[9fans] System call mania - Plan9

This is a discussion on [9fans] System call mania - Plan9 ; Hello. I was recently looking in libc when I noticed a few things about system calls: 1) brk and sbrk are implemented atop brk_, which is not documented 2) the seek function seems to be a system call, but the ...

+ Reply to Thread
Results 1 to 7 of 7

Thread: [9fans] System call mania

  1. [9fans] System call mania

    Hello. I was recently looking in libc when I noticed a few things
    about system calls:

    1) brk and sbrk are implemented atop brk_, which is not documented
    2) the seek function seems to be a system call, but the alpha folder
    defines a function seek which calls _seek
    3) Some items in that file have underscores, like BRK_ and _READ

    This leads me to the following question: what exactly are the system
    calls in Plan 9?

    Thanks.



  2. Re: [9fans] System call mania

    On 3/22/08, Pietro Gagliardi wrote:
    > Hello. I was recently looking in libc when I noticed a few things
    > about system calls:
    >
    > 1) brk and sbrk are implemented atop brk_, which is not documented
    > 2) the seek function seems to be a system call, but the alpha folder
    > defines a function seek which calls _seek
    > 3) Some items in that file have underscores, like BRK_ and _READ
    >
    > This leads me to the following question: what exactly are the system
    > calls in Plan 9?
    >


    grep -i syscall /sys/src/9/port

    iru


  3. Re: [9fans] System call mania

    That confirmed one suspicion: the REAL calls are those who are simply
    named in /sys/src/libc/9syscall/sys.h, and that the USER-LEVEL calls
    are a big mess in libc. Thanks!

    On Mar 22, 2008, at 12:16 AM, Iruata Souza wrote:

    > On 3/22/08, Pietro Gagliardi wrote:
    >> Hello. I was recently looking in libc when I noticed a few things
    >> about system calls:
    >>
    >> 1) brk and sbrk are implemented atop brk_, which is not documented
    >> 2) the seek function seems to be a system call, but the alpha folder
    >> defines a function seek which calls _seek
    >> 3) Some items in that file have underscores, like BRK_ and _READ
    >>
    >> This leads me to the following question: what exactly are the system
    >> calls in Plan 9?
    >>

    >
    > grep -i syscall /sys/src/9/port
    >
    > iru
    >




  4. Re: [9fans] System call mania

    On 3/22/08, Pietro Gagliardi wrote:
    > That confirmed one suspicion: the REAL calls are those who are simply
    > named in /sys/src/libc/9syscall/sys.h, and that the USER-LEVEL calls
    > are a big mess in libc. Thanks!
    >


    doesn't the name system call suggests you something?

    iru


  5. Re: [9fans] System call mania

    It suggests to me that these calls are the lowest level of
    communication with the kernel. I once thought that all system calls
    could be called by a program :-P

    On Mar 22, 2008, at 12:41 AM, Iruata Souza wrote:

    > On 3/22/08, Pietro Gagliardi wrote:
    >> That confirmed one suspicion: the REAL calls are those who are simply
    >> named in /sys/src/libc/9syscall/sys.h, and that the USER-LEVEL calls
    >> are a big mess in libc. Thanks!
    >>

    >
    > doesn't the name system call suggests you something?
    >
    > iru
    >




  6. Re: [9fans] System call mania

    On Fri, Mar 21, 2008 at 9:44 PM, Pietro Gagliardi wrote:
    > It suggests to me that these calls are the lowest level of
    > communication with the kernel. I once thought that all system calls
    > could be called by a program :-P
    >


    why do you think that they can't be?

    ron


  7. Re: [9fans] System call mania

    I was tired and I used the wrong words.

    On Mar 22, 2008, at 1:32 AM, ron minnich wrote:

    > On Fri, Mar 21, 2008 at 9:44 PM, Pietro Gagliardi
    > wrote:
    >> It suggests to me that these calls are the lowest level of
    >> communication with the kernel. I once thought that all system calls
    >> could be called by a program :-P
    >>

    >
    > why do you think that they can't be?
    >
    > ron
    >




+ Reply to Thread