[PATCH] powerpc/ps3: Set driver coherent_dma_mask

[Alan Stern]

On Thu, 19 Jul 2018, Geoff Levand wrote:

> Hi Alan,
> 
> On 07/19/2018 07:33 AM, Alan Stern wrote:
> > On Wed, 18 Jul 2018, Geoff Levand wrote:
> > 
> >> diff --git a/drivers/usb/host/ehci-ps3.c b/drivers/usb/host/ehci-ps3.c
> >> index 8c733492d8fe..454d8c624a3f 100644
> >> --- a/drivers/usb/host/ehci-ps3.c
> >> +++ b/drivers/usb/host/ehci-ps3.c
> >> @@ -86,7 +86,7 @@ static int ps3_ehci_probe(struct ps3_system_bus_device *dev)
> >>  	int result;
> >>  	struct usb_hcd *hcd;
> >>  	unsigned int virq;
> >> -	static u64 dummy_mask = DMA_BIT_MASK(32);
> >> +	static u64 dummy_mask;
> >>  
> >>  	if (usb_disabled()) {
> >>  		result = -ENODEV;
> >> @@ -131,7 +131,9 @@ static int ps3_ehci_probe(struct ps3_system_bus_device *dev)
> >>  		goto fail_irq;
> >>  	}
> >>  
> >> -	dev->core.dma_mask = &dummy_mask; /* FIXME: for improper usb code */
> >> +	dummy_mask = DMA_BIT_MASK(32);
> >> +	dev->core.dma_mask = &dummy_mask;
> >> +	dma_set_coherent_mask(&dev->core, dummy_mask);
> > 
> > What is the reason for changing a static initialization to a dynamic
> > one?  As far as I can see, the patch touches four lines of code but the
> > only real difference is addition of a single line (and removal of a 
> > comment).
> 
> I thought it would be better if all the setting was done in
> one place, that's the only reason.

All right; in that case (for the EHCI and OHCI pieces):

Acked-by: Alan Stern <stern at rowland.harvard.edu>

Alan Stern