On 23.10.2015 15:39, Kevin Wolf wrote:
> Am 21.10.2015 um 15:47 hat Max Reitz geschrieben:
>> On 21.10.2015 13:49, Alberto Garcia wrote:
>>> On Mon 19 Oct 2015 05:53:37 PM CEST, Max Reitz wrote:
>>>> And a helper function for that, which directly takes a pointer to the
>>>> BDS to be inserted instead of its node-name (which will be used for
>>>> implementing 'change' using blockdev-insert-medium).
>>>
>>> Shouldn't this update bdrv_states?
>>
>> I hate bdrv_states.
>>
>> Yes, it should. Thanks!
> 
> Once your reimplement blk_set_bs() on top of blk_insert/remove_bs(),
> this logic would replace the code in change_parent_backing_link().
> 
> Of course, I left the list update in block.c for a reason, it's meant to
> be an internal data structure, so your code accessing it from outside
> won't be much nicer. Do we actually still need bdrv_states or can we get
> rid of it in a follow-up series? If so, I wouldn't mind an ugly
> intermediate state.

I do get rid of it in "blockdev: Further BlockBackend work"* (the final
patch of that series).

Max


* http://lists.nongnu.org/archive/html/qemu-block/2015-02/msg00021.html