Libreboot thinkpad x200 vs Raptor computing IBM power9

Hard to tell @joaanaa. The more people tapping in the reserves of x230/t430/w530/t440p, the higher the prices and the lower the reserves… The pandemic made it hard to have proper sources as well. Refurbishing became cheaply made on once stable sources as well… everything got complicated on the supply chain level. At least this is my local Canadian story.

As for the secure part of the question, I already replied indirectly and more specifically on XSA-404 that hit us recently. That will depend on workarounds that can be applied, and what kind of vulnerabilities that will be found that can be mitigated without microcode updates. I would say again 10 years, but do not quote me on that. We all do not know on that. What we all know is that we need something better, though. And i’m not talking about another x86 here. We needed to build something else years ago. But that is hard.

Last time I checked, creating a Power based laptop (with a custom CPU based on OpenPower ISA) would have costed 1M USD.

That again, goes with the endeavor of already, or in parallel, having full Xen support on Power… to have Qubes support on it. That work has started, funded by community members (thanks again @Rudd-O @rspigler) and can be followed here: Port Qubes to ppc64 [3 bitcoin bounty] · Issue #4318 · QubesOS/qubes-issues · GitHub. I push for this for years, and now slow steps are happening. This is rejoicing. But that needs to continue.

And yet again, open source firmware doesn’t mean auditable (Reading hostboot, skiboot, skiroot and other parts of OpenPower made firmware shows this and documentation produced by 3mdeb while porting Power9 to coreboot is a real gem on that level. The port of coreboot on Power9 (Talos II) is still ongoing (Thanks to 3mdeb for the work and my past sweat (and blood?) for funding it), with a lot of unexpected problems along this crazy journey. The latest being the realization that Talos II doesn’t have a fully functional TPM connector… Who would have known. So no fully functional Heads as of today, but that will change soon enough.

Progress happening under Dasharo. Testers needed here for current Talos II owners: POWER9 (Talos II and Talos II Lite) coreboot beta testers needed · Issue #1018 · linuxboot/heads · GitHub and progress can be followed from the documentation section here that points to their code repositories: Building manual - Dasharo Universe . Also conferences here: Conference materials - Dasharo Universe. And yes, you do not even need to flash your Talos II to test it: https://docs.dasharo.com/variants/talos_2/installation-manual/#testing-firmware-images-without-flashing. Isn’t it totally amazing? No risk of bricking whtsoever. That is the power of open firmware and open hardware: not being locked down to anything. But that means community needs to make alternatives, right?

As said above in previous posts, if we do not do a politic move as a community to make that happen, I’m really not sure how we will get away of X86, ARM licenses for implementation (Core accelerated AES instruction set costing 1M in royalties, what? Even if RPI5 will allegedly have 16GB of ram (criticism for X230 being that its not enough on Qubes…), and that a Xen port exists for RPI4… That is not a good target either, no IOMMU as of now) and Risc-V is simply not a good candidate to create powerful machines. Again, OpenPower is the best candidate we have now being an Open ISA, open documentation (again, Thnaks to 3mdeb for their amazing low level documentations produced in the coreboot port: GitHub - 3mdeb/openpower-coreboot-docs: Documentation related to POWER9 coreboot porting effort).

Anybody being able to assemble a team with proper expertise and money, organizing things properly could create the CPU (bypassing IBM) and create real Open Hardware and defeat proprietary market in the long run…

It’s kind of a time to stop waiting for it to happen magically and invest in the world we all want to live in. I’m personally waiting for others to jump also in that ship and collaborate the same way I did in the past years, without receiving enough support to make it happen together now. But maybe the time was not right but it’s becoming now?

The question here again, is how to organize to make this happen quickly, sustainably and for good. Otherwise we should all remember OpenMoko/Neo73, One laptop per Child and PhoneBloks and other really nice projects/ideas that were too good for their times, were bought/corrupted/put back on the shelves and did not thrive.

We do not want to do the same errors this time, don’t we? Who can pilot this? I don’t know. But I would gladly be part of that team, though.

1 Like