NFS: support large reads and writes on the wire

 Most NFS server implementations allow up to 64KB reads and writes on the
 wire.  The Solaris NFS server allows up to a megabyte, for instance.

 Now the Linux NFS client supports transfer sizes up to 1MB, too.  This will
 help reduce protocol and context switch overhead on read/write intensive NFS
 workloads, and support larger atomic read and write operations on servers
 that support them.

 Test-plan:
 Connectathon and iozone on mount point with wsize=rsize>32768 over TCP.
 Tests with NFS over UDP to verify the maximum RPC payload size cap.

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

1 files changed, 35 insertions, 6 deletions

diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h
index 4dff705d2ff2..d38010ba6477 100644
--- a/include/linux/nfs_fs.h
+++ b/include/linux/nfs_fs.h
@@ -38,9 +38,6 @@
 # define NFS_DEBUG
 #endif
 
-#define NFS_MAX_FILE_IO_BUFFER_SIZE     32768
-#define NFS_DEF_FILE_IO_BUFFER_SIZE     4096
-
 /* Default timeout values */
 #define NFS_MAX_UDP_TIMEOUT     (60*HZ)
 #define NFS_MAX_TCP_TIMEOUT     (600*HZ)
@@ -462,18 +459,33 @@ static inline int nfs_wb_page(struct inode *inode, struct page* page)
  */
 extern mempool_t *nfs_wdata_mempool;
 
-static inline struct nfs_write_data *nfs_writedata_alloc(void)
+static inline struct nfs_write_data *nfs_writedata_alloc(unsigned int pagecount)
 {
         struct nfs_write_data *p = mempool_alloc(nfs_wdata_mempool, SLAB_NOFS);
+
         if (p) {
                 memset(p, 0, sizeof(*p));
                 INIT_LIST_HEAD(&p->pages);
+                if (pagecount < NFS_PAGEVEC_SIZE)
+                        p->pagevec = &p->page_array[0];
+                else {
+                        size_t size = ++pagecount * sizeof(struct page *);
+                        p->pagevec = kmalloc(size, GFP_NOFS);
+                        if (p->pagevec) {
+                                memset(p->pagevec, 0, size);
+                        } else {
+                                mempool_free(p, nfs_wdata_mempool);
+                                p = NULL;
+                        }
+                }
         }
         return p;
 }
 
 static inline void nfs_writedata_free(struct nfs_write_data *p)
 {
+        if (p && (p->pagevec != &p->page_array[0]))
+                kfree(p->pagevec);
         mempool_free(p, nfs_wdata_mempool);
 }
 
@@ -492,16 +504,33 @@ extern void  nfs_readdata_release(void *data);
  */
 extern mempool_t *nfs_rdata_mempool;
 
-static inline struct nfs_read_data *nfs_readdata_alloc(void)
+static inline struct nfs_read_data *nfs_readdata_alloc(unsigned int pagecount)
 {
         struct nfs_read_data *p = mempool_alloc(nfs_rdata_mempool, SLAB_NOFS);
-        if (p)
+
+        if (p) {
                 memset(p, 0, sizeof(*p));
+                INIT_LIST_HEAD(&p->pages);
+                if (pagecount < NFS_PAGEVEC_SIZE)
+                        p->pagevec = &p->page_array[0];
+                else {
+                        size_t size = ++pagecount * sizeof(struct page *);
+                        p->pagevec = kmalloc(size, GFP_NOFS);
+                        if (p->pagevec) {
+                                memset(p->pagevec, 0, size);
+                        } else {
+                                mempool_free(p, nfs_rdata_mempool);
+                                p = NULL;
+                        }
+                }
+        }
         return p;
 }
 
 static inline void nfs_readdata_free(struct nfs_read_data *p)
 {
+        if (p && (p->pagevec != &p->page_array[0]))
+                kfree(p->pagevec);
         mempool_free(p, nfs_rdata_mempool);
 }