On Mon, Sep 01, 2014 at 11:32:04AM +0200, Benoît Canet wrote: > The Monday 01 Sep 2014 à 17:26:09 (+0800), Liu Yuan wrote : > > On Mon, Sep 01, 2014 at 10:35:27AM +0200, Benoît Canet wrote: > > > The Monday 01 Sep 2014 à 15:43:11 (+0800), Liu Yuan wrote : > > > > For a fifo read pattern, we only have one running aio > > > > > > >(possible other cases that has less number than num_children in the > > > >future) > > > I have trouble understanding this part of the commit message could you try > > > to clarify it ? > > > > Until this patch, we have two cases, read single or read all. But later > > patch > > allow VMs to continue if some devices are down. So the discrete number 1 > > and N > > becomes a range [1, N], that is possible running requests are from 1 to N. > > Why not > > (In some other cases some children of the num_children BDS could be disabled > reducing the number of requests needed) ? >
Sounds better, I'll take yours, thanks! Yuan