Merge tag 'nfs-for-5.1-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

Pull NFS client updates from Trond Myklebust:
 "Highlights include:

  Stable fixes:
   - Fixes for NFS I/O request leakages
   - Fix error handling paths in the NFS I/O recoalescing code
   - Reinitialise NFSv4.1 sequence results before retransmitting a
     request
   - Fix a soft lockup in the delegation recovery code
   - Bulk destroy of layouts needs to be safe w.r.t. umount
   - Prevent thundering herd issues when the SUNRPC socket is not
     connected
   - Respect RPC call timeouts when retrying transmission

  Features:
   - Convert rpc auth layer to use xdr_streams
   - Config option to disable insecure RPCSEC_GSS crypto types
   - Reduce size of RPC receive buffers
   - Readdirplus optimization by cache mechanism
   - Convert SUNRPC socket send code to use iov_iter()
   - SUNRPC micro-optimisations to avoid indirect calls
   - Add support for the pNFS LAYOUTERROR operation and use it with the
     pNFS/flexfiles driver
   - Add trace events to report non-zero NFS status codes
   - Various removals of unnecessary dprintks

  Bugfixes and cleanups:
   - Fix a number of sparse warnings and documentation format warnings
   - Fix nfs_parse_devname to not modify it's argument
   - Fix potential corruption of page being written through pNFS/blocks
   - fix xfstest generic/099 failures on nfsv3
   - Avoid NFSv4.1 "false retries" when RPC calls are interrupted
   - Abort I/O early if the pNFS/flexfiles layout segment was
     invalidated
   - Avoid unnecessary pNFS/flexfiles layout invalidations"

* tag 'nfs-for-5.1-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs: (90 commits)
  SUNRPC: Take the transport send lock before binding+connecting
  SUNRPC: Micro-optimise when the task is known not to be sleeping
  SUNRPC: Check whether the task was transmitted before rebind/reconnect
  SUNRPC: Remove redundant calls to RPC_IS_QUEUED()
  SUNRPC: Clean up
  SUNRPC: Respect RPC call timeouts when retrying transmission
  SUNRPC: Fix up RPC back channel transmission
  SUNRPC: Prevent thundering herd when the socket is not connected
  SUNRPC: Allow dynamic allocation of back channel slots
  NFSv4.1: Bump the default callback session slot count to 16
  SUNRPC: Convert remaining GFP_NOIO, and GFP_NOWAIT sites in sunrpc
  NFS/flexfiles: Clean up mirror DS initialisation
  NFS/flexfiles: Remove dead code in ff_layout_mirror_valid()
  NFS/flexfile: Simplify nfs4_ff_layout_select_ds_stateid()
  NFS/flexfile: Simplify nfs4_ff_layout_ds_version()
  NFS/flexfiles: Simplify ff_layout_get_ds_cred()
  NFS/flexfiles: Simplify nfs4_ff_find_or_create_ds_client()
  NFS/flexfiles: Simplify nfs4_ff_layout_select_ds_fh()
  NFS/flexfiles: Speed up read failover when DSes are down
  NFS/flexfiles: Don't invalidate DS deviceids for being unresponsive
  ...

1 files changed, 28 insertions, 16 deletions

diff --git a/fs/nfs/file.c b/fs/nfs/file.c
index 29553fdba8af7..4899b85f9b3c2 100644
--- a/fs/nfs/file.c
+++ b/fs/nfs/file.c
@@ -89,8 +89,8 @@ EXPORT_SYMBOL_GPL(nfs_file_release);
 
 /**
  * nfs_revalidate_size - Revalidate the file size
- * @inode - pointer to inode struct
- * @file - pointer to struct file
+ * @inode: pointer to inode struct
+ * @filp: pointer to struct file
  *
  * Revalidates the file length. This is basically a wrapper around
  * nfs_revalidate_inode() that takes into account the fact that we may
@@ -276,6 +276,12 @@ EXPORT_SYMBOL_GPL(nfs_file_fsync);
  * then a modify/write/read cycle when writing to a page in the
  * page cache.
  *
+ * Some pNFS layout drivers can only read/write at a certain block
+ * granularity like all block devices and therefore we must perform
+ * read/modify/write whenever a page hasn't read yet and the data
+ * to be written there is not aligned to a block boundary and/or
+ * smaller than the block size.
+ *
  * The modify/write/read cycle may occur if a page is read before
  * being completely filled by the writer.  In this situation, the
  * page must be completely written to stable storage on the server
@@ -291,26 +297,32 @@ EXPORT_SYMBOL_GPL(nfs_file_fsync);
  * and that the new data won't completely replace the old data in
  * that range of the file.
  */
-static int nfs_want_read_modify_write(struct file *file, struct page *page,
-			loff_t pos, unsigned len)
+static bool nfs_full_page_write(struct page *page, loff_t pos, unsigned int len)
 {
 	unsigned int pglen = nfs_page_length(page);
 	unsigned int offset = pos & (PAGE_SIZE - 1);
 	unsigned int end = offset + len;
 
-	if (pnfs_ld_read_whole_page(file->f_mapping->host)) {
-		if (!PageUptodate(page))
-			return 1;
-		return 0;
-	}
+	return !pglen || (end >= pglen && !offset);
+}
 
-	if ((file->f_mode & FMODE_READ) &&	/* open for read? */
-	    !PageUptodate(page) &&		/* Uptodate? */
-	    !PagePrivate(page) &&		/* i/o request already? */
-	    pglen &&				/* valid bytes of file? */
-	    (end < pglen || offset))		/* replace all valid bytes? */
-		return 1;
-	return 0;
+static bool nfs_want_read_modify_write(struct file *file, struct page *page,
+			loff_t pos, unsigned int len)
+{
+	/*
+	 * Up-to-date pages, those with ongoing or full-page write
+	 * don't need read/modify/write
+	 */
+	if (PageUptodate(page) || PagePrivate(page) ||
+	    nfs_full_page_write(page, pos, len))
+		return false;
+
+	if (pnfs_ld_read_whole_page(file->f_mapping->host))
+		return true;
+	/* Open for reading too? */
+	if (file->f_mode & FMODE_READ)
+		return true;
+	return false;
 }
 
 /*