IPC restriction handling for user processes - Minix

This is a discussion on IPC restriction handling for user processes - Minix ; Hi again, I found the following lines in kernel/proc.c, to regulate the ipc from userspace into kernel tasks. However, those lines are disabled through the "#if 0" preprocessor statement. #if 0 if ((iskerneln(src_dst_p) && _function != SENDREC && _function != ...

+ Reply to Thread
Results 1 to 2 of 2

Thread: IPC restriction handling for user processes

  1. IPC restriction handling for user processes

    Hi again,


    I found the following lines in kernel/proc.c, to regulate the ipc from
    userspace into kernel tasks. However, those lines are disabled through
    the "#if 0" preprocessor statement.

    #if 0
    if ((iskerneln(src_dst_p) && _function != SENDREC
    && _function != RECEIVE)) {
    #if DEBUG_ENABLE_IPC_WARNINGS
    kprintf("sys_call: trap %d not allowed, caller %d, src_dst %d
    \n",
    function, proc_nr(caller_ptr), src_dst);
    #endif
    return(ETRAPDENIED); /* trap denied by mask or kernel */
    }
    #endif

    I am a little irritated about these lines due to two reasons.

    1st. why haven't those disabled lines been cropped from the source
    code, when
    2nd., Jorrit N. Herder rewrote ipc and system call handling in 2005 as
    file comments tell, especially as
    3rd., these lines seem to be superseded by the introduction of the
    "sys_map_t s_ipc_from", "sys_map_t s_ipc_to" and "sys_map_t
    s_ipc_sendrec" variables in proc.h

    So, can anyone support/refute my assumption, these lines are legacy
    and outdated?

    Thanks for your ideas,

    Alex

    ---------------------------------------------------------------------
    As I'm currently about to port Minix to a research processor platform,
    some more questions like this may arise in future...

  2. Re: IPC restriction handling for user processes

    Just to bring my question up again...

+ Reply to Thread