* [RFC PATCH v2 0/6] Btrfs: stop abusing current->journal_info for direct I/O @ 2018-08-28 0:03 Omar Sandoval 2018-08-28 0:03 ` [RFC PATCH v2 1/6] fs: pass iocb to direct I/O get_block() Omar Sandoval ` (5 more replies) 0 siblings, 6 replies; 7+ messages in thread From: Omar Sandoval @ 2018-08-28 0:03 UTC (permalink / raw) To: linux-fsdevel, linux-btrfs, Al Viro; +Cc: David Sterba, kernel-team From: Omar Sandoval <osandov@fb.com> Hi, This is a different approach from v1 [1] of this series to stop abusing current->journal_info in Btrfs. This approach unifies everything to use iocb->private instead of map_bh->b_private. Patches 1 and 5 pass the iocb to a couple of callbacks which need it. Patches 2 and 3 migrates the users of b_private to use iocb->private, and patch 4 gets rid of the b_private handling in the direct I/O code. Patch 6 cleans up Btrfs. I'm not convinced that this is cleaner that my first approach, but it at least avoids growing the argument list to do_blockdev_direct_IO(), which was Al's complaint of v1. Thanks! 1: https://www.spinics.net/lists/linux-btrfs/msg77859.html Omar Sandoval (6): fs: pass iocb to direct I/O get_block() ext4: use iocb->private instead of bh->b_private ocfs2: use iocb->private instead of bh->b_private fs: stop propagating bh->b_private for direct I/O fs: pass iocb to direct I/O submit_io() Btrfs: stop abusing current->journal_info in btrfs_direct_IO() fs/affs/file.c | 9 ++++++- fs/btrfs/inode.c | 36 +++++++------------------ fs/direct-io.c | 23 +++++++--------- fs/ext2/inode.c | 9 ++++++- fs/ext4/ext4.h | 2 -- fs/ext4/inode.c | 40 ++++++++++++++++------------ fs/f2fs/data.c | 5 ++-- fs/fat/inode.c | 9 ++++++- fs/gfs2/aops.c | 5 ++-- fs/hfs/inode.c | 9 ++++++- fs/hfsplus/inode.c | 9 ++++++- fs/jfs/inode.c | 9 ++++++- fs/nilfs2/inode.c | 9 ++++++- fs/ocfs2/aops.c | 39 ++++++++++++++------------- fs/ocfs2/aops.h | 64 ++++++++++++++++++++++++++++++--------------- fs/reiserfs/inode.c | 4 +-- fs/udf/inode.c | 9 ++++++- include/linux/fs.h | 17 ++++++------ 18 files changed, 187 insertions(+), 120 deletions(-) -- 2.18.0 ^ permalink raw reply [flat|nested] 7+ messages in thread
* [RFC PATCH v2 1/6] fs: pass iocb to direct I/O get_block() 2018-08-28 0:03 [RFC PATCH v2 0/6] Btrfs: stop abusing current->journal_info for direct I/O Omar Sandoval @ 2018-08-28 0:03 ` Omar Sandoval 2018-08-28 0:03 ` [RFC PATCH v2 2/6] ext4: use iocb->private instead of bh->b_private Omar Sandoval ` (4 subsequent siblings) 5 siblings, 0 replies; 7+ messages in thread From: Omar Sandoval @ 2018-08-28 0:03 UTC (permalink / raw) To: linux-fsdevel, linux-btrfs, Al Viro; +Cc: David Sterba, kernel-team From: Omar Sandoval <osandov@fb.com> Split out dio_get_block_t which is the same as get_block_t except that it takes the iocb as well, and update fs/direct-io.c and all callers to use it. This is preparation for replacing the use of bh->b_private in the direct I/O code with iocb->private. Signed-off-by: Omar Sandoval <osandov@fb.com> --- fs/affs/file.c | 9 ++++++++- fs/btrfs/inode.c | 3 ++- fs/direct-io.c | 13 ++++++------- fs/ext2/inode.c | 9 ++++++++- fs/ext4/ext4.h | 2 -- fs/ext4/inode.c | 27 ++++++++++++++++++--------- fs/f2fs/data.c | 5 +++-- fs/fat/inode.c | 9 ++++++++- fs/gfs2/aops.c | 5 +++-- fs/hfs/inode.c | 9 ++++++++- fs/hfsplus/inode.c | 9 ++++++++- fs/jfs/inode.c | 9 ++++++++- fs/nilfs2/inode.c | 9 ++++++++- fs/ocfs2/aops.c | 15 +++++++++------ fs/reiserfs/inode.c | 4 ++-- fs/udf/inode.c | 9 ++++++++- include/linux/fs.h | 10 ++++++---- 17 files changed, 113 insertions(+), 43 deletions(-) diff --git a/fs/affs/file.c b/fs/affs/file.c index a85817f54483..66a1a5601d65 100644 --- a/fs/affs/file.c +++ b/fs/affs/file.c @@ -389,6 +389,13 @@ static void affs_write_failed(struct address_space *mapping, loff_t to) } } +static int affs_get_block_dio(struct kiocb *iocb, struct inode *inode, + sector_t block, struct buffer_head *bh_result, + int create) +{ + return affs_get_block(inode, block, bh_result, create); +} + static ssize_t affs_direct_IO(struct kiocb *iocb, struct iov_iter *iter) { @@ -406,7 +413,7 @@ affs_direct_IO(struct kiocb *iocb, struct iov_iter *iter) return 0; } - ret = blockdev_direct_IO(iocb, inode, iter, affs_get_block); + ret = blockdev_direct_IO(iocb, inode, iter, affs_get_block_dio); if (ret < 0 && iov_iter_rw(iter) == WRITE) affs_write_failed(mapping, offset + count); return ret; diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index eba61bcb9bb3..b61ea6dd9956 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -7659,7 +7659,8 @@ static int btrfs_get_blocks_direct_write(struct extent_map **map, return ret; } -static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock, +static int btrfs_get_blocks_direct(struct kiocb *iocb, struct inode *inode, + sector_t iblock, struct buffer_head *bh_result, int create) { struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb); diff --git a/fs/direct-io.c b/fs/direct-io.c index 093fb54cd316..f631aa98849b 100644 --- a/fs/direct-io.c +++ b/fs/direct-io.c @@ -82,7 +82,7 @@ struct dio_submit { int reap_counter; /* rate limit reaping */ sector_t final_block_in_request;/* doesn't change */ int boundary; /* prev block is at a boundary */ - get_block_t *get_block; /* block mapping function */ + dio_get_block_t *get_block; /* block mapping function */ dio_submit_t *submit_io; /* IO submition function */ loff_t logical_offset_in_bio; /* current first logical block in bio */ @@ -713,8 +713,8 @@ static int get_more_blocks(struct dio *dio, struct dio_submit *sdio, create = 0; } - ret = (*sdio->get_block)(dio->inode, fs_startblk, - map_bh, create); + ret = (*sdio->get_block)(dio->iocb, dio->inode, fs_startblk, + map_bh, create); /* Store for completion */ dio->private = map_bh->b_private; @@ -1170,7 +1170,7 @@ static inline int drop_refcount(struct dio *dio) static inline ssize_t do_blockdev_direct_IO(struct kiocb *iocb, struct inode *inode, struct block_device *bdev, struct iov_iter *iter, - get_block_t get_block, dio_iodone_t end_io, + dio_get_block_t get_block, dio_iodone_t end_io, dio_submit_t submit_io, int flags) { unsigned i_blkbits = READ_ONCE(inode->i_blkbits); @@ -1398,9 +1398,8 @@ do_blockdev_direct_IO(struct kiocb *iocb, struct inode *inode, ssize_t __blockdev_direct_IO(struct kiocb *iocb, struct inode *inode, struct block_device *bdev, struct iov_iter *iter, - get_block_t get_block, - dio_iodone_t end_io, dio_submit_t submit_io, - int flags) + dio_get_block_t get_block, dio_iodone_t end_io, + dio_submit_t submit_io, int flags) { /* * The block device state is needed in the end to finally diff --git a/fs/ext2/inode.c b/fs/ext2/inode.c index 71635909df3b..f390e6392238 100644 --- a/fs/ext2/inode.c +++ b/fs/ext2/inode.c @@ -930,6 +930,13 @@ static sector_t ext2_bmap(struct address_space *mapping, sector_t block) return generic_block_bmap(mapping,block,ext2_get_block); } +static int ext2_get_block_dio(struct kiocb *iocb, struct inode *inode, + sector_t iblock, struct buffer_head *bh_result, + int create) +{ + return ext2_get_block(inode, iblock, bh_result, create); +} + static ssize_t ext2_direct_IO(struct kiocb *iocb, struct iov_iter *iter) { @@ -940,7 +947,7 @@ ext2_direct_IO(struct kiocb *iocb, struct iov_iter *iter) loff_t offset = iocb->ki_pos; ssize_t ret; - ret = blockdev_direct_IO(iocb, inode, iter, ext2_get_block); + ret = blockdev_direct_IO(iocb, inode, iter, ext2_get_block_dio); if (ret < 0 && iov_iter_rw(iter) == WRITE) ext2_write_failed(mapping, offset + count); return ret; diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 7c7123f265c2..2c46c2e69b42 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -2425,8 +2425,6 @@ int ext4_get_block_unwritten(struct inode *inode, sector_t iblock, struct buffer_head *bh_result, int create); int ext4_get_block(struct inode *inode, sector_t iblock, struct buffer_head *bh_result, int create); -int ext4_dio_get_block(struct inode *inode, sector_t iblock, - struct buffer_head *bh_result, int create); int ext4_da_get_block_prep(struct inode *inode, sector_t iblock, struct buffer_head *bh, int create); int ext4_walk_page_buffers(handle_t *handle, diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 4efe77286ecd..18ad91b1c8f6 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -850,8 +850,9 @@ static int ext4_get_block_trans(struct inode *inode, sector_t iblock, } /* Get block function for DIO reads and writes to inodes without extents */ -int ext4_dio_get_block(struct inode *inode, sector_t iblock, - struct buffer_head *bh, int create) +static int ext4_dio_get_block(struct kiocb *iocb, struct inode *inode, + sector_t iblock, struct buffer_head *bh, + int create) { /* We don't expect handle for direct IO */ WARN_ON_ONCE(ext4_journal_current_handle()); @@ -866,8 +867,11 @@ int ext4_dio_get_block(struct inode *inode, sector_t iblock, * blocks are not allocated yet. The extent will be converted to written * after IO is complete. */ -static int ext4_dio_get_block_unwritten_async(struct inode *inode, - sector_t iblock, struct buffer_head *bh_result, int create) +static int ext4_dio_get_block_unwritten_async(struct kiocb *iocb, + struct inode *inode, + sector_t iblock, + struct buffer_head *bh_result, + int create) { int ret; @@ -905,8 +909,11 @@ static int ext4_dio_get_block_unwritten_async(struct inode *inode, * blocks are not allocated yet. The extent will be converted to written * after IO is complete by ext4_direct_IO_write(). */ -static int ext4_dio_get_block_unwritten_sync(struct inode *inode, - sector_t iblock, struct buffer_head *bh_result, int create) +static int ext4_dio_get_block_unwritten_sync(struct kiocb *iocb, + struct inode *inode, + sector_t iblock, + struct buffer_head *bh_result, + int create) { int ret; @@ -927,8 +934,10 @@ static int ext4_dio_get_block_unwritten_sync(struct inode *inode, return ret; } -static int ext4_dio_get_block_overwrite(struct inode *inode, sector_t iblock, - struct buffer_head *bh_result, int create) +static int ext4_dio_get_block_overwrite(struct kiocb *iocb, struct inode *inode, + sector_t iblock, + struct buffer_head *bh_result, + int create) { int ret; @@ -3663,7 +3672,7 @@ static ssize_t ext4_direct_IO_write(struct kiocb *iocb, struct iov_iter *iter) loff_t offset = iocb->ki_pos; size_t count = iov_iter_count(iter); int overwrite = 0; - get_block_t *get_block_func = NULL; + dio_get_block_t *get_block_func = NULL; int dio_flags = 0; loff_t final_size = offset + count; int orphan = 0; diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index 8f931d699287..a465d7877886 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -1232,8 +1232,9 @@ static int get_data_block(struct inode *inode, sector_t iblock, NO_CHECK_TYPE); } -static int get_data_block_dio(struct inode *inode, sector_t iblock, - struct buffer_head *bh_result, int create) +static int get_data_block_dio(struct kiocb *iocb, struct inode *inode, + sector_t iblock, struct buffer_head *bh_result, + int create) { return __get_data_block(inode, iblock, bh_result, create, F2FS_GET_BLOCK_DEFAULT, NULL, diff --git a/fs/fat/inode.c b/fs/fat/inode.c index bfd589ea74c0..b2caca4ed6bc 100644 --- a/fs/fat/inode.c +++ b/fs/fat/inode.c @@ -251,6 +251,13 @@ static int fat_write_end(struct file *file, struct address_space *mapping, return err; } +static int fat_get_block_dio(struct kiocb *iocb, struct inode *inode, + sector_t iblock, struct buffer_head *bh_result, + int create) +{ + return fat_get_block(inode, iblock, bh_result, create); +} + static ssize_t fat_direct_IO(struct kiocb *iocb, struct iov_iter *iter) { struct file *file = iocb->ki_filp; @@ -279,7 +286,7 @@ static ssize_t fat_direct_IO(struct kiocb *iocb, struct iov_iter *iter) * FAT need to use the DIO_LOCKING for avoiding the race * condition of fat_get_block() and ->truncate(). */ - ret = blockdev_direct_IO(iocb, inode, iter, fat_get_block); + ret = blockdev_direct_IO(iocb, inode, iter, fat_get_block_dio); if (ret < 0 && iov_iter_rw(iter) == WRITE) fat_write_failed(mapping, offset + count); diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c index 35f5ee23566d..cb908153559b 100644 --- a/fs/gfs2/aops.c +++ b/fs/gfs2/aops.c @@ -82,8 +82,9 @@ static int gfs2_get_block_noalloc(struct inode *inode, sector_t lblock, return 0; } -static int gfs2_get_block_direct(struct inode *inode, sector_t lblock, - struct buffer_head *bh_result, int create) +static int gfs2_get_block_direct(struct kiocb *iocb, struct inode *inode, + sector_t lblock, struct buffer_head *bh_result, + int create) { return gfs2_block_map(inode, lblock, bh_result, 0); } diff --git a/fs/hfs/inode.c b/fs/hfs/inode.c index 2a16111d312f..6986024c8ba0 100644 --- a/fs/hfs/inode.c +++ b/fs/hfs/inode.c @@ -126,6 +126,13 @@ static int hfs_releasepage(struct page *page, gfp_t mask) return res ? try_to_free_buffers(page) : 0; } +static int hfs_get_block_dio(struct kiocb *iocb, struct inode *inode, + sector_t block, struct buffer_head *bh_result, + int create) +{ + return hfs_get_block(inode, block, bh_result, create); +} + static ssize_t hfs_direct_IO(struct kiocb *iocb, struct iov_iter *iter) { struct file *file = iocb->ki_filp; @@ -134,7 +141,7 @@ static ssize_t hfs_direct_IO(struct kiocb *iocb, struct iov_iter *iter) size_t count = iov_iter_count(iter); ssize_t ret; - ret = blockdev_direct_IO(iocb, inode, iter, hfs_get_block); + ret = blockdev_direct_IO(iocb, inode, iter, hfs_get_block_dio); /* * In case of error extending write may have instantiated a few diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c index c824f702feec..0e1659bc934f 100644 --- a/fs/hfsplus/inode.c +++ b/fs/hfsplus/inode.c @@ -124,6 +124,13 @@ static int hfsplus_releasepage(struct page *page, gfp_t mask) return res ? try_to_free_buffers(page) : 0; } +static int hfsplus_get_block_dio(struct kiocb *iocb, struct inode *inode, + sector_t iblock, struct buffer_head *bh_result, + int create) +{ + return hfsplus_get_block(inode, iblock, bh_result, create); +} + static ssize_t hfsplus_direct_IO(struct kiocb *iocb, struct iov_iter *iter) { struct file *file = iocb->ki_filp; @@ -132,7 +139,7 @@ static ssize_t hfsplus_direct_IO(struct kiocb *iocb, struct iov_iter *iter) size_t count = iov_iter_count(iter); ssize_t ret; - ret = blockdev_direct_IO(iocb, inode, iter, hfsplus_get_block); + ret = blockdev_direct_IO(iocb, inode, iter, hfsplus_get_block_dio); /* * In case of error extending write may have instantiated a few diff --git a/fs/jfs/inode.c b/fs/jfs/inode.c index 054cc761b426..e255bd5b15dc 100644 --- a/fs/jfs/inode.c +++ b/fs/jfs/inode.c @@ -332,6 +332,13 @@ static sector_t jfs_bmap(struct address_space *mapping, sector_t block) return generic_block_bmap(mapping, block, jfs_get_block); } +static int jfs_get_block_dio(struct kiocb *iocb, struct inode *ip, + sector_t lblock, struct buffer_head *bh_result, + int create) +{ + return jfs_get_block(ip, lblock, bh_result, create); +} + static ssize_t jfs_direct_IO(struct kiocb *iocb, struct iov_iter *iter) { struct file *file = iocb->ki_filp; @@ -340,7 +347,7 @@ static ssize_t jfs_direct_IO(struct kiocb *iocb, struct iov_iter *iter) size_t count = iov_iter_count(iter); ssize_t ret; - ret = blockdev_direct_IO(iocb, inode, iter, jfs_get_block); + ret = blockdev_direct_IO(iocb, inode, iter, jfs_get_block_dio); /* * In case of error extending write may have instantiated a few diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c index 6a612d832e7d..10f1b063fb4d 100644 --- a/fs/nilfs2/inode.c +++ b/fs/nilfs2/inode.c @@ -300,6 +300,13 @@ static int nilfs_write_end(struct file *file, struct address_space *mapping, return err ? : copied; } +static int nilfs_get_block_dio(struct kiocb *iocb, struct inode *inode, + sector_t blkoff, struct buffer_head *bh_result, + int create) +{ + return nilfs_get_block(inode, blkoff, bh_result, create); +} + static ssize_t nilfs_direct_IO(struct kiocb *iocb, struct iov_iter *iter) { @@ -309,7 +316,7 @@ nilfs_direct_IO(struct kiocb *iocb, struct iov_iter *iter) return 0; /* Needs synchronization with the cleaner */ - return blockdev_direct_IO(iocb, inode, iter, nilfs_get_block); + return blockdev_direct_IO(iocb, inode, iter, nilfs_get_block_dio); } const struct address_space_operations nilfs_aops = { diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c index 302cd7caa4a7..93ca23c56b07 100644 --- a/fs/ocfs2/aops.c +++ b/fs/ocfs2/aops.c @@ -134,8 +134,9 @@ static int ocfs2_symlink_get_block(struct inode *inode, sector_t iblock, return err; } -static int ocfs2_lock_get_block(struct inode *inode, sector_t iblock, - struct buffer_head *bh_result, int create) +static int ocfs2_lock_get_block(struct kiocb *iocb, struct inode *inode, + sector_t iblock, struct buffer_head *bh_result, + int create) { int ret = 0; struct ocfs2_inode_info *oi = OCFS2_I(inode); @@ -2143,8 +2144,9 @@ static void ocfs2_dio_free_write_ctx(struct inode *inode, * called like this: dio->get_blocks(dio->inode, fs_startblk, * fs_count, map_bh, dio->rw == WRITE); */ -static int ocfs2_dio_wr_get_block(struct inode *inode, sector_t iblock, - struct buffer_head *bh_result, int create) +static int ocfs2_dio_wr_get_block(struct kiocb *iocb, struct inode *inode, + sector_t iblock, + struct buffer_head *bh_result, int create) { struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); struct ocfs2_inode_info *oi = OCFS2_I(inode); @@ -2171,7 +2173,8 @@ static int ocfs2_dio_wr_get_block(struct inode *inode, sector_t iblock, if (pos + total_len <= i_size_read(inode)) { /* This is the fast path for re-write. */ - ret = ocfs2_lock_get_block(inode, iblock, bh_result, create); + ret = ocfs2_lock_get_block(iocb, inode, iblock, bh_result, + create); if (buffer_mapped(bh_result) && !buffer_new(bh_result) && ret == 0) @@ -2427,7 +2430,7 @@ static ssize_t ocfs2_direct_IO(struct kiocb *iocb, struct iov_iter *iter) struct file *file = iocb->ki_filp; struct inode *inode = file->f_mapping->host; struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); - get_block_t *get_block; + dio_get_block_t *get_block; /* * Fallback to buffered I/O if we see an inode without diff --git a/fs/reiserfs/inode.c b/fs/reiserfs/inode.c index 132ec4406ed0..55ac918dbc73 100644 --- a/fs/reiserfs/inode.c +++ b/fs/reiserfs/inode.c @@ -511,8 +511,8 @@ static int reiserfs_get_block_create_0(struct inode *inode, sector_t block, * This is special helper for reiserfs_get_block in case we are executing * direct_IO request. */ -static int reiserfs_get_blocks_direct_io(struct inode *inode, - sector_t iblock, +static int reiserfs_get_blocks_direct_io(struct kiocb *iocb, + struct inode *inode, sector_t iblock, struct buffer_head *bh_result, int create) { diff --git a/fs/udf/inode.c b/fs/udf/inode.c index 9915a58fbabd..5ef1c72fa1ab 100644 --- a/fs/udf/inode.c +++ b/fs/udf/inode.c @@ -206,6 +206,13 @@ static int udf_write_begin(struct file *file, struct address_space *mapping, return ret; } +static int udf_get_block_dio(struct kiocb *iocb, struct inode *inode, + sector_t block, struct buffer_head *bh_result, + int create) +{ + return udf_get_block(inode, block, bh_result, create); +} + static ssize_t udf_direct_IO(struct kiocb *iocb, struct iov_iter *iter) { struct file *file = iocb->ki_filp; @@ -214,7 +221,7 @@ static ssize_t udf_direct_IO(struct kiocb *iocb, struct iov_iter *iter) size_t count = iov_iter_count(iter); ssize_t ret; - ret = blockdev_direct_IO(iocb, inode, iter, udf_get_block); + ret = blockdev_direct_IO(iocb, inode, iter, udf_get_block_dio); if (unlikely(ret < 0 && iov_iter_rw(iter) == WRITE)) udf_write_failed(mapping, iocb->ki_pos + count); return ret; diff --git a/include/linux/fs.h b/include/linux/fs.h index 805bf22898cf..85db69835023 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -80,6 +80,9 @@ typedef __kernel_rwf_t rwf_t; struct buffer_head; typedef int (get_block_t)(struct inode *inode, sector_t iblock, struct buffer_head *bh_result, int create); +typedef int (dio_get_block_t)(struct kiocb *iocb, struct inode *inode, + sector_t iblock, struct buffer_head *bh_result, + int create); typedef int (dio_iodone_t)(struct kiocb *iocb, loff_t offset, ssize_t bytes, void *private); @@ -3017,14 +3020,13 @@ void dio_warn_stale_pagecache(struct file *filp); ssize_t __blockdev_direct_IO(struct kiocb *iocb, struct inode *inode, struct block_device *bdev, struct iov_iter *iter, - get_block_t get_block, - dio_iodone_t end_io, dio_submit_t submit_io, - int flags); + dio_get_block_t get_block, dio_iodone_t end_io, + dio_submit_t submit_io, int flags); static inline ssize_t blockdev_direct_IO(struct kiocb *iocb, struct inode *inode, struct iov_iter *iter, - get_block_t get_block) + dio_get_block_t get_block) { return __blockdev_direct_IO(iocb, inode, inode->i_sb->s_bdev, iter, get_block, NULL, NULL, DIO_LOCKING | DIO_SKIP_HOLES); -- 2.18.0 ^ permalink raw reply related [flat|nested] 7+ messages in thread
* [RFC PATCH v2 2/6] ext4: use iocb->private instead of bh->b_private 2018-08-28 0:03 [RFC PATCH v2 0/6] Btrfs: stop abusing current->journal_info for direct I/O Omar Sandoval 2018-08-28 0:03 ` [RFC PATCH v2 1/6] fs: pass iocb to direct I/O get_block() Omar Sandoval @ 2018-08-28 0:03 ` Omar Sandoval 2018-08-28 0:03 ` [RFC PATCH v2 3/6] ocfs2: " Omar Sandoval ` (3 subsequent siblings) 5 siblings, 0 replies; 7+ messages in thread From: Omar Sandoval @ 2018-08-28 0:03 UTC (permalink / raw) To: linux-fsdevel, linux-btrfs, Al Viro Cc: David Sterba, kernel-team, Theodore Ts'o, Andreas Dilger From: Omar Sandoval <osandov@fb.com> As part of simplifying all of the private data passed around for direct I/O, bh->b_private will no longer be passed to dio_iodone_t. iocb is still available there, however, so convert ext4 to use it. Note that ext4_file_write_iter() also uses iocb->private, but ext4_direct_IO_write() resets it to NULL after reading it. Also note that the comment above ext4_should_dioread_nolock() is no longer accurate. It seems that it should be possible to remove the data journaling restriction now? Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Andreas Dilger <adilger.kernel@dilger.ca> Signed-off-by: Omar Sandoval <osandov@fb.com> --- fs/ext4/inode.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 18ad91b1c8f6..841d79919cef 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -884,18 +884,16 @@ static int ext4_dio_get_block_unwritten_async(struct kiocb *iocb, /* * When doing DIO using unwritten extents, we need io_end to convert * unwritten extents to written on IO completion. We allocate io_end - * once we spot unwritten extent and store it in b_private. Generic - * DIO code keeps b_private set and furthermore passes the value to - * our completion callback in 'private' argument. + * once we spot unwritten extent and store it in iocb->private. */ if (!ret && buffer_unwritten(bh_result)) { - if (!bh_result->b_private) { + if (!iocb->private) { ext4_io_end_t *io_end; io_end = ext4_init_io_end(inode, GFP_KERNEL); if (!io_end) return -ENOMEM; - bh_result->b_private = io_end; + iocb->private = io_end; ext4_set_io_unwritten_flag(inode, io_end); } set_buffer_defer_completion(bh_result); @@ -3617,7 +3615,7 @@ const struct iomap_ops ext4_iomap_ops = { static int ext4_end_io_dio(struct kiocb *iocb, loff_t offset, ssize_t size, void *private) { - ext4_io_end_t *io_end = private; + ext4_io_end_t *io_end = iocb->private; /* if not async direct IO just return */ if (!io_end) -- 2.18.0 ^ permalink raw reply related [flat|nested] 7+ messages in thread
* [RFC PATCH v2 3/6] ocfs2: use iocb->private instead of bh->b_private 2018-08-28 0:03 [RFC PATCH v2 0/6] Btrfs: stop abusing current->journal_info for direct I/O Omar Sandoval 2018-08-28 0:03 ` [RFC PATCH v2 1/6] fs: pass iocb to direct I/O get_block() Omar Sandoval 2018-08-28 0:03 ` [RFC PATCH v2 2/6] ext4: use iocb->private instead of bh->b_private Omar Sandoval @ 2018-08-28 0:03 ` Omar Sandoval 2018-08-28 0:03 ` [RFC PATCH v2 4/6] fs: stop propagating bh->b_private for direct I/O Omar Sandoval ` (2 subsequent siblings) 5 siblings, 0 replies; 7+ messages in thread From: Omar Sandoval @ 2018-08-28 0:03 UTC (permalink / raw) To: linux-fsdevel, linux-btrfs, Al Viro Cc: David Sterba, kernel-team, Mark Fasheh, Joel Becker From: Omar Sandoval <osandov@fb.com> As part of simplifying all of the private data passed around for direct I/O, bh->b_private will no longer be passed to dio_iodone_t. Instead, filesystems should use iocb->private. ocfs2 already uses iocb->private for storing a couple of flag bits, but we can use it as a tagged pointer and hide all of the messiness in helpers. Cc: Mark Fasheh <mark@fasheh.com> Cc: Joel Becker <jlbec@evilplan.org> Signed-off-by: Omar Sandoval <osandov@fb.com> --- fs/ocfs2/aops.c | 19 ++++++++------- fs/ocfs2/aops.h | 64 +++++++++++++++++++++++++++++++++---------------- 2 files changed, 54 insertions(+), 29 deletions(-) diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c index 93ca23c56b07..fc4a18b6ad3c 100644 --- a/fs/ocfs2/aops.c +++ b/fs/ocfs2/aops.c @@ -2104,12 +2104,13 @@ struct ocfs2_dio_write_ctxt { }; static struct ocfs2_dio_write_ctxt * -ocfs2_dio_alloc_write_ctx(struct buffer_head *bh, int *alloc) +ocfs2_dio_alloc_write_ctx(struct kiocb *iocb, int *alloc) { - struct ocfs2_dio_write_ctxt *dwc = NULL; + struct ocfs2_dio_write_ctxt *dwc; - if (bh->b_private) - return bh->b_private; + dwc = ocfs2_iocb_private(iocb); + if (dwc) + return dwc; dwc = kmalloc(sizeof(struct ocfs2_dio_write_ctxt), GFP_NOFS); if (dwc == NULL) @@ -2118,7 +2119,7 @@ ocfs2_dio_alloc_write_ctx(struct buffer_head *bh, int *alloc) dwc->dw_zero_count = 0; dwc->dw_orphaned = 0; dwc->dw_writer_pid = task_pid_nr(current); - bh->b_private = dwc; + ocfs2_iocb_set_private(iocb, dwc); *alloc = 1; return dwc; @@ -2184,7 +2185,7 @@ static int ocfs2_dio_wr_get_block(struct kiocb *iocb, struct inode *inode, bh_result->b_state = 0; } - dwc = ocfs2_dio_alloc_write_ctx(bh_result, &first_get_block); + dwc = ocfs2_dio_alloc_write_ctx(iocb, &first_get_block); if (unlikely(dwc == NULL)) { ret = -ENOMEM; mlog_errno(ret); @@ -2408,6 +2409,7 @@ static int ocfs2_dio_end_io(struct kiocb *iocb, ssize_t bytes, void *private) { + struct ocfs2_dio_write_ctxt *dwc; struct inode *inode = file_inode(iocb->ki_filp); int level; int ret = 0; @@ -2415,8 +2417,9 @@ static int ocfs2_dio_end_io(struct kiocb *iocb, /* this io's submitter should not have unlocked this before we could */ BUG_ON(!ocfs2_iocb_is_rw_locked(iocb)); - if (bytes > 0 && private) - ret = ocfs2_dio_end_io_write(inode, private, offset, bytes); + dwc = ocfs2_iocb_private(iocb); + if (bytes > 0 && dwc) + ret = ocfs2_dio_end_io_write(inode, dwc, offset, bytes); ocfs2_iocb_clear_rw_locked(iocb); diff --git a/fs/ocfs2/aops.h b/fs/ocfs2/aops.h index 3494a62ed749..2c3219e0c010 100644 --- a/fs/ocfs2/aops.h +++ b/fs/ocfs2/aops.h @@ -63,32 +63,54 @@ int ocfs2_size_fits_inline_data(struct buffer_head *di_bh, u64 new_size); int ocfs2_get_block(struct inode *inode, sector_t iblock, struct buffer_head *bh_result, int create); -/* all ocfs2_dio_end_io()'s fault */ -#define ocfs2_iocb_is_rw_locked(iocb) \ - test_bit(0, (unsigned long *)&iocb->private) + +/* + * Direct I/O uses iocb->private as a tagged pointer. The bottom two bits + * defined below are used for communication between ocfs2_dio_end_io() and + * ocfs2_file_write/read_iter(). + */ +#define OCFS2_IOCB_RW_LOCK 1 +#define OCFS2_IOCB_RW_LOCK_LEVEL 2 + +static inline void *ocfs2_iocb_private(struct kiocb *iocb) +{ + return (void *)((unsigned long)iocb->private & ~3); +} + +static inline void ocfs2_iocb_set_private(struct kiocb *iocb, void *private) +{ + iocb->private = (void *)(((unsigned long)iocb->private & 3) | + ((unsigned long)private & ~3)); +} + +static inline bool ocfs2_iocb_is_rw_locked(struct kiocb *iocb) +{ + return (unsigned long)iocb->private & OCFS2_IOCB_RW_LOCK; +} + static inline void ocfs2_iocb_set_rw_locked(struct kiocb *iocb, int level) { - set_bit(0, (unsigned long *)&iocb->private); + unsigned long private = (unsigned long)iocb->private; + + private |= OCFS2_IOCB_RW_LOCK; if (level) - set_bit(1, (unsigned long *)&iocb->private); + private |= OCFS2_IOCB_RW_LOCK_LEVEL; else - clear_bit(1, (unsigned long *)&iocb->private); + private &= ~OCFS2_IOCB_RW_LOCK_LEVEL; + iocb->private = (void *)private; } -/* - * Using a named enum representing lock types in terms of #N bit stored in - * iocb->private, which is going to be used for communication between - * ocfs2_dio_end_io() and ocfs2_file_write/read_iter(). - */ -enum ocfs2_iocb_lock_bits { - OCFS2_IOCB_RW_LOCK = 0, - OCFS2_IOCB_RW_LOCK_LEVEL, - OCFS2_IOCB_NUM_LOCKS -}; - -#define ocfs2_iocb_clear_rw_locked(iocb) \ - clear_bit(OCFS2_IOCB_RW_LOCK, (unsigned long *)&iocb->private) -#define ocfs2_iocb_rw_locked_level(iocb) \ - test_bit(OCFS2_IOCB_RW_LOCK_LEVEL, (unsigned long *)&iocb->private) +static inline void ocfs2_iocb_clear_rw_locked(struct kiocb *iocb) +{ + unsigned long private = (unsigned long)iocb->private; + + private &= ~OCFS2_IOCB_RW_LOCK; + iocb->private = (void *)private; +} + +static inline bool ocfs2_iocb_rw_locked_level(struct kiocb *iocb) +{ + return (unsigned long)iocb->private & OCFS2_IOCB_RW_LOCK_LEVEL; +} #endif /* OCFS2_FILE_H */ -- 2.18.0 ^ permalink raw reply related [flat|nested] 7+ messages in thread
* [RFC PATCH v2 4/6] fs: stop propagating bh->b_private for direct I/O 2018-08-28 0:03 [RFC PATCH v2 0/6] Btrfs: stop abusing current->journal_info for direct I/O Omar Sandoval ` (2 preceding siblings ...) 2018-08-28 0:03 ` [RFC PATCH v2 3/6] ocfs2: " Omar Sandoval @ 2018-08-28 0:03 ` Omar Sandoval 2018-08-28 0:03 ` [RFC PATCH v2 5/6] fs: pass iocb to direct I/O submit_io() Omar Sandoval 2018-08-28 0:03 ` [RFC PATCH v2 6/6] Btrfs: stop abusing current->journal_info in btrfs_direct_IO() Omar Sandoval 5 siblings, 0 replies; 7+ messages in thread From: Omar Sandoval @ 2018-08-28 0:03 UTC (permalink / raw) To: linux-fsdevel, linux-btrfs, Al Viro; +Cc: David Sterba, kernel-team From: Omar Sandoval <osandov@fb.com> Currently, the direct I/O code saves the value of bh->b_private set by the filesystem and passes it to the end_io callback. However, struct kiocb already has a ->private member which can be used for this purpose, with the added benefit of being available before get_block is called, too. The only users of the bh->b_private functionality have been converted to use iocb->private, so stop passing it around. Signed-off-by: Omar Sandoval <osandov@fb.com> --- fs/direct-io.c | 7 +------ fs/ext4/inode.c | 3 +-- fs/ocfs2/aops.c | 5 +---- include/linux/fs.h | 3 +-- 4 files changed, 4 insertions(+), 14 deletions(-) diff --git a/fs/direct-io.c b/fs/direct-io.c index f631aa98849b..80e488afe6c6 100644 --- a/fs/direct-io.c +++ b/fs/direct-io.c @@ -122,8 +122,6 @@ struct dio { loff_t i_size; /* i_size when submitted */ dio_iodone_t *end_io; /* IO completion function */ - void *private; /* copy from map_bh.b_private */ - /* BIO completion state */ spinlock_t bio_lock; /* protects BIO fields below */ int page_errors; /* errno from get_user_pages() */ @@ -288,7 +286,7 @@ static ssize_t dio_complete(struct dio *dio, ssize_t ret, unsigned int flags) if (dio->end_io) { // XXX: ki_pos?? - err = dio->end_io(dio->iocb, offset, ret, dio->private); + err = dio->end_io(dio->iocb, offset, ret); if (err) ret = err; } @@ -716,9 +714,6 @@ static int get_more_blocks(struct dio *dio, struct dio_submit *sdio, ret = (*sdio->get_block)(dio->iocb, dio->inode, fs_startblk, map_bh, create); - /* Store for completion */ - dio->private = map_bh->b_private; - if (ret == 0 && buffer_defer_completion(map_bh)) ret = dio_set_defer_completion(dio); } diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 841d79919cef..0f42793765bf 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -3612,8 +3612,7 @@ const struct iomap_ops ext4_iomap_ops = { .iomap_end = ext4_iomap_end, }; -static int ext4_end_io_dio(struct kiocb *iocb, loff_t offset, - ssize_t size, void *private) +static int ext4_end_io_dio(struct kiocb *iocb, loff_t offset, ssize_t size) { ext4_io_end_t *io_end = iocb->private; diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c index fc4a18b6ad3c..c1232df20be5 100644 --- a/fs/ocfs2/aops.c +++ b/fs/ocfs2/aops.c @@ -2404,10 +2404,7 @@ static int ocfs2_dio_end_io_write(struct inode *inode, * particularly interested in the aio/dio case. We use the rw_lock DLM lock * to protect io on one node from truncation on another. */ -static int ocfs2_dio_end_io(struct kiocb *iocb, - loff_t offset, - ssize_t bytes, - void *private) +static int ocfs2_dio_end_io(struct kiocb *iocb, loff_t offset, ssize_t bytes) { struct ocfs2_dio_write_ctxt *dwc; struct inode *inode = file_inode(iocb->ki_filp); diff --git a/include/linux/fs.h b/include/linux/fs.h index 85db69835023..f1a235f0fa21 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -83,8 +83,7 @@ typedef int (get_block_t)(struct inode *inode, sector_t iblock, typedef int (dio_get_block_t)(struct kiocb *iocb, struct inode *inode, sector_t iblock, struct buffer_head *bh_result, int create); -typedef int (dio_iodone_t)(struct kiocb *iocb, loff_t offset, - ssize_t bytes, void *private); +typedef int (dio_iodone_t)(struct kiocb *iocb, loff_t offset, ssize_t bytes); #define MAY_EXEC 0x00000001 #define MAY_WRITE 0x00000002 -- 2.18.0 ^ permalink raw reply related [flat|nested] 7+ messages in thread
* [RFC PATCH v2 5/6] fs: pass iocb to direct I/O submit_io() 2018-08-28 0:03 [RFC PATCH v2 0/6] Btrfs: stop abusing current->journal_info for direct I/O Omar Sandoval ` (3 preceding siblings ...) 2018-08-28 0:03 ` [RFC PATCH v2 4/6] fs: stop propagating bh->b_private for direct I/O Omar Sandoval @ 2018-08-28 0:03 ` Omar Sandoval 2018-08-28 0:03 ` [RFC PATCH v2 6/6] Btrfs: stop abusing current->journal_info in btrfs_direct_IO() Omar Sandoval 5 siblings, 0 replies; 7+ messages in thread From: Omar Sandoval @ 2018-08-28 0:03 UTC (permalink / raw) To: linux-fsdevel, linux-btrfs, Al Viro; +Cc: David Sterba, kernel-team From: Omar Sandoval <osandov@fb.com> Btrfs abuses current->journal_info in btrfs_direct_IO() in order to pass around some state to get_block() and submit_io(). However, iocb->private is free for Btrfs to use, we just need it passed to submit_io(). Btrfs is the only user of submit_io(), so this doesn't affect any other filesystems. Signed-off-by: Omar Sandoval <osandov@fb.com> --- fs/btrfs/inode.c | 4 ++-- fs/direct-io.c | 3 ++- include/linux/fs.h | 4 ++-- 3 files changed, 6 insertions(+), 5 deletions(-) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index b61ea6dd9956..6efa6a6e3e20 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -8427,8 +8427,8 @@ static int btrfs_submit_direct_hook(struct btrfs_dio_private *dip) return 0; } -static void btrfs_submit_direct(struct bio *dio_bio, struct inode *inode, - loff_t file_offset) +static void btrfs_submit_direct(struct kiocb *iocb, struct bio *dio_bio, + struct inode *inode, loff_t file_offset) { struct btrfs_dio_private *dip = NULL; struct bio *bio = NULL; diff --git a/fs/direct-io.c b/fs/direct-io.c index 80e488afe6c6..aa367e70456d 100644 --- a/fs/direct-io.c +++ b/fs/direct-io.c @@ -473,7 +473,8 @@ static inline void dio_bio_submit(struct dio *dio, struct dio_submit *sdio) dio->bio_disk = bio->bi_disk; if (sdio->submit_io) { - sdio->submit_io(bio, dio->inode, sdio->logical_offset_in_bio); + sdio->submit_io(dio->iocb, bio, dio->inode, + sdio->logical_offset_in_bio); dio->bio_cookie = BLK_QC_T_NONE; } else dio->bio_cookie = submit_bio(bio); diff --git a/include/linux/fs.h b/include/linux/fs.h index f1a235f0fa21..daf1df811f67 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -3003,8 +3003,8 @@ extern int generic_file_open(struct inode * inode, struct file * filp); extern int nonseekable_open(struct inode * inode, struct file * filp); #ifdef CONFIG_BLOCK -typedef void (dio_submit_t)(struct bio *bio, struct inode *inode, - loff_t file_offset); +typedef void (dio_submit_t)(struct kiocb *iocb, struct bio *bio, + struct inode *inode, loff_t file_offset); enum { /* need locking between buffered and direct access */ -- 2.18.0 ^ permalink raw reply related [flat|nested] 7+ messages in thread
* [RFC PATCH v2 6/6] Btrfs: stop abusing current->journal_info in btrfs_direct_IO() 2018-08-28 0:03 [RFC PATCH v2 0/6] Btrfs: stop abusing current->journal_info for direct I/O Omar Sandoval ` (4 preceding siblings ...) 2018-08-28 0:03 ` [RFC PATCH v2 5/6] fs: pass iocb to direct I/O submit_io() Omar Sandoval @ 2018-08-28 0:03 ` Omar Sandoval 5 siblings, 0 replies; 7+ messages in thread From: Omar Sandoval @ 2018-08-28 0:03 UTC (permalink / raw) To: linux-fsdevel, linux-btrfs, Al Viro; +Cc: David Sterba, kernel-team From: Omar Sandoval <osandov@fb.com> Now that we can pass around the struct btrfs_dio_data through the different callbacks generically, we don't need to shove it in current->journal_info. Signed-off-by: Omar Sandoval <osandov@fb.com> --- fs/btrfs/inode.c | 29 ++++++----------------------- 1 file changed, 6 insertions(+), 23 deletions(-) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 6efa6a6e3e20..38a41e9d6e93 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -7654,7 +7654,6 @@ static int btrfs_get_blocks_direct_write(struct extent_map **map, WARN_ON(dio_data->reserve < len); dio_data->reserve -= len; dio_data->unsubmitted_oe_range_end = start + len; - current->journal_info = dio_data; out: return ret; } @@ -7666,7 +7665,7 @@ static int btrfs_get_blocks_direct(struct kiocb *iocb, struct inode *inode, struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb); struct extent_map *em; struct extent_state *cached_state = NULL; - struct btrfs_dio_data *dio_data = NULL; + struct btrfs_dio_data *dio_data = iocb->private; u64 start = iblock << inode->i_blkbits; u64 lockstart, lockend; u64 len = bh_result->b_size; @@ -7681,25 +7680,13 @@ static int btrfs_get_blocks_direct(struct kiocb *iocb, struct inode *inode, lockstart = start; lockend = start + len - 1; - if (current->journal_info) { - /* - * Need to pull our outstanding extents and set journal_info to NULL so - * that anything that needs to check if there's a transaction doesn't get - * confused. - */ - dio_data = current->journal_info; - current->journal_info = NULL; - } - /* * If this errors out it's because we couldn't invalidate pagecache for * this range and we need to fallback to buffered. */ if (lock_extent_direct(inode, lockstart, lockend, &cached_state, - create)) { - ret = -ENOTBLK; - goto err; - } + create)) + return -ENOTBLK; em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, start, len, 0); if (IS_ERR(em)) { @@ -7767,9 +7754,6 @@ static int btrfs_get_blocks_direct(struct kiocb *iocb, struct inode *inode, unlock_err: clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, lockend, unlock_bits, 1, 0, &cached_state); -err: - if (dio_data) - current->journal_info = dio_data; return ret; } @@ -8470,7 +8454,7 @@ static void btrfs_submit_direct(struct kiocb *iocb, struct bio *dio_bio, * time by btrfs_direct_IO(). */ if (write) { - struct btrfs_dio_data *dio_data = current->journal_info; + struct btrfs_dio_data *dio_data = iocb->private; dio_data->unsubmitted_oe_range_end = dip->logical_offset + dip->bytes; @@ -8612,13 +8596,13 @@ static ssize_t btrfs_direct_IO(struct kiocb *iocb, struct iov_iter *iter) /* * We need to know how many extents we reserved so that we can * do the accounting properly if we go over the number we - * originally calculated. Abuse current->journal_info for this. + * originally calculated. */ dio_data.reserve = round_up(count, fs_info->sectorsize); dio_data.unsubmitted_oe_range_start = (u64)offset; dio_data.unsubmitted_oe_range_end = (u64)offset; - current->journal_info = &dio_data; + iocb->private = &dio_data; down_read(&BTRFS_I(inode)->dio_sem); } else if (test_bit(BTRFS_INODE_READDIO_NEED_LOCK, &BTRFS_I(inode)->runtime_flags)) { @@ -8633,7 +8617,6 @@ static ssize_t btrfs_direct_IO(struct kiocb *iocb, struct iov_iter *iter) btrfs_submit_direct, flags); if (iov_iter_rw(iter) == WRITE) { up_read(&BTRFS_I(inode)->dio_sem); - current->journal_info = NULL; if (ret < 0 && ret != -EIOCBQUEUED) { if (dio_data.reserve) btrfs_delalloc_release_space(inode, data_reserved, -- 2.18.0 ^ permalink raw reply related [flat|nested] 7+ messages in thread
end of thread, other threads:[~2018-08-28 3:52 UTC | newest] Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2018-08-28 0:03 [RFC PATCH v2 0/6] Btrfs: stop abusing current->journal_info for direct I/O Omar Sandoval 2018-08-28 0:03 ` [RFC PATCH v2 1/6] fs: pass iocb to direct I/O get_block() Omar Sandoval 2018-08-28 0:03 ` [RFC PATCH v2 2/6] ext4: use iocb->private instead of bh->b_private Omar Sandoval 2018-08-28 0:03 ` [RFC PATCH v2 3/6] ocfs2: " Omar Sandoval 2018-08-28 0:03 ` [RFC PATCH v2 4/6] fs: stop propagating bh->b_private for direct I/O Omar Sandoval 2018-08-28 0:03 ` [RFC PATCH v2 5/6] fs: pass iocb to direct I/O submit_io() Omar Sandoval 2018-08-28 0:03 ` [RFC PATCH v2 6/6] Btrfs: stop abusing current->journal_info in btrfs_direct_IO() Omar Sandoval
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).