> YAMAMOTO Takashi wrote:
> >> YAMAMOTO Takashi wrote:
> >>>>> what you are proposing seems like another kind of open/close to me.
> >>>>> VOP_OPEN and VOP_CLOSE are appropriate ways to handle them.
> >>>>>
> >>>>> YAMAMOTO Takashi
> >>>> can you explain what VOP_OPEN() should we add flags to in case of swap
> >>>> device or raidframe? I'm not sure I follow your logic just yet...
> >>>>
> >>>> -e.
> >>> in the case of swap, the one in swap_on().  i'm not sure about raidframe.
> >>>
> >>> YAMAMOTO Takashi
> >> raidframe can be a problem with your approach I'm afraid.
> > 
> > why it isn't a problem with yours?
> > 
> > YAMAMOTO Takashi
> 
> because to get the 'vp' for the device we don't have to rely on VOP_OPEN
> and can use vfinddev(),

do you mean that raidframe doesn't have to involve VOP_OPEN or d_open?

> or the interface can simply be changed to take a
> different parameter.

which interface are you talking about?

> btw: did you mean adding another parameter, or adding a flag to the
> 'mode'?
> 
> -e.

i prefer another argument, but i don't think it is very important.

YAMAMOTO Takashi