linux-ext4 - [PATCH 2/2] ext4: use raw i_version value for ea

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]

Message-Id: <20180426112027.11258-3-guaneryu@gmail.com>
Date:   Thu, 26 Apr 2018 19:20:27 +0800
From:   Eryu Guan <guaneryu@...il.com>
To:     linux-ext4@...r.kernel.org
Cc:     "Darrick J. Wong" <darrick.wong@...cle.com>,
        Theodore Ts'o <tytso@....edu>,
        Tahsin Erdogan <tahsin@...gle.com>,
        Eryu Guan <guaneryu@...il.com>
Subject: [PATCH 2/2] ext4: use raw i_version value for ea_inode

Currently, creating large xattr (e.g. 2k) in ea_inode would cause
ea_inode refcount corruption, e.g.

  Pass 4: Checking reference counts
  Extended attribute inode 13 ref count is 0, should be 1. Fix? no

This is because that we save the lower 32bit of refcount in
inode->i_version and store it in raw_inode->i_disk_version on disk.
But since commit ee73f9a52a34 ("ext4: convert to new i_version
API"), we load/store modified i_disk_version from/to disk instead of
raw value, which causes on-disk ea_inode refcount corruption.

Fix it by loading/storing raw i_version/i_disk_version, because it's
a self-managed value in this case.

Fixes: ee73f9a52a34 ("ext4: convert to new i_version API")
Cc: Tahsin Erdogan <tahsin@...gle.com>
Signed-off-by: Eryu Guan <guaneryu@...il.com>
---
 fs/ext4/inode.c | 24 ++++++++++++++++++++++--
 1 file changed, 22 insertions(+), 2 deletions(-)

diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 1e50c5efae67..0eb64e8f9602 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -4724,6 +4724,26 @@ int ext4_get_projid(struct inode *inode, kprojid_t *projid)
 	return 0;
 }
 
+/*
+ * ext4 has self-managed i_version for ea inodes, it stores the lower 32bit of
+ * refcount in i_version, so use raw values if inode has EXT4_EA_INODE_FL flag
+ * set.
+ */
+static inline void ext4_inode_set_iversion_queried(struct inode *inode, u64 val)
+{
+	if (unlikely(EXT4_I(inode)->i_flags & EXT4_EA_INODE_FL))
+		inode_set_iversion_raw(inode, val);
+	else
+		inode_set_iversion_queried(inode, val);
+}
+static inline u64 ext4_inode_peek_iversion(const struct inode *inode)
+{
+	if (unlikely(EXT4_I(inode)->i_flags & EXT4_EA_INODE_FL))
+		return inode_peek_iversion_raw(inode);
+	else
+		return inode_peek_iversion(inode);
+}
+
 struct inode *ext4_iget(struct super_block *sb, unsigned long ino)
 {
 	struct ext4_iloc iloc;
@@ -4910,7 +4930,7 @@ struct inode *ext4_iget(struct super_block *sb, unsigned long ino)
 				ivers |=
 		    (__u64)(le32_to_cpu(raw_inode->i_version_hi)) << 32;
 		}
-		inode_set_iversion_queried(inode, ivers);
+		ext4_inode_set_iversion_queried(inode, ivers);
 	}
 
 	ret = 0;
@@ -5196,7 +5216,7 @@ static int ext4_do_update_inode(handle_t *handle,
 	}
 
 	if (likely(!test_opt2(inode->i_sb, HURD_COMPAT))) {
-		u64 ivers = inode_peek_iversion(inode);
+		u64 ivers = ext4_inode_peek_iversion(inode);
 
 		raw_inode->i_disk_version = cpu_to_le32(ivers);
 		if (ei->i_extra_isize) {
-- 
2.14.3