00:31:33 Kate Hendle: Welcome to the OSF Tech Talk! Please keep yourself on mute and put any questions for the speakers here in the chat! 00:32:34 Archna Haylock: you can find all the recordings/slides/chat transcripts here: https://www.opencompute.org/events/past-events 00:32:49 Ron Williams: Thanks! 01:01:54 Sainath: how can community developers test their work on real hardware? 01:02:44 Arthur Heymans: Not a question, but awesome talk! Thanks Ron 01:02:55 Archna Haylock: if anyone else has a question, please unmute and ask! 01:03:42 Chris : Yeah - thanks Ron, great talk, great shirt. 01:04:07 ron minnich: Glad you liked the shirt :-) 01:05:14 Ron Williams: Hi Folks. I don't have a mic available. I'm sure the CS folks at my University would love to play with some of this. So we can buy boards off the shelf now? 01:05:16 Adonay: what framework/orchestration tool is used for reserving these donated systems? 01:06:12 Adonay: referring to OCP community lab infrastructure here 01:06:28 Ron Williams: OK. I need to talk to the CS folks. I work at the Vet School at UC Davis. So it's going to be a bit of networking. 01:06:51 Ron Williams: Awesome, Thanks! 01:07:13 Bryan Cantrill: Hopefully it's a smoother process than when we bought a Tioga Pass from Wiwynn several years ago 01:07:32 Bryan Cantrill: It was absolutely brutal -- I'm convinced that we were the only ones who bought one 01:07:45 bruningh: Does the OCP community lab allow IPMI RMCP access to the system manager ? 01:07:55 Bryan Cantrill: Huge thanks to David Hendricks, without whom it wouldn't have happened at all 01:08:56 ron minnich: David has a way of getting stuff done :-) 01:09:32 Bryan Cantrill: Wiwynn acted like we were trying to buy a human cadaver 01:09:52 Nill Ge 葛士建 (ByteDance): Thanks Ron and all, great talk! 01:10:21 Bryan Cantrill: Thank you Ron! 01:10:22 David Hendricks: Thanks :-) Everything's gotta start somewhere, it's really the community that has driven this to fruition. 01:10:35 archagam: Nice talk Ron!! 01:11:42 ron minnich: Thanks, glad it worked out :-) 01:12:10 ron minnich: This status slide is just fantastic 01:15:37 archagam: 👋 01:20:09 ron minnich: This talk is really showing collaboration! 01:21:51 Sainath: when will all the work presented here available upstream? 01:22:10 Sainath: be available* 01:22:13 Arthur Heymans: All the deltalake work is upstream 01:25:28 jonathan zhang: SPR-SP support and SPR-SP based server support will be posted upstream as soon as SPR-SP product is launched from Intel. 01:26:31 bruningh: On the TPM how many PCR registers are needed out of the 127 allowed? 01:27:27 Arthur Heymans: So I think 1 is used by CBnT and 1 in coreboot. As PCR are append only that works 01:28:02 ron minnich: We LinuxBIOS OG’s can’t resist the name :-) 01:28:04 jonathan zhang: @bruningh The DeltaLake server in OCP community lab has OpenBMC as well. The communication between OpenBMC and host firmware is mostly through IPMI. The OpenBMC is open source as well. 01:28:41 Chris : coreboot uses four PCRs, if vboot is enabled. CBnT uses PCR-0 01:28:56 Daocheng Bu(Amos)@Intel: Thanks Syspro for ACPI related support from Xeon-SP coreboot! 01:29:35 ron minnich: Yep! 01:32:31 marc jones: lol, I just read that. I have a typo! *sheesh* 01:33:57 Lowell Dennis: I'm hearing a weird echo! 01:34:00 Sainath: hearing echo 01:34:02 Arun Koshy : echo 01:34:27 bruningh: Yes, the PCR question has been cleared up. Thanks! 01:35:05 Chris : Sorry for the echo - not sure what went wrong. 01:38:35 Sainath: How to participate in the regular OSF meetings? 01:39:01 Sainath: are they open for individual developers to attend? 01:39:22 archagam: https://global.gotomeeting.com/join/829370213 01:39:32 archagam: https://global.gotomeeting.com/join/829370213 01:39:48 Archna Haylock: @sainath, please subscribe and see call details here: https://www.opencompute.org/projects/open-system-firmware 01:39:56 Archna Haylock: anyone can particpate! 01:41:25 Bryan Cantrill: I presume this is not, in fact, Intel Confiential? 01:41:34 jonathan zhang: @Bryan I think DeltaLake will give you a nice surprise. Please feel free to contact me if there is anything I could help. 01:41:52 ron minnich: Yeah would be good to fix those markings :-) 01:42:28 Chris : Not confidential anymore now I guess. :) 01:42:57 Bryan Cantrill: @Jonaathan: I almost want to try to buy one just to see how the process has improved. ;) 01:43:11 Bryan Cantrill: Can I buy one with a credit card? 01:43:29 Bryan Cantrill: (Paymet was the first of many, many problems that we had.) 01:43:42 jonathan zhang: @Bryan I believe so. I could get you in touch with WW sales. 01:44:04 Chris : @Bryan - you can also get access to one of ours if you only want to try it out :) 01:44:18 Chris : We also enabled the OCP Test & Validation Group on our DeltaLakes. 01:44:55 Johnny Lin: Hi Bryan, Wiwynn will improve the process for sure, sorry it was not working so well from the past. :-( 01:45:35 jonathan zhang: Thanks @Chris. Yes, with the global supply shortage and with us buying so many servers, it may take months for you to get a server on your desk. In the mean time, as Chris said, he could arrange remote access. 01:45:56 Chris : Sure. Feel free to ping me for that. 01:49:02 jonathan zhang: @Bryan : DeltaLake docuemntation: https://review.coreboot.org/plugins/gitiles/coreboot/+/refs/heads/master/Documentation/mainboard/ocp/deltalake.md 02:10:32 smith: It's never me, im a cat. 02:13:58 ron minnich: “Avoid copy-left issues”? 02:16:08 Arthur Heymans: Very 'open' :-p 02:17:23 ron minnich: Modules -> late binding -> vuln 02:18:21 Sheng: there seems to be a fundamental differences between an open source based solution and "open sourced based" closed source distributed& modified solution 02:19:16 bruningh: Does SBOM include what patent license are required to use the software? 02:20:23 Adonay: I think you're referring to the difference between an open-source product vs open-source business model; i.e., Open-core. @sheng 02:24:05 Chris : coreboot actually implements SBOM already with open-source tooling to generate and export the SBOM ingredients. https://review.coreboot.org/c/coreboot/+/63639 02:24:28 Sheng: @adonay how would you define an open-source product? Almost all software products contain part of open source code in today's wor 02:24:38 Sheng: *world 02:26:27 Adonay: agreed! 02:27:18 ron minnich: When I tested all the URLs in repos a few years back, 25% were stale. Could that be a problem here? How long does this URL need to be valid? 02:31:43 Adonay: This is a good question from the chat... 02:31:44 Adonay: Does SBOM include what patent license are required to use the software? 02:33:41 Sheng: @Adonay to be fair, a healthy open source ecosystem requires a community of open-source business model, which i am surprise to see how coreboot grows to today's size thanks to those contributions 02:34:25 Steve Klabnik: 🎊🎊🎊 02:36:32 Corvinux: 🎊🎊🎊 02:39:27 ron minnich: In the beginning,SMM is what happens when really bad software and really bad hardware design co-exist 02:39:31 ron minnich: :-) 02:40:46 vincent zimmer: some ancient history on SMM https://www.amazon.com/Intels-Architecture-Designing-Applications-McGraw-Hill/dp/0079113362 02:46:01 bruningh: On the SBOM presentation, it mentioned APSEED, they have patents on IPMI/BMC. Others do AWS etc, Dell filed 30 so far this year-2022, HPe 20. If open source is going to compete with the patents they have to implement the patents, there are 600+ IPMI BMC patents. Could SBOM identify what patents so I can tell my customers they are free and clear of patents? Hank Bruning 02:48:01 ron minnich: Hank, sounds like oxide’s decision not to use BMC makes a lot of sense? 02:50:29 ron minnich: Doesn’t sound unachievable, we did this back in linuxbios days :-) 02:54:11 ron minnich: By which I mean the phased stuff 02:54:13 Chris : Great talk Bryan :) Thanks! 02:54:16 Ron Williams: Great talk! Exciting stuff. 02:54:28 Steve Klabnik: https://github.com/oxidecomputer/ <- for convenience 02:54:38 ron minnich: Glad you made it Bryan thanks — do you think the SBOM stuff affects you at some point 02:54:41 Sheng: really impressive presentation! 02:54:41 Ron Williams: Thanks Steve! 02:54:51 bruningh: How do you power management , turn on on PIC slots, blades, the AdvancedTCA model without a BMC? 02:55:13 Dan: For the phased approach, to avoid reconfiguring something like your SPI controller once the OS starts, is configuration pushed into SPI flash? Or is that an example of a peripheral that just stays in the start up configuration? 02:55:52 vincent zimmer: interesting approach. this means that the oxide model is hosted VM's only w/ Bhyve. No bare-metal service offering? Latter often need to expose industry std. interfaces to work w/ shrink-wrap iso's, etc 02:55:54 David Hendricks: +1 to open-source PSP :-) 02:57:10 moritzfago: Where do you want Oxide to be in 10 years time? Sun 2.0? 02:58:47 Bryan Cantrill: With my apologies, I need to run to a meeting -- thank you, everyone! 02:58:54 Chris : Thanks Bryan! 03:00:05 Hash: Thanks Bryan! 03:07:47 bruningh: Are there hooks to DMTF ? 03:10:43 Adonay: Is the intent to use this CI tool for open-source fw products? 03:11:11 Adonay: Or is it a solution that others will have to pay for local deployment? 03:16:07 ron minnich: Uh oh 03:20:14 ron minnich: question: do you think direct TFA -> Linux handoff might be possible in future? 03:25:05 Arthur Heymans: x86 has a way to pass ACPI via the Linux zero page, without depending on the legacy presence (lowest 1M) and without depending on UEFI. Something similar would be great on ARM, you get ACPI without a complicated UEFI dependency :-) 03:31:05 ron minnich: question: does neoverse work with upstream kernels now? 03:31:20 ron minnich: question: please get me these talk links for showcasing on linuxboot.org? 03:33:41 ron minnich: this is wonderful thnx