9p: Fix the incorrect update of inode size in v9fs_file_write()
Abhishek Kulkarni [Tue, 22 Sep 2009 16:34:05 +0000 (11:34 -0500)]
When using the cache=loose flags, the inode's size was not being
updated correctly on a remote write. Thus subsequent reads of
the whole file resulted in a truncated read. Fix it.

Signed-off-by: Abhishek Kulkarni <adkulkar@umail.iu.edu>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>

fs/9p/vfs_file.c

index de7690e..cafaa46 100644 (file)
@@ -32,6 +32,7 @@
 #include <linux/string.h>
 #include <linux/inet.h>
 #include <linux/list.h>
+#include <linux/pagemap.h>
 #include <asm/uaccess.h>
 #include <linux/idr.h>
 #include <net/9p/9p.h>
@@ -210,6 +211,7 @@ v9fs_file_write(struct file *filp, const char __user * data,
        struct p9_client *clnt;
        struct inode *inode = filp->f_path.dentry->d_inode;
        int origin = *offset;
+       unsigned long pg_start, pg_end;
 
        P9_DPRINTK(P9_DEBUG_VFS, "data %p count %d offset %x\n", data,
                (int)count, (int)*offset);
@@ -225,7 +227,7 @@ v9fs_file_write(struct file *filp, const char __user * data,
                if (count < rsize)
                        rsize = count;
 
-               n = p9_client_write(fid, NULL, data+total, *offset+total,
+               n = p9_client_write(fid, NULL, data+total, origin+total,
                                                                        rsize);
                if (n <= 0)
                        break;
@@ -234,13 +236,12 @@ v9fs_file_write(struct file *filp, const char __user * data,
        } while (count > 0);
 
        if (total > 0) {
-               invalidate_inode_pages2_range(inode->i_mapping, origin,
-                                                               origin+total);
+               pg_start = origin >> PAGE_CACHE_SHIFT;
+               pg_end = (origin + total - 1) >> PAGE_CACHE_SHIFT;
+               invalidate_inode_pages2_range(inode->i_mapping, pg_start,
+                                                               pg_end);
                *offset += total;
-       }
-
-       if (*offset > i_size_read(inode)) {
-               i_size_write(inode, *offset);
+               i_size_write(inode, i_size_read(inode) + total);
                inode->i_blocks = (i_size_read(inode) + 512 - 1) >> 9;
        }