Red Hat had their usual shindig in May, the Red Hat Summit. At the Summit, they showed off RHEL10 running on RISC-V. Cool!
Later, they made a developer preview available. You can look at https://developers.redhat.com/products/rhel-riscv and https://access.redhat.com/articles/7117031 to get started.
Now, I have the Milk-V Jupiter board. It tends to run Bianbu from the SpacemiT team, and the SpacemiT team actually have reasonably good documentation for how to flash the board with a later version, so on and so forth. I will try and record here my experience in attempting to get this Dev Preview installed on the Milk-V Jupiter.
Right off the bat, Red Hat ships a .img that is XZ compressed. This does not work with TitanTools, so immediately you are presented with the problem of uncompressing the image and recompressing it in a useful format so that TitanTools can read it. -1 to Red Hat before I have even started. I know Red Hat has a habit of being completely tone-deaf, and this is no exception. Once the .img was zipped up and I could begin looking at it with TitanTools. Sidenote: the source archive is just all the SRPMs for what is in the .img file. Why bother - that is not useful. You could have saved a ton of space and just listed it all in a simple textfile if the packages (and the source RPMS) are available from the Customer Portal. Anyway, TitanTools read the .zip and extracts the image into the temporary flashing folder and.. there is a .img and nothing else. Where's the bootfs.ext4, rootfs.ext4, env.bin, factory folder, .itb and .json files? RHEL is not going to use the same environment setup as Bianbu, surely? There is no details on what it expects by way of partition layout. -5 for that Red Hat. You assume the "es_fs" command is available in u-boot, and there is no indication you have encountered a system that does not. You don't ship the .itb with the u-boot that you use, or the env.bin that contains your environment to allow anyone getting up and running.
So, trying to analyse the .img and see if anything is actually available within the image that could be used.
Disk /dev/loop0: 12,01 GiB, 12893290496 bytes, 25182208 sectors Units: sectors of 1 * 512 = 512 bytes Sector size (logical/physical): 512 bytes / 512 bytes I/O size (minimum/optimal): 512 bytes / 512 bytes Disklabel type: gpt Disk identifier: A9F0B08E-F15C-4A50-97BA-AD094302BB0D Device Start End Sectors Size Type /dev/loop0p1 2048 104447 102400 50M EFI System /dev/loop0p2 104448 25182174 25077727 12G Linux root (RISC-V-64)
An EFI partition and a root-fs. There's nothing in here that I can use. I am sure I could craft the right partition_universal.json file to flash this onto my board, but without at least the env.bin, I'd probably be looking at a paper-weight. The key thing here is that I would likely get further (and I have done) putting Debian Trixie on this system than trying the Red Hat Developer Preview. Until Red Hat can provide something, you know - useful - trying RHEL10 is dead in the water. I am sufficiently knowledgeable that I can handle even a new architecture like RISC-V with not too much trouble. But here, they've assumed "everyone will use EXACTLY the one system we used and nothing else". That is decidedly not how the world works.
Summary: Red Hat's grade - 1/10. Your loss, not mine. I can utilise this system to improve on Debian rather than RHEL.