Fam Zheng <address@hidden> 於 2018-09-12 17:34 寫道：
On Tue, 08/28 11:17, address@hidden wrote:
> From: yuchenlin <address@hidden>
>
> There is a rare case which the size of last compressed cluster
> is larger than the cluster size, which will cause the file is
> not aligned at the sector boundary.
I don't understand. Doesn't it mean that if you force the alignment by
truncating out the extra bytes, some data is lost?
You can take qcow2_co_pwritev_compressed in block/qcow2.c as an example.
The bdrv_getlength will return the length in bytes which is always a multiple of BDRV_SECTOR_SIZE.
After truncates this size, the vmdk is extended to align sector size.
>
> Signed-off-by: yuchenlin <address@hidden>
> ---
> block/vmdk.c | 18 ++++++++++++++++++
> 1 file changed, 18 insertions(+)
>
> diff --git a/block/vmdk.c b/block/vmdk.c
> index a9d0084e36..a8ae7c65d2 100644
> --- a/block/vmdk.c
> +++ b/block/vmdk.c
> @@ -1698,6 +1698,24 @@ static int coroutine_fn
> vmdk_co_pwritev_compressed(BlockDriverState *bs, uint64_t offset,
> uint64_t bytes, QEMUIOVector *qiov)
> {
> + if (bytes == 0) {
Where is this bytes == 0 condition from?
From the end of convert_do_copy in qemu-img.c.
if (s->compressed && !s->ret) {
/* signal EOF to align */
ret = blk_pwrite_compressed(s->target, 0, NULL, 0);
if (ret < 0) {
return ret;
}
}
It signals the EOF to the block driver.
> + /* align end of file to a sector boundary. */
> + BDRVVmdkState *s = bs->opaque;
> + int i, ret;
> + int64_t length;
> +
> + for (i = 0; i < s->num_extents; i++) {
> + length = bdrv_getlength(s->extents[i].file->bs);
> + if (length < 0) {
> + return length;
> + }
> + ret = bdrv_truncate(s->extents[i].file, length, PREALLOC_MODE_OFF, NULL);
> + if (ret < 0) {
> + return ret;
> + }
> + }
> + return 0;
> + }
> return vmdk_co_pwritev(bs, offset, bytes, qiov, 0);
> }
>
> --
> 2.17.0
>
Fam
yuchenlin