NFS: Flesh out nfs_invalidate_page()

In the case of a call to truncate_inode_pages(), we should really try to
cancel any pending writes on the page.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

5 files changed, 60 insertions, 25 deletions

diff --git a/fs/nfs/file.c b/fs/nfs/file.c
index 63154070145a..106ef0dec04d 100644
--- a/fs/nfs/file.c
+++ b/fs/nfs/file.c
@@ -303,7 +303,11 @@ static int nfs_commit_write(struct file *file, struct page *page, unsigned offse
 
 static void nfs_invalidate_page(struct page *page, unsigned long offset)
 {
-	/* FIXME: we really should cancel any unstarted writes on this page */
+	struct inode *inode = page->mapping->host;
+
+	/* Cancel any unstarted writes on this page */
+	if (offset == 0)
+		nfs_sync_inode_wait(inode, page->index, 1, FLUSH_INVALIDATE);
 }
 
 static int nfs_release_page(struct page *page, gfp_t gfp)
diff --git a/fs/nfs/pagelist.c b/fs/nfs/pagelist.c
index 106aca388ebc..656481c0daa3 100644
--- a/fs/nfs/pagelist.c
+++ b/fs/nfs/pagelist.c
@@ -325,6 +325,7 @@ out:
 
 /**
  * nfs_scan_list - Scan a list for matching requests
+ * @nfsi: NFS inode
  * @head: One of the NFS inode request lists
  * @dst: Destination list
  * @idx_start: lower bound of page->index to scan
@@ -336,14 +337,15 @@ out:
  * The requests are *not* checked to ensure that they form a contiguous set.
  * You must be holding the inode's req_lock when calling this function
  */
-int
-nfs_scan_list(struct list_head *head, struct list_head *dst,
-	      unsigned long idx_start, unsigned int npages)
+int nfs_scan_list(struct nfs_inode *nfsi, struct list_head *head,
+		struct list_head *dst, unsigned long idx_start,
+		unsigned int npages)
 {
-	struct list_head	*pos, *tmp;
-	struct nfs_page		*req;
-	unsigned long		idx_end;
-	int			res;
+	struct nfs_page *pgvec[NFS_SCAN_MAXENTRIES];
+	struct nfs_page *req;
+	unsigned long idx_end;
+	int found, i;
+	int res;
 
 	res = 0;
 	if (npages == 0)
@@ -351,21 +353,28 @@ nfs_scan_list(struct list_head *head, struct list_head *dst,
 	else
 		idx_end = idx_start + npages - 1;
 
-	list_for_each_safe(pos, tmp, head) {
-
-		req = nfs_list_entry(pos);
-
-		if (req->wb_index < idx_start)
-			continue;
-		if (req->wb_index > idx_end)
+	for (;;) {
+		found = radix_tree_gang_lookup(&nfsi->nfs_page_tree,
+				(void **)&pgvec[0], idx_start,
+				NFS_SCAN_MAXENTRIES);
+		if (found <= 0)
 			break;
+		for (i = 0; i < found; i++) {
+			req = pgvec[i];
+			if (req->wb_index > idx_end)
+				goto out;
+			idx_start = req->wb_index + 1;
+			if (req->wb_list_head != head)
+				continue;
+			if (nfs_set_page_writeback_locked(req)) {
+				nfs_list_remove_request(req);
+				nfs_list_add_request(req, dst);
+				res++;
+			}
+		}
 
-		if (!nfs_set_page_writeback_locked(req))
-			continue;
-		nfs_list_remove_request(req);
-		nfs_list_add_request(req, dst);
-		res++;
 	}
+out:
 	return res;
 }
 
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index a515ec714bb6..e03abbd8302e 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -579,6 +579,17 @@ static int nfs_wait_on_requests(struct inode *inode, unsigned long idx_start, un
 	return ret;
 }
 
+static void nfs_cancel_requests(struct list_head *head)
+{
+	struct nfs_page *req;
+	while(!list_empty(head)) {
+		req = nfs_list_entry(head->next);
+		nfs_list_remove_request(req);
+		nfs_inode_remove_request(req);
+		nfs_clear_page_writeback(req);
+	}
+}
+
 /*
  * nfs_scan_dirty - Scan an inode for dirty requests
  * @inode: NFS inode to scan
@@ -623,7 +634,7 @@ nfs_scan_commit(struct inode *inode, struct list_head *dst, unsigned long idx_st
 	int res = 0;
 
 	if (nfsi->ncommit != 0) {
-		res = nfs_scan_list(&nfsi->commit, dst, idx_start, npages);
+		res = nfs_scan_list(nfsi, &nfsi->commit, dst, idx_start, npages);
 		nfsi->ncommit -= res;
 		if ((nfsi->ncommit == 0) != list_empty(&nfsi->commit))
 			printk(KERN_ERR "NFS: desynchronized value of nfs_i.ncommit.\n");
@@ -1491,15 +1502,25 @@ int nfs_sync_inode_wait(struct inode *inode, unsigned long idx_start,
 		pages = nfs_scan_dirty(inode, &head, idx_start, npages);
 		if (pages != 0) {
 			spin_unlock(&nfsi->req_lock);
-			ret = nfs_flush_list(inode, &head, pages, how);
+			if (how & FLUSH_INVALIDATE)
+				nfs_cancel_requests(&head);
+			else
+				ret = nfs_flush_list(inode, &head, pages, how);
 			spin_lock(&nfsi->req_lock);
 			continue;
 		}
 		if (nocommit)
 			break;
-		pages = nfs_scan_commit(inode, &head, 0, 0);
+		pages = nfs_scan_commit(inode, &head, idx_start, npages);
 		if (pages == 0)
 			break;
+		if (how & FLUSH_INVALIDATE) {
+			spin_unlock(&nfsi->req_lock);
+			nfs_cancel_requests(&head);
+			spin_lock(&nfsi->req_lock);
+			continue;
+		}
+		pages += nfs_scan_commit(inode, &head, 0, 0);
 		spin_unlock(&nfsi->req_lock);
 		ret = nfs_commit_list(inode, &head, how);
 		spin_lock(&nfsi->req_lock);
diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h
index 1b524b9f982a..fc48135621ed 100644
--- a/include/linux/nfs_fs.h
+++ b/include/linux/nfs_fs.h
@@ -61,6 +61,7 @@
 #define FLUSH_LOWPRI		8	/* low priority background flush */
 #define FLUSH_HIGHPRI		16	/* high priority memory reclaim flush */
 #define FLUSH_NOCOMMIT		32	/* Don't send the NFSv3/v4 COMMIT */
+#define FLUSH_INVALIDATE	64	/* Invalidate the page cache */
 
 #ifdef __KERNEL__
 
diff --git a/include/linux/nfs_page.h b/include/linux/nfs_page.h
index 66e2ed658527..8cadb0a77a7a 100644
--- a/include/linux/nfs_page.h
+++ b/include/linux/nfs_page.h
@@ -63,8 +63,8 @@ extern	void nfs_release_request(struct nfs_page *req);
 
 extern  int nfs_scan_lock_dirty(struct nfs_inode *nfsi, struct list_head *dst,
 				unsigned long idx_start, unsigned int npages);
-extern	int nfs_scan_list(struct list_head *, struct list_head *,
-			  unsigned long, unsigned int);
+extern	int nfs_scan_list(struct nfs_inode *nfsi, struct list_head *head, struct list_head *dst,
+			  unsigned long idx_start, unsigned int npages);
 extern	int nfs_coalesce_requests(struct list_head *, struct list_head *,
 				  unsigned int);
 extern  int nfs_wait_on_request(struct nfs_page *);