drm/udl: initial UDL driver (v4)udl-v4

Supports fb_defio,
supports KMS dumb interface
supports 24bpp via conversion to 16bpp, hw can do this better.

rebase onto new mode interfaces, fix some whitespace

12 files changed, 2300 insertions, 0 deletions

diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
index 2418429a9836..8de5b39c4aca 100644
--- a/drivers/gpu/drm/Kconfig
+++ b/drivers/gpu/drm/Kconfig
@@ -165,3 +165,11 @@ source "drivers/gpu/drm/vmwgfx/Kconfig"
 
 source "drivers/gpu/drm/gma500/Kconfig"
 
+config DRM_UDL
+	tristate "DisplayLink"
+	depends on DRM && USB
+	select FB_SYS_FILLRECT
+	select FB_SYS_COPYAREA
+	select FB_SYS_IMAGEBLIT
+	help
+	  UDL
diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
index 0cde1b80fdb1..7f2029726db4 100644
--- a/drivers/gpu/drm/Makefile
+++ b/drivers/gpu/drm/Makefile
@@ -37,4 +37,5 @@ obj-$(CONFIG_DRM_VIA)	+=via/
 obj-$(CONFIG_DRM_NOUVEAU) +=nouveau/
 obj-$(CONFIG_DRM_EXYNOS) +=exynos/
 obj-$(CONFIG_DRM_GMA500) += gma500/
+obj-$(CONFIG_DRM_UDL) += udl/
 obj-y			+= i2c/
diff --git a/drivers/gpu/drm/udl/Makefile b/drivers/gpu/drm/udl/Makefile
new file mode 100644
index 000000000000..05c7481bfd40
--- /dev/null
+++ b/drivers/gpu/drm/udl/Makefile
@@ -0,0 +1,6 @@
+
+ccflags-y := -Iinclude/drm
+
+udl-y := udl_drv.o udl_modeset.o udl_connector.o udl_encoder.o udl_main.o udl_fb.o udl_transfer.o udl_gem.o
+
+obj-$(CONFIG_DRM_UDL) := udl.o
diff --git a/drivers/gpu/drm/udl/udl_connector.c b/drivers/gpu/drm/udl/udl_connector.c
new file mode 100644
index 000000000000..4f5a01c2d70c
--- /dev/null
+++ b/drivers/gpu/drm/udl/udl_connector.c
@@ -0,0 +1,127 @@
+#include "drmP.h"
+#include "drm_crtc.h"
+#include "drm_edid.h"
+#include "drm_crtc_helper.h"
+#include "udl_drv.h"
+
+/* dummy connector to just get EDID,
+   all UDL appear to have a DVI-D */
+
+static u8 *udl_get_edid(struct udl_device *udl)
+{
+	u8 *block;
+	char rbuf[3];
+	int ret, i;
+
+	block = kmalloc(EDID_LENGTH, GFP_KERNEL);
+	if (block == NULL)
+		return NULL;
+
+	for (i = 0; i < EDID_LENGTH; i++) {
+		ret = usb_control_msg(udl->udev,
+				      usb_rcvctrlpipe(udl->udev, 0), (0x02),
+				      (0x80 | (0x02 << 5)), i << 8, 0xA1, rbuf, 2,
+				      HZ);
+		if (ret < 1) {
+			DRM_ERROR("Read EDID byte %d failed err %x\n", i, ret);
+			i--;
+			goto error;
+		}
+		block[i] = rbuf[1];
+	}
+
+	return block;
+
+error:
+	kfree(block);
+	return NULL;
+}
+
+static int udl_get_modes(struct drm_connector *connector)
+{
+	struct udl_device *udl = connector->dev->dev_private;
+	struct edid *edid;
+	int ret;
+
+	edid = (struct edid *)udl_get_edid(udl);
+
+	connector->display_info.raw_edid = (char *)edid;
+
+	drm_mode_connector_update_edid_property(connector, edid);
+	ret = drm_add_edid_modes(connector, edid);
+	connector->display_info.raw_edid = NULL;
+	kfree(edid);
+	return ret;
+}
+
+static int udl_mode_valid(struct drm_connector *connector,
+			  struct drm_display_mode *mode)
+{
+	return 0;
+}
+
+static enum drm_connector_status
+udl_detect(struct drm_connector *connector, bool force)
+{
+	return connector_status_connected;
+}
+
+struct drm_encoder *udl_best_single_encoder(struct drm_connector *connector)
+{
+	int enc_id = connector->encoder_ids[0];
+	struct drm_mode_object *obj;
+	struct drm_encoder *encoder;
+
+	obj = drm_mode_object_find(connector->dev, enc_id, DRM_MODE_OBJECT_ENCODER);
+	if (!obj)
+		return NULL;
+	encoder = obj_to_encoder(obj);
+	return encoder;
+}
+
+int udl_connector_set_property(struct drm_connector *connector, struct drm_property *property,
+			       uint64_t val)
+{
+	return 0;
+}
+
+static void udl_connector_destroy(struct drm_connector *connector)
+{
+	drm_sysfs_connector_remove(connector);
+	drm_connector_cleanup(connector);
+	kfree(connector);
+}
+
+struct drm_connector_helper_funcs udl_connector_helper_funcs = {
+	.get_modes = udl_get_modes,
+	.mode_valid = udl_mode_valid,
+	.best_encoder = udl_best_single_encoder,
+};
+
+struct drm_connector_funcs udl_connector_funcs = {
+	.dpms = drm_helper_connector_dpms,
+	.detect = udl_detect,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.destroy = udl_connector_destroy,
+	.set_property = udl_connector_set_property,
+};
+
+int udl_connector_init(struct drm_device *dev, struct drm_encoder *encoder)
+{
+	struct drm_connector *connector;
+
+	connector = kzalloc(sizeof(struct drm_connector), GFP_KERNEL);
+	if (!connector)
+		return -ENOMEM;
+
+	drm_connector_init(dev, connector, &udl_connector_funcs, DRM_MODE_CONNECTOR_DVII);
+	drm_connector_helper_add(connector, &udl_connector_helper_funcs);
+
+	drm_sysfs_connector_add(connector);
+	drm_mode_connector_attach_encoder(connector, encoder);
+
+	drm_connector_attach_property(connector,
+				      dev->mode_config.dirty_info_property,
+				      1);
+	return 0;
+}
diff --git a/drivers/gpu/drm/udl/udl_drv.c b/drivers/gpu/drm/udl/udl_drv.c
new file mode 100644
index 000000000000..bc35398a3146
--- /dev/null
+++ b/drivers/gpu/drm/udl/udl_drv.c
@@ -0,0 +1,85 @@
+#include <linux/module.h>
+#include "drm_usb.h"
+#include "udl_drv.h"
+
+static struct drm_driver driver;
+
+static struct usb_device_id id_table[] = {
+	{.idVendor = 0x17e9, .match_flags = USB_DEVICE_ID_MATCH_VENDOR,},
+	{},
+};
+MODULE_DEVICE_TABLE(usb, id_table);
+
+MODULE_LICENSE("GPL");
+
+static int udl_usb_probe(struct usb_interface *interface,
+			 const struct usb_device_id *id)
+{
+	return drm_get_usb_dev(interface, id, &driver);
+}
+
+static void udl_usb_disconnect(struct usb_interface *interface)
+{
+	struct drm_device *dev = usb_get_intfdata(interface);
+	drm_put_dev(dev);
+}
+
+static struct vm_operations_struct udl_gem_vm_ops = {
+	.fault = udl_gem_fault,
+	.open = drm_gem_vm_open,
+	.close = drm_gem_vm_close,
+};
+
+static const struct file_operations udl_driver_fops = {
+	.owner = THIS_MODULE,
+	.open = drm_open,
+	.mmap = drm_gem_mmap,
+	.poll = drm_poll,
+	.read = drm_read,
+	.unlocked_ioctl	= drm_ioctl,
+	.release = drm_release,
+	.fasync = drm_fasync,
+	.llseek = noop_llseek,
+};
+
+static struct drm_driver driver = {
+	.driver_features = DRIVER_MODESET | DRIVER_GEM,
+	.load = udl_driver_load,
+	.unload = udl_driver_unload,
+
+	/* gem hooks */
+	.gem_init_object = udl_gem_init_object,
+	.gem_free_object = udl_gem_free_object,
+	.gem_vm_ops = &udl_gem_vm_ops,
+
+	.dumb_create = udl_dumb_create,
+	.dumb_map_offset = udl_gem_mmap,
+	.dumb_destroy = udl_dumb_destroy,
+	.fops = &udl_driver_fops,
+	.name = DRIVER_NAME,
+	.desc = DRIVER_DESC,
+	.date = DRIVER_DATE,
+	.major = DRIVER_MAJOR,
+	.minor = DRIVER_MINOR,
+	.patchlevel = DRIVER_PATCHLEVEL,
+};
+
+static struct usb_driver udl_driver = {
+	.name = "udl",
+	.probe = udl_usb_probe,
+	.disconnect = udl_usb_disconnect,
+	.id_table = id_table,
+};
+
+static int __init udl_init(void)
+{
+	return drm_usb_init(&driver, &udl_driver);
+}
+
+static void __exit udl_exit(void)
+{
+	drm_usb_exit(&driver, &udl_driver);
+}
+
+module_init(udl_init);
+module_exit(udl_exit);
diff --git a/drivers/gpu/drm/udl/udl_drv.h b/drivers/gpu/drm/udl/udl_drv.h
new file mode 100644
index 000000000000..2ed1931e1009
--- /dev/null
+++ b/drivers/gpu/drm/udl/udl_drv.h
@@ -0,0 +1,129 @@
+#ifndef UDL_DRV_H
+#define UDL_DRV_H
+
+#include <linux/usb.h>
+
+#define DRIVER_NAME		"udl"
+#define DRIVER_DESC		"DisplayLink"
+#define DRIVER_DATE		"20101215"
+
+#define DRIVER_MAJOR		0
+#define DRIVER_MINOR		0
+#define DRIVER_PATCHLEVEL	0
+
+struct udl_device;
+
+struct urb_node {
+	struct list_head entry;
+	struct udl_device *dev;
+	struct delayed_work release_urb_work;
+	struct urb *urb;
+};
+
+
+struct urb_list {
+	struct list_head list;
+	spinlock_t lock;
+	struct semaphore limit_sem;
+	int available;
+	int count;
+	size_t size;
+};
+
+struct udl_fbdev;
+
+struct udl_device {
+	struct device *dev;
+	struct drm_device *ddev;
+	struct usb_device *udev;
+
+	int sku_pixel_limit;
+
+	struct urb_list urbs;
+	atomic_t lost_pixels; /* 1 = a render op failed. Need screen refresh */
+	bool virtualized;
+
+	struct udl_fbdev *fbdev;
+	char mode_buf[1024];
+	uint32_t mode_buf_len;
+	void *backing_buffer;
+	atomic_t bytes_rendered; /* raw pixel-bytes driver asked to render */
+	atomic_t bytes_identical; /* saved effort with backbuffer comparison */
+	atomic_t bytes_sent; /* to usb, after compression including overhead */
+	atomic_t cpu_kcycles_used; /* transpired during pixel processing */
+};
+
+struct udl_gem_object {
+	struct drm_gem_object base;
+	struct page **pages;
+	void *vmapping;
+};
+
+#define to_udl_bo(x) container_of(x, struct udl_gem_object, base)
+
+struct udl_framebuffer {
+	struct drm_framebuffer base;
+	struct udl_gem_object *obj;
+	bool active_16; /* active on the 16-bit channel */
+};
+
+#define to_udl_fb(x) container_of(x, struct udl_framebuffer, base)
+
+/* modeset */
+int udl_modeset_init(struct drm_device *dev);
+void udl_modeset_cleanup(struct drm_device *dev);
+int udl_connector_init(struct drm_device *dev, struct drm_encoder *encoder);
+
+struct drm_encoder *udl_encoder_init(struct drm_device *dev);
+
+struct urb *udl_get_urb(struct drm_device *dev);
+
+int udl_submit_urb(struct drm_device *dev, struct urb *urb, size_t len);
+void udl_urb_completion(struct urb *urb);
+
+int udl_driver_load(struct drm_device *dev, unsigned long flags);
+int udl_driver_unload(struct drm_device *dev);
+
+int udl_fbdev_init(struct drm_device *dev);
+void udl_fbdev_cleanup(struct drm_device *dev);
+struct drm_framebuffer *
+udl_fb_user_fb_create(struct drm_device *dev,
+		      struct drm_file *file,
+		      struct drm_mode_fb_cmd2 *mode_cmd);
+
+int udl_render_hline(struct drm_device *dev, int bpp, struct urb **urb_ptr,
+		     const char *front, char **urb_buf_ptr,
+		     u32 byte_offset, u32 byte_width,
+		     int *ident_ptr, int *sent_ptr);
+
+int udl_dumb_create(struct drm_file *file_priv,
+		    struct drm_device *dev,
+		    struct drm_mode_create_dumb *args);
+int udl_gem_mmap(struct drm_file *file_priv, struct drm_device *dev,
+		 uint32_t handle, uint64_t *offset);
+int udl_dumb_destroy(struct drm_file *file_priv, struct drm_device *dev,
+		     uint32_t handle);
+
+int udl_gem_init_object(struct drm_gem_object *obj);
+void udl_gem_free_object(struct drm_gem_object *gem_obj);
+struct udl_gem_object *udl_gem_alloc_object(struct drm_device *dev,
+					    size_t size);
+
+int udl_gem_vmap(struct udl_gem_object *obj);
+void udl_gem_vunmap(struct udl_gem_object *obj);
+int udl_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf);
+
+int udl_handle_damage(struct udl_framebuffer *fb, int x, int y,
+		      int width, int height);
+
+#define CMD_WRITE_RAW8   "\xAF\x60" /**< 8 bit raw write command. */
+#define CMD_WRITE_RL8    "\xAF\x61" /**< 8 bit run length command. */
+#define CMD_WRITE_COPY8  "\xAF\x62" /**< 8 bit copy command. */
+#define CMD_WRITE_RLX8   "\xAF\x63" /**< 8 bit extended run length command. */
+
+#define CMD_WRITE_RAW16  "\xAF\x68" /**< 16 bit raw write command. */
+#define CMD_WRITE_RL16   "\xAF\x69" /**< 16 bit run length command. */
+#define CMD_WRITE_COPY16 "\xAF\x6A" /**< 16 bit copy command. */
+#define CMD_WRITE_RLX16  "\xAF\x6B" /**< 16 bit extended run length command. */
+
+#endif
diff --git a/drivers/gpu/drm/udl/udl_encoder.c b/drivers/gpu/drm/udl/udl_encoder.c
new file mode 100644
index 000000000000..4ecdb35889fb
--- /dev/null
+++ b/drivers/gpu/drm/udl/udl_encoder.c
@@ -0,0 +1,68 @@
+#include "drmP.h"
+#include "drm_crtc.h"
+#include "drm_crtc_helper.h"
+#include "udl_drv.h"
+
+/* dummy encoder */
+void udl_enc_destroy(struct drm_encoder *encoder)
+{
+	drm_encoder_cleanup(encoder);
+	kfree(encoder);
+}
+
+static void udl_encoder_disable(struct drm_encoder *encoder)
+{
+}
+
+static bool udl_mode_fixup(struct drm_encoder *encoder,
+			   struct drm_display_mode *mode,
+			   struct drm_display_mode *adjusted_mode)
+{
+	return true;
+}
+
+static void udl_encoder_prepare(struct drm_encoder *encoder)
+{
+}
+
+static void udl_encoder_commit(struct drm_encoder *encoder)
+{
+}
+
+static void udl_encoder_mode_set(struct drm_encoder *encoder,
+				 struct drm_display_mode *mode,
+				 struct drm_display_mode *adjusted_mode)
+{
+}
+
+static void
+udl_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+}
+
+static const struct drm_encoder_helper_funcs udl_helper_funcs = {
+	.dpms = udl_encoder_dpms,
+	.mode_fixup = udl_mode_fixup,
+	.prepare = udl_encoder_prepare,
+	.mode_set = udl_encoder_mode_set,
+	.commit = udl_encoder_commit,
+	.disable = udl_encoder_disable,
+};
+
+static const struct drm_encoder_funcs udl_enc_funcs = {
+	.destroy = udl_enc_destroy,
+};
+
+struct drm_encoder *udl_encoder_init(struct drm_device *dev)
+{
+	struct drm_encoder *encoder;
+
+	encoder = kzalloc(sizeof(struct drm_encoder), GFP_KERNEL);
+	if (!encoder)
+		return NULL;
+
+	drm_encoder_init(dev, encoder, &udl_enc_funcs, DRM_MODE_ENCODER_TMDS);
+	drm_encoder_helper_add(encoder, &udl_helper_funcs);
+	encoder->possible_crtcs = 1;
+	return encoder;
+}
diff --git a/drivers/gpu/drm/udl/udl_fb.c b/drivers/gpu/drm/udl/udl_fb.c
new file mode 100644
index 000000000000..6b63289305bd
--- /dev/null
+++ b/drivers/gpu/drm/udl/udl_fb.c
@@ -0,0 +1,622 @@
+/*
+ * Copyright © 2010 David Airlie
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *     David Airlie
+ */
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/fb.h>
+
+#include "drmP.h"
+#include "drm.h"
+#include "drm_crtc.h"
+#include "drm_crtc_helper.h"
+#include "udl_drv.h"
+
+#include "drm_fb_helper.h"
+
+#define DL_DEFIO_WRITE_DELAY    5 /* fb_deferred_io.delay in jiffies */
+
+static int fb_defio = 1;  /* Optionally enable experimental fb_defio mmap support */
+static int fb_bpp = 16;
+
+module_param(fb_bpp, int, S_IWUSR | S_IRUSR | S_IWGRP | S_IRGRP);
+module_param(fb_defio, int, S_IWUSR | S_IRUSR | S_IWGRP | S_IRGRP);
+
+struct udl_fbdev {
+	struct drm_fb_helper helper;
+	struct udl_framebuffer ufb;
+	struct list_head fbdev_list;
+	int fb_count;
+};
+
+#define DL_ALIGN_UP(x, a) ALIGN(x, a)
+#define DL_ALIGN_DOWN(x, a) ALIGN(x-(a-1), a)
+
+/** Read the red component (0..255) of a 32 bpp colour. */
+#define DLO_RGB_GETRED(col) (uint8_t)((col) & 0xFF)
+
+/** Read the green component (0..255) of a 32 bpp colour. */
+#define DLO_RGB_GETGRN(col) (uint8_t)(((col) >> 8) & 0xFF)
+
+/** Read the blue component (0..255) of a 32 bpp colour. */
+#define DLO_RGB_GETBLU(col) (uint8_t)(((col) >> 16) & 0xFF)
+
+/** Return red/green component of a 16 bpp colour number. */
+#define DLO_RG16(red, grn) (uint8_t)((((red) & 0xF8) | ((grn) >> 5)) & 0xFF)
+
+/** Return green/blue component of a 16 bpp colour number. */
+#define DLO_GB16(grn, blu) (uint8_t)(((((grn) & 0x1C) << 3) | ((blu) >> 3)) & 0xFF)
+
+/** Return 8 bpp colour number from red, green and blue components. */
+#define DLO_RGB8(red, grn, blu) ((((red) << 5) | (((grn) & 3) << 3) | ((blu) & 7)) & 0xFF)
+
+static uint8_t rgb8(uint32_t col)
+{
+	uint8_t red = DLO_RGB_GETRED(col);
+	uint8_t grn = DLO_RGB_GETGRN(col);
+	uint8_t blu = DLO_RGB_GETBLU(col);
+
+	return DLO_RGB8(red, grn, blu);
+}
+
+static uint16_t rgb16(uint32_t col)
+{
+	uint8_t red = DLO_RGB_GETRED(col);
+	uint8_t grn = DLO_RGB_GETGRN(col);
+	uint8_t blu = DLO_RGB_GETBLU(col);
+
+	return (DLO_RG16(red, grn) << 8) + DLO_GB16(grn, blu);
+}
+
+
+/*
+ * NOTE: fb_defio.c is holding info->fbdefio.mutex
+ *   Touching ANY framebuffer memory that triggers a page fault
+ *   in fb_defio will cause a deadlock, when it also tries to
+ *   grab the same mutex.
+ */
+static void udlfb_dpy_deferred_io(struct fb_info *info,
+				  struct list_head *pagelist)
+{
+	struct page *cur;
+	struct fb_deferred_io *fbdefio = info->fbdefio;
+	struct udl_fbdev *ufbdev = info->par;
+	struct drm_device *dev = ufbdev->ufb.base.dev;
+	struct udl_device *udl = dev->dev_private;
+	struct urb *urb;
+	char *cmd;
+	cycles_t start_cycles, end_cycles;
+	int bytes_sent = 0;
+	int bytes_identical = 0;
+	int bytes_rendered = 0;
+
+	if (!fb_defio)
+		return;
+
+	/* if (!atomic_read(&dev->usb_active))
+	   return;*/
+
+	start_cycles = get_cycles();
+
+	urb = udl_get_urb(dev);
+	if (!urb)
+		return;
+
+	cmd = urb->transfer_buffer;
+
+	/* walk the written page list and render each to device */
+	list_for_each_entry(cur, &fbdefio->pagelist, lru) {
+
+		if (udl_render_hline(dev, (ufbdev->ufb.base.bits_per_pixel / 8),
+				  &urb, (char *) info->fix.smem_start,
+				  &cmd, cur->index << PAGE_SHIFT,
+				  PAGE_SIZE, &bytes_identical, &bytes_sent))
+			goto error;
+		bytes_rendered += PAGE_SIZE;
+	}
+
+	if (cmd > (char *) urb->transfer_buffer) {
+		/* Send partial buffer remaining before exiting */
+		int len = cmd - (char *) urb->transfer_buffer;
+		udl_submit_urb(dev, urb, len);
+		bytes_sent += len;
+	} else
+		udl_urb_completion(urb);
+
+error:
+	atomic_add(bytes_sent, &udl->bytes_sent);
+	atomic_add(bytes_identical, &udl->bytes_identical);
+	atomic_add(bytes_rendered, &udl->bytes_rendered);
+	end_cycles = get_cycles();
+	atomic_add(((unsigned int) ((end_cycles - start_cycles)
+		    >> 10)), /* Kcycles */
+		   &udl->cpu_kcycles_used);
+}
+
+int udl_handle_damage(struct udl_framebuffer *fb, int x, int y,
+		      int width, int height)
+{
+	struct drm_device *dev = fb->base.dev;
+	struct udl_device *udl = dev->dev_private;
+	int i, ret;
+	char *cmd;
+	cycles_t start_cycles, end_cycles;
+	int bytes_sent = 0;
+	int bytes_identical = 0;
+	struct urb *urb;
+	int aligned_x;
+	int bpp = (fb->base.bits_per_pixel / 8);
+
+	if (!fb->active_16)
+		return 0;
+
+	if (!fb->obj->vmapping)
+		udl_gem_vmap(fb->obj);
+
+	start_cycles = get_cycles();
+
+	aligned_x = DL_ALIGN_DOWN(x, sizeof(unsigned long));
+	width = DL_ALIGN_UP(width + (x-aligned_x), sizeof(unsigned long));
+	x = aligned_x;
+
+	if ((width <= 0) ||
+	    (x + width > fb->base.width) ||
+	    (y + height > fb->base.height))
+		return -EINVAL;
+
+//	if (!atomic_read(&dev->usb_active))
+//		return 0;
+
+	urb = udl_get_urb(dev);
+	if (!urb)
+		return 0;
+	cmd = urb->transfer_buffer;
+
+	for (i = y; i < y + height ; i++) {
+		const int line_offset = fb->base.pitches[0] * i;
+		const int byte_offset = line_offset + (x * bpp);
+
+		if (udl_render_hline(dev, bpp, &urb,
+				     (char *) fb->obj->vmapping,
+				     &cmd, byte_offset, width * bpp,
+				     &bytes_identical, &bytes_sent))
+			goto error;
+	}
+
+	if (cmd > (char *) urb->transfer_buffer) {
+		/* Send partial buffer remaining before exiting */
+		int len = cmd - (char *) urb->transfer_buffer;
+		ret = udl_submit_urb(dev, urb, len);
+		bytes_sent += len;
+	} else
+		udl_urb_completion(urb);
+
+error:
+	atomic_add(bytes_sent, &udl->bytes_sent);
+	atomic_add(bytes_identical, &udl->bytes_identical);
+	atomic_add(width*height*bpp, &udl->bytes_rendered);
+	end_cycles = get_cycles();
+	atomic_add(((unsigned int) ((end_cycles - start_cycles)
+		    >> 10)), /* Kcycles */
+		   &udl->cpu_kcycles_used);
+
+	return 0;
+}
+
+static int udl_fb_mmap(struct fb_info *info, struct vm_area_struct *vma)
+{
+	unsigned long start = vma->vm_start;
+	unsigned long size = vma->vm_end - vma->vm_start;
+	unsigned long offset = vma->vm_pgoff << PAGE_SHIFT;
+	unsigned long page, pos;
+
+	if (offset + size > info->fix.smem_len)
+		return -EINVAL;
+
+	pos = (unsigned long)info->fix.smem_start + offset;
+
+	pr_notice("mmap() framebuffer addr:%lu size:%lu\n",
+		  pos, size);
+
+	while (size > 0) {
+		page = vmalloc_to_pfn((void *)pos);
+		if (remap_pfn_range(vma, start, page, PAGE_SIZE, PAGE_SHARED))
+			return -EAGAIN;
+
+		start += PAGE_SIZE;
+		pos += PAGE_SIZE;
+		if (size > PAGE_SIZE)
+			size -= PAGE_SIZE;
+		else
+			size = 0;
+	}
+
+	vma->vm_flags |= VM_RESERVED;	/* avoid to swap out this VMA */
+	return 0;
+}
+
+static void udl_fb_fillrect(struct fb_info *info, const struct fb_fillrect *rect)
+{
+	struct udl_fbdev *ufbdev = info->par;
+
+	sys_fillrect(info, rect);
+
+	udl_handle_damage(&ufbdev->ufb, rect->dx, rect->dy, rect->width,
+			  rect->height);
+}
+
+static void udl_fb_copyarea(struct fb_info *info, const struct fb_copyarea *region)
+{
+	struct udl_fbdev *ufbdev = info->par;
+
+	sys_copyarea(info, region);
+
+	udl_handle_damage(&ufbdev->ufb, region->dx, region->dy, region->width,
+			  region->height);
+}
+
+static void udl_fb_imageblit(struct fb_info *info, const struct fb_image *image)
+{
+	struct udl_fbdev *ufbdev = info->par;
+
+	sys_imageblit(info, image);
+
+	udl_handle_damage(&ufbdev->ufb, image->dx, image->dy, image->width,
+			  image->height);
+}
+
+/*
+ * It's common for several clients to have framebuffer open simultaneously.
+ * e.g. both fbcon and X. Makes things interesting.
+ * Assumes caller is holding info->lock (for open and release at least)
+ */
+static int udl_fb_open(struct fb_info *info, int user)
+{
+	struct udl_fbdev *ufbdev = info->par;
+	struct drm_device *dev = ufbdev->ufb.base.dev;
+	struct udl_device *udl = dev->dev_private;
+
+	/*
+	 * fbcon aggressively connects to first framebuffer it finds,
+	 * preventing other clients (X) from working properly. Usually
+	 * not what the user wants. Fail by default with option to enable.
+	 */
+	//	if ((user == 0) & (!console))
+	//		return -EBUSY;
+
+	/* If the USB device is gone, we don't accept new opens */
+	if (udl->virtualized)
+		return -ENODEV;
+
+	ufbdev->fb_count++;
+
+	if (fb_defio && (info->fbdefio == NULL)) {
+		/* enable defio at last moment if not disabled by client */
+
+		struct fb_deferred_io *fbdefio;
+
+		fbdefio = kmalloc(sizeof(struct fb_deferred_io), GFP_KERNEL);
+
+		if (fbdefio) {
+			fbdefio->delay = DL_DEFIO_WRITE_DELAY;
+			fbdefio->deferred_io = udlfb_dpy_deferred_io;
+		}
+
+		info->fbdefio = fbdefio;
+		fb_deferred_io_init(info);
+	}
+
+	pr_notice("open /dev/fb%d user=%d fb_info=%p count=%d\n",
+		  info->node, user, info, ufbdev->fb_count);
+
+	return 0;
+}
+
+
+/*
+ * Assumes caller is holding info->lock mutex (for open and release at least)
+ */
+static int udl_fb_release(struct fb_info *info, int user)
+{
+	struct udl_fbdev *ufbdev = info->par;
+
+	ufbdev->fb_count--;
+
+	if ((ufbdev->fb_count == 0) && (info->fbdefio)) {
+		fb_deferred_io_cleanup(info);
+		kfree(info->fbdefio);
+		info->fbdefio = NULL;
+		info->fbops->fb_mmap = udl_fb_mmap;
+	}
+
+	pr_warn("released /dev/fb%d user=%d count=%d\n",
+		info->node, user, ufbdev->fb_count);
+
+	return 0;
+}
+
+static struct fb_ops udlfb_ops = {
+	.owner = THIS_MODULE,
+	.fb_check_var = drm_fb_helper_check_var,
+	.fb_set_par = drm_fb_helper_set_par,
+	.fb_fillrect = udl_fb_fillrect,
+	.fb_copyarea = udl_fb_copyarea,
+	.fb_imageblit = udl_fb_imageblit,
+	.fb_pan_display = drm_fb_helper_pan_display,
+	.fb_blank = drm_fb_helper_blank,
+	.fb_setcmap = drm_fb_helper_setcmap,
+	.fb_debug_enter = drm_fb_helper_debug_enter,
+	.fb_debug_leave = drm_fb_helper_debug_leave,
+	.fb_mmap = udl_fb_mmap,
+	.fb_open = udl_fb_open,
+	.fb_release = udl_fb_release,
+};
+
+void udl_crtc_fb_gamma_set(struct drm_crtc *crtc, u16 red, u16 green,
+			   u16 blue, int regno)
+{
+}
+
+void udl_crtc_fb_gamma_get(struct drm_crtc *crtc, u16 *red, u16 *green,
+			     u16 *blue, int regno)
+{
+	*red = 0;
+	*green = 0;
+	*blue = 0;
+}
+
+static int udl_user_framebuffer_dirty(struct drm_framebuffer *fb,
+				      struct drm_file *file,
+				      unsigned flags, unsigned color,
+				      struct drm_clip_rect *clips,
+				      unsigned num_clips)
+{
+	struct udl_framebuffer *ufb = to_udl_fb(fb);
+	int i;
+
+	if (!ufb->active_16)
+		return 0;
+
+	for (i = 0; i < num_clips; i++) {
+		udl_handle_damage(ufb, clips[i].x1, clips[i].y1,
+				  clips[i].x2 - clips[i].x1,
+				  clips[i].y2 - clips[i].y1);
+	}
+	return 0;
+}
+
+static void udl_user_framebuffer_destroy(struct drm_framebuffer *fb)
+{
+	struct udl_framebuffer *ufb = to_udl_fb(fb);
+
+	if (ufb->obj)
+		drm_gem_object_unreference_unlocked(&ufb->obj->base);
+
+	drm_framebuffer_cleanup(fb);
+	kfree(ufb);
+}
+
+static const struct drm_framebuffer_funcs udlfb_funcs = {
+	.destroy = udl_user_framebuffer_destroy,
+	.dirty = udl_user_framebuffer_dirty,
+	.create_handle = NULL,
+};
+
+
+static int
+udl_framebuffer_init(struct drm_device *dev,
+		     struct udl_framebuffer *ufb,
+		     struct drm_mode_fb_cmd2 *mode_cmd,
+		     struct udl_gem_object *obj)
+{
+	int ret;
+
+	ufb->obj = obj;
+	ret = drm_framebuffer_init(dev, &ufb->base, &udlfb_funcs);
+	drm_helper_mode_fill_fb_struct(&ufb->base, mode_cmd);
+	return ret;
+}
+
+
+static int udlfb_create(struct udl_fbdev *ufbdev,
+			struct drm_fb_helper_surface_size *sizes)
+{
+	struct drm_device *dev = ufbdev->helper.dev;
+	struct fb_info *info;
+	struct device *device = &dev->usbdev->dev;
+	struct drm_framebuffer *fb;
+	struct drm_mode_fb_cmd2 mode_cmd;
+	struct udl_gem_object *obj;
+	uint32_t size;
+	int ret = 0;
+
+	if (sizes->surface_bpp == 24)
+		sizes->surface_bpp = 32;
+
+	mode_cmd.width = sizes->surface_width;
+	mode_cmd.height = sizes->surface_height;
+	mode_cmd.pitches[0] = mode_cmd.width * ((sizes->surface_bpp + 7) / 8);
+
+	mode_cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp,
+							  sizes->surface_depth);
+
+	size = mode_cmd.pitches[0] * mode_cmd.height;
+	size = ALIGN(size, PAGE_SIZE);
+
+	obj = udl_gem_alloc_object(dev, size);
+	if (!obj)
+		goto out;
+
+	ret = udl_gem_vmap(obj);
+	if (ret) {
+		DRM_ERROR("failed to vmap fb\n");
+		goto out_gfree;
+	}
+
+	info = framebuffer_alloc(0, device);
+	if (!info) {
+		ret = -ENOMEM;
+		goto out_gfree;
+	}
+	info->par = ufbdev;
+
+	ret = udl_framebuffer_init(dev, &ufbdev->ufb, &mode_cmd, obj);
+	if (ret)
+		goto out_gfree;
+
+	fb = &ufbdev->ufb.base;
+
+	ufbdev->helper.fb = fb;
+	ufbdev->helper.fbdev = info;
+
+	strcpy(info->fix.id, "udldrmfb");
+
+	info->screen_base = ufbdev->ufb.obj->vmapping;
+	info->fix.smem_len = size;
+	info->fix.smem_start = (unsigned long)ufbdev->ufb.obj->vmapping;
+
+	info->flags = FBINFO_DEFAULT | FBINFO_CAN_FORCE_OUTPUT;
+	info->fbops = &udlfb_ops;
+	drm_fb_helper_fill_fix(info, fb->pitches[0], fb->depth);
+	drm_fb_helper_fill_var(info, &ufbdev->helper, sizes->fb_width, sizes->fb_height);
+
+	ret = fb_alloc_cmap(&info->cmap, 256, 0);
+	if (ret) {
+		ret = -ENOMEM;
+		goto out_gfree;
+	}
+
+
+	DRM_DEBUG_KMS("allocated %dx%d vmal %p\n",
+		      fb->width, fb->height,
+		      ufbdev->ufb.obj->vmapping);
+
+	return ret;
+out_gfree:
+	drm_gem_object_unreference(&ufbdev->ufb.obj->base);
+out:
+	return ret;
+}
+
+static int udl_fb_find_or_create_single(struct drm_fb_helper *helper,
+					struct drm_fb_helper_surface_size *sizes)
+{
+	struct udl_fbdev *ufbdev = (struct udl_fbdev *)helper;
+	int new_fb = 0;
+	int ret;
+
+	if (!helper->fb) {
+		ret = udlfb_create(ufbdev, sizes);
+		if (ret)
+			return ret;
+
+		new_fb = 1;
+	}
+	return new_fb;
+}
+
+static struct drm_fb_helper_funcs udl_fb_helper_funcs = {
+	.gamma_set = udl_crtc_fb_gamma_set,
+	.gamma_get = udl_crtc_fb_gamma_get,
+	.fb_probe = udl_fb_find_or_create_single,
+};
+
+static void udl_fbdev_destroy(struct drm_device *dev,
+			      struct udl_fbdev *ufbdev)
+{
+	struct fb_info *info;
+	if (ufbdev->helper.fbdev) {
+		info = ufbdev->helper.fbdev;
+		unregister_framebuffer(info);
+		if (info->cmap.len)
+			fb_dealloc_cmap(&info->cmap);
+		framebuffer_release(info);
+	}
+	drm_fb_helper_fini(&ufbdev->helper);
+	drm_framebuffer_cleanup(&ufbdev->ufb.base);
+	drm_gem_object_unreference_unlocked(&ufbdev->ufb.obj->base);
+}
+
+int udl_fbdev_init(struct drm_device *dev)
+{
+	struct udl_device *udl = dev->dev_private;
+	int bpp_sel = fb_bpp;
+	struct udl_fbdev *ufbdev;
+	int ret;
+
+	ufbdev = kzalloc(sizeof(struct udl_fbdev), GFP_KERNEL);
+	if (!ufbdev)
+		return -ENOMEM;
+
+	udl->fbdev = ufbdev;
+	ufbdev->helper.funcs = &udl_fb_helper_funcs;
+
+	ret = drm_fb_helper_init(dev, &ufbdev->helper,
+				 1, 1);
+	if (ret) {
+		kfree(ufbdev);
+		return ret;
+
+	}
+
+	drm_fb_helper_single_add_all_connectors(&ufbdev->helper);
+	drm_fb_helper_initial_config(&ufbdev->helper, bpp_sel);
+	return 0;
+}
+
+void udl_fbdev_cleanup(struct drm_device *dev)
+{
+	struct udl_device *udl = dev->dev_private;
+	if (!udl->fbdev)
+		return;
+
+	udl_fbdev_destroy(dev, udl->fbdev);
+	kfree(udl->fbdev);
+	udl->fbdev = NULL;
+}
+
+struct drm_framebuffer *
+udl_fb_user_fb_create(struct drm_device *dev,
+		   struct drm_file *file,
+		   struct drm_mode_fb_cmd2 *mode_cmd)
+{
+	struct drm_gem_object *obj;
+	struct udl_framebuffer *ufb;
+	int ret;
+
+	obj = drm_gem_object_lookup(dev, file, mode_cmd->handles[0]);
+	if (obj == NULL)
+		return ERR_PTR(-ENOENT);
+
+	ufb = kzalloc(sizeof(*ufb), GFP_KERNEL);
+	if (ufb == NULL)
+		return ERR_PTR(-ENOMEM);
+
+	ret = udl_framebuffer_init(dev, ufb, mode_cmd, to_udl_bo(obj));
+	if (ret) {
+		kfree(ufb);
+		return ERR_PTR(-EINVAL);
+	}
+	return &ufb->base;
+}
diff --git a/drivers/gpu/drm/udl/udl_gem.c b/drivers/gpu/drm/udl/udl_gem.c
new file mode 100644
index 000000000000..78dd7144a3b1
--- /dev/null
+++ b/drivers/gpu/drm/udl/udl_gem.c
@@ -0,0 +1,274 @@
+#include "drmP.h"
+#include "udl_drv.h"
+#include <linux/shmem_fs.h>
+
+struct udl_gem_object *udl_gem_alloc_object(struct drm_device *dev,
+					    size_t size)
+{
+	struct udl_gem_object *obj;
+
+	obj = kzalloc(sizeof(*obj), GFP_KERNEL);
+	if (obj == NULL)
+		return NULL;
+
+	if (drm_gem_object_init(dev, &obj->base, size) != 0) {
+		kfree(obj);
+		return NULL;
+	}
+
+	return obj;
+}
+
+static int
+udl_gem_create(struct drm_file *file,
+	       struct drm_device *dev,
+	       uint64_t size,
+	       uint32_t *handle_p)
+{
+	struct udl_gem_object *obj;
+	int ret;
+	u32 handle;
+
+	size = roundup(size, PAGE_SIZE);
+
+	obj = udl_gem_alloc_object(dev, size);
+	if (obj == NULL)
+		return -ENOMEM;
+
+	ret = drm_gem_handle_create(file, &obj->base, &handle);
+	if (ret) {
+		drm_gem_object_release(&obj->base);
+		kfree(obj);
+		return ret;
+	}
+
+	drm_gem_object_unreference(&obj->base);
+	*handle_p = handle;
+	return 0;
+}
+
+int udl_dumb_create(struct drm_file *file,
+		    struct drm_device *dev,
+		    struct drm_mode_create_dumb *args)
+{
+	args->pitch = args->width * ((args->bpp + 1) / 8);
+	args->size = args->pitch * args->height;
+	return udl_gem_create(file, dev,
+			      args->size, &args->handle);
+}
+
+int udl_dumb_destroy(struct drm_file *file, struct drm_device *dev,
+		     uint32_t handle)
+{
+	return drm_gem_handle_delete(file, handle);
+}
+
+static int
+udl_gem_create_mmap_offset(struct udl_gem_object *obj)
+{
+	struct drm_device *dev = obj->base.dev;
+	struct drm_gem_mm *mm = dev->mm_private;
+	struct drm_map_list *list;
+	struct drm_local_map *map;
+	int ret = 0;
+
+	/* Set the object up for mmap'ing */
+	list = &obj->base.map_list;
+	list->map = kzalloc(sizeof(struct drm_map_list), GFP_KERNEL);
+	if (!list->map)
+		return -ENOMEM;
+
+	map = list->map;
+	map->type = _DRM_GEM;
+	map->size = obj->base.size;
+	map->handle = obj;
+
+	/* Get a DRM GEM mmap offset allocated... */
+	list->file_offset_node = drm_mm_search_free(&mm->offset_manager,
+						    obj->base.size / PAGE_SIZE,
+						    0, 0);
+	if (!list->file_offset_node) {
+		DRM_ERROR("failed to allocate offset for bo %d\n",
+			  obj->base.name);
+		ret = -ENOSPC;
+		goto out_free_list;
+	}
+
+	list->file_offset_node = drm_mm_get_block(list->file_offset_node,
+						  obj->base.size / PAGE_SIZE,
+						  0);
+	if (!list->file_offset_node) {
+		ret = -ENOMEM;
+		goto out_free_list;
+	}
+
+	list->hash.key = list->file_offset_node->start;
+	ret = drm_ht_insert_item(&mm->offset_hash, &list->hash);
+	if (ret) {
+		DRM_ERROR("failed to add to map hash\n");
+		goto out_free_mm;
+	}
+
+	return 0;
+
+out_free_mm:
+	drm_mm_put_block(list->file_offset_node);
+out_free_list:
+	kfree(list->map);
+	list->map = NULL;
+
+	return ret;
+}
+
+
+int udl_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
+{
+	struct udl_gem_object *obj = to_udl_bo(vma->vm_private_data);
+	struct page *page;
+	unsigned int page_offset;
+	int ret = 0;
+
+	page_offset = ((unsigned long)vmf->virtual_address - vma->vm_start) >>
+		PAGE_SHIFT;
+
+	if (!obj->pages)
+		return VM_FAULT_SIGBUS;
+
+	page = obj->pages[page_offset];
+	ret = vm_insert_page(vma, (unsigned long)vmf->virtual_address, page);
+	switch (ret) {
+	case -EAGAIN:
+		set_need_resched();
+	case 0:
+	case -ERESTARTSYS:
+		return VM_FAULT_NOPAGE;
+	case -ENOMEM:
+		return VM_FAULT_OOM;
+	default:
+		return VM_FAULT_SIGBUS;
+	}
+}
+
+int udl_gem_init_object(struct drm_gem_object *obj)
+{
+	BUG();
+
+	return 0;
+}
+
+static int udl_gem_get_pages(struct udl_gem_object *obj, gfp_t gfpmask)
+{
+	int page_count, i;
+	struct page *page;
+	struct inode *inode;
+	struct address_space *mapping;
+
+	if (obj->pages)
+		return 0;
+
+	page_count = obj->base.size / PAGE_SIZE;
+	BUG_ON(obj->pages != NULL);
+	obj->pages = drm_malloc_ab(page_count, sizeof(struct page *));
+	if (obj->pages == NULL)
+		return -ENOMEM;
+
+	inode = obj->base.filp->f_path.dentry->d_inode;
+	mapping = inode->i_mapping;
+	gfpmask |= mapping_gfp_mask(mapping);
+
+	for (i = 0; i < page_count; i++) {
+		page = shmem_read_mapping_page_gfp(mapping, i, gfpmask);
+		if (IS_ERR(page))
+			goto err_pages;
+		obj->pages[i] = page;
+	}
+
+	return 0;
+err_pages:
+	while (i--)
+		page_cache_release(obj->pages[i]);
+	drm_free_large(obj->pages);
+	obj->pages = NULL;
+	return PTR_ERR(page);
+}
+
+static void udl_gem_put_pages(struct udl_gem_object *obj)
+{
+	int page_count = obj->base.size / PAGE_SIZE;
+	int i;
+
+	for (i = 0; i < page_count; i++)
+		page_cache_release(obj->pages[i]);
+
+	drm_free_large(obj->pages);
+	obj->pages = NULL;
+}
+
+int udl_gem_vmap(struct udl_gem_object *obj)
+{
+	int page_count = obj->base.size / PAGE_SIZE;
+	int ret;
+
+	ret = udl_gem_get_pages(obj, GFP_KERNEL);
+	if (ret)
+		return ret;
+
+	obj->vmapping = vmap(obj->pages, page_count, 0, PAGE_KERNEL);
+	if (!obj->vmapping)
+		return -ENOMEM;
+	return 0;
+}
+
+void udl_gem_vunmap(struct udl_gem_object *obj)
+{
+	if (obj->vmapping)
+		vunmap(obj->vmapping);
+
+	udl_gem_put_pages(obj);
+}
+
+void udl_gem_free_object(struct drm_gem_object *gem_obj)
+{
+	struct udl_gem_object *obj = to_udl_bo(gem_obj);
+
+	if (obj->vmapping)
+		udl_gem_vunmap(obj);
+
+	if (obj->pages)
+		udl_gem_put_pages(obj);
+}
+
+/* the dumb interface doesn't work with the GEM straight MMAP
+   interface, it expects to do MMAP on the drm fd, like normal */
+int udl_gem_mmap(struct drm_file *file, struct drm_device *dev,
+		 uint32_t handle, uint64_t *offset)
+{
+	struct udl_gem_object *gobj;
+	struct drm_gem_object *obj;
+	int ret = 0;
+
+	mutex_lock(&dev->struct_mutex);
+	obj = drm_gem_object_lookup(dev, file, handle);
+	if (obj == NULL) {
+		ret = -ENOENT;
+		goto unlock;
+	}
+	gobj = to_udl_bo(obj);
+
+	ret = udl_gem_get_pages(gobj, GFP_KERNEL);
+	if (ret)
+		return ret;
+	if (!gobj->base.map_list.map) {
+		ret = udl_gem_create_mmap_offset(gobj);
+		if (ret)
+			goto out;
+	}
+
+	*offset = (u64)gobj->base.map_list.hash.key << PAGE_SHIFT;
+
+out:
+	drm_gem_object_unreference(&gobj->base);
+unlock:
+	mutex_unlock(&dev->struct_mutex);
+	return ret;
+}
diff --git a/drivers/gpu/drm/udl/udl_main.c b/drivers/gpu/drm/udl/udl_main.c
new file mode 100644
index 000000000000..ae0ab74a08da
--- /dev/null
+++ b/drivers/gpu/drm/udl/udl_main.c
@@ -0,0 +1,320 @@
+#include "drmP.h"
+#include "udl_drv.h"
+
+/* -BULK_SIZE as per usb-skeleton. Can we get full page and avoid overhead? */
+#define BULK_SIZE 512
+
+#define MAX_TRANSFER (PAGE_SIZE*16 - BULK_SIZE)
+#define WRITES_IN_FLIGHT (4)
+#define MAX_VENDOR_DESCRIPTOR_SIZE 256
+
+#define GET_URB_TIMEOUT	HZ
+#define FREE_URB_TIMEOUT (HZ*2)
+
+static int udl_parse_vendor_descriptor(struct drm_device *dev,
+				       struct usb_device *usbdev)
+{
+	struct udl_device *udl = dev->dev_private;
+	char *desc;
+	char *buf;
+	char *desc_end;
+
+	u8 total_len = 0;
+
+	buf = kzalloc(MAX_VENDOR_DESCRIPTOR_SIZE, GFP_KERNEL);
+	if (!buf)
+		return false;
+	desc = buf;
+
+	total_len = usb_get_descriptor(usbdev, 0x5f, /* vendor specific */
+				    0, desc, MAX_VENDOR_DESCRIPTOR_SIZE);
+	if (total_len > 5) {
+		DRM_INFO("vendor descriptor length:%x data:%02x %02x %02x %02x" \
+			"%02x %02x %02x %02x %02x %02x %02x\n",
+			total_len, desc[0],
+			desc[1], desc[2], desc[3], desc[4], desc[5], desc[6],
+			desc[7], desc[8], desc[9], desc[10]);
+
+		if ((desc[0] != total_len) || /* descriptor length */
+		    (desc[1] != 0x5f) ||   /* vendor descriptor type */
+		    (desc[2] != 0x01) ||   /* version (2 bytes) */
+		    (desc[3] != 0x00) ||
+		    (desc[4] != total_len - 2)) /* length after type */
+			goto unrecognized;
+
+		desc_end = desc + total_len;
+		desc += 5; /* the fixed header we've already parsed */
+
+		while (desc < desc_end) {
+			u8 length;
+			u16 key;
+
+			key = *((u16 *) desc);
+			desc += sizeof(u16);
+			length = *desc;
+			desc++;
+
+			switch (key) {
+			case 0x0200: { /* max_area */
+				u32 max_area;
+				max_area = le32_to_cpu(*((u32 *)desc));
+				DRM_DEBUG("DL chip limited to %d pixel modes\n",
+					max_area);
+				udl->sku_pixel_limit = max_area;
+				break;
+			}
+			default:
+				break;
+			}
+			desc += length;
+		}
+	}
+
+	goto success;
+
+unrecognized:
+	/* allow udlfb to load for now even if firmware unrecognized */
+	DRM_ERROR("Unrecognized vendor firmware descriptor\n");
+
+success:
+	kfree(buf);
+	return true;
+}
+
+static void udl_release_urb_work(struct work_struct *work)
+{
+	struct urb_node *unode = container_of(work, struct urb_node,
+					      release_urb_work.work);
+
+	up(&unode->dev->urbs.limit_sem);
+}
+
+void udl_urb_completion(struct urb *urb)
+{
+	struct urb_node *unode = urb->context;
+	struct udl_device *udl = unode->dev;
+	unsigned long flags;
+
+	/* sync/async unlink faults aren't errors */
+	if (urb->status) {
+		if (!(urb->status == -ENOENT ||
+		    urb->status == -ECONNRESET ||
+		    urb->status == -ESHUTDOWN)) {
+			DRM_ERROR("%s - nonzero write bulk status received: %d\n",
+				__func__, urb->status);
+			atomic_set(&udl->lost_pixels, 1);
+		}
+	}
+
+	urb->transfer_buffer_length = udl->urbs.size; /* reset to actual */
+
+	spin_lock_irqsave(&udl->urbs.lock, flags);
+	list_add_tail(&unode->entry, &udl->urbs.list);
+	udl->urbs.available++;
+	spin_unlock_irqrestore(&udl->urbs.lock, flags);
+
+#if 0
+	/*
+	 * When using fb_defio, we deadlock if up() is called
+	 * while another is waiting. So queue to another process.
+	 */
+	if (fb_defio)
+		schedule_delayed_work(&unode->release_urb_work, 0);
+	else
+#endif
+		up(&udl->urbs.limit_sem);
+}
+
+static void udl_free_urb_list(struct drm_device *dev)
+{
+	struct udl_device *udl = dev->dev_private;
+	int count = udl->urbs.count;
+	struct list_head *node;
+	struct urb_node *unode;
+	struct urb *urb;
+	int ret;
+	unsigned long flags;
+
+	DRM_DEBUG("Waiting for completes and freeing all render urbs\n");
+
+	/* keep waiting and freeing, until we've got 'em all */
+	while (count--) {
+
+		/* Getting interrupted means a leak, but ok at shutdown*/
+		ret = down_interruptible(&udl->urbs.limit_sem);
+		if (ret)
+			break;
+
+		spin_lock_irqsave(&udl->urbs.lock, flags);
+
+		node = udl->urbs.list.next; /* have reserved one with sem */
+		list_del_init(node);
+
+		spin_unlock_irqrestore(&udl->urbs.lock, flags);
+
+		unode = list_entry(node, struct urb_node, entry);
+		urb = unode->urb;
+
+		/* Free each separately allocated piece */
+		usb_free_coherent(urb->dev, udl->urbs.size,
+				  urb->transfer_buffer, urb->transfer_dma);
+		usb_free_urb(urb);
+		kfree(node);
+	}
+
+}
+
+static int udl_alloc_urb_list(struct drm_device *dev, int count, size_t size)
+{
+	struct udl_device *udl = dev->dev_private;
+	int i = 0;
+	struct urb *urb;
+	struct urb_node *unode;
+	char *buf;
+
+	spin_lock_init(&udl->urbs.lock);
+
+	udl->urbs.size = size;
+	INIT_LIST_HEAD(&udl->urbs.list);
+
+	while (i < count) {
+		unode = kzalloc(sizeof(struct urb_node), GFP_KERNEL);
+		if (!unode)
+			break;
+		unode->dev = udl;
+
+		INIT_DELAYED_WORK(&unode->release_urb_work,
+			  udl_release_urb_work);
+
+		urb = usb_alloc_urb(0, GFP_KERNEL);
+		if (!urb) {
+			kfree(unode);
+			break;
+		}
+		unode->urb = urb;
+
+		buf = usb_alloc_coherent(udl->udev, MAX_TRANSFER, GFP_KERNEL,
+					 &urb->transfer_dma);
+		if (!buf) {
+			kfree(unode);
+			usb_free_urb(urb);
+			break;
+		}
+
+		/* urb->transfer_buffer_length set to actual before submit */
+		usb_fill_bulk_urb(urb, udl->udev, usb_sndbulkpipe(udl->udev, 1),
+			buf, size, udl_urb_completion, unode);
+		urb->transfer_flags |= URB_NO_TRANSFER_DMA_MAP;
+
+		list_add_tail(&unode->entry, &udl->urbs.list);
+
+		i++;
+	}
+
+	sema_init(&udl->urbs.limit_sem, i);
+	udl->urbs.count = i;
+	udl->urbs.available = i;
+
+	DRM_DEBUG("allocated %d %d byte urbs\n", i, (int) size);
+
+	return i;
+}
+
+struct urb *udl_get_urb(struct drm_device *dev)
+{
+	struct udl_device *udl = dev->dev_private;
+	int ret = 0;
+	struct list_head *entry;
+	struct urb_node *unode;
+	struct urb *urb = NULL;
+	unsigned long flags;
+
+	/* Wait for an in-flight buffer to complete and get re-queued */
+	ret = down_timeout(&udl->urbs.limit_sem, GET_URB_TIMEOUT);
+	if (ret) {
+		atomic_set(&udl->lost_pixels, 1);
+		DRM_INFO("wait for urb interrupted: %x available: %d\n",
+		       ret, udl->urbs.available);
+		goto error;
+	}
+
+	spin_lock_irqsave(&udl->urbs.lock, flags);
+
+	BUG_ON(list_empty(&udl->urbs.list)); /* reserved one with limit_sem */
+	entry = udl->urbs.list.next;
+	list_del_init(entry);
+	udl->urbs.available--;
+
+	spin_unlock_irqrestore(&udl->urbs.lock, flags);
+
+	unode = list_entry(entry, struct urb_node, entry);
+	urb = unode->urb;
+
+error:
+	return urb;
+}
+
+int udl_submit_urb(struct drm_device *dev, struct urb *urb, size_t len)
+{
+	struct udl_device *udl = dev->dev_private;
+	int ret;
+
+	BUG_ON(len > udl->urbs.size);
+
+	urb->transfer_buffer_length = len; /* set to actual payload len */
+	ret = usb_submit_urb(urb, GFP_ATOMIC);
+	if (ret) {
+		udl_urb_completion(urb); /* because no one else will */
+		atomic_set(&udl->lost_pixels, 1);
+		DRM_ERROR("usb_submit_urb error %x\n", ret);
+	}
+	return ret;
+}
+
+int udl_driver_load(struct drm_device *dev, unsigned long flags)
+{
+	struct udl_device *udl;
+	int ret;
+
+	DRM_DEBUG("\n");
+	udl = kzalloc(sizeof(struct udl_device), GFP_KERNEL);
+	if (!udl)
+		return -ENOMEM;
+
+	udl->ddev = dev;
+	udl->udev = dev->usbdev;
+	dev->dev_private = udl;
+
+	if (!udl_parse_vendor_descriptor(dev, dev->usbdev)) {
+		DRM_ERROR("firmware not recognized. Assume incompatible device\n");
+		goto err;
+	}
+
+	if (!udl_alloc_urb_list(dev, WRITES_IN_FLIGHT, MAX_TRANSFER)) {
+		ret = -ENOMEM;
+		DRM_ERROR("udl_alloc_urb_list failed\n");
+		goto err;
+	}
+
+	DRM_DEBUG("\n");
+	ret = udl_modeset_init(dev);
+
+	ret = udl_fbdev_init(dev);
+	return 0;
+err:
+	kfree(udl);
+	DRM_ERROR("%d\n", ret);
+	return ret;
+}
+
+int udl_driver_unload(struct drm_device *dev)
+{
+	struct udl_device *udl = dev->dev_private;
+
+	udl_free_urb_list(dev);
+
+	udl_fbdev_cleanup(dev);
+	udl_modeset_cleanup(dev);
+	kfree(udl);
+	return 0;
+}
diff --git a/drivers/gpu/drm/udl/udl_modeset.c b/drivers/gpu/drm/udl/udl_modeset.c
new file mode 100644
index 000000000000..6336c094b24d
--- /dev/null
+++ b/drivers/gpu/drm/udl/udl_modeset.c
@@ -0,0 +1,402 @@
+#include "drmP.h"
+#include "drm_crtc.h"
+#include "drm_crtc_helper.h"
+#include "udl_drv.h"
+
+/*
+ * All DisplayLink bulk operations start with 0xAF, followed by specific code
+ * All operations are written to buffers which then later get sent to device
+ */
+static char *udl_set_register(char *buf, u8 reg, u8 val)
+{
+	*buf++ = 0xAF;
+	*buf++ = 0x20;
+	*buf++ = reg;
+	*buf++ = val;
+	return buf;
+}
+
+static char *udl_vidreg_lock(char *buf)
+{
+	return udl_set_register(buf, 0xFF, 0x00);
+}
+
+static char *udl_vidreg_unlock(char *buf)
+{
+	return udl_set_register(buf, 0xFF, 0xFF);
+}
+
+/*
+ * On/Off for driving the DisplayLink framebuffer to the display
+ *  0x00 H and V sync on
+ *  0x01 H and V sync off (screen blank but powered)
+ *  0x07 DPMS powerdown (requires modeset to come back)
+ */
+static char *udl_enable_hvsync(char *buf, bool enable)
+{
+	if (enable)
+		return udl_set_register(buf, 0x1F, 0x00);
+	else
+		return udl_set_register(buf, 0x1F, 0x07);
+}
+
+static char *udl_set_color_depth(char *buf, u8 selection)
+{
+	return udl_set_register(buf, 0x00, selection);
+}
+
+static char *udl_set_base16bpp(char *wrptr, u32 base)
+{
+	/* the base pointer is 16 bits wide, 0x20 is hi byte. */
+	wrptr = udl_set_register(wrptr, 0x20, base >> 16);
+	wrptr = udl_set_register(wrptr, 0x21, base >> 8);
+	return udl_set_register(wrptr, 0x22, base);
+}
+
+/*
+ * DisplayLink HW has separate 16bpp and 8bpp framebuffers.
+ * In 24bpp modes, the low 323 RGB bits go in the 8bpp framebuffer
+ */
+static char *udl_set_base8bpp(char *wrptr, u32 base)
+{
+	wrptr = udl_set_register(wrptr, 0x26, base >> 16);
+	wrptr = udl_set_register(wrptr, 0x27, base >> 8);
+	return udl_set_register(wrptr, 0x28, base);
+}
+
+static char *udl_set_register_16(char *wrptr, u8 reg, u16 value)
+{
+	wrptr = udl_set_register(wrptr, reg, value >> 8);
+	return udl_set_register(wrptr, reg+1, value);
+}
+
+/*
+ * This is kind of weird because the controller takes some
+ * register values in a different byte order than other registers.
+ */
+static char *udl_set_register_16be(char *wrptr, u8 reg, u16 value)
+{
+	wrptr = udl_set_register(wrptr, reg, value);
+	return udl_set_register(wrptr, reg+1, value >> 8);
+}
+
+/*
+ * LFSR is linear feedback shift register. The reason we have this is
+ * because the display controller needs to minimize the clock depth of
+ * various counters used in the display path. So this code reverses the
+ * provided value into the lfsr16 value by counting backwards to get
+ * the value that needs to be set in the hardware comparator to get the
+ * same actual count. This makes sense once you read above a couple of
+ * times and think about it from a hardware perspective.
+ */
+static u16 udl_lfsr16(u16 actual_count)
+{
+	u32 lv = 0xFFFF; /* This is the lfsr value that the hw starts with */
+
+	while (actual_count--) {
+		lv =	 ((lv << 1) |
+			(((lv >> 15) ^ (lv >> 4) ^ (lv >> 2) ^ (lv >> 1)) & 1))
+			& 0xFFFF;
+	}
+
+	return (u16) lv;
+}
+
+/*
+ * This does LFSR conversion on the value that is to be written.
+ * See LFSR explanation above for more detail.
+ */
+static char *udl_set_register_lfsr16(char *wrptr, u8 reg, u16 value)
+{
+	return udl_set_register_16(wrptr, reg, udl_lfsr16(value));
+}
+
+/*
+ * This takes a standard fbdev screeninfo struct and all of its monitor mode
+ * details and converts them into the DisplayLink equivalent register commands.
+  ERR(vreg(dev,               0x00, (color_depth == 16) ? 0 : 1));
+  ERR(vreg_lfsr16(dev,        0x01, xDisplayStart));
+  ERR(vreg_lfsr16(dev,        0x03, xDisplayEnd));
+  ERR(vreg_lfsr16(dev,        0x05, yDisplayStart));
+  ERR(vreg_lfsr16(dev,        0x07, yDisplayEnd));
+  ERR(vreg_lfsr16(dev,        0x09, xEndCount));
+  ERR(vreg_lfsr16(dev,        0x0B, hSyncStart));
+  ERR(vreg_lfsr16(dev,        0x0D, hSyncEnd));
+  ERR(vreg_big_endian(dev,    0x0F, hPixels));
+  ERR(vreg_lfsr16(dev,        0x11, yEndCount));
+  ERR(vreg_lfsr16(dev,        0x13, vSyncStart));
+  ERR(vreg_lfsr16(dev,        0x15, vSyncEnd));
+  ERR(vreg_big_endian(dev,    0x17, vPixels));
+  ERR(vreg_little_endian(dev, 0x1B, pixelClock5KHz));
+
+  ERR(vreg(dev,               0x1F, 0));
+
+  ERR(vbuf(dev, WRITE_VIDREG_UNLOCK, DSIZEOF(WRITE_VIDREG_UNLOCK)));
+ */
+static char *udl_set_vid_cmds(char *wrptr, struct drm_display_mode *mode)
+{
+	u16 xds, yds;
+	u16 xde, yde;
+	u16 yec;
+
+	/* x display start */
+	xds = mode->crtc_htotal - mode->crtc_hsync_start;
+	wrptr = udl_set_register_lfsr16(wrptr, 0x01, xds);
+	/* x display end */
+	xde = xds + mode->crtc_hdisplay;
+	wrptr = udl_set_register_lfsr16(wrptr, 0x03, xde);
+
+	/* y display start */
+	yds = mode->crtc_vtotal - mode->crtc_vsync_start;
+	wrptr = udl_set_register_lfsr16(wrptr, 0x05, yds);
+	/* y display end */
+	yde = yds + mode->crtc_vdisplay;
+	wrptr = udl_set_register_lfsr16(wrptr, 0x07, yde);
+
+	/* x end count is active + blanking - 1 */
+	wrptr = udl_set_register_lfsr16(wrptr, 0x09,
+					mode->crtc_htotal - 1);
+
+	/* libdlo hardcodes hsync start to 1 */
+	wrptr = udl_set_register_lfsr16(wrptr, 0x0B, 1);
+
+	/* hsync end is width of sync pulse + 1 */
+	wrptr = udl_set_register_lfsr16(wrptr, 0x0D,
+					mode->crtc_hsync_end - mode->crtc_hsync_start + 1);
+
+	/* hpixels is active pixels */
+	wrptr = udl_set_register_16(wrptr, 0x0F, mode->hdisplay);
+
+	/* yendcount is vertical active + vertical blanking */
+	yec = mode->crtc_vtotal;
+	wrptr = udl_set_register_lfsr16(wrptr, 0x11, yec);
+
+	/* libdlo hardcodes vsync start to 0 */
+	wrptr = udl_set_register_lfsr16(wrptr, 0x13, 0);
+
+	/* vsync end is width of vsync pulse */
+	wrptr = udl_set_register_lfsr16(wrptr, 0x15, mode->crtc_vsync_end - mode->crtc_vsync_start);
+
+	/* vpixels is active pixels */
+	wrptr = udl_set_register_16(wrptr, 0x17, mode->crtc_vdisplay);
+
+	wrptr = udl_set_register_16be(wrptr, 0x1B,
+				      mode->clock / 5);
+
+	return wrptr;
+}
+
+static int udl_crtc_write_mode_to_hw(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	struct udl_device *udl = dev->dev_private;
+	struct urb *urb;
+	char *buf;
+	int retval;
+
+	urb = udl_get_urb(dev);
+	if (!urb)
+		return -ENOMEM;
+
+	buf = (char *)urb->transfer_buffer;
+
+	memcpy(buf, udl->mode_buf, udl->mode_buf_len);
+	retval = udl_submit_urb(dev, urb, udl->mode_buf_len);
+	DRM_INFO("write mode info %d\n", udl->mode_buf_len);
+	return retval;
+}
+
+
+static void udl_crtc_dpms(struct drm_crtc *crtc, int mode)
+{
+	struct drm_device *dev = crtc->dev;
+	struct udl_device *udl = dev->dev_private;
+	int retval;
+
+	if (mode == DRM_MODE_DPMS_OFF) {
+		char *buf;
+		struct urb *urb;
+		urb = udl_get_urb(dev);
+		if (!urb)
+			return;
+
+		buf = (char *)urb->transfer_buffer;
+		buf = udl_vidreg_lock(buf);
+		buf = udl_enable_hvsync(buf, false);
+		buf = udl_vidreg_unlock(buf);
+
+		retval = udl_submit_urb(dev, urb, buf - (char *)
+					urb->transfer_buffer);
+	} else {
+		if (udl->mode_buf_len == 0) {
+			DRM_ERROR("Trying to enable DPMS with no mode\n");
+			return;
+		}
+		udl_crtc_write_mode_to_hw(crtc);
+	}
+
+}
+
+static bool udl_crtc_mode_fixup(struct drm_crtc *crtc,
+				  struct drm_display_mode *mode,
+				  struct drm_display_mode *adjusted_mode)
+
+{
+	return true;
+}
+
+#if 0
+static int
+udl_pipe_set_base_atomic(struct drm_crtc *crtc, struct drm_framebuffer *fb,
+			   int x, int y, enum mode_set_atomic state)
+{
+	return 0;
+}
+
+static int
+udl_pipe_set_base(struct drm_crtc *crtc, int x, int y,
+		    struct drm_framebuffer *old_fb)
+{
+	return 0;
+}
+#endif
+
+static int udl_crtc_mode_set(struct drm_crtc *crtc,
+			       struct drm_display_mode *mode,
+			       struct drm_display_mode *adjusted_mode,
+			       int x, int y,
+			       struct drm_framebuffer *old_fb)
+
+{
+	struct drm_device *dev = crtc->dev;
+	struct udl_framebuffer *ufb = to_udl_fb(crtc->fb);
+	struct udl_device *udl = dev->dev_private;
+	char *buf;
+	char *wrptr;
+	int color_depth = 0;
+
+	buf = (char *)udl->mode_buf;
+
+	if (crtc->fb->bits_per_pixel != 16)
+		color_depth = 1;
+
+	/* This first section has to do with setting the base address on the
+	* controller * associated with the display. There are 2 base
+	* pointers, currently, we only * use the 16 bpp segment.
+	*/
+	wrptr = udl_vidreg_lock(buf);
+	wrptr = udl_set_color_depth(wrptr, color_depth);
+	/* set base for 16bpp segment to 0 */
+	wrptr = udl_set_base16bpp(wrptr, 0);
+	/* set base for 8bpp segment to end of fb */
+	printk("8bpp offset %08x\n", 2 * mode->vdisplay * mode->hdisplay);
+	wrptr = udl_set_base8bpp(wrptr, 2 * mode->vdisplay * mode->hdisplay); //0x35e000);
+
+	wrptr = udl_set_vid_cmds(wrptr, adjusted_mode);
+	wrptr = udl_enable_hvsync(wrptr, true);
+	wrptr = udl_vidreg_unlock(wrptr);
+	DRM_INFO("setup mode info %ld\n", wrptr-buf);
+
+	ufb->active_16 = true;
+	if (old_fb) {
+		struct udl_framebuffer *uold_fb = to_udl_fb(old_fb);
+		uold_fb->active_16 = false;
+	}
+	udl->mode_buf_len = wrptr - buf;
+
+	/* damage all of it */
+	udl_handle_damage(ufb, 0, 0, ufb->base.width, ufb->base.height);
+	return 0;
+}
+
+
+static void udl_crtc_disable(struct drm_crtc *crtc)
+{
+
+
+}
+
+static void udl_crtc_destroy(struct drm_crtc *crtc)
+{
+	drm_crtc_cleanup(crtc);
+	kfree(crtc);
+}
+
+static void udl_load_lut(struct drm_crtc *crtc)
+{
+}
+
+static void udl_crtc_prepare(struct drm_crtc *crtc)
+{
+}
+
+static void udl_crtc_commit(struct drm_crtc *crtc)
+{
+	udl_crtc_dpms(crtc, DRM_MODE_DPMS_ON);
+}
+
+static struct drm_crtc_helper_funcs udl_helper_funcs = {
+	.dpms = udl_crtc_dpms,
+	.mode_fixup = udl_crtc_mode_fixup,
+	.mode_set = udl_crtc_mode_set,
+	.prepare = udl_crtc_prepare,
+	.commit = udl_crtc_commit,
+	.disable = udl_crtc_disable,
+	.load_lut = udl_load_lut,
+};
+
+static const struct drm_crtc_funcs udl_crtc_funcs = {
+	.set_config = drm_crtc_helper_set_config,
+	.destroy = udl_crtc_destroy,
+};
+
+int udl_crtc_init(struct drm_device *dev)
+{
+	struct drm_crtc *crtc;
+
+	crtc = kzalloc(sizeof(struct drm_crtc) + sizeof(struct drm_connector *), GFP_KERNEL);
+	if (crtc == NULL)
+		return -ENOMEM;
+
+	drm_crtc_init(dev, crtc, &udl_crtc_funcs);
+	drm_crtc_helper_add(crtc, &udl_helper_funcs);
+
+	return 0;
+}
+
+static const struct drm_mode_config_funcs udl_mode_funcs = {
+	.fb_create = udl_fb_user_fb_create,
+	.output_poll_changed = NULL,
+};
+
+int udl_modeset_init(struct drm_device *dev)
+{
+	struct drm_encoder *encoder;
+	drm_mode_config_init(dev);
+
+	dev->mode_config.min_width = 640;
+	dev->mode_config.min_height = 480;
+
+	dev->mode_config.max_width = 2048;
+	dev->mode_config.max_height = 2048;
+
+	dev->mode_config.prefer_shadow = 0;
+	dev->mode_config.preferred_depth = 24;
+
+	dev->mode_config.funcs = (void *)&udl_mode_funcs;
+
+	drm_mode_create_dirty_info_property(dev);
+
+	udl_crtc_init(dev);
+
+	encoder = udl_encoder_init(dev);
+
+	udl_connector_init(dev, encoder);
+
+	return 0;
+}
+
+void udl_modeset_cleanup(struct drm_device *dev)
+{
+	drm_mode_config_cleanup(dev);
+}
diff --git a/drivers/gpu/drm/udl/udl_transfer.c b/drivers/gpu/drm/udl/udl_transfer.c
new file mode 100644
index 000000000000..0cd14a03bc71
--- /dev/null
+++ b/drivers/gpu/drm/udl/udl_transfer.c
@@ -0,0 +1,258 @@
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/fb.h>
+#include <linux/prefetch.h>
+
+#include "drmP.h"
+#include "udl_drv.h"
+
+#define MAX_CMD_PIXELS		255
+
+#define RLX_HEADER_BYTES	7
+#define MIN_RLX_PIX_BYTES       4
+#define MIN_RLX_CMD_BYTES	(RLX_HEADER_BYTES + MIN_RLX_PIX_BYTES)
+
+#define RLE_HEADER_BYTES	6
+#define MIN_RLE_PIX_BYTES	3
+#define MIN_RLE_CMD_BYTES	(RLE_HEADER_BYTES + MIN_RLE_PIX_BYTES)
+
+#define RAW_HEADER_BYTES	6
+#define MIN_RAW_PIX_BYTES	2
+#define MIN_RAW_CMD_BYTES	(RAW_HEADER_BYTES + MIN_RAW_PIX_BYTES)
+
+/*
+ * Trims identical data from front and back of line
+ * Sets new front buffer address and width
+ * And returns byte count of identical pixels
+ * Assumes CPU natural alignment (unsigned long)
+ * for back and front buffer ptrs and width
+ */
+static int udl_trim_hline(const u8 *bback, const u8 **bfront, int *width_bytes)
+{
+	int j, k;
+	const unsigned long *back = (const unsigned long *) bback;
+	const unsigned long *front = (const unsigned long *) *bfront;
+	const int width = *width_bytes / sizeof(unsigned long);
+	int identical = width;
+	int start = width;
+	int end = width;
+
+	prefetch((void *) front);
+	prefetch((void *) back);
+
+	for (j = 0; j < width; j++) {
+		if (back[j] != front[j]) {
+			start = j;
+			break;
+		}
+	}
+
+	for (k = width - 1; k > j; k--) {
+		if (back[k] != front[k]) {
+			end = k+1;
+			break;
+		}
+	}
+
+	identical = start + (width - end);
+	*bfront = (u8 *) &front[start];
+	*width_bytes = (end - start) * sizeof(unsigned long);
+
+	return identical * sizeof(unsigned long);
+}
+
+static inline u16 pixel32_to_be16p(const uint8_t *pixel)
+{
+	uint32_t pix = *(uint32_t *)pixel;
+	u16 retval;
+
+	retval =  (((pix >> 3) & 0x001f) |
+		   ((pix >> 6) & 0x03e0) |
+		   ((pix >> 9) & 0x7c00));
+	return retval;
+}
+
+/*
+ * Render a command stream for an encoded horizontal line segment of pixels.
+ *
+ * A command buffer holds several commands.
+ * It always begins with a fresh command header
+ * (the protocol doesn't require this, but we enforce it to allow
+ * multiple buffers to be potentially encoded and sent in parallel).
+ * A single command encodes one contiguous horizontal line of pixels
+ *
+ * The function relies on the client to do all allocation, so that
+ * rendering can be done directly to output buffers (e.g. USB URBs).
+ * The function fills the supplied command buffer, providing information
+ * on where it left off, so the client may call in again with additional
+ * buffers if the line will take several buffers to complete.
+ *
+ * A single command can transmit a maximum of 256 pixels,
+ * regardless of the compression ratio (protocol design limit).
+ * To the hardware, 0 for a size byte means 256
+ *
+ * Rather than 256 pixel commands which are either rl or raw encoded,
+ * the rlx command simply assumes alternating raw and rl spans within one cmd.
+ * This has a slightly larger header overhead, but produces more even results.
+ * It also processes all data (read and write) in a single pass.
+ * Performance benchmarks of common cases show it having just slightly better
+ * compression than 256 pixel raw or rle commands, with similar CPU consumpion.
+ * But for very rl friendly data, will compress not quite as well.
+ */
+static void udl_compress_hline16(
+	const u8 **pixel_start_ptr,
+	const u8 *const pixel_end,
+	uint32_t *device_address_ptr,
+	uint8_t **command_buffer_ptr,
+	const uint8_t *const cmd_buffer_end, int bpp)
+{
+	const u8 *pixel = *pixel_start_ptr;
+	uint32_t dev_addr  = *device_address_ptr;
+	uint8_t *cmd = *command_buffer_ptr;
+
+	while ((pixel_end > pixel) &&
+	       (cmd_buffer_end - MIN_RLX_CMD_BYTES > cmd)) {
+		uint8_t *raw_pixels_count_byte = 0;
+		uint8_t *cmd_pixels_count_byte = 0;
+		const u8 *raw_pixel_start = 0;
+		const u8 *cmd_pixel_start, *cmd_pixel_end = 0;
+
+		prefetchw((void *) cmd); /* pull in one cache line at least */
+
+		*cmd++ = 0xaf;
+		*cmd++ = 0x6b;
+		*cmd++ = (uint8_t) ((dev_addr >> 16) & 0xFF);
+		*cmd++ = (uint8_t) ((dev_addr >> 8) & 0xFF);
+		*cmd++ = (uint8_t) ((dev_addr) & 0xFF);
+
+		cmd_pixels_count_byte = cmd++; /*  we'll know this later */
+		cmd_pixel_start = pixel;
+
+		raw_pixels_count_byte = cmd++; /*  we'll know this later */
+		raw_pixel_start = pixel;
+
+		cmd_pixel_end = pixel + (min(MAX_CMD_PIXELS + 1,
+			min((int)(pixel_end - pixel) / bpp,
+			    (int)(cmd_buffer_end - cmd) / 2))) * bpp;
+
+		prefetch_range((void *) pixel, (cmd_pixel_end - pixel) * bpp);
+
+		while (pixel < cmd_pixel_end) {
+			const u8 * const repeating_pixel = pixel;
+
+			if (bpp == 2)
+				*(uint16_t *)cmd = cpu_to_be16p((uint16_t *)pixel);
+			else if (bpp == 4)
+				*(uint16_t *)cmd = pixel32_to_be16p(pixel);
+
+			cmd += 2;
+			pixel += bpp;
+
+			if (unlikely((pixel < cmd_pixel_end) &&
+				     (!memcmp(pixel, repeating_pixel, bpp)))) {
+				/* go back and fill in raw pixel count */
+				*raw_pixels_count_byte = (((repeating_pixel -
+						raw_pixel_start) / bpp) + 1) & 0xFF;
+
+				while ((pixel < cmd_pixel_end)
+				       && (!memcmp(pixel, repeating_pixel, bpp))) {
+					pixel += bpp;
+				}
+
+				/* immediately after raw data is repeat byte */
+				*cmd++ = (((pixel - repeating_pixel) / bpp) - 1) & 0xFF;
+
+				/* Then start another raw pixel span */
+				raw_pixel_start = pixel;
+				raw_pixels_count_byte = cmd++;
+			}
+		}
+
+		if (pixel > raw_pixel_start) {
+			/* finalize last RAW span */
+			*raw_pixels_count_byte = ((pixel-raw_pixel_start) / bpp) & 0xFF;
+		}
+
+		*cmd_pixels_count_byte = ((pixel - cmd_pixel_start) / bpp) & 0xFF;
+		dev_addr += ((pixel - cmd_pixel_start) / bpp) * 2;
+	}
+
+	if (cmd_buffer_end <= MIN_RLX_CMD_BYTES + cmd) {
+		/* Fill leftover bytes with no-ops */
+		if (cmd_buffer_end > cmd)
+			memset(cmd, 0xAF, cmd_buffer_end - cmd);
+		cmd = (uint8_t *) cmd_buffer_end;
+	}
+
+	*command_buffer_ptr = cmd;
+	*pixel_start_ptr = pixel;
+	*device_address_ptr = dev_addr;
+
+	return;
+}
+
+/*
+ * There are 3 copies of every pixel: The front buffer that the fbdev
+ * client renders to, the actual framebuffer across the USB bus in hardware
+ * (that we can only write to, slowly, and can never read), and (optionally)
+ * our shadow copy that tracks what's been sent to that hardware buffer.
+ */
+int udl_render_hline(struct drm_device *dev, int bpp, struct urb **urb_ptr,
+		     const char *front, char **urb_buf_ptr,
+		     u32 byte_offset, u32 byte_width,
+		     int *ident_ptr, int *sent_ptr)
+{
+	struct udl_device *udl = dev->dev_private;
+	const u8 *line_start, *line_end, *next_pixel;
+	u32 base16 = 0 + (byte_offset / bpp) * 2;
+	struct urb *urb = *urb_ptr;
+	u8 *cmd = *urb_buf_ptr;
+	u8 *cmd_end = (u8 *) urb->transfer_buffer + urb->transfer_buffer_length;
+
+	line_start = (u8 *) (front + byte_offset);
+	next_pixel = line_start;
+	line_end = next_pixel + byte_width;
+
+	if (udl->backing_buffer) {
+		int offset;
+		const u8 *back_start = (u8 *) (udl->backing_buffer
+						+ byte_offset);
+
+		*ident_ptr += udl_trim_hline(back_start, &next_pixel,
+			&byte_width);
+
+		offset = next_pixel - line_start;
+		line_end = next_pixel + byte_width;
+		base16 += offset;
+		back_start += offset;
+		line_start += offset;
+
+		memcpy((char *)back_start, (char *) line_start,
+		       byte_width);
+	}
+
+	while (next_pixel < line_end) {
+
+		udl_compress_hline16(&next_pixel,
+			     line_end, &base16,
+			     (u8 **) &cmd, (u8 *) cmd_end, bpp);
+
+		if (cmd >= cmd_end) {
+			int len = cmd - (u8 *) urb->transfer_buffer;
+			if (udl_submit_urb(dev, urb, len))
+				return 1; /* lost pixels is set */
+			*sent_ptr += len;
+			urb = udl_get_urb(dev);
+			if (!urb)
+				return 1; /* lost_pixels is set */
+			*urb_ptr = urb;
+			cmd = urb->transfer_buffer;
+			cmd_end = &cmd[urb->transfer_buffer_length];
+		}
+	}
+
+	*urb_buf_ptr = cmd;
+
+	return 0;
+}
+