**** BEGIN LOGGING AT Sun Mar 07 02:59:57 2021 Mar 07 03:00:05 I am going to connect it to BELA for extra ordinary sounds, i.e. not extraordinary! Mar 07 05:02:56 maybe I should do a psuedo guitar hmm Mar 07 06:08:57 hey! Mar 07 15:44:26 my BBAI takes a long time to boot caused by "Timed out waiting for device /dev/ttyGS0.". What does this refer to and is there a way to disable it? Mar 07 15:49:23 lolleren: I'm guessing you disabled the usb gadget? Mar 07 15:50:22 that device is the usb serial gadget, I'd guess this error is from its getty service instance Mar 07 15:50:55 try: sudo systemctl disable serial-getty@ttyGS0.service Mar 07 15:58:08 Yes I disabled several services some time ago. probably also usb gadget. Thanks, I'll try that Mar 07 20:10:06 Hello! I'm very excited about the BeagleV. I would like to ask if everything is on schedule, and whether the first boards will arrive next month. Mar 07 20:11:44 has anything ever been delivered on time? Mar 07 20:16:15 there at least seem to be prototypes, pdp7's twitter shows he's working on bringing up linux Mar 07 20:16:50 or at least having a fight with u-boot on one... maybe linux is already running fine on it Mar 07 21:07:31 :) Thanks for this Mar 07 21:17:06 mark[m]2: there will be 300 beta BeagleV's that Seeed is planning to ship first week of April. They will be sent developers selected from the list of those that filled out the register-interest form. https://docs.google.com/forms/d/e/1FAIpQLScCwms7bcJDSkM1kDDKL6aGpVlRbLskv41lqO3XDfvyV6yZZw/viewform Mar 07 21:17:59 Currently, 5.10 kernel with changes from StarFive is booting on the BeagleV. Fu Wei from the Fedora team has a basic Fedora image booting on it. Mar 07 21:18:04 :) Thank you. This is great news. Mar 07 21:18:31 "kernel with changes from StarFive" <-- that sounds scary Mar 07 21:18:42 Did you fill out the form? We are looking for people that have experience in linux and uboot and other low level software. Mar 07 21:19:29 Yes, we are using their fork of uboot, opensbi and linux. The current dev environment is based on SiFive Freedom U SDK which uses buildroot Mar 07 21:19:54 The goal is to get the changes upstreamed and maintainers from all those projects are on the beta list Mar 07 21:20:04 mru: that sounds perfectly normal for a new SoC Mar 07 21:20:11 mainline support takes time Mar 07 21:20:15 normal and also horrible Mar 07 21:20:16 In April we should finally start getting their feedback on how much needs to be changed for upstreaming Mar 07 21:20:18 I'm a long time Linux user... but I'm not a dev. I'll need to wait a while - probably until the version with the GPU comes out. Mar 07 21:20:55 the good thing is that the RISC-V arch maintainers in u-boot and Linux are highly motivated for affordable hardware so we should be able to get them heavily involved Mar 07 21:21:23 pdp7: getting feedback only after things are working instead of while they were being worked on sounds like a recipe for having to rewrite a lot? Mar 07 21:21:23 :D Mar 07 21:22:52 zmatt: yes, i would have liked to have gotten the upstream maintainers involved months ago but StarFive has not yet understood how valuable that can be. The upside though is that affordable Linux capable RISC-V boards are rare so they upstream developers are highly motivated to get things working. Mar 07 21:23:35 I'm working with StarFive to try to get as much documentation as possible to be released because that is what will really enable the upstream developers to get the StarFive SoC well supported Mar 07 21:23:44 indeed Mar 07 21:23:50 I do agree that a lot of what StarFive software team has done will probably get replaced Mar 07 21:24:32 the good thing is that Fu Wei of Fedora has been involved with bring up and he understands the upstream RISC-V work that is being done and is trying to bridge that gap ahead of time Mar 07 21:26:38 * mark[m]2 sent a long message: < https://matrix.org/_matrix/media/r0/download/matrix.org/JwNQjKadvycTBjuzyayXRDCy/message.txt > Mar 07 21:26:41 Currently I'm working with StarFive to try to help them figure out how to get permission to include register level documentation relesaed publicly for the peripherals they licensed. The SiFive core will be fully documented but we are trying to navigate the other IP vendors. The good news is that most of those peripherals already have upstream drivers in Linux and uboot. But still I want to get as much public as possible Mar 07 21:27:25 Fedora RISC-V in general is very strong, similar to Debian. They key peices for BeagleV is getting the SoC well supported in uboot and linu Mar 07 21:27:29 *linux Mar 07 21:27:53 Fu Wei has been involved with the uboot and Linux bring up for the SoC so getting Fedora running is natural for him Mar 07 21:28:24 We (beagleboard.org) would like to see many options documentated and working like Debian, Fedora, buildroot, yocto, etc Mar 07 21:28:54 the biggest challenge is getting the SoC well supported as those projects already support RISC-V pretty well Mar 07 21:28:54 :) Absolutely Mar 07 21:29:44 public documentation is indeed important, especially for the risc-v audience, given how they're carrying the "open" banner Mar 07 21:30:08 Included in the beta will be u-boot and tianocore developers that have been working on RISC-V UEFI support so, in the future, it may be possible to support generic RISC-V distro images in the future. Mar 07 21:30:42 That would be great Mar 07 21:31:53 StarFive has CC licensed doc for the core IP from StarFive. The bigger challenge is getting permission to publish register maps for the synopsys and cadence IP. Some things are simple like I2C and SPI but other things are more complex like USB. We should learn more in the next few weeks as to what can be published. Mar 07 21:33:16 There is also some IP that is from StarFive that they will be able release documentation themselves, including some simple things like GPIO but also more complicated things like the ISP Mar 07 21:34:09 I am hoping that in the next few weeks StarFive will make their uboot, linux and opensbi repos public. That at least needs to be done when the beta starts shipping. The device tree bindings will shed more light on the IP that is used. Mar 07 21:34:34 I still don't really understand what IP vendors have to gain from keeping documentation confidential Mar 07 21:34:39 I agree Mar 07 21:35:01 It is interesting what Allwinner does where they let the datasheets leak Mar 07 21:35:13 "whoops, how did that happen" Mar 07 21:35:31 they're probably afraid that someone will use the docs as evidence of patent infringement Mar 07 21:35:32 so all the Allwinner SoCs have datasheets linked from sunxi and even kernel.org but many of those have watermarks Mar 07 21:35:58 mru: -.- Mar 07 21:37:40 StarFive is trying to be careful and get all the necessary permission from the IP vendors they used. I think for peripehials were the IP vendor says no, then hopefully we can at least state the IP and the version and the device tree bindings will make it easy to know the specifics because the peripherals is already supported upstream. For example, the Gigabit Ethernet uses the existing stmmac driver. Mar 07 21:38:29 the existence of public drivers makes keeping documentation secret even more bizarre Mar 07 21:38:30 But I want StarFive to first try to get as much permission as possible to just release the register level documentation Mar 07 21:38:57 Yeah, I think part of it may just be that most SoC companies don't even think to ask Mar 07 21:39:53 Before we met StarFive, they were not thinking about making any documentation public as the normal situation is they would just provide documentation to their direct customers under NDA. And then they don't have to do all the extra work of figuring out exactly what can be published Mar 07 21:40:58 StarFive has agreed to sell the SoC into distribution as that is part of the BeagleBoard.org philosophy so we helping guide them as to how to operate under the situation where the public is buying the SoC and therefore public documentation is needed. Mar 07 21:42:14 I think the issue is the flow is often IP vendor sells IP to SoC vendor under NDA, SoC vendor sells chip to board level customer under NDA. No one in that chain is motivated to release public documentation Mar 07 21:42:27 reminds me of a talk that talked about Sun's process of making Solaris open source, which is quite a pain with random bits of licensed third-party code for which they either had to renegotiate the license or replace the code Mar 07 21:42:38 That does not work for BeagleBoard.org so we are trying to help them become more open. Mar 07 21:43:58 I think one advantage a giant semi like Texas Instruments has is that they own a lot more of the IP in the Sitara and they also have the experience and power to demand the IP vendors they do use let them publish register level documentation. Mar 07 21:44:35 StarFive is cautious, and I think rightly so, about what they have permission to release in their documentation. Mar 07 21:44:40 and even they evidently don't always suceed, e.g. the musb code isn't in the TRMs Mar 07 21:44:48 *core Mar 07 21:44:50 Interesting, I was wondering about that Mar 07 21:46:39 crypto accelerators also aren't documented, security features like firewalls are inconsistently documented, debug stuff is often poorly documented Mar 07 21:47:06 I think the next few weeks for StarFive will be lots of emails back and forth with Synopsys and Cadence to see exactly what they can publish Mar 07 21:47:59 I guess we'll see how that turns out Mar 07 21:48:32 I have been encouraging StarFive to model their documentation and code releases after SiFive (which they were spun out of). SiFive has done a great job with tricky things like zero stage and first stage bootloaders https://www.sifive.com/blog/an-open-source-release-of-the-freedom-u540-c000s-bootloader Mar 07 21:48:53 the SiFive SoC manuals are also quite good https://static.dev.sifive.com/FU540-C000-v1.0.pdf Mar 07 21:49:24 not TRM level, but they have enough details for it to be well understood and supported Mar 07 21:50:10 yeah TI-quality documentation is probably a ton of effort Mar 07 21:53:03 the latest TRM of the J721E (DRA829/TDA4VM/AM752x) is 18155 pages Mar 07 21:57:15 I cannot wait! See... Mar 07 21:57:29 'wait and see,' is all I do. I getting good at it! Mar 07 22:07:46 I am only mediocre at is as my last sentence is me not being cute but actual poor english. Mar 07 22:07:48 Blah. Mar 07 22:08:05 * set_ says no more typing today! Mar 07 22:08:46 Still, I cannot wait! Mar 08 01:14:10 Mo' typing is not what i typed earlier. What is w/ all this chatter/clatter? Mar 08 01:14:28 clickity/clackity. Mo'. Mar 08 01:14:34 Done. Sorry. Mar 08 01:15:01 Are there any new docs. on the BeagleV yet? Mar 08 01:15:20 outside of...the books about RISC-V? Mar 08 01:21:50 I had an odd rendition of typing but I got the 'meat' of the situation. Sorry. Mar 08 01:34:23 Also...what is the deal w/ apt? Why does it only allow for updating/upgrading on one terminal? Is that normal or universal? Mar 08 01:44:39 you can't run two apt commands at the same time, if that's what you mean Mar 08 01:45:41 since apt is acting on the entire system, it would be unsafe to have two apt operations going on in parallel Mar 08 01:45:55 so it prevents that from happening by accident Mar 08 02:07:16 Right. Mar 08 02:07:25 Okay. Mar 08 02:07:28 Just wondering here. Mar 08 02:15:47 I am currently installing boto3-core. Ask me why...blah. Mar 08 02:18:31 Oh...no questions. Here are answers! MachineKit.io needs a newer new. So, I am going to pitch in and learn while performing. Mar 08 02:18:32 ... Mar 08 02:19:01 boto3-core is a needed lib. from Google Codelabs. Mar 08 02:19:36 So, that is the reasoning. Flawed? Maybe. Will it work? Maybe. Do I care about my sensory overload. NO! **** ENDING LOGGING AT Mon Mar 08 02:59:57 2021