Joel Fernandes  

@joel_linux

my thoughts on Linux kernel, Android, electronics, math, CS, maker community, space technology, politics, bitcoin. (all opinions my own). I work at Google.

Vrijeme pridruživanja: svibanj 2009.

Tweetovi

Blokirali ste korisnika/cu @joel_linux

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @joel_linux

  1. Prikvačeni tweet
    22. ruj 2019.

    Sneak peak at a slide from my RCU talk at . This "Life cycle of a grace period" block diagram serves as a reference for me too now ;-)

    Poništi
  2. proslijedio/la je Tweet
    29. sij

    This is the best talks on locks I've ever seen. I can't believe how much covers in 41mins: memory/compiler memory models, atomics, futexes. Wow 🔥 (via 's Best of 2019 tech talks list)

    Poništi
  3. proslijedio/la je Tweet
    22. sij

    Yet another C quiz! Can you tell the output just by looking at the code? int main(void) { fork(); putchar('x'); fork(); return 0; } gcc prog.c && ./a.out Assume standard up to date Linux glibc implementation.

    Prikaži ovu nit
    Poništi
  4. proslijedio/la je Tweet
    19. sij

    This is 🤯 using historical data to predict container performance and using cgroups to control CPU placement. CFS makes an attempt at all of these things (except using PMCs to guide placement) but it’s always going to be imperfect because...

    Prikaži ovu nit
    Poništi
  5. proslijedio/la je Tweet
    14. sij

    "A Thread Synchronization Model for the PREEMPT_RT Linux Kernel " Dear Mr Bristot de Oliveira, Thank you for submitting your manuscript to Journal of Systems Architecture. I am pleased to inform you that your manuscript has been accepted for publication. /me opens a beer.

    Poništi
  6. proslijedio/la je Tweet
    17. sij
    Poništi
  7. proslijedio/la je Tweet
    Poništi
  8. proslijedio/la je Tweet

    nobody: Tom Steyer at the end of the debate:

    Prikaži ovu nit
    Poništi
  9. 6. sij

    The timing problem with the "spinlocks and scheduler" post that Linus is beating up could also be solved by restartable sequences (rseq). But hey I agree with him that spinlocks in userspace seem like a bad idea, vs mutex.

    Poništi
  10. 14. pro 2019.

    ...Memory, performance and tracing projects with focus on kernel/system work. CrOS is personally advocated by Linus Torvalds as the future of Desktop Linux. Here we go, USB booting my first CrOS device over the weekend on a Samsung Chromebook. ;-).

    Prikaži ovu nit
    Poništi
  11. 14. pro 2019.

    Starting a new job at Google working on ChromeOS (CrOS) devices. Frankly I'm excited about portage and about working on x86 machines more now (versus when I was at Android). I will be working on Android as well, due to Android apps running on CrOS. Likely going to be working on..

    Prikaži ovu nit
    Poništi
  12. proslijedio/la je Tweet
    29. stu 2019.
    Poništi
  13. proslijedio/la je Tweet
    14. stu 2019.

    You may already be running Linux v5.3! I'm excited about heap auto-initialization, pidfd_open(), global -Wimplicit-fallthrough, x86 CR4 & CR0 pinning, more kfree() sanity checks, arm64 default KASLR, and hardware security embargo documentation:

    Poništi
  14. 12. stu 2019.

    I am really enjoying my Linux kernel mentorship experience at both and LKMP and I'm getting good people. They are new but they are good and my evaluation tasks are confirming that. Looking forward to working with them!

    Poništi
  15. proslijedio/la je Tweet
    5. stu 2019.
    Prikaži ovu nit
    Poništi
  16. proslijedio/la je Tweet
    31. lis 2019.

    Slides of my talk on Core Scheduling for at 2019. They were online Tuesday night already, but I've kept updating them until ~30 mins before the talk (yeah... I know!), so you may want to re-check/re-download them

    Poništi
  17. proslijedio/la je Tweet
    25. lis 2019.

    [$] Redesigned workqueues for io_uring

    Poništi
  18. proslijedio/la je Tweet
    23. lis 2019.

    Up for crashing any kernels? Repo with 3000 syzkaller reproducers for kernel bugs: I like the ARTS name (Autogenerated Regression TestS). Announce email:

    Poništi
  19. 18. lis 2019.

    Some notes on modeling non-determinism using : "Modeling (lack of) store ordering using PlusCal – and a wishlist"

    Poništi
  20. proslijedio/la je Tweet
    18. lis 2019.

    Here’s a great write up about the Rome memory system. It’s always amazing to see these articles put the pieces together.

    Prikaži ovu nit
    Poništi
  21. proslijedio/la je Tweet
    15. lis 2019.
    Poništi

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·