Re: [Qemu-devel] [PATCH v8 10/14] migration: add postcopy migration of dirty bitmaps

[John Snow]

On 11/16/2017 05:24 AM, Vladimir Sementsov-Ogievskiy wrote:
>>> +    /* if a block is zero we need to flush here since the network
>>> +     * bandwidth is now a lot higher than the storage device bandwidth.
>>> +     * thus if we queue zero blocks we slow down the migration. */
>> Can you elaborate on this for me?
> 
> it comes from migration/block.c, as it was a prototype for dirty bitmaps
> migration.
> May be the original thought was to give destination storage more time to
> handle
> write-zeros? It may make sense if it can't do it fast but really writes
> zeros.
> 
>>
>>> +    if (flags & DIRTY_BITMAP_MIG_FLAG_ZEROES) {
>>> +        qemu_fflush(f);
>>> +    } else {
>>> +        qemu_put_be64(f, buf_size);
>>> +        qemu_put_buffer(f, buf, buf_size);
>>> +    }
>>> +
>>> +    g_free(buf);
>>> +} 

Peter, Stefan; do you have any insight on the reason for this blurb's
appearance in migration/block.c? Looks like it showed up in
323004a39d4d8d33c744a5b108f80bfe6402fca3

I'm not sure I understand what the concern was that lead to this, unless
it's simply that without a flush() we can't guarantee progress.