[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH v4 7/8] block: core copy-on-read logic
From: |
Marcelo Tosatti |
Subject: |
Re: [Qemu-devel] [PATCH v4 7/8] block: core copy-on-read logic |
Date: |
Fri, 2 Dec 2011 15:13:19 -0200 |
User-agent: |
Mutt/1.5.21 (2010-09-15) |
On Wed, Nov 23, 2011 at 11:47:57AM +0000, Stefan Hajnoczi wrote:
> Signed-off-by: Stefan Hajnoczi <address@hidden>
> ---
> block.c | 72
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
> trace-events | 1 +
> 2 files changed, 73 insertions(+), 0 deletions(-)
>
> diff --git a/block.c b/block.c
> index c30c8f2..a598a19 100644
> --- a/block.c
> +++ b/block.c
> @@ -1469,6 +1469,61 @@ int bdrv_pwrite_sync(BlockDriverState *bs, int64_t
> offset,
> return 0;
> }
>
> +static int coroutine_fn bdrv_co_copy_on_readv(BlockDriverState *bs,
> + int64_t sector_num, int nb_sectors, QEMUIOVector *qiov)
> +{
> + /* Perform I/O through a temporary buffer so that users who scribble over
> + * their read buffer while the operation is in progress do not end up
> + * modifying the image file. This is critical for zero-copy guest I/O
> + * where anything might happen inside guest memory.
> + */
> + void *bounce_buffer;
> +
> + struct iovec iov;
> + QEMUIOVector bounce_qiov;
> + int64_t cluster_sector_num;
> + int cluster_nb_sectors;
> + size_t skip_bytes;
> + int ret;
> +
> + /* Cover entire cluster so no additional backing file I/O is required
> when
> + * allocating cluster in the image file.
> + */
> + round_to_clusters(bs, sector_num, nb_sectors,
> + &cluster_sector_num, &cluster_nb_sectors);
> +
> + trace_bdrv_co_copy_on_readv(bs, sector_num, nb_sectors,
> + cluster_sector_num, cluster_nb_sectors);
> +
> + iov.iov_len = cluster_nb_sectors * BDRV_SECTOR_SIZE;
> + iov.iov_base = bounce_buffer = qemu_blockalign(bs, iov.iov_len);
> + qemu_iovec_init_external(&bounce_qiov, &iov, 1);
> +
> + ret = bs->drv->bdrv_co_readv(bs, cluster_sector_num, cluster_nb_sectors,
> + &bounce_qiov);
> + if (ret < 0) {
> + goto err;
> + }
> +
> + ret = bs->drv->bdrv_co_writev(bs, cluster_sector_num, cluster_nb_sectors,
> + &bounce_qiov);
> + if (ret < 0) {
> + /* It might be okay to ignore write errors for guest requests. If
> this
> + * is a deliberate copy-on-read then we don't want to ignore the
> error.
> + * Simply report it in all cases.
> + */
> + goto err;
> + }
> +
> + skip_bytes = (sector_num - cluster_sector_num) * BDRV_SECTOR_SIZE;
> + qemu_iovec_from_buffer(qiov, bounce_buffer + skip_bytes,
> + nb_sectors * BDRV_SECTOR_SIZE);
> +
> +err:
> + qemu_vfree(bounce_buffer);
> + return ret;
> +}
> +
> /*
> * Handle a read request in coroutine context
> */
> @@ -1496,7 +1551,24 @@ static int coroutine_fn
> bdrv_co_do_readv(BlockDriverState *bs,
> }
>
> tracked_request_begin(&req, bs, sector_num, nb_sectors, false);
> +
> + if (bs->copy_on_read) {
> + int pnum;
> +
> + ret = bdrv_co_is_allocated(bs, sector_num, nb_sectors, &pnum);
> + if (ret < 0) {
> + goto out;
> + }
Stefan,
It is not clear where support for shared backing files would fit.
Let us consider the following block copy example:
1) Original chain:
[ BASE ] -> [ IMAGE-1 ] -> [ IMAGE-2 ] -> [ IMAGE-3 ]
2) Final chain:
[ BASE ] -> [ IMAGE-3 ]
I was talking to Kevin and we don't have code/monitor command to the
switch from 1) to 2). But that is a separate issue.
Question is, how do you plan to stream the contents of IMAGE-1 and
IMAGE-2 (but not BASE), into IMAGE-3 ? That is an important use case.
Also, do you have status on the external COW file work, for raw images?
Thanks
- Re: [Qemu-devel] [PATCH v4 7/8] block: core copy-on-read logic,
Marcelo Tosatti <=