Re: usb and no bus_dma(9)

[Eric Haszlakiewicz <erh%nimenees.com@localhost>]

On Tue, Mar 15, 2011 at 03:17:17AM +0900, Izumi Tsutsui wrote:
> > > The point is "usb* at slhci?" is usb but it doesn't use DMA,
> > > unlike uhci/ohci/ehci.
> > 
> > That seems like a rather poor reason for causing all usb devices to know 
> > about
> > slhci or the architecture it runs on then.
> 
> What causes all usb devices to know about slhci?

Oops, I misread that patch and thought it was something that
applied to one specific usb device.

> Which lines in the patch are actually unacceptable for you?
> 
> N[EOU]HCI are defined in [eou]hci.h generated by config(1),
> #if N[EOU]HCI conditional is only required in usb_mem.c,
> and only USB host drivers call usb_allocmem() etc. in usb_mem.c
> if neccesary.
> 
> Anyway, it doesn't look slhci.c requires functions in usb_mem.c
> at all so it might be better to add usb_dma attribute
> for [eou]hci that require usb_mem.c, as Manuel suggested.
> 
> Probably it might also be required to refactor
> usb_dma_block_t to allow USB hosts without DMA,
> but I'd like to suggest a simple and acceptable compromise,
> rather than paranoiac and unreachable perfect goals.

Now that I've cleared some of the fog away from by brain, using the
usb_dma attribute (or a USB_NO_BUS_DMA ifdef, if necessary) sounds ok.

Sorry about the confusion.

eric