[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH][RFC][0/2] REF+/REF- optimization
From: |
Frediano Ziglio |
Subject: |
Re: [Qemu-devel] [PATCH][RFC][0/2] REF+/REF- optimization |
Date: |
Thu, 15 Sep 2011 09:24:29 +0200 |
2011/9/14 Kevin Wolf <address@hidden>:
...
>
>>>>> But let's measure the effects first, I suspect that for cluster
>>>>> allocation it doesn't help much because every REF- comes with a REF+.
>>>>>
>>>>
>>>> That's 50% of effort if REF- clusters are far from REF+ :)
>>>
>>> I would expect that the next REF+ allocates exactly the REF- cluster.
>>> But you still have a point, we save the write on REF- and combine it
>>> with the REF+ write.
>>>
>>
>> This is still a TODO for REF+ patch.
>
> Actually, I was talking about the qcow2_cache_entry_mark_dirty_wb() case
> without any other change. You get it automatically then.
>
No, I forgot I already thought about this. REF- cluster get not reused
on normal operation, but reach 0 only during snapshot delete. This
cause if it reach 0 it means that it was 1 but if it was 1 copied flag
would be 1 and so there is no reason to decrement counter (reductio ab
asburdum).
Frediano
Re: [Qemu-devel] [PATCH][RFC][0/2] REF+/REF- optimization, Frediano Ziglio, 2011/09/13