Re: 2.6.34-rc3-git6: Reported regressions from 2.6.33

[Al Viro <viro-3bDd1+5oDREiFSDQTTA3OLVCufUGDwFn@public.gmane.org>]

On Wed, Apr 07, 2010 at 03:22:20PM -0700, Linus Torvalds wrote:
> 
> 
> On Wed, 7 Apr 2010, Al Viro wrote:
> > 
> > No, it's not the same thing; the fix is to have nfs ->d_revalidate()
> > return an error on failing open attempt (in insane codepath that has
> > ->d_revalidate() handling open()).  Confirmed to work by reporter...
> 
> Ok, can you do the proper changelog description and sign-offs etc?

[PATCH] Have nfs ->d_revalidate() report errors properly

	If nfs atomic open implementation ends up doing open request from
->d_revalidate() codepath and gets an error from server, return that error
to caller explicitly and don't bother with lookup_instantiate_filp() at all.
->d_revalidate() can return an error itself just fine...

Signed-off-by: Al Viro <viro-RmSDqhL/yNMiFSDQTTA3OLVCufUGDwFn@public.gmane.org>
---
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index d79a7b3..fe0cd9e 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -2068,8 +2068,7 @@ nfs4_open_revalidate(struct inode *dir, struct dentry *dentry, int openflags, st
 			case -EDQUOT:
 			case -ENOSPC:
 			case -EROFS:
-				lookup_instantiate_filp(nd, (struct dentry *)state, NULL);
-				return 1;
+				return PTR_ERR(state);
 			default:
 				goto out_drop;
 		}