vserver 1.9.5.x5
[linux-2.6.git] / kernel / compat.c
index 57cc48e..f14fbde 100644 (file)
@@ -20,6 +20,7 @@
 #include <linux/futex.h>       /* for FUTEX_WAIT */
 #include <linux/syscalls.h>
 #include <linux/unistd.h>
+#include <linux/security.h>
 
 #include <asm/uaccess.h>
 
@@ -160,10 +161,39 @@ asmlinkage long compat_sys_times(struct compat_tms __user *tbuf)
         */
        if (tbuf) {
                struct compat_tms tmp;
-               tmp.tms_utime = compat_jiffies_to_clock_t(current->utime);
-               tmp.tms_stime = compat_jiffies_to_clock_t(current->stime);
-               tmp.tms_cutime = compat_jiffies_to_clock_t(current->cutime);
-               tmp.tms_cstime = compat_jiffies_to_clock_t(current->cstime);
+               struct task_struct *tsk = current;
+               struct task_struct *t;
+               cputime_t utime, stime, cutime, cstime;
+
+               read_lock(&tasklist_lock);
+               utime = tsk->signal->utime;
+               stime = tsk->signal->stime;
+               t = tsk;
+               do {
+                       utime = cputime_add(utime, t->utime);
+                       stime = cputime_add(stime, t->stime);
+                       t = next_thread(t);
+               } while (t != tsk);
+
+               /*
+                * While we have tasklist_lock read-locked, no dying thread
+                * can be updating current->signal->[us]time.  Instead,
+                * we got their counts included in the live thread loop.
+                * However, another thread can come in right now and
+                * do a wait call that updates current->signal->c[us]time.
+                * To make sure we always see that pair updated atomically,
+                * we take the siglock around fetching them.
+                */
+               spin_lock_irq(&tsk->sighand->siglock);
+               cutime = tsk->signal->cutime;
+               cstime = tsk->signal->cstime;
+               spin_unlock_irq(&tsk->sighand->siglock);
+               read_unlock(&tasklist_lock);
+
+               tmp.tms_utime = compat_jiffies_to_clock_t(cputime_to_jiffies(utime));
+               tmp.tms_stime = compat_jiffies_to_clock_t(cputime_to_jiffies(stime));
+               tmp.tms_cutime = compat_jiffies_to_clock_t(cputime_to_jiffies(cutime));
+               tmp.tms_cstime = compat_jiffies_to_clock_t(cputime_to_jiffies(cstime));
                if (copy_to_user(tbuf, &tmp, sizeof(tmp)))
                        return -EFAULT;
        }
@@ -310,7 +340,7 @@ asmlinkage long compat_sys_getrlimit (unsigned int resource,
        return ret;
 }
 
-static long put_compat_rusage(struct compat_rusage __user *ru, struct rusage *r)
+int put_compat_rusage(const struct rusage *r, struct compat_rusage __user *ru)
 {
        if (!access_ok(VERIFY_WRITE, ru, sizeof(*ru)) ||
            __put_user(r->ru_utime.tv_sec, &ru->ru_utime.tv_sec) ||
@@ -348,7 +378,7 @@ asmlinkage long compat_sys_getrusage(int who, struct compat_rusage __user *ru)
        if (ret)
                return ret;
 
-       if (put_compat_rusage(ru, &r))
+       if (put_compat_rusage(&r, ru))
                return -EFAULT;
 
        return 0;
@@ -374,7 +404,7 @@ compat_sys_wait4(compat_pid_t pid, compat_uint_t __user *stat_addr, int options,
                set_fs (old_fs);
 
                if (ret > 0) {
-                       if (put_compat_rusage(ru, &r)) 
+                       if (put_compat_rusage(&r, ru))
                                return -EFAULT;
                        if (stat_addr && put_user(status, stat_addr))
                                return -EFAULT;
@@ -383,48 +413,58 @@ compat_sys_wait4(compat_pid_t pid, compat_uint_t __user *stat_addr, int options,
        }
 }
 
-asmlinkage long compat_sys_sched_setaffinity(compat_pid_t pid, 
+static int compat_get_user_cpu_mask(compat_ulong_t __user *user_mask_ptr,
+                                   unsigned len, cpumask_t *new_mask)
+{
+       unsigned long *k;
+
+       if (len < sizeof(cpumask_t))
+               memset(new_mask, 0, sizeof(cpumask_t));
+       else if (len > sizeof(cpumask_t))
+               len = sizeof(cpumask_t);
+
+       k = cpus_addr(*new_mask);
+       return compat_get_bitmap(k, user_mask_ptr, len * 8);
+}
+
+asmlinkage long compat_sys_sched_setaffinity(compat_pid_t pid,
                                             unsigned int len,
                                             compat_ulong_t __user *user_mask_ptr)
 {
-       unsigned long kern_mask;
-       mm_segment_t old_fs;
-       int ret;
-
-       if (get_user(kern_mask, user_mask_ptr))
-               return -EFAULT;
+       cpumask_t new_mask;
+       int retval;
 
-       old_fs = get_fs();
-       set_fs(KERNEL_DS);
-       ret = sys_sched_setaffinity(pid,
-                                   sizeof(kern_mask),
-                                   (unsigned long __user *) &kern_mask);
-       set_fs(old_fs);
+       retval = compat_get_user_cpu_mask(user_mask_ptr, len, &new_mask);
+       if (retval)
+               return retval;
 
-       return ret;
+       return sched_setaffinity(pid, new_mask);
 }
 
 asmlinkage long compat_sys_sched_getaffinity(compat_pid_t pid, unsigned int len,
                                             compat_ulong_t __user *user_mask_ptr)
 {
-       unsigned long kern_mask;
-       mm_segment_t old_fs;
        int ret;
+       cpumask_t mask;
+       unsigned long *k;
+       unsigned int min_length = sizeof(cpumask_t);
 
-       old_fs = get_fs();
-       set_fs(KERNEL_DS);
-       ret = sys_sched_getaffinity(pid,
-                                   sizeof(kern_mask),
-                                   (unsigned long __user *) &kern_mask);
-       set_fs(old_fs);
+       if (NR_CPUS <= BITS_PER_COMPAT_LONG)
+               min_length = sizeof(compat_ulong_t);
 
-       if (ret > 0) {
-               ret = sizeof(compat_ulong_t);
-               if (put_user(kern_mask, user_mask_ptr))
-                       return -EFAULT;
-       }
+       if (len < min_length)
+               return -EINVAL;
 
-       return ret;
+       ret = sched_getaffinity(pid, &mask);
+       if (ret < 0)
+               return ret;
+
+       k = cpus_addr(mask);
+       ret = compat_put_bitmap(user_mask_ptr, k, min_length * 8);
+       if (ret)
+               return ret;
+
+       return min_length;
 }
 
 static int get_compat_itimerspec(struct itimerspec *dst, 
@@ -445,7 +485,7 @@ static int put_compat_itimerspec(struct compat_itimerspec __user *dst,
        return 0;
 } 
 
-long compat_timer_settime(timer_t timer_id, int flags, 
+long compat_sys_timer_settime(timer_t timer_id, int flags,
                          struct compat_itimerspec __user *new, 
                          struct compat_itimerspec __user *old)
 { 
@@ -468,7 +508,7 @@ long compat_timer_settime(timer_t timer_id, int flags,
        return err;
 } 
 
-long compat_timer_gettime(timer_t timer_id,
+long compat_sys_timer_gettime(timer_t timer_id,
                struct compat_itimerspec __user *setting)
 { 
        long err;
@@ -485,7 +525,7 @@ long compat_timer_gettime(timer_t timer_id,
        return err;
 } 
 
-long compat_clock_settime(clockid_t which_clock,
+long compat_sys_clock_settime(clockid_t which_clock,
                struct compat_timespec __user *tp)
 {
        long err;
@@ -502,7 +542,7 @@ long compat_clock_settime(clockid_t which_clock,
        return err;
 } 
 
-long compat_clock_gettime(clockid_t which_clock,
+long compat_sys_clock_gettime(clockid_t which_clock,
                struct compat_timespec __user *tp)
 {
        long err;
@@ -519,7 +559,7 @@ long compat_clock_gettime(clockid_t which_clock,
        return err;
 } 
 
-long compat_clock_getres(clockid_t which_clock,
+long compat_sys_clock_getres(clockid_t which_clock,
                struct compat_timespec __user *tp)
 {
        long err;
@@ -531,12 +571,12 @@ long compat_clock_getres(clockid_t which_clock,
        err = sys_clock_getres(which_clock,
                               (struct timespec __user *) &ts);
        set_fs(oldfs);
-       if (!err && put_compat_timespec(&ts, tp))
+       if (!err && tp && put_compat_timespec(&ts, tp))
                return -EFAULT; 
        return err;
 } 
 
-long compat_clock_nanosleep(clockid_t which_clock, int flags,
+long compat_sys_clock_nanosleep(clockid_t which_clock, int flags,
                            struct compat_timespec __user *rqtp,
                            struct compat_timespec __user *rmtp)
 {
@@ -561,3 +601,208 @@ long compat_clock_nanosleep(clockid_t which_clock, int flags,
 
 /* timer_create is architecture specific because it needs sigevent conversion */
 
+long compat_get_bitmap(unsigned long *mask, compat_ulong_t __user *umask,
+                      unsigned long bitmap_size)
+{
+       int i, j;
+       unsigned long m;
+       compat_ulong_t um;
+       unsigned long nr_compat_longs;
+
+       /* align bitmap up to nearest compat_long_t boundary */
+       bitmap_size = ALIGN(bitmap_size, BITS_PER_COMPAT_LONG);
+
+       if (verify_area(VERIFY_READ, umask, bitmap_size / 8))
+               return -EFAULT;
+
+       nr_compat_longs = BITS_TO_COMPAT_LONGS(bitmap_size);
+
+       for (i = 0; i < BITS_TO_LONGS(bitmap_size); i++) {
+               m = 0;
+
+               for (j = 0; j < sizeof(m)/sizeof(um); j++) {
+                       /*
+                        * We dont want to read past the end of the userspace
+                        * bitmap. We must however ensure the end of the
+                        * kernel bitmap is zeroed.
+                        */
+                       if (nr_compat_longs-- > 0) {
+                               if (__get_user(um, umask))
+                                       return -EFAULT;
+                       } else {
+                               um = 0;
+                       }
+
+                       umask++;
+                       m |= (long)um << (j * BITS_PER_COMPAT_LONG);
+               }
+               *mask++ = m;
+       }
+
+       return 0;
+}
+
+long compat_put_bitmap(compat_ulong_t __user *umask, unsigned long *mask,
+                      unsigned long bitmap_size)
+{
+       int i, j;
+       unsigned long m;
+       compat_ulong_t um;
+       unsigned long nr_compat_longs;
+
+       /* align bitmap up to nearest compat_long_t boundary */
+       bitmap_size = ALIGN(bitmap_size, BITS_PER_COMPAT_LONG);
+
+       if (verify_area(VERIFY_WRITE, umask, bitmap_size / 8))
+               return -EFAULT;
+
+       nr_compat_longs = BITS_TO_COMPAT_LONGS(bitmap_size);
+
+       for (i = 0; i < BITS_TO_LONGS(bitmap_size); i++) {
+               m = *mask++;
+
+               for (j = 0; j < sizeof(m)/sizeof(um); j++) {
+                       um = m;
+
+                       /*
+                        * We dont want to write past the end of the userspace
+                        * bitmap.
+                        */
+                       if (nr_compat_longs-- > 0) {
+                               if (__put_user(um, umask))
+                                       return -EFAULT;
+                       }
+
+                       umask++;
+                       m >>= 4*sizeof(um);
+                       m >>= 4*sizeof(um);
+               }
+       }
+
+       return 0;
+}
+
+void
+sigset_from_compat (sigset_t *set, compat_sigset_t *compat)
+{
+       switch (_NSIG_WORDS) {
+#if defined (__COMPAT_ENDIAN_SWAP__)
+       case 4: set->sig[3] = compat->sig[7] | (((long)compat->sig[6]) << 32 );
+       case 3: set->sig[2] = compat->sig[5] | (((long)compat->sig[4]) << 32 );
+       case 2: set->sig[1] = compat->sig[3] | (((long)compat->sig[2]) << 32 );
+       case 1: set->sig[0] = compat->sig[1] | (((long)compat->sig[0]) << 32 );
+#else
+       case 4: set->sig[3] = compat->sig[6] | (((long)compat->sig[7]) << 32 );
+       case 3: set->sig[2] = compat->sig[4] | (((long)compat->sig[5]) << 32 );
+       case 2: set->sig[1] = compat->sig[2] | (((long)compat->sig[3]) << 32 );
+       case 1: set->sig[0] = compat->sig[0] | (((long)compat->sig[1]) << 32 );
+#endif
+       }
+}
+
+asmlinkage long
+compat_sys_rt_sigtimedwait (compat_sigset_t __user *uthese,
+               struct compat_siginfo __user *uinfo,
+               struct compat_timespec __user *uts, compat_size_t sigsetsize)
+{
+       compat_sigset_t s32;
+       sigset_t s;
+       int sig;
+       struct timespec t;
+       siginfo_t info;
+       long ret, timeout = 0;
+
+       if (sigsetsize != sizeof(sigset_t))
+               return -EINVAL;
+
+       if (copy_from_user(&s32, uthese, sizeof(compat_sigset_t)))
+               return -EFAULT;
+       sigset_from_compat(&s, &s32);
+       sigdelsetmask(&s,sigmask(SIGKILL)|sigmask(SIGSTOP));
+       signotset(&s);
+
+       if (uts) {
+               if (get_compat_timespec (&t, uts))
+                       return -EFAULT;
+               if (t.tv_nsec >= 1000000000L || t.tv_nsec < 0
+                               || t.tv_sec < 0)
+                       return -EINVAL;
+       }
+
+       spin_lock_irq(&current->sighand->siglock);
+       sig = dequeue_signal(current, &s, &info);
+       if (!sig) {
+               timeout = MAX_SCHEDULE_TIMEOUT;
+               if (uts)
+                       timeout = timespec_to_jiffies(&t)
+                               +(t.tv_sec || t.tv_nsec);
+               if (timeout) {
+                       current->real_blocked = current->blocked;
+                       sigandsets(&current->blocked, &current->blocked, &s);
+
+                       recalc_sigpending();
+                       spin_unlock_irq(&current->sighand->siglock);
+
+                       current->state = TASK_INTERRUPTIBLE;
+                       timeout = schedule_timeout(timeout);
+
+                       spin_lock_irq(&current->sighand->siglock);
+                       sig = dequeue_signal(current, &s, &info);
+                       current->blocked = current->real_blocked;
+                       siginitset(&current->real_blocked, 0);
+                       recalc_sigpending();
+               }
+       }
+       spin_unlock_irq(&current->sighand->siglock);
+
+       if (sig) {
+               ret = sig;
+               if (uinfo) {
+                       if (copy_siginfo_to_user32(uinfo, &info))
+                               ret = -EFAULT;
+               }
+       }else {
+               ret = timeout?-EINTR:-EAGAIN;
+       }
+       return ret;
+
+}
+
+#ifdef __ARCH_WANT_COMPAT_SYS_TIME
+
+/* compat_time_t is a 32 bit "long" and needs to get converted. */
+
+asmlinkage long compat_sys_time(compat_time_t __user * tloc)
+{
+       compat_time_t i;
+       struct timeval tv;
+
+       do_gettimeofday(&tv);
+       i = tv.tv_sec;
+
+       if (tloc) {
+               if (put_user(i,tloc))
+                       i = -EFAULT;
+       }
+       return i;
+}
+
+asmlinkage long compat_sys_stime(compat_time_t __user *tptr)
+{
+       struct timespec tv;
+       int err;
+
+       if (get_user(tv.tv_sec, tptr))
+               return -EFAULT;
+
+       tv.tv_nsec = 0;
+
+       err = security_settime(&tv, NULL);
+       if (err)
+               return err;
+
+       do_settimeofday(&tv);
+       return 0;
+}
+
+#endif /* __ARCH_WANT_COMPAT_SYS_TIME */