[PATCH V3 (resend) 3/7] CIFS: Add support for direct I/O read

[Long Li <longli@linuxonhyperv.com>]

From: Long Li <longli@microsoft.com>

With direct I/O read, we transfer the data directly from transport layer to
the user data buffer.

Change in v3: add support for kernel AIO

Signed-off-by: Long Li <longli@microsoft.com>
---
 fs/cifs/cifsfs.h   |   1 +
 fs/cifs/cifsglob.h |   5 ++
 fs/cifs/file.c     | 210 +++++++++++++++++++++++++++++++++++++++++++++--------
 3 files changed, 187 insertions(+), 29 deletions(-)

diff --git a/fs/cifs/cifsfs.h b/fs/cifs/cifsfs.h
index f047e87..ed5479c 100644
--- a/fs/cifs/cifsfs.h
+++ b/fs/cifs/cifsfs.h
@@ -101,6 +101,7 @@ extern int cifs_open(struct inode *inode, struct file *file);
 extern int cifs_close(struct inode *inode, struct file *file);
 extern int cifs_closedir(struct inode *inode, struct file *file);
 extern ssize_t cifs_user_readv(struct kiocb *iocb, struct iov_iter *to);
+extern ssize_t cifs_direct_readv(struct kiocb *iocb, struct iov_iter *to);
 extern ssize_t cifs_strict_readv(struct kiocb *iocb, struct iov_iter *to);
 extern ssize_t cifs_user_writev(struct kiocb *iocb, struct iov_iter *from);
 extern ssize_t cifs_strict_writev(struct kiocb *iocb, struct iov_iter *from);
diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h
index 9dcaed0..2131fec 100644
--- a/fs/cifs/cifsglob.h
+++ b/fs/cifs/cifsglob.h
@@ -1172,6 +1172,11 @@ struct cifs_aio_ctx {
 	unsigned int		len;
 	unsigned int		total_len;
 	bool			should_dirty;
+	/*
+	 * Indicates if this aio_ctx is for direct_io,
+	 * If yes, iter is a copy of the user passed iov_iter
+	 */
+	bool			direct_io;
 };
 
 struct cifs_readdata;
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index 8d41ca7..6a939fa 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -2965,7 +2965,6 @@ cifs_uncached_readdata_release(struct kref *refcount)
 	kref_put(&rdata->ctx->refcount, cifs_aio_ctx_release);
 	for (i = 0; i < rdata->nr_pages; i++) {
 		put_page(rdata->pages[i]);
-		rdata->pages[i] = NULL;
 	}
 	cifs_readdata_release(refcount);
 }
@@ -3004,7 +3003,7 @@ cifs_readdata_to_iov(struct cifs_readdata *rdata, struct iov_iter *iter)
 	return remaining ? -EFAULT : 0;
 }
 
-static void collect_uncached_read_data(struct cifs_aio_ctx *ctx);
+static void collect_uncached_read_data(struct cifs_readdata *rdata, struct cifs_aio_ctx *ctx);
 
 static void
 cifs_uncached_readv_complete(struct work_struct *work)
@@ -3013,7 +3012,7 @@ cifs_uncached_readv_complete(struct work_struct *work)
 						struct cifs_readdata, work);
 
 	complete(&rdata->done);
-	collect_uncached_read_data(rdata->ctx);
+	collect_uncached_read_data(rdata, rdata->ctx);
 	/* the below call can possibly free the last ref to aio ctx */
 	kref_put(&rdata->refcount, cifs_uncached_readdata_release);
 }
@@ -3103,6 +3102,9 @@ cifs_send_async_read(loff_t offset, size_t len, struct cifsFileInfo *open_file,
 	int rc;
 	pid_t pid;
 	struct TCP_Server_Info *server;
+	struct page **pagevec;
+	size_t start;
+	struct iov_iter direct_iov = ctx->iter;
 
 	server = tlink_tcon(open_file->tlink)->ses->server;
 
@@ -3111,6 +3113,9 @@ cifs_send_async_read(loff_t offset, size_t len, struct cifsFileInfo *open_file,
 	else
 		pid = current->tgid;
 
+	if (ctx->direct_io)
+		iov_iter_advance(&direct_iov, offset - ctx->pos);
+
 	do {
 		rc = server->ops->wait_mtu_credits(server, cifs_sb->rsize,
 						   &rsize, &credits);
@@ -3118,20 +3123,56 @@ cifs_send_async_read(loff_t offset, size_t len, struct cifsFileInfo *open_file,
 			break;
 
 		cur_len = min_t(const size_t, len, rsize);
-		npages = DIV_ROUND_UP(cur_len, PAGE_SIZE);
 
-		/* allocate a readdata struct */
-		rdata = cifs_readdata_alloc(npages,
+		if (ctx->direct_io) {
+
+			cur_len = iov_iter_get_pages_alloc(
+					&direct_iov, &pagevec,
+					cur_len, &start);
+			if (cur_len < 0) {
+				cifs_dbg(VFS,
+					"couldn't get user pages (cur_len=%zd)"
+					" iter type %d"
+					" iov_offset %zd count %zd\n",
+					cur_len, direct_iov.type, direct_iov.iov_offset,
+					direct_iov.count);
+				dump_stack();
+				break;
+			}
+			iov_iter_advance(&direct_iov, cur_len);
+
+			rdata = cifs_readdata_direct_alloc(
+					pagevec, cifs_uncached_readv_complete);
+			if (!rdata) {
+				add_credits_and_wake_if(server, credits, 0);
+				rc = -ENOMEM;
+				break;
+			}
+
+			npages = (cur_len + start + PAGE_SIZE-1) / PAGE_SIZE;
+			rdata->page_offset = start;
+			rdata->tailsz = npages > 1 ?
+				cur_len-(PAGE_SIZE-start)-(npages-2)*PAGE_SIZE :
+				cur_len;
+
+		} else {
+
+			npages = DIV_ROUND_UP(cur_len, PAGE_SIZE);
+			/* allocate a readdata struct */
+			rdata = cifs_readdata_alloc(npages,
 					    cifs_uncached_readv_complete);
-		if (!rdata) {
-			add_credits_and_wake_if(server, credits, 0);
-			rc = -ENOMEM;
-			break;
-		}
+			if (!rdata) {
+				add_credits_and_wake_if(server, credits, 0);
+				rc = -ENOMEM;
+				break;
+			}
 
-		rc = cifs_read_allocate_pages(rdata, npages);
-		if (rc)
-			goto error;
+			rc = cifs_read_allocate_pages(rdata, npages);
+			if (rc)
+				goto error;
+
+			rdata->tailsz = PAGE_SIZE;
+		}
 
 		rdata->cfile = cifsFileInfo_get(open_file);
 		rdata->nr_pages = npages;
@@ -3139,7 +3180,6 @@ cifs_send_async_read(loff_t offset, size_t len, struct cifsFileInfo *open_file,
 		rdata->bytes = cur_len;
 		rdata->pid = pid;
 		rdata->pagesz = PAGE_SIZE;
-		rdata->tailsz = PAGE_SIZE;
 		rdata->read_into_pages = cifs_uncached_read_into_pages;
 		rdata->copy_into_pages = cifs_uncached_copy_into_pages;
 		rdata->credits = credits;
@@ -3153,13 +3193,17 @@ cifs_send_async_read(loff_t offset, size_t len, struct cifsFileInfo *open_file,
 		if (rc) {
 			add_credits_and_wake_if(server, rdata->credits, 0);
 			kref_put(&rdata->refcount,
-				 cifs_uncached_readdata_release);
-			if (rc == -EAGAIN)
+				cifs_uncached_readdata_release);
+			if (rc == -EAGAIN) {
+				iov_iter_revert(&direct_iov, cur_len);
 				continue;
+			}
 			break;
 		}
 
-		list_add_tail(&rdata->list, rdata_list);
+		/* Add to aio pending list if it's not there */
+		if (rdata_list)
+			list_add_tail(&rdata->list, rdata_list);
 		offset += cur_len;
 		len -= cur_len;
 	} while (len > 0);
@@ -3168,7 +3212,7 @@ cifs_send_async_read(loff_t offset, size_t len, struct cifsFileInfo *open_file,
 }
 
 static void
-collect_uncached_read_data(struct cifs_aio_ctx *ctx)
+collect_uncached_read_data(struct cifs_readdata *uncached_rdata, struct cifs_aio_ctx *ctx)
 {
 	struct cifs_readdata *rdata, *tmp;
 	struct iov_iter *to = &ctx->iter;
@@ -3211,10 +3255,12 @@ collect_uncached_read_data(struct cifs_aio_ctx *ctx)
 				 * reading.
 				 */
 				if (got_bytes && got_bytes < rdata->bytes) {
-					rc = cifs_readdata_to_iov(rdata, to);
+					rc = 0;
+					if (!ctx->direct_io)
+						rc = cifs_readdata_to_iov(rdata, to);
 					if (rc) {
 						kref_put(&rdata->refcount,
-						cifs_uncached_readdata_release);
+							cifs_uncached_readdata_release);
 						continue;
 					}
 				}
@@ -3228,28 +3274,32 @@ collect_uncached_read_data(struct cifs_aio_ctx *ctx)
 				list_splice(&tmp_list, &ctx->list);
 
 				kref_put(&rdata->refcount,
-					 cifs_uncached_readdata_release);
+					cifs_uncached_readdata_release);
 				goto again;
 			} else if (rdata->result)
 				rc = rdata->result;
-			else
+			else if (!ctx->direct_io)
 				rc = cifs_readdata_to_iov(rdata, to);
 
 			/* if there was a short read -- discard anything left */
 			if (rdata->got_bytes && rdata->got_bytes < rdata->bytes)
 				rc = -ENODATA;
+
+			ctx->total_len += rdata->got_bytes;
 		}
 		list_del_init(&rdata->list);
 		kref_put(&rdata->refcount, cifs_uncached_readdata_release);
 	}
 
-	for (i = 0; i < ctx->npages; i++) {
-		if (ctx->should_dirty)
-			set_page_dirty(ctx->bv[i].bv_page);
-		put_page(ctx->bv[i].bv_page);
-	}
+	if (!ctx->direct_io) {
+		for (i = 0; i < ctx->npages; i++) {
+			if (ctx->should_dirty)
+				set_page_dirty(ctx->bv[i].bv_page);
+			put_page(ctx->bv[i].bv_page);
+		}
 
-	ctx->total_len = ctx->len - iov_iter_count(to);
+		ctx->total_len = ctx->len - iov_iter_count(to);
+	}
 
 	cifs_stats_bytes_read(tcon, ctx->total_len);
 
@@ -3267,6 +3317,108 @@ collect_uncached_read_data(struct cifs_aio_ctx *ctx)
 		complete(&ctx->done);
 }
 
+ssize_t cifs_direct_readv(struct kiocb *iocb, struct iov_iter *to)
+{
+	size_t len;
+	struct file *file;
+	struct cifs_sb_info *cifs_sb;
+	struct cifsFileInfo *cfile;
+	struct cifs_tcon *tcon;
+	ssize_t rc, total_read = 0;
+	struct TCP_Server_Info *server;
+	loff_t offset = iocb->ki_pos;
+	pid_t pid;
+	struct cifs_aio_ctx *ctx;
+
+	/*
+	 * iov_iter_get_pages_alloc() doesn't work with ITER_KVEC,
+	 * fall back to data copy read path
+	 * this could be improved by getting pages directly in ITER_KVEC
+	 */
+	if (to->type & ITER_KVEC) {
+		cifs_dbg(FYI, "use non-direct cifs_user_readv for kvec I/O\n");
+		return cifs_user_readv(iocb, to);
+	}
+
+	len = iov_iter_count(to);
+	if (!len)
+		return 0;
+
+	file = iocb->ki_filp;
+	cifs_sb = CIFS_FILE_SB(file);
+	cfile = file->private_data;
+	tcon = tlink_tcon(cfile->tlink);
+	server = tcon->ses->server;
+
+	if (!server->ops->async_readv)
+		return -ENOSYS;
+
+	if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_RWPIDFORWARD)
+		pid = cfile->pid;
+	else
+		pid = current->tgid;
+
+	if ((file->f_flags & O_ACCMODE) == O_WRONLY)
+		cifs_dbg(FYI, "attempting read on write only file instance\n");
+
+	ctx = cifs_aio_ctx_alloc();
+	if (!ctx)
+		return -ENOMEM;
+
+	ctx->cfile = cifsFileInfo_get(cfile);
+
+	if (!is_sync_kiocb(iocb))
+		ctx->iocb = iocb;
+
+	if (to->type == ITER_IOVEC)
+		ctx->should_dirty = true;
+
+	ctx->pos = offset;
+	ctx->direct_io = true;
+	ctx->iter = *to;
+	ctx->len = len;
+
+	/* grab a lock here due to read response handlers can access ctx */
+	mutex_lock(&ctx->aio_mutex);
+
+	rc = cifs_send_async_read(offset, len, cfile, cifs_sb, &ctx->list, ctx);
+
+	/* if at least one read request send succeeded, then reset rc */
+	if (!list_empty(&ctx->list))
+		rc = 0;
+
+	mutex_unlock(&ctx->aio_mutex);
+
+	if (rc) {
+		kref_put(&ctx->refcount, cifs_aio_ctx_release);
+		return rc;
+	}
+
+	if (!is_sync_kiocb(iocb)) {
+		kref_put(&ctx->refcount, cifs_aio_ctx_release);
+		return -EIOCBQUEUED;
+	}
+
+	rc = wait_for_completion_killable(&ctx->done);
+	if (rc) {
+		mutex_lock(&ctx->aio_mutex);
+		ctx->rc = rc = -EINTR;
+		total_read = ctx->total_len;
+		mutex_unlock(&ctx->aio_mutex);
+	} else {
+		rc = ctx->rc;
+		total_read = ctx->total_len;
+	}
+
+	kref_put(&ctx->refcount, cifs_aio_ctx_release);
+
+	if (total_read) {
+		iocb->ki_pos += total_read;
+		return total_read;
+	}
+	return rc;
+}
+
 ssize_t cifs_user_readv(struct kiocb *iocb, struct iov_iter *to)
 {
 	struct file *file = iocb->ki_filp;
-- 
2.7.4