On 10.04.2017 10:42, Kevin Wolf wrote:
> Am 07.04.2017 um 19:10 hat Max Reitz geschrieben:
>> One case I'd be especially interested in are of course 4 kB subclusters
>> for 64 kB clusters (because 4 kB is a usual page size and can be
>> configured to be the block size of a guest device; and because 64 kB
>> simply is the standard cluster size of qcow2 images nowadays[1]...).
> 
> Why should the current default cluster size be an argument for anything?
> 64k is a tradeoff between small COW size and large allocation
> granularity that seemed to work well (and it used to be the maximum
> cluster size originally, so it seemed safer than 128k).
> 
> With subclusters, we have a completely different situation and need to
> find a new default that works best. I'm relatively sure that 64k are too
> small under the new conditions.
> 
> Also, undefined reference: [1] (I was hoping to find a better argument
> there...)

I meant [1] to be basically what you just said.

Well, the thing is that 64 kB is still better than 32 kB. :-P

Max