* Wei Yang (richardw.y...@linux.intel.com) wrote:
> The check of writev_buffer is in qemu_fflush, which means it is not
> harmful if it is NULL.
>
> And removing it will make the code consistent since all other
> add_to_iovec() is called without the check.
>
> Signed-off-by: Wei Yang <richardw.y...@linux.intel.com>
So I think the real reason here is commit 11808bb which removed
support for non-iovec writers.
Reviewed-by: Dr. David Alan Gilbert <dgilb...@redhat.com>
> ---
> migration/qemu-file.c | 4 +---
> 1 file changed, 1 insertion(+), 3 deletions(-)
>
> diff --git a/migration/qemu-file.c b/migration/qemu-file.c
> index 0431585502..35c22605dd 100644
> --- a/migration/qemu-file.c
> +++ b/migration/qemu-file.c
> @@ -717,9 +717,7 @@ ssize_t qemu_put_compression_data(QEMUFile *f, z_stream
> *stream,
> }
>
> qemu_put_be32(f, blen);
> - if (f->ops->writev_buffer) {
> - add_to_iovec(f, f->buf + f->buf_index, blen, false);
> - }
> + add_to_iovec(f, f->buf + f->buf_index, blen, false);
> f->buf_index += blen;
> if (f->buf_index == IO_BUF_SIZE) {
> qemu_fflush(f);
> --
> 2.17.1
>
--
Dr. David Alan Gilbert / dgilb...@redhat.com / Manchester, UK
