Am 13.12.2013 um 14:22 hat Kevin Wolf geschrieben:
> This is going to become the bdrv_co_do_preadv() equivalent for writes.
> In this patch, however, just a function taking byte offsets is created,
> it doesn't align anything yet.
>
> Signed-off-by: Kevin Wolf <kw...@redhat.com>
> ---
> block.c | 23 +++++++++++++++++------
> 1 file changed, 17 insertions(+), 6 deletions(-)
>
> diff --git a/block.c b/block.c
> index 385fb8a..a80db2e 100644
> --- a/block.c
> +++ b/block.c
> @@ -3010,8 +3010,8 @@ static int coroutine_fn
> bdrv_aligned_pwritev(BlockDriverState *bs,
> /*
> * Handle a write request in coroutine context
> */
> -static int coroutine_fn bdrv_co_do_writev(BlockDriverState *bs,
> - int64_t sector_num, int nb_sectors, QEMUIOVector *qiov,
> +static int coroutine_fn bdrv_co_do_pwritev(BlockDriverState *bs,
> + int64_t offset, unsigned int bytes, QEMUIOVector *qiov,
> BdrvRequestFlags flags)
> {
> int ret;
> @@ -3022,21 +3022,32 @@ static int coroutine_fn
> bdrv_co_do_writev(BlockDriverState *bs,
> if (bs->read_only) {
> return -EACCES;
> }
> - if (bdrv_check_request(bs, sector_num, nb_sectors)) {
> + if (bdrv_check_byte_request(bs, offset, bytes)) {
> return -EIO;
> }
>
> /* throttling disk I/O */
> if (bs->io_limits_enabled) {
> - bdrv_io_limits_intercept(bs, nb_sectors, true);
> + bdrv_io_limits_intercept(bs, bytes << BDRV_SECTOR_BITS, true);
> }
Oh nice, this shifts in the wrong direction.
If somebody feels like writing a test case, something testing that I/O
throttling actually throttles would be useful...
Kevin
