Message ID | f4abda9c0c835d9a50b644fdbec8d43269f6b0f7.1565609891.git.mbobrowski@mbobrowski.org (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | ext4: direct IO via iomap infrastructure | expand |
please add linux-xfs to the cc list for the whole series. Besides touching xfs itself it is also mentioned in MAINTAINERS for the iomap code. On Mon, Aug 12, 2019 at 10:53:11PM +1000, Matthew Bobrowski wrote: > - if (size <= 0) > - return size; > + if (error || !size) > + return error ? error : size; This should be: if (error) return error; if (!size) return 0; > if (flags & IOMAP_DIO_COW) { > - error = xfs_reflink_end_cow(ip, offset, size); > - if (error) > + ret = xfs_reflink_end_cow(ip, offset, size); > + if (ret) I think we can just keep reusing error here. > +typedef int (iomap_dio_end_io_t)(struct kiocb *iocb, ssize_t size, > + ssize_t error, unsigned int flags); error should be an int and not a ssize_t.
On Mon, Aug 12, 2019 at 10:18:32AM -0700, Christoph Hellwig wrote: > please add linux-xfs to the cc list for the whole series. Besides > touching xfs itself it is also mentioned in MAINTAINERS for the iomap > code. Firstly, thank you for the review, highly appreciated! Secondly, not a problem, will do. > On Mon, Aug 12, 2019 at 10:53:11PM +1000, Matthew Bobrowski wrote: > > - if (size <= 0) > > - return size; > > + if (error || !size) > > + return error ? error : size; > > This should be: > > if (error) > return error; > if (!size) > return 0; OK. > > if (flags & IOMAP_DIO_COW) { > > - error = xfs_reflink_end_cow(ip, offset, size); > > - if (error) > > + ret = xfs_reflink_end_cow(ip, offset, size); > > + if (ret) > > I think we can just keep reusing error here. Ah yes, that will work. > > +typedef int (iomap_dio_end_io_t)(struct kiocb *iocb, ssize_t size, > > + ssize_t error, unsigned int flags); > > error should be an int and not a ssize_t. Updated. --M
diff --git a/fs/iomap/direct-io.c b/fs/iomap/direct-io.c index 10517cea9682..2ccf1c6460d4 100644 --- a/fs/iomap/direct-io.c +++ b/fs/iomap/direct-io.c @@ -77,13 +77,10 @@ static ssize_t iomap_dio_complete(struct iomap_dio *dio) loff_t offset = iocb->ki_pos; ssize_t ret; - if (dio->end_io) { - ret = dio->end_io(iocb, - dio->error ? dio->error : dio->size, - dio->flags); - } else { + if (dio->end_io) + ret = dio->end_io(iocb, dio->size, dio->error, dio->flags); + else ret = dio->error; - } if (likely(!ret)) { ret = dio->size; diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c index 28101bbc0b78..f2bc3ac4a60e 100644 --- a/fs/xfs/xfs_file.c +++ b/fs/xfs/xfs_file.c @@ -369,21 +369,22 @@ static int xfs_dio_write_end_io( struct kiocb *iocb, ssize_t size, + ssize_t error, unsigned flags) { struct inode *inode = file_inode(iocb->ki_filp); struct xfs_inode *ip = XFS_I(inode); loff_t offset = iocb->ki_pos; unsigned int nofs_flag; - int error = 0; + int ret = 0; trace_xfs_end_io_direct_write(ip, offset, size); if (XFS_FORCED_SHUTDOWN(ip->i_mount)) return -EIO; - if (size <= 0) - return size; + if (error || !size) + return error ? error : size; /* * Capture amount written on completion as we can't reliably account @@ -399,8 +400,8 @@ xfs_dio_write_end_io( nofs_flag = memalloc_nofs_save(); if (flags & IOMAP_DIO_COW) { - error = xfs_reflink_end_cow(ip, offset, size); - if (error) + ret = xfs_reflink_end_cow(ip, offset, size); + if (ret) goto out; } @@ -411,7 +412,7 @@ xfs_dio_write_end_io( * they are converted. */ if (flags & IOMAP_DIO_UNWRITTEN) { - error = xfs_iomap_write_unwritten(ip, offset, size, true); + ret = xfs_iomap_write_unwritten(ip, offset, size, true); goto out; } @@ -430,14 +431,14 @@ xfs_dio_write_end_io( if (offset + size > i_size_read(inode)) { i_size_write(inode, offset + size); spin_unlock(&ip->i_flags_lock); - error = xfs_setfilesize(ip, offset, size); + ret = xfs_setfilesize(ip, offset, size); } else { spin_unlock(&ip->i_flags_lock); } out: memalloc_nofs_restore(nofs_flag); - return error; + return ret; } /* diff --git a/include/linux/iomap.h b/include/linux/iomap.h index bc499ceae392..900284e5c06c 100644 --- a/include/linux/iomap.h +++ b/include/linux/iomap.h @@ -188,8 +188,8 @@ sector_t iomap_bmap(struct address_space *mapping, sector_t bno, */ #define IOMAP_DIO_UNWRITTEN (1 << 0) /* covers unwritten extent(s) */ #define IOMAP_DIO_COW (1 << 1) /* covers COW extent(s) */ -typedef int (iomap_dio_end_io_t)(struct kiocb *iocb, ssize_t ret, - unsigned flags); +typedef int (iomap_dio_end_io_t)(struct kiocb *iocb, ssize_t size, + ssize_t error, unsigned int flags); ssize_t iomap_dio_rw(struct kiocb *iocb, struct iov_iter *iter, const struct iomap_ops *ops, iomap_dio_end_io_t end_io); int iomap_dio_iopoll(struct kiocb *kiocb, bool spin);
This patch modifies the calling convention for the iomap ->end_io() callback. Rather than passing either dio->error or dio->size as the 'size' argument, we instead pass both dio->error and dio->size values separately. In the instance that an error occurred during a write, we currently cannot determine whether any blocks have been allocated beyond the current EOF and data has subsequently been written to these blocks within the ->end_io() callback. As a result, we cannot judge whether we should take the truncate failed write path. Having both dio->error and dio->size will allow us to perform such checks within this callback. Signed-off-by: Matthew Bobrowski <mbobrowski@mbobrowski.org> --- fs/iomap/direct-io.c | 9 +++------ fs/xfs/xfs_file.c | 17 +++++++++-------- include/linux/iomap.h | 4 ++-- 3 files changed, 14 insertions(+), 16 deletions(-)