Hi Andrew,
Please send this to upstream. This fixes a regression of splice() on nilfs2.
Thanks,
Ryusuke Konishi
On Mon, Feb 8, 2021 at 8:35 PM Ryusuke Konishi
<konishi.ryus...@gmail.com> wrote:
>
> From: Joachim Henke <joachim.he...@t-systems.com>
>
> Since 5.10, splice() or sendfile() to NILFS2 return EINVAL. This was
> caused by commit 36e2c7421f02 ("fs: don't allow splice read/write
> without explicit ops").
>
> This patch initializes the splice_write field in file_operations, like
> most file systems do, to restore the functionality.
>
> Signed-off-by: Joachim Henke <joachim.he...@t-systems.com>
> Signed-off-by: Ryusuke Konishi <konishi.ryus...@gmail.com>
> Tested-by: Ryusuke Konishi <konishi.ryus...@gmail.com>
> Cc: sta...@vger.kernel.org # 5.10+
> ---
> fs/nilfs2/file.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/fs/nilfs2/file.c b/fs/nilfs2/file.c
> index 64bc81363c6c..e1bd592ce700 100644
> --- a/fs/nilfs2/file.c
> +++ b/fs/nilfs2/file.c
> @@ -141,6 +141,7 @@ static int nilfs_file_mmap(struct file *file, struct
> vm_area_struct *vma)
> /* .release = nilfs_release_file, */
> .fsync = nilfs_sync_file,
> .splice_read = generic_file_splice_read,
> + .splice_write = iter_file_splice_write,
> };
>
> const struct inode_operations nilfs_file_inode_operations = {
> --
> 1.8.3.1
>
