| 3 Apr 2024 |
 fgaz | In reply to @alex:tunstall.xyz
Sure. Would it belong best on a separate page, given that it's quite long? Since [0x4A6F] added a GHC redlink I assume that's the intention | 07:53:03 |
 Alex | In reply to @fgaz:matrix.org
Since [0x4A6F] added a GHC redlink I assume that's the intention Ah it wasn't showing up as as red link because I wasn't signed in apparently. | 08:05:07 |
 [0x4A6F] | And it wasn't saved because of that calc-captcha... | 16:16:40 |
| [0x4A6F] | * And it wasn't saved because of that calc-captcha... -> https://wiki.nixos.org/wiki/RISC-V/GHC | 16:17:09 |
| [0x4A6F] | Thanks Alex for that writeup and your work on cross bootstraping GHC. ❤️ | 16:17:31 |
| Alex | In reply to @0x4a6f:matrix.org
And it wasn't saved because of that calc-captcha... -> https://wiki.nixos.org/wiki/RISC-V/GHC Pain. I was going to copy it over to the wiki anyway. I guess I might touch it up a bit to make it more flake-friendly. | 16:43:31 |
| Alex | In reply to @0x4a6f:matrix.org
Thanks Alex for that writeup and your work on cross bootstraping GHC. ❤️ I still need to figure out what's causing the registerised segfaults, so it's not done yet. | 16:44:39 |
| Alex | The rustc build is segfaulting on native builds too 🙃
Seems to be https://github.com/rust-lang/rust/issues/117022
Now beginning to wonder if all the GHC segfaults aren't also kernel-related... | 18:02:35 |
| Alex |
- rustc segfaults due to the 6.5 kernel(?).
- uboot-tools is needed to build the 6.8 kernel.
- python3-cryptography is needed to build uboot-tools.
- rustc is needed to build python3-cryptography.
I'm glad I've never GCed on this system, because I might need to run an older kernel to upgrade the kernel... | 19:50:19 |
 hive ⬡ | https://github.com/NixOS/nixpkgs/issues/301340 | 20:01:35 |
| hive ⬡ | do you think all of the segfault errors on my builds are related to this kernel as well ? | 20:02:02 |
| Alex | In reply to @hive:the-apothecary.club
do you think all of the segfault errors on my builds are related to this kernel as well ? I doubt it, because QEMU uses the host system's kernel. | 20:02:25 |
| hive ⬡ | ah ok | 20:02:35 |
| Alex | I guess you could try using a different version if you're on 6.5 though?
It just might work. I don't really understand the root cause of my segfaults well enough to say. | 20:03:23 |
| hive ⬡ | i'm on 6.8 | 20:03:44 |
| 4 Apr 2024 |
|  Guilhem joined the room. | 09:22:20 |
 hexa | anyway, debian is building on hifive unmatched 😄 | 09:23:54 |
| hexa | my employer seems to host a bunch of these 👀 | 09:24:45 |
| fgaz | hexa: as far as I know our main limitation is rack space (ping Mic92). the upcoming milk-v oasis seems like a more efficient use of that space | 09:31:27 |
| hexa | huh? rack space where? | 09:31:43 |
| hexa | TUM? | 09:32:01 |
 Pratham Patel (you can mention me) | In reply to @fgaz:matrix.org
hexa: as far as I know our main limitation is rack space (ping Mic92). the upcoming milk-v oasis seems like a more efficient use of that space The third RISC-V SoM from Sipeed is gonna have the same SG2380 SoC that the Oasis will too. Not sure on the max memory that they will ship. So for CPU-bound compiles, this SoM should be better if $ and space are an issue.
https://twitter.com/sipeedio/status/1774644666375524659
| 09:34:19 |
| fgaz | In reply to @hexa:lossy.network
huh? rack space where? In the nix community rack. Last time we discussed this I think that was the candidate for hosting a builder. I don't know more than that | 09:35:32 |
 Julien | And the pioneer is a no go ? | 09:36:50 |
| Pratham Patel (you can mention me) | In reply to @julienmalka:matrix.org
And the pioneer is a no go ? Yes. | 09:37:02 |
| Pratham Patel (you can mention me) | All machines with the C910 core are only good for test builds and cannot be "trusted" (not in context of a backdoor but it is not compliant to the spec). | 09:38:00 |
| Pratham Patel (you can mention me) | The "v2" of C920 is supposed to be more compliant with the spec but no one has had hands-on with it yet. | 09:38:48 |
|  @shalokshalom:kde.org joined the room. | 10:05:05 |
| @shalokshalom:kde.org | Hi there. I heard about the attempts to bootstrap the GHC on NixOS RISC-V. I guess you tried the LLVM backend? Sorry if thats a bit naive, I guess there is a very good reason, why this wouldnt work. The Github ticket around this issue also mentions, that this is possible. Is the backend not capable to compile itself? | 10:31:29 |
| @shalokshalom:kde.org | * Hi there. I heard about the attempts to bootstrap the GHC on NixOS RISC-V. I guess you tried the LLVM backend? Sorry if thats a bit naive, I guess there is a very good reason, why this wouldnt work. The Github ticket around this issue also mentions, that this is possible. Is the LLVM backend not capable to compile GHC? | 10:32:42 |
