[RFC PATCH 02/11] vfs: Set IOCB_WRITE in iocbs that we're going to write from

From: David Howells
Date: Fri Jun 30 2023 - 11:19:30 EST


IOCB_WRITE is set by aio, io_uring and cachefiles before submitting a write
operation to the VFS, but it isn't set by, say, the write() system call.

Fix this by adding an extra argument to init_sync_kiocb() to indicate the
direction and setting that to READ or WRITE, which will cause IOCB_WRITE to
be set as appropriate.

Whilst we're at it, rename init_sync_kiocb() to init_kiocb().

This will allow drivers to use IOCB_WRITE instead of the iterator data
source to determine the I/O direction.

Signed-off-by: David Howells <dhowells@xxxxxxxxxx>
cc: Christoph Hellwig <hch@xxxxxx>
cc: Jens Axboe <axboe@xxxxxxxxx>
cc: Christian Brauner <christian@xxxxxxxxxx>
cc: Alexander Viro <viro@xxxxxxxxxxxxxxxxxx>
cc: linux-block@xxxxxxxxxxxxxxx
cc: linux-fsdevel@xxxxxxxxxxxxxxx
---
fs/btrfs/ioctl.c | 4 ++--
fs/read_write.c | 10 +++++-----
fs/seq_file.c | 2 +-
fs/splice.c | 2 +-
include/linux/fs.h | 6 +++++-
mm/filemap.c | 2 +-
mm/page_io.c | 4 ++--
7 files changed, 17 insertions(+), 13 deletions(-)

diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index a895d105464b..15870337dd26 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -4422,7 +4422,7 @@ static int btrfs_ioctl_encoded_read(struct file *file, void __user *argp,
if (ret < 0)
goto out_iov;

- init_sync_kiocb(&kiocb, file);
+ init_kiocb(&kiocb, file, READ);
kiocb.ki_pos = pos;

ret = btrfs_encoded_read(&kiocb, &iter, &args);
@@ -4523,7 +4523,7 @@ static int btrfs_ioctl_encoded_write(struct file *file, void __user *argp, bool
if (ret < 0)
goto out_end_write;

- init_sync_kiocb(&kiocb, file);
+ init_kiocb(&kiocb, file, WRITE);
ret = kiocb_set_rw_flags(&kiocb, 0);
if (ret)
goto out_end_write;
diff --git a/fs/read_write.c b/fs/read_write.c
index b07de77ef126..6fe517047095 100644
--- a/fs/read_write.c
+++ b/fs/read_write.c
@@ -382,7 +382,7 @@ static ssize_t new_sync_read(struct file *filp, char __user *buf, size_t len, lo
struct iov_iter iter;
ssize_t ret;

- init_sync_kiocb(&kiocb, filp);
+ init_kiocb(&kiocb, filp, READ);
kiocb.ki_pos = (ppos ? *ppos : 0);
iov_iter_ubuf(&iter, ITER_DEST, buf, len);

@@ -422,7 +422,7 @@ ssize_t __kernel_read(struct file *file, void *buf, size_t count, loff_t *pos)
if (unlikely(!file->f_op->read_iter || file->f_op->read))
return warn_unsupported(file, "read");

- init_sync_kiocb(&kiocb, file);
+ init_kiocb(&kiocb, file, READ);
kiocb.ki_pos = pos ? *pos : 0;
iov_iter_kvec(&iter, ITER_DEST, &iov, 1, iov.iov_len);
ret = file->f_op->read_iter(&kiocb, &iter);
@@ -484,7 +484,7 @@ static ssize_t new_sync_write(struct file *filp, const char __user *buf, size_t
struct iov_iter iter;
ssize_t ret;

- init_sync_kiocb(&kiocb, filp);
+ init_kiocb(&kiocb, filp, WRITE);
kiocb.ki_pos = (ppos ? *ppos : 0);
iov_iter_ubuf(&iter, ITER_SOURCE, (void __user *)buf, len);

@@ -512,7 +512,7 @@ ssize_t __kernel_write_iter(struct file *file, struct iov_iter *from, loff_t *po
if (unlikely(!file->f_op->write_iter || file->f_op->write))
return warn_unsupported(file, "write");

- init_sync_kiocb(&kiocb, file);
+ init_kiocb(&kiocb, file, WRITE);
kiocb.ki_pos = pos ? *pos : 0;
ret = file->f_op->write_iter(&kiocb, from);
if (ret > 0) {
@@ -723,7 +723,7 @@ static ssize_t do_iter_readv_writev(struct file *filp, struct iov_iter *iter,
struct kiocb kiocb;
ssize_t ret;

- init_sync_kiocb(&kiocb, filp);
+ init_kiocb(&kiocb, filp, type);
ret = kiocb_set_rw_flags(&kiocb, flags);
if (ret)
return ret;
diff --git a/fs/seq_file.c b/fs/seq_file.c
index f5fdaf3b1572..1ee6ffc630da 100644
--- a/fs/seq_file.c
+++ b/fs/seq_file.c
@@ -155,7 +155,7 @@ ssize_t seq_read(struct file *file, char __user *buf, size_t size, loff_t *ppos)
struct iov_iter iter;
ssize_t ret;

- init_sync_kiocb(&kiocb, file);
+ init_kiocb(&kiocb, file, READ);
iov_iter_init(&iter, ITER_DEST, &iov, 1, size);

kiocb.ki_pos = *ppos;
diff --git a/fs/splice.c b/fs/splice.c
index 004eb1c4ce31..867357ebb2c3 100644
--- a/fs/splice.c
+++ b/fs/splice.c
@@ -362,7 +362,7 @@ ssize_t copy_splice_read(struct file *in, loff_t *ppos,

/* Do the I/O */
iov_iter_bvec(&to, ITER_DEST, bv, npages, len);
- init_sync_kiocb(&kiocb, in);
+ init_kiocb(&kiocb, in, READ);
kiocb.ki_pos = *ppos;
ret = call_read_iter(in, &kiocb, &to);

diff --git a/include/linux/fs.h b/include/linux/fs.h
index d4b67bdeb53e..466eba253502 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -2017,13 +2017,17 @@ static inline bool HAS_UNMAPPED_ID(struct mnt_idmap *idmap,
!vfsgid_valid(i_gid_into_vfsgid(idmap, inode));
}

-static inline void init_sync_kiocb(struct kiocb *kiocb, struct file *filp)
+static inline void init_kiocb(struct kiocb *kiocb, struct file *filp,
+ unsigned int rw)
{
*kiocb = (struct kiocb) {
.ki_filp = filp,
.ki_flags = filp->f_iocb_flags,
.ki_ioprio = get_current_ioprio(),
};
+
+ if (rw == WRITE)
+ kiocb->ki_flags |= IOCB_WRITE;
}

static inline void kiocb_clone(struct kiocb *kiocb, struct kiocb *kiocb_src,
diff --git a/mm/filemap.c b/mm/filemap.c
index 9e44a49bbd74..cd763122d2a2 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -2905,7 +2905,7 @@ ssize_t filemap_splice_read(struct file *in, loff_t *ppos,
if (unlikely(*ppos >= in->f_mapping->host->i_sb->s_maxbytes))
return 0;

- init_sync_kiocb(&iocb, in);
+ init_kiocb(&iocb, in, READ);
iocb.ki_pos = *ppos;

/* Work out how much data we can actually add into the pipe */
diff --git a/mm/page_io.c b/mm/page_io.c
index 684cd3c7b59b..85cbadaf7395 100644
--- a/mm/page_io.c
+++ b/mm/page_io.c
@@ -312,7 +312,7 @@ static void swap_writepage_fs(struct page *page, struct writeback_control *wbc)
}
if (!sio) {
sio = mempool_alloc(sio_pool, GFP_NOIO);
- init_sync_kiocb(&sio->iocb, swap_file);
+ init_kiocb(&sio->iocb, swap_file, WRITE);
sio->iocb.ki_complete = sio_write_complete;
sio->iocb.ki_pos = pos;
sio->pages = 0;
@@ -443,7 +443,7 @@ static void swap_readpage_fs(struct page *page,
}
if (!sio) {
sio = mempool_alloc(sio_pool, GFP_KERNEL);
- init_sync_kiocb(&sio->iocb, sis->swap_file);
+ init_kiocb(&sio->iocb, sis->swap_file, READ);
sio->iocb.ki_pos = pos;
sio->iocb.ki_complete = sio_read_complete;
sio->pages = 0;