USB: ehci: tegra: Align DMA transfers to 32 bytes

The Tegra2 USB controller doesn't properly deal with misaligned DMA
buffers, causing corruption.  This is especially prevalent with USB
network adapters, where skbuff alignment is often in the middle of a
4-byte dword.

To avoid this, allocate a temporary buffer for the DMA if the provided
buffer isn't sufficiently aligned.

Signed-off-by: Robert Morell <rmorell@nvidia.com>
Signed-off-by: Benoit Goby <benoit@android.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

drivers/usb/host/ehci-tegra.c

@@ -23,6 +23,8 @@
 #include <linux/usb/otg.h>
 #include <mach/usb_phy.h>
 
+#define TEGRA_USB_DMA_ALIGN 32
+
 struct tegra_ehci_hcd {
 	struct ehci_hcd *ehci;
 	struct tegra_usb_phy *phy;
@@ -383,6 +385,92 @@ static int tegra_ehci_bus_resume(struct usb_hcd *hcd)
 }
 #endif
 
+struct temp_buffer {
+	void *kmalloc_ptr;
+	void *old_xfer_buffer;
+	u8 data[0];
+};
+
+static void free_temp_buffer(struct urb *urb)
+{
+	enum dma_data_direction dir;
+	struct temp_buffer *temp;
+
+	if (!(urb->transfer_flags & URB_ALIGNED_TEMP_BUFFER))
+		return;
+
+	dir = usb_urb_dir_in(urb) ? DMA_FROM_DEVICE : DMA_TO_DEVICE;
+
+	temp = container_of(urb->transfer_buffer, struct temp_buffer,
+			    data);
+
+	if (dir == DMA_FROM_DEVICE)
+		memcpy(temp->old_xfer_buffer, temp->data,
+		       urb->transfer_buffer_length);
+	urb->transfer_buffer = temp->old_xfer_buffer;
+	kfree(temp->kmalloc_ptr);
+
+	urb->transfer_flags &= ~URB_ALIGNED_TEMP_BUFFER;
+}
+
+static int alloc_temp_buffer(struct urb *urb, gfp_t mem_flags)
+{
+	enum dma_data_direction dir;
+	struct temp_buffer *temp, *kmalloc_ptr;
+	size_t kmalloc_size;
+
+	if (urb->num_sgs || urb->sg ||
+	    urb->transfer_buffer_length == 0 ||
+	    !((uintptr_t)urb->transfer_buffer & (TEGRA_USB_DMA_ALIGN - 1)))
+		return 0;
+
+	dir = usb_urb_dir_in(urb) ? DMA_FROM_DEVICE : DMA_TO_DEVICE;
+
+	/* Allocate a buffer with enough padding for alignment */
+	kmalloc_size = urb->transfer_buffer_length +
+		sizeof(struct temp_buffer) + TEGRA_USB_DMA_ALIGN - 1;
+
+	kmalloc_ptr = kmalloc(kmalloc_size, mem_flags);
+	if (!kmalloc_ptr)
+		return -ENOMEM;
+
+	/* Position our struct temp_buffer such that data is aligned */
+	temp = PTR_ALIGN(kmalloc_ptr + 1, TEGRA_USB_DMA_ALIGN) - 1;
+
+	temp->kmalloc_ptr = kmalloc_ptr;
+	temp->old_xfer_buffer = urb->transfer_buffer;
+	if (dir == DMA_TO_DEVICE)
+		memcpy(temp->data, urb->transfer_buffer,
+		       urb->transfer_buffer_length);
+	urb->transfer_buffer = temp->data;
+
+	urb->transfer_flags |= URB_ALIGNED_TEMP_BUFFER;
+
+	return 0;
+}
+
+static int tegra_ehci_map_urb_for_dma(struct usb_hcd *hcd, struct urb *urb,
+				      gfp_t mem_flags)
+{
+	int ret;
+
+	ret = alloc_temp_buffer(urb, mem_flags);
+	if (ret)
+		return ret;
+
+	ret = usb_hcd_map_urb_for_dma(hcd, urb, mem_flags);
+	if (ret)
+		free_temp_buffer(urb);
+
+	return ret;
+}
+
+static void tegra_ehci_unmap_urb_for_dma(struct usb_hcd *hcd, struct urb *urb)
+{
+	usb_hcd_unmap_urb_for_dma(hcd, urb);
+	free_temp_buffer(urb);
+}
+
 static const struct hc_driver tegra_ehci_hc_driver = {
 	.description		= hcd_name,
 	.product_desc		= "Tegra EHCI Host Controller",
@@ -398,6 +486,8 @@ static const struct hc_driver tegra_ehci_hc_driver = {
 	.shutdown		= tegra_ehci_shutdown,
 	.urb_enqueue		= ehci_urb_enqueue,
 	.urb_dequeue		= ehci_urb_dequeue,
+	.map_urb_for_dma	= tegra_ehci_map_urb_for_dma,
+	.unmap_urb_for_dma	= tegra_ehci_unmap_urb_for_dma,
 	.endpoint_disable	= ehci_endpoint_disable,
 	.endpoint_reset		= ehci_endpoint_reset,
 	.get_frame_number	= ehci_get_frame,