Commit 5e3cc1ee authored by Hou Tao's avatar Hou Tao Committed by Dominique Martinet

9p: use inode->i_lock to protect i_size_write() under 32-bit

Use inode->i_lock to protect i_size_write(), else i_size_read() in
generic_fillattr() may loop infinitely in read_seqcount_begin() when
multiple processes invoke v9fs_vfs_getattr() or v9fs_vfs_getattr_dotl()
simultaneously under 32-bit SMP environment, and a soft lockup will be
triggered as show below:

  watchdog: BUG: soft lockup - CPU#5 stuck for 22s! [stat:2217]
  Modules linked in:
  CPU: 5 PID: 2217 Comm: stat Not tainted 5.0.0-rc1-00005-g7f702faf5a9e #4
  Hardware name: Generic DT based system
  PC is at generic_fillattr+0x104/0x108
  LR is at 0xec497f00
  pc : [<802b8898>]    lr : [<ec497f00>]    psr: 200c0013
  sp : ec497e20  ip : ed608030  fp : ec497e3c
  r10: 00000000  r9 : ec497f00  r8 : ed608030
  r7 : ec497ebc  r6 : ec497f00  r5 : ee5c1550  r4 : ee005780
  r3 : 0000052d  r2 : 00000000  r1 : ec497f00  r0 : ed608030
  Flags: nzCv  IRQs on  FIQs on  Mode SVC_32  ISA ARM  Segment none
  Control: 10c5387d  Table: ac48006a  DAC: 00000051
  CPU: 5 PID: 2217 Comm: stat Not tainted 5.0.0-rc1-00005-g7f702faf5a9e #4
  Hardware name: Generic DT based system
  Backtrace:
  [<8010d974>] (dump_backtrace) from [<8010dc88>] (show_stack+0x20/0x24)
  [<8010dc68>] (show_stack) from [<80a1d194>] (dump_stack+0xb0/0xdc)
  [<80a1d0e4>] (dump_stack) from [<80109f34>] (show_regs+0x1c/0x20)
  [<80109f18>] (show_regs) from [<801d0a80>] (watchdog_timer_fn+0x280/0x2f8)
  [<801d0800>] (watchdog_timer_fn) from [<80198658>] (__hrtimer_run_queues+0x18c/0x380)
  [<801984cc>] (__hrtimer_run_queues) from [<80198e60>] (hrtimer_run_queues+0xb8/0xf0)
  [<80198da8>] (hrtimer_run_queues) from [<801973e8>] (run_local_timers+0x28/0x64)
  [<801973c0>] (run_local_timers) from [<80197460>] (update_process_times+0x3c/0x6c)
  [<80197424>] (update_process_times) from [<801ab2b8>] (tick_nohz_handler+0xe0/0x1bc)
  [<801ab1d8>] (tick_nohz_handler) from [<80843050>] (arch_timer_handler_virt+0x38/0x48)
  [<80843018>] (arch_timer_handler_virt) from [<80180a64>] (handle_percpu_devid_irq+0x8c/0x240)
  [<801809d8>] (handle_percpu_devid_irq) from [<8017ac20>] (generic_handle_irq+0x34/0x44)
  [<8017abec>] (generic_handle_irq) from [<8017b344>] (__handle_domain_irq+0x6c/0xc4)
  [<8017b2d8>] (__handle_domain_irq) from [<801022e0>] (gic_handle_irq+0x4c/0x88)
  [<80102294>] (gic_handle_irq) from [<80101a30>] (__irq_svc+0x70/0x98)
  [<802b8794>] (generic_fillattr) from [<8056b284>] (v9fs_vfs_getattr_dotl+0x74/0xa4)
  [<8056b210>] (v9fs_vfs_getattr_dotl) from [<802b8904>] (vfs_getattr_nosec+0x68/0x7c)
  [<802b889c>] (vfs_getattr_nosec) from [<802b895c>] (vfs_getattr+0x44/0x48)
  [<802b8918>] (vfs_getattr) from [<802b8a74>] (vfs_statx+0x9c/0xec)
  [<802b89d8>] (vfs_statx) from [<802b9428>] (sys_lstat64+0x48/0x78)
  [<802b93e0>] (sys_lstat64) from [<80101000>] (ret_fast_syscall+0x0/0x28)

[dominique.martinet@cea.fr: updated comment to not refer to a function
in another subsystem]
Link: http://lkml.kernel.org/r/20190124063514.8571-2-houtao1@huawei.com
Cc: stable@vger.kernel.org
Fixes: 7549ae3e ("9p: Use the i_size_[read, write]() macros instead of using inode->i_size directly.")
Reported-by: default avatarXing Gaopeng <xingaopeng@huawei.com>
Signed-off-by: default avatarHou Tao <houtao1@huawei.com>
Signed-off-by: default avatarDominique Martinet <dominique.martinet@cea.fr>
parent 3bbe8b1a
...@@ -40,6 +40,9 @@ ...@@ -40,6 +40,9 @@
*/ */
#define P9_LOCK_TIMEOUT (30*HZ) #define P9_LOCK_TIMEOUT (30*HZ)
/* flags for v9fs_stat2inode() & v9fs_stat2inode_dotl() */
#define V9FS_STAT2INODE_KEEP_ISIZE 1
extern struct file_system_type v9fs_fs_type; extern struct file_system_type v9fs_fs_type;
extern const struct address_space_operations v9fs_addr_operations; extern const struct address_space_operations v9fs_addr_operations;
extern const struct file_operations v9fs_file_operations; extern const struct file_operations v9fs_file_operations;
...@@ -61,8 +64,10 @@ int v9fs_init_inode(struct v9fs_session_info *v9ses, ...@@ -61,8 +64,10 @@ int v9fs_init_inode(struct v9fs_session_info *v9ses,
struct inode *inode, umode_t mode, dev_t); struct inode *inode, umode_t mode, dev_t);
void v9fs_evict_inode(struct inode *inode); void v9fs_evict_inode(struct inode *inode);
ino_t v9fs_qid2ino(struct p9_qid *qid); ino_t v9fs_qid2ino(struct p9_qid *qid);
void v9fs_stat2inode(struct p9_wstat *, struct inode *, struct super_block *); void v9fs_stat2inode(struct p9_wstat *stat, struct inode *inode,
void v9fs_stat2inode_dotl(struct p9_stat_dotl *, struct inode *); struct super_block *sb, unsigned int flags);
void v9fs_stat2inode_dotl(struct p9_stat_dotl *stat, struct inode *inode,
unsigned int flags);
int v9fs_dir_release(struct inode *inode, struct file *filp); int v9fs_dir_release(struct inode *inode, struct file *filp);
int v9fs_file_open(struct inode *inode, struct file *file); int v9fs_file_open(struct inode *inode, struct file *file);
void v9fs_inode2stat(struct inode *inode, struct p9_wstat *stat); void v9fs_inode2stat(struct inode *inode, struct p9_wstat *stat);
...@@ -83,4 +88,18 @@ static inline void v9fs_invalidate_inode_attr(struct inode *inode) ...@@ -83,4 +88,18 @@ static inline void v9fs_invalidate_inode_attr(struct inode *inode)
} }
int v9fs_open_to_dotl_flags(int flags); int v9fs_open_to_dotl_flags(int flags);
static inline void v9fs_i_size_write(struct inode *inode, loff_t i_size)
{
/*
* 32-bit need the lock, concurrent updates could break the
* sequences and make i_size_read() loop forever.
* 64-bit updates are atomic and can skip the locking.
*/
if (sizeof(i_size) > sizeof(long))
spin_lock(&inode->i_lock);
i_size_write(inode, i_size);
if (sizeof(i_size) > sizeof(long))
spin_unlock(&inode->i_lock);
}
#endif #endif
...@@ -446,7 +446,11 @@ v9fs_file_write_iter(struct kiocb *iocb, struct iov_iter *from) ...@@ -446,7 +446,11 @@ v9fs_file_write_iter(struct kiocb *iocb, struct iov_iter *from)
i_size = i_size_read(inode); i_size = i_size_read(inode);
if (iocb->ki_pos > i_size) { if (iocb->ki_pos > i_size) {
inode_add_bytes(inode, iocb->ki_pos - i_size); inode_add_bytes(inode, iocb->ki_pos - i_size);
i_size_write(inode, iocb->ki_pos); /*
* Need to serialize against i_size_write() in
* v9fs_stat2inode()
*/
v9fs_i_size_write(inode, iocb->ki_pos);
} }
return retval; return retval;
} }
......
...@@ -538,7 +538,7 @@ static struct inode *v9fs_qid_iget(struct super_block *sb, ...@@ -538,7 +538,7 @@ static struct inode *v9fs_qid_iget(struct super_block *sb,
if (retval) if (retval)
goto error; goto error;
v9fs_stat2inode(st, inode, sb); v9fs_stat2inode(st, inode, sb, 0);
v9fs_cache_inode_get_cookie(inode); v9fs_cache_inode_get_cookie(inode);
unlock_new_inode(inode); unlock_new_inode(inode);
return inode; return inode;
...@@ -1092,7 +1092,7 @@ v9fs_vfs_getattr(const struct path *path, struct kstat *stat, ...@@ -1092,7 +1092,7 @@ v9fs_vfs_getattr(const struct path *path, struct kstat *stat,
if (IS_ERR(st)) if (IS_ERR(st))
return PTR_ERR(st); return PTR_ERR(st);
v9fs_stat2inode(st, d_inode(dentry), dentry->d_sb); v9fs_stat2inode(st, d_inode(dentry), dentry->d_sb, 0);
generic_fillattr(d_inode(dentry), stat); generic_fillattr(d_inode(dentry), stat);
p9stat_free(st); p9stat_free(st);
...@@ -1170,12 +1170,13 @@ static int v9fs_vfs_setattr(struct dentry *dentry, struct iattr *iattr) ...@@ -1170,12 +1170,13 @@ static int v9fs_vfs_setattr(struct dentry *dentry, struct iattr *iattr)
* @stat: Plan 9 metadata (mistat) structure * @stat: Plan 9 metadata (mistat) structure
* @inode: inode to populate * @inode: inode to populate
* @sb: superblock of filesystem * @sb: superblock of filesystem
* @flags: control flags (e.g. V9FS_STAT2INODE_KEEP_ISIZE)
* *
*/ */
void void
v9fs_stat2inode(struct p9_wstat *stat, struct inode *inode, v9fs_stat2inode(struct p9_wstat *stat, struct inode *inode,
struct super_block *sb) struct super_block *sb, unsigned int flags)
{ {
umode_t mode; umode_t mode;
char ext[32]; char ext[32];
...@@ -1216,10 +1217,11 @@ v9fs_stat2inode(struct p9_wstat *stat, struct inode *inode, ...@@ -1216,10 +1217,11 @@ v9fs_stat2inode(struct p9_wstat *stat, struct inode *inode,
mode = p9mode2perm(v9ses, stat); mode = p9mode2perm(v9ses, stat);
mode |= inode->i_mode & ~S_IALLUGO; mode |= inode->i_mode & ~S_IALLUGO;
inode->i_mode = mode; inode->i_mode = mode;
i_size_write(inode, stat->length);
if (!(flags & V9FS_STAT2INODE_KEEP_ISIZE))
v9fs_i_size_write(inode, stat->length);
/* not real number of blocks, but 512 byte ones ... */ /* not real number of blocks, but 512 byte ones ... */
inode->i_blocks = (i_size_read(inode) + 512 - 1) >> 9; inode->i_blocks = (stat->length + 512 - 1) >> 9;
v9inode->cache_validity &= ~V9FS_INO_INVALID_ATTR; v9inode->cache_validity &= ~V9FS_INO_INVALID_ATTR;
} }
...@@ -1416,9 +1418,9 @@ int v9fs_refresh_inode(struct p9_fid *fid, struct inode *inode) ...@@ -1416,9 +1418,9 @@ int v9fs_refresh_inode(struct p9_fid *fid, struct inode *inode)
{ {
int umode; int umode;
dev_t rdev; dev_t rdev;
loff_t i_size;
struct p9_wstat *st; struct p9_wstat *st;
struct v9fs_session_info *v9ses; struct v9fs_session_info *v9ses;
unsigned int flags;
v9ses = v9fs_inode2v9ses(inode); v9ses = v9fs_inode2v9ses(inode);
st = p9_client_stat(fid); st = p9_client_stat(fid);
...@@ -1431,16 +1433,13 @@ int v9fs_refresh_inode(struct p9_fid *fid, struct inode *inode) ...@@ -1431,16 +1433,13 @@ int v9fs_refresh_inode(struct p9_fid *fid, struct inode *inode)
if ((inode->i_mode & S_IFMT) != (umode & S_IFMT)) if ((inode->i_mode & S_IFMT) != (umode & S_IFMT))
goto out; goto out;
spin_lock(&inode->i_lock);
/* /*
* We don't want to refresh inode->i_size, * We don't want to refresh inode->i_size,
* because we may have cached data * because we may have cached data
*/ */
i_size = inode->i_size; flags = (v9ses->cache == CACHE_LOOSE || v9ses->cache == CACHE_FSCACHE) ?
v9fs_stat2inode(st, inode, inode->i_sb); V9FS_STAT2INODE_KEEP_ISIZE : 0;
if (v9ses->cache == CACHE_LOOSE || v9ses->cache == CACHE_FSCACHE) v9fs_stat2inode(st, inode, inode->i_sb, flags);
inode->i_size = i_size;
spin_unlock(&inode->i_lock);
out: out:
p9stat_free(st); p9stat_free(st);
kfree(st); kfree(st);
......
...@@ -143,7 +143,7 @@ static struct inode *v9fs_qid_iget_dotl(struct super_block *sb, ...@@ -143,7 +143,7 @@ static struct inode *v9fs_qid_iget_dotl(struct super_block *sb,
if (retval) if (retval)
goto error; goto error;
v9fs_stat2inode_dotl(st, inode); v9fs_stat2inode_dotl(st, inode, 0);
v9fs_cache_inode_get_cookie(inode); v9fs_cache_inode_get_cookie(inode);
retval = v9fs_get_acl(inode, fid); retval = v9fs_get_acl(inode, fid);
if (retval) if (retval)
...@@ -496,7 +496,7 @@ v9fs_vfs_getattr_dotl(const struct path *path, struct kstat *stat, ...@@ -496,7 +496,7 @@ v9fs_vfs_getattr_dotl(const struct path *path, struct kstat *stat,
if (IS_ERR(st)) if (IS_ERR(st))
return PTR_ERR(st); return PTR_ERR(st);
v9fs_stat2inode_dotl(st, d_inode(dentry)); v9fs_stat2inode_dotl(st, d_inode(dentry), 0);
generic_fillattr(d_inode(dentry), stat); generic_fillattr(d_inode(dentry), stat);
/* Change block size to what the server returned */ /* Change block size to what the server returned */
stat->blksize = st->st_blksize; stat->blksize = st->st_blksize;
...@@ -607,11 +607,13 @@ int v9fs_vfs_setattr_dotl(struct dentry *dentry, struct iattr *iattr) ...@@ -607,11 +607,13 @@ int v9fs_vfs_setattr_dotl(struct dentry *dentry, struct iattr *iattr)
* v9fs_stat2inode_dotl - populate an inode structure with stat info * v9fs_stat2inode_dotl - populate an inode structure with stat info
* @stat: stat structure * @stat: stat structure
* @inode: inode to populate * @inode: inode to populate
* @flags: ctrl flags (e.g. V9FS_STAT2INODE_KEEP_ISIZE)
* *
*/ */
void void
v9fs_stat2inode_dotl(struct p9_stat_dotl *stat, struct inode *inode) v9fs_stat2inode_dotl(struct p9_stat_dotl *stat, struct inode *inode,
unsigned int flags)
{ {
umode_t mode; umode_t mode;
struct v9fs_inode *v9inode = V9FS_I(inode); struct v9fs_inode *v9inode = V9FS_I(inode);
...@@ -631,7 +633,8 @@ v9fs_stat2inode_dotl(struct p9_stat_dotl *stat, struct inode *inode) ...@@ -631,7 +633,8 @@ v9fs_stat2inode_dotl(struct p9_stat_dotl *stat, struct inode *inode)
mode |= inode->i_mode & ~S_IALLUGO; mode |= inode->i_mode & ~S_IALLUGO;
inode->i_mode = mode; inode->i_mode = mode;
i_size_write(inode, stat->st_size); if (!(flags & V9FS_STAT2INODE_KEEP_ISIZE))
v9fs_i_size_write(inode, stat->st_size);
inode->i_blocks = stat->st_blocks; inode->i_blocks = stat->st_blocks;
} else { } else {
if (stat->st_result_mask & P9_STATS_ATIME) { if (stat->st_result_mask & P9_STATS_ATIME) {
...@@ -661,8 +664,9 @@ v9fs_stat2inode_dotl(struct p9_stat_dotl *stat, struct inode *inode) ...@@ -661,8 +664,9 @@ v9fs_stat2inode_dotl(struct p9_stat_dotl *stat, struct inode *inode)
} }
if (stat->st_result_mask & P9_STATS_RDEV) if (stat->st_result_mask & P9_STATS_RDEV)
inode->i_rdev = new_decode_dev(stat->st_rdev); inode->i_rdev = new_decode_dev(stat->st_rdev);
if (stat->st_result_mask & P9_STATS_SIZE) if (!(flags & V9FS_STAT2INODE_KEEP_ISIZE) &&
i_size_write(inode, stat->st_size); stat->st_result_mask & P9_STATS_SIZE)
v9fs_i_size_write(inode, stat->st_size);
if (stat->st_result_mask & P9_STATS_BLOCKS) if (stat->st_result_mask & P9_STATS_BLOCKS)
inode->i_blocks = stat->st_blocks; inode->i_blocks = stat->st_blocks;
} }
...@@ -928,9 +932,9 @@ v9fs_vfs_get_link_dotl(struct dentry *dentry, ...@@ -928,9 +932,9 @@ v9fs_vfs_get_link_dotl(struct dentry *dentry,
int v9fs_refresh_inode_dotl(struct p9_fid *fid, struct inode *inode) int v9fs_refresh_inode_dotl(struct p9_fid *fid, struct inode *inode)
{ {
loff_t i_size;
struct p9_stat_dotl *st; struct p9_stat_dotl *st;
struct v9fs_session_info *v9ses; struct v9fs_session_info *v9ses;
unsigned int flags;
v9ses = v9fs_inode2v9ses(inode); v9ses = v9fs_inode2v9ses(inode);
st = p9_client_getattr_dotl(fid, P9_STATS_ALL); st = p9_client_getattr_dotl(fid, P9_STATS_ALL);
...@@ -942,16 +946,13 @@ int v9fs_refresh_inode_dotl(struct p9_fid *fid, struct inode *inode) ...@@ -942,16 +946,13 @@ int v9fs_refresh_inode_dotl(struct p9_fid *fid, struct inode *inode)
if ((inode->i_mode & S_IFMT) != (st->st_mode & S_IFMT)) if ((inode->i_mode & S_IFMT) != (st->st_mode & S_IFMT))
goto out; goto out;
spin_lock(&inode->i_lock);
/* /*
* We don't want to refresh inode->i_size, * We don't want to refresh inode->i_size,
* because we may have cached data * because we may have cached data
*/ */
i_size = inode->i_size; flags = (v9ses->cache == CACHE_LOOSE || v9ses->cache == CACHE_FSCACHE) ?
v9fs_stat2inode_dotl(st, inode); V9FS_STAT2INODE_KEEP_ISIZE : 0;
if (v9ses->cache == CACHE_LOOSE || v9ses->cache == CACHE_FSCACHE) v9fs_stat2inode_dotl(st, inode, flags);
inode->i_size = i_size;
spin_unlock(&inode->i_lock);
out: out:
kfree(st); kfree(st);
return 0; return 0;
......
...@@ -172,7 +172,7 @@ static struct dentry *v9fs_mount(struct file_system_type *fs_type, int flags, ...@@ -172,7 +172,7 @@ static struct dentry *v9fs_mount(struct file_system_type *fs_type, int flags,
goto release_sb; goto release_sb;
} }
d_inode(root)->i_ino = v9fs_qid2ino(&st->qid); d_inode(root)->i_ino = v9fs_qid2ino(&st->qid);
v9fs_stat2inode_dotl(st, d_inode(root)); v9fs_stat2inode_dotl(st, d_inode(root), 0);
kfree(st); kfree(st);
} else { } else {
struct p9_wstat *st = NULL; struct p9_wstat *st = NULL;
...@@ -183,7 +183,7 @@ static struct dentry *v9fs_mount(struct file_system_type *fs_type, int flags, ...@@ -183,7 +183,7 @@ static struct dentry *v9fs_mount(struct file_system_type *fs_type, int flags,
} }
d_inode(root)->i_ino = v9fs_qid2ino(&st->qid); d_inode(root)->i_ino = v9fs_qid2ino(&st->qid);
v9fs_stat2inode(st, d_inode(root), sb); v9fs_stat2inode(st, d_inode(root), sb, 0);
p9stat_free(st); p9stat_free(st);
kfree(st); kfree(st);
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment