Merge branch 'cifs-3.2' of git://git.samba.org/jlayton/linux into temp-3.2-jeff

fs/cifs/cifsfs.c

@@ -74,7 +74,7 @@ module_param(cifs_min_small, int, 0);
 MODULE_PARM_DESC(cifs_min_small, "Small network buffers in pool. Default: 30 "
 				 "Range: 2 to 256");
 unsigned int cifs_max_pending = CIFS_MAX_REQ;
-module_param(cifs_max_pending, int, 0);
+module_param(cifs_max_pending, int, 0444);
 MODULE_PARM_DESC(cifs_max_pending, "Simultaneous requests to server. "
 				   "Default: 50 Range: 2 to 256");
 unsigned short echo_retries = 5;

fs/cifs/cifsglob.h

@@ -291,7 +291,13 @@ struct TCP_Server_Info {
 	bool	sec_kerberosu2u;	/* supports U2U Kerberos */
 	bool	sec_kerberos;		/* supports plain Kerberos */
 	bool	sec_mskerberos;		/* supports legacy MS Kerberos */
+	bool	large_buf;		/* is current buffer large? */
 	struct delayed_work	echo; /* echo ping workqueue job */
+	struct kvec *iov;	/* reusable kvec array for receives */
+	unsigned int nr_iov;	/* number of kvecs in array */
+	char	*smallbuf;	/* pointer to current "small" buffer */
+	char	*bigbuf;	/* pointer to current "big" buffer */
+	unsigned int total_read; /* total amount of data read in this pass */
 #ifdef CONFIG_CIFS_FSCACHE
 	struct fscache_cookie   *fscache; /* client index cache cookie */
 #endif
@@ -650,8 +656,24 @@ static inline void cifs_stats_bytes_read(struct cifs_tcon *tcon,
 struct mid_q_entry;
 
 /*
- * This is the prototype for the mid callback function. When creating one,
- * take special care to avoid deadlocks. Things to bear in mind:
+ * This is the prototype for the mid receive function. This function is for
+ * receiving the rest of the SMB frame, starting with the WordCount (which is
+ * just after the MID in struct smb_hdr). Note:
+ *
+ * - This will be called by cifsd, with no locks held.
+ * - The mid will still be on the pending_mid_q.
+ * - mid->resp_buf will point to the current buffer.
+ *
+ * Returns zero on a successful receive, or an error. The receive state in
+ * the TCP_Server_Info will also be updated.
+ */
+typedef int (mid_receive_t)(struct TCP_Server_Info *server,
+			    struct mid_q_entry *mid);
+
+/*
+ * This is the prototype for the mid callback function. This is called once the
+ * mid has been received off of the socket. When creating one, take special
+ * care to avoid deadlocks. Things to bear in mind:
  *
  * - it will be called by cifsd, with no locks held
  * - the mid will be removed from any lists
@@ -669,9 +691,10 @@ struct mid_q_entry {
 	unsigned long when_sent; /* time when smb send finished */
 	unsigned long when_received; /* when demux complete (taken off wire) */
 #endif
+	mid_receive_t *receive; /* call receive callback */
 	mid_callback_t *callback; /* call completion callback */
 	void *callback_data;	  /* general purpose pointer for callback */
-	struct smb_hdr *resp_buf;	/* response buffer */
+	struct smb_hdr *resp_buf;	/* pointer to received SMB header */
 	int midState;	/* wish this were enum but can not pass to wait_event */
 	__u8 command;	/* smb command code */
 	bool largeBuf:1;	/* if valid response, is pointer to large buf */

fs/cifs/cifspdu.h

@@ -1089,9 +1089,7 @@ typedef struct smb_com_read_rsp {
 	__le16 DataLengthHigh;
 	__u64 Reserved2;
 	__u16 ByteCount;
-	__u8 Pad;		/* BB check for whether padded to DWORD
-				   boundary and optimum performance here */
-	char Data[1];
+	/* read response data immediately follows */
 } __attribute__((packed)) READ_RSP;
 
 typedef struct locking_andx_range {

fs/cifs/cifsproto.h

@@ -69,8 +69,9 @@ extern struct mid_q_entry *AllocMidQEntry(const struct smb_hdr *smb_buffer,
 					struct TCP_Server_Info *server);
 extern void DeleteMidQEntry(struct mid_q_entry *midEntry);
 extern int cifs_call_async(struct TCP_Server_Info *server, struct kvec *iov,
-			   unsigned int nvec, mid_callback_t *callback,
-			   void *cbdata, bool ignore_pend);
+			   unsigned int nvec, mid_receive_t *receive,
+			   mid_callback_t *callback, void *cbdata,
+			   bool ignore_pend);
 extern int SendReceive(const unsigned int /* xid */ , struct cifs_ses *,
 			struct smb_hdr * /* input */ ,
 			struct smb_hdr * /* out */ ,
@@ -153,6 +154,12 @@ extern struct cifs_ntsd *get_cifs_acl(struct cifs_sb_info *, struct inode *,
 extern int set_cifs_acl(struct cifs_ntsd *, __u32, struct inode *,
 				const char *, int);
 
+extern void dequeue_mid(struct mid_q_entry *mid, bool malformed);
+extern int cifs_read_from_socket(struct TCP_Server_Info *server, char *buf,
+		     unsigned int to_read);
+extern int cifs_readv_from_socket(struct TCP_Server_Info *server,
+		struct kvec *iov_orig, unsigned int nr_segs,
+		unsigned int to_read);
 extern void cifs_setup_cifs_sb(struct smb_vol *pvolume_info,
 			       struct cifs_sb_info *cifs_sb);
 extern int cifs_match_super(struct super_block *, void *);
@@ -442,6 +449,24 @@ extern int E_md4hash(const unsigned char *passwd, unsigned char *p16);
 extern int SMBencrypt(unsigned char *passwd, const unsigned char *c8,
 			unsigned char *p24);
 
+/* asynchronous read support */
+struct cifs_readdata {
+	struct cifsFileInfo		*cfile;
+	struct address_space		*mapping;
+	__u64				offset;
+	unsigned int			bytes;
+	pid_t				pid;
+	int				result;
+	struct list_head		pages;
+	struct work_struct		work;
+	unsigned int			nr_iov;
+	struct kvec			iov[1];
+};
+
+struct cifs_readdata *cifs_readdata_alloc(unsigned int nr_pages);
+void cifs_readdata_free(struct cifs_readdata *rdata);
+int cifs_async_readv(struct cifs_readdata *rdata);
+
 /* asynchronous write support */
 struct cifs_writedata {
 	struct kref			refcount;

fs/cifs/cifssmb.c

@@ -33,6 +33,8 @@
 #include <linux/slab.h>
 #include <linux/posix_acl_xattr.h>
 #include <linux/pagemap.h>
+#include <linux/swap.h>
+#include <linux/task_io_accounting_ops.h>
 #include <asm/uaccess.h>
 #include "cifspdu.h"
 #include "cifsglob.h"
@@ -40,6 +42,7 @@
 #include "cifsproto.h"
 #include "cifs_unicode.h"
 #include "cifs_debug.h"
+#include "fscache.h"
 
 #ifdef CONFIG_CIFS_POSIX
 static struct {
@@ -83,6 +86,9 @@ static struct {
 #endif /* CONFIG_CIFS_WEAK_PW_HASH */
 #endif /* CIFS_POSIX */
 
+/* Forward declarations */
+static void cifs_readv_complete(struct work_struct *work);
+
 /* Mark as invalid, all open files on tree connections since they
    were closed when session to server was lost */
 static void mark_open_files_invalid(struct cifs_tcon *pTcon)
@@ -737,7 +743,8 @@ CIFSSMBEcho(struct TCP_Server_Info *server)
 	iov.iov_base = smb;
 	iov.iov_len = be32_to_cpu(smb->hdr.smb_buf_length) + 4;
 
-	rc = cifs_call_async(server, &iov, 1, cifs_echo_callback, server, true);
+	rc = cifs_call_async(server, &iov, 1, NULL, cifs_echo_callback,
+			     server, true);
 	if (rc)
 		cFYI(1, "Echo request failed: %d", rc);
 
@@ -1374,6 +1381,359 @@ CIFSSMBOpen(const int xid, struct cifs_tcon *tcon,
 	return rc;
 }
 
+struct cifs_readdata *
+cifs_readdata_alloc(unsigned int nr_pages)
+{
+	struct cifs_readdata *rdata;
+
+	/* readdata + 1 kvec for each page */
+	rdata = kzalloc(sizeof(*rdata) +
+			sizeof(struct kvec) * nr_pages, GFP_KERNEL);
+	if (rdata != NULL) {
+		INIT_WORK(&rdata->work, cifs_readv_complete);
+		INIT_LIST_HEAD(&rdata->pages);
+	}
+	return rdata;
+}
+
+void
+cifs_readdata_free(struct cifs_readdata *rdata)
+{
+	cifsFileInfo_put(rdata->cfile);
+	kfree(rdata);
+}
+
+/*
+ * Discard any remaining data in the current SMB. To do this, we borrow the
+ * current bigbuf.
+ */
+static int
+cifs_readv_discard(struct TCP_Server_Info *server, struct mid_q_entry *mid)
+{
+	READ_RSP *rsp = (READ_RSP *)server->smallbuf;
+	unsigned int rfclen = be32_to_cpu(rsp->hdr.smb_buf_length);
+	int remaining = rfclen + 4 - server->total_read;
+	struct cifs_readdata *rdata = mid->callback_data;
+
+	while (remaining > 0) {
+		int length;
+
+		length = cifs_read_from_socket(server, server->bigbuf,
+				min_t(unsigned int, remaining,
+					CIFSMaxBufSize + MAX_CIFS_HDR_SIZE));
+		if (length < 0)
+			return length;
+		server->total_read += length;
+		remaining -= length;
+	}
+
+	dequeue_mid(mid, rdata->result);
+	return 0;
+}
+
+static int
+cifs_readv_receive(struct TCP_Server_Info *server, struct mid_q_entry *mid)
+{
+	int length, len;
+	unsigned int data_offset, remaining, data_len;
+	struct cifs_readdata *rdata = mid->callback_data;
+	READ_RSP *rsp = (READ_RSP *)server->smallbuf;
+	unsigned int rfclen = be32_to_cpu(rsp->hdr.smb_buf_length) + 4;
+	u64 eof;
+	pgoff_t eof_index;
+	struct page *page, *tpage;
+
+	cFYI(1, "%s: mid=%u offset=%llu bytes=%u", __func__,
+		mid->mid, rdata->offset, rdata->bytes);
+
+	/*
+	 * read the rest of READ_RSP header (sans Data array), or whatever we
+	 * can if there's not enough data. At this point, we've read down to
+	 * the Mid.
+	 */
+	len = min_t(unsigned int, rfclen, sizeof(*rsp)) -
+			sizeof(struct smb_hdr) + 1;
+
+	rdata->iov[0].iov_base = server->smallbuf + sizeof(struct smb_hdr) - 1;
+	rdata->iov[0].iov_len = len;
+
+	length = cifs_readv_from_socket(server, rdata->iov, 1, len);
+	if (length < 0)
+		return length;
+	server->total_read += length;
+
+	/* Was the SMB read successful? */
+	rdata->result = map_smb_to_linux_error(&rsp->hdr, false);
+	if (rdata->result != 0) {
+		cFYI(1, "%s: server returned error %d", __func__,
+			rdata->result);
+		return cifs_readv_discard(server, mid);
+	}
+
+	/* Is there enough to get to the rest of the READ_RSP header? */
+	if (server->total_read < sizeof(READ_RSP)) {
+		cFYI(1, "%s: server returned short header. got=%u expected=%zu",
+			__func__, server->total_read, sizeof(READ_RSP));
+		rdata->result = -EIO;
+		return cifs_readv_discard(server, mid);
+	}
+
+	data_offset = le16_to_cpu(rsp->DataOffset) + 4;
+	if (data_offset < server->total_read) {
+		/*
+		 * win2k8 sometimes sends an offset of 0 when the read
+		 * is beyond the EOF. Treat it as if the data starts just after
+		 * the header.
+		 */
+		cFYI(1, "%s: data offset (%u) inside read response header",
+			__func__, data_offset);
+		data_offset = server->total_read;
+	} else if (data_offset > MAX_CIFS_SMALL_BUFFER_SIZE) {
+		/* data_offset is beyond the end of smallbuf */
+		cFYI(1, "%s: data offset (%u) beyond end of smallbuf",
+			__func__, data_offset);
+		rdata->result = -EIO;
+		return cifs_readv_discard(server, mid);
+	}
+
+	cFYI(1, "%s: total_read=%u data_offset=%u", __func__,
+		server->total_read, data_offset);
+
+	len = data_offset - server->total_read;
+	if (len > 0) {
+		/* read any junk before data into the rest of smallbuf */
+		rdata->iov[0].iov_base = server->smallbuf + server->total_read;
+		rdata->iov[0].iov_len = len;
+		length = cifs_readv_from_socket(server, rdata->iov, 1, len);
+		if (length < 0)
+			return length;
+		server->total_read += length;
+	}
+
+	/* set up first iov for signature check */
+	rdata->iov[0].iov_base = server->smallbuf;
+	rdata->iov[0].iov_len = server->total_read;
+	cFYI(1, "0: iov_base=%p iov_len=%zu",
+		rdata->iov[0].iov_base, rdata->iov[0].iov_len);
+
+	/* how much data is in the response? */
+	data_len = le16_to_cpu(rsp->DataLengthHigh) << 16;
+	data_len += le16_to_cpu(rsp->DataLength);
+	if (data_offset + data_len > rfclen) {
+		/* data_len is corrupt -- discard frame */
+		rdata->result = -EIO;
+		return cifs_readv_discard(server, mid);
+	}
+
+	/* marshal up the page array */
+	len = 0;
+	remaining = data_len;
+	rdata->nr_iov = 1;
+
+	/* determine the eof that the server (probably) has */
+	eof = CIFS_I(rdata->mapping->host)->server_eof;
+	eof_index = eof ? (eof - 1) >> PAGE_CACHE_SHIFT : 0;
+	cFYI(1, "eof=%llu eof_index=%lu", eof, eof_index);
+
+	list_for_each_entry_safe(page, tpage, &rdata->pages, lru) {
+		if (remaining >= PAGE_CACHE_SIZE) {
+			/* enough data to fill the page */
+			rdata->iov[rdata->nr_iov].iov_base = kmap(page);
+			rdata->iov[rdata->nr_iov].iov_len = PAGE_CACHE_SIZE;
+			cFYI(1, "%u: idx=%lu iov_base=%p iov_len=%zu",
+				rdata->nr_iov, page->index,
+				rdata->iov[rdata->nr_iov].iov_base,
+				rdata->iov[rdata->nr_iov].iov_len);
+			++rdata->nr_iov;
+			len += PAGE_CACHE_SIZE;
+			remaining -= PAGE_CACHE_SIZE;
+		} else if (remaining > 0) {
+			/* enough for partial page, fill and zero the rest */
+			rdata->iov[rdata->nr_iov].iov_base = kmap(page);
+			rdata->iov[rdata->nr_iov].iov_len = remaining;
+			cFYI(1, "%u: idx=%lu iov_base=%p iov_len=%zu",
+				rdata->nr_iov, page->index,
+				rdata->iov[rdata->nr_iov].iov_base,
+				rdata->iov[rdata->nr_iov].iov_len);
+			memset(rdata->iov[rdata->nr_iov].iov_base + remaining,
+				'\0', PAGE_CACHE_SIZE - remaining);
+			++rdata->nr_iov;
+			len += remaining;
+			remaining = 0;
+		} else if (page->index > eof_index) {
+			/*
+			 * The VFS will not try to do readahead past the
+			 * i_size, but it's possible that we have outstanding
+			 * writes with gaps in the middle and the i_size hasn't
+			 * caught up yet. Populate those with zeroed out pages
+			 * to prevent the VFS from repeatedly attempting to
+			 * fill them until the writes are flushed.
+			 */
+			zero_user(page, 0, PAGE_CACHE_SIZE);
+			list_del(&page->lru);
+			lru_cache_add_file(page);
+			flush_dcache_page(page);
+			SetPageUptodate(page);
+			unlock_page(page);
+			page_cache_release(page);
+		} else {
+			/* no need to hold page hostage */
+			list_del(&page->lru);
+			lru_cache_add_file(page);
+			unlock_page(page);
+			page_cache_release(page);
+		}
+	}
+
+	/* issue the read if we have any iovecs left to fill */
+	if (rdata->nr_iov > 1) {
+		length = cifs_readv_from_socket(server, &rdata->iov[1],
+						rdata->nr_iov - 1, len);
+		if (length < 0)
+			return length;
+		server->total_read += length;
+	} else {
+		length = 0;
+	}
+
+	rdata->bytes = length;
+
+	cFYI(1, "total_read=%u rfclen=%u remaining=%u", server->total_read,
+		rfclen, remaining);
+
+	/* discard anything left over */
+	if (server->total_read < rfclen)
+		return cifs_readv_discard(server, mid);
+
+	dequeue_mid(mid, false);
+	return length;
+}
+
+static void
+cifs_readv_complete(struct work_struct *work)
+{
+	struct cifs_readdata *rdata = container_of(work,
+						struct cifs_readdata, work);
+	struct page *page, *tpage;
+
+	list_for_each_entry_safe(page, tpage, &rdata->pages, lru) {
+		list_del(&page->lru);
+		lru_cache_add_file(page);
+		kunmap(page);
+
+		if (rdata->result == 0) {
+			flush_dcache_page(page);
+			SetPageUptodate(page);
+		}
+
+		unlock_page(page);
+
+		if (rdata->result == 0)
+			cifs_readpage_to_fscache(rdata->mapping->host, page);
+
+		page_cache_release(page);
+	}
+	cifs_readdata_free(rdata);
+}
+
+static void
+cifs_readv_callback(struct mid_q_entry *mid)
+{
+	struct cifs_readdata *rdata = mid->callback_data;
+	struct cifs_tcon *tcon = tlink_tcon(rdata->cfile->tlink);
+	struct TCP_Server_Info *server = tcon->ses->server;
+
+	cFYI(1, "%s: mid=%u state=%d result=%d bytes=%u", __func__,
+		mid->mid, mid->midState, rdata->result, rdata->bytes);
+
+	switch (mid->midState) {
+	case MID_RESPONSE_RECEIVED:
+		/* result already set, check signature */
+		if (server->sec_mode &
+		    (SECMODE_SIGN_REQUIRED | SECMODE_SIGN_ENABLED)) {
+			if (cifs_verify_signature(rdata->iov, rdata->nr_iov,
+					  server, mid->sequence_number + 1))
+				cERROR(1, "Unexpected SMB signature");
+		}
+		/* FIXME: should this be counted toward the initiating task? */
+		task_io_account_read(rdata->bytes);
+		cifs_stats_bytes_read(tcon, rdata->bytes);
+		break;
+	case MID_REQUEST_SUBMITTED:
+	case MID_RETRY_NEEDED:
+		rdata->result = -EAGAIN;
+		break;
+	default:
+		rdata->result = -EIO;
+	}
+
+	queue_work(system_nrt_wq, &rdata->work);
+	DeleteMidQEntry(mid);
+	atomic_dec(&server->inFlight);
+	wake_up(&server->request_q);
+}
+
+/* cifs_async_readv - send an async write, and set up mid to handle result */
+int
+cifs_async_readv(struct cifs_readdata *rdata)
+{
+	int rc;
+	READ_REQ *smb = NULL;
+	int wct;
+	struct cifs_tcon *tcon = tlink_tcon(rdata->cfile->tlink);
+
+	cFYI(1, "%s: offset=%llu bytes=%u", __func__,
+		rdata->offset, rdata->bytes);
+
+	if (tcon->ses->capabilities & CAP_LARGE_FILES)
+		wct = 12;
+	else {
+		wct = 10; /* old style read */
+		if ((rdata->offset >> 32) > 0)  {
+			/* can not handle this big offset for old */
+			return -EIO;
+		}
+	}
+
+	rc = small_smb_init(SMB_COM_READ_ANDX, wct, tcon, (void **)&smb);
+	if (rc)
+		return rc;
+
+	smb->hdr.Pid = cpu_to_le16((__u16)rdata->pid);
+	smb->hdr.PidHigh = cpu_to_le16((__u16)(rdata->pid >> 16));
+
+	smb->AndXCommand = 0xFF;	/* none */
+	smb->Fid = rdata->cfile->netfid;
+	smb->OffsetLow = cpu_to_le32(rdata->offset & 0xFFFFFFFF);
+	if (wct == 12)
+		smb->OffsetHigh = cpu_to_le32(rdata->offset >> 32);
+	smb->Remaining = 0;
+	smb->MaxCount = cpu_to_le16(rdata->bytes & 0xFFFF);
+	smb->MaxCountHigh = cpu_to_le32(rdata->bytes >> 16);
+	if (wct == 12)
+		smb->ByteCount = 0;
+	else {
+		/* old style read */
+		struct smb_com_readx_req *smbr =
+			(struct smb_com_readx_req *)smb;
+		smbr->ByteCount = 0;
+	}
+
+	/* 4 for RFC1001 length + 1 for BCC */
+	rdata->iov[0].iov_base = smb;
+	rdata->iov[0].iov_len = be32_to_cpu(smb->hdr.smb_buf_length) + 4;
+
+	rc = cifs_call_async(tcon->ses->server, rdata->iov, 1,
+			     cifs_readv_receive, cifs_readv_callback,
+			     rdata, false);
+
+	if (rc == 0)
+		cifs_stats_inc(&tcon->num_reads);
+
+	cifs_small_buf_release(smb);
+	return rc;
+}
+
 int
 CIFSSMBRead(const int xid, struct cifs_io_parms *io_parms, unsigned int *nbytes,
 	    char **buf, int *pbuf_type)
@@ -1834,7 +2194,7 @@ cifs_async_writev(struct cifs_writedata *wdata)
 
 	kref_get(&wdata->refcount);
 	rc = cifs_call_async(tcon->ses->server, iov, wdata->nr_pages + 1,
-			     cifs_writev_callback, wdata, false);
+			     NULL, cifs_writev_callback, wdata, false);
 
 	if (rc == 0)
 		cifs_stats_inc(&tcon->num_writes);

fs/cifs/file.c

@@ -32,6 +32,7 @@
 #include <linux/delay.h>
 #include <linux/mount.h>
 #include <linux/slab.h>
+#include <linux/swap.h>
 #include <asm/div64.h>
 #include "cifsfs.h"
 #include "cifspdu.h"
@@ -1757,6 +1758,7 @@ cifs_iovec_read(struct file *file, const struct iovec *iov,
 	struct smb_com_read_rsp *pSMBr;
 	struct cifs_io_parms io_parms;
 	char *read_data;
+	unsigned int rsize;
 	__u32 pid;
 
 	if (!nr_segs)
@@ -1769,6 +1771,9 @@ cifs_iovec_read(struct file *file, const struct iovec *iov,
 	xid = GetXid();
 	cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
 
+	/* FIXME: set up handlers for larger reads and/or convert to async */
+	rsize = min_t(unsigned int, cifs_sb->rsize, CIFSMaxBufSize);
+
 	open_file = file->private_data;
 	pTcon = tlink_tcon(open_file->tlink);
 
@@ -1781,7 +1786,7 @@ cifs_iovec_read(struct file *file, const struct iovec *iov,
 		cFYI(1, "attempting read on write only file instance");
 
 	for (total_read = 0; total_read < len; total_read += bytes_read) {
-		cur_len = min_t(const size_t, len - total_read, cifs_sb->rsize);
+		cur_len = min_t(const size_t, len - total_read, rsize);
 		rc = -EAGAIN;
 		read_data = NULL;
 
@@ -1873,6 +1878,7 @@ static ssize_t cifs_read(struct file *file, char *read_data, size_t read_size,
 	unsigned int bytes_read = 0;
 	unsigned int total_read;
 	unsigned int current_read_size;
+	unsigned int rsize;
 	struct cifs_sb_info *cifs_sb;
 	struct cifs_tcon *pTcon;
 	int xid;
@@ -1885,6 +1891,9 @@ static ssize_t cifs_read(struct file *file, char *read_data, size_t read_size,
 	xid = GetXid();
 	cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
 
+	/* FIXME: set up handlers for larger reads and/or convert to async */
+	rsize = min_t(unsigned int, cifs_sb->rsize, CIFSMaxBufSize);
+
 	if (file->private_data == NULL) {
 		rc = -EBADF;
 		FreeXid(xid);
@@ -1904,8 +1913,8 @@ static ssize_t cifs_read(struct file *file, char *read_data, size_t read_size,
 	for (total_read = 0, current_offset = read_data;
 	     read_size > total_read;
 	     total_read += bytes_read, current_offset += bytes_read) {
-		current_read_size = min_t(uint, read_size - total_read,
-					  cifs_sb->rsize);
+		current_read_size = min_t(uint, read_size - total_read, rsize);
+
 		/* For windows me and 9x we do not want to request more
 		than it negotiated since it will refuse the read then */
 		if ((pTcon->ses) &&
@@ -2000,82 +2009,24 @@ int cifs_file_mmap(struct file *file, struct vm_area_struct *vma)
 	return rc;
 }
 
-
-static void cifs_copy_cache_pages(struct address_space *mapping,
-	struct list_head *pages, int bytes_read, char *data)
-{
-	struct page *page;
-	char *target;
-
-	while (bytes_read > 0) {
-		if (list_empty(pages))
-			break;
-
-		page = list_entry(pages->prev, struct page, lru);
-		list_del(&page->lru);
-
-		if (add_to_page_cache_lru(page, mapping, page->index,
-				      GFP_KERNEL)) {
-			page_cache_release(page);
-			cFYI(1, "Add page cache failed");
-			data += PAGE_CACHE_SIZE;
-			bytes_read -= PAGE_CACHE_SIZE;
-			continue;
-		}
-		page_cache_release(page);
-
-		target = kmap_atomic(page, KM_USER0);
-
-		if (PAGE_CACHE_SIZE > bytes_read) {
-			memcpy(target, data, bytes_read);
-			/* zero the tail end of this partial page */
-			memset(target + bytes_read, 0,
-			       PAGE_CACHE_SIZE - bytes_read);
-			bytes_read = 0;
-		} else {
-			memcpy(target, data, PAGE_CACHE_SIZE);
-			bytes_read -= PAGE_CACHE_SIZE;
-		}
-		kunmap_atomic(target, KM_USER0);
-
-		flush_dcache_page(page);
-		SetPageUptodate(page);
-		unlock_page(page);
-		data += PAGE_CACHE_SIZE;
-
-		/* add page to FS-Cache */
-		cifs_readpage_to_fscache(mapping->host, page);
-	}
-	return;
-}
-
 static int cifs_readpages(struct file *file, struct address_space *mapping,
 	struct list_head *page_list, unsigned num_pages)
 {
-	int rc = -EACCES;
-	int xid;
-	loff_t offset;
-	struct page *page;
-	struct cifs_sb_info *cifs_sb;
-	struct cifs_tcon *pTcon;
-	unsigned int bytes_read = 0;
-	unsigned int read_size, i;
-	char *smb_read_data = NULL;
-	struct smb_com_read_rsp *pSMBr;
-	struct cifsFileInfo *open_file;
-	struct cifs_io_parms io_parms;
-	int buf_type = CIFS_NO_BUFFER;
-	__u32 pid;
+	int rc;
+	struct list_head tmplist;
+	struct cifsFileInfo *open_file = file->private_data;
+	struct cifs_sb_info *cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
+	unsigned int rsize = cifs_sb->rsize;
+	pid_t pid;
 
-	xid = GetXid();
-	if (file->private_data == NULL) {
-		rc = -EBADF;
-		FreeXid(xid);
-		return rc;
-	}
-	open_file = file->private_data;
-	cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
-	pTcon = tlink_tcon(open_file->tlink);
+	/*
+	 * Give up immediately if rsize is too small to read an entire page.
+	 * The VFS will fall back to readpage. We should never reach this
+	 * point however since we set ra_pages to 0 when the rsize is smaller
+	 * than a cache page.
+	 */
+	if (unlikely(rsize < PAGE_CACHE_SIZE))
+		return 0;
 
 	/*
 	 * Reads as many pages as possible from fscache. Returns -ENOBUFS
@@ -2084,125 +2035,127 @@ static int cifs_readpages(struct file *file, struct address_space *mapping,
 	rc = cifs_readpages_from_fscache(mapping->host, mapping, page_list,
 					 &num_pages);
 	if (rc == 0)
-		goto read_complete;
+		return rc;
 
-	cFYI(DBG2, "rpages: num pages %d", num_pages);
 	if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_RWPIDFORWARD)
 		pid = open_file->pid;
 	else
 		pid = current->tgid;
 
-	for (i = 0; i < num_pages; ) {
-		unsigned contig_pages;
-		struct page *tmp_page;
-		unsigned long expected_index;
+	rc = 0;
+	INIT_LIST_HEAD(&tmplist);
+
+	cFYI(1, "%s: file=%p mapping=%p num_pages=%u", __func__, file,
+		mapping, num_pages);
 
-		if (list_empty(page_list))
-			break;
+	/*
+	 * Start with the page at end of list and move it to private
+	 * list. Do the same with any following pages until we hit
+	 * the rsize limit, hit an index discontinuity, or run out of
+	 * pages. Issue the async read and then start the loop again
+	 * until the list is empty.
+	 *
+	 * Note that list order is important. The page_list is in
+	 * the order of declining indexes. When we put the pages in
+	 * the rdata->pages, then we want them in increasing order.
+	 */
+	while (!list_empty(page_list)) {
+		unsigned int bytes = PAGE_CACHE_SIZE;
+		unsigned int expected_index;
+		unsigned int nr_pages = 1;
+		loff_t offset;
+		struct page *page, *tpage;
+		struct cifs_readdata *rdata;
 
 		page = list_entry(page_list->prev, struct page, lru);
+
+		/*
+		 * Lock the page and put it in the cache. Since no one else
+		 * should have access to this page, we're safe to simply set
+		 * PG_locked without checking it first.
+		 */
+		__set_page_locked(page);
+		rc = add_to_page_cache_locked(page, mapping,
+					      page->index, GFP_KERNEL);
+
+		/* give up if we can't stick it in the cache */
+		if (rc) {
+			__clear_page_locked(page);
+			break;
+		}
+
+		/* move first page to the tmplist */
 		offset = (loff_t)page->index << PAGE_CACHE_SHIFT;
+		list_move_tail(&page->lru, &tmplist);
 
-		/* count adjacent pages that we will read into */
-		contig_pages = 0;
-		expected_index =
-			list_entry(page_list->prev, struct page, lru)->index;
-		list_for_each_entry_reverse(tmp_page, page_list, lru) {
-			if (tmp_page->index == expected_index) {
-				contig_pages++;
-				expected_index++;
-			} else
+		/* now try and add more pages onto the request */
+		expected_index = page->index + 1;
+		list_for_each_entry_safe_reverse(page, tpage, page_list, lru) {
+			/* discontinuity ? */
+			if (page->index != expected_index)
 				break;
+
+			/* would this page push the read over the rsize? */
+			if (bytes + PAGE_CACHE_SIZE > rsize)
+				break;
+
+			__set_page_locked(page);
+			if (add_to_page_cache_locked(page, mapping,
+						page->index, GFP_KERNEL)) {
+				__clear_page_locked(page);
+				break;
+			}
+			list_move_tail(&page->lru, &tmplist);
+			bytes += PAGE_CACHE_SIZE;
+			expected_index++;
+			nr_pages++;
 		}
-		if (contig_pages + i >  num_pages)
-			contig_pages = num_pages - i;
-
-		/* for reads over a certain size could initiate async
-		   read ahead */
-
-		read_size = contig_pages * PAGE_CACHE_SIZE;
-		/* Read size needs to be in multiples of one page */
-		read_size = min_t(const unsigned int, read_size,
-				  cifs_sb->rsize & PAGE_CACHE_MASK);
-		cFYI(DBG2, "rpages: read size 0x%x  contiguous pages %d",
-				read_size, contig_pages);
-		rc = -EAGAIN;
-		while (rc == -EAGAIN) {
+
+		rdata = cifs_readdata_alloc(nr_pages);
+		if (!rdata) {
+			/* best to give up if we're out of mem */
+			list_for_each_entry_safe(page, tpage, &tmplist, lru) {
+				list_del(&page->lru);
+				lru_cache_add_file(page);
+				unlock_page(page);
+				page_cache_release(page);
+			}
+			rc = -ENOMEM;
+			break;
+		}
+
+		spin_lock(&cifs_file_list_lock);
+		cifsFileInfo_get(open_file);
+		spin_unlock(&cifs_file_list_lock);
+		rdata->cfile = open_file;
+		rdata->mapping = mapping;
+		rdata->offset = offset;
+		rdata->bytes = bytes;
+		rdata->pid = pid;
+		list_splice_init(&tmplist, &rdata->pages);
+
+		do {
 			if (open_file->invalidHandle) {
 				rc = cifs_reopen_file(open_file, true);
 				if (rc != 0)
-					break;
+					continue;
 			}
-			io_parms.netfid = open_file->netfid;
-			io_parms.pid = pid;
-			io_parms.tcon = pTcon;
-			io_parms.offset = offset;
-			io_parms.length = read_size;
-			rc = CIFSSMBRead(xid, &io_parms, &bytes_read,
-					 &smb_read_data, &buf_type);
-			/* BB more RC checks ? */
-			if (rc == -EAGAIN) {
-				if (smb_read_data) {
-					if (buf_type == CIFS_SMALL_BUFFER)
-						cifs_small_buf_release(smb_read_data);
-					else if (buf_type == CIFS_LARGE_BUFFER)
-						cifs_buf_release(smb_read_data);
-					smb_read_data = NULL;
-				}
-			}
-		}
-		if ((rc < 0) || (smb_read_data == NULL)) {
-			cFYI(1, "Read error in readpages: %d", rc);
-			break;
-		} else if (bytes_read > 0) {
-			task_io_account_read(bytes_read);
-			pSMBr = (struct smb_com_read_rsp *)smb_read_data;
-			cifs_copy_cache_pages(mapping, page_list, bytes_read,
-				smb_read_data + 4 /* RFC1001 hdr */ +
-				le16_to_cpu(pSMBr->DataOffset));
-
-			i +=  bytes_read >> PAGE_CACHE_SHIFT;
-			cifs_stats_bytes_read(pTcon, bytes_read);
-			if ((bytes_read & PAGE_CACHE_MASK) != bytes_read) {
-				i++; /* account for partial page */
-
-				/* server copy of file can have smaller size
-				   than client */
-				/* BB do we need to verify this common case ?
-				   this case is ok - if we are at server EOF
-				   we will hit it on next read */
+			rc = cifs_async_readv(rdata);
+		} while (rc == -EAGAIN);
 
-				/* break; */
+		if (rc != 0) {
+			list_for_each_entry_safe(page, tpage, &rdata->pages,
+						 lru) {
+				list_del(&page->lru);
+				lru_cache_add_file(page);
+				unlock_page(page);
+				page_cache_release(page);
 			}
-		} else {
-			cFYI(1, "No bytes read (%d) at offset %lld . "
-				"Cleaning remaining pages from readahead list",
-				bytes_read, offset);
-			/* BB turn off caching and do new lookup on
-			   file size at server? */
+			cifs_readdata_free(rdata);
 			break;
 		}
-		if (smb_read_data) {
-			if (buf_type == CIFS_SMALL_BUFFER)
-				cifs_small_buf_release(smb_read_data);
-			else if (buf_type == CIFS_LARGE_BUFFER)
-				cifs_buf_release(smb_read_data);
-			smb_read_data = NULL;
-		}
-		bytes_read = 0;
-	}
-
-/* need to free smb_read_data buf before exit */
-	if (smb_read_data) {
-		if (buf_type == CIFS_SMALL_BUFFER)
-			cifs_small_buf_release(smb_read_data);
-		else if (buf_type == CIFS_LARGE_BUFFER)
-			cifs_buf_release(smb_read_data);
-		smb_read_data = NULL;
 	}
 
-read_complete:
-	FreeXid(xid);
 	return rc;
 }
 

fs/cifs/transport.c

@@ -26,6 +26,7 @@
 #include <linux/wait.h>
 #include <linux/net.h>
 #include <linux/delay.h>
+#include <linux/freezer.h>
 #include <asm/uaccess.h>
 #include <asm/processor.h>
 #include <linux/mempool.h>
@@ -324,7 +325,7 @@ wait_for_response(struct TCP_Server_Info *server, struct mid_q_entry *midQ)
 {
 	int error;
 
-	error = wait_event_killable(server->response_q,
+	error = wait_event_freezekillable(server->response_q,
 				    midQ->midState != MID_REQUEST_SUBMITTED);
 	if (error < 0)
 		return -ERESTARTSYS;
@@ -339,8 +340,8 @@ wait_for_response(struct TCP_Server_Info *server, struct mid_q_entry *midQ)
  */
 int
 cifs_call_async(struct TCP_Server_Info *server, struct kvec *iov,
-		unsigned int nvec, mid_callback_t *callback, void *cbdata,
-		bool ignore_pend)
+		unsigned int nvec, mid_receive_t *receive,
+		mid_callback_t *callback, void *cbdata, bool ignore_pend)
 {
 	int rc;
 	struct mid_q_entry *mid;
@@ -374,6 +375,7 @@ cifs_call_async(struct TCP_Server_Info *server, struct kvec *iov,
 		goto out_err;
 	}
 
+	mid->receive = receive;
 	mid->callback = callback;
 	mid->callback_data = cbdata;
 	mid->midState = MID_REQUEST_SUBMITTED;

include/linux/freezer.h

@@ -134,10 +134,25 @@ static inline void set_freezable_with_signal(void)
 }
 
 /*
- * Freezer-friendly wrappers around wait_event_interruptible() and
- * wait_event_interruptible_timeout(), originally defined in <linux/wait.h>
+ * Freezer-friendly wrappers around wait_event_interruptible(),
+ * wait_event_killable() and wait_event_interruptible_timeout(), originally
+ * defined in <linux/wait.h>
  */
 
+#define wait_event_freezekillable(wq, condition)			\
+({									\
+	int __retval;							\
+	do {								\
+		__retval = wait_event_killable(wq, 			\
+				(condition) || freezing(current));	\
+		if (__retval && !freezing(current))			\
+			break;						\
+		else if (!(condition))					\
+			__retval = -ERESTARTSYS;			\
+	} while (try_to_freeze());					\
+	__retval;							\
+})
+
 #define wait_event_freezable(wq, condition)				\
 ({									\
 	int __retval;							\