Make smbfs with UNIX extensions get file disk usage count right.

We need to convert from bytes to blocks, and we also need to
handle the fact that old smbd's will round up the disk usage
bytecount to even multiples of 1MB, and gracefully fall back
on an estimate based on the file size instead in that case.

Acked by Samuel Thibault who noted the problem in the first
place, and Jeremy Allison wrt smbd behavior.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>

diff --git a/fs/smbfs/proc.c b/fs/smbfs/proc.c
index 8096d74922eca2757947eb961979c387c3f5ba42..418c0c36ac8b62bf677d88b526fdf074447c0638 100644 (file)
--- a/fs/smbfs/proc.c
+++ b/fs/smbfs/proc.c
@@ -2076,6 +2076,8 @@ out:
 
 void smb_decode_unix_basic(struct smb_fattr *fattr, char *p)
 {
+       u64 size, disk_bytes;
+
        /* FIXME: verify nls support. all is sent as utf8? */
 
        fattr->f_unix = 1;
@@ -2093,8 +2095,19 @@ void smb_decode_unix_basic(struct smb_fattr *fattr, char *p)
        /* 84 L permissions */
        /* 92 L link count */
 
-       fattr->f_size = LVAL(p, 0);
-       fattr->f_blocks = LVAL(p, 8);
+       size = LVAL(p, 0);
+       disk_bytes = LVAL(p, 8);
+
+       /*
+        * Some samba versions round up on-disk byte usage
+        * to 1MB boundaries, making it useless. When seeing
+        * that, use the size instead.
+        */
+       if (!(disk_bytes & 0xfffff))
+               disk_bytes = size+511;
+
+       fattr->f_size = size;
+       fattr->f_blocks = disk_bytes >> 9;
        fattr->f_ctime = smb_ntutc2unixutc(LVAL(p, 16));
        fattr->f_atime = smb_ntutc2unixutc(LVAL(p, 24));
        fattr->f_mtime = smb_ntutc2unixutc(LVAL(p, 32));