Posts by jbqueru@floss.social
(DIR) Post #Auk9tXx07eF0LMEaq8 by jbqueru@floss.social
2025-06-03T05:23:42Z
0 likes, 0 repeats
@foone @cinebox @root The firmware in the 6301 in the Atari ST's keyboard even has commands to remotely peek, poke, and execute code. I'm not sure it's powerful enough to run Doom, though.
(DIR) Post #AuuSDTi9KsiVk4XakK by jbqueru@floss.social
2025-06-08T04:39:57Z
0 likes, 0 repeats
@foone I hear you. I can write Z80, 6502 or 68000 assembly easily, but don't ask me to write x86_64, aarch64 or riscv64.
(DIR) Post #Av3bcBW4AEfaNf6lgO by jbqueru@floss.social
2025-06-12T14:37:26Z
0 likes, 0 repeats
@foone When I worked in the Android team at Google, we used to count the number of Base-64 decoders, those would sprout everywhere...
(DIR) Post #Av3cDbaYBslBxP36lk by jbqueru@floss.social
2025-06-12T14:44:26Z
0 likes, 0 repeats
@foone Fun factoid: on 68000, the MOVEM instruction (M->R) does one extra read beyond the end of the actual data. That means that it's not safe to use at the very end of a block of physical memory. Depending on implementations, that might result in having 2 versions of memcpy.E.g. on Atari ST, you can't simply copy the ROM to RAM by doing a pair of MOVEM in a loop, the last read crashes with a bus error when it reads beyond the end of the ROM.
(DIR) Post #AvY0n0LqyjyhuPCwbo by jbqueru@floss.social
2025-06-27T05:41:01Z
0 likes, 0 repeats
From what I'm able to gather, getting reasonable support from major Linux distros for a RISC-V home machine boils down to getting one of the JH7110 boards.That's about the speed of a Raspberry Pi 3, or about 20% of the speed of my 2011 MacBook Air. The advantage of those RISC-V boards is that they can be had with 8GB of RAM.Still, that's not high performance by any stretch of the imagination. Speaking of imagination (pun!), it's not clear whether the GPU in JH7110 is supported.
(DIR) Post #AvY0n1WWcY6nXnCysi by jbqueru@floss.social
2025-06-27T05:49:36Z
0 likes, 0 repeats
Now, I understand that RISC-V is claimed to be Free (i.e. unencumbered by licensing), but I'm not sure what that does for me as a user.In practice, for financial reasons or otherwise, there's no practical way for me for evolve my CPU or to pay someone to do that for me.Even if that Freedom allows to have multiple vendors I'm the ecosystem, those vendors are so imcompatible with one another that migration requires a full reinstall, which I can also do onto any non-RISC-V system.
(DIR) Post #AvY0n9VkvIFMJiuTnE by jbqueru@floss.social
2025-06-27T05:58:16Z
0 likes, 0 repeats
Look, I like to have alternatives, and I'm willing to suffer through suboptimal systems, but even I have limits.In the current state, I see that RISC-V systems are slower, harder to install, have worse drivers, worse OS support, worse app supports.Now, I get the performance part, RISC-V can't use any secret sauce. I also get the app part, that's someone else's problem.But the OS-level issues are inexcusable after 10 years, RISC-V leadership should get their act togethere there.
(DIR) Post #AvY1r0rZ7QjubjW39M by jbqueru@floss.social
2025-06-27T06:50:09Z
0 likes, 0 repeats
@brettm yup, that vision is my hope as well (i.e. the desired end state), I hope that someone makes it their mission to solve one or more of the problems currently getting in the way.Performance is so bad, though, that it needs to evolve not by percentage or even multiples, but by an order of magnitude or more.
(DIR) Post #AvwxhEfirq9bRJPeIy by jbqueru@floss.social
2025-07-09T07:32:49Z
0 likes, 0 repeats
@bortzmeyer @bagder Now, of course, if you run cat or more or curl as suid root (with full capabilities), things happen, but that's quite literally going to be true of anything that processes files, and it's not the fault of those tools (unless they install themselves like that by default, of course).
(DIR) Post #AvzrXOPvlNpIW8pTZg by jbqueru@floss.social
2025-07-10T17:09:42Z
0 likes, 0 repeats
@futurebird @dalias @CatDragon It's always fun to follow the same reasoning with cars: "my great-grandparents didn't have cars and they did just fine, so we don't need cars." (and the trams and trains and corner shops that they relied on don't exist any more).
(DIR) Post #AwCmR47Mckxx2jtSAC by jbqueru@floss.social
2025-06-20T14:27:36Z
0 likes, 1 repeats
It's common to say that the "free" in #FreeSoftware is closer to "free as in speech" than to "free as in beer," and that is true.However, "free as in speech" only resonates so far with me. To me, Free Software is even closer to free will than to free speech, in the sense that, with Free Software, I have control over my actions in how they relate to that software, i.e. how I have control to decide how to modify it, how to use it, whether to maintain it, whether to deploy it or distribute it.
(DIR) Post #AwCmRCBCeMn42xkdG4 by jbqueru@floss.social
2025-06-20T14:34:11Z
0 likes, 0 repeats
A situation that appears more and more, though, is that "with great power comes great responsibility." When using Free Software, you might only have yourself to rely on to maintain the software you use.Don't get me wrong, most maintainers will go through great lengths to help their users, but that's from being kind and decent human beings, not from being contractually required to.When using Free Software, it's also "free as in puppy": no upfront cost, but maintainance is on you.
(DIR) Post #AwCmRKDclFTqymwg8u by jbqueru@floss.social
2025-06-20T14:37:42Z
0 likes, 0 repeats
In my little world of #RetroComputing, I have stumbled across minor issues in some of the tools that I use.I've had two cases so far where maintainers very quickly took my suggestion and made it available in their main branch. I've also had one where I haven't heard back from the maintainer, i.e. where I'm going to need to maintain my own fork (for #RiscV compatibility). And that's OK. Beyond being glad that I got help, I'm glad that the original authors allow me also to do the work myself.
(DIR) Post #AwcN8RHP6yPJzGIGZs by jbqueru@floss.social
2025-07-29T07:02:20Z
0 likes, 0 repeats
@foone Amusingly, the same would also happen in C.
(DIR) Post #Ax270etfyJzzOPW1FA by jbqueru@floss.social
2025-08-10T17:04:27Z
0 likes, 0 repeats
@debian I don't always install new OSes withing 24 hours of release, but, when I do, it's Debian.Almost flawless here, and it wasn't hard to work around the one small glitch I faced.
(DIR) Post #Ax2kVbsp7i17EdrmOe by jbqueru@floss.social
2025-08-10T18:49:34Z
0 likes, 0 repeats
Me, looking at the 8008/8080/Z80 instruction sets: "bah, why do you need conditional calls and conditional returns when the conditional jumps are already more than enough?"Me, coding in 68000: "I really wish I had conditional calls and conditional returns, relying on conditional jumps makes the code harder to read than necessary."I am never happy.#RetroComputing #Assembly
(DIR) Post #Ax362lPcoV5XNmrKd6 by jbqueru@floss.social
2025-08-11T04:28:31Z
0 likes, 0 repeats
@tomjennings That's definitely an interesting approach, though one that might be hurting branch prediction on modern processors.8008 didn't have it, 8080 has PCHL, named JP (HL) in Z80... and RISC-V has JALR, as if it needed to be made more confusing (and on RISC-V that's the only way to do jumps longer than 1MB if I'm reading correctly).
(DIR) Post #Ax44C4asrQBojHOee0 by jbqueru@floss.social
2025-08-11T04:32:03Z
0 likes, 0 repeats
@tomjennings And 8008 is more microcontroller than microprocessor, its stack doesn't store arbitrary data, so you can't just push the jump address and RET to it.
(DIR) Post #Ax4IajbtPh1kFOWRWq by jbqueru@floss.social
2025-08-11T18:22:43Z
0 likes, 0 repeats
@foone Is that limitation still in exFAT?
(DIR) Post #AxGS4IPCM5Gu2rH4Ua by jbqueru@floss.social
2025-08-17T15:06:31Z
0 likes, 1 repeats
@futurebird I've worked in companies where servers weren't uniform, some were meant to compute things, others to handle large data sets in RAM, others to store persistent data. The nodes that were mean to compute were called compute nodes, and collectively, compute resources.