On Fri, Jul 16, 2010 at 5:17 PM, Kevin Wolf <kw...@redhat.com> wrote:
> bdrv_commit copies the image to its backing file sector by sector, which
> is (surprise!) relatively slow. Let's take a larger buffer and handle more
> sectors at once if possible.
>
> With a 1G qcow2 file, this brought the time bdrv_commit takes down from
> 5:06 min to 1:14 min for me.
Nice performance improvement!
> Signed-off-by: Kevin Wolf <kw...@redhat.com>
> ---
> block.c | 35 ++++++++++++++++++-----------------
> 1 files changed, 18 insertions(+), 17 deletions(-)
>
> diff --git a/block.c b/block.c
> index 65cf4dc..12dbc43 100644
> --- a/block.c
> +++ b/block.c
> @@ -729,9 +729,9 @@ int bdrv_commit(BlockDriverState *bs)
> {
> BlockDriver *drv = bs->drv;
> int64_t i, total_sectors;
> - int n, j, ro, open_flags;
> + int n, ro, open_flags;
> int ret = 0, rw_ret = 0;
> - unsigned char sector[BDRV_SECTOR_SIZE];
> + uint8_t *buf;
> char filename[1024];
> BlockDriverState *bs_rw, *bs_ro;
>
> @@ -774,25 +774,26 @@ int bdrv_commit(BlockDriverState *bs)
> }
>
> total_sectors = bdrv_getlength(bs) >> BDRV_SECTOR_BITS;
> + buf = qemu_malloc(2048 * BDRV_SECTOR_SIZE);
> +
> for (i = 0; i < total_sectors;) {
> - if (drv->bdrv_is_allocated(bs, i, 65536, &n)) {
> - for(j = 0; j < n; j++) {
> - if (bdrv_read(bs, i, sector, 1) != 0) {
> - ret = -EIO;
> - goto ro_cleanup;
> - }
> + if (drv->bdrv_is_allocated(bs, i, 2048, &n)) {
>
> - if (bdrv_write(bs->backing_hd, i, sector, 1) != 0) {
> - ret = -EIO;
> - goto ro_cleanup;
> - }
> - i++;
> - }
> - } else {
> - i += n;
> + if (bdrv_read(bs, i, buf, n) != 0) {
> + ret = -EIO;
> + goto ro_cleanup;
> + }
> +
> + if (bdrv_write(bs->backing_hd, i, buf, n) != 0) {
> + ret = -EIO;
> + goto ro_cleanup;
> + }
> }
> + i += n;
> }
>
> + qemu_free(buf);
Is buf being freed in the error case ro_cleanup label?
Stefan
> +
> if (drv->bdrv_make_empty) {
> ret = drv->bdrv_make_empty(bs);
> bdrv_flush(bs);
> --
> 1.7.1.1
>
>
>
