Merge Linus's staging merge point into staging-next
authorGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Wed, 6 Dec 2017 14:27:17 +0000 (15:27 +0100)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Wed, 6 Dec 2017 14:27:17 +0000 (15:27 +0100)
This resolves the merge issue pointed out by Stephen in
drivers/iio/adc/meson_saradc.c.

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
1  2 
MAINTAINERS
drivers/iio/adc/meson_saradc.c
drivers/iio/health/max30102.c
drivers/iio/industrialio-core.c
drivers/iio/proximity/sx9500.c
drivers/staging/ccree/ssi_hash.c
drivers/staging/comedi/drivers/ni_atmio.c
drivers/staging/ncpfs/inode.c

diff --cc MAINTAINERS
Simple merge
index f0b6502a8904f44fd64b6bd5d29e46ae230ff576,36047147ce7c727003a0f08df34880302ebd2e08..48fdbb3a061b10a051411a166c47097432b576ac
@@@ -732,8 -762,9 +755,7 @@@ static int meson_sar_adc_hw_enable(stru
  err_adc_clk:
        regmap_update_bits(priv->regmap, MESON_SAR_ADC_REG3,
                           MESON_SAR_ADC_REG3_ADC_EN, 0);
-       regmap_update_bits(priv->regmap, MESON_SAR_ADC_REG11,
-                          MESON_SAR_ADC_REG11_BANDGAP_EN, 0);
+       meson_sar_adc_set_bandgap(indio_dev, false);
 -      clk_disable_unprepare(priv->sana_clk);
 -err_sana_clk:
        clk_disable_unprepare(priv->core_clk);
  err_core_clk:
        regulator_disable(priv->vref);
@@@ -756,9 -787,10 +778,9 @@@ static int meson_sar_adc_hw_disable(str
  
        regmap_update_bits(priv->regmap, MESON_SAR_ADC_REG3,
                           MESON_SAR_ADC_REG3_ADC_EN, 0);
-       regmap_update_bits(priv->regmap, MESON_SAR_ADC_REG11,
-                          MESON_SAR_ADC_REG11_BANDGAP_EN, 0);
+       meson_sar_adc_set_bandgap(indio_dev, false);
  
 -      clk_disable_unprepare(priv->sana_clk);
        clk_disable_unprepare(priv->core_clk);
  
        regulator_disable(priv->vref);
index 490dfbaf35ac25ddb52ab08edffa885980ab3918,147a8c14235f3f8b39302c8f78beec811f6871dd..15ccadc74891f73dca409acb0806cb4136646da9
@@@ -482,22 -355,23 +482,22 @@@ static int max30102_read_raw(struct iio
        switch (mask) {
        case IIO_CHAN_INFO_RAW:
                /*
 -               * Temperature reading can only be acquired while engine
 -               * is running
 +               * Temperature reading can only be acquired when not in
 +               * shutdown; leave shutdown briefly when buffer not running
                 */
                mutex_lock(&indio_dev->mlock);
 -
                if (!iio_buffer_enabled(indio_dev))
 -                      ret = -EBUSY;
 -              else {
 -                      ret = max30102_get_temp(data, val);
 -                      if (!ret)
 -                              ret = IIO_VAL_INT;
 -              }
 -
 +                      ret = max30102_get_temp(data, val, true);
 +              else
 +                      ret = max30102_get_temp(data, val, false);
                mutex_unlock(&indio_dev->mlock);
 +              if (ret)
 +                      return ret;
 +
 +              ret = IIO_VAL_INT;
                break;
        case IIO_CHAN_INFO_SCALE:
-               *val = 1;  /* 0.0625 */
+               *val = 1000;  /* 62.5 */
                *val2 = 16;
                ret = IIO_VAL_FRACTIONAL;
                break;
Simple merge
Simple merge
Simple merge
index 129f1937fa2c11527633c98a8840c2aae011fa0e,0000000000000000000000000000000000000000..41de88cdc053fa4c6f8921cd2ee29f1a911642a4
mode 100644,000000..100644
--- /dev/null
@@@ -1,1066 -1,0 +1,1066 @@@
-       *flags |= MS_NODIRATIME;
 +/*
 + *  inode.c
 + *
 + *  Copyright (C) 1995, 1996 by Volker Lendecke
 + *  Modified for big endian by J.F. Chadima and David S. Miller
 + *  Modified 1997 Peter Waltenberg, Bill Hawes, David Woodhouse for 2.1 dcache
 + *  Modified 1998 Wolfram Pienkoss for NLS
 + *  Modified 2000 Ben Harris, University of Cambridge for NFS NS meta-info
 + *
 + */
 +
 +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 +
 +#include <linux/module.h>
 +
 +#include <linux/uaccess.h>
 +#include <asm/byteorder.h>
 +
 +#include <linux/time.h>
 +#include <linux/kernel.h>
 +#include <linux/mm.h>
 +#include <linux/string.h>
 +#include <linux/stat.h>
 +#include <linux/errno.h>
 +#include <linux/file.h>
 +#include <linux/fcntl.h>
 +#include <linux/slab.h>
 +#include <linux/vmalloc.h>
 +#include <linux/init.h>
 +#include <linux/vfs.h>
 +#include <linux/mount.h>
 +#include <linux/seq_file.h>
 +#include <linux/sched/signal.h>
 +#include <linux/namei.h>
 +
 +#include <net/sock.h>
 +
 +#include "ncp_fs.h"
 +#include "getopt.h"
 +
 +#define NCP_DEFAULT_FILE_MODE 0600
 +#define NCP_DEFAULT_DIR_MODE 0700
 +#define NCP_DEFAULT_TIME_OUT 10
 +#define NCP_DEFAULT_RETRY_COUNT 20
 +
 +static void ncp_evict_inode(struct inode *);
 +static void ncp_put_super(struct super_block *);
 +static int  ncp_statfs(struct dentry *, struct kstatfs *);
 +static int  ncp_show_options(struct seq_file *, struct dentry *);
 +
 +static struct kmem_cache * ncp_inode_cachep;
 +
 +static struct inode *ncp_alloc_inode(struct super_block *sb)
 +{
 +      struct ncp_inode_info *ei;
 +      ei = (struct ncp_inode_info *)kmem_cache_alloc(ncp_inode_cachep, GFP_KERNEL);
 +      if (!ei)
 +              return NULL;
 +      return &ei->vfs_inode;
 +}
 +
 +static void ncp_i_callback(struct rcu_head *head)
 +{
 +      struct inode *inode = container_of(head, struct inode, i_rcu);
 +      kmem_cache_free(ncp_inode_cachep, NCP_FINFO(inode));
 +}
 +
 +static void ncp_destroy_inode(struct inode *inode)
 +{
 +      call_rcu(&inode->i_rcu, ncp_i_callback);
 +}
 +
 +static void init_once(void *foo)
 +{
 +      struct ncp_inode_info *ei = (struct ncp_inode_info *) foo;
 +
 +      mutex_init(&ei->open_mutex);
 +      inode_init_once(&ei->vfs_inode);
 +}
 +
 +static int init_inodecache(void)
 +{
 +      ncp_inode_cachep = kmem_cache_create("ncp_inode_cache",
 +                                           sizeof(struct ncp_inode_info),
 +                                           0, (SLAB_RECLAIM_ACCOUNT|
 +                                              SLAB_MEM_SPREAD|SLAB_ACCOUNT),
 +                                           init_once);
 +      if (ncp_inode_cachep == NULL)
 +              return -ENOMEM;
 +      return 0;
 +}
 +
 +static void destroy_inodecache(void)
 +{
 +      /*
 +       * Make sure all delayed rcu free inodes are flushed before we
 +       * destroy cache.
 +       */
 +      rcu_barrier();
 +      kmem_cache_destroy(ncp_inode_cachep);
 +}
 +
 +static int ncp_remount(struct super_block *sb, int *flags, char* data)
 +{
 +      sync_filesystem(sb);
-       sb->s_flags |= MS_NODIRATIME;   /* probably even noatime */
++      *flags |= SB_NODIRATIME;
 +      return 0;
 +}
 +
 +static const struct super_operations ncp_sops =
 +{
 +      .alloc_inode    = ncp_alloc_inode,
 +      .destroy_inode  = ncp_destroy_inode,
 +      .drop_inode     = generic_delete_inode,
 +      .evict_inode    = ncp_evict_inode,
 +      .put_super      = ncp_put_super,
 +      .statfs         = ncp_statfs,
 +      .remount_fs     = ncp_remount,
 +      .show_options   = ncp_show_options,
 +};
 +
 +/*
 + * Fill in the ncpfs-specific information in the inode.
 + */
 +static void ncp_update_dirent(struct inode *inode, struct ncp_entry_info *nwinfo)
 +{
 +      NCP_FINFO(inode)->DosDirNum = nwinfo->i.DosDirNum;
 +      NCP_FINFO(inode)->dirEntNum = nwinfo->i.dirEntNum;
 +      NCP_FINFO(inode)->volNumber = nwinfo->volume;
 +}
 +
 +void ncp_update_inode(struct inode *inode, struct ncp_entry_info *nwinfo)
 +{
 +      ncp_update_dirent(inode, nwinfo);
 +      NCP_FINFO(inode)->nwattr = nwinfo->i.attributes;
 +      NCP_FINFO(inode)->access = nwinfo->access;
 +      memcpy(NCP_FINFO(inode)->file_handle, nwinfo->file_handle,
 +                      sizeof(nwinfo->file_handle));
 +      ncp_dbg(1, "updated %s, volnum=%d, dirent=%u\n",
 +              nwinfo->i.entryName, NCP_FINFO(inode)->volNumber,
 +              NCP_FINFO(inode)->dirEntNum);
 +}
 +
 +static void ncp_update_dates(struct inode *inode, struct nw_info_struct *nwi)
 +{
 +      /* NFS namespace mode overrides others if it's set. */
 +      ncp_dbg(1, "(%s) nfs.mode=0%o\n", nwi->entryName, nwi->nfs.mode);
 +      if (nwi->nfs.mode) {
 +              /* XXX Security? */
 +              inode->i_mode = nwi->nfs.mode;
 +      }
 +
 +      inode->i_blocks = (i_size_read(inode) + NCP_BLOCK_SIZE - 1) >> NCP_BLOCK_SHIFT;
 +
 +      inode->i_mtime.tv_sec = ncp_date_dos2unix(nwi->modifyTime, nwi->modifyDate);
 +      inode->i_ctime.tv_sec = ncp_date_dos2unix(nwi->creationTime, nwi->creationDate);
 +      inode->i_atime.tv_sec = ncp_date_dos2unix(0, nwi->lastAccessDate);
 +      inode->i_atime.tv_nsec = 0;
 +      inode->i_mtime.tv_nsec = 0;
 +      inode->i_ctime.tv_nsec = 0;
 +}
 +
 +static void ncp_update_attrs(struct inode *inode, struct ncp_entry_info *nwinfo)
 +{
 +      struct nw_info_struct *nwi = &nwinfo->i;
 +      struct ncp_server *server = NCP_SERVER(inode);
 +
 +      if (nwi->attributes & aDIR) {
 +              inode->i_mode = server->m.dir_mode;
 +              /* for directories dataStreamSize seems to be some
 +                 Object ID ??? */
 +              i_size_write(inode, NCP_BLOCK_SIZE);
 +      } else {
 +              u32 size;
 +
 +              inode->i_mode = server->m.file_mode;
 +              size = le32_to_cpu(nwi->dataStreamSize);
 +              i_size_write(inode, size);
 +#ifdef CONFIG_NCPFS_EXTRAS
 +              if ((server->m.flags & (NCP_MOUNT_EXTRAS|NCP_MOUNT_SYMLINKS)) 
 +               && (nwi->attributes & aSHARED)) {
 +                      switch (nwi->attributes & (aHIDDEN|aSYSTEM)) {
 +                              case aHIDDEN:
 +                                      if (server->m.flags & NCP_MOUNT_SYMLINKS) {
 +                                              if (/* (size >= NCP_MIN_SYMLINK_SIZE)
 +                                               && */ (size <= NCP_MAX_SYMLINK_SIZE)) {
 +                                                      inode->i_mode = (inode->i_mode & ~S_IFMT) | S_IFLNK;
 +                                                      NCP_FINFO(inode)->flags |= NCPI_KLUDGE_SYMLINK;
 +                                                      break;
 +                                              }
 +                                      }
 +                                      /* FALLTHROUGH */
 +                              case 0:
 +                                      if (server->m.flags & NCP_MOUNT_EXTRAS)
 +                                              inode->i_mode |= S_IRUGO;
 +                                      break;
 +                              case aSYSTEM:
 +                                      if (server->m.flags & NCP_MOUNT_EXTRAS)
 +                                              inode->i_mode |= (inode->i_mode >> 2) & S_IXUGO;
 +                                      break;
 +                              /* case aSYSTEM|aHIDDEN: */
 +                              default:
 +                                      /* reserved combination */
 +                                      break;
 +                      }
 +              }
 +#endif
 +      }
 +      if (nwi->attributes & aRONLY) inode->i_mode &= ~S_IWUGO;
 +}
 +
 +void ncp_update_inode2(struct inode* inode, struct ncp_entry_info *nwinfo)
 +{
 +      NCP_FINFO(inode)->flags = 0;
 +      if (!atomic_read(&NCP_FINFO(inode)->opened)) {
 +              NCP_FINFO(inode)->nwattr = nwinfo->i.attributes;
 +              ncp_update_attrs(inode, nwinfo);
 +      }
 +
 +      ncp_update_dates(inode, &nwinfo->i);
 +      ncp_update_dirent(inode, nwinfo);
 +}
 +
 +/*
 + * Fill in the inode based on the ncp_entry_info structure.  Used only for brand new inodes.
 + */
 +static void ncp_set_attr(struct inode *inode, struct ncp_entry_info *nwinfo)
 +{
 +      struct ncp_server *server = NCP_SERVER(inode);
 +
 +      NCP_FINFO(inode)->flags = 0;
 +      
 +      ncp_update_attrs(inode, nwinfo);
 +
 +      ncp_dbg(2, "inode->i_mode = %u\n", inode->i_mode);
 +
 +      set_nlink(inode, 1);
 +      inode->i_uid = server->m.uid;
 +      inode->i_gid = server->m.gid;
 +
 +      ncp_update_dates(inode, &nwinfo->i);
 +      ncp_update_inode(inode, nwinfo);
 +}
 +
 +#if defined(CONFIG_NCPFS_EXTRAS) || defined(CONFIG_NCPFS_NFS_NS)
 +static const struct inode_operations ncp_symlink_inode_operations = {
 +      .get_link       = page_get_link,
 +      .setattr        = ncp_notify_change,
 +};
 +#endif
 +
 +/*
 + * Get a new inode.
 + */
 +struct inode * 
 +ncp_iget(struct super_block *sb, struct ncp_entry_info *info)
 +{
 +      struct inode *inode;
 +
 +      if (info == NULL) {
 +              pr_err("%s: info is NULL\n", __func__);
 +              return NULL;
 +      }
 +
 +      inode = new_inode(sb);
 +      if (inode) {
 +              atomic_set(&NCP_FINFO(inode)->opened, info->opened);
 +
 +              inode->i_ino = info->ino;
 +              ncp_set_attr(inode, info);
 +              if (S_ISREG(inode->i_mode)) {
 +                      inode->i_op = &ncp_file_inode_operations;
 +                      inode->i_fop = &ncp_file_operations;
 +              } else if (S_ISDIR(inode->i_mode)) {
 +                      inode->i_op = &ncp_dir_inode_operations;
 +                      inode->i_fop = &ncp_dir_operations;
 +#ifdef CONFIG_NCPFS_NFS_NS
 +              } else if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode) || S_ISFIFO(inode->i_mode) || S_ISSOCK(inode->i_mode)) {
 +                      init_special_inode(inode, inode->i_mode,
 +                              new_decode_dev(info->i.nfs.rdev));
 +#endif
 +#if defined(CONFIG_NCPFS_EXTRAS) || defined(CONFIG_NCPFS_NFS_NS)
 +              } else if (S_ISLNK(inode->i_mode)) {
 +                      inode->i_op = &ncp_symlink_inode_operations;
 +                      inode_nohighmem(inode);
 +                      inode->i_data.a_ops = &ncp_symlink_aops;
 +#endif
 +              } else {
 +                      make_bad_inode(inode);
 +              }
 +              insert_inode_hash(inode);
 +      } else
 +              pr_err("%s: iget failed!\n", __func__);
 +      return inode;
 +}
 +
 +static void
 +ncp_evict_inode(struct inode *inode)
 +{
 +      truncate_inode_pages_final(&inode->i_data);
 +      clear_inode(inode);
 +
 +      if (S_ISDIR(inode->i_mode)) {
 +              ncp_dbg(2, "put directory %ld\n", inode->i_ino);
 +      }
 +
 +      if (ncp_make_closed(inode) != 0) {
 +              /* We can't do anything but complain. */
 +              pr_err("%s: could not close\n", __func__);
 +      }
 +}
 +
 +static void ncp_stop_tasks(struct ncp_server *server) {
 +      struct sock* sk = server->ncp_sock->sk;
 +
 +      lock_sock(sk);
 +      sk->sk_error_report = server->error_report;
 +      sk->sk_data_ready   = server->data_ready;
 +      sk->sk_write_space  = server->write_space;
 +      release_sock(sk);
 +      del_timer_sync(&server->timeout_tm);
 +
 +      flush_work(&server->rcv.tq);
 +      if (sk->sk_socket->type == SOCK_STREAM)
 +              flush_work(&server->tx.tq);
 +      else
 +              flush_work(&server->timeout_tq);
 +}
 +
 +static int  ncp_show_options(struct seq_file *seq, struct dentry *root)
 +{
 +      struct ncp_server *server = NCP_SBP(root->d_sb);
 +      unsigned int tmp;
 +
 +      if (!uid_eq(server->m.uid, GLOBAL_ROOT_UID))
 +              seq_printf(seq, ",uid=%u",
 +                         from_kuid_munged(&init_user_ns, server->m.uid));
 +      if (!gid_eq(server->m.gid, GLOBAL_ROOT_GID))
 +              seq_printf(seq, ",gid=%u",
 +                         from_kgid_munged(&init_user_ns, server->m.gid));
 +      if (!uid_eq(server->m.mounted_uid, GLOBAL_ROOT_UID))
 +              seq_printf(seq, ",owner=%u",
 +                         from_kuid_munged(&init_user_ns, server->m.mounted_uid));
 +      tmp = server->m.file_mode & S_IALLUGO;
 +      if (tmp != NCP_DEFAULT_FILE_MODE)
 +              seq_printf(seq, ",mode=0%o", tmp);
 +      tmp = server->m.dir_mode & S_IALLUGO;
 +      if (tmp != NCP_DEFAULT_DIR_MODE)
 +              seq_printf(seq, ",dirmode=0%o", tmp);
 +      if (server->m.time_out != NCP_DEFAULT_TIME_OUT * HZ / 100) {
 +              tmp = server->m.time_out * 100 / HZ;
 +              seq_printf(seq, ",timeout=%u", tmp);
 +      }
 +      if (server->m.retry_count != NCP_DEFAULT_RETRY_COUNT)
 +              seq_printf(seq, ",retry=%u", server->m.retry_count);
 +      if (server->m.flags != 0)
 +              seq_printf(seq, ",flags=%lu", server->m.flags);
 +      if (server->m.wdog_pid != NULL)
 +              seq_printf(seq, ",wdogpid=%u", pid_vnr(server->m.wdog_pid));
 +
 +      return 0;
 +}
 +
 +static const struct ncp_option ncp_opts[] = {
 +      { "uid",        OPT_INT,        'u' },
 +      { "gid",        OPT_INT,        'g' },
 +      { "owner",      OPT_INT,        'o' },
 +      { "mode",       OPT_INT,        'm' },
 +      { "dirmode",    OPT_INT,        'd' },
 +      { "timeout",    OPT_INT,        't' },
 +      { "retry",      OPT_INT,        'r' },
 +      { "flags",      OPT_INT,        'f' },
 +      { "wdogpid",    OPT_INT,        'w' },
 +      { "ncpfd",      OPT_INT,        'n' },
 +      { "infofd",     OPT_INT,        'i' },  /* v5 */
 +      { "version",    OPT_INT,        'v' },
 +      { NULL,         0,              0 } };
 +
 +static int ncp_parse_options(struct ncp_mount_data_kernel *data, char *options) {
 +      int optval;
 +      char *optarg;
 +      unsigned long optint;
 +      int version = 0;
 +      int ret;
 +
 +      data->flags = 0;
 +      data->int_flags = 0;
 +      data->mounted_uid = GLOBAL_ROOT_UID;
 +      data->wdog_pid = NULL;
 +      data->ncp_fd = ~0;
 +      data->time_out = NCP_DEFAULT_TIME_OUT;
 +      data->retry_count = NCP_DEFAULT_RETRY_COUNT;
 +      data->uid = GLOBAL_ROOT_UID;
 +      data->gid = GLOBAL_ROOT_GID;
 +      data->file_mode = NCP_DEFAULT_FILE_MODE;
 +      data->dir_mode = NCP_DEFAULT_DIR_MODE;
 +      data->info_fd = -1;
 +      data->mounted_vol[0] = 0;
 +      
 +      while ((optval = ncp_getopt("ncpfs", &options, ncp_opts, NULL, &optarg, &optint)) != 0) {
 +              ret = optval;
 +              if (ret < 0)
 +                      goto err;
 +              switch (optval) {
 +                      case 'u':
 +                              data->uid = make_kuid(current_user_ns(), optint);
 +                              if (!uid_valid(data->uid)) {
 +                                      ret = -EINVAL;
 +                                      goto err;
 +                              }
 +                              break;
 +                      case 'g':
 +                              data->gid = make_kgid(current_user_ns(), optint);
 +                              if (!gid_valid(data->gid)) {
 +                                      ret = -EINVAL;
 +                                      goto err;
 +                              }
 +                              break;
 +                      case 'o':
 +                              data->mounted_uid = make_kuid(current_user_ns(), optint);
 +                              if (!uid_valid(data->mounted_uid)) {
 +                                      ret = -EINVAL;
 +                                      goto err;
 +                              }
 +                              break;
 +                      case 'm':
 +                              data->file_mode = optint;
 +                              break;
 +                      case 'd':
 +                              data->dir_mode = optint;
 +                              break;
 +                      case 't':
 +                              data->time_out = optint;
 +                              break;
 +                      case 'r':
 +                              data->retry_count = optint;
 +                              break;
 +                      case 'f':
 +                              data->flags = optint;
 +                              break;
 +                      case 'w':
 +                              data->wdog_pid = find_get_pid(optint);
 +                              break;
 +                      case 'n':
 +                              data->ncp_fd = optint;
 +                              break;
 +                      case 'i':
 +                              data->info_fd = optint;
 +                              break;
 +                      case 'v':
 +                              ret = -ECHRNG;
 +                              if (optint < NCP_MOUNT_VERSION_V4)
 +                                      goto err;
 +                              if (optint > NCP_MOUNT_VERSION_V5)
 +                                      goto err;
 +                              version = optint;
 +                              break;
 +                      
 +              }
 +      }
 +      return 0;
 +err:
 +      put_pid(data->wdog_pid);
 +      data->wdog_pid = NULL;
 +      return ret;
 +}
 +
 +static int ncp_fill_super(struct super_block *sb, void *raw_data, int silent)
 +{
 +      struct ncp_mount_data_kernel data;
 +      struct ncp_server *server;
 +      struct inode *root_inode;
 +      struct socket *sock;
 +      int error;
 +      int default_bufsize;
 +#ifdef CONFIG_NCPFS_PACKET_SIGNING
 +      int options;
 +#endif
 +      struct ncp_entry_info finfo;
 +
 +      memset(&data, 0, sizeof(data));
 +      server = kzalloc(sizeof(struct ncp_server), GFP_KERNEL);
 +      if (!server)
 +              return -ENOMEM;
 +      sb->s_fs_info = server;
 +
 +      error = -EFAULT;
 +      if (raw_data == NULL)
 +              goto out;
 +      switch (*(int*)raw_data) {
 +              case NCP_MOUNT_VERSION:
 +                      {
 +                              struct ncp_mount_data* md = (struct ncp_mount_data*)raw_data;
 +
 +                              data.flags = md->flags;
 +                              data.int_flags = NCP_IMOUNT_LOGGEDIN_POSSIBLE;
 +                              data.mounted_uid = make_kuid(current_user_ns(), md->mounted_uid);
 +                              data.wdog_pid = find_get_pid(md->wdog_pid);
 +                              data.ncp_fd = md->ncp_fd;
 +                              data.time_out = md->time_out;
 +                              data.retry_count = md->retry_count;
 +                              data.uid = make_kuid(current_user_ns(), md->uid);
 +                              data.gid = make_kgid(current_user_ns(), md->gid);
 +                              data.file_mode = md->file_mode;
 +                              data.dir_mode = md->dir_mode;
 +                              data.info_fd = -1;
 +                              memcpy(data.mounted_vol, md->mounted_vol,
 +                                      NCP_VOLNAME_LEN+1);
 +                      }
 +                      break;
 +              case NCP_MOUNT_VERSION_V4:
 +                      {
 +                              struct ncp_mount_data_v4* md = (struct ncp_mount_data_v4*)raw_data;
 +
 +                              data.flags = md->flags;
 +                              data.mounted_uid = make_kuid(current_user_ns(), md->mounted_uid);
 +                              data.wdog_pid = find_get_pid(md->wdog_pid);
 +                              data.ncp_fd = md->ncp_fd;
 +                              data.time_out = md->time_out;
 +                              data.retry_count = md->retry_count;
 +                              data.uid = make_kuid(current_user_ns(), md->uid);
 +                              data.gid = make_kgid(current_user_ns(), md->gid);
 +                              data.file_mode = md->file_mode;
 +                              data.dir_mode = md->dir_mode;
 +                              data.info_fd = -1;
 +                      }
 +                      break;
 +              default:
 +                      error = -ECHRNG;
 +                      if (memcmp(raw_data, "vers", 4) == 0) {
 +                              error = ncp_parse_options(&data, raw_data);
 +                      }
 +                      if (error)
 +                              goto out;
 +                      break;
 +      }
 +      error = -EINVAL;
 +      if (!uid_valid(data.mounted_uid) || !uid_valid(data.uid) ||
 +          !gid_valid(data.gid))
 +              goto out;
 +      sock = sockfd_lookup(data.ncp_fd, &error);
 +      if (!sock)
 +              goto out;
 +
 +      if (sock->type == SOCK_STREAM)
 +              default_bufsize = 0xF000;
 +      else
 +              default_bufsize = 1024;
 +
++      sb->s_flags |= SB_NODIRATIME;   /* probably even noatime */
 +      sb->s_maxbytes = 0xFFFFFFFFU;
 +      sb->s_blocksize = 1024; /* Eh...  Is this correct? */
 +      sb->s_blocksize_bits = 10;
 +      sb->s_magic = NCP_SUPER_MAGIC;
 +      sb->s_op = &ncp_sops;
 +      sb->s_d_op = &ncp_dentry_operations;
 +
 +      server = NCP_SBP(sb);
 +      memset(server, 0, sizeof(*server));
 +
 +      error = super_setup_bdi(sb);
 +      if (error)
 +              goto out_fput;
 +
 +      server->ncp_sock = sock;
 +      
 +      if (data.info_fd != -1) {
 +              struct socket *info_sock = sockfd_lookup(data.info_fd, &error);
 +              if (!info_sock)
 +                      goto out_fput;
 +              server->info_sock = info_sock;
 +              error = -EBADFD;
 +              if (info_sock->type != SOCK_STREAM)
 +                      goto out_fput2;
 +      }
 +
 +/*    server->lock = 0;       */
 +      mutex_init(&server->mutex);
 +      server->packet = NULL;
 +/*    server->buffer_size = 0;        */
 +/*    server->conn_status = 0;        */
 +/*    server->root_dentry = NULL;     */
 +/*    server->root_setuped = 0;       */
 +      mutex_init(&server->root_setup_lock);
 +#ifdef CONFIG_NCPFS_PACKET_SIGNING
 +/*    server->sign_wanted = 0;        */
 +/*    server->sign_active = 0;        */
 +#endif
 +      init_rwsem(&server->auth_rwsem);
 +      server->auth.auth_type = NCP_AUTH_NONE;
 +/*    server->auth.object_name_len = 0;       */
 +/*    server->auth.object_name = NULL;        */
 +/*    server->auth.object_type = 0;           */
 +/*    server->priv.len = 0;                   */
 +/*    server->priv.data = NULL;               */
 +
 +      server->m = data;
 +      /* Although anything producing this is buggy, it happens
 +         now because of PATH_MAX changes.. */
 +      if (server->m.time_out < 1) {
 +              server->m.time_out = 10;
 +              pr_info("You need to recompile your ncpfs utils..\n");
 +      }
 +      server->m.time_out = server->m.time_out * HZ / 100;
 +      server->m.file_mode = (server->m.file_mode & S_IRWXUGO) | S_IFREG;
 +      server->m.dir_mode = (server->m.dir_mode & S_IRWXUGO) | S_IFDIR;
 +
 +#ifdef CONFIG_NCPFS_NLS
 +      /* load the default NLS charsets */
 +      server->nls_vol = load_nls_default();
 +      server->nls_io = load_nls_default();
 +#endif /* CONFIG_NCPFS_NLS */
 +
 +      atomic_set(&server->dentry_ttl, 0);     /* no caching */
 +
 +      INIT_LIST_HEAD(&server->tx.requests);
 +      mutex_init(&server->rcv.creq_mutex);
 +      server->tx.creq         = NULL;
 +      server->rcv.creq        = NULL;
 +
 +      timer_setup(&server->timeout_tm, ncpdgram_timeout_call, 0);
 +#undef NCP_PACKET_SIZE
 +#define NCP_PACKET_SIZE 131072
 +      error = -ENOMEM;
 +      server->packet_size = NCP_PACKET_SIZE;
 +      server->packet = vmalloc(NCP_PACKET_SIZE);
 +      if (server->packet == NULL)
 +              goto out_nls;
 +      server->txbuf = vmalloc(NCP_PACKET_SIZE);
 +      if (server->txbuf == NULL)
 +              goto out_packet;
 +      server->rxbuf = vmalloc(NCP_PACKET_SIZE);
 +      if (server->rxbuf == NULL)
 +              goto out_txbuf;
 +
 +      lock_sock(sock->sk);
 +      server->data_ready      = sock->sk->sk_data_ready;
 +      server->write_space     = sock->sk->sk_write_space;
 +      server->error_report    = sock->sk->sk_error_report;
 +      sock->sk->sk_user_data  = server;
 +      sock->sk->sk_data_ready   = ncp_tcp_data_ready;
 +      sock->sk->sk_error_report = ncp_tcp_error_report;
 +      if (sock->type == SOCK_STREAM) {
 +              server->rcv.ptr = (unsigned char*)&server->rcv.buf;
 +              server->rcv.len = 10;
 +              server->rcv.state = 0;
 +              INIT_WORK(&server->rcv.tq, ncp_tcp_rcv_proc);
 +              INIT_WORK(&server->tx.tq, ncp_tcp_tx_proc);
 +              sock->sk->sk_write_space = ncp_tcp_write_space;
 +      } else {
 +              INIT_WORK(&server->rcv.tq, ncpdgram_rcv_proc);
 +              INIT_WORK(&server->timeout_tq, ncpdgram_timeout_proc);
 +      }
 +      release_sock(sock->sk);
 +
 +      ncp_lock_server(server);
 +      error = ncp_connect(server);
 +      ncp_unlock_server(server);
 +      if (error < 0)
 +              goto out_rxbuf;
 +      ncp_dbg(1, "NCP_SBP(sb) = %p\n", NCP_SBP(sb));
 +
 +      error = -EMSGSIZE;      /* -EREMOTESIDEINCOMPATIBLE */
 +#ifdef CONFIG_NCPFS_PACKET_SIGNING
 +      if (ncp_negotiate_size_and_options(server, default_bufsize,
 +              NCP_DEFAULT_OPTIONS, &(server->buffer_size), &options) == 0)
 +      {
 +              if (options != NCP_DEFAULT_OPTIONS)
 +              {
 +                      if (ncp_negotiate_size_and_options(server, 
 +                              default_bufsize,
 +                              options & 2, 
 +                              &(server->buffer_size), &options) != 0)
 +                              
 +                      {
 +                              goto out_disconnect;
 +                      }
 +              }
 +              ncp_lock_server(server);
 +              if (options & 2)
 +                      server->sign_wanted = 1;
 +              ncp_unlock_server(server);
 +      }
 +      else 
 +#endif        /* CONFIG_NCPFS_PACKET_SIGNING */
 +      if (ncp_negotiate_buffersize(server, default_bufsize,
 +                                   &(server->buffer_size)) != 0)
 +              goto out_disconnect;
 +      ncp_dbg(1, "bufsize = %d\n", server->buffer_size);
 +
 +      memset(&finfo, 0, sizeof(finfo));
 +      finfo.i.attributes      = aDIR;
 +      finfo.i.dataStreamSize  = 0;    /* ignored */
 +      finfo.i.dirEntNum       = 0;
 +      finfo.i.DosDirNum       = 0;
 +#ifdef CONFIG_NCPFS_SMALLDOS
 +      finfo.i.NSCreator       = NW_NS_DOS;
 +#endif
 +      finfo.volume            = NCP_NUMBER_OF_VOLUMES;
 +      /* set dates of mountpoint to Jan 1, 1986; 00:00 */
 +      finfo.i.creationTime    = finfo.i.modifyTime
 +                              = cpu_to_le16(0x0000);
 +      finfo.i.creationDate    = finfo.i.modifyDate
 +                              = finfo.i.lastAccessDate
 +                              = cpu_to_le16(0x0C21);
 +      finfo.i.nameLen         = 0;
 +      finfo.i.entryName[0]    = '\0';
 +
 +      finfo.opened            = 0;
 +      finfo.ino               = 2;    /* tradition */
 +
 +      server->name_space[finfo.volume] = NW_NS_DOS;
 +
 +      error = -ENOMEM;
 +        root_inode = ncp_iget(sb, &finfo);
 +        if (!root_inode)
 +              goto out_disconnect;
 +      ncp_dbg(1, "root vol=%d\n", NCP_FINFO(root_inode)->volNumber);
 +      sb->s_root = d_make_root(root_inode);
 +        if (!sb->s_root)
 +              goto out_disconnect;
 +      return 0;
 +
 +out_disconnect:
 +      ncp_lock_server(server);
 +      ncp_disconnect(server);
 +      ncp_unlock_server(server);
 +out_rxbuf:
 +      ncp_stop_tasks(server);
 +      vfree(server->rxbuf);
 +out_txbuf:
 +      vfree(server->txbuf);
 +out_packet:
 +      vfree(server->packet);
 +out_nls:
 +#ifdef CONFIG_NCPFS_NLS
 +      unload_nls(server->nls_io);
 +      unload_nls(server->nls_vol);
 +#endif
 +      mutex_destroy(&server->rcv.creq_mutex);
 +      mutex_destroy(&server->root_setup_lock);
 +      mutex_destroy(&server->mutex);
 +out_fput2:
 +      if (server->info_sock)
 +              sockfd_put(server->info_sock);
 +out_fput:
 +      sockfd_put(sock);
 +out:
 +      put_pid(data.wdog_pid);
 +      sb->s_fs_info = NULL;
 +      kfree(server);
 +      return error;
 +}
 +
 +static void delayed_free(struct rcu_head *p)
 +{
 +      struct ncp_server *server = container_of(p, struct ncp_server, rcu);
 +#ifdef CONFIG_NCPFS_NLS
 +      /* unload the NLS charsets */
 +      unload_nls(server->nls_vol);
 +      unload_nls(server->nls_io);
 +#endif /* CONFIG_NCPFS_NLS */
 +      kfree(server);
 +}
 +
 +static void ncp_put_super(struct super_block *sb)
 +{
 +      struct ncp_server *server = NCP_SBP(sb);
 +
 +      ncp_lock_server(server);
 +      ncp_disconnect(server);
 +      ncp_unlock_server(server);
 +
 +      ncp_stop_tasks(server);
 +
 +      mutex_destroy(&server->rcv.creq_mutex);
 +      mutex_destroy(&server->root_setup_lock);
 +      mutex_destroy(&server->mutex);
 +
 +      if (server->info_sock)
 +              sockfd_put(server->info_sock);
 +      sockfd_put(server->ncp_sock);
 +      kill_pid(server->m.wdog_pid, SIGTERM, 1);
 +      put_pid(server->m.wdog_pid);
 +
 +      kfree(server->priv.data);
 +      kfree(server->auth.object_name);
 +      vfree(server->rxbuf);
 +      vfree(server->txbuf);
 +      vfree(server->packet);
 +      call_rcu(&server->rcu, delayed_free);
 +}
 +
 +static int ncp_statfs(struct dentry *dentry, struct kstatfs *buf)
 +{
 +      struct dentry* d;
 +      struct inode* i;
 +      struct ncp_inode_info* ni;
 +      struct ncp_server* s;
 +      struct ncp_volume_info vi;
 +      struct super_block *sb = dentry->d_sb;
 +      int err;
 +      __u8 dh;
 +      
 +      d = sb->s_root;
 +      if (!d) {
 +              goto dflt;
 +      }
 +      i = d_inode(d);
 +      if (!i) {
 +              goto dflt;
 +      }
 +      ni = NCP_FINFO(i);
 +      if (!ni) {
 +              goto dflt;
 +      }
 +      s = NCP_SBP(sb);
 +      if (!s) {
 +              goto dflt;
 +      }
 +      if (!s->m.mounted_vol[0]) {
 +              goto dflt;
 +      }
 +
 +      err = ncp_dirhandle_alloc(s, ni->volNumber, ni->DosDirNum, &dh);
 +      if (err) {
 +              goto dflt;
 +      }
 +      err = ncp_get_directory_info(s, dh, &vi);
 +      ncp_dirhandle_free(s, dh);
 +      if (err) {
 +              goto dflt;
 +      }
 +      buf->f_type = NCP_SUPER_MAGIC;
 +      buf->f_bsize = vi.sectors_per_block * 512;
 +      buf->f_blocks = vi.total_blocks;
 +      buf->f_bfree = vi.free_blocks;
 +      buf->f_bavail = vi.free_blocks;
 +      buf->f_files = vi.total_dir_entries;
 +      buf->f_ffree = vi.available_dir_entries;
 +      buf->f_namelen = 12;
 +      return 0;
 +
 +      /* We cannot say how much disk space is left on a mounted
 +         NetWare Server, because free space is distributed over
 +         volumes, and the current user might have disk quotas. So
 +         free space is not that simple to determine. Our decision
 +         here is to err conservatively. */
 +
 +dflt:;
 +      buf->f_type = NCP_SUPER_MAGIC;
 +      buf->f_bsize = NCP_BLOCK_SIZE;
 +      buf->f_blocks = 0;
 +      buf->f_bfree = 0;
 +      buf->f_bavail = 0;
 +      buf->f_namelen = 12;
 +      return 0;
 +}
 +
 +int ncp_notify_change(struct dentry *dentry, struct iattr *attr)
 +{
 +      struct inode *inode = d_inode(dentry);
 +      int result = 0;
 +      __le32 info_mask;
 +      struct nw_modify_dos_info info;
 +      struct ncp_server *server;
 +
 +      result = -EIO;
 +
 +      server = NCP_SERVER(inode);
 +      if (!server)    /* How this could happen? */
 +              goto out;
 +
 +      result = -EPERM;
 +      if (IS_DEADDIR(d_inode(dentry)))
 +              goto out;
 +
 +      /* ageing the dentry to force validation */
 +      ncp_age_dentry(server, dentry);
 +
 +      result = setattr_prepare(dentry, attr);
 +      if (result < 0)
 +              goto out;
 +
 +      result = -EPERM;
 +      if ((attr->ia_valid & ATTR_UID) && !uid_eq(attr->ia_uid, server->m.uid))
 +              goto out;
 +
 +      if ((attr->ia_valid & ATTR_GID) && !gid_eq(attr->ia_gid, server->m.gid))
 +              goto out;
 +
 +      if (((attr->ia_valid & ATTR_MODE) &&
 +           (attr->ia_mode &
 +            ~(S_IFREG | S_IFDIR | S_IRWXUGO))))
 +              goto out;
 +
 +      info_mask = 0;
 +      memset(&info, 0, sizeof(info));
 +
 +#if 1 
 +        if ((attr->ia_valid & ATTR_MODE) != 0)
 +        {
 +              umode_t newmode = attr->ia_mode;
 +
 +              info_mask |= DM_ATTRIBUTES;
 +
 +                if (S_ISDIR(inode->i_mode)) {
 +                      newmode &= server->m.dir_mode;
 +              } else {
 +#ifdef CONFIG_NCPFS_EXTRAS                    
 +                      if (server->m.flags & NCP_MOUNT_EXTRAS) {
 +                              /* any non-default execute bit set */
 +                              if (newmode & ~server->m.file_mode & S_IXUGO)
 +                                      info.attributes |= aSHARED | aSYSTEM;
 +                              /* read for group/world and not in default file_mode */
 +                              else if (newmode & ~server->m.file_mode & S_IRUGO)
 +                                      info.attributes |= aSHARED;
 +                      } else
 +#endif
 +                              newmode &= server->m.file_mode;                 
 +                }
 +                if (newmode & S_IWUGO)
 +                      info.attributes &= ~(aRONLY|aRENAMEINHIBIT|aDELETEINHIBIT);
 +                else
 +                      info.attributes |=  (aRONLY|aRENAMEINHIBIT|aDELETEINHIBIT);
 +
 +#ifdef CONFIG_NCPFS_NFS_NS
 +              if (ncp_is_nfs_extras(server, NCP_FINFO(inode)->volNumber)) {
 +                      result = ncp_modify_nfs_info(server,
 +                                                   NCP_FINFO(inode)->volNumber,
 +                                                   NCP_FINFO(inode)->dirEntNum,
 +                                                   attr->ia_mode, 0);
 +                      if (result != 0)
 +                              goto out;
 +                      info.attributes &= ~(aSHARED | aSYSTEM);
 +                      {
 +                              /* mark partial success */
 +                              struct iattr tmpattr;
 +                              
 +                              tmpattr.ia_valid = ATTR_MODE;
 +                              tmpattr.ia_mode = attr->ia_mode;
 +
 +                              setattr_copy(inode, &tmpattr);
 +                              mark_inode_dirty(inode);
 +                      }
 +              }
 +#endif
 +        }
 +#endif
 +
 +      /* Do SIZE before attributes, otherwise mtime together with size does not work...
 +       */
 +      if ((attr->ia_valid & ATTR_SIZE) != 0) {
 +              int written;
 +
 +              ncp_dbg(1, "trying to change size to %llu\n", attr->ia_size);
 +
 +              if ((result = ncp_make_open(inode, O_WRONLY)) < 0) {
 +                      result = -EACCES;
 +                      goto out;
 +              }
 +              ncp_write_kernel(NCP_SERVER(inode), NCP_FINFO(inode)->file_handle,
 +                        attr->ia_size, 0, "", &written);
 +
 +              /* According to ndir, the changes only take effect after
 +                 closing the file */
 +              ncp_inode_close(inode);
 +              result = ncp_make_closed(inode);
 +              if (result)
 +                      goto out;
 +
 +              if (attr->ia_size != i_size_read(inode)) {
 +                      truncate_setsize(inode, attr->ia_size);
 +                      mark_inode_dirty(inode);
 +              }
 +      }
 +      if ((attr->ia_valid & ATTR_CTIME) != 0) {
 +              info_mask |= (DM_CREATE_TIME | DM_CREATE_DATE);
 +              ncp_date_unix2dos(attr->ia_ctime.tv_sec,
 +                           &info.creationTime, &info.creationDate);
 +      }
 +      if ((attr->ia_valid & ATTR_MTIME) != 0) {
 +              info_mask |= (DM_MODIFY_TIME | DM_MODIFY_DATE);
 +              ncp_date_unix2dos(attr->ia_mtime.tv_sec,
 +                                &info.modifyTime, &info.modifyDate);
 +      }
 +      if ((attr->ia_valid & ATTR_ATIME) != 0) {
 +              __le16 dummy;
 +              info_mask |= (DM_LAST_ACCESS_DATE);
 +              ncp_date_unix2dos(attr->ia_atime.tv_sec,
 +                                &dummy, &info.lastAccessDate);
 +      }
 +      if (info_mask != 0) {
 +              result = ncp_modify_file_or_subdir_dos_info(NCP_SERVER(inode),
 +                                    inode, info_mask, &info);
 +              if (result != 0) {
 +                      if (info_mask == (DM_CREATE_TIME | DM_CREATE_DATE)) {
 +                              /* NetWare seems not to allow this. I
 +                                 do not know why. So, just tell the
 +                                 user everything went fine. This is
 +                                 a terrible hack, but I do not know
 +                                 how to do this correctly. */
 +                              result = 0;
 +                      } else
 +                              goto out;
 +              }
 +#ifdef CONFIG_NCPFS_STRONG            
 +              if ((!result) && (info_mask & DM_ATTRIBUTES))
 +                      NCP_FINFO(inode)->nwattr = info.attributes;
 +#endif
 +      }
 +      if (result)
 +              goto out;
 +
 +      setattr_copy(inode, attr);
 +      mark_inode_dirty(inode);
 +
 +out:
 +      if (result > 0)
 +              result = -EACCES;
 +      return result;
 +}
 +
 +static struct dentry *ncp_mount(struct file_system_type *fs_type,
 +      int flags, const char *dev_name, void *data)
 +{
 +      return mount_nodev(fs_type, flags, data, ncp_fill_super);
 +}
 +
 +static struct file_system_type ncp_fs_type = {
 +      .owner          = THIS_MODULE,
 +      .name           = "ncpfs",
 +      .mount          = ncp_mount,
 +      .kill_sb        = kill_anon_super,
 +      .fs_flags       = FS_BINARY_MOUNTDATA,
 +};
 +MODULE_ALIAS_FS("ncpfs");
 +
 +static int __init init_ncp_fs(void)
 +{
 +      int err;
 +      ncp_dbg(1, "called\n");
 +
 +      err = init_inodecache();
 +      if (err)
 +              goto out1;
 +      err = register_filesystem(&ncp_fs_type);
 +      if (err)
 +              goto out;
 +      return 0;
 +out:
 +      destroy_inodecache();
 +out1:
 +      return err;
 +}
 +
 +static void __exit exit_ncp_fs(void)
 +{
 +      ncp_dbg(1, "called\n");
 +      unregister_filesystem(&ncp_fs_type);
 +      destroy_inodecache();
 +}
 +
 +module_init(init_ncp_fs)
 +module_exit(exit_ncp_fs)
 +MODULE_LICENSE("GPL");