1 files changed, 128 insertions, 77 deletions

diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index cbefe1f1f9fe..f0a81e631ae6 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -107,7 +107,7 @@ static inline int cifs_open_inode_helper(struct inode *inode, struct file *file,
 
 	/* want handles we can use to read with first
 	   in the list so we do not have to walk the
-	   list to search for one in prepare_write */
+	   list to search for one in write_begin */
 	if ((file->f_flags & O_ACCMODE) == O_WRONLY) {
 		list_add_tail(&pCifsFile->flist,
 			      &pCifsInode->openFileList);
@@ -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->tidStatus != CifsNeedReconnect) {
+			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"));
@@ -915,7 +922,7 @@ ssize_t cifs_user_write(struct file *file, const char __user *write_data,
 }
 
 static ssize_t cifs_write(struct file *file, const char *write_data,
-	size_t write_size, loff_t *poffset)
+			  size_t write_size, loff_t *poffset)
 {
 	int rc = 0;
 	unsigned int bytes_written = 0;
@@ -1065,6 +1072,7 @@ struct cifsFileInfo *find_readable_file(struct cifsInodeInfo *cifs_inode)
 struct cifsFileInfo *find_writable_file(struct cifsInodeInfo *cifs_inode)
 {
 	struct cifsFileInfo *open_file;
+	bool any_available = false;
 	int rc;
 
 	/* Having a null inode here (because mapping->host was set to zero by
@@ -1080,8 +1088,10 @@ struct cifsFileInfo *find_writable_file(struct cifsInodeInfo *cifs_inode)
 	read_lock(&GlobalSMBSeslock);
 refind_writable:
 	list_for_each_entry(open_file, &cifs_inode->openFileList, flist) {
-		if (open_file->closePend)
+		if (open_file->closePend ||
+		    (!any_available && open_file->pid != current->tgid))
 			continue;
+
 		if (open_file->pfile &&
 		    ((open_file->pfile->f_flags & O_RDWR) ||
 		     (open_file->pfile->f_flags & O_WRONLY))) {
@@ -1131,6 +1141,11 @@ refind_writable:
 			   of the loop here. */
 		}
 	}
+	/* couldn't find useable FH with same pid, try any available */
+	if (!any_available) {
+		any_available = true;
+		goto refind_writable;
+	}
 	read_unlock(&GlobalSMBSeslock);
 	return NULL;
 }
@@ -1396,7 +1411,10 @@ retry:
 			if ((wbc->nr_to_write -= n_iov) <= 0)
 				done = 1;
 			index = next;
-		}
+		} else
+			/* Need to re-find the pages we skipped */
+			index = pvec.pages[0]->index + 1;
+
 		pagevec_release(&pvec);
 	}
 	if (!scanned && !done) {
@@ -1447,49 +1465,56 @@ static int cifs_writepage(struct page *page, struct writeback_control *wbc)
 	return rc;
 }
 
-static int cifs_commit_write(struct file *file, struct page *page,
-	unsigned offset, unsigned to)
+static int cifs_write_end(struct file *file, struct address_space *mapping,
+			loff_t pos, unsigned len, unsigned copied,
+			struct page *page, void *fsdata)
 {
-	int xid;
-	int rc = 0;
-	struct inode *inode = page->mapping->host;
-	loff_t position = ((loff_t)page->index << PAGE_CACHE_SHIFT) + to;
-	char *page_data;
+	int rc;
+	struct inode *inode = mapping->host;
 
-	xid = GetXid();
-	cFYI(1, ("commit write for page %p up to position %lld for %d",
-		 page, position, to));
-	spin_lock(&inode->i_lock);
-	if (position > inode->i_size)
-		i_size_write(inode, position);
+	cFYI(1, ("write_end for page %p from pos %lld with %d bytes",
+		 page, pos, copied));
+
+	if (PageChecked(page)) {
+		if (copied == len)
+			SetPageUptodate(page);
+		ClearPageChecked(page);
+	} else if (!PageUptodate(page) && copied == PAGE_CACHE_SIZE)
+		SetPageUptodate(page);
 
-	spin_unlock(&inode->i_lock);
 	if (!PageUptodate(page)) {
-		position =  ((loff_t)page->index << PAGE_CACHE_SHIFT) + offset;
-		/* can not rely on (or let) writepage write this data */
-		if (to < offset) {
-			cFYI(1, ("Illegal offsets, can not copy from %d to %d",
-				offset, to));
-			FreeXid(xid);
-			return rc;
-		}
+		char *page_data;
+		unsigned offset = pos & (PAGE_CACHE_SIZE - 1);
+		int xid;
+
+		xid = GetXid();
 		/* this is probably better than directly calling
 		   partialpage_write since in this function the file handle is
 		   known which we might as well	leverage */
 		/* BB check if anything else missing out of ppw
 		   such as updating last write time */
 		page_data = kmap(page);
-		rc = cifs_write(file, page_data + offset, to-offset,
-				&position);
-		if (rc > 0)
-			rc = 0;
-		/* else if (rc < 0) should we set writebehind rc? */
+		rc = cifs_write(file, page_data + offset, copied, &pos);
+		/* if (rc < 0) should we set writebehind rc? */
 		kunmap(page);
+
+		FreeXid(xid);
 	} else {
+		rc = copied;
+		pos += copied;
 		set_page_dirty(page);
 	}
 
-	FreeXid(xid);
+	if (rc > 0) {
+		spin_lock(&inode->i_lock);
+		if (pos > inode->i_size)
+			i_size_write(inode, pos);
+		spin_unlock(&inode->i_lock);
+	}
+
+	unlock_page(page);
+	page_cache_release(page);
+
 	return rc;
 }
 
@@ -1780,7 +1805,7 @@ static void cifs_copy_cache_pages(struct address_space *mapping,
 		SetPageUptodate(page);
 		unlock_page(page);
 		if (!pagevec_add(plru_pvec, page))
-			__pagevec_lru_add(plru_pvec);
+			__pagevec_lru_add_file(plru_pvec);
 		data += PAGE_CACHE_SIZE;
 	}
 	return;
@@ -1813,7 +1838,7 @@ static int cifs_readpages(struct file *file, struct address_space *mapping,
 	pTcon = cifs_sb->tcon;
 
 	pagevec_init(&lru_pvec, 0);
-		cFYI(DBG2, ("rpages: num pages %d", num_pages));
+	cFYI(DBG2, ("rpages: num pages %d", num_pages));
 	for (i = 0; i < num_pages; ) {
 		unsigned contig_pages;
 		struct page *tmp_page;
@@ -1914,7 +1939,7 @@ static int cifs_readpages(struct file *file, struct address_space *mapping,
 		bytes_read = 0;
 	}
 
-	pagevec_lru_add(&lru_pvec);
+	pagevec_lru_add_file(&lru_pvec);
 
 /* need to free smb_read_data buf before exit */
 	if (smb_read_data) {
@@ -2035,50 +2060,76 @@ bool is_size_safe_to_change(struct cifsInodeInfo *cifsInode, __u64 end_of_file)
 		return true;
 }
 
-static int cifs_prepare_write(struct file *file, struct page *page,
-	unsigned from, unsigned to)
+static int cifs_write_begin(struct file *file, struct address_space *mapping,
+			loff_t pos, unsigned len, unsigned flags,
+			struct page **pagep, void **fsdata)
 {
-	int rc = 0;
+	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;
-	loff_t offset;
+	struct page *page;
+	int rc = 0;
 
-	cFYI(1, ("prepare write for page %p from %d to %d", page, from, to));
-	if (PageUptodate(page))
-		return 0;
+	cFYI(1, ("write_begin from %lld len %d", (long long)pos, len));
 
-	/* If we are writing a full page it will be up to date,
-	   no need to read from the server */
-	if ((to == PAGE_CACHE_SIZE) && (from == 0)) {
-		SetPageUptodate(page);
-		return 0;
+	page = __grab_cache_page(mapping, index);
+	if (!page) {
+		rc = -ENOMEM;
+		goto out;
 	}
 
-	offset = (loff_t)page->index << PAGE_CACHE_SHIFT;
-	i_size = i_size_read(page->mapping->host);
+	if (PageUptodate(page))
+		goto out;
+
+	/*
+	 * 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;
+
+	/*
+	 * 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;
+		}
+	}
 
-	if ((offset >= i_size) ||
-	    ((from == 0) && (offset + to) >= i_size)) {
+	if ((file->f_flags & O_ACCMODE) != O_WRONLY) {
 		/*
-		 * We don't need to read data beyond the end of the file.
-		 * zero it, and set the page uptodate
+		 * 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.
 		 */
-		simple_prepare_write(file, page, from, to);
-		SetPageUptodate(page);
-	} else if ((file->f_flags & O_ACCMODE) != O_WRONLY) {
-		/* might as well read a page, it is fast enough */
-		rc = cifs_readpage_worker(file, page, &offset);
+		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 commit_write so is fine */
+		   this will be written out by write_end so is fine */
 	}
-
-	/* we do not need to pass errors back
-	   e.g. if we do not have read access to the file
-	   because cifs_commit_write will do the right thing.  -- shaggy */
-
-	return 0;
+out:
+	*pagep = page;
+	return rc;
 }
 
 const struct address_space_operations cifs_addr_ops = {
@@ -2086,8 +2137,8 @@ const struct address_space_operations cifs_addr_ops = {
 	.readpages = cifs_readpages,
 	.writepage = cifs_writepage,
 	.writepages = cifs_writepages,
-	.prepare_write = cifs_prepare_write,
-	.commit_write = cifs_commit_write,
+	.write_begin = cifs_write_begin,
+	.write_end = cifs_write_end,
 	.set_page_dirty = __set_page_dirty_nobuffers,
 	/* .sync_page = cifs_sync_page, */
 	/* .direct_IO = */
@@ -2102,8 +2153,8 @@ const struct address_space_operations cifs_addr_ops_smallbuf = {
 	.readpage = cifs_readpage,
 	.writepage = cifs_writepage,
 	.writepages = cifs_writepages,
-	.prepare_write = cifs_prepare_write,
-	.commit_write = cifs_commit_write,
+	.write_begin = cifs_write_begin,
+	.write_end = cifs_write_end,
 	.set_page_dirty = __set_page_dirty_nobuffers,
 	/* .sync_page = cifs_sync_page, */
 	/* .direct_IO = */