steady refinement and incremental amount of work. But I don't have any big flashy, exciting things that we did this awesome thing. But really the storage stack is awesome. It's we've just been polishing and polishing. Perfect. I know BPF has done a bunch. Oh, Amir, you want to come talk about file systems? Better you and me. I don't know what's going on. Yeah, so nothing new is file system either. I don't think that we have any new file systems. I was looking at the last year's schedule to see if anything comes out of these gatherings. And you know, we have several types of talks in these meetings. We have the occasional new thing someone is trying to propose. And we have the reoccurring topics that we bring up every year, hopefully with some progress. So I specifically looked at the things that reoccurring. And maybe some stuff that was new last year and really evolved. So I hope I didn't miss much. But we have, we usually have a discussion about file system and stable trees. And we're going to have one. We have had one of those for the past four at least meetings and we have one this time. And there was progress, but we're going to show the progress in the meetup. Also a general talk about maintaining file systems. Joseph Tech gave this talk last year, but we don't have a specific talk about this this year, but we do have a new VFS Co-Maintenance. So that's a progress. Christian Barano, if you don't know, know. What else? There was a talk about IDMAP mounts and there was certainly progress in this area over a year. There was the traditional talk about FSA info. And there's one this year and nothing happened on that front yet. This year it will happen. And there was a talk and there's going to be a talk about FS test and BLK test and KDevOps. Where's Luis? I don't see him. Hey, Luis. There has been progress. KDevOps is a growing community now. And you will hear about that. There was some talk about extending Stadex and FSID for better fests. I mean, things have happened, but I don't know how much, not much. That's it for... Oh, there was a NetFS. There was a talk about NetFS. And I don't know if we talked about IOMAP conversions, but certainly things happened with IOMAP conversion this year. XT4 got a bit more support. I hear that XT2 is queued up for IOMAP conversion. But for me, the most important thing that happened is that XFS, which is the spearhead of IOMAP, has detected some design issues, actually, that were tackled with stale mapping and such. And that was important because it's important that we have XFS doing this work for the rest of us. So thank you about that, XFS users. And that's it. What else? So what we have... Yeah, there was a talk about file system notification for remote file systems, but that didn't make any progress. There's going to be one talk about file system notification this year, but something else. What are we going to have this year? Apart from assorted topics. I'm not going to list them all. We have a sort of a Fuse mini conference because a lot of exciting things happened in Fuse recently or new development. So we have a few developers here. And this afternoon, we're going to have a sort of a Fuse mini conference. That's going to be interesting. We have a VFS mini conference, I think, Wednesday with the new VFS call maintainer. Tuesday. Thank you. And I'm just going to look at the schedule and I missed something. Yeah, we have a new interest coming up with the image, how do you call it? Images, file system images. A lot is going on with distributions that are distributing file system images and new technologies like Composer Fest that is going to be discussed this year. So new stuff is happening there. Next up. Yeah, so quite a lot has been going on in the memory management land. And I will not list all of that because I would certainly forget about something. But first of all, one and a big change from the last year is a new memory management process where Andrew has started using Git workflow and it seemed to be working pretty well, at least from what I've heard. So this is a nice success story after the last year. Apart from that, multi-generation LRU has lifted off and that's also a result of the last year discussion which is also really great. And yeah, Folios are still moving forward. I think that Matthew is seeing more contributors to that work which is also nice. And M-MAPSAM, Scatillability has moved forward some not yet there, but RCU, Hotpath is almost implemented. It's just waiting to be finalized to the measurable state but Surin is doing a really great job in approaching that goal. So yeah, I think that since the last year, nice success stories and yeah, it seems like we almost have to refuse some of the proposals this year but in the end we've managed to fill up the schedule without rejecting anybody and there's still one slot so if there is something popping up from discussions then feel free to reach out to me and we will be more than happy to put something on the agenda and I can see Andrew lifting his hand. So yeah, and another really major change is the Maple Tree being merged. So yeah, anybody from the BPF lands want to? All right, so nothing happened in BPF either since last year. No, so like a lot of the development is driven by new use cases in the kernel. For example, we see more evolving around the security angle by supporting LSMs. Also tracing and scheduling is a big topic which also helps to improve the core infrastructure of BPF itself, of the verifier to make it more easy to use for people to consume. I'm sure I will forget a lot of things but one thing that makes me very happy as a co-maintenor is also that we have a wonderful BPFCI which is evolving also into different architectures, tons of more tests which helps to stabilize the infrastructure around BPF in the kernel. A lot of the improvements are also around K-funks to be able to make it easier for users, for example, around file systems or in other areas like the HID improvements to be able to consume it in their own and to be able to develop it in their own subsystems further, to be more independent of BPF itself, to evolve it more independently. Yeah, so I think. Excellent. Thanks, guys. Jens is not here so he doesn't get to speak. There's been a lot of IOU ring stuff. I know he keeps talking about it. I printed a listen. Do we have anybody who can speak? I mean, I know we do. Pavel, are you around? All right, fine. Don't volunteer. Anyway, lots of work in the IOU ring, too. It's going to be on the list. So lots of cool stuff going on, especially the, I don't know, I'm really excited about the few stuff, too. All right, still way ahead of schedule. I'm doing great. Why did I think this was going to take 45 minutes? Absolutely. Yeah, there's a click on the bottom. You take this one. I'll take this one. Thank you. So the question I keep getting a lot that I don't have an answer to you is for the BPF people. So Alexei being here, I will move next to you so you can answer it. I have a load of people basically from academia whose question is, now that we have Rust in the kernel, why do we need BPF? I think there's an answer, but I haven't been able to successfully articulate it. I need help. How about you will listen to my presentation in half an hour, and I think it will answer it. Perfect. Okay, so I guess we're just going to roll right into the next thing.