OpenCloudOS-Kernel/fs/cifs/cifsfs.c

1562 lines
41 KiB
C
Raw Normal View History

/*
* fs/cifs/cifsfs.c
*
* Copyright (C) International Business Machines Corp., 2002,2008
* Author(s): Steve French (sfrench@us.ibm.com)
*
* Common Internet FileSystem (CIFS) client
*
* This library is free software; you can redistribute it and/or modify
* it under the terms of the GNU Lesser General Public License as published
* by the Free Software Foundation; either version 2.1 of the License, or
* (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See
* the GNU Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public License
* along with this library; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*/
/* Note that BB means BUGBUG (ie something to fix eventually) */
#include <linux/module.h>
#include <linux/fs.h>
#include <linux/mount.h>
#include <linux/slab.h>
#include <linux/init.h>
#include <linux/list.h>
#include <linux/seq_file.h>
#include <linux/vfs.h>
#include <linux/mempool.h>
#include <linux/delay.h>
#include <linux/kthread.h>
#include <linux/freezer.h>
#include <linux/namei.h>
#include <linux/random.h>
#include <linux/uuid.h>
#include <linux/xattr.h>
#include <net/ipv6.h>
#include "cifsfs.h"
#include "cifspdu.h"
#define DECLARE_GLOBALS_HERE
#include "cifsglob.h"
#include "cifsproto.h"
#include "cifs_debug.h"
#include "cifs_fs_sb.h"
#include <linux/mm.h>
#include <linux/key-type.h>
#include "cifs_spnego.h"
#include "fscache.h"
#include "smb2pdu.h"
int cifsFYI = 0;
bool traceSMB;
bool enable_oplocks = true;
bool linuxExtEnabled = true;
bool lookupCacheEnabled = true;
bool disable_legacy_dialects; /* false by default */
unsigned int global_secflags = CIFSSEC_DEF;
/* unsigned int ntlmv2_support = 0; */
unsigned int sign_CIFS_PDUs = 1;
static const struct super_operations cifs_super_ops;
unsigned int CIFSMaxBufSize = CIFS_MAX_MSGSIZE;
module_param(CIFSMaxBufSize, uint, 0444);
MODULE_PARM_DESC(CIFSMaxBufSize, "Network buffer size (not including header) "
"for CIFS requests. "
"Default: 16384 Range: 8192 to 130048");
unsigned int cifs_min_rcv = CIFS_MIN_RCV_POOL;
module_param(cifs_min_rcv, uint, 0444);
MODULE_PARM_DESC(cifs_min_rcv, "Network buffers in pool. Default: 4 Range: "
"1 to 64");
unsigned int cifs_min_small = 30;
module_param(cifs_min_small, uint, 0444);
MODULE_PARM_DESC(cifs_min_small, "Small network buffers in pool. Default: 30 "
"Range: 2 to 256");
unsigned int cifs_max_pending = CIFS_MAX_REQ;
module_param(cifs_max_pending, uint, 0444);
MODULE_PARM_DESC(cifs_max_pending, "Simultaneous requests to server for "
"CIFS/SMB1 dialect (N/A for SMB3) "
"Default: 32767 Range: 2 to 32767.");
module_param(enable_oplocks, bool, 0644);
MODULE_PARM_DESC(enable_oplocks, "Enable or disable oplocks. Default: y/Y/1");
module_param(disable_legacy_dialects, bool, 0644);
MODULE_PARM_DESC(disable_legacy_dialects, "To improve security it may be "
"helpful to restrict the ability to "
"override the default dialects (SMB2.1, "
"SMB3 and SMB3.02) on mount with old "
"dialects (CIFS/SMB1 and SMB2) since "
"vers=1.0 (CIFS/SMB1) and vers=2.0 are weaker"
" and less secure. Default: n/N/0");
extern mempool_t *cifs_sm_req_poolp;
extern mempool_t *cifs_req_poolp;
extern mempool_t *cifs_mid_poolp;
struct workqueue_struct *cifsiod_wq;
CIFS: fix oplock break deadlocks When the final cifsFileInfo_put() is called from cifsiod and an oplock break work is queued, lockdep complains loudly: ============================================= [ INFO: possible recursive locking detected ] 4.11.0+ #21 Not tainted --------------------------------------------- kworker/0:2/78 is trying to acquire lock: ("cifsiod"){++++.+}, at: flush_work+0x215/0x350 but task is already holding lock: ("cifsiod"){++++.+}, at: process_one_work+0x255/0x8e0 other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock("cifsiod"); lock("cifsiod"); *** DEADLOCK *** May be due to missing lock nesting notation 2 locks held by kworker/0:2/78: #0: ("cifsiod"){++++.+}, at: process_one_work+0x255/0x8e0 #1: ((&wdata->work)){+.+...}, at: process_one_work+0x255/0x8e0 stack backtrace: CPU: 0 PID: 78 Comm: kworker/0:2 Not tainted 4.11.0+ #21 Workqueue: cifsiod cifs_writev_complete Call Trace: dump_stack+0x85/0xc2 __lock_acquire+0x17dd/0x2260 ? match_held_lock+0x20/0x2b0 ? trace_hardirqs_off_caller+0x86/0x130 ? mark_lock+0xa6/0x920 lock_acquire+0xcc/0x260 ? lock_acquire+0xcc/0x260 ? flush_work+0x215/0x350 flush_work+0x236/0x350 ? flush_work+0x215/0x350 ? destroy_worker+0x170/0x170 __cancel_work_timer+0x17d/0x210 ? ___preempt_schedule+0x16/0x18 cancel_work_sync+0x10/0x20 cifsFileInfo_put+0x338/0x7f0 cifs_writedata_release+0x2a/0x40 ? cifs_writedata_release+0x2a/0x40 cifs_writev_complete+0x29d/0x850 ? preempt_count_sub+0x18/0xd0 process_one_work+0x304/0x8e0 worker_thread+0x9b/0x6a0 kthread+0x1b2/0x200 ? process_one_work+0x8e0/0x8e0 ? kthread_create_on_node+0x40/0x40 ret_from_fork+0x31/0x40 This is a real warning. Since the oplock is queued on the same workqueue this can deadlock if there is only one worker thread active for the workqueue (which will be the case during memory pressure when the rescuer thread is handling it). Furthermore, there is at least one other kind of hang possible due to the oplock break handling if there is only worker. (This can be reproduced without introducing memory pressure by having passing 1 for the max_active parameter of cifsiod.) cifs_oplock_break() can wait indefintely in the filemap_fdatawait() while the cifs_writev_complete() work is blocked: sysrq: SysRq : Show Blocked State task PC stack pid father kworker/0:1 D 0 16 2 0x00000000 Workqueue: cifsiod cifs_oplock_break Call Trace: __schedule+0x562/0xf40 ? mark_held_locks+0x4a/0xb0 schedule+0x57/0xe0 io_schedule+0x21/0x50 wait_on_page_bit+0x143/0x190 ? add_to_page_cache_lru+0x150/0x150 __filemap_fdatawait_range+0x134/0x190 ? do_writepages+0x51/0x70 filemap_fdatawait_range+0x14/0x30 filemap_fdatawait+0x3b/0x40 cifs_oplock_break+0x651/0x710 ? preempt_count_sub+0x18/0xd0 process_one_work+0x304/0x8e0 worker_thread+0x9b/0x6a0 kthread+0x1b2/0x200 ? process_one_work+0x8e0/0x8e0 ? kthread_create_on_node+0x40/0x40 ret_from_fork+0x31/0x40 dd D 0 683 171 0x00000000 Call Trace: __schedule+0x562/0xf40 ? mark_held_locks+0x29/0xb0 schedule+0x57/0xe0 io_schedule+0x21/0x50 wait_on_page_bit+0x143/0x190 ? add_to_page_cache_lru+0x150/0x150 __filemap_fdatawait_range+0x134/0x190 ? do_writepages+0x51/0x70 filemap_fdatawait_range+0x14/0x30 filemap_fdatawait+0x3b/0x40 filemap_write_and_wait+0x4e/0x70 cifs_flush+0x6a/0xb0 filp_close+0x52/0xa0 __close_fd+0xdc/0x150 SyS_close+0x33/0x60 entry_SYSCALL_64_fastpath+0x1f/0xbe Showing all locks held in the system: 2 locks held by kworker/0:1/16: #0: ("cifsiod"){.+.+.+}, at: process_one_work+0x255/0x8e0 #1: ((&cfile->oplock_break)){+.+.+.}, at: process_one_work+0x255/0x8e0 Showing busy workqueues and worker pools: workqueue cifsiod: flags=0xc pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=1/1 in-flight: 16:cifs_oplock_break delayed: cifs_writev_complete, cifs_echo_request pool 0: cpus=0 node=0 flags=0x0 nice=0 hung=0s workers=3 idle: 750 3 Fix these problems by creating a a new workqueue (with a rescuer) for the oplock break work. Signed-off-by: Rabin Vincent <rabinv@axis.com> Signed-off-by: Steve French <smfrench@gmail.com> CC: Stable <stable@vger.kernel.org>
2017-05-03 23:54:01 +08:00
struct workqueue_struct *cifsoplockd_wq;
__u32 cifs_lock_secret;
/*
* Bumps refcount for cifs super block.
* Note that it should be only called if a referece to VFS super block is
* already held, e.g. in open-type syscalls context. Otherwise it can race with
* atomic_dec_and_test in deactivate_locked_super.
*/
void
cifs_sb_active(struct super_block *sb)
{
struct cifs_sb_info *server = CIFS_SB(sb);
if (atomic_inc_return(&server->active) == 1)
atomic_inc(&sb->s_active);
}
void
cifs_sb_deactive(struct super_block *sb)
{
struct cifs_sb_info *server = CIFS_SB(sb);
if (atomic_dec_and_test(&server->active))
deactivate_super(sb);
}
static int
cifs_read_super(struct super_block *sb)
{
struct inode *inode;
struct cifs_sb_info *cifs_sb;
struct cifs_tcon *tcon;
int rc = 0;
cifs_sb = CIFS_SB(sb);
tcon = cifs_sb_master_tcon(cifs_sb);
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_POSIXACL)
Rename superblock flags (MS_xyz -> SB_xyz) This is a pure automated search-and-replace of the internal kernel superblock flags. The s_flags are now called SB_*, with the names and the values for the moment mirroring the MS_* flags that they're equivalent to. Note how the MS_xyz flags are the ones passed to the mount system call, while the SB_xyz flags are what we then use in sb->s_flags. The script to do this was: # places to look in; re security/*: it generally should *not* be # touched (that stuff parses mount(2) arguments directly), but # there are two places where we really deal with superblock flags. FILES="drivers/mtd drivers/staging/lustre fs ipc mm \ include/linux/fs.h include/uapi/linux/bfs_fs.h \ security/apparmor/apparmorfs.c security/apparmor/include/lib.h" # the list of MS_... constants SYMS="RDONLY NOSUID NODEV NOEXEC SYNCHRONOUS REMOUNT MANDLOCK \ DIRSYNC NOATIME NODIRATIME BIND MOVE REC VERBOSE SILENT \ POSIXACL UNBINDABLE PRIVATE SLAVE SHARED RELATIME KERNMOUNT \ I_VERSION STRICTATIME LAZYTIME SUBMOUNT NOREMOTELOCK NOSEC BORN \ ACTIVE NOUSER" SED_PROG= for i in $SYMS; do SED_PROG="$SED_PROG -e s/MS_$i/SB_$i/g"; done # we want files that contain at least one of MS_..., # with fs/namespace.c and fs/pnode.c excluded. L=$(for i in $SYMS; do git grep -w -l MS_$i $FILES; done| sort|uniq|grep -v '^fs/namespace.c'|grep -v '^fs/pnode.c') for f in $L; do sed -i $f $SED_PROG; done Requested-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-11-28 05:05:09 +08:00
sb->s_flags |= SB_POSIXACL;
if (tcon->snapshot_time)
sb->s_flags |= SB_RDONLY;
if (tcon->ses->capabilities & tcon->ses->server->vals->cap_large_files)
sb->s_maxbytes = MAX_LFS_FILESIZE;
else
sb->s_maxbytes = MAX_NON_LFS;
/* BB FIXME fix time_gran to be larger for LANMAN sessions */
sb->s_time_gran = 100;
sb->s_magic = CIFS_MAGIC_NUMBER;
sb->s_op = &cifs_super_ops;
sb->s_xattr = cifs_xattr_handlers;
rc = super_setup_bdi(sb);
if (rc)
goto out_no_root;
/* tune readahead according to rsize */
sb->s_bdi->ra_pages = cifs_sb->rsize / PAGE_SIZE;
sb->s_blocksize = CIFS_MAX_MSGSIZE;
sb->s_blocksize_bits = 14; /* default 2**14 = CIFS_MAX_MSGSIZE */
inode = cifs_root_iget(sb);
if (IS_ERR(inode)) {
rc = PTR_ERR(inode);
goto out_no_root;
}
if (tcon->nocase)
sb->s_d_op = &cifs_ci_dentry_ops;
else
sb->s_d_op = &cifs_dentry_ops;
sb->s_root = d_make_root(inode);
if (!sb->s_root) {
rc = -ENOMEM;
goto out_no_root;
}
#ifdef CONFIG_CIFS_NFSD_EXPORT
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_SERVER_INUM) {
cifs_dbg(FYI, "export ops supported\n");
sb->s_export_op = &cifs_export_ops;
}
#endif /* CONFIG_CIFS_NFSD_EXPORT */
return 0;
out_no_root:
cifs_dbg(VFS, "%s: get root inode failed\n", __func__);
return rc;
}
static void cifs_kill_sb(struct super_block *sb)
{
struct cifs_sb_info *cifs_sb = CIFS_SB(sb);
kill_anon_super(sb);
cifs_umount(cifs_sb);
}
static int
cifs_statfs(struct dentry *dentry, struct kstatfs *buf)
{
struct super_block *sb = dentry->d_sb;
struct cifs_sb_info *cifs_sb = CIFS_SB(sb);
struct cifs_tcon *tcon = cifs_sb_master_tcon(cifs_sb);
struct TCP_Server_Info *server = tcon->ses->server;
unsigned int xid;
int rc = 0;
xid = get_xid();
if (le32_to_cpu(tcon->fsAttrInfo.MaxPathNameComponentLength) > 0)
buf->f_namelen =
le32_to_cpu(tcon->fsAttrInfo.MaxPathNameComponentLength);
else
buf->f_namelen = PATH_MAX;
buf->f_fsid.val[0] = tcon->vol_serial_number;
/* are using part of create time for more randomness, see man statfs */
buf->f_fsid.val[1] = (int)le64_to_cpu(tcon->vol_create_time);
buf->f_files = 0; /* undefined */
buf->f_ffree = 0; /* unlimited */
if (server->ops->queryfs)
rc = server->ops->queryfs(xid, tcon, buf);
free_xid(xid);
return 0;
}
static long cifs_fallocate(struct file *file, int mode, loff_t off, loff_t len)
{
struct cifs_sb_info *cifs_sb = CIFS_FILE_SB(file);
struct cifs_tcon *tcon = cifs_sb_master_tcon(cifs_sb);
struct TCP_Server_Info *server = tcon->ses->server;
if (server->ops->fallocate)
return server->ops->fallocate(file, tcon, mode, off, len);
return -EOPNOTSUPP;
}
static int cifs_permission(struct inode *inode, int mask)
{
struct cifs_sb_info *cifs_sb;
cifs_sb = CIFS_SB(inode->i_sb);
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_PERM) {
if ((mask & MAY_EXEC) && !execute_ok(inode))
return -EACCES;
else
return 0;
} else /* file mode might have been restricted at mount time
on the client (above and beyond ACL on servers) for
servers which do not support setting and viewing mode bits,
so allowing client to check permissions is useful */
return generic_permission(inode, mask);
}
static struct kmem_cache *cifs_inode_cachep;
static struct kmem_cache *cifs_req_cachep;
static struct kmem_cache *cifs_mid_cachep;
static struct kmem_cache *cifs_sm_req_cachep;
mempool_t *cifs_sm_req_poolp;
mempool_t *cifs_req_poolp;
mempool_t *cifs_mid_poolp;
static struct inode *
cifs_alloc_inode(struct super_block *sb)
{
struct cifsInodeInfo *cifs_inode;
cifs_inode = kmem_cache_alloc(cifs_inode_cachep, GFP_KERNEL);
if (!cifs_inode)
return NULL;
cifs_inode->cifsAttrs = 0x20; /* default */
cifs_inode->time = 0;
/*
* Until the file is open and we have gotten oplock info back from the
* server, can not assume caching of file data or metadata.
*/
cifs_set_oplock_level(cifs_inode, 0);
cifs_inode->flags = 0;
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: Sachin Prabhu <sprabhu@redhat.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: Pavel Shilovsky <piastry@etersoft.ru> Signed-off-by: Steve French <smfrench@gmail.com>
2014-03-12 00:11:47 +08:00
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->server_eof = 0;
cifs_inode->uniqueid = 0;
cifs_inode->createtime = 0;
cifs_inode->epoch = 0;
generate_random_uuid(cifs_inode->lease_key);
/*
* Can not set i_flags here - they get immediately overwritten to zero
* by the VFS.
*/
/* cifs_inode->vfs_inode.i_flags = S_NOATIME | S_NOCMTIME; */
INIT_LIST_HEAD(&cifs_inode->openFileList);
INIT_LIST_HEAD(&cifs_inode->llist);
return &cifs_inode->vfs_inode;
}
2011-01-07 14:49:49 +08:00
static void cifs_i_callback(struct rcu_head *head)
{
struct inode *inode = container_of(head, struct inode, i_rcu);
kmem_cache_free(cifs_inode_cachep, CIFS_I(inode));
}
static void
cifs_destroy_inode(struct inode *inode)
{
2011-01-07 14:49:49 +08:00
call_rcu(&inode->i_rcu, cifs_i_callback);
}
static void
cifs_evict_inode(struct inode *inode)
{
mm + fs: store shadow entries in page cache Reclaim will be leaving shadow entries in the page cache radix tree upon evicting the real page. As those pages are found from the LRU, an iput() can lead to the inode being freed concurrently. At this point, reclaim must no longer install shadow pages because the inode freeing code needs to ensure the page tree is really empty. Add an address_space flag, AS_EXITING, that the inode freeing code sets under the tree lock before doing the final truncate. Reclaim will check for this flag before installing shadow pages. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Rik van Riel <riel@redhat.com> Reviewed-by: Minchan Kim <minchan@kernel.org> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Bob Liu <bob.liu@oracle.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Greg Thelen <gthelen@google.com> Cc: Hugh Dickins <hughd@google.com> Cc: Jan Kara <jack@suse.cz> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Luigi Semenzato <semenzato@google.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Metin Doslu <metin@citusdata.com> Cc: Michel Lespinasse <walken@google.com> Cc: Ozgun Erdogan <ozgun@citusdata.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Roman Gushchin <klamm@yandex-team.ru> Cc: Ryan Mallon <rmallon@gmail.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vlastimil Babka <vbabka@suse.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-04-04 05:47:49 +08:00
truncate_inode_pages_final(&inode->i_data);
clear_inode(inode);
cifs_fscache_release_inode_cookie(inode);
}
static void
cifs_show_address(struct seq_file *s, struct TCP_Server_Info *server)
{
struct sockaddr_in *sa = (struct sockaddr_in *) &server->dstaddr;
struct sockaddr_in6 *sa6 = (struct sockaddr_in6 *) &server->dstaddr;
seq_puts(s, ",addr=");
switch (server->dstaddr.ss_family) {
case AF_INET:
seq_printf(s, "%pI4", &sa->sin_addr.s_addr);
break;
case AF_INET6:
seq_printf(s, "%pI6", &sa6->sin6_addr.s6_addr);
if (sa6->sin6_scope_id)
seq_printf(s, "%%%u", sa6->sin6_scope_id);
break;
default:
seq_puts(s, "(unknown)");
}
if (server->rdma)
seq_puts(s, ",rdma");
}
static void
cifs_show_security(struct seq_file *s, struct cifs_ses *ses)
{
if (ses->sectype == Unspecified) {
if (ses->user_name == NULL)
seq_puts(s, ",sec=none");
return;
}
seq_puts(s, ",sec=");
switch (ses->sectype) {
case LANMAN:
seq_puts(s, "lanman");
break;
case NTLMv2:
seq_puts(s, "ntlmv2");
break;
case NTLM:
seq_puts(s, "ntlm");
break;
case Kerberos:
seq_puts(s, "krb5");
break;
case RawNTLMSSP:
seq_puts(s, "ntlmssp");
break;
default:
/* shouldn't ever happen */
seq_puts(s, "unknown");
break;
}
if (ses->sign)
seq_puts(s, "i");
}
static void
cifs_show_cache_flavor(struct seq_file *s, struct cifs_sb_info *cifs_sb)
{
seq_puts(s, ",cache=");
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_STRICT_IO)
seq_puts(s, "strict");
else if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_DIRECT_IO)
seq_puts(s, "none");
else
seq_puts(s, "loose");
}
static void
cifs_show_nls(struct seq_file *s, struct nls_table *cur)
{
struct nls_table *def;
/* Display iocharset= option if it's not default charset */
def = load_nls_default();
if (def != cur)
seq_printf(s, ",iocharset=%s", cur->charset);
unload_nls(def);
}
/*
* cifs_show_options() is for displaying mount options in /proc/mounts.
* Not all settable options are displayed but most of the important
* ones are.
*/
static int
cifs_show_options(struct seq_file *s, struct dentry *root)
{
struct cifs_sb_info *cifs_sb = CIFS_SB(root->d_sb);
struct cifs_tcon *tcon = cifs_sb_master_tcon(cifs_sb);
struct sockaddr *srcaddr;
srcaddr = (struct sockaddr *)&tcon->ses->server->srcaddr;
fs: create and use seq_show_option for escaping Many file systems that implement the show_options hook fail to correctly escape their output which could lead to unescaped characters (e.g. new lines) leaking into /proc/mounts and /proc/[pid]/mountinfo files. This could lead to confusion, spoofed entries (resulting in things like systemd issuing false d-bus "mount" notifications), and who knows what else. This looks like it would only be the root user stepping on themselves, but it's possible weird things could happen in containers or in other situations with delegated mount privileges. Here's an example using overlay with setuid fusermount trusting the contents of /proc/mounts (via the /etc/mtab symlink). Imagine the use of "sudo" is something more sneaky: $ BASE="ovl" $ MNT="$BASE/mnt" $ LOW="$BASE/lower" $ UP="$BASE/upper" $ WORK="$BASE/work/ 0 0 none /proc fuse.pwn user_id=1000" $ mkdir -p "$LOW" "$UP" "$WORK" $ sudo mount -t overlay -o "lowerdir=$LOW,upperdir=$UP,workdir=$WORK" none /mnt $ cat /proc/mounts none /root/ovl/mnt overlay rw,relatime,lowerdir=ovl/lower,upperdir=ovl/upper,workdir=ovl/work/ 0 0 none /proc fuse.pwn user_id=1000 0 0 $ fusermount -u /proc $ cat /proc/mounts cat: /proc/mounts: No such file or directory This fixes the problem by adding new seq_show_option and seq_show_option_n helpers, and updating the vulnerable show_option handlers to use them as needed. Some, like SELinux, need to be open coded due to unusual existing escape mechanisms. [akpm@linux-foundation.org: add lost chunk, per Kees] [keescook@chromium.org: seq_show_option should be using const parameters] Signed-off-by: Kees Cook <keescook@chromium.org> Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Acked-by: Jan Kara <jack@suse.com> Acked-by: Paul Moore <paul@paul-moore.com> Cc: J. R. Okajima <hooanon05g@gmail.com> Signed-off-by: Kees Cook <keescook@chromium.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-09-05 06:44:57 +08:00
seq_show_option(s, "vers", tcon->ses->server->vals->version_string);
cifs_show_security(s, tcon->ses);
cifs_show_cache_flavor(s, cifs_sb);
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MULTIUSER)
seq_puts(s, ",multiuser");
else if (tcon->ses->user_name)
fs: create and use seq_show_option for escaping Many file systems that implement the show_options hook fail to correctly escape their output which could lead to unescaped characters (e.g. new lines) leaking into /proc/mounts and /proc/[pid]/mountinfo files. This could lead to confusion, spoofed entries (resulting in things like systemd issuing false d-bus "mount" notifications), and who knows what else. This looks like it would only be the root user stepping on themselves, but it's possible weird things could happen in containers or in other situations with delegated mount privileges. Here's an example using overlay with setuid fusermount trusting the contents of /proc/mounts (via the /etc/mtab symlink). Imagine the use of "sudo" is something more sneaky: $ BASE="ovl" $ MNT="$BASE/mnt" $ LOW="$BASE/lower" $ UP="$BASE/upper" $ WORK="$BASE/work/ 0 0 none /proc fuse.pwn user_id=1000" $ mkdir -p "$LOW" "$UP" "$WORK" $ sudo mount -t overlay -o "lowerdir=$LOW,upperdir=$UP,workdir=$WORK" none /mnt $ cat /proc/mounts none /root/ovl/mnt overlay rw,relatime,lowerdir=ovl/lower,upperdir=ovl/upper,workdir=ovl/work/ 0 0 none /proc fuse.pwn user_id=1000 0 0 $ fusermount -u /proc $ cat /proc/mounts cat: /proc/mounts: No such file or directory This fixes the problem by adding new seq_show_option and seq_show_option_n helpers, and updating the vulnerable show_option handlers to use them as needed. Some, like SELinux, need to be open coded due to unusual existing escape mechanisms. [akpm@linux-foundation.org: add lost chunk, per Kees] [keescook@chromium.org: seq_show_option should be using const parameters] Signed-off-by: Kees Cook <keescook@chromium.org> Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Acked-by: Jan Kara <jack@suse.com> Acked-by: Paul Moore <paul@paul-moore.com> Cc: J. R. Okajima <hooanon05g@gmail.com> Signed-off-by: Kees Cook <keescook@chromium.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-09-05 06:44:57 +08:00
seq_show_option(s, "username", tcon->ses->user_name);
if (tcon->ses->domainName)
fs: create and use seq_show_option for escaping Many file systems that implement the show_options hook fail to correctly escape their output which could lead to unescaped characters (e.g. new lines) leaking into /proc/mounts and /proc/[pid]/mountinfo files. This could lead to confusion, spoofed entries (resulting in things like systemd issuing false d-bus "mount" notifications), and who knows what else. This looks like it would only be the root user stepping on themselves, but it's possible weird things could happen in containers or in other situations with delegated mount privileges. Here's an example using overlay with setuid fusermount trusting the contents of /proc/mounts (via the /etc/mtab symlink). Imagine the use of "sudo" is something more sneaky: $ BASE="ovl" $ MNT="$BASE/mnt" $ LOW="$BASE/lower" $ UP="$BASE/upper" $ WORK="$BASE/work/ 0 0 none /proc fuse.pwn user_id=1000" $ mkdir -p "$LOW" "$UP" "$WORK" $ sudo mount -t overlay -o "lowerdir=$LOW,upperdir=$UP,workdir=$WORK" none /mnt $ cat /proc/mounts none /root/ovl/mnt overlay rw,relatime,lowerdir=ovl/lower,upperdir=ovl/upper,workdir=ovl/work/ 0 0 none /proc fuse.pwn user_id=1000 0 0 $ fusermount -u /proc $ cat /proc/mounts cat: /proc/mounts: No such file or directory This fixes the problem by adding new seq_show_option and seq_show_option_n helpers, and updating the vulnerable show_option handlers to use them as needed. Some, like SELinux, need to be open coded due to unusual existing escape mechanisms. [akpm@linux-foundation.org: add lost chunk, per Kees] [keescook@chromium.org: seq_show_option should be using const parameters] Signed-off-by: Kees Cook <keescook@chromium.org> Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Acked-by: Jan Kara <jack@suse.com> Acked-by: Paul Moore <paul@paul-moore.com> Cc: J. R. Okajima <hooanon05g@gmail.com> Signed-off-by: Kees Cook <keescook@chromium.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-09-05 06:44:57 +08:00
seq_show_option(s, "domain", tcon->ses->domainName);
if (srcaddr->sa_family != AF_UNSPEC) {
struct sockaddr_in *saddr4;
struct sockaddr_in6 *saddr6;
saddr4 = (struct sockaddr_in *)srcaddr;
saddr6 = (struct sockaddr_in6 *)srcaddr;
if (srcaddr->sa_family == AF_INET6)
seq_printf(s, ",srcaddr=%pI6c",
&saddr6->sin6_addr);
else if (srcaddr->sa_family == AF_INET)
seq_printf(s, ",srcaddr=%pI4",
&saddr4->sin_addr.s_addr);
else
seq_printf(s, ",srcaddr=BAD-AF:%i",
(int)(srcaddr->sa_family));
}
seq_printf(s, ",uid=%u",
from_kuid_munged(&init_user_ns, cifs_sb->mnt_uid));
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_OVERR_UID)
seq_puts(s, ",forceuid");
else
seq_puts(s, ",noforceuid");
seq_printf(s, ",gid=%u",
from_kgid_munged(&init_user_ns, cifs_sb->mnt_gid));
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_OVERR_GID)
seq_puts(s, ",forcegid");
else
seq_puts(s, ",noforcegid");
cifs_show_address(s, tcon->ses->server);
if (!tcon->unix_ext)
seq_printf(s, ",file_mode=0%ho,dir_mode=0%ho",
cifs_sb->mnt_file_mode,
cifs_sb->mnt_dir_mode);
cifs_show_nls(s, cifs_sb->local_nls);
if (tcon->seal)
seq_puts(s, ",seal");
if (tcon->nocase)
seq_puts(s, ",nocase");
if (tcon->retry)
seq_puts(s, ",hard");
else
seq_puts(s, ",soft");
if (tcon->use_persistent)
seq_puts(s, ",persistenthandles");
else if (tcon->use_resilient)
seq_puts(s, ",resilienthandles");
if (tcon->posix_extensions)
seq_puts(s, ",posix");
else if (tcon->unix_ext)
seq_puts(s, ",unix");
else
seq_puts(s, ",nounix");
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_POSIX_PATHS)
seq_puts(s, ",posixpaths");
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_SET_UID)
seq_puts(s, ",setuids");
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_UID_FROM_ACL)
seq_puts(s, ",idsfromsid");
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_SERVER_INUM)
seq_puts(s, ",serverino");
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_RWPIDFORWARD)
seq_puts(s, ",rwpidforward");
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NOPOSIXBRL)
seq_puts(s, ",forcemand");
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_XATTR)
seq_puts(s, ",nouser_xattr");
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MAP_SPECIAL_CHR)
seq_puts(s, ",mapchars");
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MAP_SFM_CHR)
seq_puts(s, ",mapposix");
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_UNX_EMUL)
seq_puts(s, ",sfu");
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_BRL)
seq_puts(s, ",nobrl");
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_HANDLE_CACHE)
seq_puts(s, ",nohandlecache");
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_CIFS_ACL)
seq_puts(s, ",cifsacl");
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_DYNPERM)
seq_puts(s, ",dynperm");
Rename superblock flags (MS_xyz -> SB_xyz) This is a pure automated search-and-replace of the internal kernel superblock flags. The s_flags are now called SB_*, with the names and the values for the moment mirroring the MS_* flags that they're equivalent to. Note how the MS_xyz flags are the ones passed to the mount system call, while the SB_xyz flags are what we then use in sb->s_flags. The script to do this was: # places to look in; re security/*: it generally should *not* be # touched (that stuff parses mount(2) arguments directly), but # there are two places where we really deal with superblock flags. FILES="drivers/mtd drivers/staging/lustre fs ipc mm \ include/linux/fs.h include/uapi/linux/bfs_fs.h \ security/apparmor/apparmorfs.c security/apparmor/include/lib.h" # the list of MS_... constants SYMS="RDONLY NOSUID NODEV NOEXEC SYNCHRONOUS REMOUNT MANDLOCK \ DIRSYNC NOATIME NODIRATIME BIND MOVE REC VERBOSE SILENT \ POSIXACL UNBINDABLE PRIVATE SLAVE SHARED RELATIME KERNMOUNT \ I_VERSION STRICTATIME LAZYTIME SUBMOUNT NOREMOTELOCK NOSEC BORN \ ACTIVE NOUSER" SED_PROG= for i in $SYMS; do SED_PROG="$SED_PROG -e s/MS_$i/SB_$i/g"; done # we want files that contain at least one of MS_..., # with fs/namespace.c and fs/pnode.c excluded. L=$(for i in $SYMS; do git grep -w -l MS_$i $FILES; done| sort|uniq|grep -v '^fs/namespace.c'|grep -v '^fs/pnode.c') for f in $L; do sed -i $f $SED_PROG; done Requested-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-11-28 05:05:09 +08:00
if (root->d_sb->s_flags & SB_POSIXACL)
seq_puts(s, ",acl");
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MF_SYMLINKS)
seq_puts(s, ",mfsymlinks");
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_FSCACHE)
seq_puts(s, ",fsc");
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NOSSYNC)
seq_puts(s, ",nostrictsync");
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_PERM)
seq_puts(s, ",noperm");
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_CIFS_BACKUPUID)
seq_printf(s, ",backupuid=%u",
from_kuid_munged(&init_user_ns,
cifs_sb->mnt_backupuid));
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_CIFS_BACKUPGID)
seq_printf(s, ",backupgid=%u",
from_kgid_munged(&init_user_ns,
cifs_sb->mnt_backupgid));
seq_printf(s, ",rsize=%u", cifs_sb->rsize);
seq_printf(s, ",wsize=%u", cifs_sb->wsize);
seq_printf(s, ",echo_interval=%lu",
tcon->ses->server->echo_interval / HZ);
if (tcon->snapshot_time)
seq_printf(s, ",snapshot=%llu", tcon->snapshot_time);
/* convert actimeo and display it in seconds */
seq_printf(s, ",actimeo=%lu", cifs_sb->actimeo / HZ);
return 0;
}
static void cifs_umount_begin(struct super_block *sb)
{
struct cifs_sb_info *cifs_sb = CIFS_SB(sb);
struct cifs_tcon *tcon;
if (cifs_sb == NULL)
return;
tcon = cifs_sb_master_tcon(cifs_sb);
spin_lock(&cifs_tcp_ses_lock);
if ((tcon->tc_count > 1) || (tcon->tidStatus == CifsExiting)) {
/* we have other mounts to same share or we have
already tried to force umount this and woken up
all waiting network requests, nothing to do */
spin_unlock(&cifs_tcp_ses_lock);
return;
} else if (tcon->tc_count == 1)
tcon->tidStatus = CifsExiting;
spin_unlock(&cifs_tcp_ses_lock);
/* cancel_brl_requests(tcon); */ /* BB mark all brl mids as exiting */
2005-11-10 07:21:09 +08:00
/* cancel_notify_requests(tcon); */
if (tcon->ses && tcon->ses->server) {
cifs_dbg(FYI, "wake up tasks now - umount begin not complete\n");
wake_up_all(&tcon->ses->server->request_q);
wake_up_all(&tcon->ses->server->response_q);
msleep(1); /* yield */
/* we have to kick the requests once more */
wake_up_all(&tcon->ses->server->response_q);
msleep(1);
}
return;
}
#ifdef CONFIG_CIFS_STATS2
static int cifs_show_stats(struct seq_file *s, struct dentry *root)
{
/* BB FIXME */
return 0;
}
#endif
static int cifs_remount(struct super_block *sb, int *flags, char *data)
{
fs: push sync_filesystem() down to the file system's remount_fs() Previously, the no-op "mount -o mount /dev/xxx" operation when the file system is already mounted read-write causes an implied, unconditional syncfs(). This seems pretty stupid, and it's certainly documented or guaraunteed to do this, nor is it particularly useful, except in the case where the file system was mounted rw and is getting remounted read-only. However, it's possible that there might be some file systems that are actually depending on this behavior. In most file systems, it's probably fine to only call sync_filesystem() when transitioning from read-write to read-only, and there are some file systems where this is not needed at all (for example, for a pseudo-filesystem or something like romfs). Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Cc: linux-fsdevel@vger.kernel.org Cc: Christoph Hellwig <hch@infradead.org> Cc: Artem Bityutskiy <dedekind1@gmail.com> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Evgeniy Dushistov <dushistov@mail.ru> Cc: Jan Kara <jack@suse.cz> Cc: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp> Cc: Anders Larsen <al@alarsen.net> Cc: Phillip Lougher <phillip@squashfs.org.uk> Cc: Kees Cook <keescook@chromium.org> Cc: Mikulas Patocka <mikulas@artax.karlin.mff.cuni.cz> Cc: Petr Vandrovec <petr@vandrovec.name> Cc: xfs@oss.sgi.com Cc: linux-btrfs@vger.kernel.org Cc: linux-cifs@vger.kernel.org Cc: samba-technical@lists.samba.org Cc: codalist@coda.cs.cmu.edu Cc: linux-ext4@vger.kernel.org Cc: linux-f2fs-devel@lists.sourceforge.net Cc: fuse-devel@lists.sourceforge.net Cc: cluster-devel@redhat.com Cc: linux-mtd@lists.infradead.org Cc: jfs-discussion@lists.sourceforge.net Cc: linux-nfs@vger.kernel.org Cc: linux-nilfs@vger.kernel.org Cc: linux-ntfs-dev@lists.sourceforge.net Cc: ocfs2-devel@oss.oracle.com Cc: reiserfs-devel@vger.kernel.org
2014-03-13 22:14:33 +08:00
sync_filesystem(sb);
Rename superblock flags (MS_xyz -> SB_xyz) This is a pure automated search-and-replace of the internal kernel superblock flags. The s_flags are now called SB_*, with the names and the values for the moment mirroring the MS_* flags that they're equivalent to. Note how the MS_xyz flags are the ones passed to the mount system call, while the SB_xyz flags are what we then use in sb->s_flags. The script to do this was: # places to look in; re security/*: it generally should *not* be # touched (that stuff parses mount(2) arguments directly), but # there are two places where we really deal with superblock flags. FILES="drivers/mtd drivers/staging/lustre fs ipc mm \ include/linux/fs.h include/uapi/linux/bfs_fs.h \ security/apparmor/apparmorfs.c security/apparmor/include/lib.h" # the list of MS_... constants SYMS="RDONLY NOSUID NODEV NOEXEC SYNCHRONOUS REMOUNT MANDLOCK \ DIRSYNC NOATIME NODIRATIME BIND MOVE REC VERBOSE SILENT \ POSIXACL UNBINDABLE PRIVATE SLAVE SHARED RELATIME KERNMOUNT \ I_VERSION STRICTATIME LAZYTIME SUBMOUNT NOREMOTELOCK NOSEC BORN \ ACTIVE NOUSER" SED_PROG= for i in $SYMS; do SED_PROG="$SED_PROG -e s/MS_$i/SB_$i/g"; done # we want files that contain at least one of MS_..., # with fs/namespace.c and fs/pnode.c excluded. L=$(for i in $SYMS; do git grep -w -l MS_$i $FILES; done| sort|uniq|grep -v '^fs/namespace.c'|grep -v '^fs/pnode.c') for f in $L; do sed -i $f $SED_PROG; done Requested-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-11-28 05:05:09 +08:00
*flags |= SB_NODIRATIME;
return 0;
}
static int cifs_drop_inode(struct inode *inode)
{
struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
/* no serverino => unconditional eviction */
return !(cifs_sb->mnt_cifs_flags & CIFS_MOUNT_SERVER_INUM) ||
generic_drop_inode(inode);
}
static const struct super_operations cifs_super_ops = {
.statfs = cifs_statfs,
.alloc_inode = cifs_alloc_inode,
.destroy_inode = cifs_destroy_inode,
.drop_inode = cifs_drop_inode,
.evict_inode = cifs_evict_inode,
/* .delete_inode = cifs_delete_inode, */ /* Do not need above
function unless later we add lazy close of inodes or unless the
kernel forgets to call us with the same number of releases (closes)
as opens */
.show_options = cifs_show_options,
2005-11-10 07:21:09 +08:00
.umount_begin = cifs_umount_begin,
.remount_fs = cifs_remount,
#ifdef CONFIG_CIFS_STATS2
.show_stats = cifs_show_stats,
#endif
};
/*
* Get root dentry from superblock according to prefix path mount option.
* Return dentry with refcount + 1 on success and NULL otherwise.
*/
static struct dentry *
cifs_get_root(struct smb_vol *vol, struct super_block *sb)
{
struct dentry *dentry;
struct cifs_sb_info *cifs_sb = CIFS_SB(sb);
char *full_path = NULL;
char *s, *p;
char sep;
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_USE_PREFIX_PATH)
return dget(sb->s_root);
full_path = cifs_build_path_to_root(vol, cifs_sb,
cifs_sb_master_tcon(cifs_sb), 0);
if (full_path == NULL)
return ERR_PTR(-ENOMEM);
cifs_dbg(FYI, "Get root dentry for %s\n", full_path);
sep = CIFS_DIR_SEP(cifs_sb);
dentry = dget(sb->s_root);
p = s = full_path;
do {
struct inode *dir = d_inode(dentry);
struct dentry *child;
if (!dir) {
dput(dentry);
dentry = ERR_PTR(-ENOENT);
break;
}
cifs: ensure that cifs_get_root() only traverses directories Kjell Braden reported this oops: [ 833.211970] BUG: unable to handle kernel NULL pointer dereference at (null) [ 833.212816] IP: [< (null)>] (null) [ 833.213280] PGD 1b9b2067 PUD e9f7067 PMD 0 [ 833.213874] Oops: 0010 [#1] SMP [ 833.214344] CPU 0 [ 833.214458] Modules linked in: des_generic md4 nls_utf8 cifs vboxvideo drm snd_intel8x0 snd_ac97_codec ac97_bus snd_pcm snd_seq_midi snd_rawmidi snd_seq_midi_event snd_seq bnep rfcomm snd_timer bluetooth snd_seq_device ppdev snd vboxguest parport_pc joydev mac_hid soundcore snd_page_alloc psmouse i2c_piix4 serio_raw lp parport usbhid hid e1000 [ 833.215629] [ 833.215629] Pid: 1752, comm: mount.cifs Not tainted 3.0.0-rc7-bisectcifs-fec11dd9a0+ #18 innotek GmbH VirtualBox/VirtualBox [ 833.215629] RIP: 0010:[<0000000000000000>] [< (null)>] (null) [ 833.215629] RSP: 0018:ffff8800119c9c50 EFLAGS: 00010282 [ 833.215629] RAX: ffffffffa02186c0 RBX: ffff88000c427780 RCX: 0000000000000000 [ 833.215629] RDX: 0000000000000000 RSI: ffff88000c427780 RDI: ffff88000c4362e8 [ 833.215629] RBP: ffff8800119c9c88 R08: ffff88001fc15e30 R09: 00000000d69515c7 [ 833.215629] R10: ffffffffa0201972 R11: ffff88000e8f6a28 R12: ffff88000c4362e8 [ 833.215629] R13: 0000000000000000 R14: 0000000000000000 R15: ffff88001181aaa6 [ 833.215629] FS: 00007f2986171700(0000) GS:ffff88001fc00000(0000) knlGS:0000000000000000 [ 833.215629] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b [ 833.215629] CR2: 0000000000000000 CR3: 000000001b982000 CR4: 00000000000006f0 [ 833.215629] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [ 833.215629] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 [ 833.215629] Process mount.cifs (pid: 1752, threadinfo ffff8800119c8000, task ffff88001c1c16f0) [ 833.215629] Stack: [ 833.215629] ffffffff8116a9b5 ffff8800119c9c88 ffffffff81178075 0000000000000286 [ 833.215629] 0000000000000000 ffff88000c4276c0 ffff8800119c9ce8 ffff8800119c9cc8 [ 833.215629] ffffffff8116b06e ffff88001bc6fc00 ffff88000c4276c0 ffff88000c4276c0 [ 833.215629] Call Trace: [ 833.215629] [<ffffffff8116a9b5>] ? d_alloc_and_lookup+0x45/0x90 [ 833.215629] [<ffffffff81178075>] ? d_lookup+0x35/0x60 [ 833.215629] [<ffffffff8116b06e>] __lookup_hash.part.14+0x9e/0xc0 [ 833.215629] [<ffffffff8116b1d6>] lookup_one_len+0x146/0x1e0 [ 833.215629] [<ffffffff815e4f7e>] ? _raw_spin_lock+0xe/0x20 [ 833.215629] [<ffffffffa01eef0d>] cifs_do_mount+0x26d/0x500 [cifs] [ 833.215629] [<ffffffff81163bd3>] mount_fs+0x43/0x1b0 [ 833.215629] [<ffffffff8117d41a>] vfs_kern_mount+0x6a/0xd0 [ 833.215629] [<ffffffff8117e584>] do_kern_mount+0x54/0x110 [ 833.215629] [<ffffffff8117fdc2>] do_mount+0x262/0x840 [ 833.215629] [<ffffffff81108a0e>] ? __get_free_pages+0xe/0x50 [ 833.215629] [<ffffffff8117f9ca>] ? copy_mount_options+0x3a/0x180 [ 833.215629] [<ffffffff8118075d>] sys_mount+0x8d/0xe0 [ 833.215629] [<ffffffff815ece82>] system_call_fastpath+0x16/0x1b [ 833.215629] Code: Bad RIP value. [ 833.215629] RIP [< (null)>] (null) [ 833.215629] RSP <ffff8800119c9c50> [ 833.215629] CR2: 0000000000000000 [ 833.238525] ---[ end trace ec00758b8d44f529 ]--- When walking down the path on the server, it's possible to hit a symlink. The path walking code assumes that the caller will handle that situation properly, but cifs_get_root() isn't set up for it. This patch prevents the oops by simply returning an error. A better solution would be to try and chase the symlinks here, but that's fairly complicated to handle. Fixes: https://bugzilla.kernel.org/show_bug.cgi?id=53221 Reported-and-tested-by: Kjell Braden <afflux@pentabarf.de> Cc: stable <stable@vger.kernel.org> Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
2013-02-02 04:11:01 +08:00
if (!S_ISDIR(dir->i_mode)) {
dput(dentry);
dentry = ERR_PTR(-ENOTDIR);
break;
}
/* skip separators */
while (*s == sep)
s++;
if (!*s)
break;
p = s++;
/* next separator */
while (*s && *s != sep)
s++;
child = lookup_one_len_unlocked(p, dentry, s - p);
dput(dentry);
dentry = child;
} while (!IS_ERR(dentry));
kfree(full_path);
return dentry;
}
static int cifs_set_super(struct super_block *sb, void *data)
{
struct cifs_mnt_data *mnt_data = data;
sb->s_fs_info = mnt_data->cifs_sb;
return set_anon_super(sb, NULL);
}
static struct dentry *
cifs_smb3_do_mount(struct file_system_type *fs_type,
int flags, const char *dev_name, void *data, bool is_smb3)
{
int rc;
struct super_block *sb;
struct cifs_sb_info *cifs_sb;
struct smb_vol *volume_info;
struct cifs_mnt_data mnt_data;
struct dentry *root;
cifs_dbg(FYI, "Devname: %s flags: %d\n", dev_name, flags);
volume_info = cifs_get_volume_info((char *)data, dev_name, is_smb3);
if (IS_ERR(volume_info))
return ERR_CAST(volume_info);
cifs_sb = kzalloc(sizeof(struct cifs_sb_info), GFP_KERNEL);
if (cifs_sb == NULL) {
root = ERR_PTR(-ENOMEM);
goto out_nls;
}
cifs_sb->mountdata = kstrndup(data, PAGE_SIZE, GFP_KERNEL);
if (cifs_sb->mountdata == NULL) {
root = ERR_PTR(-ENOMEM);
goto out_free;
}
rc = cifs_setup_cifs_sb(volume_info, cifs_sb);
if (rc) {
root = ERR_PTR(rc);
goto out_free;
}
rc = cifs_mount(cifs_sb, volume_info);
if (rc) {
Rename superblock flags (MS_xyz -> SB_xyz) This is a pure automated search-and-replace of the internal kernel superblock flags. The s_flags are now called SB_*, with the names and the values for the moment mirroring the MS_* flags that they're equivalent to. Note how the MS_xyz flags are the ones passed to the mount system call, while the SB_xyz flags are what we then use in sb->s_flags. The script to do this was: # places to look in; re security/*: it generally should *not* be # touched (that stuff parses mount(2) arguments directly), but # there are two places where we really deal with superblock flags. FILES="drivers/mtd drivers/staging/lustre fs ipc mm \ include/linux/fs.h include/uapi/linux/bfs_fs.h \ security/apparmor/apparmorfs.c security/apparmor/include/lib.h" # the list of MS_... constants SYMS="RDONLY NOSUID NODEV NOEXEC SYNCHRONOUS REMOUNT MANDLOCK \ DIRSYNC NOATIME NODIRATIME BIND MOVE REC VERBOSE SILENT \ POSIXACL UNBINDABLE PRIVATE SLAVE SHARED RELATIME KERNMOUNT \ I_VERSION STRICTATIME LAZYTIME SUBMOUNT NOREMOTELOCK NOSEC BORN \ ACTIVE NOUSER" SED_PROG= for i in $SYMS; do SED_PROG="$SED_PROG -e s/MS_$i/SB_$i/g"; done # we want files that contain at least one of MS_..., # with fs/namespace.c and fs/pnode.c excluded. L=$(for i in $SYMS; do git grep -w -l MS_$i $FILES; done| sort|uniq|grep -v '^fs/namespace.c'|grep -v '^fs/pnode.c') for f in $L; do sed -i $f $SED_PROG; done Requested-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-11-28 05:05:09 +08:00
if (!(flags & SB_SILENT))
cifs_dbg(VFS, "cifs_mount failed w/return code = %d\n",
rc);
root = ERR_PTR(rc);
goto out_free;
}
mnt_data.vol = volume_info;
mnt_data.cifs_sb = cifs_sb;
mnt_data.flags = flags;
/* BB should we make this contingent on mount parm? */
Rename superblock flags (MS_xyz -> SB_xyz) This is a pure automated search-and-replace of the internal kernel superblock flags. The s_flags are now called SB_*, with the names and the values for the moment mirroring the MS_* flags that they're equivalent to. Note how the MS_xyz flags are the ones passed to the mount system call, while the SB_xyz flags are what we then use in sb->s_flags. The script to do this was: # places to look in; re security/*: it generally should *not* be # touched (that stuff parses mount(2) arguments directly), but # there are two places where we really deal with superblock flags. FILES="drivers/mtd drivers/staging/lustre fs ipc mm \ include/linux/fs.h include/uapi/linux/bfs_fs.h \ security/apparmor/apparmorfs.c security/apparmor/include/lib.h" # the list of MS_... constants SYMS="RDONLY NOSUID NODEV NOEXEC SYNCHRONOUS REMOUNT MANDLOCK \ DIRSYNC NOATIME NODIRATIME BIND MOVE REC VERBOSE SILENT \ POSIXACL UNBINDABLE PRIVATE SLAVE SHARED RELATIME KERNMOUNT \ I_VERSION STRICTATIME LAZYTIME SUBMOUNT NOREMOTELOCK NOSEC BORN \ ACTIVE NOUSER" SED_PROG= for i in $SYMS; do SED_PROG="$SED_PROG -e s/MS_$i/SB_$i/g"; done # we want files that contain at least one of MS_..., # with fs/namespace.c and fs/pnode.c excluded. L=$(for i in $SYMS; do git grep -w -l MS_$i $FILES; done| sort|uniq|grep -v '^fs/namespace.c'|grep -v '^fs/pnode.c') for f in $L; do sed -i $f $SED_PROG; done Requested-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-11-28 05:05:09 +08:00
flags |= SB_NODIRATIME | SB_NOATIME;
sb = sget(fs_type, cifs_match_super, cifs_set_super, flags, &mnt_data);
if (IS_ERR(sb)) {
root = ERR_CAST(sb);
cifs_umount(cifs_sb);
goto out;
}
if (sb->s_root) {
cifs_dbg(FYI, "Use existing superblock\n");
cifs_umount(cifs_sb);
} else {
rc = cifs_read_super(sb);
if (rc) {
root = ERR_PTR(rc);
goto out_super;
}
Rename superblock flags (MS_xyz -> SB_xyz) This is a pure automated search-and-replace of the internal kernel superblock flags. The s_flags are now called SB_*, with the names and the values for the moment mirroring the MS_* flags that they're equivalent to. Note how the MS_xyz flags are the ones passed to the mount system call, while the SB_xyz flags are what we then use in sb->s_flags. The script to do this was: # places to look in; re security/*: it generally should *not* be # touched (that stuff parses mount(2) arguments directly), but # there are two places where we really deal with superblock flags. FILES="drivers/mtd drivers/staging/lustre fs ipc mm \ include/linux/fs.h include/uapi/linux/bfs_fs.h \ security/apparmor/apparmorfs.c security/apparmor/include/lib.h" # the list of MS_... constants SYMS="RDONLY NOSUID NODEV NOEXEC SYNCHRONOUS REMOUNT MANDLOCK \ DIRSYNC NOATIME NODIRATIME BIND MOVE REC VERBOSE SILENT \ POSIXACL UNBINDABLE PRIVATE SLAVE SHARED RELATIME KERNMOUNT \ I_VERSION STRICTATIME LAZYTIME SUBMOUNT NOREMOTELOCK NOSEC BORN \ ACTIVE NOUSER" SED_PROG= for i in $SYMS; do SED_PROG="$SED_PROG -e s/MS_$i/SB_$i/g"; done # we want files that contain at least one of MS_..., # with fs/namespace.c and fs/pnode.c excluded. L=$(for i in $SYMS; do git grep -w -l MS_$i $FILES; done| sort|uniq|grep -v '^fs/namespace.c'|grep -v '^fs/pnode.c') for f in $L; do sed -i $f $SED_PROG; done Requested-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-11-28 05:05:09 +08:00
sb->s_flags |= SB_ACTIVE;
}
root = cifs_get_root(volume_info, sb);
if (IS_ERR(root))
goto out_super;
cifs_dbg(FYI, "dentry root is: %p\n", root);
goto out;
out_super:
deactivate_locked_super(sb);
out:
cifs_cleanup_volume_info(volume_info);
return root;
out_free:
kfree(cifs_sb->prepath);
kfree(cifs_sb->mountdata);
kfree(cifs_sb);
out_nls:
unload_nls(volume_info->local_nls);
goto out;
}
static struct dentry *
smb3_do_mount(struct file_system_type *fs_type,
int flags, const char *dev_name, void *data)
{
return cifs_smb3_do_mount(fs_type, flags, dev_name, data, true);
}
static struct dentry *
cifs_do_mount(struct file_system_type *fs_type,
int flags, const char *dev_name, void *data)
{
return cifs_smb3_do_mount(fs_type, flags, dev_name, data, false);
}
static ssize_t
cifs_loose_read_iter(struct kiocb *iocb, struct iov_iter *iter)
{
ssize_t rc;
struct inode *inode = file_inode(iocb->ki_filp);
if (iocb->ki_filp->f_flags & O_DIRECT)
return cifs_user_readv(iocb, iter);
rc = cifs_revalidate_mapping(inode);
if (rc)
return rc;
return generic_file_read_iter(iocb, iter);
}
static ssize_t cifs_file_write_iter(struct kiocb *iocb, struct iov_iter *from)
{
struct inode *inode = file_inode(iocb->ki_filp);
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: Sachin Prabhu <sprabhu@redhat.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: Pavel Shilovsky <piastry@etersoft.ru> Signed-off-by: Steve French <smfrench@gmail.com>
2014-03-12 00:11:47 +08:00
struct cifsInodeInfo *cinode = CIFS_I(inode);
ssize_t written;
int rc;
if (iocb->ki_filp->f_flags & O_DIRECT) {
written = cifs_user_writev(iocb, from);
if (written > 0 && CIFS_CACHE_READ(cinode)) {
cifs_zap_mapping(inode);
cifs_dbg(FYI,
"Set no oplock for inode=%p after a write operation\n",
inode);
cinode->oplock = 0;
}
return written;
}
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: Sachin Prabhu <sprabhu@redhat.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: Pavel Shilovsky <piastry@etersoft.ru> Signed-off-by: Steve French <smfrench@gmail.com>
2014-03-12 00:11:47 +08:00
written = cifs_get_writer(cinode);
if (written)
return written;
written = generic_file_write_iter(iocb, from);
if (CIFS_CACHE_WRITE(CIFS_I(inode)))
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: Sachin Prabhu <sprabhu@redhat.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: Pavel Shilovsky <piastry@etersoft.ru> Signed-off-by: Steve French <smfrench@gmail.com>
2014-03-12 00:11:47 +08:00
goto out;
rc = filemap_fdatawrite(inode->i_mapping);
if (rc)
cifs_dbg(FYI, "cifs_file_write_iter: %d rc on %p inode\n",
rc, inode);
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: Sachin Prabhu <sprabhu@redhat.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: Pavel Shilovsky <piastry@etersoft.ru> Signed-off-by: Steve French <smfrench@gmail.com>
2014-03-12 00:11:47 +08:00
out:
cifs_put_writer(cinode);
return written;
}
static loff_t cifs_llseek(struct file *file, loff_t offset, int whence)
{
/*
* whence == SEEK_END || SEEK_DATA || SEEK_HOLE => we must revalidate
* the cached file length
*/
if (whence != SEEK_SET && whence != SEEK_CUR) {
int rc;
struct inode *inode = file_inode(file);
/*
* We need to be sure that all dirty pages are written and the
* server has the newest file length.
*/
if (!CIFS_CACHE_READ(CIFS_I(inode)) && inode->i_mapping &&
inode->i_mapping->nrpages != 0) {
rc = filemap_fdatawait(inode->i_mapping);
if (rc) {
mapping_set_error(inode->i_mapping, rc);
return rc;
}
}
/*
* Some applications poll for the file length in this strange
* way so we must seek to end on non-oplocked files by
* setting the revalidate time to zero.
*/
CIFS_I(inode)->time = 0;
rc = cifs_revalidate_file_attr(file);
if (rc < 0)
return (loff_t)rc;
}
return generic_file_llseek(file, offset, whence);
}
static int
cifs_setlease(struct file *file, long arg, struct file_lock **lease, void **priv)
{
/*
* Note that this is called by vfs setlease with i_lock held to
* protect *lease from going away.
*/
struct inode *inode = file_inode(file);
struct cifsFileInfo *cfile = file->private_data;
if (!(S_ISREG(inode->i_mode)))
return -EINVAL;
/* Check if file is oplocked if this is request for new lease */
if (arg == F_UNLCK ||
((arg == F_RDLCK) && CIFS_CACHE_READ(CIFS_I(inode))) ||
((arg == F_WRLCK) && CIFS_CACHE_WRITE(CIFS_I(inode))))
return generic_setlease(file, arg, lease, priv);
else if (tlink_tcon(cfile->tlink)->local_lease &&
!CIFS_CACHE_READ(CIFS_I(inode)))
/*
* If the server claims to support oplock on this file, then we
* still need to check oplock even if the local_lease mount
* option is set, but there are servers which do not support
* oplock for which this mount option may be useful if the user
* knows that the file won't be changed on the server by anyone
* else.
*/
return generic_setlease(file, arg, lease, priv);
else
return -EAGAIN;
}
struct file_system_type cifs_fs_type = {
.owner = THIS_MODULE,
.name = "cifs",
.mount = cifs_do_mount,
.kill_sb = cifs_kill_sb,
/* .fs_flags */
};
MODULE_ALIAS_FS("cifs");
static struct file_system_type smb3_fs_type = {
.owner = THIS_MODULE,
.name = "smb3",
.mount = smb3_do_mount,
.kill_sb = cifs_kill_sb,
/* .fs_flags */
};
MODULE_ALIAS_FS("smb3");
MODULE_ALIAS("smb3");
const struct inode_operations cifs_dir_inode_ops = {
.create = cifs_create,
.atomic_open = cifs_atomic_open,
.lookup = cifs_lookup,
.getattr = cifs_getattr,
.unlink = cifs_unlink,
.link = cifs_hardlink,
.mkdir = cifs_mkdir,
.rmdir = cifs_rmdir,
.rename = cifs_rename2,
.permission = cifs_permission,
.setattr = cifs_setattr,
.symlink = cifs_symlink,
.mknod = cifs_mknod,
.listxattr = cifs_listxattr,
};
const struct inode_operations cifs_file_inode_ops = {
.setattr = cifs_setattr,
.getattr = cifs_getattr,
.permission = cifs_permission,
.listxattr = cifs_listxattr,
};
const struct inode_operations cifs_symlink_inode_ops = {
.get_link = cifs_get_link,
.permission = cifs_permission,
.listxattr = cifs_listxattr,
};
static int cifs_clone_file_range(struct file *src_file, loff_t off,
struct file *dst_file, loff_t destoff, u64 len)
{
struct inode *src_inode = file_inode(src_file);
struct inode *target_inode = file_inode(dst_file);
struct cifsFileInfo *smb_file_src = src_file->private_data;
struct cifsFileInfo *smb_file_target = dst_file->private_data;
struct cifs_tcon *target_tcon = tlink_tcon(smb_file_target->tlink);
unsigned int xid;
int rc;
cifs_dbg(FYI, "clone range\n");
xid = get_xid();
if (!src_file->private_data || !dst_file->private_data) {
rc = -EBADF;
cifs_dbg(VFS, "missing cifsFileInfo on copy range src file\n");
goto out;
}
/*
* Note: cifs case is easier than btrfs since server responsible for
* checks for proper open modes and file type and if it wants
* server could even support copy of range where source = target
*/
lock_two_nondirectories(target_inode, src_inode);
if (len == 0)
len = src_inode->i_size - off;
cifs_dbg(FYI, "about to flush pages\n");
/* should we flush first and last page first */
truncate_inode_pages_range(&target_inode->i_data, destoff,
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 20:29:47 +08:00
PAGE_ALIGN(destoff + len)-1);
if (target_tcon->ses->server->ops->duplicate_extents)
rc = target_tcon->ses->server->ops->duplicate_extents(xid,
smb_file_src, smb_file_target, off, len, destoff);
else
rc = -EOPNOTSUPP;
/* force revalidate of size and timestamps of target file now
that target is updated on the server */
CIFS_I(target_inode)->time = 0;
/* although unlocking in the reverse order from locking is not
strictly necessary here it is a little cleaner to be consistent */
unlock_two_nondirectories(src_inode, target_inode);
out:
free_xid(xid);
return rc;
}
ssize_t cifs_file_copychunk_range(unsigned int xid,
struct file *src_file, loff_t off,
struct file *dst_file, loff_t destoff,
size_t len, unsigned int flags)
{
struct inode *src_inode = file_inode(src_file);
struct inode *target_inode = file_inode(dst_file);
struct cifsFileInfo *smb_file_src;
struct cifsFileInfo *smb_file_target;
struct cifs_tcon *src_tcon;
struct cifs_tcon *target_tcon;
ssize_t rc;
cifs_dbg(FYI, "copychunk range\n");
if (src_inode == target_inode) {
rc = -EINVAL;
goto out;
}
if (!src_file->private_data || !dst_file->private_data) {
rc = -EBADF;
cifs_dbg(VFS, "missing cifsFileInfo on copy range src file\n");
goto out;
}
rc = -EXDEV;
smb_file_target = dst_file->private_data;
smb_file_src = src_file->private_data;
src_tcon = tlink_tcon(smb_file_src->tlink);
target_tcon = tlink_tcon(smb_file_target->tlink);
if (src_tcon->ses != target_tcon->ses) {
cifs_dbg(VFS, "source and target of copy not on same server\n");
goto out;
}
/*
* Note: cifs case is easier than btrfs since server responsible for
* checks for proper open modes and file type and if it wants
* server could even support copy of range where source = target
*/
lock_two_nondirectories(target_inode, src_inode);
cifs_dbg(FYI, "about to flush pages\n");
/* should we flush first and last page first */
truncate_inode_pages(&target_inode->i_data, 0);
if (target_tcon->ses->server->ops->copychunk_range)
rc = target_tcon->ses->server->ops->copychunk_range(xid,
smb_file_src, smb_file_target, off, len, destoff);
else
rc = -EOPNOTSUPP;
/* force revalidate of size and timestamps of target file now
* that target is updated on the server
*/
CIFS_I(target_inode)->time = 0;
/* although unlocking in the reverse order from locking is not
* strictly necessary here it is a little cleaner to be consistent
*/
unlock_two_nondirectories(src_inode, target_inode);
out:
return rc;
}
/*
* Directory operations under CIFS/SMB2/SMB3 are synchronous, so fsync()
* is a dummy operation.
*/
static int cifs_dir_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
cifs_dbg(FYI, "Sync directory - name: %pD datasync: 0x%x\n",
file, datasync);
return 0;
}
static ssize_t cifs_copy_file_range(struct file *src_file, loff_t off,
struct file *dst_file, loff_t destoff,
size_t len, unsigned int flags)
{
unsigned int xid = get_xid();
ssize_t rc;
rc = cifs_file_copychunk_range(xid, src_file, off, dst_file, destoff,
len, flags);
free_xid(xid);
return rc;
}
const struct file_operations cifs_file_ops = {
.read_iter = cifs_loose_read_iter,
.write_iter = cifs_file_write_iter,
.open = cifs_open,
.release = cifs_close,
.lock = cifs_lock,
.fsync = cifs_fsync,
.flush = cifs_flush,
.mmap = cifs_file_mmap,
.splice_read = generic_file_splice_read,
.splice_write = iter_file_splice_write,
.llseek = cifs_llseek,
.unlocked_ioctl = cifs_ioctl,
.copy_file_range = cifs_copy_file_range,
.clone_file_range = cifs_clone_file_range,
.setlease = cifs_setlease,
.fallocate = cifs_fallocate,
};
const struct file_operations cifs_file_strict_ops = {
.read_iter = cifs_strict_readv,
.write_iter = cifs_strict_writev,
.open = cifs_open,
.release = cifs_close,
.lock = cifs_lock,
.fsync = cifs_strict_fsync,
.flush = cifs_flush,
.mmap = cifs_file_strict_mmap,
.splice_read = generic_file_splice_read,
.splice_write = iter_file_splice_write,
.llseek = cifs_llseek,
.unlocked_ioctl = cifs_ioctl,
.copy_file_range = cifs_copy_file_range,
.clone_file_range = cifs_clone_file_range,
.setlease = cifs_setlease,
.fallocate = cifs_fallocate,
};
const struct file_operations cifs_file_direct_ops = {
/* BB reevaluate whether they can be done with directio, no cache */
.read_iter = cifs_user_readv,
.write_iter = cifs_user_writev,
.open = cifs_open,
.release = cifs_close,
.lock = cifs_lock,
.fsync = cifs_fsync,
.flush = cifs_flush,
.mmap = cifs_file_mmap,
.splice_read = generic_file_splice_read,
.splice_write = iter_file_splice_write,
.unlocked_ioctl = cifs_ioctl,
.copy_file_range = cifs_copy_file_range,
.clone_file_range = cifs_clone_file_range,
.llseek = cifs_llseek,
.setlease = cifs_setlease,
.fallocate = cifs_fallocate,
};
const struct file_operations cifs_file_nobrl_ops = {
.read_iter = cifs_loose_read_iter,
.write_iter = cifs_file_write_iter,
.open = cifs_open,
.release = cifs_close,
.fsync = cifs_fsync,
.flush = cifs_flush,
.mmap = cifs_file_mmap,
.splice_read = generic_file_splice_read,
.splice_write = iter_file_splice_write,
.llseek = cifs_llseek,
.unlocked_ioctl = cifs_ioctl,
.copy_file_range = cifs_copy_file_range,
.clone_file_range = cifs_clone_file_range,
.setlease = cifs_setlease,
.fallocate = cifs_fallocate,
};
const struct file_operations cifs_file_strict_nobrl_ops = {
.read_iter = cifs_strict_readv,
.write_iter = cifs_strict_writev,
.open = cifs_open,
.release = cifs_close,
.fsync = cifs_strict_fsync,
.flush = cifs_flush,
.mmap = cifs_file_strict_mmap,
.splice_read = generic_file_splice_read,
.splice_write = iter_file_splice_write,
.llseek = cifs_llseek,
.unlocked_ioctl = cifs_ioctl,
.copy_file_range = cifs_copy_file_range,
.clone_file_range = cifs_clone_file_range,
.setlease = cifs_setlease,
.fallocate = cifs_fallocate,
};
const struct file_operations cifs_file_direct_nobrl_ops = {
/* BB reevaluate whether they can be done with directio, no cache */
.read_iter = cifs_user_readv,
.write_iter = cifs_user_writev,
.open = cifs_open,
.release = cifs_close,
.fsync = cifs_fsync,
.flush = cifs_flush,
.mmap = cifs_file_mmap,
.splice_read = generic_file_splice_read,
.splice_write = iter_file_splice_write,
.unlocked_ioctl = cifs_ioctl,
.copy_file_range = cifs_copy_file_range,
.clone_file_range = cifs_clone_file_range,
.llseek = cifs_llseek,
.setlease = cifs_setlease,
.fallocate = cifs_fallocate,
};
const struct file_operations cifs_dir_ops = {
.iterate_shared = cifs_readdir,
.release = cifs_closedir,
.read = generic_read_dir,
.unlocked_ioctl = cifs_ioctl,
.copy_file_range = cifs_copy_file_range,
.clone_file_range = cifs_clone_file_range,
.llseek = generic_file_llseek,
.fsync = cifs_dir_fsync,
};
static void
cifs_init_once(void *inode)
{
struct cifsInodeInfo *cifsi = inode;
inode_init_once(&cifsi->vfs_inode);
init_rwsem(&cifsi->lock_sem);
}
static int __init
cifs_init_inodecache(void)
{
cifs_inode_cachep = kmem_cache_create("cifs_inode_cache",
sizeof(struct cifsInodeInfo),
0, (SLAB_RECLAIM_ACCOUNT|
2016-01-15 07:18:21 +08:00
SLAB_MEM_SPREAD|SLAB_ACCOUNT),
cifs_init_once);
if (cifs_inode_cachep == NULL)
return -ENOMEM;
return 0;
}
static void
cifs_destroy_inodecache(void)
{
/*
* Make sure all delayed rcu free inodes are flushed before we
* destroy cache.
*/
rcu_barrier();
kmem_cache_destroy(cifs_inode_cachep);
}
static int
cifs_init_request_bufs(void)
{
/*
* SMB2 maximum header size is bigger than CIFS one - no problems to
* allocate some more bytes for CIFS.
*/
size_t max_hdr_size = MAX_SMB2_HDR_SIZE;
if (CIFSMaxBufSize < 8192) {
/* Buffer size can not be smaller than 2 * PATH_MAX since maximum
Unicode path name has to fit in any SMB/CIFS path based frames */
CIFSMaxBufSize = 8192;
} else if (CIFSMaxBufSize > 1024*127) {
CIFSMaxBufSize = 1024 * 127;
} else {
CIFSMaxBufSize &= 0x1FE00; /* Round size to even 512 byte mult*/
}
/*
cifs_dbg(VFS, "CIFSMaxBufSize %d 0x%x\n",
CIFSMaxBufSize, CIFSMaxBufSize);
*/
cifs_req_cachep = kmem_cache_create_usercopy("cifs_request",
CIFSMaxBufSize + max_hdr_size, 0,
SLAB_HWCACHE_ALIGN, 0,
CIFSMaxBufSize + max_hdr_size,
NULL);
if (cifs_req_cachep == NULL)
return -ENOMEM;
if (cifs_min_rcv < 1)
cifs_min_rcv = 1;
else if (cifs_min_rcv > 64) {
cifs_min_rcv = 64;
cifs_dbg(VFS, "cifs_min_rcv set to maximum (64)\n");
}
cifs_req_poolp = mempool_create_slab_pool(cifs_min_rcv,
cifs_req_cachep);
if (cifs_req_poolp == NULL) {
kmem_cache_destroy(cifs_req_cachep);
return -ENOMEM;
}
/* MAX_CIFS_SMALL_BUFFER_SIZE bytes is enough for most SMB responses and
almost all handle based requests (but not write response, nor is it
sufficient for path based requests). A smaller size would have
been more efficient (compacting multiple slab items on one 4k page)
for the case in which debug was on, but this larger size allows
more SMBs to use small buffer alloc and is still much more
efficient to alloc 1 per page off the slab compared to 17K (5page)
alloc of large cifs buffers even when page debugging is on */
cifs_sm_req_cachep = kmem_cache_create_usercopy("cifs_small_rq",
MAX_CIFS_SMALL_BUFFER_SIZE, 0, SLAB_HWCACHE_ALIGN,
0, MAX_CIFS_SMALL_BUFFER_SIZE, NULL);
if (cifs_sm_req_cachep == NULL) {
mempool_destroy(cifs_req_poolp);
kmem_cache_destroy(cifs_req_cachep);
return -ENOMEM;
}
if (cifs_min_small < 2)
cifs_min_small = 2;
else if (cifs_min_small > 256) {
cifs_min_small = 256;
cifs_dbg(FYI, "cifs_min_small set to maximum (256)\n");
}
cifs_sm_req_poolp = mempool_create_slab_pool(cifs_min_small,
cifs_sm_req_cachep);
if (cifs_sm_req_poolp == NULL) {
mempool_destroy(cifs_req_poolp);
kmem_cache_destroy(cifs_req_cachep);
kmem_cache_destroy(cifs_sm_req_cachep);
return -ENOMEM;
}
return 0;
}
static void
cifs_destroy_request_bufs(void)
{
mempool_destroy(cifs_req_poolp);
kmem_cache_destroy(cifs_req_cachep);
mempool_destroy(cifs_sm_req_poolp);
kmem_cache_destroy(cifs_sm_req_cachep);
}
static int
cifs_init_mids(void)
{
cifs_mid_cachep = kmem_cache_create("cifs_mpx_ids",
sizeof(struct mid_q_entry), 0,
SLAB_HWCACHE_ALIGN, NULL);
if (cifs_mid_cachep == NULL)
return -ENOMEM;
/* 3 is a reasonable minimum number of simultaneous operations */
cifs_mid_poolp = mempool_create_slab_pool(3, cifs_mid_cachep);
if (cifs_mid_poolp == NULL) {
kmem_cache_destroy(cifs_mid_cachep);
return -ENOMEM;
}
return 0;
}
static void
cifs_destroy_mids(void)
{
mempool_destroy(cifs_mid_poolp);
kmem_cache_destroy(cifs_mid_cachep);
}
static int __init
init_cifs(void)
{
int rc = 0;
cifs_proc_init();
INIT_LIST_HEAD(&cifs_tcp_ses_list);
#ifdef CONFIG_CIFS_DNOTIFY_EXPERIMENTAL /* unused temporarily */
INIT_LIST_HEAD(&GlobalDnotifyReqList);
INIT_LIST_HEAD(&GlobalDnotifyRsp_Q);
#endif /* was needed for dnotify, and will be needed for inotify when VFS fix */
/*
* Initialize Global counters
*/
atomic_set(&sesInfoAllocCount, 0);
atomic_set(&tconInfoAllocCount, 0);
atomic_set(&tcpSesAllocCount, 0);
atomic_set(&tcpSesReconnectCount, 0);
atomic_set(&tconInfoReconnectCount, 0);
atomic_set(&bufAllocCount, 0);
atomic_set(&smBufAllocCount, 0);
#ifdef CONFIG_CIFS_STATS2
atomic_set(&totBufAllocCount, 0);
atomic_set(&totSmBufAllocCount, 0);
#endif /* CONFIG_CIFS_STATS2 */
atomic_set(&midCount, 0);
GlobalCurrentXid = 0;
GlobalTotalActiveXid = 0;
GlobalMaxActiveXid = 0;
spin_lock_init(&cifs_tcp_ses_lock);
spin_lock_init(&GlobalMid_Lock);
cifs_lock_secret = get_random_u32();
if (cifs_max_pending < 2) {
cifs_max_pending = 2;
cifs_dbg(FYI, "cifs_max_pending set to min of 2\n");
} else if (cifs_max_pending > CIFS_MAX_REQ) {
cifs_max_pending = CIFS_MAX_REQ;
cifs_dbg(FYI, "cifs_max_pending set to max of %u\n",
CIFS_MAX_REQ);
}
cifsiod_wq = alloc_workqueue("cifsiod", WQ_FREEZABLE|WQ_MEM_RECLAIM, 0);
if (!cifsiod_wq) {
rc = -ENOMEM;
goto out_clean_proc;
}
CIFS: fix oplock break deadlocks When the final cifsFileInfo_put() is called from cifsiod and an oplock break work is queued, lockdep complains loudly: ============================================= [ INFO: possible recursive locking detected ] 4.11.0+ #21 Not tainted --------------------------------------------- kworker/0:2/78 is trying to acquire lock: ("cifsiod"){++++.+}, at: flush_work+0x215/0x350 but task is already holding lock: ("cifsiod"){++++.+}, at: process_one_work+0x255/0x8e0 other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock("cifsiod"); lock("cifsiod"); *** DEADLOCK *** May be due to missing lock nesting notation 2 locks held by kworker/0:2/78: #0: ("cifsiod"){++++.+}, at: process_one_work+0x255/0x8e0 #1: ((&wdata->work)){+.+...}, at: process_one_work+0x255/0x8e0 stack backtrace: CPU: 0 PID: 78 Comm: kworker/0:2 Not tainted 4.11.0+ #21 Workqueue: cifsiod cifs_writev_complete Call Trace: dump_stack+0x85/0xc2 __lock_acquire+0x17dd/0x2260 ? match_held_lock+0x20/0x2b0 ? trace_hardirqs_off_caller+0x86/0x130 ? mark_lock+0xa6/0x920 lock_acquire+0xcc/0x260 ? lock_acquire+0xcc/0x260 ? flush_work+0x215/0x350 flush_work+0x236/0x350 ? flush_work+0x215/0x350 ? destroy_worker+0x170/0x170 __cancel_work_timer+0x17d/0x210 ? ___preempt_schedule+0x16/0x18 cancel_work_sync+0x10/0x20 cifsFileInfo_put+0x338/0x7f0 cifs_writedata_release+0x2a/0x40 ? cifs_writedata_release+0x2a/0x40 cifs_writev_complete+0x29d/0x850 ? preempt_count_sub+0x18/0xd0 process_one_work+0x304/0x8e0 worker_thread+0x9b/0x6a0 kthread+0x1b2/0x200 ? process_one_work+0x8e0/0x8e0 ? kthread_create_on_node+0x40/0x40 ret_from_fork+0x31/0x40 This is a real warning. Since the oplock is queued on the same workqueue this can deadlock if there is only one worker thread active for the workqueue (which will be the case during memory pressure when the rescuer thread is handling it). Furthermore, there is at least one other kind of hang possible due to the oplock break handling if there is only worker. (This can be reproduced without introducing memory pressure by having passing 1 for the max_active parameter of cifsiod.) cifs_oplock_break() can wait indefintely in the filemap_fdatawait() while the cifs_writev_complete() work is blocked: sysrq: SysRq : Show Blocked State task PC stack pid father kworker/0:1 D 0 16 2 0x00000000 Workqueue: cifsiod cifs_oplock_break Call Trace: __schedule+0x562/0xf40 ? mark_held_locks+0x4a/0xb0 schedule+0x57/0xe0 io_schedule+0x21/0x50 wait_on_page_bit+0x143/0x190 ? add_to_page_cache_lru+0x150/0x150 __filemap_fdatawait_range+0x134/0x190 ? do_writepages+0x51/0x70 filemap_fdatawait_range+0x14/0x30 filemap_fdatawait+0x3b/0x40 cifs_oplock_break+0x651/0x710 ? preempt_count_sub+0x18/0xd0 process_one_work+0x304/0x8e0 worker_thread+0x9b/0x6a0 kthread+0x1b2/0x200 ? process_one_work+0x8e0/0x8e0 ? kthread_create_on_node+0x40/0x40 ret_from_fork+0x31/0x40 dd D 0 683 171 0x00000000 Call Trace: __schedule+0x562/0xf40 ? mark_held_locks+0x29/0xb0 schedule+0x57/0xe0 io_schedule+0x21/0x50 wait_on_page_bit+0x143/0x190 ? add_to_page_cache_lru+0x150/0x150 __filemap_fdatawait_range+0x134/0x190 ? do_writepages+0x51/0x70 filemap_fdatawait_range+0x14/0x30 filemap_fdatawait+0x3b/0x40 filemap_write_and_wait+0x4e/0x70 cifs_flush+0x6a/0xb0 filp_close+0x52/0xa0 __close_fd+0xdc/0x150 SyS_close+0x33/0x60 entry_SYSCALL_64_fastpath+0x1f/0xbe Showing all locks held in the system: 2 locks held by kworker/0:1/16: #0: ("cifsiod"){.+.+.+}, at: process_one_work+0x255/0x8e0 #1: ((&cfile->oplock_break)){+.+.+.}, at: process_one_work+0x255/0x8e0 Showing busy workqueues and worker pools: workqueue cifsiod: flags=0xc pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=1/1 in-flight: 16:cifs_oplock_break delayed: cifs_writev_complete, cifs_echo_request pool 0: cpus=0 node=0 flags=0x0 nice=0 hung=0s workers=3 idle: 750 3 Fix these problems by creating a a new workqueue (with a rescuer) for the oplock break work. Signed-off-by: Rabin Vincent <rabinv@axis.com> Signed-off-by: Steve French <smfrench@gmail.com> CC: Stable <stable@vger.kernel.org>
2017-05-03 23:54:01 +08:00
cifsoplockd_wq = alloc_workqueue("cifsoplockd",
WQ_FREEZABLE|WQ_MEM_RECLAIM, 0);
if (!cifsoplockd_wq) {
rc = -ENOMEM;
goto out_destroy_cifsiod_wq;
}
rc = cifs_fscache_register();
if (rc)
CIFS: fix oplock break deadlocks When the final cifsFileInfo_put() is called from cifsiod and an oplock break work is queued, lockdep complains loudly: ============================================= [ INFO: possible recursive locking detected ] 4.11.0+ #21 Not tainted --------------------------------------------- kworker/0:2/78 is trying to acquire lock: ("cifsiod"){++++.+}, at: flush_work+0x215/0x350 but task is already holding lock: ("cifsiod"){++++.+}, at: process_one_work+0x255/0x8e0 other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock("cifsiod"); lock("cifsiod"); *** DEADLOCK *** May be due to missing lock nesting notation 2 locks held by kworker/0:2/78: #0: ("cifsiod"){++++.+}, at: process_one_work+0x255/0x8e0 #1: ((&wdata->work)){+.+...}, at: process_one_work+0x255/0x8e0 stack backtrace: CPU: 0 PID: 78 Comm: kworker/0:2 Not tainted 4.11.0+ #21 Workqueue: cifsiod cifs_writev_complete Call Trace: dump_stack+0x85/0xc2 __lock_acquire+0x17dd/0x2260 ? match_held_lock+0x20/0x2b0 ? trace_hardirqs_off_caller+0x86/0x130 ? mark_lock+0xa6/0x920 lock_acquire+0xcc/0x260 ? lock_acquire+0xcc/0x260 ? flush_work+0x215/0x350 flush_work+0x236/0x350 ? flush_work+0x215/0x350 ? destroy_worker+0x170/0x170 __cancel_work_timer+0x17d/0x210 ? ___preempt_schedule+0x16/0x18 cancel_work_sync+0x10/0x20 cifsFileInfo_put+0x338/0x7f0 cifs_writedata_release+0x2a/0x40 ? cifs_writedata_release+0x2a/0x40 cifs_writev_complete+0x29d/0x850 ? preempt_count_sub+0x18/0xd0 process_one_work+0x304/0x8e0 worker_thread+0x9b/0x6a0 kthread+0x1b2/0x200 ? process_one_work+0x8e0/0x8e0 ? kthread_create_on_node+0x40/0x40 ret_from_fork+0x31/0x40 This is a real warning. Since the oplock is queued on the same workqueue this can deadlock if there is only one worker thread active for the workqueue (which will be the case during memory pressure when the rescuer thread is handling it). Furthermore, there is at least one other kind of hang possible due to the oplock break handling if there is only worker. (This can be reproduced without introducing memory pressure by having passing 1 for the max_active parameter of cifsiod.) cifs_oplock_break() can wait indefintely in the filemap_fdatawait() while the cifs_writev_complete() work is blocked: sysrq: SysRq : Show Blocked State task PC stack pid father kworker/0:1 D 0 16 2 0x00000000 Workqueue: cifsiod cifs_oplock_break Call Trace: __schedule+0x562/0xf40 ? mark_held_locks+0x4a/0xb0 schedule+0x57/0xe0 io_schedule+0x21/0x50 wait_on_page_bit+0x143/0x190 ? add_to_page_cache_lru+0x150/0x150 __filemap_fdatawait_range+0x134/0x190 ? do_writepages+0x51/0x70 filemap_fdatawait_range+0x14/0x30 filemap_fdatawait+0x3b/0x40 cifs_oplock_break+0x651/0x710 ? preempt_count_sub+0x18/0xd0 process_one_work+0x304/0x8e0 worker_thread+0x9b/0x6a0 kthread+0x1b2/0x200 ? process_one_work+0x8e0/0x8e0 ? kthread_create_on_node+0x40/0x40 ret_from_fork+0x31/0x40 dd D 0 683 171 0x00000000 Call Trace: __schedule+0x562/0xf40 ? mark_held_locks+0x29/0xb0 schedule+0x57/0xe0 io_schedule+0x21/0x50 wait_on_page_bit+0x143/0x190 ? add_to_page_cache_lru+0x150/0x150 __filemap_fdatawait_range+0x134/0x190 ? do_writepages+0x51/0x70 filemap_fdatawait_range+0x14/0x30 filemap_fdatawait+0x3b/0x40 filemap_write_and_wait+0x4e/0x70 cifs_flush+0x6a/0xb0 filp_close+0x52/0xa0 __close_fd+0xdc/0x150 SyS_close+0x33/0x60 entry_SYSCALL_64_fastpath+0x1f/0xbe Showing all locks held in the system: 2 locks held by kworker/0:1/16: #0: ("cifsiod"){.+.+.+}, at: process_one_work+0x255/0x8e0 #1: ((&cfile->oplock_break)){+.+.+.}, at: process_one_work+0x255/0x8e0 Showing busy workqueues and worker pools: workqueue cifsiod: flags=0xc pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=1/1 in-flight: 16:cifs_oplock_break delayed: cifs_writev_complete, cifs_echo_request pool 0: cpus=0 node=0 flags=0x0 nice=0 hung=0s workers=3 idle: 750 3 Fix these problems by creating a a new workqueue (with a rescuer) for the oplock break work. Signed-off-by: Rabin Vincent <rabinv@axis.com> Signed-off-by: Steve French <smfrench@gmail.com> CC: Stable <stable@vger.kernel.org>
2017-05-03 23:54:01 +08:00
goto out_destroy_cifsoplockd_wq;
rc = cifs_init_inodecache();
if (rc)
goto out_unreg_fscache;
rc = cifs_init_mids();
if (rc)
goto out_destroy_inodecache;
rc = cifs_init_request_bufs();
if (rc)
goto out_destroy_mids;
#ifdef CONFIG_CIFS_UPCALL
rc = init_cifs_spnego();
if (rc)
goto out_destroy_request_bufs;
#endif /* CONFIG_CIFS_UPCALL */
#ifdef CONFIG_CIFS_ACL
rc = init_cifs_idmap();
if (rc)
goto out_register_key_type;
#endif /* CONFIG_CIFS_ACL */
rc = register_filesystem(&cifs_fs_type);
if (rc)
goto out_init_cifs_idmap;
rc = register_filesystem(&smb3_fs_type);
if (rc) {
unregister_filesystem(&cifs_fs_type);
goto out_init_cifs_idmap;
}
return 0;
out_init_cifs_idmap:
#ifdef CONFIG_CIFS_ACL
exit_cifs_idmap();
out_register_key_type:
#endif
#ifdef CONFIG_CIFS_UPCALL
exit_cifs_spnego();
out_destroy_request_bufs:
#endif
cifs_destroy_request_bufs();
out_destroy_mids:
cifs_destroy_mids();
out_destroy_inodecache:
cifs_destroy_inodecache();
out_unreg_fscache:
cifs_fscache_unregister();
CIFS: fix oplock break deadlocks When the final cifsFileInfo_put() is called from cifsiod and an oplock break work is queued, lockdep complains loudly: ============================================= [ INFO: possible recursive locking detected ] 4.11.0+ #21 Not tainted --------------------------------------------- kworker/0:2/78 is trying to acquire lock: ("cifsiod"){++++.+}, at: flush_work+0x215/0x350 but task is already holding lock: ("cifsiod"){++++.+}, at: process_one_work+0x255/0x8e0 other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock("cifsiod"); lock("cifsiod"); *** DEADLOCK *** May be due to missing lock nesting notation 2 locks held by kworker/0:2/78: #0: ("cifsiod"){++++.+}, at: process_one_work+0x255/0x8e0 #1: ((&wdata->work)){+.+...}, at: process_one_work+0x255/0x8e0 stack backtrace: CPU: 0 PID: 78 Comm: kworker/0:2 Not tainted 4.11.0+ #21 Workqueue: cifsiod cifs_writev_complete Call Trace: dump_stack+0x85/0xc2 __lock_acquire+0x17dd/0x2260 ? match_held_lock+0x20/0x2b0 ? trace_hardirqs_off_caller+0x86/0x130 ? mark_lock+0xa6/0x920 lock_acquire+0xcc/0x260 ? lock_acquire+0xcc/0x260 ? flush_work+0x215/0x350 flush_work+0x236/0x350 ? flush_work+0x215/0x350 ? destroy_worker+0x170/0x170 __cancel_work_timer+0x17d/0x210 ? ___preempt_schedule+0x16/0x18 cancel_work_sync+0x10/0x20 cifsFileInfo_put+0x338/0x7f0 cifs_writedata_release+0x2a/0x40 ? cifs_writedata_release+0x2a/0x40 cifs_writev_complete+0x29d/0x850 ? preempt_count_sub+0x18/0xd0 process_one_work+0x304/0x8e0 worker_thread+0x9b/0x6a0 kthread+0x1b2/0x200 ? process_one_work+0x8e0/0x8e0 ? kthread_create_on_node+0x40/0x40 ret_from_fork+0x31/0x40 This is a real warning. Since the oplock is queued on the same workqueue this can deadlock if there is only one worker thread active for the workqueue (which will be the case during memory pressure when the rescuer thread is handling it). Furthermore, there is at least one other kind of hang possible due to the oplock break handling if there is only worker. (This can be reproduced without introducing memory pressure by having passing 1 for the max_active parameter of cifsiod.) cifs_oplock_break() can wait indefintely in the filemap_fdatawait() while the cifs_writev_complete() work is blocked: sysrq: SysRq : Show Blocked State task PC stack pid father kworker/0:1 D 0 16 2 0x00000000 Workqueue: cifsiod cifs_oplock_break Call Trace: __schedule+0x562/0xf40 ? mark_held_locks+0x4a/0xb0 schedule+0x57/0xe0 io_schedule+0x21/0x50 wait_on_page_bit+0x143/0x190 ? add_to_page_cache_lru+0x150/0x150 __filemap_fdatawait_range+0x134/0x190 ? do_writepages+0x51/0x70 filemap_fdatawait_range+0x14/0x30 filemap_fdatawait+0x3b/0x40 cifs_oplock_break+0x651/0x710 ? preempt_count_sub+0x18/0xd0 process_one_work+0x304/0x8e0 worker_thread+0x9b/0x6a0 kthread+0x1b2/0x200 ? process_one_work+0x8e0/0x8e0 ? kthread_create_on_node+0x40/0x40 ret_from_fork+0x31/0x40 dd D 0 683 171 0x00000000 Call Trace: __schedule+0x562/0xf40 ? mark_held_locks+0x29/0xb0 schedule+0x57/0xe0 io_schedule+0x21/0x50 wait_on_page_bit+0x143/0x190 ? add_to_page_cache_lru+0x150/0x150 __filemap_fdatawait_range+0x134/0x190 ? do_writepages+0x51/0x70 filemap_fdatawait_range+0x14/0x30 filemap_fdatawait+0x3b/0x40 filemap_write_and_wait+0x4e/0x70 cifs_flush+0x6a/0xb0 filp_close+0x52/0xa0 __close_fd+0xdc/0x150 SyS_close+0x33/0x60 entry_SYSCALL_64_fastpath+0x1f/0xbe Showing all locks held in the system: 2 locks held by kworker/0:1/16: #0: ("cifsiod"){.+.+.+}, at: process_one_work+0x255/0x8e0 #1: ((&cfile->oplock_break)){+.+.+.}, at: process_one_work+0x255/0x8e0 Showing busy workqueues and worker pools: workqueue cifsiod: flags=0xc pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=1/1 in-flight: 16:cifs_oplock_break delayed: cifs_writev_complete, cifs_echo_request pool 0: cpus=0 node=0 flags=0x0 nice=0 hung=0s workers=3 idle: 750 3 Fix these problems by creating a a new workqueue (with a rescuer) for the oplock break work. Signed-off-by: Rabin Vincent <rabinv@axis.com> Signed-off-by: Steve French <smfrench@gmail.com> CC: Stable <stable@vger.kernel.org>
2017-05-03 23:54:01 +08:00
out_destroy_cifsoplockd_wq:
destroy_workqueue(cifsoplockd_wq);
out_destroy_cifsiod_wq:
destroy_workqueue(cifsiod_wq);
out_clean_proc:
cifs_proc_clean();
return rc;
}
static void __exit
exit_cifs(void)
{
cifs_dbg(NOISY, "exit_smb3\n");
unregister_filesystem(&cifs_fs_type);
unregister_filesystem(&smb3_fs_type);
cifs_dfs_release_automount_timer();
#ifdef CONFIG_CIFS_ACL
exit_cifs_idmap();
#endif
#ifdef CONFIG_CIFS_UPCALL
exit_cifs_spnego();
#endif
cifs_destroy_request_bufs();
cifs_destroy_mids();
cifs_destroy_inodecache();
cifs_fscache_unregister();
CIFS: fix oplock break deadlocks When the final cifsFileInfo_put() is called from cifsiod and an oplock break work is queued, lockdep complains loudly: ============================================= [ INFO: possible recursive locking detected ] 4.11.0+ #21 Not tainted --------------------------------------------- kworker/0:2/78 is trying to acquire lock: ("cifsiod"){++++.+}, at: flush_work+0x215/0x350 but task is already holding lock: ("cifsiod"){++++.+}, at: process_one_work+0x255/0x8e0 other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock("cifsiod"); lock("cifsiod"); *** DEADLOCK *** May be due to missing lock nesting notation 2 locks held by kworker/0:2/78: #0: ("cifsiod"){++++.+}, at: process_one_work+0x255/0x8e0 #1: ((&wdata->work)){+.+...}, at: process_one_work+0x255/0x8e0 stack backtrace: CPU: 0 PID: 78 Comm: kworker/0:2 Not tainted 4.11.0+ #21 Workqueue: cifsiod cifs_writev_complete Call Trace: dump_stack+0x85/0xc2 __lock_acquire+0x17dd/0x2260 ? match_held_lock+0x20/0x2b0 ? trace_hardirqs_off_caller+0x86/0x130 ? mark_lock+0xa6/0x920 lock_acquire+0xcc/0x260 ? lock_acquire+0xcc/0x260 ? flush_work+0x215/0x350 flush_work+0x236/0x350 ? flush_work+0x215/0x350 ? destroy_worker+0x170/0x170 __cancel_work_timer+0x17d/0x210 ? ___preempt_schedule+0x16/0x18 cancel_work_sync+0x10/0x20 cifsFileInfo_put+0x338/0x7f0 cifs_writedata_release+0x2a/0x40 ? cifs_writedata_release+0x2a/0x40 cifs_writev_complete+0x29d/0x850 ? preempt_count_sub+0x18/0xd0 process_one_work+0x304/0x8e0 worker_thread+0x9b/0x6a0 kthread+0x1b2/0x200 ? process_one_work+0x8e0/0x8e0 ? kthread_create_on_node+0x40/0x40 ret_from_fork+0x31/0x40 This is a real warning. Since the oplock is queued on the same workqueue this can deadlock if there is only one worker thread active for the workqueue (which will be the case during memory pressure when the rescuer thread is handling it). Furthermore, there is at least one other kind of hang possible due to the oplock break handling if there is only worker. (This can be reproduced without introducing memory pressure by having passing 1 for the max_active parameter of cifsiod.) cifs_oplock_break() can wait indefintely in the filemap_fdatawait() while the cifs_writev_complete() work is blocked: sysrq: SysRq : Show Blocked State task PC stack pid father kworker/0:1 D 0 16 2 0x00000000 Workqueue: cifsiod cifs_oplock_break Call Trace: __schedule+0x562/0xf40 ? mark_held_locks+0x4a/0xb0 schedule+0x57/0xe0 io_schedule+0x21/0x50 wait_on_page_bit+0x143/0x190 ? add_to_page_cache_lru+0x150/0x150 __filemap_fdatawait_range+0x134/0x190 ? do_writepages+0x51/0x70 filemap_fdatawait_range+0x14/0x30 filemap_fdatawait+0x3b/0x40 cifs_oplock_break+0x651/0x710 ? preempt_count_sub+0x18/0xd0 process_one_work+0x304/0x8e0 worker_thread+0x9b/0x6a0 kthread+0x1b2/0x200 ? process_one_work+0x8e0/0x8e0 ? kthread_create_on_node+0x40/0x40 ret_from_fork+0x31/0x40 dd D 0 683 171 0x00000000 Call Trace: __schedule+0x562/0xf40 ? mark_held_locks+0x29/0xb0 schedule+0x57/0xe0 io_schedule+0x21/0x50 wait_on_page_bit+0x143/0x190 ? add_to_page_cache_lru+0x150/0x150 __filemap_fdatawait_range+0x134/0x190 ? do_writepages+0x51/0x70 filemap_fdatawait_range+0x14/0x30 filemap_fdatawait+0x3b/0x40 filemap_write_and_wait+0x4e/0x70 cifs_flush+0x6a/0xb0 filp_close+0x52/0xa0 __close_fd+0xdc/0x150 SyS_close+0x33/0x60 entry_SYSCALL_64_fastpath+0x1f/0xbe Showing all locks held in the system: 2 locks held by kworker/0:1/16: #0: ("cifsiod"){.+.+.+}, at: process_one_work+0x255/0x8e0 #1: ((&cfile->oplock_break)){+.+.+.}, at: process_one_work+0x255/0x8e0 Showing busy workqueues and worker pools: workqueue cifsiod: flags=0xc pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=1/1 in-flight: 16:cifs_oplock_break delayed: cifs_writev_complete, cifs_echo_request pool 0: cpus=0 node=0 flags=0x0 nice=0 hung=0s workers=3 idle: 750 3 Fix these problems by creating a a new workqueue (with a rescuer) for the oplock break work. Signed-off-by: Rabin Vincent <rabinv@axis.com> Signed-off-by: Steve French <smfrench@gmail.com> CC: Stable <stable@vger.kernel.org>
2017-05-03 23:54:01 +08:00
destroy_workqueue(cifsoplockd_wq);
destroy_workqueue(cifsiod_wq);
cifs_proc_clean();
}
MODULE_AUTHOR("Steve French <sfrench@us.ibm.com>");
MODULE_LICENSE("GPL"); /* combination of LGPL + GPL source behaves as GPL */
MODULE_DESCRIPTION
("VFS to access servers complying with the SNIA CIFS Specification "
"e.g. Samba and Windows");
MODULE_VERSION(CIFS_VERSION);
MODULE_SOFTDEP("pre: arc4");
MODULE_SOFTDEP("pre: des");
MODULE_SOFTDEP("pre: ecb");
MODULE_SOFTDEP("pre: hmac");
MODULE_SOFTDEP("pre: md4");
MODULE_SOFTDEP("pre: md5");
MODULE_SOFTDEP("pre: nls");
MODULE_SOFTDEP("pre: aes");
MODULE_SOFTDEP("pre: cmac");
MODULE_SOFTDEP("pre: sha256");
MODULE_SOFTDEP("pre: sha512");
MODULE_SOFTDEP("pre: aead2");
MODULE_SOFTDEP("pre: ccm");
module_init(init_cifs)
module_exit(exit_cifs)