Commit c11f1df5 authored by Sachin Prabhu's avatar Sachin Prabhu Committed by Steve French

cifs: Wait for writebacks to complete before attempting write.

Problem reported in Red Hat bz 1040329 for strict writes where we cache
only when we hold oplock and write direct to the server when we don't.

When we receive an oplock break, we first change the oplock value for
the inode in cifsInodeInfo->oplock to indicate that we no longer hold
the oplock before we enqueue a task to flush changes to the backing
device. Once we have completed flushing the changes, we return the
oplock to the server.

There are 2 ways here where we can have data corruption
1) While we flush changes to the backing device as part of the oplock
break, we can have processes write to the file. These writes check for
the oplock, find none and attempt to write directly to the server.
These direct writes made while we are flushing from cache could be
overwritten by data being flushed from the cache causing data
corruption.
2) While a thread runs in cifs_strict_writev, the machine could receive
and process an oplock break after the thread has checked the oplock and
found that it allows us to cache and before we have made changes to the
cache. In that case, we end up with a dirty page in cache when we
shouldn't have any. This will be flushed later and will overwrite all
subsequent writes to the part of the file represented by this page.

Before making any writes to the server, we need to confirm that we are
not in the process of flushing data to the server and if we are, we
should wait until the process is complete before we attempt the write.
We should also wait for existing writes to complete before we process
an oplock break request which changes oplock values.

We add a version specific  downgrade_oplock() operation to allow for
differences in the oplock values set for the different smb versions.

Cc: stable@vger.kernel.org
Signed-off-by: default avatarSachin Prabhu <sprabhu@redhat.com>
Reviewed-by: default avatarJeff Layton <jlayton@redhat.com>
Reviewed-by: default avatarPavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: default avatarSteve French <smfrench@gmail.com>
parent 0f689a33
...@@ -253,6 +253,11 @@ cifs_alloc_inode(struct super_block *sb) ...@@ -253,6 +253,11 @@ cifs_alloc_inode(struct super_block *sb)
cifs_set_oplock_level(cifs_inode, 0); cifs_set_oplock_level(cifs_inode, 0);
cifs_inode->delete_pending = false; cifs_inode->delete_pending = false;
cifs_inode->invalid_mapping = false; cifs_inode->invalid_mapping = false;
clear_bit(CIFS_INODE_PENDING_OPLOCK_BREAK, &cifs_inode->flags);
clear_bit(CIFS_INODE_PENDING_WRITERS, &cifs_inode->flags);
clear_bit(CIFS_INODE_DOWNGRADE_OPLOCK_TO_L2, &cifs_inode->flags);
spin_lock_init(&cifs_inode->writers_lock);
cifs_inode->writers = 0;
cifs_inode->vfs_inode.i_blkbits = 14; /* 2**14 = CIFS_MAX_MSGSIZE */ cifs_inode->vfs_inode.i_blkbits = 14; /* 2**14 = CIFS_MAX_MSGSIZE */
cifs_inode->server_eof = 0; cifs_inode->server_eof = 0;
cifs_inode->uniqueid = 0; cifs_inode->uniqueid = 0;
...@@ -732,19 +737,26 @@ static ssize_t cifs_file_aio_write(struct kiocb *iocb, const struct iovec *iov, ...@@ -732,19 +737,26 @@ static ssize_t cifs_file_aio_write(struct kiocb *iocb, const struct iovec *iov,
unsigned long nr_segs, loff_t pos) unsigned long nr_segs, loff_t pos)
{ {
struct inode *inode = file_inode(iocb->ki_filp); struct inode *inode = file_inode(iocb->ki_filp);
struct cifsInodeInfo *cinode = CIFS_I(inode);
ssize_t written; ssize_t written;
int rc; int rc;
written = cifs_get_writer(cinode);
if (written)
return written;
written = generic_file_aio_write(iocb, iov, nr_segs, pos); written = generic_file_aio_write(iocb, iov, nr_segs, pos);
if (CIFS_CACHE_WRITE(CIFS_I(inode))) if (CIFS_CACHE_WRITE(CIFS_I(inode)))
return written; goto out;
rc = filemap_fdatawrite(inode->i_mapping); rc = filemap_fdatawrite(inode->i_mapping);
if (rc) if (rc)
cifs_dbg(FYI, "cifs_file_aio_write: %d rc on %p inode\n", cifs_dbg(FYI, "cifs_file_aio_write: %d rc on %p inode\n",
rc, inode); rc, inode);
out:
cifs_put_writer(cinode);
return written; return written;
} }
......
...@@ -228,6 +228,8 @@ struct smb_version_operations { ...@@ -228,6 +228,8 @@ struct smb_version_operations {
/* verify the message */ /* verify the message */
int (*check_message)(char *, unsigned int); int (*check_message)(char *, unsigned int);
bool (*is_oplock_break)(char *, struct TCP_Server_Info *); bool (*is_oplock_break)(char *, struct TCP_Server_Info *);
void (*downgrade_oplock)(struct TCP_Server_Info *,
struct cifsInodeInfo *, bool);
/* process transaction2 response */ /* process transaction2 response */
bool (*check_trans2)(struct mid_q_entry *, struct TCP_Server_Info *, bool (*check_trans2)(struct mid_q_entry *, struct TCP_Server_Info *,
char *, int); char *, int);
...@@ -1113,6 +1115,12 @@ struct cifsInodeInfo { ...@@ -1113,6 +1115,12 @@ struct cifsInodeInfo {
unsigned int epoch; /* used to track lease state changes */ unsigned int epoch; /* used to track lease state changes */
bool delete_pending; /* DELETE_ON_CLOSE is set */ bool delete_pending; /* DELETE_ON_CLOSE is set */
bool invalid_mapping; /* pagecache is invalid */ bool invalid_mapping; /* pagecache is invalid */
unsigned long flags;
#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 */
spinlock_t writers_lock;
unsigned int writers; /* Number of writers on this inode */
unsigned long time; /* jiffies of last update of inode */ unsigned long time; /* jiffies of last update of inode */
u64 server_eof; /* current file size on server -- protected by i_lock */ u64 server_eof; /* current file size on server -- protected by i_lock */
u64 uniqueid; /* server inode number */ u64 uniqueid; /* server inode number */
......
...@@ -127,6 +127,9 @@ extern u64 cifs_UnixTimeToNT(struct timespec); ...@@ -127,6 +127,9 @@ extern u64 cifs_UnixTimeToNT(struct timespec);
extern struct timespec cnvrtDosUnixTm(__le16 le_date, __le16 le_time, extern struct timespec cnvrtDosUnixTm(__le16 le_date, __le16 le_time,
int offset); int offset);
extern void cifs_set_oplock_level(struct cifsInodeInfo *cinode, __u32 oplock); extern void cifs_set_oplock_level(struct cifsInodeInfo *cinode, __u32 oplock);
extern int cifs_get_writer(struct cifsInodeInfo *cinode);
extern void cifs_put_writer(struct cifsInodeInfo *cinode);
extern void cifs_done_oplock_break(struct cifsInodeInfo *cinode);
extern int cifs_unlock_range(struct cifsFileInfo *cfile, extern int cifs_unlock_range(struct cifsFileInfo *cfile,
struct file_lock *flock, const unsigned int xid); struct file_lock *flock, const unsigned int xid);
extern int cifs_push_mandatory_locks(struct cifsFileInfo *cfile); extern int cifs_push_mandatory_locks(struct cifsFileInfo *cfile);
......
...@@ -2621,12 +2621,20 @@ cifs_strict_writev(struct kiocb *iocb, const struct iovec *iov, ...@@ -2621,12 +2621,20 @@ cifs_strict_writev(struct kiocb *iocb, const struct iovec *iov,
struct cifs_tcon *tcon = tlink_tcon(cfile->tlink); struct cifs_tcon *tcon = tlink_tcon(cfile->tlink);
ssize_t written; ssize_t written;
written = cifs_get_writer(cinode);
if (written)
return written;
if (CIFS_CACHE_WRITE(cinode)) { if (CIFS_CACHE_WRITE(cinode)) {
if (cap_unix(tcon->ses) && if (cap_unix(tcon->ses) &&
(CIFS_UNIX_FCNTL_CAP & le64_to_cpu(tcon->fsUnixInfo.Capability)) (CIFS_UNIX_FCNTL_CAP & le64_to_cpu(tcon->fsUnixInfo.Capability))
&& ((cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NOPOSIXBRL) == 0)) && ((cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NOPOSIXBRL) == 0)) {
return generic_file_aio_write(iocb, iov, nr_segs, pos); written = generic_file_aio_write(
return cifs_writev(iocb, iov, nr_segs, pos); iocb, iov, nr_segs, pos);
goto out;
}
written = cifs_writev(iocb, iov, nr_segs, pos);
goto out;
} }
/* /*
* For non-oplocked files in strict cache mode we need to write the data * For non-oplocked files in strict cache mode we need to write the data
...@@ -2646,6 +2654,8 @@ cifs_strict_writev(struct kiocb *iocb, const struct iovec *iov, ...@@ -2646,6 +2654,8 @@ cifs_strict_writev(struct kiocb *iocb, const struct iovec *iov,
inode); inode);
cinode->oplock = 0; cinode->oplock = 0;
} }
out:
cifs_put_writer(cinode);
return written; return written;
} }
...@@ -3621,6 +3631,13 @@ static int cifs_launder_page(struct page *page) ...@@ -3621,6 +3631,13 @@ static int cifs_launder_page(struct page *page)
return rc; return rc;
} }
static int
cifs_pending_writers_wait(void *unused)
{
schedule();
return 0;
}
void cifs_oplock_break(struct work_struct *work) void cifs_oplock_break(struct work_struct *work)
{ {
struct cifsFileInfo *cfile = container_of(work, struct cifsFileInfo, struct cifsFileInfo *cfile = container_of(work, struct cifsFileInfo,
...@@ -3628,8 +3645,15 @@ void cifs_oplock_break(struct work_struct *work) ...@@ -3628,8 +3645,15 @@ void cifs_oplock_break(struct work_struct *work)
struct inode *inode = cfile->dentry->d_inode; struct inode *inode = cfile->dentry->d_inode;
struct cifsInodeInfo *cinode = CIFS_I(inode); struct cifsInodeInfo *cinode = CIFS_I(inode);
struct cifs_tcon *tcon = tlink_tcon(cfile->tlink); struct cifs_tcon *tcon = tlink_tcon(cfile->tlink);
struct TCP_Server_Info *server = tcon->ses->server;
int rc = 0; int rc = 0;
wait_on_bit(&cinode->flags, CIFS_INODE_PENDING_WRITERS,
cifs_pending_writers_wait, TASK_UNINTERRUPTIBLE);
server->ops->downgrade_oplock(server, cinode,
test_bit(CIFS_INODE_DOWNGRADE_OPLOCK_TO_L2, &cinode->flags));
if (!CIFS_CACHE_WRITE(cinode) && CIFS_CACHE_READ(cinode) && if (!CIFS_CACHE_WRITE(cinode) && CIFS_CACHE_READ(cinode) &&
cifs_has_mand_locks(cinode)) { cifs_has_mand_locks(cinode)) {
cifs_dbg(FYI, "Reset oplock to None for inode=%p due to mand locks\n", cifs_dbg(FYI, "Reset oplock to None for inode=%p due to mand locks\n",
...@@ -3666,6 +3690,7 @@ void cifs_oplock_break(struct work_struct *work) ...@@ -3666,6 +3690,7 @@ void cifs_oplock_break(struct work_struct *work)
cinode); cinode);
cifs_dbg(FYI, "Oplock release rc = %d\n", rc); cifs_dbg(FYI, "Oplock release rc = %d\n", rc);
} }
cifs_done_oplock_break(cinode);
} }
/* /*
......
...@@ -466,8 +466,22 @@ is_valid_oplock_break(char *buffer, struct TCP_Server_Info *srv) ...@@ -466,8 +466,22 @@ is_valid_oplock_break(char *buffer, struct TCP_Server_Info *srv)
cifs_dbg(FYI, "file id match, oplock break\n"); cifs_dbg(FYI, "file id match, oplock break\n");
pCifsInode = CIFS_I(netfile->dentry->d_inode); pCifsInode = CIFS_I(netfile->dentry->d_inode);
cifs_set_oplock_level(pCifsInode, set_bit(CIFS_INODE_PENDING_OPLOCK_BREAK,
pSMB->OplockLevel ? OPLOCK_READ : 0); &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);
queue_work(cifsiod_wq, queue_work(cifsiod_wq,
&netfile->oplock_break); &netfile->oplock_break);
netfile->oplock_break_cancelled = false; netfile->oplock_break_cancelled = false;
...@@ -551,6 +565,62 @@ void cifs_set_oplock_level(struct cifsInodeInfo *cinode, __u32 oplock) ...@@ -551,6 +565,62 @@ void cifs_set_oplock_level(struct cifsInodeInfo *cinode, __u32 oplock)
cinode->oplock = 0; cinode->oplock = 0;
} }
static int
cifs_oplock_break_wait(void *unused)
{
schedule();
return signal_pending(current) ? -ERESTARTSYS : 0;
}
/*
* We wait for oplock breaks to be processed before we attempt to perform
* writes.
*/
int cifs_get_writer(struct cifsInodeInfo *cinode)
{
int rc;
start:
rc = wait_on_bit(&cinode->flags, CIFS_INODE_PENDING_OPLOCK_BREAK,
cifs_oplock_break_wait, TASK_KILLABLE);
if (rc)
return rc;
spin_lock(&cinode->writers_lock);
if (!cinode->writers)
set_bit(CIFS_INODE_PENDING_WRITERS, &cinode->flags);
cinode->writers++;
/* Check to see if we have started servicing an oplock break */
if (test_bit(CIFS_INODE_PENDING_OPLOCK_BREAK, &cinode->flags)) {
cinode->writers--;
if (cinode->writers == 0) {
clear_bit(CIFS_INODE_PENDING_WRITERS, &cinode->flags);
wake_up_bit(&cinode->flags, CIFS_INODE_PENDING_WRITERS);
}
spin_unlock(&cinode->writers_lock);
goto start;
}
spin_unlock(&cinode->writers_lock);
return 0;
}
void cifs_put_writer(struct cifsInodeInfo *cinode)
{
spin_lock(&cinode->writers_lock);
cinode->writers--;
if (cinode->writers == 0) {
clear_bit(CIFS_INODE_PENDING_WRITERS, &cinode->flags);
wake_up_bit(&cinode->flags, CIFS_INODE_PENDING_WRITERS);
}
spin_unlock(&cinode->writers_lock);
}
void cifs_done_oplock_break(struct cifsInodeInfo *cinode)
{
clear_bit(CIFS_INODE_PENDING_OPLOCK_BREAK, &cinode->flags);
wake_up_bit(&cinode->flags, CIFS_INODE_PENDING_OPLOCK_BREAK);
}
bool bool
backup_cred(struct cifs_sb_info *cifs_sb) backup_cred(struct cifs_sb_info *cifs_sb)
{ {
......
...@@ -372,6 +372,16 @@ coalesce_t2(char *second_buf, struct smb_hdr *target_hdr) ...@@ -372,6 +372,16 @@ coalesce_t2(char *second_buf, struct smb_hdr *target_hdr)
return 0; return 0;
} }
static void
cifs_downgrade_oplock(struct TCP_Server_Info *server,
struct cifsInodeInfo *cinode, bool set_level2)
{
if (set_level2)
cifs_set_oplock_level(cinode, OPLOCK_READ);
else
cifs_set_oplock_level(cinode, 0);
}
static bool static bool
cifs_check_trans2(struct mid_q_entry *mid, struct TCP_Server_Info *server, cifs_check_trans2(struct mid_q_entry *mid, struct TCP_Server_Info *server,
char *buf, int malformed) char *buf, int malformed)
...@@ -1019,6 +1029,7 @@ struct smb_version_operations smb1_operations = { ...@@ -1019,6 +1029,7 @@ struct smb_version_operations smb1_operations = {
.clear_stats = cifs_clear_stats, .clear_stats = cifs_clear_stats,
.print_stats = cifs_print_stats, .print_stats = cifs_print_stats,
.is_oplock_break = is_valid_oplock_break, .is_oplock_break = is_valid_oplock_break,
.downgrade_oplock = cifs_downgrade_oplock,
.check_trans2 = cifs_check_trans2, .check_trans2 = cifs_check_trans2,
.need_neg = cifs_need_neg, .need_neg = cifs_need_neg,
.negotiate = cifs_negotiate, .negotiate = cifs_negotiate,
......
...@@ -575,9 +575,21 @@ smb2_is_valid_oplock_break(char *buffer, struct TCP_Server_Info *server) ...@@ -575,9 +575,21 @@ smb2_is_valid_oplock_break(char *buffer, struct TCP_Server_Info *server)
else else
cfile->oplock_break_cancelled = false; cfile->oplock_break_cancelled = false;
server->ops->set_oplock_level(cinode, set_bit(CIFS_INODE_PENDING_OPLOCK_BREAK,
rsp->OplockLevel ? SMB2_OPLOCK_LEVEL_II : 0, &cinode->flags);
0, NULL);
/*
* 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);
queue_work(cifsiod_wq, &cfile->oplock_break); queue_work(cifsiod_wq, &cfile->oplock_break);
......
...@@ -904,6 +904,17 @@ smb2_query_symlink(const unsigned int xid, struct cifs_tcon *tcon, ...@@ -904,6 +904,17 @@ smb2_query_symlink(const unsigned int xid, struct cifs_tcon *tcon,
return rc; return rc;
} }
static void
smb2_downgrade_oplock(struct TCP_Server_Info *server,
struct cifsInodeInfo *cinode, bool set_level2)
{
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);
}
static void static void
smb2_set_oplock_level(struct cifsInodeInfo *cinode, __u32 oplock, smb2_set_oplock_level(struct cifsInodeInfo *cinode, __u32 oplock,
unsigned int epoch, bool *purge_cache) unsigned int epoch, bool *purge_cache)
...@@ -1110,6 +1121,7 @@ struct smb_version_operations smb20_operations = { ...@@ -1110,6 +1121,7 @@ struct smb_version_operations smb20_operations = {
.clear_stats = smb2_clear_stats, .clear_stats = smb2_clear_stats,
.print_stats = smb2_print_stats, .print_stats = smb2_print_stats,
.is_oplock_break = smb2_is_valid_oplock_break, .is_oplock_break = smb2_is_valid_oplock_break,
.downgrade_oplock = smb2_downgrade_oplock,
.need_neg = smb2_need_neg, .need_neg = smb2_need_neg,
.negotiate = smb2_negotiate, .negotiate = smb2_negotiate,
.negotiate_wsize = smb2_negotiate_wsize, .negotiate_wsize = smb2_negotiate_wsize,
...@@ -1184,6 +1196,7 @@ struct smb_version_operations smb21_operations = { ...@@ -1184,6 +1196,7 @@ struct smb_version_operations smb21_operations = {
.clear_stats = smb2_clear_stats, .clear_stats = smb2_clear_stats,
.print_stats = smb2_print_stats, .print_stats = smb2_print_stats,
.is_oplock_break = smb2_is_valid_oplock_break, .is_oplock_break = smb2_is_valid_oplock_break,
.downgrade_oplock = smb2_downgrade_oplock,
.need_neg = smb2_need_neg, .need_neg = smb2_need_neg,
.negotiate = smb2_negotiate, .negotiate = smb2_negotiate,
.negotiate_wsize = smb2_negotiate_wsize, .negotiate_wsize = smb2_negotiate_wsize,
...@@ -1259,6 +1272,7 @@ struct smb_version_operations smb30_operations = { ...@@ -1259,6 +1272,7 @@ struct smb_version_operations smb30_operations = {
.print_stats = smb2_print_stats, .print_stats = smb2_print_stats,
.dump_share_caps = smb2_dump_share_caps, .dump_share_caps = smb2_dump_share_caps,
.is_oplock_break = smb2_is_valid_oplock_break, .is_oplock_break = smb2_is_valid_oplock_break,
.downgrade_oplock = smb2_downgrade_oplock,
.need_neg = smb2_need_neg, .need_neg = smb2_need_neg,
.negotiate = smb2_negotiate, .negotiate = smb2_negotiate,
.negotiate_wsize = smb2_negotiate_wsize, .negotiate_wsize = smb2_negotiate_wsize,
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment