[PATCH] use kzalloc and kcalloc in core fs code
[safe/jmp/linux-2.6] / fs / compat.c
index 8c66570..ef5a077 100644 (file)
@@ -44,6 +44,8 @@
 #include <linux/nfsd/syscall.h>
 #include <linux/personality.h>
 #include <linux/rwsem.h>
+#include <linux/acct.h>
+#include <linux/mm.h>
 
 #include <net/sock.h>          /* siocdevprivate_ioctl */
 
@@ -51,6 +53,8 @@
 #include <asm/mmu_context.h>
 #include <asm/ioctls.h>
 
+extern void sigset_from_compat(sigset_t *set, compat_sigset_t *compat);
+
 /*
  * Not all architectures have sys_utime, so implement this in terms
  * of sys_utimes.
@@ -66,28 +70,33 @@ asmlinkage long compat_sys_utime(char __user *filename, struct compat_utimbuf __
                tv[0].tv_usec = 0;
                tv[1].tv_usec = 0;
        }
-       return do_utimes(filename, t ? tv : NULL);
+       return do_utimes(AT_FDCWD, filename, t ? tv : NULL);
 }
 
-asmlinkage long compat_sys_utimes(char __user *filename, struct compat_timeval __user *t)
+asmlinkage long compat_sys_futimesat(unsigned int dfd, char __user *filename, struct compat_timeval __user *t)
 {
        struct timeval tv[2];
 
-       if (t) { 
+       if (t) {
                if (get_user(tv[0].tv_sec, &t[0].tv_sec) ||
                    get_user(tv[0].tv_usec, &t[0].tv_usec) ||
                    get_user(tv[1].tv_sec, &t[1].tv_sec) ||
                    get_user(tv[1].tv_usec, &t[1].tv_usec))
-                       return -EFAULT; 
-       } 
-       return do_utimes(filename, t ? tv : NULL);
+                       return -EFAULT;
+       }
+       return do_utimes(dfd, filename, t ? tv : NULL);
+}
+
+asmlinkage long compat_sys_utimes(char __user *filename, struct compat_timeval __user *t)
+{
+       return compat_sys_futimesat(AT_FDCWD, filename, t);
 }
 
 asmlinkage long compat_sys_newstat(char __user * filename,
                struct compat_stat __user *statbuf)
 {
        struct kstat stat;
-       int error = vfs_stat(filename, &stat);
+       int error = vfs_stat_fd(AT_FDCWD, filename, &stat);
 
        if (!error)
                error = cp_compat_stat(&stat, statbuf);
@@ -98,12 +107,35 @@ asmlinkage long compat_sys_newlstat(char __user * filename,
                struct compat_stat __user *statbuf)
 {
        struct kstat stat;
-       int error = vfs_lstat(filename, &stat);
+       int error = vfs_lstat_fd(AT_FDCWD, filename, &stat);
+
+       if (!error)
+               error = cp_compat_stat(&stat, statbuf);
+       return error;
+}
+
+#ifndef __ARCH_WANT_STAT64
+asmlinkage long compat_sys_newfstatat(unsigned int dfd, char __user *filename,
+               struct compat_stat __user *statbuf, int flag)
+{
+       struct kstat stat;
+       int error = -EINVAL;
+
+       if ((flag & ~AT_SYMLINK_NOFOLLOW) != 0)
+               goto out;
+
+       if (flag & AT_SYMLINK_NOFOLLOW)
+               error = vfs_lstat_fd(dfd, filename, &stat);
+       else
+               error = vfs_stat_fd(dfd, filename, &stat);
 
        if (!error)
                error = cp_compat_stat(&stat, statbuf);
+
+out:
        return error;
 }
+#endif
 
 asmlinkage long compat_sys_newfstat(unsigned int fd,
                struct compat_stat __user * statbuf)
@@ -166,8 +198,8 @@ asmlinkage long compat_sys_statfs(const char __user *path, struct compat_statfs
        if (!error) {
                struct kstatfs tmp;
                error = vfs_statfs(nd.dentry->d_inode->i_sb, &tmp);
-               if (!error && put_compat_statfs(buf, &tmp))
-                       error = -EFAULT;
+               if (!error)
+                       error = put_compat_statfs(buf, &tmp);
                path_release(&nd);
        }
        return error;
@@ -184,8 +216,8 @@ asmlinkage long compat_sys_fstatfs(unsigned int fd, struct compat_statfs __user
        if (!file)
                goto out;
        error = vfs_statfs(file->f_dentry->d_inode->i_sb, &tmp);
-       if (!error && put_compat_statfs(buf, &tmp))
-               error = -EFAULT;
+       if (!error)
+               error = put_compat_statfs(buf, &tmp);
        fput(file);
 out:
        return error;
@@ -234,8 +266,8 @@ asmlinkage long compat_sys_statfs64(const char __user *path, compat_size_t sz, s
        if (!error) {
                struct kstatfs tmp;
                error = vfs_statfs(nd.dentry->d_inode->i_sb, &tmp);
-               if (!error && put_compat_statfs64(buf, &tmp))
-                       error = -EFAULT;
+               if (!error)
+                       error = put_compat_statfs64(buf, &tmp);
                path_release(&nd);
        }
        return error;
@@ -255,8 +287,8 @@ asmlinkage long compat_sys_fstatfs64(unsigned int fd, compat_size_t sz, struct c
        if (!file)
                goto out;
        error = vfs_statfs(file->f_dentry->d_inode->i_sb, &tmp);
-       if (!error && put_compat_statfs64(buf, &tmp))
-               error = -EFAULT;
+       if (!error)
+               error = put_compat_statfs64(buf, &tmp);
        fput(file);
 out:
        return error;
@@ -266,7 +298,6 @@ out:
 
 #define IOCTL_HASHSIZE 256
 static struct ioctl_trans *ioctl32_hash_table[IOCTL_HASHSIZE];
-static DECLARE_RWSEM(ioctl32_sem);
 
 extern struct ioctl_trans ioctl_start[];
 extern int ioctl_table_size;
@@ -388,14 +419,10 @@ asmlinkage long compat_sys_ioctl(unsigned int fd, unsigned int cmd,
                break;
        }
 
-       /* When register_ioctl32_conversion is finally gone remove
-          this lock! -AK */
-       down_read(&ioctl32_sem);
        for (t = ioctl32_hash_table[ioctl32_hash(cmd)]; t; t = t->next) {
                if (t->cmd == cmd)
                        goto found_handler;
        }
-       up_read(&ioctl32_sem);
 
        if (S_ISSOCK(filp->f_dentry->d_inode->i_mode) &&
            cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15)) {
@@ -415,11 +442,9 @@ asmlinkage long compat_sys_ioctl(unsigned int fd, unsigned int cmd,
                lock_kernel();
                error = t->handler(fd, cmd, arg, filp);
                unlock_kernel();
-               up_read(&ioctl32_sem);
                goto out_fput;
        }
 
-       up_read(&ioctl32_sem);
  do_ioctl:
        error = vfs_ioctl(filp, fd, cmd, arg);
  out_fput:
@@ -499,9 +524,21 @@ asmlinkage long compat_sys_fcntl64(unsigned int fd, unsigned int cmd,
                ret = sys_fcntl(fd, cmd, (unsigned long)&f);
                set_fs(old_fs);
                if (cmd == F_GETLK && ret == 0) {
-                       if ((f.l_start >= COMPAT_OFF_T_MAX) ||
-                           ((f.l_start + f.l_len) > COMPAT_OFF_T_MAX))
+                       /* GETLK was successfule and we need to return the data...
+                        * but it needs to fit in the compat structure.
+                        * l_start shouldn't be too big, unless the original
+                        * start + end is greater than COMPAT_OFF_T_MAX, in which
+                        * case the app was asking for trouble, so we return
+                        * -EOVERFLOW in that case.
+                        * l_len could be too big, in which case we just truncate it,
+                        * and only allow the app to see that part of the conflicting
+                        * lock that might make sense to it anyway
+                        */
+
+                       if (f.l_start > COMPAT_OFF_T_MAX)
                                ret = -EOVERFLOW;
+                       if (f.l_len > COMPAT_OFF_T_MAX)
+                               f.l_len = COMPAT_OFF_T_MAX;
                        if (ret == 0)
                                ret = put_compat_flock(&f, compat_ptr(arg));
                }
@@ -520,9 +557,11 @@ asmlinkage long compat_sys_fcntl64(unsigned int fd, unsigned int cmd,
                                (unsigned long)&f);
                set_fs(old_fs);
                if (cmd == F_GETLK64 && ret == 0) {
-                       if ((f.l_start >= COMPAT_LOFF_T_MAX) ||
-                           ((f.l_start + f.l_len) > COMPAT_LOFF_T_MAX))
+                       /* need to return lock information - see above for commentary */
+                       if (f.l_start > COMPAT_LOFF_T_MAX)
                                ret = -EOVERFLOW;
+                       if (f.l_len > COMPAT_LOFF_T_MAX)
+                               f.l_len = COMPAT_LOFF_T_MAX;
                        if (ret == 0)
                                ret = put_compat_flock64(&f, compat_ptr(arg));
                }
@@ -1175,7 +1214,7 @@ static ssize_t compat_do_readv_writev(int type, struct file *file,
        }
 
        ret = rw_verify_area(type, file, pos, tot_len);
-       if (ret)
+       if (ret < 0)
                goto out;
 
        fnv = NULL;
@@ -1281,7 +1320,17 @@ out:
 asmlinkage long
 compat_sys_open(const char __user *filename, int flags, int mode)
 {
-       return do_sys_open(filename, flags, mode);
+       return do_sys_open(AT_FDCWD, filename, flags, mode);
+}
+
+/*
+ * Exactly like fs/open.c:sys_openat(), except that it doesn't set the
+ * O_LARGEFILE flag.
+ */
+asmlinkage long
+compat_sys_openat(unsigned int dfd, const char __user *filename, int flags, int mode)
+{
+       return do_sys_open(dfd, filename, flags, mode);
 }
 
 /*
@@ -1427,10 +1476,9 @@ int compat_do_execve(char * filename,
        int i;
 
        retval = -ENOMEM;
-       bprm = kmalloc(sizeof(*bprm), GFP_KERNEL);
+       bprm = kzalloc(sizeof(*bprm), GFP_KERNEL);
        if (!bprm)
                goto out_ret;
-       memset(bprm, 0, sizeof(*bprm));
 
        file = open_exec(filename);
        retval = PTR_ERR(file);
@@ -1487,6 +1535,7 @@ int compat_do_execve(char * filename,
 
                /* execve success */
                security_bprm_free(bprm);
+               acct_update_integrals(current);
                kfree(bprm);
                return retval;
        }
@@ -1527,7 +1576,7 @@ out_ret:
  * Ooo, nasty.  We need here to frob 32-bit unsigned longs to
  * 64-bit unsigned longs.
  */
-static inline
+static
 int compat_get_fd_set(unsigned long nr, compat_ulong_t __user *ufdset,
                        unsigned long *fdset)
 {
@@ -1560,7 +1609,7 @@ int compat_get_fd_set(unsigned long nr, compat_ulong_t __user *ufdset,
        return 0;
 }
 
-static inline
+static
 void compat_set_fd_set(unsigned long nr, compat_ulong_t __user *ufdset,
                        unsigned long *fdset)
 {
@@ -1611,40 +1660,22 @@ static void select_bits_free(void *bits, int size)
 #define MAX_SELECT_SECONDS \
        ((unsigned long) (MAX_SCHEDULE_TIMEOUT / HZ)-1)
 
-asmlinkage long
-compat_sys_select(int n, compat_ulong_t __user *inp, compat_ulong_t __user *outp,
-               compat_ulong_t __user *exp, struct compat_timeval __user *tvp)
+int compat_core_sys_select(int n, compat_ulong_t __user *inp,
+       compat_ulong_t __user *outp, compat_ulong_t __user *exp, s64 *timeout)
 {
        fd_set_bits fds;
        char *bits;
-       long timeout;
        int size, max_fdset, ret = -EINVAL;
-
-       timeout = MAX_SCHEDULE_TIMEOUT;
-       if (tvp) {
-               time_t sec, usec;
-
-               if (!access_ok(VERIFY_READ, tvp, sizeof(*tvp))
-                   || __get_user(sec, &tvp->tv_sec)
-                   || __get_user(usec, &tvp->tv_usec)) {
-                       ret = -EFAULT;
-                       goto out_nofds;
-               }
-
-               if (sec < 0 || usec < 0)
-                       goto out_nofds;
-
-               if ((unsigned long) sec < MAX_SELECT_SECONDS) {
-                       timeout = ROUND_UP(usec, 1000000/HZ);
-                       timeout += sec * (unsigned long) HZ;
-               }
-       }
+       struct fdtable *fdt;
 
        if (n < 0)
                goto out_nofds;
 
        /* max_fdset can increase, so grab it once to avoid race */
-       max_fdset = current->files->max_fdset;
+       rcu_read_lock();
+       fdt = files_fdtable(current->files);
+       max_fdset = fdt->max_fdset;
+       rcu_read_unlock();
        if (n > max_fdset)
                n = max_fdset;
 
@@ -1673,19 +1704,7 @@ compat_sys_select(int n, compat_ulong_t __user *inp, compat_ulong_t __user *outp
        zero_fd_set(n, fds.res_out);
        zero_fd_set(n, fds.res_ex);
 
-       ret = do_select(n, &fds, &timeout);
-
-       if (tvp && !(current->personality & STICKY_TIMEOUTS)) {
-               time_t sec = 0, usec = 0;
-               if (timeout) {
-                       sec = timeout / HZ;
-                       usec = timeout % HZ;
-                       usec *= (1000000/HZ);
-               }
-               if (put_user(sec, &tvp->tv_sec) ||
-                   put_user(usec, &tvp->tv_usec))
-                       ret = -EFAULT;
-       }
+       ret = do_select(n, &fds, timeout);
 
        if (ret < 0)
                goto out;
@@ -1706,6 +1725,237 @@ out_nofds:
        return ret;
 }
 
+asmlinkage long compat_sys_select(int n, compat_ulong_t __user *inp,
+       compat_ulong_t __user *outp, compat_ulong_t __user *exp,
+       struct compat_timeval __user *tvp)
+{
+       s64 timeout = -1;
+       struct compat_timeval tv;
+       int ret;
+
+       if (tvp) {
+               if (copy_from_user(&tv, tvp, sizeof(tv)))
+                       return -EFAULT;
+
+               if (tv.tv_sec < 0 || tv.tv_usec < 0)
+                       return -EINVAL;
+
+               /* Cast to u64 to make GCC stop complaining */
+               if ((u64)tv.tv_sec >= (u64)MAX_INT64_SECONDS)
+                       timeout = -1;   /* infinite */
+               else {
+                       timeout = ROUND_UP(tv.tv_usec, 1000000/HZ);
+                       timeout += tv.tv_sec * HZ;
+               }
+       }
+
+       ret = compat_core_sys_select(n, inp, outp, exp, &timeout);
+
+       if (tvp) {
+               struct compat_timeval rtv;
+
+               if (current->personality & STICKY_TIMEOUTS)
+                       goto sticky;
+               rtv.tv_usec = jiffies_to_usecs(do_div((*(u64*)&timeout), HZ));
+               rtv.tv_sec = timeout;
+               if (compat_timeval_compare(&rtv, &tv) >= 0)
+                       rtv = tv;
+               if (copy_to_user(tvp, &rtv, sizeof(rtv))) {
+sticky:
+                       /*
+                        * If an application puts its timeval in read-only
+                        * memory, we don't want the Linux-specific update to
+                        * the timeval to cause a fault after the select has
+                        * completed successfully. However, because we're not
+                        * updating the timeval, we can't restart the system
+                        * call.
+                        */
+                       if (ret == -ERESTARTNOHAND)
+                               ret = -EINTR;
+               }
+       }
+
+       return ret;
+}
+
+#ifdef TIF_RESTORE_SIGMASK
+asmlinkage long compat_sys_pselect7(int n, compat_ulong_t __user *inp,
+       compat_ulong_t __user *outp, compat_ulong_t __user *exp,
+       struct compat_timespec __user *tsp, compat_sigset_t __user *sigmask,
+       compat_size_t sigsetsize)
+{
+       compat_sigset_t ss32;
+       sigset_t ksigmask, sigsaved;
+       s64 timeout = MAX_SCHEDULE_TIMEOUT;
+       struct compat_timespec ts;
+       int ret;
+
+       if (tsp) {
+               if (copy_from_user(&ts, tsp, sizeof(ts)))
+                       return -EFAULT;
+
+               if (ts.tv_sec < 0 || ts.tv_nsec < 0)
+                       return -EINVAL;
+       }
+
+       if (sigmask) {
+               if (sigsetsize != sizeof(compat_sigset_t))
+                       return -EINVAL;
+               if (copy_from_user(&ss32, sigmask, sizeof(ss32)))
+                       return -EFAULT;
+               sigset_from_compat(&ksigmask, &ss32);
+
+               sigdelsetmask(&ksigmask, sigmask(SIGKILL)|sigmask(SIGSTOP));
+               sigprocmask(SIG_SETMASK, &ksigmask, &sigsaved);
+       }
+
+       do {
+               if (tsp) {
+                       if ((unsigned long)ts.tv_sec < MAX_SELECT_SECONDS) {
+                               timeout = ROUND_UP(ts.tv_nsec, 1000000000/HZ);
+                               timeout += ts.tv_sec * (unsigned long)HZ;
+                               ts.tv_sec = 0;
+                               ts.tv_nsec = 0;
+                       } else {
+                               ts.tv_sec -= MAX_SELECT_SECONDS;
+                               timeout = MAX_SELECT_SECONDS * HZ;
+                       }
+               }
+
+               ret = compat_core_sys_select(n, inp, outp, exp, &timeout);
+
+       } while (!ret && !timeout && tsp && (ts.tv_sec || ts.tv_nsec));
+
+       if (tsp && !(current->personality & STICKY_TIMEOUTS)) {
+               struct compat_timespec rts;
+
+               rts.tv_sec = timeout / HZ;
+               rts.tv_nsec = (timeout % HZ) * (NSEC_PER_SEC/HZ);
+               if (rts.tv_nsec >= NSEC_PER_SEC) {
+                       rts.tv_sec++;
+                       rts.tv_nsec -= NSEC_PER_SEC;
+               }
+               if (compat_timespec_compare(&rts, &ts) >= 0)
+                       rts = ts;
+               copy_to_user(tsp, &rts, sizeof(rts));
+       }
+
+       if (ret == -ERESTARTNOHAND) {
+               /*
+                * Don't restore the signal mask yet. Let do_signal() deliver
+                * the signal on the way back to userspace, before the signal
+                * mask is restored.
+                */
+               if (sigmask) {
+                       memcpy(&current->saved_sigmask, &sigsaved,
+                                       sizeof(sigsaved));
+                       set_thread_flag(TIF_RESTORE_SIGMASK);
+               }
+       } else if (sigmask)
+               sigprocmask(SIG_SETMASK, &sigsaved, NULL);
+
+       return ret;
+}
+
+asmlinkage long compat_sys_pselect6(int n, compat_ulong_t __user *inp,
+       compat_ulong_t __user *outp, compat_ulong_t __user *exp,
+       struct compat_timespec __user *tsp, void __user *sig)
+{
+       compat_size_t sigsetsize = 0;
+       compat_uptr_t up = 0;
+
+       if (sig) {
+               if (!access_ok(VERIFY_READ, sig,
+                               sizeof(compat_uptr_t)+sizeof(compat_size_t)) ||
+                       __get_user(up, (compat_uptr_t __user *)sig) ||
+                       __get_user(sigsetsize,
+                               (compat_size_t __user *)(sig+sizeof(up))))
+                       return -EFAULT;
+       }
+       return compat_sys_pselect7(n, inp, outp, exp, tsp, compat_ptr(up),
+                                       sigsetsize);
+}
+
+asmlinkage long compat_sys_ppoll(struct pollfd __user *ufds,
+       unsigned int nfds, struct compat_timespec __user *tsp,
+       const compat_sigset_t __user *sigmask, compat_size_t sigsetsize)
+{
+       compat_sigset_t ss32;
+       sigset_t ksigmask, sigsaved;
+       struct compat_timespec ts;
+       s64 timeout = -1;
+       int ret;
+
+       if (tsp) {
+               if (copy_from_user(&ts, tsp, sizeof(ts)))
+                       return -EFAULT;
+
+               /* We assume that ts.tv_sec is always lower than
+                  the number of seconds that can be expressed in
+                  an s64. Otherwise the compiler bitches at us */
+               timeout = ROUND_UP(ts.tv_nsec, 1000000000/HZ);
+               timeout += ts.tv_sec * HZ;
+       }
+
+       if (sigmask) {
+               if (sigsetsize |= sizeof(compat_sigset_t))
+                       return -EINVAL;
+               if (copy_from_user(&ss32, sigmask, sizeof(ss32)))
+                       return -EFAULT;
+               sigset_from_compat(&ksigmask, &ss32);
+
+               sigdelsetmask(&ksigmask, sigmask(SIGKILL)|sigmask(SIGSTOP));
+               sigprocmask(SIG_SETMASK, &ksigmask, &sigsaved);
+       }
+
+       ret = do_sys_poll(ufds, nfds, &timeout);
+
+       /* We can restart this syscall, usually */
+       if (ret == -EINTR) {
+               /*
+                * Don't restore the signal mask yet. Let do_signal() deliver
+                * the signal on the way back to userspace, before the signal
+                * mask is restored.
+                */
+               if (sigmask) {
+                       memcpy(&current->saved_sigmask, &sigsaved,
+                               sizeof(sigsaved));
+                       set_thread_flag(TIF_RESTORE_SIGMASK);
+               }
+               ret = -ERESTARTNOHAND;
+       } else if (sigmask)
+               sigprocmask(SIG_SETMASK, &sigsaved, NULL);
+
+       if (tsp && timeout >= 0) {
+               struct compat_timespec rts;
+
+               if (current->personality & STICKY_TIMEOUTS)
+                       goto sticky;
+               /* Yes, we know it's actually an s64, but it's also positive. */
+               rts.tv_nsec = jiffies_to_usecs(do_div((*(u64*)&timeout), HZ)) *
+                                       1000;
+               rts.tv_sec = timeout;
+               if (compat_timespec_compare(&rts, &ts) >= 0)
+                       rts = ts;
+               if (copy_to_user(tsp, &rts, sizeof(rts))) {
+sticky:
+                       /*
+                        * If an application puts its timeval in read-only
+                        * memory, we don't want the Linux-specific update to
+                        * the timeval to cause a fault after the select has
+                        * completed successfully. However, because we're not
+                        * updating the timeval, we can't restart the system
+                        * call.
+                        */
+                       if (ret == -ERESTARTNOHAND && timeout >= 0)
+                               ret = -EINTR;
+               }
+       }
+
+       return ret;
+}
+#endif /* TIF_RESTORE_SIGMASK */
+
 #if defined(CONFIG_NFSD) || defined(CONFIG_NFSD_MODULE)
 /* Stuff for NFS server syscalls... */
 struct compat_nfsctl_svc {
@@ -1919,9 +2169,12 @@ asmlinkage long compat_sys_nfsservctl(int cmd, struct compat_nfsctl_arg __user *
 
        default:
                err = -EINVAL;
-               goto done;
+               break;
        }
 
+       if (err)
+               goto done;
+
        oldfs = get_fs();
        set_fs(KERNEL_DS);
        /* The __user pointer casts are valid because of the set_fs() */