Hi,

On Wed, Mar 15, 2017 at 05:17:46PM +0000, David Woodhouse wrote:

From: Brijesh Singh <redacted>

To support pci resource mapping from userspace, pci_mmap_page_range
implementation must be done for that platform. This support was
broken for arm64.

This patch copies existing implementation from arm to
enable sysfs mmap.

It's not so much "broken" as "not currently supported".

[...]

+#define HAVE_PCI_MMAP
+extern int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
+			       enum pci_mmap_state mmap_state, int write_combine);
+

Per the prior attempt at this [1], we only want to expose the sysfs
interface, and not the legacy procfs interface, and need the two
decoupled [2].

... or has something changed in the mean time, so that this only exposes
the sysfs interface?

[...]

+int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
+			enum pci_mmap_state mmap_state, int write_combine)
+{
+	if (mmap_state == pci_mmap_io)
+		return -EINVAL;
+
+	if (write_combine)
+		vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot);
+	else
+		vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);

... as per Will's comment in [3], the latter of these should use
pgprot_device.

Thanks,
Mark.

[1] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-April/421948.html
[2] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-April/423083.html
[3] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-April/422571.html