linux-kernel - [PATCH 04/27] afs: Always include dir in bulk status fetch from afs_do

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 29 May 2020 23:00:35 +0100
From:   David Howells <dhowells@...hat.com>
To:     linux-afs@...ts.infradead.org
Cc:     Dave Botsch <botsch@....cornell.edu>,
        Jeffrey Altman <jaltman@...istor.com>, dhowells@...hat.com,
        linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: [PATCH 04/27] afs: Always include dir in bulk status fetch from
 afs_do_lookup()

When a lookup is done in an AFS directory, the filesystem will speculate
and fetch up to 49 other statuses for files in the same directory and fetch
those as well, turning them into inodes or updating inodes that already
exist.

However, occasionally, a callback break might go missing due to NAT timing
out, but the afs filesystem doesn't then realise that the directory is not
up to date.

Alleviate this by using one of the status slots to check the directory in
which the lookup is being done.

Reported-by: Dave Botsch <botsch@....cornell.edu>
Suggested-by: Jeffrey Altman <jaltman@...istor.com>
Signed-off-by: David Howells <dhowells@...hat.com>
---

 fs/afs/dir.c |    9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/fs/afs/dir.c b/fs/afs/dir.c
index d1e1caa23c8b..3c486340b220 100644
--- a/fs/afs/dir.c
+++ b/fs/afs/dir.c
@@ -658,7 +658,8 @@ static struct inode *afs_do_lookup(struct inode *dir, struct dentry *dentry,
 
 	cookie->ctx.actor = afs_lookup_filldir;
 	cookie->name = dentry->d_name;
-	cookie->nr_fids = 1; /* slot 0 is saved for the fid we actually want */
+	cookie->nr_fids = 2; /* slot 0 is saved for the fid we actually want
+			      * and slot 1 for the directory */
 
 	read_seqlock_excl(&dvnode->cb_lock);
 	dcbi = rcu_dereference_protected(dvnode->cb_interest,
@@ -709,7 +710,11 @@ static struct inode *afs_do_lookup(struct inode *dir, struct dentry *dentry,
 	if (!cookie->inodes)
 		goto out_s;
 
-	for (i = 1; i < cookie->nr_fids; i++) {
+	cookie->fids[1] = dvnode->fid;
+	cookie->statuses[1].cb_break = afs_calc_vnode_cb_break(dvnode);
+	cookie->inodes[1] = igrab(&dvnode->vfs_inode);
+
+	for (i = 2; i < cookie->nr_fids; i++) {
 		scb = &cookie->statuses[i];
 
 		/* Find any inodes that already exist and get their