power management routines for NAND driver

Thomas Gleixner tglx at linutronix.de
Thu Aug 11 12:25:18 EDT 2005


On Thu, 2005-08-11 at 19:25 +0400, Vitaly Wool wrote:
> >I have no idea whats the state of those PM functions and why ARM is not
> >using them.
> Well, they are not at all widely used, only one MIPS board and x86 APM 
> code use them.

Well, those functions are marked deprecated. The device model is the
right way to do it. the SA1100 driver got it right apparently.

> >It's discussable, whether suspend is supposed to fail, when a flash
> >operation is in progress or just waits until the operation is finished,
> >which is usually only a matter of a couple of ms.
> >
> What if introduce an additional parameter ('nowait')?

If its necessary and worth the work :)

> I'd prefer probably to use platform_device and device_driver structures 
> and put suspend/resume functions into device_driver. Those functions 
> will first call mtd->suspend/mtd->resume and then shut down/re-enable 
> the clocks.
> (that's more or less how it's done in drivers/mtd/maps/sa1100-flash.c)
> mtd->suspend and mtd->resume need to be implemented in nand_base then.
> Does that sound reasonable?

Sounds like a plan.

tglx






More information about the linux-mtd mailing list