Re: oops when calling dm_set_disp()

[Dean Roehrich <roehrich@xxxxxxx>]

>From:  Damian Hazen <dhazen@xxxxxxx>
>It looks like the vfs superblock structure is not being filled in
>completely until after the mount event is responded to.  If 
>dm_set_disp() is called before responding, it causes an oops in 
>dm_handle_to_ip() when it attempts to get the dentry here:
>
>
>         /* Now that the mutex is released, wait until we have access to the
>            inode.
>         */
>
>         sb = fsrp->fr_sb;
>         if (fidp->fid_len == 0) {       /* filesystem handle */
>=>                ip = sb->s_root->d_inode;
>                 igrab(ip);
>
>
>If you wait and call dm_set_disp() after responding, there's still a 
>race with linvfs_fill_super() to get s_root set.  Would it be possible 
>to move up setting s_root to somewhere before dm_send_mount_event() is 
>called?  It's nice to be able to set disposition and event lists before 
>turning the mount event loose to prevent anyone sneaking into the 
>filesystem before we're ready.


This fixes the panic in my test.  Would you try this please?

Dean


Index: 2.4.x-xfs-e/fs/dmapi/dmapi_register.c
===================================================================
--- 2.4.x-xfs-e.orig/fs/dmapi/dmapi_register.c  2005-02-18 12:30:10.000000000 
-0600
+++ 2.4.x-xfs-e/fs/dmapi/dmapi_register.c       2005-03-07 10:12:14.000000000 
-0600
@@ -507,6 +507,7 @@ dm_handle_to_ip(
        struct super_block *sb;
        struct inode    *ip;
        int             filetype;
+       struct filesystem_dmapi_operations *dmapiops;
 
        if ((fsrp = dm_find_fsreg_and_lock(&handlep->ha_fsid, &lc)) == NULL)
                return(NULL);
@@ -549,17 +550,11 @@ dm_handle_to_ip(
        */
 
        sb = fsrp->fr_sb;
-       if (fidp->fid_len == 0) {       /* filesystem handle */
-               ip = sb->s_root->d_inode;
-               igrab(ip);
-       } else {                        /* file object handle */
-               struct filesystem_dmapi_operations *dmapiops;
-               error = -ENOSYS;
-               dmapiops = dm_fsys_ops_by_fstype(sb->s_type);
-               ASSERT(dmapiops);
-               if (dmapiops->fh_to_inode)
-                       error = dmapiops->fh_to_inode(sb, &ip, (void*)fidp);
-       }
+       error = -ENOSYS;
+       dmapiops = dm_fsys_ops_by_fstype(sb->s_type);
+       ASSERT(dmapiops);
+       if (dmapiops->fh_to_inode)
+               error = dmapiops->fh_to_inode(sb, &ip, (void*)fidp);
 
        lc = mutex_spinlock(&fsrp->fr_lock);
 
Index: 2.4.x-xfs-e/fs/xfs/xfs_dmapi.c
===================================================================
--- 2.4.x-xfs-e.orig/fs/xfs/xfs_dmapi.c 2005-03-03 12:52:56.000000000 -0600
+++ 2.4.x-xfs-e/fs/xfs/xfs_dmapi.c      2005-03-07 10:17:51.000000000 -0600
@@ -3291,9 +3291,13 @@ xfs_dm_fh_to_inode(
 
        *ip = NULL;
        memcpy(&fid, dmfsfid, sizeof(*dmfsfid));
-       ASSERT(fid.fid_len);
-       VFS_VGET(vfsp, &vp, &fid, error);
-       if (vp && (error == 0))
+       if (fid.fid_len) {      /* file object handle */
+               VFS_VGET(vfsp, &vp, &fid, error);
+       }
+       else {                  /* filesystem handle */
+               VFS_ROOT(vfsp, &vp, error);
+       }
+       if(vp && (error == 0))
                *ip = LINVFS_GET_IP(vp);
        return -error; /* Return negative error to DMAPI */
 }