[PATCH 4.14 041/166] CIFS: Properly process SMB3 lease breaks

From: Greg Kroah-Hartman
Date: Tue Sep 29 2020 - 08:41:06 EST

Next message: Greg Kroah-Hartman: "[PATCH 4.14 045/166] seqlock: Require WRITE_ONCE surrounding raw_seqcount_barrier"
Previous message: Greg Kroah-Hartman: "[PATCH 4.14 042/166] kernel/sys.c: avoid copying possible padding bytes in copy_to_user"
In reply to: Greg Kroah-Hartman: "[PATCH 4.14 042/166] kernel/sys.c: avoid copying possible padding bytes in copy_to_user"
Next in thread: Greg Kroah-Hartman: "[PATCH 4.14 045/166] seqlock: Require WRITE_ONCE surrounding raw_seqcount_barrier"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

From: Pavel Shilovsky <pshilov@xxxxxxxxxxxxx>

[ Upstream commit 9bd4540836684013aaad6070a65d6fcdd9006625 ]

Currenly we doesn't assume that a server may break a lease
from RWH to RW which causes us setting a wrong lease state
on a file and thus mistakenly flushing data and byte-range
locks and purging cached data on the client. This leads to
performance degradation because subsequent IOs go directly
to the server.

Fix this by propagating new lease state and epoch values
to the oplock break handler through cifsFileInfo structure
and removing the use of cifsInodeInfo flags for that. It
allows to avoid some races of several lease/oplock breaks
using those flags in parallel.

Signed-off-by: Pavel Shilovsky <pshilov@xxxxxxxxxxxxx>
Signed-off-by: Steve French <stfrench@xxxxxxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
fs/cifs/cifsglob.h | 9 ++++++---
fs/cifs/file.c | 10 +++++++---
fs/cifs/misc.c | 17 +++--------------
fs/cifs/smb1ops.c | 8 +++-----
fs/cifs/smb2misc.c | 32 +++++++-------------------------
fs/cifs/smb2ops.c | 44 ++++++++++++++++++++++++++++++--------------
fs/cifs/smb2pdu.h | 2 +-
7 files changed, 57 insertions(+), 65 deletions(-)

diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h
index 600bb838c15b8..f166fcb48ac0e 100644
--- a/fs/cifs/cifsglob.h
+++ b/fs/cifs/cifsglob.h
@@ -246,8 +246,9 @@ struct smb_version_operations {
int (*check_message)(char *, unsigned int, struct TCP_Server_Info *);
bool (*is_oplock_break)(char *, struct TCP_Server_Info *);
int (*handle_cancelled_mid)(char *, struct TCP_Server_Info *);
- void (*downgrade_oplock)(struct TCP_Server_Info *,
- struct cifsInodeInfo *, bool);
+ void (*downgrade_oplock)(struct TCP_Server_Info *server,
+ struct cifsInodeInfo *cinode, __u32 oplock,
+ unsigned int epoch, bool *purge_cache);
/* process transaction2 response */
bool (*check_trans2)(struct mid_q_entry *, struct TCP_Server_Info *,
char *, int);
@@ -1092,6 +1093,8 @@ struct cifsFileInfo {
unsigned int f_flags;
bool invalidHandle:1; /* file closed via session abend */
bool oplock_break_cancelled:1;
+ unsigned int oplock_epoch; /* epoch from the lease break */
+ __u32 oplock_level; /* oplock/lease level from the lease break */
int count;
spinlock_t file_info_lock; /* protects four flag/count fields above */
struct mutex fh_mutex; /* prevents reopen race after dead ses*/
@@ -1223,7 +1226,7 @@ struct cifsInodeInfo {
unsigned int epoch; /* used to track lease state changes */
#define CIFS_INODE_PENDING_OPLOCK_BREAK (0) /* oplock break in progress */
#define CIFS_INODE_PENDING_WRITERS (1) /* Writes in progress */
-#define CIFS_INODE_DOWNGRADE_OPLOCK_TO_L2 (2) /* Downgrade oplock to L2 */
+#define CIFS_INODE_FLAG_UNUSED (2) /* Unused flag */
#define CIFS_INO_DELETE_PENDING (3) /* delete pending on server */
#define CIFS_INO_INVALID_MAPPING (4) /* pagecache is invalid */
#define CIFS_INO_LOCK (5) /* lock bit for synchronization */
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index 72e7cbfb325a6..0981731132ec0 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -4135,12 +4135,13 @@ void cifs_oplock_break(struct work_struct *work)
struct cifs_tcon *tcon = tlink_tcon(cfile->tlink);
struct TCP_Server_Info *server = tcon->ses->server;
int rc = 0;
+ bool purge_cache = false;

wait_on_bit(&cinode->flags, CIFS_INODE_PENDING_WRITERS,
TASK_UNINTERRUPTIBLE);

- server->ops->downgrade_oplock(server, cinode,
- test_bit(CIFS_INODE_DOWNGRADE_OPLOCK_TO_L2, &cinode->flags));
+ server->ops->downgrade_oplock(server, cinode, cfile->oplock_level,
+ cfile->oplock_epoch, &purge_cache);

if (!CIFS_CACHE_WRITE(cinode) && CIFS_CACHE_READ(cinode) &&
cifs_has_mand_locks(cinode)) {
@@ -4155,18 +4156,21 @@ void cifs_oplock_break(struct work_struct *work)
else
break_lease(inode, O_WRONLY);
rc = filemap_fdatawrite(inode->i_mapping);
- if (!CIFS_CACHE_READ(cinode)) {
+ if (!CIFS_CACHE_READ(cinode) || purge_cache) {
rc = filemap_fdatawait(inode->i_mapping);
mapping_set_error(inode->i_mapping, rc);
cifs_zap_mapping(inode);
}
cifs_dbg(FYI, "Oplock flush inode %p rc %d\n", inode, rc);
+ if (CIFS_CACHE_WRITE(cinode))
+ goto oplock_break_ack;
}

rc = cifs_push_locks(cfile);
if (rc)
cifs_dbg(VFS, "Push locks rc = %d\n", rc);

+oplock_break_ack:
/*
* releasing stale oplock after recent reconnect of smb session using
* a now incorrect file handle is not a data integrity issue but do
diff --git a/fs/cifs/misc.c b/fs/cifs/misc.c
index 76f1649ab444f..d0e024856c0d4 100644
--- a/fs/cifs/misc.c
+++ b/fs/cifs/misc.c
@@ -473,21 +473,10 @@ is_valid_oplock_break(char *buffer, struct TCP_Server_Info *srv)
set_bit(CIFS_INODE_PENDING_OPLOCK_BREAK,
&pCifsInode->flags);

- /*
- * Set flag if the server downgrades the oplock
- * to L2 else clear.
- */
- if (pSMB->OplockLevel)
- set_bit(
- CIFS_INODE_DOWNGRADE_OPLOCK_TO_L2,
- &pCifsInode->flags);
- else
- clear_bit(
- CIFS_INODE_DOWNGRADE_OPLOCK_TO_L2,
- &pCifsInode->flags);
-
- cifs_queue_oplock_break(netfile);
+ netfile->oplock_epoch = 0;
+ netfile->oplock_level = pSMB->OplockLevel;
netfile->oplock_break_cancelled = false;
+ cifs_queue_oplock_break(netfile);

spin_unlock(&tcon->open_file_lock);
spin_unlock(&cifs_tcp_ses_lock);
diff --git a/fs/cifs/smb1ops.c b/fs/cifs/smb1ops.c
index 483458340b10c..9b271ae641c19 100644
--- a/fs/cifs/smb1ops.c
+++ b/fs/cifs/smb1ops.c
@@ -379,12 +379,10 @@ coalesce_t2(char *second_buf, struct smb_hdr *target_hdr)

static void
cifs_downgrade_oplock(struct TCP_Server_Info *server,
- struct cifsInodeInfo *cinode, bool set_level2)
+ struct cifsInodeInfo *cinode, __u32 oplock,
+ unsigned int epoch, bool *purge_cache)
{
- if (set_level2)
- cifs_set_oplock_level(cinode, OPLOCK_READ);
- else
- cifs_set_oplock_level(cinode, 0);
+ cifs_set_oplock_level(cinode, oplock);
}

static bool
diff --git a/fs/cifs/smb2misc.c b/fs/cifs/smb2misc.c
index ff2ad15f67d63..0c6e5450ff765 100644
--- a/fs/cifs/smb2misc.c
+++ b/fs/cifs/smb2misc.c
@@ -496,7 +496,7 @@ smb2_tcon_has_lease(struct cifs_tcon *tcon, struct smb2_lease_break *rsp,

cifs_dbg(FYI, "found in the open list\n");
cifs_dbg(FYI, "lease key match, lease break 0x%x\n",
- le32_to_cpu(rsp->NewLeaseState));
+ lease_state);

if (ack_req)
cfile->oplock_break_cancelled = false;
@@ -505,17 +505,8 @@ smb2_tcon_has_lease(struct cifs_tcon *tcon, struct smb2_lease_break *rsp,

set_bit(CIFS_INODE_PENDING_OPLOCK_BREAK, &cinode->flags);

- /*
- * Set or clear flags depending on the lease state being READ.
- * HANDLE caching flag should be added when the client starts
- * to defer closing remote file handles with HANDLE leases.
- */
- if (lease_state & SMB2_LEASE_READ_CACHING_HE)
- set_bit(CIFS_INODE_DOWNGRADE_OPLOCK_TO_L2,
- &cinode->flags);
- else
- clear_bit(CIFS_INODE_DOWNGRADE_OPLOCK_TO_L2,
- &cinode->flags);
+ cfile->oplock_epoch = le16_to_cpu(rsp->Epoch);
+ cfile->oplock_level = lease_state;

cifs_queue_oplock_break(cfile);
kfree(lw);
@@ -538,7 +529,7 @@ smb2_tcon_has_lease(struct cifs_tcon *tcon, struct smb2_lease_break *rsp,

cifs_dbg(FYI, "found in the pending open list\n");
cifs_dbg(FYI, "lease key match, lease break 0x%x\n",
- le32_to_cpu(rsp->NewLeaseState));
+ lease_state);

open->oplock = lease_state;
}
@@ -650,18 +641,9 @@ smb2_is_valid_oplock_break(char *buffer, struct TCP_Server_Info *server)
set_bit(CIFS_INODE_PENDING_OPLOCK_BREAK,
&cinode->flags);

- /*
- * Set flag if the server downgrades the oplock
- * to L2 else clear.
- */
- if (rsp->OplockLevel)
- set_bit(
- CIFS_INODE_DOWNGRADE_OPLOCK_TO_L2,
- &cinode->flags);
- else
- clear_bit(
- CIFS_INODE_DOWNGRADE_OPLOCK_TO_L2,
- &cinode->flags);
+ cfile->oplock_epoch = 0;
+ cfile->oplock_level = rsp->OplockLevel;
+
spin_unlock(&cfile->file_info_lock);

cifs_queue_oplock_break(cfile);
diff --git a/fs/cifs/smb2ops.c b/fs/cifs/smb2ops.c
index b46fdb2b8d349..90d4288907a61 100644
--- a/fs/cifs/smb2ops.c
+++ b/fs/cifs/smb2ops.c
@@ -1935,22 +1935,38 @@ static long smb3_fallocate(struct file *file, struct cifs_tcon *tcon, int mode,

static void
smb2_downgrade_oplock(struct TCP_Server_Info *server,
- struct cifsInodeInfo *cinode, bool set_level2)
+ struct cifsInodeInfo *cinode, __u32 oplock,
+ unsigned int epoch, bool *purge_cache)
{
- if (set_level2)
- server->ops->set_oplock_level(cinode, SMB2_OPLOCK_LEVEL_II,
- 0, NULL);
- else
- server->ops->set_oplock_level(cinode, 0, 0, NULL);
+ server->ops->set_oplock_level(cinode, oplock, 0, NULL);
}

static void
-smb21_downgrade_oplock(struct TCP_Server_Info *server,
- struct cifsInodeInfo *cinode, bool set_level2)
+smb21_set_oplock_level(struct cifsInodeInfo *cinode, __u32 oplock,
+ unsigned int epoch, bool *purge_cache);
+
+static void
+smb3_downgrade_oplock(struct TCP_Server_Info *server,
+ struct cifsInodeInfo *cinode, __u32 oplock,
+ unsigned int epoch, bool *purge_cache)
{
- server->ops->set_oplock_level(cinode,
- set_level2 ? SMB2_LEASE_READ_CACHING_HE :
- 0, 0, NULL);
+ unsigned int old_state = cinode->oplock;
+ unsigned int old_epoch = cinode->epoch;
+ unsigned int new_state;
+
+ if (epoch > old_epoch) {
+ smb21_set_oplock_level(cinode, oplock, 0, NULL);
+ cinode->epoch = epoch;
+ }
+
+ new_state = cinode->oplock;
+ *purge_cache = false;
+
+ if ((old_state & CIFS_CACHE_READ_FLG) != 0 &&
+ (new_state & CIFS_CACHE_READ_FLG) == 0)
+ *purge_cache = true;
+ else if (old_state == new_state && (epoch - old_epoch > 1))
+ *purge_cache = true;
}

static void
@@ -2953,7 +2969,7 @@ struct smb_version_operations smb21_operations = {
.print_stats = smb2_print_stats,
.is_oplock_break = smb2_is_valid_oplock_break,
.handle_cancelled_mid = smb2_handle_cancelled_mid,
- .downgrade_oplock = smb21_downgrade_oplock,
+ .downgrade_oplock = smb2_downgrade_oplock,
.need_neg = smb2_need_neg,
.negotiate = smb2_negotiate,
.negotiate_wsize = smb2_negotiate_wsize,
@@ -3048,7 +3064,7 @@ struct smb_version_operations smb30_operations = {
.dump_share_caps = smb2_dump_share_caps,
.is_oplock_break = smb2_is_valid_oplock_break,
.handle_cancelled_mid = smb2_handle_cancelled_mid,
- .downgrade_oplock = smb21_downgrade_oplock,
+ .downgrade_oplock = smb3_downgrade_oplock,
.need_neg = smb2_need_neg,
.negotiate = smb2_negotiate,
.negotiate_wsize = smb2_negotiate_wsize,
@@ -3153,7 +3169,7 @@ struct smb_version_operations smb311_operations = {
.dump_share_caps = smb2_dump_share_caps,
.is_oplock_break = smb2_is_valid_oplock_break,
.handle_cancelled_mid = smb2_handle_cancelled_mid,
- .downgrade_oplock = smb21_downgrade_oplock,
+ .downgrade_oplock = smb3_downgrade_oplock,
.need_neg = smb2_need_neg,
.negotiate = smb2_negotiate,
.negotiate_wsize = smb2_negotiate_wsize,
diff --git a/fs/cifs/smb2pdu.h b/fs/cifs/smb2pdu.h
index bad458a2b579e..f8baa54c83008 100644
--- a/fs/cifs/smb2pdu.h
+++ b/fs/cifs/smb2pdu.h
@@ -1046,7 +1046,7 @@ struct smb2_oplock_break {
struct smb2_lease_break {
struct smb2_hdr hdr;
__le16 StructureSize; /* Must be 44 */
- __le16 Reserved;
+ __le16 Epoch;
__le32 Flags;
__u8 LeaseKey[16];
__le32 CurrentLeaseState;
--
2.25.1

Next message: Greg Kroah-Hartman: "[PATCH 4.14 045/166] seqlock: Require WRITE_ONCE surrounding raw_seqcount_barrier"
Previous message: Greg Kroah-Hartman: "[PATCH 4.14 042/166] kernel/sys.c: avoid copying possible padding bytes in copy_to_user"
In reply to: Greg Kroah-Hartman: "[PATCH 4.14 042/166] kernel/sys.c: avoid copying possible padding bytes in copy_to_user"
Next in thread: Greg Kroah-Hartman: "[PATCH 4.14 045/166] seqlock: Require WRITE_ONCE surrounding raw_seqcount_barrier"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]