ceph_pagelist_encode_string() will not fail in reserved case,
also, we do not check err code here, so remove unnecessary
assignment.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: "Yan, Zheng" <[email protected]>
Signed-off-by: Ilya Dryomov <[email protected]>
We print session's refcount in debug message inside
ceph_put_mds_session() and get_session(), so we don't have to
print it in con_get()/__ceph_lookup_mds_session()/con_put().
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: "Yan, Zheng" <[email protected]>
Signed-off-by: Ilya Dryomov <[email protected]>
Actually, total amount of available minor number
for a single major is MINORMASK + 1. So expand
minor range when registering chrdev region.
Signed-off-by: Chengguang Xu <[email protected]>
Acked-by: Wu Hao <[email protected]>
Acked-by: Alan Tull <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
We have already checked flag IOCB_DIRECT in the sanity
check of flag IOCB_NOWAIT, so don't have to check it
again here.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Chao Yu <[email protected]>
Signed-off-by: Jaegeuk Kim <[email protected]>
When we fail from allocating inode/space, we back out
the change we already did. In a special case which has
exceeded soft limit by the change, we should also check
time limit and reset it properly.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Jan Kara <[email protected]>
We need to check return code only when calling ->read_dqblk(),
so fix it properly.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Jan Kara <[email protected]>
Just remove unused include <linux/badblocks.h> from
fs/block_dev.c.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Bart Van Assche <[email protected]>
Signed-off-by: Al Viro <[email protected]>
unlikely has already included in IS_ERR(),
so just remove redundant unlikely annotation.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Jonathan Cameron <[email protected]>
The function comment of __register_chrdev_region()
is out of date, so update it based on the code.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
register_chrdev_region() carefully checks minor range
before calling __register_chrdev_region() but there is
another path from alloc_chrdev_region() which does not
check the range properly. So add a check for given minor
range in __register_chrdev_region().
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
When sb->s_root is NULL dput() will do nothing,
so jump to label 'free_node_inode' instead of lable
'free_root_inode' when failing from d_make_root().
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Chao Yu <[email protected]>
Signed-off-by: Jaegeuk Kim <[email protected]>
unlikely has already included in IS_ERR(),
so just remove redundant unlikely annotation.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Mike Snitzer <[email protected]>
unlikely has already included in IS_ERR(),
so just remove redundant unlikely annotation.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Mike Snitzer <[email protected]>
unlikely has already included in IS_ERR(),
so just remove redundant unlikely annotation.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Mike Snitzer <[email protected]>
unlikely has already included in IS_ERR(), so just
remove redundant likely/unlikely annotation.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Horia Geantă <[email protected]>
Signed-off-by: Herbert Xu <[email protected]>
unlikely has already included in IS_ERR(),
so just remove redundant likely annotation.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Actually, total amount of available minor number for a single major is
MINORMASK + 1. So expand minor range when registering chrdev region.
Signed-off-by: Chengguang Xu <[email protected]>
[wsa: fixed typo in commit message]
Signed-off-by: Wolfram Sang <[email protected]>
Actually, total amount of available minor number
for a single major is MINORMARK + 1. So expand
minor range when registering chrdev region.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Alex Williamson <[email protected]>
Actually, total amount of available minor number
for a single major is MINORMARK + 1. So expand
minor range when registering chrdev region.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Alex Williamson <[email protected]>
Actually, total amount of available minor number
for a single major is MINORMARK + 1. So expand
minor range when registering chrdev region.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Alex Williamson <[email protected]>
Actually, total amount of available minor number
for a single major is MINORMARK + 1. So expand
minor range when registering chrdev region.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Alex Williamson <[email protected]>
unlikely has already included in IS_ERR(),
so just remove it.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Gao Xiang <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
When calling register_blkdev() with specified major
device number, the return code is 0 on success.
So it seems not correct direct assign return code to
variable major_num in this case.
Tested-by: Michael Schmitz <[email protected]>
Reviewed-by: Geert Uytterhoeven <[email protected]>
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Jens Axboe <[email protected]>
Some error cases like failing from d_make_root() will
cause double iput because d_make_root() also does iput
in its error path.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Chao Yu <[email protected]>
Reviewed-by: Gao Xiang <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
It is not necessary to check idev->info several times under
mutex lock, so just remove redundant check.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Should jump to lable err_infoopen when idev->info is NULL
in uio_open().
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Set proper return code when failing from allocating
memory in ext2_fill_super().
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Jan Kara <[email protected]>
If new mode is the same as old mode we don't have to reset
inode mode in the rest of the code, so compare old and new
mode before setting update_mode flag.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Dave Kleikamp <[email protected]>
There is redundant assighment of variable i in
ceph_mdsmap_get_random_mds(), just remvoe it.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Ilya Dryomov <[email protected]>
mempool_destroy() can handle NULL pointer correctly,
so there is no need to check NULL pointer before calling
mempool_destroy().
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Shaohua Li <[email protected]>
blk_mq_init_queue() will not return NULL pointer to its caller,
so it's better to replace IS_ERR_OR_NULL using IS_ERR in loop_add().
If in the future things change to check NULL pointer inside loop_add(),
we should return -ENOMEM as return code instead of PTR_ERR(NULL).
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Jens Axboe <[email protected]>
Add __exit annotation to cleanup helper which
is only called once in the module.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Jens Axboe <[email protected]>
Add __exit annotation to cleanup helper which is only
called once in the module.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Sagi Grimberg <[email protected]>
Signed-off-by: Christoph Hellwig <[email protected]>
If new mode is the same as old mode we don't have to reset
inode mode in the rest of the code, so compare old and new
mode before setting update_mode flag.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Theodore Ts'o <[email protected]>
ext4_xattr_destroy_cache() can handle NULL pointer correctly,
so there is no need to check NULL pointer before calling
ext4_xattr_destroy_cache().
Reviewed-by: Jan Kara <[email protected]>
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Theodore Ts'o <[email protected]>
Reusable parameter of mb_cache_entry_create() is bool type,
so it's better to set true instead of 1.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Jan Kara <[email protected]>
ext2_xattr_destroy_cache() can handle NULL pointer correctly,
so there is no need to check NULL pointer before calling
ext2_xattr_destroy_cache().
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Jan Kara <[email protected]>
If filesystem has already mounted as read-only, then we don't have
to do it again.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Jan Kara <[email protected]>
dma_pool_destroy() can handle NULL pointer correctly, so there is
no need to check NULL pointer before calling dma_pool_destroy().
Signed-off-by: Chengguang Xu <[email protected]>
Acked-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
mempool_destroy() can handle NULL pointer correctly, so there is no
need to check NULL pointer before calling mempool_destroy().
Signed-off-by: Chengguang Xu <[email protected]>
Acked-by: Manish Rangankar <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
dma_pool_destroy() can handle NULL pointer correctly, so there is no
need to check NULL pointer before calling dma_pool_destroy().
Signed-off-by: Chengguang Xu <[email protected]>
Acked-by: Manish Rangankar <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
There is no functional change but it seems better to get size by calling
posix_acl_xattr_size() instead of calling posix_acl_to_xattr() with
NULL buffer argument. Additionally, remove unnecessary assignments.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Miklos Szeredi <[email protected]>
Do the snap check first in ceph_set_acl(), so we can avoid
unnecessary operations when the inode has snap.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: "Yan, Zheng" <[email protected]>
Signed-off-by: Ilya Dryomov <[email protected]>
Add __init/__exit annotation to init/cleanup helpers
which are only called once in the module.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Ilya Dryomov <[email protected]>
Signed-off-by: Ilya Dryomov <[email protected]>
__cap_delay_requeue() only requeue inode which does not
have CEPH_I_FLUSH flag, so avoid reset cap hold timeout
for that inode.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: "Yan, Zheng" <[email protected]>
Signed-off-by: Ilya Dryomov <[email protected]>
There are some cases can cause memory leak when parsing
option 'osdname'.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Al Viro <[email protected]>
default_acl and acl of newly created inode will be initiated
as ACL_NOT_CACHED in vfs function inode_init_always() and later
will be updated by calling xxx_init_acl() in specific filesystems.
Howerver, when default_acl and acl are NULL then they keep the value
of ACL_NOT_CACHED, this patch tries to cache NULL for acl/default_acl
in this case.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Mike Marshall <[email protected]>
default_acl and acl of newly created inode will be initiated as
ACL_NOT_CACHED in vfs function inode_init_always() and later will be
updated by calling xxx_init_acl() in specific filesystems. However,
when default_acl and acl are NULL then they keep the value of
ACL_NOT_CACHED. This patch changes the code to cache NULL for acl /
default_acl in this case to save unnecessary ACL lookup attempt.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Theodore Ts'o <[email protected]>
Reviewed-by: Jan Kara <[email protected]>
Currently we show mount option "io_bits=%u" as "io_size=%uKB",
it will cause option parsing problem(unrecognized mount option)
in remount.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Chao Yu <[email protected]>
Signed-off-by: Jaegeuk Kim <[email protected]>
If macro CONFIG_QUOTA is not enabled then mount option flag
of usrquota/grpquota will not be set, so we can remove some
building macro check safely in ext2_shwo_options().
Additionally, I think it's better to define EXT2_MOUNT_DAX
regardless macro CONFIG_FS_DAX is enabled just like acl/xattr.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Jan Kara <[email protected]>
kmem_cache_destroy() can handle NULL pointer correctly, so there is
no need to check NULL pointer before calling kmem_cache_destroy().
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Jan Höppner <[email protected]>
Signed-off-by: Martin Schwidefsky <[email protected]>
Add new helper erofs_get_fault_rate() to get fault rate instead of
directly getting it from sbi, so we can remove the macro check
surrounding it.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Chao Yu <[email protected]>
Reviewed-by: Gao Xiang <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Define a dummpy function of erofs_build_fault_attr() when macro
CONFIG_EROFS_FAULT_INJECTION is disabled, so that we don't have to
check the macro in calling place. Based on above adjustment,
do proper code cleanup for option parsing of fault_injection.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Chao Yu <[email protected]>
Reviewed-by: Gao Xiang <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Define a dummy function of time_to_inject()/erofs_show_injection_info(),
so that we don't have to check macro CONFIG_EROFS_FAULT_INJECTION in
calling place.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Chao Yu <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
It's a little bit strange when fault_injection related
option fail with -EINVAL which was already disabled
from config, so surround all fault_injection related option
parsing code using CONFIG_EROFS_FAULT_INJECTION. Meanwhile,
slightly change warning message to keep consistency with
option POSIX_ACL and FS_XATTR.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Chao Yu <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
It's a little bit strange when fault_injection related
options fail with -EINVAL which were already disabled
from config, so surround all fault_injection related option
parsing code using CONFIG_F2FS_FAULT_INJECTION. Meanwhile,
slightly change warning message to keep consistency with
option POSIX_ACL and FS_XATTR.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Chao Yu <[email protected]>
Signed-off-by: Jaegeuk Kim <[email protected]>
default_acl and acl of newly created inode will be initiated
as ACL_NOT_CACHED in vfs function inode_init_always() and later
will be updated by calling xxx_init_acl() in specific filesystems.
Howerver, when default_acl and acl are NULL then they keep the value
of ACL_NOT_CACHED, this patch tries to cache NULL for acl/default_acl
in this case.
Signed-off-by: Chengguang Xu <[email protected]>
Acked-by: Chao Yu <[email protected]>
Signed-off-by: Jaegeuk Kim <[email protected]>
We treat quota option as usrquota, so remove quota option from ignore
list.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Dave Kleikamp <[email protected]>
default_acl and acl of newly created inode will be initiated
as ACL_NOT_CACHED in vfs function inode_init_always() and later
will be updated by calling xxx_init_acl() in specific filesystems.
Howerver, when default_acl and acl are NULL then they keep the value
of ACL_NOT_CACHED, this patch tries to cache NULL for acl/default_acl
in this case.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Dave Kleikamp <[email protected]>
Add additinal sanity check for irregular case(e.g. corruption).
If size of extended attribution is smaller than size of acl header,
then return -EINVAL.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Chao Yu <[email protected]>
Signed-off-by: Jaegeuk Kim <[email protected]>
default_acl and acl of newly created inode will be initiated as
ACL_NOT_CACHED in vfs function inode_init_always() and later will be
updated by calling xxx_init_acl() in specific filesystems. However,
when default_acl and acl are NULL, then they keep the value of
ACL_NOT_CACHED. This patch changes the code to cache NULL for acl /
default_acl in this case to save unnecessary ACL lookup in future.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Jan Kara <[email protected]>
kmem_cache_destroy() can handle NULL pointer correctly, so there is no
need to check NULL pointer before calling kmem_cache_destroy()
Signed-off-by: Chengguang Xu <[email protected]>
Acked-by: Johannes Thumshirn <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
kmem_cache_destroy() can handle NULL pointer correctly, so there is
no need to check e->icq_cache before calling kmem_cache_destroy().
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Jens Axboe <[email protected]>
Because blk_do_io_stat() only does a judgement about the request
contributes to IO statistics, it better changes return type to bool.
Signed-off-by: Chengguang Xu <[email protected]>
Signed-off-by: Jens Axboe <[email protected]>
The code of ceph_unreserve_caps() and error handling in
ceph_reserve_caps() are duplicated, so introduce a helper
__ceph_unreserve_caps() to reduce duplicated code.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: "Yan, Zheng" <[email protected]>
Signed-off-by: Ilya Dryomov <[email protected]>
We do not check return code for __do_request() in all callers,
so change to void return type.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: "Yan, Zheng" <[email protected]>
Signed-off-by: Ilya Dryomov <[email protected]>
ceph_setattr() finally calls vfs function inode_newsize_ok()
to do offset validation and that is based on sb->s_maxbytes.
Because we set sb->s_maxbytes to MAX_LFS_FILESIZE to through
VFS check and do proper offset validation in cephfs level,
we need adding proper offset validation before calling
inode_newsize_ok().
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: "Yan, Zheng" <[email protected]>
Signed-off-by: Ilya Dryomov <[email protected]>
It's only coding style fix not functinal change. When if/else has only
one statement then the braces are not needed.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Nikolay Borisov <[email protected]>
Reviewed-by: David Sterba <[email protected]>
Signed-off-by: David Sterba <[email protected]>
It's not good to override the error code when failing from
btrfs_getxattr() in btrfs_get_acl() because it hides the real reason of
the failure.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Nikolay Borisov <[email protected]>
Reviewed-by: David Sterba <[email protected]>
Signed-off-by: David Sterba <[email protected]>
There is no chance to get into -ERANGE error condition because we first
call btrfs_getxattr to get the length of the attribute, then we do a
subsequent call with the size from the first call. Between the 2 calls
the size shouldn't change. So remove the unnecessary -ERANGE error
check.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Nikolay Borisov <[email protected]>
Reviewed-by: David Sterba <[email protected]>
Signed-off-by: David Sterba <[email protected]>
In btrfs_get_acl() the first call of btr_getxattr() is for getting the
length of attribute, the value buffer is never used in this case. So
it's better to replace empty string with NULL.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Nikolay Borisov <[email protected]>
Reviewed-by: David Sterba <[email protected]>
Signed-off-by: David Sterba <[email protected]>
The caller of btrfs_get_acl() checks error condition so there is no
impact from this change. In practice there is no chance to get into
default case of switch statement because VFS has already checked the
type.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Nikolay Borisov <[email protected]>
Reviewed-by: David Sterba <[email protected]>
Signed-off-by: David Sterba <[email protected]>
If the offset is larger or equal to both real file size and
max file size, then return -EFBIG.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: "Yan, Zheng" <[email protected]>
Signed-off-by: Ilya Dryomov <[email protected]>
If the range is larger than both real file size and limit of
max file size, then return -EFBIG.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: "Yan, Zheng" <[email protected]>
Signed-off-by: Ilya Dryomov <[email protected]>
In order to not bother to VFS and other specific filesystems,
we decided to do offset validation inside ceph kernel client,
so just simply set sb->s_maxbytes to MAX_LFS_FILESIZE so that
it can successfully pass VFS check. We add new field max_file_size
in ceph_fs_client to store real file size limit and doing proper
check based on it.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: "Yan, Zheng" <[email protected]>
Signed-off-by: Ilya Dryomov <[email protected]>
When file num exceeds quota limit, should call d_drop to drop
dentry from cache as well.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: "Yan, Zheng" <[email protected]>
Signed-off-by: Ilya Dryomov <[email protected]>
When file num exceeds quota limit or fails from ceph_per_init_acls()
should call d_drop to drop dentry from cache as well.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: "Yan, Zheng" <[email protected]>
Signed-off-by: Ilya Dryomov <[email protected]>
In order to return correct error code should replace variable ret
using err in error case.
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: "Yan, Zheng" <[email protected]>
Signed-off-by: Ilya Dryomov <[email protected]>
There is no reason to continue option parsing after detecting
bad option.
[ Return match_int() errors from ceph_parse_options() to match the
behaviour of parse_rbd_opts_token() and parse_fsopt_token(). ]
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: Ilya Dryomov <[email protected]>
Signed-off-by: Ilya Dryomov <[email protected]>
It's better to restore ctime as well in the case of restoring old mode
in ceph_set_acl().
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: "Yan, Zheng" <[email protected]>
Signed-off-by: Ilya Dryomov <[email protected]>
When the size of acl extended attribution is larger than pre-allocated
value buffer size, we will hit error '-ERANGE' and it's probabaly caused
by concurrent get/set acl from different clients. In this case, current
logic just sets acl to NULL so that we cannot get proper information but
the operation looks successful.
This patch adds retry logic for error -ERANGE and return -EIO if fail
from the retry. Additionally, print real errno when failing from
__ceph_getxattr().
Signed-off-by: Chengguang Xu <[email protected]>
Reviewed-by: "Yan, Zheng" <[email protected]>
Signed-off-by: Ilya Dryomov <[email protected]>
Chengguang Xu