Re: [PATCH] NFS: fix subtle change in COMMIT behavior

[Christoph Hellwig <hch@lst.de>]

[-- Attachment #1: Type: text/plain, Size: 52 bytes --]

The patch below passes block layout testing for me:

[-- Attachment #2: 0001-pnfs-blocklayout-fix-end-calculation-in-pnfs_num_con.patch --]
[-- Type: text/x-patch, Size: 1017 bytes --]

From: Christoph Hellwig <hch@lst.de>
Subject: pnfs/blocklayout: fix end calculation in pnfs_num_cont_bytes

Use the number of pages in the pagecache mapping instead of the
number of pnfs requests which is only slightly related.

Reported-by: Weston Andros Adamson <dros@primarydata.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 fs/nfs/blocklayout/blocklayout.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/nfs/blocklayout/blocklayout.c b/fs/nfs/blocklayout/blocklayout.c
index 5228f20..98d4900 100644
--- a/fs/nfs/blocklayout/blocklayout.c
+++ b/fs/nfs/blocklayout/blocklayout.c
@@ -812,7 +812,7 @@ static u64 pnfs_num_cont_bytes(struct inode *inode, pgoff_t idx)
 
 	/* Optimize common case that writes from 0 to end of file */
 	end = DIV_ROUND_UP(i_size_read(inode), PAGE_CACHE_SIZE);
-	if (end != NFS_I(inode)->npages) {
+	if (end != inode->i_mapping->nr_pages) {
 		rcu_read_lock();
 		end = page_cache_next_hole(mapping, idx + 1, ULONG_MAX);
 		rcu_read_unlock();
-- 
1.9.1