Posts by moody@hj.9fs.net
(DIR) Post #B6m6wxjbWSqBwMbqAC by moody@hj.9fs.net
0 likes, 0 repeats
And this is about all it can do at the moment, for some reason it seems to be unaligning its own stack and then expecting to SSE to and from it. I tried throwing -mstackrealign just to see if it would figure itself out but that doesn't seem to do anything. I specifically make sure the stack is 16byte aligned in the newlib entrypoint, and I can verify that it starts aligned at function start so I'm thinking this is just a bug. Is this masked on linux? Does it transparently fix sse alignment fuckups? I fought with this more or less all day so forgive me taking the easy way out and asking the void before I take a nice long break from this.
(DIR) Post #B6mO6MOetg8IFusZX6 by moody@hj.9fs.net
1 likes, 1 repeats
flan% x86_64-9front-plan9-gcc --sysroot /mnt/term/home/moody/src/newlib/sysroot -c blah.ccc1: out of memory allocating 18446744073709551615 bytes after a total of 274432 bytes
(DIR) Post #B6mOa1wFWahhUdjrhw by moody@hj.9fs.net
1 likes, 0 repeats
@kirby@freerobuxextremist.com In this economy?
(DIR) Post #B928vG7k3ty5vfPnge by moody@hj.9fs.net
0 likes, 0 repeats
flan% wc -l mlkem.c1961 mlkem.cflan% mk6c -FTVw main.c6c -FTVw notrandombytes.c6c -FTVw mlkem.c6c -FTVw verify.c6c -FTVw randombytes.car vu mlkem.amd64.a mlkem.6 verify.6 randombytes.6ar: creating mlkem.amd64.aa - mlkem.6a - verify.6a - randombytes.66l -o 6.out main.6 notrandombytes.6 mlkem.amd64.aflan% 6.outShared secret: 776c74df301f8d82525e8ebbb40095cd2e92df6dc933e7866259f531c7350ad5[MLKEM-512] OKShared secret: e91377840e6b6694eaa9f01c97ff68874e8b0c520b00c2cde37c4fc239626e70[MLKEM-768] OKShared secret: 5d9e235fccb2b3499a5f490a56e3f0d3fd9b58bda28b690f91b57b88a5a80b90[MLKEM-1024] OKflan%
(DIR) Post #B95E2PKlIfm5lWc436 by moody@hj.9fs.net
1 likes, 0 repeats
Of course I forked mlkem-native right before they added more test vector coverage and now it's catching bugs I should have caught during my trimming....
(DIR) Post #B96aQg8o73QYwujXwO by moody@hj.9fs.net
0 likes, 0 repeats
Fixed.... phew.
(DIR) Post #B98gikGMVxpmzEBXjk by moody@hj.9fs.net
0 likes, 0 repeats
The ported and trimmed ml-kem code now runs and passes all of the wycheproof test vectors:https://shithub.us/moody/mlkem-wheat/8cd56dd073eff14a4be0b3bdd1d1ef718a1031de/wych.c
(DIR) Post #B9EilNfkKO1oeHCc76 by moody@hj.9fs.net
1 likes, 0 repeats
After trying to get a hold of anyone in the wireguard project to take a look at my implementation for 9front over the past 2-3 months (and one email of intent with a rough sekeleton PoC over a year ago) I have heard not a thing. To be clear there's no expectations here, but after the FreeBSD debacle I get the impression they want to ensure some bar of quality for implementations with the wireguard name so I figured I'd ask. Ah well.
(DIR) Post #B9N1ctJwgtU5LtnZzc by moody@hj.9fs.net
0 likes, 0 repeats
I am staring down a line of 9front libsec yaks.
(DIR) Post #B9TADXOp7bdCf1pDnM by moody@hj.9fs.net
0 likes, 0 repeats
lol
(DIR) Post #B9yMPrGzmOB54YYNYO by moody@hj.9fs.net
1 likes, 0 repeats
9front has hit 12k commits: https://build.9front.org/
(DIR) Post #BA0Huo1qGMKXQKEpQO by moody@hj.9fs.net
0 likes, 0 repeats
Still wondering what 'major contributions' to 9front have been made by sourcehut staff: https://sourcehut.org/foss-contributions/Would love to be proven wrong, but from my look at the git log I just don't see it.
(DIR) Post #BA0ItJz4aBUJuFnb16 by moody@hj.9fs.net
0 likes, 0 repeats
@lanodan@queer.hacktivis.me Is that a contribution to the project though? To add a half baked builder, refuse patches to make it better and then delete it?
(DIR) Post #BA0KnTgbnThQATVILA by moody@hj.9fs.net
0 likes, 0 repeats
Typically 9front's style is that for those (Thanks) lines the patch was written almost entirely or majority by the individual being thanked. I will leave the reader to evaluate if this patch is worthy of the inclusion of the 9front project in that list.CC: @lanodan@queer.hacktivis.me
(DIR) Post #BA0gC89d5Tc1fHft0y by moody@hj.9fs.net
0 likes, 0 repeats
I'm sorry, I am fundamentally a miserable person and sometimes I fail to keep that isolated.
(DIR) Post #BA2YBIBvsCuXBhlFNQ by moody@hj.9fs.net
0 likes, 0 repeats
Composition of stuff on the web is a bit harder on my eyes due to the mix of the red underline for spell check and the more hot pink of the enter boxes. Not a complaint just an observation.
(DIR) Post #BA2YUHGZXMyVovz4W8 by moody@hj.9fs.net
1 likes, 0 repeats
I don't get why linux heads get a bit upset when you point out that linux sucks, like do you guys not realize that most of the people using their computer for doing modern things don't quite enjoy it? You are allowed to use software in spite of it, just ask anyone using windows 11. I'm sorry that the cool software hippie thing you tied your personality to is now run by people kissing the boot but not acknowledging that is a pretty bad look.
(DIR) Post #BA2jlYMeJrcPGfGbfk by moody@hj.9fs.net
1 likes, 0 repeats
@chansecodina@sunny.garden I don't think you have to say anything to be clear, no one does in my book. All I ask is for folks to be aware of it. I still see a lot of folks talking about linux like its still how it was 15-20 years ago and now that Linux is more accessible than ever (in part to such companies) I think newcomers are getting the wrong idea. I guess what I am advocating for is for communities that were/are built on the ideals that brought them to Linux to continue on to something better. Maybe that means un-fucking userspace, maybe that means a new kernel, maybe that means forking. I don't know what the answer is. I just work on what I think the solution is.I'm just tired of the user community continuing to use the hippie marketing and continuing to think the GPL actually worked. I don't think its wrong to say "switching to linux from windows is better" but I think the buck stops there.I know its sad, and I am being a bit flippant, it just feels like the writing had been on the wall for a while.
(DIR) Post #BA2pvdCRnO2ysxgS80 by moody@hj.9fs.net
0 likes, 0 repeats
And since we're laying our cards on the table, I'm personally tired of people working on niche projects acting like their project is going to be some kind of replacement for Linux while still getting to be an exclusive club for the coolest kids on the block.Fair play and point taken. I agree, and I recognize how crazy it is that I can grab nearly any computer, boot linux on it and mostly expect a working system now. I haven't been around super long but I did get to experience the before times of looking up supported chips and hardware.I think an ideal world, a group that went through the dark times of reverse engineering wifi chipsets and doing things the hard ways, would stand on their own ground and refuse to merge support for chipsets that are openly hostile to others doing the same once they had an upper hand. I understand their reasoning, upstream support results in a better experience for end users, an out of tree module or a forked kernel for a specific chip sucks. Its one thing to say "well if someone brings in a patch for this oddware buggy crap then maybe" and then another if someone calls your bluff and actually does that (corporate or not).Maybe in that ideal world this pushback and refusal would nudge users in the right direction, I think we did see a bit of this. People buy AMD cards for linux, the dd-wrt open router list nudged folks, same with wifi chips. The problem is just that finding out after you have the hardware in hand sucks, but if nothing changes I can only see the same steps being recreated.
(DIR) Post #BA3EK11FHkOKbrbWvg by moody@hj.9fs.net
0 likes, 0 repeats
@dalias@hachyderm.io Oh hey something I actually know about this time! I haven't worked on UCA but I did do normalization for 9front, both NFC and NFD[0]. I wrote a paper and did a talk about my experiences working on this stuff for 9front to try and get other folks there to be more interested in Unicode. As of right now the NFD code I wrote passes all the test vectors (and I've been testing against the v18 beta right now, still no issues). I don't fully recall the license conditions for musl but all the code I did for 9front is MIT if you wanted to compare notes :).[0] https://git.9front.org/plan9front/9front/5b902d52edaa25e5f430f024c1a1b9b56f94e3e0/sys/src/libc/ucd/runenorm.cI can link the paper if you think that would help too, would just have to dig it up.