Re: [PATCH] pci-dma: Fix x86 dma_alloc_coherent to fully clear all pages returned

[Tim Chen <tim.c.chen@linux.intel.com>]

On Wed, 2015-02-18 at 15:45 -0500, Alan Stern wrote:
> On Wed, 18 Feb 2015, Tim Chen wrote:
> 
> > > Have you tried doing the experiments I suggested in
> > > 
> > >         http://marc.info/?l=linux-usb&m=142272448620716&w=2
> > > 
> > > to determine where the problem occurs?
> > > 
> > 
> > I was bogged down with other things lately and I haven't got a chance to
> > test that.  But as you said, there's very few places where xhci 
> > call this memory allocation.  So I think the problem has been fairly
> > narrowed down for the XHCI folks.
> 
> I disagree.  _You_ reported the error.  How can you expect other people 
> to figure out where it is with no help from you?
> 
> I looked briefly at the xhci-hcd DMA allocation code.  It does not 
> contain any obvious mistakes.
> 
> Alan Stern
> 

The error and my quick fix is right here.  And xhci still needs to be
fixed properly.  I'll send out the patch below in a proper patch.

Tim

diff --git a/drivers/usb/host/xhci-mem.c b/drivers/usb/host/xhci-mem.c
index 5cb3d7a..39e7196 100644
--- a/drivers/usb/host/xhci-mem.c
+++ b/drivers/usb/host/xhci-mem.c
@@ -1658,7 +1658,7 @@ static int scratchpad_alloc(struct xhci_hcd *xhci,
gfp_t flags)
                goto fail_sp;

        xhci->scratchpad->sp_array = dma_alloc_coherent(dev,
-                                    num_sp * sizeof(u64),
+                                    PAGE_ALIGN(num_sp * sizeof(u64)),
                                     &xhci->scratchpad->sp_dma, flags);
        if (!xhci->scratchpad->sp_array)