Merge tag 'topic/drm-misc-2016-06-01' of git://anongit.freedesktop.org/drm-intel into drm-next

Frist -misc pull for 4.8, with pretty much just random all over plus a few
more lockless gem BO patches acked/reviewed by driver maintainers.

I'm starting a bit earlier this time around because there's a few invasive
patch series to land (nonblocking atomic prep work, fence prep work,
rst/sphinx kerneldoc finally happening) and I need a baseline with all the
branches merged.

* tag 'topic/drm-misc-2016-06-01' of git://anongit.freedesktop.org/drm-intel: (21 commits)
  drm/vc4: Use lockless gem BO free callback
  drm/vc4: Use drm_gem_object_unreference_unlocked
  drm: Initialize a linear gamma table by default
  drm/vgem: Use lockless gem BO free callback
  drm/qxl: Don't set a gamma table size
  drm/msm: Nuke dummy gamma_set/get functions
  drm/cirrus: Drop redundnant gamma size check
  drm/fb-helper: Remove dead code in setcolreg
  drm/mediatek: Use lockless gem BO free callback
  drm/hisilicon: Use lockless gem BO free callback
  drm/hlcd: Use lockless gem BO free callback
  vga_switcheroo: Support deferred probing of audio clients
  vga_switcheroo: Add helper for deferred probing
  virtio-gpu: fix output lookup
  drm/doc: Unify KMS Locking docs
  drm/atomic-helper: Do not call ->mode_fixup for CRTC which will be disabled
  Fix annoyingly awkward typo in drm_edid_load.c
  drm/doc: Drop vblank_disable_allow wording
  drm: use seqlock for vblank time/count
  drm/mm: avoid possible null pointer dereference
  ...

Documentation/DocBook/gpu.tmpl

@@ -1092,22 +1092,6 @@ int max_width, max_height;</synopsis>
         operation.
       </para>
     </sect2>
-    <sect2>
-      <title>Locking</title>
-      <para>
-        Beside some lookup structures with their own locking (which is hidden
-	behind the interface functions) most of the modeset state is protected
-	by the <code>dev-&lt;mode_config.lock</code> mutex and additionally
-	per-crtc locks to allow cursor updates, pageflips and similar operations
-	to occur concurrently with background tasks like output detection.
-	Operations which cross domains like a full modeset always grab all
-	locks. Drivers there need to protect resources shared between crtcs with
-	additional locking. They also need to be careful to always grab the
-	relevant crtc locks if a modset functions touches crtc state, e.g. for
-	load detection (which does only grab the <code>mode_config.lock</code>
-	to allow concurrent screen updates on live crtcs).
-      </para>
-    </sect2>
   </sect1>
 
   <!-- Internals: kms initialization and cleanup -->
@@ -2845,14 +2829,7 @@ void (*disable_vblank) (struct drm_device *dev, int crtc);</synopsis>
     <para>
       Drivers must initialize the vertical blanking handling core with a call to
       <function>drm_vblank_init</function> in their
-      <methodname>load</methodname> operation. The function will set the struct
-      <structname>drm_device</structname>
-      <structfield>vblank_disable_allowed</structfield> field to 0. This will
-      keep vertical blanking interrupts enabled permanently until the first mode
-      set operation, where <structfield>vblank_disable_allowed</structfield> is
-      set to 1. The reason behind this is not clear. Drivers can set the field
-      to 1 after <function>calling drm_vblank_init</function> to make vertical
-      blanking interrupts dynamically managed from the beginning.
+      <methodname>load</methodname> operation.
     </para>
     <para>
       Vertical blanking interrupts can be enabled by the DRM core or by drivers

MAINTAINERS

@@ -3854,6 +3854,9 @@ T:	git git://people.freedesktop.org/~airlied/linux
 S:	Maintained
 F:	drivers/gpu/drm/
 F:	drivers/gpu/vga/
+F:	Documentation/devicetree/bindings/display/
+F:	Documentation/devicetree/bindings/gpu/
+F:	Documentation/devicetree/bindings/video/
 F:	Documentation/DocBook/gpu.*
 F:	include/drm/
 F:	include/uapi/drm/

drivers/gpu/drm/arm/hdlcd_drv.c

@@ -316,7 +316,7 @@ static struct drm_driver hdlcd_driver = {
 	.get_vblank_counter = drm_vblank_no_hw_counter,
 	.enable_vblank = hdlcd_enable_vblank,
 	.disable_vblank = hdlcd_disable_vblank,
-	.gem_free_object = drm_gem_cma_free_object,
+	.gem_free_object_unlocked = drm_gem_cma_free_object,
 	.gem_vm_ops = &drm_gem_cma_vm_ops,
 	.dumb_create = drm_gem_cma_dumb_create,
 	.dumb_map_offset = drm_gem_cma_dumb_map_offset,

drivers/gpu/drm/cirrus/cirrus_mode.c

@@ -331,9 +331,6 @@ static void cirrus_crtc_gamma_set(struct drm_crtc *crtc, u16 *red, u16 *green,
 	struct cirrus_crtc *cirrus_crtc = to_cirrus_crtc(crtc);
 	int i;
 
-	if (size != CIRRUS_LUT_SIZE)
-		return;
-
 	for (i = 0; i < CIRRUS_LUT_SIZE; i++) {
 		cirrus_crtc->lut_r[i] = red[i];
 		cirrus_crtc->lut_g[i] = green[i];

drivers/gpu/drm/drm_atomic_helper.c

@@ -414,6 +414,9 @@ mode_fixup(struct drm_atomic_state *state)
 	for_each_crtc_in_state(state, crtc, crtc_state, i) {
 		const struct drm_crtc_helper_funcs *funcs;
 
+		if (!crtc_state->enable)
+			continue;
+
 		if (!crtc_state->mode_changed &&
 		    !crtc_state->connectors_changed)
 			continue;

drivers/gpu/drm/drm_crtc.c

@@ -5139,6 +5139,9 @@ EXPORT_SYMBOL(drm_mode_connector_attach_encoder);
 int drm_mode_crtc_set_gamma_size(struct drm_crtc *crtc,
 				 int gamma_size)
 {
+	uint16_t *r_base, *g_base, *b_base;
+	int i;
+
 	crtc->gamma_size = gamma_size;
 
 	crtc->gamma_store = kcalloc(gamma_size, sizeof(uint16_t) * 3,
@@ -5148,6 +5151,16 @@ int drm_mode_crtc_set_gamma_size(struct drm_crtc *crtc,
 		return -ENOMEM;
 	}
 
+	r_base = crtc->gamma_store;
+	g_base = r_base + gamma_size;
+	b_base = g_base + gamma_size;
+	for (i = 0; i < gamma_size; i++) {
+		r_base[i] = i << 8;
+		g_base[i] = i << 8;
+		b_base[i] = i << 8;
+	}
+
+
 	return 0;
 }
 EXPORT_SYMBOL(drm_mode_crtc_set_gamma_size);

drivers/gpu/drm/drm_edid_load.c

@@ -271,7 +271,7 @@ int drm_load_edid_firmware(struct drm_connector *connector)
 	 * by commas, search through the list looking for one that
 	 * matches the connector.
 	 *
-	 * If there's one or more that don't't specify a connector, keep
+	 * If there's one or more that doesn't specify a connector, keep
 	 * the last one found one as a fallback.
 	 */
 	fwstr = kstrdup(edid_firmware, GFP_KERNEL);

drivers/gpu/drm/drm_fb_helper.c

@@ -1042,7 +1042,6 @@ static int setcolreg(struct drm_crtc *crtc, u16 red, u16 green,
 {
 	struct drm_fb_helper *fb_helper = info->par;
 	struct drm_framebuffer *fb = fb_helper->fb;
-	int pindex;
 
 	if (info->fix.visual == FB_VISUAL_TRUECOLOR) {
 		u32 *palette;
@@ -1074,38 +1073,10 @@ static int setcolreg(struct drm_crtc *crtc, u16 red, u16 green,
 		    !fb_helper->funcs->gamma_get))
 		return -EINVAL;
 
-	pindex = regno;
+	WARN_ON(fb->bits_per_pixel != 8);
 
-	if (fb->bits_per_pixel == 16) {
-		pindex = regno << 3;
+	fb_helper->funcs->gamma_set(crtc, red, green, blue, regno);
 
-		if (fb->depth == 16 && regno > 63)
-			return -EINVAL;
-		if (fb->depth == 15 && regno > 31)
-			return -EINVAL;
-
-		if (fb->depth == 16) {
-			u16 r, g, b;
-			int i;
-			if (regno < 32) {
-				for (i = 0; i < 8; i++)
-					fb_helper->funcs->gamma_set(crtc, red,
-						green, blue, pindex + i);
-			}
-
-			fb_helper->funcs->gamma_get(crtc, &r,
-						    &g, &b,
-						    pindex >> 1);
-
-			for (i = 0; i < 4; i++)
-				fb_helper->funcs->gamma_set(crtc, r,
-							    green, b,
-							    (pindex >> 1) + i);
-		}
-	}
-
-	if (fb->depth != 16)
-		fb_helper->funcs->gamma_set(crtc, red, green, blue, pindex);
 	return 0;
 }
 

drivers/gpu/drm/drm_irq.c

@@ -42,10 +42,6 @@
 #include <linux/vgaarb.h>
 #include <linux/export.h>
 
-/* Access macro for slots in vblank timestamp ringbuffer. */
-#define vblanktimestamp(dev, pipe, count) \
-	((dev)->vblank[pipe].time[(count) % DRM_VBLANKTIME_RBSIZE])
-
 /* Retry timestamp calculation up to 3 times to satisfy
  * drm_timestamp_precision before giving up.
  */
@@ -82,29 +78,15 @@ static void store_vblank(struct drm_device *dev, unsigned int pipe,
 			 struct timeval *t_vblank, u32 last)
 {
 	struct drm_vblank_crtc *vblank = &dev->vblank[pipe];
-	u32 tslot;
 
 	assert_spin_locked(&dev->vblank_time_lock);
 
 	vblank->last = last;
 
-	/* All writers hold the spinlock, but readers are serialized by
-	 * the latching of vblank->count below.
-	 */
-	tslot = vblank->count + vblank_count_inc;
-	vblanktimestamp(dev, pipe, tslot) = *t_vblank;
-
-	/*
-	 * vblank timestamp updates are protected on the write side with
-	 * vblank_time_lock, but on the read side done locklessly using a
-	 * sequence-lock on the vblank counter. Ensure correct ordering using
-	 * memory barrriers. We need the barrier both before and also after the
-	 * counter update to synchronize with the next timestamp write.
-	 * The read-side barriers for this are in drm_vblank_count_and_time.
-	 */
-	smp_wmb();
+	write_seqlock(&vblank->seqlock);
+	vblank->time = *t_vblank;
 	vblank->count += vblank_count_inc;
-	smp_wmb();
+	write_sequnlock(&vblank->seqlock);
 }
 
 /**
@@ -205,7 +187,7 @@ static void drm_update_vblank_count(struct drm_device *dev, unsigned int pipe,
 		const struct timeval *t_old;
 		u64 diff_ns;
 
-		t_old = &vblanktimestamp(dev, pipe, vblank->count);
+		t_old = &vblank->time;
 		diff_ns = timeval_to_ns(&t_vblank) - timeval_to_ns(t_old);
 
 		/*
@@ -239,49 +221,6 @@ static void drm_update_vblank_count(struct drm_device *dev, unsigned int pipe,
 		diff = 1;
 	}
 
-	/*
-	 * FIMXE: Need to replace this hack with proper seqlocks.
-	 *
-	 * Restrict the bump of the software vblank counter to a safe maximum
-	 * value of +1 whenever there is the possibility that concurrent readers
-	 * of vblank timestamps could be active at the moment, as the current
-	 * implementation of the timestamp caching and updating is not safe
-	 * against concurrent readers for calls to store_vblank() with a bump
-	 * of anything but +1. A bump != 1 would very likely return corrupted
-	 * timestamps to userspace, because the same slot in the cache could
-	 * be concurrently written by store_vblank() and read by one of those
-	 * readers without the read-retry logic detecting the collision.
-	 *
-	 * Concurrent readers can exist when we are called from the
-	 * drm_vblank_off() or drm_vblank_on() functions and other non-vblank-
-	 * irq callers. However, all those calls to us are happening with the
-	 * vbl_lock locked to prevent drm_vblank_get(), so the vblank refcount
-	 * can't increase while we are executing. Therefore a zero refcount at
-	 * this point is safe for arbitrary counter bumps if we are called
-	 * outside vblank irq, a non-zero count is not 100% safe. Unfortunately
-	 * we must also accept a refcount of 1, as whenever we are called from
-	 * drm_vblank_get() -> drm_vblank_enable() the refcount will be 1 and
-	 * we must let that one pass through in order to not lose vblank counts
-	 * during vblank irq off - which would completely defeat the whole
-	 * point of this routine.
-	 *
-	 * Whenever we are called from vblank irq, we have to assume concurrent
-	 * readers exist or can show up any time during our execution, even if
-	 * the refcount is currently zero, as vblank irqs are usually only
-	 * enabled due to the presence of readers, and because when we are called
-	 * from vblank irq we can't hold the vbl_lock to protect us from sudden
-	 * bumps in vblank refcount. Therefore also restrict bumps to +1 when
-	 * called from vblank irq.
-	 */
-	if ((diff > 1) && (atomic_read(&vblank->refcount) > 1 ||
-	    (flags & DRM_CALLED_FROM_VBLIRQ))) {
-		DRM_DEBUG_VBL("clamping vblank bump to 1 on crtc %u: diffr=%u "
-			      "refcount %u, vblirq %u\n", pipe, diff,
-			      atomic_read(&vblank->refcount),
-			      (flags & DRM_CALLED_FROM_VBLIRQ) != 0);
-		diff = 1;
-	}
-
 	DRM_DEBUG_VBL("updating vblank count on crtc %u:"
 		      " current=%u, diff=%u, hw=%u hw_last=%u\n",
 		      pipe, vblank->count, diff, cur_vblank, vblank->last);
@@ -417,6 +356,7 @@ int drm_vblank_init(struct drm_device *dev, unsigned int num_crtcs)
 		init_waitqueue_head(&vblank->queue);
 		setup_timer(&vblank->disable_timer, vblank_disable_fn,
 			    (unsigned long)vblank);
+		seqlock_init(&vblank->seqlock);
 	}
 
 	DRM_INFO("Supports vblank timestamp caching Rev 2 (21.10.2013).\n");
@@ -986,25 +926,19 @@ u32 drm_vblank_count_and_time(struct drm_device *dev, unsigned int pipe,
 			      struct timeval *vblanktime)
 {
 	struct drm_vblank_crtc *vblank = &dev->vblank[pipe];
-	int count = DRM_TIMESTAMP_MAXRETRIES;
-	u32 cur_vblank;
+	u32 vblank_count;
+	unsigned int seq;
 
 	if (WARN_ON(pipe >= dev->num_crtcs))
 		return 0;
 
-	/*
-	 * Vblank timestamps are read lockless. To ensure consistency the vblank
-	 * counter is rechecked and ordering is ensured using memory barriers.
-	 * This works like a seqlock. The write-side barriers are in store_vblank.
-	 */
 	do {
-		cur_vblank = vblank->count;
-		smp_rmb();
-		*vblanktime = vblanktimestamp(dev, pipe, cur_vblank);
-		smp_rmb();
-	} while (cur_vblank != vblank->count && --count > 0);
+		seq = read_seqbegin(&vblank->seqlock);
+		vblank_count = vblank->count;
+		*vblanktime = vblank->time;
+	} while (read_seqretry(&vblank->seqlock, seq));
 
-	return cur_vblank;
+	return vblank_count;
 }
 EXPORT_SYMBOL(drm_vblank_count_and_time);
 

drivers/gpu/drm/drm_mm.c

@@ -179,12 +179,14 @@ static void drm_mm_insert_helper(struct drm_mm_node *hole_node,
 int drm_mm_reserve_node(struct drm_mm *mm, struct drm_mm_node *node)
 {
 	struct drm_mm_node *hole;
-	u64 end = node->start + node->size;
+	u64 end;
 	u64 hole_start;
 	u64 hole_end;
 
 	BUG_ON(node == NULL);
 
+	end = node->start + node->size;
+
 	/* Find the relevant hole to add our node to */
 	drm_mm_for_each_hole(hole, mm, hole_start, hole_end) {
 		if (hole_start > node->start || hole_end < end)

drivers/gpu/drm/drm_modeset_lock.c

@@ -30,12 +30,12 @@
  *
  * As KMS moves toward more fine grained locking, and atomic ioctl where
  * userspace can indirectly control locking order, it becomes necessary
- * to use ww_mutex and acquire-contexts to avoid deadlocks.  But because
+ * to use &ww_mutex and acquire-contexts to avoid deadlocks.  But because
  * the locking is more distributed around the driver code, we want a bit
  * of extra utility/tracking out of our acquire-ctx.  This is provided
  * by drm_modeset_lock / drm_modeset_acquire_ctx.
  *
- * For basic principles of ww_mutex, see: Documentation/locking/ww-mutex-design.txt
+ * For basic principles of &ww_mutex, see: Documentation/locking/ww-mutex-design.txt
  *
  * The basic usage pattern is to:
  *
@@ -51,6 +51,13 @@
  *     ... do stuff ...
  *     drm_modeset_drop_locks(&ctx);
  *     drm_modeset_acquire_fini(&ctx);
+ *
+ *  On top of of these per-object locks using &ww_mutex there's also an overall
+ *  dev->mode_config.lock, for protecting everything else. Mostly this means
+ *  probe state of connectors, and preventing hotplug add/removal of connectors.
+ *
+ *  Finally there's a bunch of dedicated locks to protect drm core internal
+ *  lists and lookup data structures.
  */
 
 /**

drivers/gpu/drm/gma500/psb_intel_display.c

@@ -491,7 +491,6 @@ void psb_intel_crtc_init(struct drm_device *dev, int pipe,
 	struct drm_psb_private *dev_priv = dev->dev_private;
 	struct gma_crtc *gma_crtc;
 	int i;
-	uint16_t *r_base, *g_base, *b_base;
 
 	/* We allocate a extra array of drm_connector pointers
 	 * for fbdev after the crtc */
@@ -519,16 +518,10 @@ void psb_intel_crtc_init(struct drm_device *dev, int pipe,
 	gma_crtc->pipe = pipe;
 	gma_crtc->plane = pipe;
 
-	r_base = gma_crtc->base.gamma_store;
-	g_base = r_base + 256;
-	b_base = g_base + 256;
 	for (i = 0; i < 256; i++) {
 		gma_crtc->lut_r[i] = i;
 		gma_crtc->lut_g[i] = i;
 		gma_crtc->lut_b[i] = i;
-		r_base[i] = i << 8;
-		g_base[i] = i << 8;
-		b_base[i] = i << 8;
 
 		gma_crtc->lut_adj[i] = 0;
 	}

drivers/gpu/drm/hisilicon/kirin/kirin_drm_drv.c

@@ -173,7 +173,7 @@ static struct drm_driver kirin_drm_driver = {
 	.fops			= &kirin_drm_fops,
 	.set_busid		= drm_platform_set_busid,
 
-	.gem_free_object	= drm_gem_cma_free_object,
+	.gem_free_object_unlocked = drm_gem_cma_free_object,
 	.gem_vm_ops		= &drm_gem_cma_vm_ops,
 	.dumb_create		= kirin_gem_cma_dumb_create,
 	.dumb_map_offset	= drm_gem_cma_dumb_map_offset,

drivers/gpu/drm/i915/i915_drv.c

@@ -35,11 +35,9 @@
 #include "i915_trace.h"
 #include "intel_drv.h"
 
-#include <linux/apple-gmux.h>
 #include <linux/console.h>
 #include <linux/module.h>
 #include <linux/pm_runtime.h>
-#include <linux/vgaarb.h>
 #include <linux/vga_switcheroo.h>
 #include <drm/drm_crtc_helper.h>
 
@@ -1030,13 +1028,7 @@ static int i915_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	if (PCI_FUNC(pdev->devfn))
 		return -ENODEV;
 
-	/*
-	 * apple-gmux is needed on dual GPU MacBook Pro
-	 * to probe the panel if we're the inactive GPU.
-	 */
-	if (IS_ENABLED(CONFIG_VGA_ARB) && IS_ENABLED(CONFIG_VGA_SWITCHEROO) &&
-	    apple_gmux_present() && pdev != vga_default_device() &&
-	    !vga_switcheroo_handler_flags())
+	if (vga_switcheroo_client_probe_defer(pdev))
 		return -EPROBE_DEFER;
 
 	return drm_get_pci_dev(pdev, ent, &driver);

drivers/gpu/drm/mediatek/mtk_drm_drv.c

@@ -243,7 +243,7 @@ static struct drm_driver mtk_drm_driver = {
 	.enable_vblank = mtk_drm_crtc_enable_vblank,
 	.disable_vblank = mtk_drm_crtc_disable_vblank,
 
-	.gem_free_object = mtk_drm_gem_free_object,
+	.gem_free_object_unlocked = mtk_drm_gem_free_object,
 	.gem_vm_ops = &drm_gem_cma_vm_ops,
 	.dumb_create = mtk_drm_gem_dumb_create,
 	.dumb_map_offset = mtk_drm_gem_dumb_map_offset,

drivers/gpu/drm/msm/msm_fbdev.c

@@ -184,21 +184,7 @@ static int msm_fbdev_create(struct drm_fb_helper *helper,
 	return ret;
 }
 
-static void msm_crtc_fb_gamma_set(struct drm_crtc *crtc,
-		u16 red, u16 green, u16 blue, int regno)
-{
-	DBG("fbdev: set gamma");
-}
-
-static void msm_crtc_fb_gamma_get(struct drm_crtc *crtc,
-		u16 *red, u16 *green, u16 *blue, int regno)
-{
-	DBG("fbdev: get gamma");
-}
-
 static const struct drm_fb_helper_funcs msm_fb_helper_funcs = {
-	.gamma_set = msm_crtc_fb_gamma_set,
-	.gamma_get = msm_crtc_fb_gamma_get,
 	.fb_probe = msm_fbdev_create,
 };
 

drivers/gpu/drm/nouveau/nouveau_drm.c

@@ -22,13 +22,11 @@
  * Authors: Ben Skeggs
  */
 
-#include <linux/apple-gmux.h>
 #include <linux/console.h>
 #include <linux/delay.h>
 #include <linux/module.h>
 #include <linux/pci.h>
 #include <linux/pm_runtime.h>
-#include <linux/vgaarb.h>
 #include <linux/vga_switcheroo.h>
 
 #include "drmP.h"
@@ -315,13 +313,7 @@ static int nouveau_drm_probe(struct pci_dev *pdev,
 	bool boot = false;
 	int ret;
 
-	/*
-	 * apple-gmux is needed on dual GPU MacBook Pro
-	 * to probe the panel if we're the inactive GPU.
-	 */
-	if (IS_ENABLED(CONFIG_VGA_ARB) && IS_ENABLED(CONFIG_VGA_SWITCHEROO) &&
-	    apple_gmux_present() && pdev != vga_default_device() &&
-	    !vga_switcheroo_handler_flags())
+	if (vga_switcheroo_client_probe_defer(pdev))
 		return -EPROBE_DEFER;
 
 	/* remove conflicting drivers (vesafb, efifb etc) */

drivers/gpu/drm/qxl/qxl_display.c

@@ -730,7 +730,6 @@ static int qdev_crtc_init(struct drm_device *dev, int crtc_id)
 
 	drm_crtc_init(dev, &qxl_crtc->base, &qxl_crtc_funcs);
 	qxl_crtc->index = crtc_id;
-	drm_mode_crtc_set_gamma_size(&qxl_crtc->base, 256);
 	drm_crtc_helper_add(&qxl_crtc->base, &qxl_crtc_helper_funcs);
 	return 0;
 }

drivers/gpu/drm/radeon/radeon_drv.c

@@ -34,11 +34,9 @@
 #include "radeon_drv.h"
 
 #include <drm/drm_pciids.h>
-#include <linux/apple-gmux.h>
 #include <linux/console.h>
 #include <linux/module.h>
 #include <linux/pm_runtime.h>
-#include <linux/vgaarb.h>
 #include <linux/vga_switcheroo.h>
 #include <drm/drm_gem.h>
 
@@ -340,13 +338,7 @@ static int radeon_pci_probe(struct pci_dev *pdev,
 	if (ret == -EPROBE_DEFER)
 		return ret;
 
-	/*
-	 * apple-gmux is needed on dual GPU MacBook Pro
-	 * to probe the panel if we're the inactive GPU.
-	 */
-	if (IS_ENABLED(CONFIG_VGA_ARB) && IS_ENABLED(CONFIG_VGA_SWITCHEROO) &&
-	    apple_gmux_present() && pdev != vga_default_device() &&
-	    !vga_switcheroo_handler_flags())
+	if (vga_switcheroo_client_probe_defer(pdev))
 		return -EPROBE_DEFER;
 
 	/* Get rid of things like offb */

drivers/gpu/drm/vc4/vc4_bo.c

@@ -291,8 +291,6 @@ static void vc4_bo_cache_free_old(struct drm_device *dev)
 
 /* Called on the last userspace/kernel unreference of the BO.  Returns
  * it to the BO cache if possible, otherwise frees it.
- *
- * Note that this is called with the struct_mutex held.
  */
 void vc4_free_object(struct drm_gem_object *gem_bo)
 {

drivers/gpu/drm/vc4/vc4_drv.c

@@ -99,7 +99,7 @@ static struct drm_driver vc4_drm_driver = {
 #endif
 
 	.gem_create_object = vc4_create_object,
-	.gem_free_object = vc4_free_object,
+	.gem_free_object_unlocked = vc4_free_object,
 	.gem_vm_ops = &drm_gem_cma_vm_ops,
 
 	.prime_handle_to_fd = drm_gem_prime_handle_to_fd,

drivers/gpu/drm/vc4/vc4_gem.c

@@ -53,10 +53,8 @@ vc4_free_hang_state(struct drm_device *dev, struct vc4_hang_state *state)
 {
 	unsigned int i;
 
-	mutex_lock(&dev->struct_mutex);
 	for (i = 0; i < state->user_state.bo_count; i++)
-		drm_gem_object_unreference(state->bo[i]);
-	mutex_unlock(&dev->struct_mutex);
+		drm_gem_object_unreference_unlocked(state->bo[i]);
 
 	kfree(state);
 }
@@ -687,11 +685,9 @@ vc4_complete_exec(struct drm_device *dev, struct vc4_exec_info *exec)
 	struct vc4_dev *vc4 = to_vc4_dev(dev);
 	unsigned i;
 
-	/* Need the struct lock for drm_gem_object_unreference(). */
-	mutex_lock(&dev->struct_mutex);
 	if (exec->bo) {
 		for (i = 0; i < exec->bo_count; i++)
-			drm_gem_object_unreference(&exec->bo[i]->base);
+			drm_gem_object_unreference_unlocked(&exec->bo[i]->base);
 		kfree(exec->bo);
 	}
 
@@ -699,9 +695,8 @@ vc4_complete_exec(struct drm_device *dev, struct vc4_exec_info *exec)
 		struct vc4_bo *bo = list_first_entry(&exec->unref_list,
 						     struct vc4_bo, unref_head);
 		list_del(&bo->unref_head);
-		drm_gem_object_unreference(&bo->base.base);
+		drm_gem_object_unreference_unlocked(&bo->base.base);
 	}
-	mutex_unlock(&dev->struct_mutex);
 
 	mutex_lock(&vc4->power_lock);
 	if (--vc4->power_refcount == 0)

drivers/gpu/drm/vgem/vgem_drv.c

@@ -235,7 +235,7 @@ static const struct file_operations vgem_driver_fops = {
 
 static struct drm_driver vgem_driver = {
 	.driver_features		= DRIVER_GEM,
-	.gem_free_object		= vgem_gem_free_object,
+	.gem_free_object_unlocked	= vgem_gem_free_object,
 	.gem_vm_ops			= &vgem_gem_vm_ops,
 	.ioctls				= vgem_ioctls,
 	.fops				= &vgem_driver_fops,

drivers/gpu/drm/virtio/virtgpu_plane.c

@@ -63,11 +63,17 @@ static void virtio_gpu_plane_atomic_update(struct drm_plane *plane,
 {
 	struct drm_device *dev = plane->dev;
 	struct virtio_gpu_device *vgdev = dev->dev_private;
-	struct virtio_gpu_output *output = drm_crtc_to_virtio_gpu_output(plane->crtc);
+	struct virtio_gpu_output *output = NULL;
 	struct virtio_gpu_framebuffer *vgfb;
 	struct virtio_gpu_object *bo;
 	uint32_t handle;
 
+	if (plane->state->crtc)
+		output = drm_crtc_to_virtio_gpu_output(plane->state->crtc);
+	if (old_state->crtc)
+		output = drm_crtc_to_virtio_gpu_output(old_state->crtc);
+	WARN_ON(!output);
+
 	if (plane->state->fb) {
 		vgfb = to_virtio_gpu_framebuffer(plane->state->fb);
 		bo = gem_to_virtio_gpu_obj(vgfb->obj);

drivers/gpu/vga/vga_switcheroo.c

@@ -30,6 +30,7 @@
 
 #define pr_fmt(fmt) "vga_switcheroo: " fmt
 
+#include <linux/apple-gmux.h>
 #include <linux/console.h>
 #include <linux/debugfs.h>
 #include <linux/fb.h>
@@ -308,7 +309,8 @@ static int register_client(struct pci_dev *pdev,
  *
  * Register vga client (GPU). Enable vga_switcheroo if another GPU and a
  * handler have already registered. The power state of the client is assumed
- * to be ON.
+ * to be ON. Beforehand, vga_switcheroo_client_probe_defer() shall be called
+ * to ensure that all prerequisites are met.
  *
  * Return: 0 on success, -ENOMEM on memory allocation error.
  */
@@ -329,7 +331,8 @@ EXPORT_SYMBOL(vga_switcheroo_register_client);
  * @id: client identifier
  *
  * Register audio client (audio device on a GPU). The power state of the
- * client is assumed to be ON.
+ * client is assumed to be ON. Beforehand, vga_switcheroo_client_probe_defer()
+ * shall be called to ensure that all prerequisites are met.
  *
  * Return: 0 on success, -ENOMEM on memory allocation error.
  */
@@ -375,6 +378,33 @@ find_active_client(struct list_head *head)
 	return NULL;
 }
 
+/**
+ * vga_switcheroo_client_probe_defer() - whether to defer probing a given client
+ * @pdev: client pci device
+ *
+ * Determine whether any prerequisites are not fulfilled to probe a given
+ * client. Drivers shall invoke this early on in their ->probe callback
+ * and return %-EPROBE_DEFER if it evaluates to %true. Thou shalt not
+ * register the client ere thou hast called this.
+ *
+ * Return: %true if probing should be deferred, otherwise %false.
+ */
+bool vga_switcheroo_client_probe_defer(struct pci_dev *pdev)
+{
+	if ((pdev->class >> 16) == PCI_BASE_CLASS_DISPLAY) {
+		/*
+		 * apple-gmux is needed on pre-retina MacBook Pro
+		 * to probe the panel if pdev is the inactive GPU.
+		 */
+		if (apple_gmux_present() && pdev != vga_default_device() &&
+		    !vgasr_priv.handler_flags)
+			return true;
+	}
+
+	return false;
+}
+EXPORT_SYMBOL(vga_switcheroo_client_probe_defer);
+
 /**
  * vga_switcheroo_get_client_state() - obtain power state of a given client
  * @pdev: client pci device

include/drm/drmP.h

@@ -52,6 +52,7 @@
 #include <linux/poll.h>
 #include <linux/ratelimit.h>
 #include <linux/sched.h>
+#include <linux/seqlock.h>
 #include <linux/slab.h>
 #include <linux/types.h>
 #include <linux/vmalloc.h>
@@ -392,11 +393,6 @@ struct drm_master {
 	void *driver_priv;
 };
 
-/* Size of ringbuffer for vblank timestamps. Just double-buffer
- * in initial implementation.
- */
-#define DRM_VBLANKTIME_RBSIZE 2
-
 /* Flags and return codes for get_vblank_timestamp() driver function. */
 #define DRM_CALLED_FROM_VBLIRQ 1
 #define DRM_VBLANKTIME_SCANOUTPOS_METHOD (1 << 0)
@@ -725,10 +721,10 @@ struct drm_vblank_crtc {
 	wait_queue_head_t queue;	/**< VBLANK wait queue */
 	struct timer_list disable_timer;		/* delayed disable timer */
 
-	/* vblank counter, protected by dev->vblank_time_lock for writes */
-	u32 count;
-	/* vblank timestamps, protected by dev->vblank_time_lock for writes */
-	struct timeval time[DRM_VBLANKTIME_RBSIZE];
+	seqlock_t seqlock;		/* protects vblank count and time */
+
+	u32 count;			/* vblank counter */
+	struct timeval time;		/* vblank timestamp */
 
 	atomic_t refcount;		/* number of users of vblank interruptsper crtc */
 	u32 last;			/* protected by dev->vbl_lock, used */

include/linux/vga_switcheroo.h

@@ -165,6 +165,7 @@ int vga_switcheroo_unlock_ddc(struct pci_dev *pdev);
 
 int vga_switcheroo_process_delayed_switch(void);
 
+bool vga_switcheroo_client_probe_defer(struct pci_dev *pdev);
 enum vga_switcheroo_state vga_switcheroo_get_client_state(struct pci_dev *dev);
 
 void vga_switcheroo_set_dynamic_switch(struct pci_dev *pdev, enum vga_switcheroo_state dynamic);
@@ -188,6 +189,7 @@ static inline enum vga_switcheroo_handler_flags_t vga_switcheroo_handler_flags(v
 static inline int vga_switcheroo_lock_ddc(struct pci_dev *pdev) { return -ENODEV; }
 static inline int vga_switcheroo_unlock_ddc(struct pci_dev *pdev) { return -ENODEV; }
 static inline int vga_switcheroo_process_delayed_switch(void) { return 0; }
+static inline bool vga_switcheroo_client_probe_defer(struct pci_dev *pdev) { return false; }
 static inline enum vga_switcheroo_state vga_switcheroo_get_client_state(struct pci_dev *dev) { return VGA_SWITCHEROO_ON; }
 
 static inline void vga_switcheroo_set_dynamic_switch(struct pci_dev *pdev, enum vga_switcheroo_state dynamic) {}