This general is for operating systems development, if you have a question, want to share a resource or want to show us something you made, feel free to do so here.IRC: irc.libera.chat #osdevers>The OS wikihttps://osdev.wiki/wiki/Expanded_Main_Page>Operating systems bookshttps://pages.cs.wisc.edu/~remzi/OShttps://github.com/dreamportdev/Osdev-NotesModern Operating Systems by Andrew TanenbaumOperating System Concepts by Abraham Silberschatz and James Peterson (The dinosaur book)>Computer architecture booksComputer Architecture: A Quantitative Approach by John L. Hennessy, and David A. PattersonStructured Computer Organization by Andrew S. Tanenbaum, and Todd AustinAdvanced Computer architecture by Kai Hwang>AssemblyProgramming from the Ground Up>x86 and amd64 instructionshttps://www.felixcloutier.com/x86/>RISC-V assemblyhttps://www.robertwinkler.com/proje>CThe C programming language by K&RC programming a modern approach by K.N. Kinghttps://beej.us/guide/bgc/>Open source operating systems’ source codesLinuxhttps://git.kernel.org/FreeBSDhttps://cgit.freebsd.org/ReactOShttps://git.reactos.org/Plan9(discontinued)https://9p.io/sources/plan9/9front (plan9 fork actively maintained)https://git.9front.org/plan9front/9front/HEAD/info.htmlHaikuhttps://github.com/haiku/haikuMinixhttps://git.minix3.org/index.cgi?p=xv6https://github.com/mit-pdos/xv6-riscvFreeRTOShttps://github.com/FreeRTOS/FreeRTOSZephyrhttps://github.com/zephyrproject-rtos/zephyrPrevious >>109851974
>>109867906Is there an open source toy/educational real time OS for raspi that I could study?
>>109867906how tf do you modify inodes safely when you support smp?? i hate locking around everything
>>109868987You split the locks into multiple failsafe protections: - an inode spinlock- a r/w semaphoreOr, if you're talking osdev here, you make your own rules, one variant is to use a hash table with per-bucket locks.Or, if you're a God, use CoW inodes and you never modify an inode in place, you create a new version, then atomically swap a pointer.But yes, SMP is the bane of my existence, and having finally figured it out makes it even more delicious.
Bump
I don't know much about this project, but there is this CastorOS form waterloo. I'd bundle it together with xv6 in terms of scope; it has simple implementation and loosely follow some BSD conventions.https://cs350.rcs.uwaterloo.ca/opengrok/xref/
>>109871054This is another academic OS project, HelenOS, which is a microkernel-based multiserver operating system which is used in a Czech University.https://www.helenos.orgPretty cool project.
>>109867906I am actively developing a research OS, written in a string processing language. It is bootable on x86, ARM and RISC-V (the 64 bit variant of each). I can put code directly into the terminal to JIT compile it. I can also make files with an editor and compile those too. The purpose of the OS is to explore string rewriting and GOFAI. It has no networking and it never will.I can post about it more in this thread, but I am not sure if this is a maid thread or not, and I don't want to post maids here if it is disruptive.
>>109871879Beautiful! I don't think anyone here will find any osdev discussions disruptive, even if they include maids. Hit us, I might entertain you later with my microkernel capabilities-enabled design (which is currently at the got-shell, got-network stage.
>>109867906>FreeRTOS link fixedGood job fool, finally a decent thread
>>109871879>we got maidos before maidbbs
>>109871966MAIDS OS is a loose port of parts of TempleOS from HolyC to MAIDS. I like how Terry Davis designed an operating system around a compiler, and that everything is editable and can be recompiled from within the OS itself. I also think Doldoc is really good, so I ported that too. I might not port anything else, since those two things, plus having a string processor as the system language are the main things I want from the system. Once I finish squashing bugs, I will write more programs inside it and make a Doldoc explaining the language and how to use the OS and release it to Public Domain.>>109871981I don't like operating money, and getting hosting would require this. Making an OS doesn't require operating money, because I already have the computer.
I spend longer than I'm willing to admit on making sure my implementation is cohesive and close to what the unix chads did back in the day. I never got to go deep into something like graphics acceleration or networking or SMP because I was trying too hard to do unix (code style and conventions, hierarchy of directories, iinterfaces, even variable namings and mkre from sysv6 and later 386bsd.) Should I kill myself?
>>109872414Not really, I'm not one of those "OMG UNIX U KEEP UREZELF" faggots, but I only try to borrow good parts of UNIX and stray away from the unneeded complexity. Keep in mind that you're one person trying to do the work of ... everybody who worked on UNIX since 1980s? You can't cover everything, and SMP is probably the hardest concept to get right, while graphics accelerations is heavily dependent on chips, revisions, proprietary interfaces, etc.But d00d, if you don't do a DHCP DISCOVER work ... you need to start wearing that maid apron the maid anon is trying to make us wear.
>>109872318I have spent all day making the MAIDS compiler faster, so I can test quadragrams on it.
bumpy road
>>109871977Thanks anon, the zephyr link was also broken but no one pointed it out in the last thread hehe.Writing on Libreoffice can be annoying sometimes
>>109875774I see that anon is doing Windows 98(i think?) UI on a monolithic kernel, I'm doing Windows 2k UI on a capabilities-enabled microkernel, this will be interesting.
Get on my level :D>>109876215If you manage to do that on a hobby OS, and on microkernel, hat off. Where are you currently, at what stage?
