Announcement

Collapse
No announcement yet.

Statsfs: A Proposed Linux File-System For Kernel Statistics

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Larry$ilverstein
    replied
    Originally posted by k1e0x View Post
    Having this stuff on the file system is just dumb.

    unit=value

    makes a lot more sense than

    echo "value" > filesystem/path/unit.
    Why?

    Originally posted by k1e0x View Post
    I get the argument is done but.. mistakes of the past need not carry forward.
    sysctl was removed because /proc/sys was objectively better in every way except number of syscalls required and even that will be improved if the Greg K.H. readfile() syscall gets merged. The performance of reading/writing kernel tunables is irrelevant 99.9% of the time anyway.

    You can hand wave and parade your JavaScript dev sensibilities as much as you want, but at the end of the day, every single kernel engineer involved took the opposite view to you.
    Last edited by Larry$ilverstein; 28 May 2020, 05:22 PM.

    Leave a comment:


  • k1e0x
    replied
    Originally posted by Larry$ilverstein View Post

    Have you ever used anything in /proc/sys/, like e.g. /proc/sys/vm/drop_caches or /proc/sys/vm/laptop_mode? If so, you've already been using the replacement.

    The syscall was error prone and pointless.
    Yeaaah... I think I like sysctl. Having this stuff on the file system is just dumb.

    unit=value

    makes a lot more sense than

    echo "value" > filesystem/path/unit.

    I get the argument is done but.. mistakes of the past need not carry forward.
    Last edited by k1e0x; 28 May 2020, 02:27 PM.

    Leave a comment:


  • Larry$ilverstein
    replied
    Originally posted by aht0 View Post
    What was wrong with sysctl? It was deprecated in Linux in exchange of what?
    Have you ever used anything in /proc/sys/, like e.g. /proc/sys/vm/drop_caches or /proc/sys/vm/laptop_mode? If so, you've already been using the replacement.

    The syscall was error prone and pointless.

    Leave a comment:


  • Larry$ilverstein
    replied
    Originally posted by zxy_thf View Post
    The current "mount" output is horrible for human to read.
    If you want to see the mount points of your block devices, use lsblk. Should every single tool give human friendly output, just because some people decide to bang their head against a brick wall instead of using the correct tool for the job?

    Leave a comment:


  • k1e0x
    replied
    Originally posted by Britoid View Post

    No one cares about those.
    Well this is kind of an aside point my main issue is, it just feels like a bad way to do it but as far as compat goes... macOS is Unix, and people do seem to care about that.. being as it has more commercial software than Linux you might say they care more. Making yourself more different than Unix means you may have fewer programs as people will have a harder time porting.

    True they have different display engines but for instance Qt can display on both.. so it could be things like htop, but perhaps someone would like to make a system statistic Qt program. If it roughly collects the information the same way on the backed the UI is more trivial and your software can be used by a larger user base.

    ("Sends other UNIX boxes to /dev/null" Official Apple Ad ~ 2003)
    Last edited by k1e0x; 28 May 2020, 02:10 PM.

    Leave a comment:


  • niner
    replied
    Originally posted by zxy_thf View Post
    The current "mount" output is horrible for human to read.
    I counted the output on my system and there're already 11 different file systems out there.
    findmnt is your friend.

    Leave a comment:


  • aht0
    replied
    Originally posted by Britoid View Post

    No one cares about those.
    I am glad that nobody cares about your opinion either

    Leave a comment:


  • aht0
    replied
    What was wrong with sysctl? It was deprecated in Linux in exchange of what?

    It's pretty convenient to get both stats from kernel and for altering runtime parameters .

    Leave a comment:


  • Larry$ilverstein
    replied
    Originally posted by k1e0x View Post
    It also may break compatibility with other Unix'es.
    So this somehow "breaks" compatibility, but the other 200 or so Linux-specific features don't? How is that, exactly...?

    People who write portable software for "Unixes" target the POSIX API as a baseline and then detect the availability of everything else, with portable fallbacks.

    Leave a comment:


  • Britoid
    replied
    Originally posted by k1e0x View Post

    Agree. Linux has too many pseudo filesystems. It also may break compatibility with other Unix'es.
    No one cares about those.

    Leave a comment:

Working...
X