All of lore.kernel.org
 help / color / mirror / Atom feed
From: David Howells <dhowells@redhat.com>
To: linux-afs@lists.infradead.org
Cc: Markus Suvanto <markus.suvanto@gmail.com>,
	dhowells@redhat.com, markus.suvanto@gmail.com,
	Marc Dionne <marc.dionne@auristor.com>,
	linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: [PATCH 4/6] afs: Fix incorrect triggering of sillyrename on 3rd-party invalidation
Date: Wed, 08 Sep 2021 16:58:01 +0100	[thread overview]
Message-ID: <163111668100.283156.3851669884664475428.stgit@warthog.procyon.org.uk> (raw)
In-Reply-To: <163111665183.283156.17200205573146438918.stgit@warthog.procyon.org.uk>

The AFS filesystem is currently triggering the silly-rename cleanup from
afs_d_revalidate() when it sees that a dentry has been changed by a third
party[1].  It should not be doing this as the cleanup includes deleting the
silly-rename target file on iput.

Fix this by removing the places in the d_revalidate handling that validate
anything other than the directory and the dirent.  It probably should not
be looking to validate the target inode of the dentry also.

This includes removing the point in afs_d_revalidate() where the inode that
a dentry used to point to was marked as being deleted (AFS_VNODE_DELETED).
We don't know it got deleted.  It could have been renamed or it could have
hard links remaining.

This was reproduced by cloning a git repo onto an afs volume on one
machine, switching to another machine and doing "git status", then
switching back to the first and doing "git status".  The second status
would show weird output due to ".git/index" getting deleted by the above
mentioned mechanism.

A simpler way to do it is to do:

	machine 1: touch a
	machine 2: touch b; mv -f b a
	machine 1: stat a

on an afs volume.  The bug shows up as the stat failing with ENOENT and the
file server log showing that machine 1 deleted "a".

Fixes: 79ddbfa500b3 ("afs: Implement sillyrename for unlink and rename")
Reported-by: Markus Suvanto <markus.suvanto@gmail.com>
Signed-off-by: David Howells <dhowells@redhat.com>
cc: linux-afs@lists.infradead.org
Link: https://bugzilla.kernel.org/show_bug.cgi?id=214217#c4 [1]
---

 fs/afs/dir.c |   46 +++++++---------------------------------------
 1 file changed, 7 insertions(+), 39 deletions(-)

diff --git a/fs/afs/dir.c b/fs/afs/dir.c
index a8e3ae55f1f9..4579bbda4634 100644
--- a/fs/afs/dir.c
+++ b/fs/afs/dir.c
@@ -1077,9 +1077,9 @@ static struct dentry *afs_lookup(struct inode *dir, struct dentry *dentry,
  */
 static int afs_d_revalidate_rcu(struct dentry *dentry)
 {
-	struct afs_vnode *dvnode, *vnode;
+	struct afs_vnode *dvnode;
 	struct dentry *parent;
-	struct inode *dir, *inode;
+	struct inode *dir;
 	long dir_version, de_version;
 
 	_enter("%p", dentry);
@@ -1109,18 +1109,6 @@ static int afs_d_revalidate_rcu(struct dentry *dentry)
 			return -ECHILD;
 	}
 
-	/* Check to see if the vnode referred to by the dentry still
-	 * has a callback.
-	 */
-	if (d_really_is_positive(dentry)) {
-		inode = d_inode_rcu(dentry);
-		if (inode) {
-			vnode = AFS_FS_I(inode);
-			if (!afs_check_validity(vnode))
-				return -ECHILD;
-		}
-	}
-
 	return 1; /* Still valid */
 }
 
@@ -1156,17 +1144,7 @@ static int afs_d_revalidate(struct dentry *dentry, unsigned int flags)
 	if (IS_ERR(key))
 		key = NULL;
 
-	if (d_really_is_positive(dentry)) {
-		inode = d_inode(dentry);
-		if (inode) {
-			vnode = AFS_FS_I(inode);
-			afs_validate(vnode, key);
-			if (test_bit(AFS_VNODE_DELETED, &vnode->flags))
-				goto out_bad;
-		}
-	}
-
-	/* lock down the parent dentry so we can peer at it */
+	/* Hold the parent dentry so we can peer at it */
 	parent = dget_parent(dentry);
 	dir = AFS_FS_I(d_inode(parent));
 
@@ -1175,7 +1153,7 @@ static int afs_d_revalidate(struct dentry *dentry, unsigned int flags)
 
 	if (test_bit(AFS_VNODE_DELETED, &dir->flags)) {
 		_debug("%pd: parent dir deleted", dentry);
-		goto out_bad_parent;
+		goto not_found;
 	}
 
 	/* We only need to invalidate a dentry if the server's copy changed
@@ -1201,12 +1179,12 @@ static int afs_d_revalidate(struct dentry *dentry, unsigned int flags)
 	case 0:
 		/* the filename maps to something */
 		if (d_really_is_negative(dentry))
-			goto out_bad_parent;
+			goto not_found;
 		inode = d_inode(dentry);
 		if (is_bad_inode(inode)) {
 			printk("kAFS: afs_d_revalidate: %pd2 has bad inode\n",
 			       dentry);
-			goto out_bad_parent;
+			goto not_found;
 		}
 
 		vnode = AFS_FS_I(inode);
@@ -1228,9 +1206,6 @@ static int afs_d_revalidate(struct dentry *dentry, unsigned int flags)
 			       dentry, fid.unique,
 			       vnode->fid.unique,
 			       vnode->vfs_inode.i_generation);
-			write_seqlock(&vnode->cb_lock);
-			set_bit(AFS_VNODE_DELETED, &vnode->flags);
-			write_sequnlock(&vnode->cb_lock);
 			goto not_found;
 		}
 		goto out_valid;
@@ -1245,7 +1220,7 @@ static int afs_d_revalidate(struct dentry *dentry, unsigned int flags)
 	default:
 		_debug("failed to iterate dir %pd: %d",
 		       parent, ret);
-		goto out_bad_parent;
+		goto not_found;
 	}
 
 out_valid:
@@ -1256,16 +1231,9 @@ static int afs_d_revalidate(struct dentry *dentry, unsigned int flags)
 	_leave(" = 1 [valid]");
 	return 1;
 
-	/* the dirent, if it exists, now points to a different vnode */
 not_found:
-	spin_lock(&dentry->d_lock);
-	dentry->d_flags |= DCACHE_NFSFS_RENAMED;
-	spin_unlock(&dentry->d_lock);
-
-out_bad_parent:
 	_debug("dropping dentry %pd2", dentry);
 	dput(parent);
-out_bad:
 	key_put(key);
 
 	_leave(" = 0 [bad]");



  parent reply	other threads:[~2021-09-08 15:58 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-09-08 15:57 [PATCH 0/6] afs: Fixes for 3rd party-induced data corruption David Howells
2021-09-08 15:57 ` [PATCH 1/6] afs: Fix missing put on afs_read objects and missing get on the key therein David Howells
2021-09-10 21:10   ` Marc Dionne
2021-09-08 15:57 ` [PATCH 2/6] afs: Fix page leak David Howells
2021-09-08 20:37   ` Marc Dionne
2021-09-08 15:57 ` [PATCH 3/6] afs: Add missing vnode validation checks David Howells
2021-09-08 15:58 ` David Howells [this message]
2021-09-08 15:58 ` [PATCH 5/6] afs: Fix mmap coherency vs 3rd-party changes David Howells
2021-09-08 15:58 ` [PATCH 6/6] afs: Try to avoid taking RCU read lock when checking vnode validity David Howells
2021-09-09  7:40 ` [PATCH 0/6] afs: Fixes for 3rd party-induced data corruption markus.suvanto
2021-09-10 21:19 ` [PATCH 7/6] afs: Fix corruption in reads at fpos 2G-4G from an OpenAFS server David Howells

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=163111668100.283156.3851669884664475428.stgit@warthog.procyon.org.uk \
    --to=dhowells@redhat.com \
    --cc=linux-afs@lists.infradead.org \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=marc.dionne@auristor.com \
    --cc=markus.suvanto@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.