Re: [Qemu-devel] [RFC][PATCH] qemu-img: add support for skipping zeroes in input during convert

[Peter Lieven]

Hi Stefan,

please have a look at 

[PATCHv2 1.8 0/9] qemu-img convert optimizations

This one here is obsolete.

Peter

Am 02.12.2013 um 15:13 schrieb Stefan Hajnoczi <address@hidden>:

> On Fri, Nov 22, 2013 at 03:48:30PM +0100, Peter Lieven wrote:
>> -            /* If the output image is being created as a copy on write 
>> image,
>> -               assume that sectors which are unallocated in the input image
>> -               are present in both the output's and input's base images (no
>> -               need to copy them). */
>> -            if (out_baseimg) {
>> -                ret = bdrv_is_allocated(bs[bs_i], sector_num - bs_offset,
>> -                                        n, &n1);
> 
> int coroutine_fn bdrv_is_allocated(BlockDriverState *bs, int64_t sector_num,
>                                   int nb_sectors, int *pnum)
> {
>    int64_t ret = bdrv_get_block_status(bs, sector_num, nb_sectors, pnum);
>    if (ret < 0) {
>        return ret;
>    }
>    return
>        (ret & BDRV_BLOCK_DATA) ||
>        ((ret & BDRV_BLOCK_ZERO) && !bdrv_has_zero_init(bs));
> }
> 
> bdrv_has_zero_init() returns false when bs->backing_hd != NULL.
> 
>> +            if (out_baseimg || has_zero_init) {
>> +                n = nb_sectors > INT_MAX ? INT_MAX : nb_sectors;
>> +                ret = bdrv_get_block_status(bs[bs_i], sector_num - 
>> bs_offset,
>> +                                            n, &n1);
>>                 if (ret < 0) {
>> -                    error_report("error while reading metadata for sector "
>> -                                 "%" PRId64 ": %s",
>> +                    error_report("error while reading block status of 
>> sector %" PRId64 ": %s",
>>                                  sector_num - bs_offset, strerror(-ret));
>>                     goto out;
>>                 }
>> -                if (!ret) {
>> +                /* If the output image is zero initialized, we are not 
>> working
>> +                 * on a shared base and the input is zero we can skip the 
>> next
>> +                 * n1 bytes */
>> +                if (!out_baseimg && has_zero_init && ret & BDRV_BLOCK_ZERO) 
>> {
>> +                    sector_num += n1;
>> +                    continue;
>> +                }
>> +                /* If the output image is being created as a copy on write 
>> image,
>> +                   assume that sectors which are unallocated in the input 
>> image
>> +                   are present in both the output's and input's base images 
>> (no
>> +                   need to copy them). */
>> +                if (out_baseimg && !(ret & BDRV_BLOCK_DATA)) {
>>                     sector_num += n1;
>>                     continue;
>>                 }
> 
> How are these two if statements different from bdrv_is_allocated()?
> 
> Stefan