Avoid GPF when attempting to save / restore 64bit guests. More thought

required as to best way to get M2P mfns under 64...

Signed-off-by: Steven Hand <steven@xensource.com>

diff --git a/linux-2.6-xen-sparse/drivers/xen/privcmd/privcmd.c b/linux-2.6-xen-sparse/drivers/xen/privcmd/privcmd.c
index cf0c0eb7d70917dc8d108fa3d20973ca04d808f6..7e9aceaaae42719b37c3e230236682be3d604a93 100644 (file)
--- a/linux-2.6-xen-sparse/drivers/xen/privcmd/privcmd.c
+++ b/linux-2.6-xen-sparse/drivers/xen/privcmd/privcmd.c
@@ -222,6 +222,17 @@ static int privcmd_ioctl(struct inode *inode, struct file *file,
                unsigned long *p; 
                int i; 
 
+#if defined (__x86_64__)
+               /* 
+               ** XXX SMH: the below procedure won't work for 64 since 
+               ** we don't have access to the memory which maps the M2P. 
+               ** A proper fix will probably involve moving this 
+               ** functionality to Xen - for now just return an error 
+               ** here rather than GPF'ing in the kernel. 
+               */
+               ret = -EINVAL; 
+               break; 
+#endif
                if (copy_from_user(&m, (void *)data, sizeof(m)))
                        return -EFAULT;