Re: [Qemu-devel] [PATCH] block: Produce zeros when protocols reading beyond end of file

[Kevin Wolf <kwolf@redhat.com>]

Am 05.08.2013 um 10:11 hat Asias He geschrieben:
> From: MORITA Kazutaka <morita.kazutaka@lab.ntt.co.jp>
> 
> While Asias is debugging an issue creating qcow2 images on top of
> non-file protocols.  It boils down to this example using NBD:
> 
> $ qemu-io -c 'open -g nbd+unix:///?socket=/tmp/nbd.sock' -c 'read -v 0 512'
> 
> Notice the open -g option to set bs->growable.  This means you can
> read/write beyond end of file.  Reading beyond end of file is supposed
> to produce zeroes.
> 
> We rely on this behavior in qcow2_create2() during qcow2 image
> creation.  We create a new file and then write the qcow2 header
> structure using bdrv_pwrite().  Since QCowHeader is not a multiple of
> sector size, block.c first uses bdrv_read() on the empty file to fetch
> the first sector (should be all zeroes).
> 
> Here is the output from the qemu-io NBD example above:
> 
> $ qemu-io -c 'open -g nbd+unix:///?socket=/tmp/nbd.sock' -c 'read -v 0 512'
> 00000000:  ab ab ab ab ab ab ab ab ab ab ab ab ab ab ab ab  ................
> 00000010:  ab ab ab ab ab ab ab ab ab ab ab ab ab ab ab ab  ................
> 00000020:  ab ab ab ab ab ab ab ab ab ab ab ab ab ab ab ab  ................
> ...
> 
> We are not zeroing the buffer!  As a result qcow2 image creation on top
> of protocols is not guaranteed to work even when file creation is
> supported by the protocol.
> 
> Signed-off-by: MORITA Kazutaka <morita.kazutaka@lab.ntt.co.jp>
> Signed-off-by: Asias He <asias@redhat.com>
> ---
>  block.c | 30 +++++++++++++++++++++++++++++-
>  1 file changed, 29 insertions(+), 1 deletion(-)
> 
> diff --git a/block.c b/block.c
> index 01b66d8..deaf0a0 100644
> --- a/block.c
> +++ b/block.c
> @@ -2544,7 +2544,35 @@ static int coroutine_fn bdrv_co_do_readv(BlockDriverState *bs,
>          }
>      }
>  
> -    ret = drv->bdrv_co_readv(bs, sector_num, nb_sectors, qiov);
> +    if (!bs->drv->protocol_name) {

I think !bs->growable is the right check.

Checking for the protocol name is always a hack and most times wrong.

> +        ret = drv->bdrv_co_readv(bs, sector_num, nb_sectors, qiov);
> +    } else {
> +        /* NBD doesn't support reading beyond end of file. */

This is not only for NBD, make it a neutral comment like:

/* Read zeros after EOF of growable BDSes */

> +        int64_t len, total_sectors, max_nb_sectors;
> +
> +        len = bdrv_getlength(bs);
> +        if (len < 0) {
> +            ret = len;
> +            goto out;
> +        }
> +
> +        total_sectors = len >> BDRV_SECTOR_BITS;
> +        max_nb_sectors = MAX(0, total_sectors - sector_num);
> +        if (max_nb_sectors > 0) {
> +            ret = drv->bdrv_co_readv(bs, sector_num,
> +                                     MIN(nb_sectors, max_nb_sectors), qiov);
> +        } else {
> +            ret = 0;
> +        }
> +
> +        /* Reading beyond end of file is supposed to produce zeroes */
> +        if (ret == 0 && total_sectors < sector_num + nb_sectors) {
> +            size_t offset = MAX(0, total_sectors - sector_num);
> +            size_t bytes = (sector_num + nb_sectors - offset) *
> +                            BDRV_SECTOR_SIZE;

uint64_t for both offset and bytes, size_t can be 32 bits.

> +            qemu_iovec_memset(qiov, offset * BDRV_SECTOR_SIZE, 0, bytes);
> +        }
> +    }
>  
>  out:
>      tracked_request_end(&req);

Kevin