Re: [PATCH] block/file-posix: Limit max_iov to IOV_MAX

["Daniel P. Berrangé" <berrange@redhat.com>]

CC'ing qemu-block list

On Sat, Sep 18, 2021 at 12:06:58PM +0800, lishan wrote:
> AIO read/write. The size of iocb->aio_nbytes in the kernel cannot exceed UIO_MAXIOV = 1024.
> max_segments read from the block device layer may be greater than UIO_MAXIOV,
> this causes the ioq_submit interface to return a -22(-EINVAL) error result.
> ---
>  block/file-posix.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/block/file-posix.c b/block/file-posix.c
> index d81e15efa4..137e27e47b 100644
> --- a/block/file-posix.c
> +++ b/block/file-posix.c
> @@ -1273,7 +1273,8 @@ static void raw_refresh_limits(BlockDriverState *bs, Error **errp)
>  
>          ret = hdev_get_max_segments(s->fd, &st);
>          if (ret > 0) {
> -            bs->bl.max_iov = ret;
> +            /* The maximum segment size allowed by the kernel is UIO_MAXIOV = 1024. */
> +            bs->bl.max_iov = MIN(ret, IOV_MAX);

This change matches a bug fix we've done downstream for QEMU, but it
was suggested that the upstream patch would be taking a different
approach for a more comprehensive fix.


Regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|