6 files changed, 92 insertions, 32 deletions

diff --git a/fs/cifs/CHANGES b/fs/cifs/CHANGES
index 8855331b2fba..e078b7aea143 100644
--- a/fs/cifs/CHANGES
+++ b/fs/cifs/CHANGES
@@ -8,7 +8,11 @@ handling fcntl(F_SETLEASE).  Convert cifs to using blocking tcp
 sends, and also let tcp autotune the socket send and receive buffers.
 This reduces the number of EAGAIN errors returned by TCP/IP in
 high stress workloads (and the number of retries on socket writes
-when sending large SMBWriteX requests).
+when sending large SMBWriteX requests).  Fix case in which a portion of
+data can in some cases not get written to the file on the server before the
+file is closed.  Fix DFS parsing to properly handle path consumed field,
+and to handle certain codepage conversions better.  Fix mount and
+umount race that can cause oops in mount or umount or reconnect.
 
 Version 1.54
 ------------
diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h
index f1ae1f57c30d..c57c0565547f 100644
--- a/fs/cifs/cifsglob.h
+++ b/fs/cifs/cifsglob.h
@@ -606,7 +606,15 @@ GLOBAL_EXTERN struct list_head		cifs_tcp_ses_list;
  * changes to the tcon->tidStatus should be done while holding this lock.
  */
 GLOBAL_EXTERN rwlock_t          cifs_tcp_ses_lock;
-GLOBAL_EXTERN rwlock_t GlobalSMBSeslock;  /* protects list inserts on 3 above */
+
+/*
+ * This lock protects the cifs_file->llist and cifs_file->flist
+ * list operations, and updates to some flags (cifs_file->invalidHandle)
+ * It will be moved to either use the tcon->stat_lock or equivalent later.
+ * If cifs_tcp_ses_lock and the lock below are both needed to be held, then
+ * the cifs_tcp_ses_lock must be grabbed first and released last.
+ */
+GLOBAL_EXTERN rwlock_t GlobalSMBSeslock;
 
 GLOBAL_EXTERN struct list_head GlobalOplock_Q;
 
diff --git a/fs/cifs/cifssmb.c b/fs/cifs/cifssmb.c
index bdda46dd435a..2af8626ced43 100644
--- a/fs/cifs/cifssmb.c
+++ b/fs/cifs/cifssmb.c
@@ -295,7 +295,7 @@ smb_init(int smb_command, int wct, struct cifsTconInfo *tcon,
            check for tcp and smb session status done differently
            for those three - in the calling routine */
         if (tcon) {
-                if (tcon->need_reconnect) {
+                if (tcon->tidStatus == CifsExiting) {
                         /* only tree disconnect, open, and write,
                           (and ulogoff which does not have tcon)
                           are allowed as we start force umount */
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index 6449e1aae621..f0a81e631ae6 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -488,12 +488,13 @@ int cifs_close(struct inode *inode, struct file *file)
         pTcon = cifs_sb->tcon;
         if (pSMBFile) {
                 struct cifsLockInfo *li, *tmp;
-
+                write_lock(&GlobalSMBSeslock);
                 pSMBFile->closePend = true;
                 if (pTcon) {
                         /* no sense reconnecting to close a file that is
                            already closed */
                         if (!pTcon->need_reconnect) {
+                                write_unlock(&GlobalSMBSeslock);
                                 timeout = 2;
                                 while ((atomic_read(&pSMBFile->wrtPending) != 0)
                                         && (timeout <= 2048)) {
@@ -510,12 +511,15 @@ int cifs_close(struct inode *inode, struct file *file)
                                         timeout *= 4;
                                 }
                                 if (atomic_read(&pSMBFile->wrtPending))
-                                        cERROR(1,
-                                                ("close with pending writes"));
-                                rc = CIFSSMBClose(xid, pTcon,
+                                        cERROR(1, ("close with pending write"));
+                                if (!pTcon->need_reconnect &&
+                                    !pSMBFile->invalidHandle)
+                                        rc = CIFSSMBClose(xid, pTcon,
                                                   pSMBFile->netfid);
-                        }
-                }
+                        } else
+                                write_unlock(&GlobalSMBSeslock);
+                } else
+                        write_unlock(&GlobalSMBSeslock);
 
                 /* Delete any outstanding lock records.
                    We'll lose them when the file is closed anyway. */
@@ -587,15 +591,18 @@ int cifs_closedir(struct inode *inode, struct file *file)
                 pTcon = cifs_sb->tcon;
 
                 cFYI(1, ("Freeing private data in close dir"));
+                write_lock(&GlobalSMBSeslock);
                 if (!pCFileStruct->srch_inf.endOfSearch &&
                     !pCFileStruct->invalidHandle) {
                         pCFileStruct->invalidHandle = true;
+                        write_unlock(&GlobalSMBSeslock);
                         rc = CIFSFindClose(xid, pTcon, pCFileStruct->netfid);
                         cFYI(1, ("Closing uncompleted readdir with rc %d",
                                  rc));
                         /* not much we can do if it fails anyway, ignore rc */
                         rc = 0;
-                }
+                } else
+                        write_unlock(&GlobalSMBSeslock);
                 ptmp = pCFileStruct->srch_inf.ntwrk_buf_start;
                 if (ptmp) {
                         cFYI(1, ("closedir free smb buf in srch struct"));
@@ -1468,7 +1475,11 @@ static int cifs_write_end(struct file *file, struct address_space *mapping,
         cFYI(1, ("write_end for page %p from pos %lld with %d bytes",
                  page, pos, copied));
 
-        if (!PageUptodate(page) && copied == PAGE_CACHE_SIZE)
+        if (PageChecked(page)) {
+                if (copied == len)
+                        SetPageUptodate(page);
+                ClearPageChecked(page);
+        } else if (!PageUptodate(page) && copied == PAGE_CACHE_SIZE)
                 SetPageUptodate(page);
 
         if (!PageUptodate(page)) {
@@ -2055,39 +2066,70 @@ static int cifs_write_begin(struct file *file, struct address_space *mapping,
 {
         pgoff_t index = pos >> PAGE_CACHE_SHIFT;
         loff_t offset = pos & (PAGE_CACHE_SIZE - 1);
+        loff_t page_start = pos & PAGE_MASK;
+        loff_t i_size;
+        struct page *page;
+        int rc = 0;
 
         cFYI(1, ("write_begin from %lld len %d", (long long)pos, len));
 
-        *pagep = __grab_cache_page(mapping, index);
-        if (!*pagep)
-                return -ENOMEM;
-
-        if (PageUptodate(*pagep))
-                return 0;
+        page = __grab_cache_page(mapping, index);
+        if (!page) {
+                rc = -ENOMEM;
+                goto out;
+        }
 
-        /* If we are writing a full page it will be up to date,
-           no need to read from the server */
-        if (len == PAGE_CACHE_SIZE && flags & AOP_FLAG_UNINTERRUPTIBLE)
-                return 0;
+        if (PageUptodate(page))
+                goto out;
 
-        if ((file->f_flags & O_ACCMODE) != O_WRONLY) {
-                int rc;
+        /*
+         * If we write a full page it will be up to date, no need to read from
+         * the server. If the write is short, we'll end up doing a sync write
+         * instead.
+         */
+        if (len == PAGE_CACHE_SIZE)
+                goto out;
 
-                /* might as well read a page, it is fast enough */
-                rc = cifs_readpage_worker(file, *pagep, &offset);
+        /*
+         * optimize away the read when we have an oplock, and we're not
+         * expecting to use any of the data we'd be reading in. That
+         * is, when the page lies beyond the EOF, or straddles the EOF
+         * and the write will cover all of the existing data.
+         */
+        if (CIFS_I(mapping->host)->clientCanCacheRead) {
+                i_size = i_size_read(mapping->host);
+                if (page_start >= i_size ||
+                    (offset == 0 && (pos + len) >= i_size)) {
+                        zero_user_segments(page, 0, offset,
+                                           offset + len,
+                                           PAGE_CACHE_SIZE);
+                        /*
+                         * PageChecked means that the parts of the page
+                         * to which we're not writing are considered up
+                         * to date. Once the data is copied to the
+                         * page, it can be set uptodate.
+                         */
+                        SetPageChecked(page);
+                        goto out;
+                }
+        }
 
-                /* we do not need to pass errors back
-                   e.g. if we do not have read access to the file
-                   because cifs_write_end will attempt synchronous writes
-                   -- shaggy */
+        if ((file->f_flags & O_ACCMODE) != O_WRONLY) {
+                /*
+                 * might as well read a page, it is fast enough. If we get
+                 * an error, we don't need to return it. cifs_write_end will
+                 * do a sync write instead since PG_uptodate isn't set.
+                 */
+                cifs_readpage_worker(file, page, &page_start);
         } else {
                 /* we could try using another file handle if there is one -
                    but how would we lock it to prevent close of that handle
                    racing with this read? In any case
                    this will be written out by write_end so is fine */
         }
-
-        return 0;
+out:
+        *pagep = page;
+        return rc;
 }
 
 const struct address_space_operations cifs_addr_ops = {
diff --git a/fs/cifs/misc.c b/fs/cifs/misc.c
index f108040ca1bc..8a82d076450b 100644
--- a/fs/cifs/misc.c
+++ b/fs/cifs/misc.c
@@ -555,12 +555,14 @@ is_valid_oplock_break(struct smb_hdr *buf, struct TCP_Server_Info *srv)
                                 continue;
 
                         cifs_stats_inc(&tcon->num_oplock_brks);
+                        write_lock(&GlobalSMBSeslock);
                         list_for_each(tmp2, &tcon->openFileList) {
                                 netfile = list_entry(tmp2, struct cifsFileInfo,
                                                      tlist);
                                 if (pSMB->Fid != netfile->netfid)
                                         continue;
 
+                                write_unlock(&GlobalSMBSeslock);
                                 read_unlock(&cifs_tcp_ses_lock);
                                 cFYI(1, ("file id match, oplock break"));
                                 pCifsInode = CIFS_I(netfile->pInode);
@@ -576,6 +578,7 @@ is_valid_oplock_break(struct smb_hdr *buf, struct TCP_Server_Info *srv)
 
                                 return true;
                         }
+                        write_unlock(&GlobalSMBSeslock);
                         read_unlock(&cifs_tcp_ses_lock);
                         cFYI(1, ("No matching file for oplock break"));
                         return true;
diff --git a/fs/cifs/readdir.c b/fs/cifs/readdir.c
index 58d57299f2a0..9f51f9bf0292 100644
--- a/fs/cifs/readdir.c
+++ b/fs/cifs/readdir.c
@@ -741,11 +741,14 @@ static int find_cifs_entry(const int xid, struct cifsTconInfo *pTcon,
            (index_to_find < first_entry_in_buffer)) {
                 /* close and restart search */
                 cFYI(1, ("search backing up - close and restart search"));
+                write_lock(&GlobalSMBSeslock);
                 if (!cifsFile->srch_inf.endOfSearch &&
                     !cifsFile->invalidHandle) {
                         cifsFile->invalidHandle = true;
+                        write_unlock(&GlobalSMBSeslock);
                         CIFSFindClose(xid, pTcon, cifsFile->netfid);
-                }
+                } else
+                        write_unlock(&GlobalSMBSeslock);
                 if (cifsFile->srch_inf.ntwrk_buf_start) {
                         cFYI(1, ("freeing SMB ff cache buf on search rewind"));
                         if (cifsFile->srch_inf.smallBuf)