Re: USB mass storage and ARM cache coherency

[Greg KH <greg@kroah.com>]

On Fri, Jan 29, 2010 at 10:54:34AM -0800, Matthew Dharm wrote:
> On Fri, Jan 29, 2010 at 08:51:47PM +0300, Sergei Shtylyov wrote:
> > Catalin Marinas wrote:
> > 
> > >That would be a more general solution rather than going through each HCD
> > >driver since my understanding is that flush_dcache_page() is only needed
> > >together with the mass storage support.
> > 
> >   Note that DMA capable driver can be doing some transfers in PIO mode 
> > or falling back to PIO mode if DMA mode transfer is unsuccessful (the 
> > musb driver is an example of the latter and if the DMA rewrite patches 
> > will get accepted, it'll do short transfers in PIO mode).
> 
> Given that an HCD can choose, on the fly, if it's using DMA or PIO, the HCD
> driver is the only place to reasonably put any cache-synchronization code.
> 
> That said, what do the other SCSI HCDs do?  I'm guessing the question gets
> kinda muddy there, since the other SCSI HCDs all talk directly to some
> piece of hardware, and thus are responsible for the cache management
> themselves.
> 
> Based on that, one could argue that ub and usb-storage should be doing
> this.
> 
> HOWEVER, I firmly believe that the cache-management functions belong with
> the driver that actually talks to the low-level hardware, as that's the
> only place where you can be 100% certain of what cache operations are
> needed.  After all, I think someone is working on a USB-over-IP transport,
> and trying to manage cache at the usb-storage level in that scenario is
> just silly.
> 
> So, let's put this in the HCD drivers and be done with it.

I agree, that's the place to fix this issue.

thanks,

greg k-h