| packetup | If anyone is willing to give advice please DM. | 00:13 |
|---|---|---|
| packetup | Also what are all your thoughts on https://github.com/google/skywater-pdk | 00:23 |
| lkcl | hi packetup, welcome | 10:37 |
| lkcl | well, i was studying RISC-V for about 2 years so know quite a lot about it, libre-soc is using OpenPOWER ISA i am still happy to answer | 10:39 |
| lkcl | you may be interested to know that NLnet is sponsoring bunnie huang on his "betrusted" project | 10:41 |
| lkcl | https://blog.dshr.org/2020/01/bunnie-huangs-betrusted-project.html | 10:41 |
| lkcl | https://www.bunniestudios.com/blog/?p=5706 | 10:43 |
| cesar[m]1 | Trezor is a USB bitcoin hadware wallet that also sports two-factor authentication, and also stores PGP and SSH keys. See: https://blog.trezor.io/secure-two-factor-authentication-with-trezor-u2f-e940fd5a60af | 10:56 |
| cesar[m]1 | Hardware design available at: https://github.com/trezor/trezor-hardware | 10:56 |
| lkcl | ah, that's just the PCB, cesar[m]1 | 11:03 |
| lkcl | https://github.com/trezor/trezor-hardware/tree/master/electronics/trezor_model_t | 11:03 |
| lkcl | what packetup is saying (i believe) is, how can you trust a proprietary *processor*? | 11:04 |
| lkcl | i had some people contact me in 2018 about this, the fact that FIPS-approved algorithms are inherently untrustable *and* the implementations are entirely opaque | 11:05 |
| lkcl | yet people are entrusting literally TRILLIONS in cryptocurrencies to unverifiable hardware! | 11:05 |
| cesar[m]1 | Well, being based on an ARM Cortex-M class microcontroller, the STM32F, it should at least be free of proprietary bootloaders and drivers. | 11:09 |
| cesar[m]1 | The GCC-based toolchain and software development kit also seem to be libre. | 11:10 |
| cesar[m]1 | Lacking a MMU, it won't run Linux, tought. | 11:13 |
| cesar[m]1 | I agree that verification at hardware level is desirable. | 11:16 |
| lkcl | yeyyeh, i really like the STM32F series. all of them are supported extremely well by libopencm3 which is also excellent | 11:42 |
| lkcl | ST's own libraries for the STM32 series are a bit... naff :) they're orientated around windows (yukk). libopencm3 is a bit more "sane" | 11:43 |
| cesar[m]1 | By the way, at work, there is a STM32F microcontroller in the x-ray experiment, on the stratospheric balloon. | 12:05 |
| cesar[m]1 | To generate code to configure peripherals and pins, I use the ST32CubeMX software, which do seems proprietary (but does run under Linux). | 12:08 |
| * lkcl waves to jhol | 13:57 | |
| jhol | https://www.youtube.com/channel/UCeF7JKNXOy0jpMOxpgbZcpg | 13:58 |
| lkcl | jhol: thx | 14:13 |
| packetup | lkcl: thank you for those links! Our project lead was actually brainstorming about using some sort of randomization to protect the hardware implementation. | 16:56 |
| packetup | With storage in SRAM with a removable battery. | 16:58 |
| packetup | We're looking into a SoC utilizing https://github.com/cliffordwolf/picorv32 but any any libre processor would work. | 17:01 |
| lkcl | oh yeh picorv32 is pretty neat, and small, it's a good choice, as long as you're not looking for high-performance | 18:06 |
| lkcl | randomisation is *really* tricky to do well. make damn f****g sure you find a decent cryptographer who can advise you there ok? | 18:07 |
| packetup | Yes, they're a long time contributor to GnuPG | 18:08 |
| lkcl | ah superb. | 18:09 |
| pangelo[m] | hello, apropos betrusted, this just popped up in my mailbox: https://www.crowdsupply.com/sutajio-kosagi/precursor | 19:32 |
| lkcl | nice! | 22:11 |
| lkcl | packetup: the choice of TNRG there is worth investigating. bunnie tends to know his stuff | 22:12 |
| packetup | Yes, that's one design requirement of course. | 22:19 |
| packetup | I will reach out to him. | 22:39 |
| packetup | Any idea how I could reach him? | 22:48 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!