SchOS/kernel/sys/sys_proc.c

#include "core/types.h"
#include "core/defs.h"
#include "core/param.h"
#include "arch/memlayout.h"
#include "core/hwinfo.h"
#include "arch/mmu.h"
#include "core/proc.h"
#include "arch/x86.h"
#include "core/syscall.h"
#include "uapi/posix.h"
#include "core/uaccess.h"
#include "uapi/errno.h"
#include "dev/acpi.h"
#include "fs/fat.h"
#include "core/spinlock.h"
#include "core/sleeplock.h"
#include "fs/fs.h"
#include "fs/buf.h"

addr_t sys_fork(void)      { return fork(); }
addr_t sys_wait(void)      { return wait(); }

// _exit(status)/exit_group(status): the low 8 bits of the argument are what
// wait() reports back through WEXITSTATUS.
addr_t
sys_exit(void)
{
    int code = 0;
    argint(0, &code);
    // Linux's exit is one thread's: pthread_exit, and a thread returning
    // from its start function. The program ends with exit_group -- or here,
    // when this is the last thread, or the native ABI, which has no threads.
    struct proc *p = myproc();
    if (p->syscall_map && p->tgid != p->pid && proc_thread_count(p->tgid) > 1)
        thread_exit();
    exit(W_EXITCODE(code, 0));
    return 0;
}

addr_t
sys_exit_group(void)
{
    int code = 0;
    argint(0, &code);
    exit(W_EXITCODE(code, 0));
    return 0;
}


addr_t
sys_kill(void)
{
    int pid, sig;
    argint(0, &pid);
    argint(1, &sig);

    // POSIX target selection:
    //   pid  > 0  -> that process
    //   pid == 0  -> every process in the caller's own group
    //   pid == -1 -> every process we may signal (approximated as our group)
    //   pid  < -1 -> the process group with id -pid
    if (pid == 0 || pid == -1)
        return sigsend_pgrp(myproc()->pgid, sig) < 0 ? -ESRCH : 0;
    if (pid < -1)
        return sigsend_pgrp(-pid, sig) < 0 ? -ESRCH : 0;

    // sig == 0: existence check only
    if (sig == 0)
        return proc_exists(pid) ? 0 : -ESRCH;

    // Raise the signal for real: pending bit + wakeup, handler delivery on
    // the target's next syscall boundary, default action if unhandled.
    return sigsend(pid, sig) < 0 ? -ESRCH : 0;
}

// Defined with getrusage() further down; wait4() needs it too.
static void rusage_fill(struct rusage *ru, uint64 ut, uint64 st);

// A process is its thread group: getpid is the group's id, gettid the thread's.
addr_t sys_getpid(void)  { return myproc()->tgid; }
addr_t sys_getppid(void) { return myproc()->parent ? myproc()->parent->tgid : 1; }
addr_t sys_gettid(void)  { return myproc()->pid; }

addr_t
sys_waitpid(void)
{
    int pid, options;
    addr_t statusp;
    // status is optional (NULL is standard, valid usage — waitpid(pid, NULL,
    // WNOHANG) — so this must NOT go through argptr(), which rejects a NULL/
    // unmapped pointer outright).
    if (argint(0, &pid) < 0 || argaddr(1, &statusp) < 0 || argint(2, &options) < 0)
        return -1;
    int status;
    int wpid = waitpid(pid, &status, options, 0, 0);
    if (wpid > 0 && statusp)
        put_user(status, statusp);
    return wpid;
}

// wait4(2) -- and its fourth argument is a `struct rusage *`.
//
// It was read off the stack and thrown away. So a caller that passed a buffer
// got it back untouched, and read whatever had been there before: zsh's
// `time` reported 16777216.01s user for every command, which is 2^24 seconds,
// which is uninitialised memory that happened to look like a number. The
// figure did not vary with the workload because it was never written.
//
// wait3() is the same syscall with pid -1, and it is the one shells actually
// call, so this is the path every `time` goes through.
addr_t
sys_wait4(void)
{
    int pid, options;
    addr_t statusp, rup;
    uint64 ut = 0, st = 0;

    argint(0, &pid);
    argaddr(1, &statusp);
    argint(2, &options);
    argaddr(3, &rup);

    int status;
    int wpid = waitpid(pid, &status, options, &ut, &st);
    if (wpid <= 0)
        return wpid;
    if (statusp)
        put_user(status, statusp);
    if (rup) {
        struct rusage *ru = (struct rusage*)user_ptr(rup, sizeof(*ru));
        if (ru)
            rusage_fill(ru, ut, st);
    }
    return wpid;
}

addr_t
sys_clone(void)
{
    uint64 flags, stack, ptid, ctid, tls;
    argaddr(0, &flags);
    argaddr(1, &stack);
    argaddr(2, &ptid);
    argaddr(3, &ctid);
    argaddr(4, &tls);
    return clone_proc(flags, stack, ptid, ctid, tls);
}

addr_t
sys_sleep(void)
{
    int n;
    uint ticks0;
    if (argint(0, &n) < 0)
        return -1;
    acquire(&tickslock);
    ticks0 = ticks;
    while (ticks - ticks0 < (uint)n) {
        if (myproc()->killed) { release(&tickslock); return -1; }
        sleep(&ticks, &tickslock);
    }
    release(&tickslock);
    return 0;
}

addr_t
sys_uptime(void)
{
    uint xticks;
    acquire(&tickslock);
    xticks = ticks;
    release(&tickslock);
    return xticks;
}

addr_t sys_reboot(void)
{
    cprintf("System is rebooting...\n");
    // A stick left mounted should come back readable, and marked clean.
    schfs_ksync_all();
    fat_ksync_all(1);
    // The FADT's reset register first: it is the one the board's designer
    // wired up. Then the PCI reset port and the keyboard controller, which
    // are what nearly every PC has had since there were PCs.
    acpi_reboot();
    outb(0xCF9, 0x06);
    outb(0x64, 0xFE);
    return 0;
}

addr_t sys_shutdown(void)
{
    cprintf("System is shutting down...\n");
    // S5, through the firmware's own PM1 registers and its _PTS. The port
    // writes after it are QEMU's, Bochs' and VirtualBox's fixed addresses,
    // kept for a machine whose AML did not load; a real board answers none
    // of them.
    schfs_ksync_all();
    fat_ksync_all(1);
    acpi_poweroff();
    outw(0x604, 0x2000);
    outw(0xB004, 0x2000);
    outw(0x4004, 0x3400);
    return 0;
}

// --- Sessions and process groups ---------------------------------------
// Job control is defined entirely in terms of these. They used to be stubs
// returning the caller's own pid, which made every process look like its own
// group leader and left a shell's wait loop unable to make progress.

addr_t sys_setsid(void) { return proc_setsid(); }

addr_t
sys_getsid(void)
{
    int pid;
    argint(0, &pid);
    if (pid == 0 || pid == myproc()->pid)
        return myproc()->sid;
    return proc_exists(pid) ? (addr_t)myproc()->sid : (addr_t)-ESRCH;
}

addr_t
sys_getpgid(void)
{
    int pid;
    argint(0, &pid);
    if (pid == 0 || pid == myproc()->pid)
        return myproc()->pgid;
    int g = proc_pgid(pid);
    return g < 0 ? (addr_t)-ESRCH : (addr_t)g;
}

addr_t
sys_setpgid(void)
{
    int pid, pgid;
    argint(0, &pid);
    argint(1, &pgid);
    return proc_setpgid(pid, pgid);
}

addr_t sys_getpgrp(void) { return myproc()->pgid; }

// --- User and group identity -------------------------------------------
// Tracked for real now. Nothing *enforces* permissions with these yet -- the
// on-disk format has no owner field until SchFS -- but they are no longer
// invented, so getuid() after setuid() reports what actually happened.

addr_t sys_getuid(void)  { return myproc()->uid;  }
addr_t sys_geteuid(void) { return myproc()->euid; }
addr_t sys_getgid(void)  { return myproc()->gid;  }
addr_t sys_getegid(void) { return myproc()->egid; }

// Privileged (euid 0) callers may set any id; an unprivileged caller may only
// switch between its real and saved ids. Standard POSIX rule.
addr_t
sys_setuid(void)
{
    int uid;
    argint(0, &uid);
    if (uid < 0)
        return -EINVAL;
    if (myproc()->euid == 0) {
        myproc()->uid = myproc()->euid = myproc()->suid = (uint)uid;
        return 0;
    }
    if ((uint)uid != myproc()->uid && (uint)uid != myproc()->suid)
        return -EPERM;
    myproc()->euid = (uint)uid;
    return 0;
}

addr_t
sys_setgid(void)
{
    int gid;
    argint(0, &gid);
    if (gid < 0)
        return -EINVAL;
    if (myproc()->euid == 0) {
        myproc()->gid = myproc()->egid = myproc()->sgid = (uint)gid;
        return 0;
    }
    if ((uint)gid != myproc()->gid && (uint)gid != myproc()->sgid)
        return -EPERM;
    myproc()->egid = (uint)gid;
    return 0;
}

addr_t sys_umask(void)
{
    int mask;
    argint(0, &mask);
    uint old = myproc()->umask;
    myproc()->umask = (uint)mask & 0777;
    return old;
}

addr_t
sys_setreuid(void)
{
    int ruid, euid;
    argint(0, &ruid); argint(1, &euid);
    if (myproc()->euid != 0)
        return -EPERM;
    if (ruid >= 0) myproc()->uid  = (uint)ruid;
    if (euid >= 0) myproc()->euid = (uint)euid;
    myproc()->suid = myproc()->euid;
    return 0;
}

addr_t
sys_setregid(void)
{
    int rgid, egid;
    argint(0, &rgid); argint(1, &egid);
    if (myproc()->euid != 0)
        return -EPERM;
    if (rgid >= 0) myproc()->gid  = (uint)rgid;
    if (egid >= 0) myproc()->egid = (uint)egid;
    myproc()->sgid = myproc()->egid;
    return 0;
}

addr_t
sys_setresuid(void)
{
    int r, e, sv;
    argint(0, &r); argint(1, &e); argint(2, &sv);
    if (myproc()->euid != 0)
        return -EPERM;
    if (r  >= 0) myproc()->uid  = (uint)r;
    if (e  >= 0) myproc()->euid = (uint)e;
    if (sv >= 0) myproc()->suid = (uint)sv;
    return 0;
}

addr_t
sys_setresgid(void)
{
    int r, e, sv;
    argint(0, &r); argint(1, &e); argint(2, &sv);
    if (myproc()->euid != 0)
        return -EPERM;
    if (r  >= 0) myproc()->gid  = (uint)r;
    if (e  >= 0) myproc()->egid = (uint)e;
    if (sv >= 0) myproc()->sgid = (uint)sv;
    return 0;
}

// Linux-specific; the fs id follows the effective id here.
addr_t sys_setfsuid(void) { return myproc()->euid; }
addr_t sys_setfsgid(void) { return myproc()->egid; }

addr_t
sys_getresuid(void)
{
    addr_t rp, ep, sp;
    argaddr(0, &rp); argaddr(1, &ep); argaddr(2, &sp);
    if (rp) put_user(myproc()->uid,  rp);
    if (ep) put_user(myproc()->euid, ep);
    if (sp) put_user(myproc()->suid, sp);
    return 0;
}

addr_t
sys_getresgid(void)
{
    addr_t rp, ep, sp;
    argaddr(0, &rp); argaddr(1, &ep); argaddr(2, &sp);
    if (rp) put_user(myproc()->gid,  rp);
    if (ep) put_user(myproc()->egid, ep);
    if (sp) put_user(myproc()->sgid, sp);
    return 0;
}

addr_t sys_getgroups(void)
{
    int size;
    argint(0, &size);
    return 0; // no supplementary groups
}
addr_t sys_setgroups(void) { return 0; }

// --- Process/scheduling extensions ---

// A real yield. It returned 0 without giving anything up, which cost nothing
// while every timer tick was a yield anyway; with the fair scheduler a tick
// only preempts once the slice is spent, so a spin-and-yield loop (musl's
// locks, a polling driver in userland) would hold the CPU for all of it.
addr_t sys_sched_yield(void) { yield(); return 0; }

// Every processor that started, since nothing here pins a process anywhere.
//
// This wrote "CPU 0" and returned 0 -- and the return value is the number of
// bytes of mask the kernel wrote, which musl uses to clear the rest. Zero
// meant all of it, so the mask came back empty and nproc, make -j and every
// thread pool sized from sysconf(_SC_NPROCESSORS_ONLN) saw one processor.
addr_t sys_sched_getaffinity(void)
{
    int pid, cpusetsize;
    addr_t maskp;
    uchar m[(NCPU + 63) / 64 * 8];
    argint(0, &pid); argint(1, &cpusetsize); argaddr(2, &maskp);
    if (cpusetsize < (int)sizeof m || (cpusetsize & 7))
        return -EINVAL;
    memset(m, 0, sizeof m);
    for (int i = 0; i < ncpu && i < NCPU; i++)
        if (cpus[i].started)
            m[i / 8] |= 1 << (i % 8);
    if (copy_to_user(maskp, m, sizeof m) < 0)
        return -EFAULT;
    return sizeof m;
}
addr_t sys_sched_setaffinity(void) { return 0; }

addr_t sys_prctl(void)
{
    int op;
    addr_t arg2;
    argint(0, &op);
    argaddr(1, &arg2);
    if (op == PR_SET_NAME) {
        char *name;
        if (argstr(1, &name) < 0) return -1;
        safestrcpy(myproc()->name, name, sizeof(myproc()->name));
        return 0;
    }
    if (op == PR_GET_NAME) {
        char *buf = (char*)user_ptr(arg2, 16);
        if (!buf) return -EFAULT;
        safestrcpy(buf, myproc()->name, 16);
        return 0;
    }
    return 0; // all other prctl ops: silently succeed
}

addr_t sys_tgkill(void)
{
    int tgid, tid, sig;
    argint(0, &tgid); argint(1, &tid); argint(2, &sig);
    // To one thread: musl's pthread_kill, and raise(), which is tgkill on
    // the calling thread.
    if (tid <= 0) return -EINVAL;
    if (sig == 0) return proc_exists(tid) ? 0 : -ESRCH;
    return sigsend(tid, sig) < 0 ? -ESRCH : 0;
}

addr_t sys_tkill(void)
{
    int tid, sig;
    argint(0, &tid); argint(1, &sig);
    if (tid <= 0) return 0;
    if (sig == 0) return proc_exists(tid) ? 0 : -ESRCH;
    return sigsend(tid, sig) < 0 ? -ESRCH : 0;
}

addr_t sys_pause(void)
{
    // Sleep until next timer tick and return -EINTR (no real signal delivery).
    acquire(&tickslock);
    uint t0 = ticks;
    while (ticks == t0 && !myproc()->killed)
        sleep(&ticks, &tickslock);
    release(&tickslock);
    return -EINTR;
}

// getrusage(2).
//
// This used to ignore `who` entirely and report the machine's *uptime* as the
// caller's user time. Every process of every age answered with the same
// number, and RUSAGE_CHILDREN answered with it too -- so a shell timing a
// command got the age of the machine, and `time` on a pipeline was
// meaningless rather than merely imprecise.
//
// Ticks, converted at the boundary. The kernel counts whole timer ticks
// because that is what the sampling gives it; microseconds are what the
// interface asks for, so the multiply happens here rather than pretending to a
// resolution the counter does not have.
static void
rusage_fill(struct rusage *ru, uint64 ut, uint64 st)
{
    memset(ru, 0, sizeof(*ru));
    ru->ru_utime.tv_sec  = (long)(ut / HZ);
    ru->ru_utime.tv_usec = (long)((ut % HZ) * (1000000 / HZ));
    ru->ru_stime.tv_sec  = (long)(st / HZ);
    ru->ru_stime.tv_usec = (long)((st % HZ) * (1000000 / HZ));
}

addr_t sys_getrusage(void)
{
    int who;
    struct rusage *ru;
    argint(0, &who);
    if (argptr(1, (char**)&ru, sizeof(*ru)) < 0) return -EFAULT;

    switch (who) {
    case RUSAGE_SELF:
        rusage_fill(ru, myproc()->utime, myproc()->stime);
        return 0;
    case RUSAGE_CHILDREN:
        rusage_fill(ru, myproc()->cutime, myproc()->cstime);
        return 0;
    default:
        // RUSAGE_THREAD is -1 and would be RUSAGE_SELF here, since a process
        // has exactly one thread. It is refused rather than aliased: answering
        // it would imply threads exist.
        return -EINVAL;
    }
}

addr_t sys_times(void)
{
    addr_t bufaddr;
    argaddr(0, &bufaddr);
    // The tms fields are in clock ticks, which is what the kernel counts, so
    // this one needs no conversion -- USER_HZ and HZ are both 100. They used
    // to be memset to zero, which is why `times()` agreed with nothing.
    if (bufaddr) {
        struct tms *t = (struct tms*)user_ptr(bufaddr, sizeof(*t));
        if (t) {
            t->tms_utime  = (long)myproc()->utime;
            t->tms_stime  = (long)myproc()->stime;
            t->tms_cutime = (long)myproc()->cutime;
            t->tms_cstime = (long)myproc()->cstime;
        }
    }
    // USER_HZ units, measured from the TSC.
    //
    // musl answers sysconf(_SC_CLK_TCK) with 100 unconditionally, so a shell's
    // `time` builtin divides whatever this returns by 100. Deriving it from
    // the tick count made `time` under-report whenever ticks were being lost
    // -- which is how a 20-second command came to print 3.12s.
    uint64 tsec, tusec;
    monotonic(&tsec, &tusec);
    return (addr_t)(tsec * USER_HZ + (tusec * USER_HZ) / 1000000);
}

addr_t sys_sysinfo(void)
{
    addr_t bufaddr;
    argaddr(0, &bufaddr);
    struct sysinfo *si = (struct sysinfo*)user_ptr(bufaddr, sizeof(*si));
    if (!si) return -EFAULT;
    memset(si, 0, sizeof(*si));
    // TSC, like everything else that reports elapsed time.
    uint64 upsec, upusec;
    monotonic(&upsec, &upusec);
    si->uptime    = (long)upsec;
    // Real values: the RAM the page allocator was given, and how much of it is
    // still on the free list. These used to be the constants 256MB/128MB, and
    // then PHYSTOP, an address, which counted the PCI hole as memory.
    si->totalram  = ram_total_bytes();
    si->freeram   = kalloc_free_bytes();
    // The buffer cache, so free(1) does not count it as used by nothing.
    uint nbuf, maxbuf;
    bcache_size(&nbuf, &maxbuf);
    si->bufferram = (uint64)nbuf * BSIZE;
    si->mem_unit  = 1;
    si->procs     = (uint16)proc_num();
    return 0;
}

addr_t sys_waitid(void)
{
    // Simplified: wait for any child, ignore idtype/id/options.
    int wpid = wait();
    if (wpid < 0) return -ECHILD;
    // siginfo_t pointer (arg 2) — zero-fill if provided.
    addr_t infop;
    argaddr(2, &infop);
    if (infop) {
        char *p = (char*)user_ptr(infop, 128); // sizeof(siginfo_t)
        if (p) memset(p, 0, 128);
    }
    return 0;
}