drm/nouveau/pageflip: kick flip handling out of engsw and into fence

This is all very much a policy thing, and hence will not belong in SW
after the rework.

engsw now only handles receiving the event to say "can flip now" and makes
a callback to perform the actual work.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

17 files changed, 265 insertions, 114 deletions

diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index 9a86ae3e680..1855699a1ef 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -132,7 +132,7 @@ nouveau-y += nouveau_drm.o nouveau_compat.o \
 	     nouveau_mxm.o nouveau_agp.o \
 	     nouveau_abi16.o \
 	     nouveau_bios.o \
-             nv04_fence.o nv10_fence.o nv84_fence.o nvc0_fence.o \
+             nv04_fence.o nv10_fence.o nv50_fence.o nv84_fence.o nvc0_fence.o \
              nv04_software.o nv50_software.o nvc0_software.o \
              nv04_dac.o nv04_dfp.o nv04_tv.o nv17_tv.o nv17_tv_modes.o \
              nv04_crtc.o nv04_display.o nv04_cursor.o \
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
index 6535a999015..b99d976011d 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
@@ -27,6 +27,7 @@
 #include "nouveau_drv.h"
 #include <core/mm.h>
 #include <engine/fifo.h>
+#include "nouveau_software.h"
 
 static void nvc0_fifo_isr(struct drm_device *);
 
@@ -323,8 +324,11 @@ nvc0_fifo_page_flip(struct drm_device *dev, u32 chid)
 	spin_lock_irqsave(&dev_priv->channels.lock, flags);
 	if (likely(chid >= 0 && chid < priv->base.channels)) {
 		chan = dev_priv->channels.ptr[chid];
-		if (likely(chan))
-			ret = nouveau_finish_page_flip(chan, NULL);
+		if (likely(chan)) {
+			struct nouveau_software_chan *swch =
+				chan->engctx[NVOBJ_ENGINE_SW];
+			ret = swch->flip(swch->flip_data);
+		}
 	}
 	spin_unlock_irqrestore(&dev_priv->channels.lock, flags);
 	return ret;
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
index 461fbf62492..0b356f1b686 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
@@ -27,6 +27,7 @@
 #include "nouveau_drv.h"
 #include <core/mm.h>
 #include <engine/fifo.h>
+#include "nouveau_software.h"
 
 #define NVE0_FIFO_ENGINE_NUM 32
 
@@ -302,8 +303,11 @@ nve0_fifo_page_flip(struct drm_device *dev, u32 chid)
 	spin_lock_irqsave(&dev_priv->channels.lock, flags);
 	if (likely(chid >= 0 && chid < priv->base.channels)) {
 		chan = dev_priv->channels.ptr[chid];
-		if (likely(chan))
-			ret = nouveau_finish_page_flip(chan, NULL);
+		if (likely(chan)) {
+			struct nouveau_software_chan *swch =
+				chan->engctx[NVOBJ_ENGINE_SW];
+			ret = swch->flip(swch->flip_data);
+		}
 	}
 	spin_unlock_irqrestore(&dev_priv->channels.lock, flags);
 	return ret;
@@ -315,7 +319,7 @@ nve0_fifo_isr_subfifo_intr(struct drm_device *dev, int unit)
 	u32 stat = nv_rd32(dev, 0x040108 + (unit * 0x2000));
 	u32 addr = nv_rd32(dev, 0x0400c0 + (unit * 0x2000));
 	u32 data = nv_rd32(dev, 0x0400c4 + (unit * 0x2000));
-	u32 chid = nv_rd32(dev, 0x040120 + (unit * 0x2000)) & 0x7f;
+	u32 chid = nv_rd32(dev, 0x040120 + (unit * 0x2000)) & 0xfff;
 	u32 subc = (addr & 0x00070000);
 	u32 mthd = (addr & 0x00003ffc);
 	u32 show = stat;
diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c
index 44835c46495..e0a56b27788 100644
--- a/drivers/gpu/drm/nouveau/nouveau_display.c
+++ b/drivers/gpu/drm/nouveau/nouveau_display.c
@@ -435,7 +435,7 @@ nouveau_page_flip_emit(struct nouveau_channel *chan,
 		       struct nouveau_page_flip_state *s,
 		       struct nouveau_fence **pfence)
 {
-	struct nouveau_software_chan *swch = chan->engctx[NVOBJ_ENGINE_SW];
+	struct nouveau_fence_chan *fctx = chan->fence;
 	struct drm_nouveau_private *dev_priv = chan->dev->dev_private;
 	struct drm_device *dev = chan->dev;
 	unsigned long flags;
@@ -443,7 +443,7 @@ nouveau_page_flip_emit(struct nouveau_channel *chan,
 
 	/* Queue it to the pending list */
 	spin_lock_irqsave(&dev->event_lock, flags);
-	list_add_tail(&s->head, &swch->flip);
+	list_add_tail(&s->head, &fctx->flip);
 	spin_unlock_irqrestore(&dev->event_lock, flags);
 
 	/* Synchronize with the old framebuffer */
@@ -553,20 +553,20 @@ int
 nouveau_finish_page_flip(struct nouveau_channel *chan,
 			 struct nouveau_page_flip_state *ps)
 {
-	struct nouveau_software_chan *swch = chan->engctx[NVOBJ_ENGINE_SW];
+	struct nouveau_fence_chan *fctx = chan->fence;
 	struct drm_device *dev = chan->dev;
 	struct nouveau_page_flip_state *s;
 	unsigned long flags;
 
 	spin_lock_irqsave(&dev->event_lock, flags);
 
-	if (list_empty(&swch->flip)) {
+	if (list_empty(&fctx->flip)) {
 		NV_ERROR(dev, "Unexpected pageflip in channel %d.\n", chan->id);
 		spin_unlock_irqrestore(&dev->event_lock, flags);
 		return -EINVAL;
 	}
 
-	s = list_first_entry(&swch->flip, struct nouveau_page_flip_state, head);
+	s = list_first_entry(&fctx->flip, struct nouveau_page_flip_state, head);
 	if (s->event) {
 		struct drm_pending_vblank_event *e = s->event;
 		struct timeval now;
@@ -589,6 +589,25 @@ nouveau_finish_page_flip(struct nouveau_channel *chan,
 }
 
 int
+nouveau_flip_complete(void *data)
+{
+	struct nouveau_channel *chan = data;
+	struct drm_device *dev = chan->dev;
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_page_flip_state state;
+
+	if (!nouveau_finish_page_flip(chan, &state)) {
+		if (dev_priv->card_type < NV_50) {
+			nv_set_crtc_base(dev, state.crtc, state.offset +
+					 state.y * state.pitch +
+					 state.x * state.bpp / 8);
+		}
+	}
+
+	return 0;
+}
+
+int
 nouveau_display_dumb_create(struct drm_file *file_priv, struct drm_device *dev,
 			    struct drm_mode_create_dumb *args)
 {
diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.c b/drivers/gpu/drm/nouveau/nouveau_fence.c
index 9775458aff4..a1835d710f7 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.c
@@ -54,6 +54,7 @@ nouveau_fence_context_del(struct nouveau_fence_chan *fctx)
 void
 nouveau_fence_context_new(struct nouveau_fence_chan *fctx)
 {
+	INIT_LIST_HEAD(&fctx->flip);
 	INIT_LIST_HEAD(&fctx->pending);
 	spin_lock_init(&fctx->lock);
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.h b/drivers/gpu/drm/nouveau/nouveau_fence.h
index 690f46536a7..8ee65758f24 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.h
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.h
@@ -27,6 +27,8 @@ void nouveau_fence_update(struct nouveau_channel *);
 
 struct nouveau_fence_chan {
 	struct list_head pending;
+	struct list_head flip;
+
 	spinlock_t lock;
 	u32 sequence;
 };
@@ -49,8 +51,19 @@ void nouveau_fence_context_del(struct nouveau_fence_chan *);
 int nv04_fence_create(struct drm_device *dev);
 int nv04_fence_mthd(struct nouveau_channel *, u32, u32, u32);
 
-int nv10_fence_create(struct drm_device *dev);
+int  nv10_fence_emit(struct nouveau_fence *);
+int  nv17_fence_sync(struct nouveau_fence *, struct nouveau_channel *,
+		     struct nouveau_channel *);
+u32  nv10_fence_read(struct nouveau_channel *);
+void nv10_fence_context_del(struct nouveau_channel *);
+void nv10_fence_destroy(struct drm_device *);
+int  nv10_fence_create(struct drm_device *dev);
+
+int nv50_fence_create(struct drm_device *dev);
 int nv84_fence_create(struct drm_device *dev);
 int nvc0_fence_create(struct drm_device *dev);
+u64 nvc0_fence_crtc(struct nouveau_channel *, int crtc);
+
+int nouveau_flip_complete(void *chan);
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/nouveau_software.h b/drivers/gpu/drm/nouveau/nouveau_software.h
index 709e5ac680e..2105a9eef52 100644
--- a/drivers/gpu/drm/nouveau/nouveau_software.h
+++ b/drivers/gpu/drm/nouveau/nouveau_software.h
@@ -1,6 +1,8 @@
 #ifndef __NOUVEAU_SOFTWARE_H__
 #define __NOUVEAU_SOFTWARE_H__
 
+#include "nouveau_fence.h"
+
 struct nouveau_software_priv {
 	struct nouveau_exec_engine base;
 	struct list_head vblank;
@@ -8,7 +10,9 @@ struct nouveau_software_priv {
 };
 
 struct nouveau_software_chan {
-	struct list_head flip;
+	int (*flip)(void *data);
+	void *flip_data;
+
 	struct {
 		struct list_head list;
 		u32 channel;
@@ -20,10 +24,11 @@ struct nouveau_software_chan {
 };
 
 static inline void
-nouveau_software_context_new(struct nouveau_software_chan *pch)
+nouveau_software_context_new(struct nouveau_channel *chan,
+			     struct nouveau_software_chan *pch)
 {
-	INIT_LIST_HEAD(&pch->flip);
-	INIT_LIST_HEAD(&pch->vblank.list);
+	pch->flip = nouveau_flip_complete;
+	pch->flip_data = chan;
 }
 
 static inline void
@@ -51,6 +56,5 @@ nouveau_software_class(struct drm_device *dev)
 int nv04_software_create(struct drm_device *);
 int nv50_software_create(struct drm_device *);
 int nvc0_software_create(struct drm_device *);
-u64 nvc0_software_crtc(struct nouveau_channel *, int crtc);
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/nouveau_state.c b/drivers/gpu/drm/nouveau/nouveau_state.c
index 1cdabdc7908..5c2836fbf01 100644
--- a/drivers/gpu/drm/nouveau/nouveau_state.c
+++ b/drivers/gpu/drm/nouveau/nouveau_state.c
@@ -394,7 +394,7 @@ nouveau_card_init(struct drm_device *dev)
 		case NV_40:
 		case NV_50:
 			if (dev_priv->chipset < 0x84)
-				nv10_fence_create(dev);
+				nv50_fence_create(dev);
 			else
 				nv84_fence_create(dev);
 			break;
diff --git a/drivers/gpu/drm/nouveau/nv04_software.c b/drivers/gpu/drm/nouveau/nv04_software.c
index 06c316b386a..ceeb868c7c2 100644
--- a/drivers/gpu/drm/nouveau/nv04_software.c
+++ b/drivers/gpu/drm/nouveau/nv04_software.c
@@ -41,16 +41,8 @@ struct nv04_software_chan {
 static int
 mthd_flip(struct nouveau_channel *chan, u32 class, u32 mthd, u32 data)
 {
-
-	struct nouveau_page_flip_state state;
-
-	if (!nouveau_finish_page_flip(chan, &state)) {
-		nv_set_crtc_base(chan->dev, state.crtc, state.offset +
-				 state.y * state.pitch +
-				 state.x * state.bpp / 8);
-	}
-
-	return 0;
+	struct nv04_software_chan *pch = chan->engctx[NVOBJ_ENGINE_SW];
+	return pch->base.flip(pch->base.flip_data);
 }
 
 static int
@@ -62,7 +54,7 @@ nv04_software_context_new(struct nouveau_channel *chan, int engine)
 	if (!pch)
 		return -ENOMEM;
 
-	nouveau_software_context_new(&pch->base);
+	nouveau_software_context_new(chan, &pch->base);
 	chan->engctx[engine] = pch;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nv10_fence.c b/drivers/gpu/drm/nouveau/nv10_fence.c
index 4dac16a9c7f..dd7f17dd990 100644
--- a/drivers/gpu/drm/nouveau/nv10_fence.c
+++ b/drivers/gpu/drm/nouveau/nv10_fence.c
@@ -39,7 +39,7 @@ struct nv10_fence_priv {
 	u32 sequence;
 };
 
-static int
+int
 nv10_fence_emit(struct nouveau_fence *fence)
 {
 	struct nouveau_channel *chan = fence->channel;
@@ -60,7 +60,7 @@ nv10_fence_sync(struct nouveau_fence *fence,
 	return -ENODEV;
 }
 
-static int
+int
 nv17_fence_sync(struct nouveau_fence *fence,
 		struct nouveau_channel *prev, struct nouveau_channel *chan)
 {
@@ -100,13 +100,13 @@ nv17_fence_sync(struct nouveau_fence *fence,
 	return 0;
 }
 
-static u32
+u32
 nv10_fence_read(struct nouveau_channel *chan)
 {
 	return nvchan_rd32(chan, 0x0048);
 }
 
-static void
+void
 nv10_fence_context_del(struct nouveau_channel *chan)
 {
 	struct nv10_fence_chan *fctx = chan->fence;
@@ -148,7 +148,7 @@ nv10_fence_context_new(struct nouveau_channel *chan)
 	return ret;
 }
 
-static void
+void
 nv10_fence_destroy(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c
index 9c0bb20b0dc..83419a2daa0 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.c
+++ b/drivers/gpu/drm/nouveau/nv50_display.c
@@ -32,8 +32,8 @@
 #include "nouveau_fb.h"
 #include "nouveau_fbcon.h"
 #include <core/ramht.h>
-#include "nouveau_software.h"
 #include "drm_crtc_helper.h"
+#include "nouveau_fence.h"
 
 static void nv50_display_isr(struct drm_device *);
 static void nv50_display_bh(unsigned long);
@@ -446,7 +446,7 @@ nv50_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 			else
 				OUT_RING  (chan, chan->vram_handle);
 		} else {
-			u64 offset = nvc0_software_crtc(chan, nv_crtc->index);
+			u64 offset = nvc0_fence_crtc(chan, nv_crtc->index);
 			offset += dispc->sem.offset;
 			BEGIN_NVC0(chan, 0, 0x0010, 4);
 			OUT_RING  (chan, upper_32_bits(offset));
diff --git a/drivers/gpu/drm/nouveau/nv50_fence.c b/drivers/gpu/drm/nouveau/nv50_fence.c
new file mode 100644
index 00000000000..10aa04f26b8
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nv50_fence.c
@@ -0,0 +1,125 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * 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 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 COPYRIGHT HOLDER(S) OR AUTHOR(S) 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: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "drmP.h"
+#include "nouveau_drv.h"
+#include "nouveau_dma.h"
+#include <core/ramht.h>
+#include "nouveau_fence.h"
+#include "nv50_display.h"
+
+struct nv50_fence_chan {
+	struct nouveau_fence_chan base;
+};
+
+struct nv50_fence_priv {
+	struct nouveau_fence_priv base;
+	struct nouveau_bo *bo;
+	spinlock_t lock;
+	u32 sequence;
+};
+
+static int
+nv50_fence_context_new(struct nouveau_channel *chan)
+{
+	struct drm_nouveau_private *dev_priv = chan->dev->dev_private;
+	struct nv50_fence_priv *priv = dev_priv->fence.func;
+	struct nv50_fence_chan *fctx;
+	struct ttm_mem_reg *mem = &priv->bo->bo.mem;
+	struct nouveau_gpuobj *obj;
+	int ret = 0, i;
+
+	fctx = chan->fence = kzalloc(sizeof(*fctx), GFP_KERNEL);
+	if (!fctx)
+		return -ENOMEM;
+
+	nouveau_fence_context_new(&fctx->base);
+
+	ret = nouveau_gpuobj_dma_new(chan, NV_CLASS_DMA_FROM_MEMORY,
+				     mem->start * PAGE_SIZE, mem->size,
+				     NV_MEM_ACCESS_RW,
+				     NV_MEM_TARGET_VRAM, &obj);
+	if (!ret) {
+		ret = nouveau_ramht_insert(chan, NvSema, obj);
+		nouveau_gpuobj_ref(NULL, &obj);
+	}
+
+	/* dma objects for display sync channel semaphore blocks */
+	for (i = 0; i < chan->dev->mode_config.num_crtc; i++) {
+		struct nv50_display *pdisp = nv50_display(chan->dev);
+		struct nv50_display_crtc *dispc = &pdisp->crtc[i];
+		struct nouveau_gpuobj *obj = NULL;
+
+		ret = nouveau_gpuobj_dma_new(chan, NV_CLASS_DMA_IN_MEMORY,
+					     dispc->sem.bo->bo.offset, 0x1000,
+					     NV_MEM_ACCESS_RW,
+					     NV_MEM_TARGET_VRAM, &obj);
+		if (ret)
+			break;
+
+		ret = nouveau_ramht_insert(chan, NvEvoSema0 + i, obj);
+		nouveau_gpuobj_ref(NULL, &obj);
+	}
+
+	if (ret)
+		nv10_fence_context_del(chan);
+	return ret;
+}
+
+int
+nv50_fence_create(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nv50_fence_priv *priv;
+	int ret = 0;
+
+	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	priv->base.dtor = nv10_fence_destroy;
+	priv->base.context_new = nv50_fence_context_new;
+	priv->base.context_del = nv10_fence_context_del;
+	priv->base.emit = nv10_fence_emit;
+	priv->base.read = nv10_fence_read;
+	priv->base.sync = nv17_fence_sync;
+	dev_priv->fence.func = &priv->base;
+	spin_lock_init(&priv->lock);
+
+	ret = nouveau_bo_new(dev, 4096, 0x1000, TTM_PL_FLAG_VRAM,
+			     0, 0x0000, NULL, &priv->bo);
+	if (!ret) {
+		ret = nouveau_bo_pin(priv->bo, TTM_PL_FLAG_VRAM);
+		if (!ret)
+			ret = nouveau_bo_map(priv->bo);
+		if (ret)
+			nouveau_bo_ref(NULL, &priv->bo);
+	}
+
+	if (ret == 0)
+		nouveau_bo_wr32(priv->bo, 0x000, 0x00000000);
+	else
+		nv10_fence_destroy(dev);
+	return ret;
+}
diff --git a/drivers/gpu/drm/nouveau/nv50_software.c b/drivers/gpu/drm/nouveau/nv50_software.c
index a1c06d44eeb..7c9dbe862c4 100644
--- a/drivers/gpu/drm/nouveau/nv50_software.c
+++ b/drivers/gpu/drm/nouveau/nv50_software.c
@@ -88,45 +88,23 @@ mthd_vblsem_release(struct nouveau_channel *chan, u32 class, u32 mthd, u32 data)
 static int
 mthd_flip(struct nouveau_channel *chan, u32 class, u32 mthd, u32 data)
 {
-	nouveau_finish_page_flip(chan, NULL);
-	return 0;
+	struct nv50_software_chan *pch = chan->engctx[NVOBJ_ENGINE_SW];
+	return pch->base.flip(pch->base.flip_data);
 }
 
 static int
 nv50_software_context_new(struct nouveau_channel *chan, int engine)
 {
-	struct nv50_software_priv *psw = nv_engine(chan->dev, NVOBJ_ENGINE_SW);
-	struct nv50_display *pdisp = nv50_display(chan->dev);
 	struct nv50_software_chan *pch;
-	int ret = 0, i;
 
 	pch = kzalloc(sizeof(*pch), GFP_KERNEL);
 	if (!pch)
 		return -ENOMEM;
 
-	nouveau_software_context_new(&pch->base);
+	nouveau_software_context_new(chan, &pch->base);
 	pch->base.vblank.channel = chan->ramin->addr >> 12;
 	chan->engctx[engine] = pch;
-
-	/* dma objects for display sync channel semaphore blocks */
-	for (i = 0; i < chan->dev->mode_config.num_crtc; i++) {
-		struct nv50_display_crtc *dispc = &pdisp->crtc[i];
-		struct nouveau_gpuobj *obj = NULL;
-
-		ret = nouveau_gpuobj_dma_new(chan, NV_CLASS_DMA_IN_MEMORY,
-					     dispc->sem.bo->bo.offset, 0x1000,
-					     NV_MEM_ACCESS_RW,
-					     NV_MEM_TARGET_VRAM, &obj);
-		if (ret)
-			break;
-
-		ret = nouveau_ramht_insert(chan, NvEvoSema0 + i, obj);
-		nouveau_gpuobj_ref(NULL, &obj);
-	}
-
-	if (ret)
-		psw->base.base.context_del(chan, engine);
-	return ret;
+	return 0;
 }
 
 static void
diff --git a/drivers/gpu/drm/nouveau/nv84_fence.c b/drivers/gpu/drm/nouveau/nv84_fence.c
index a1812cab19d..5ef87edb878 100644
--- a/drivers/gpu/drm/nouveau/nv84_fence.c
+++ b/drivers/gpu/drm/nouveau/nv84_fence.c
@@ -28,6 +28,7 @@
 #include <engine/fifo.h>
 #include <core/ramht.h>
 #include "nouveau_fence.h"
+#include "nv50_display.h"
 
 struct nv84_fence_chan {
 	struct nouveau_fence_chan base;
@@ -99,7 +100,7 @@ nv84_fence_context_new(struct nouveau_channel *chan)
 	struct nv84_fence_priv *priv = dev_priv->fence.func;
 	struct nv84_fence_chan *fctx;
 	struct nouveau_gpuobj *obj;
-	int ret;
+	int ret, i;
 
 	fctx = chan->fence = kzalloc(sizeof(*fctx), GFP_KERNEL);
 	if (!fctx)
@@ -117,6 +118,23 @@ nv84_fence_context_new(struct nouveau_channel *chan)
 		nv_wo32(priv->mem, chan->id * 16, 0x00000000);
 	}
 
+	/* dma objects for display sync channel semaphore blocks */
+	for (i = 0; i < chan->dev->mode_config.num_crtc; i++) {
+		struct nv50_display *pdisp = nv50_display(chan->dev);
+		struct nv50_display_crtc *dispc = &pdisp->crtc[i];
+		struct nouveau_gpuobj *obj = NULL;
+
+		ret = nouveau_gpuobj_dma_new(chan, NV_CLASS_DMA_IN_MEMORY,
+					     dispc->sem.bo->bo.offset, 0x1000,
+					     NV_MEM_ACCESS_RW,
+					     NV_MEM_TARGET_VRAM, &obj);
+		if (ret)
+			break;
+
+		ret = nouveau_ramht_insert(chan, NvEvoSema0 + i, obj);
+		nouveau_gpuobj_ref(NULL, &obj);
+	}
+
 	if (ret)
 		nv84_fence_context_del(chan);
 	return ret;
diff --git a/drivers/gpu/drm/nouveau/nvc0_fence.c b/drivers/gpu/drm/nouveau/nvc0_fence.c
index d53ae32caea..779c5ff4ed7 100644
--- a/drivers/gpu/drm/nouveau/nvc0_fence.c
+++ b/drivers/gpu/drm/nouveau/nvc0_fence.c
@@ -28,6 +28,7 @@
 #include <engine/fifo.h>
 #include <core/ramht.h>
 #include "nouveau_fence.h"
+#include "nv50_display.h"
 
 struct nvc0_fence_priv {
 	struct nouveau_fence_priv base;
@@ -38,8 +39,16 @@ struct nvc0_fence_priv {
 struct nvc0_fence_chan {
 	struct nouveau_fence_chan base;
 	struct nouveau_vma vma;
+	struct nouveau_vma dispc_vma[4];
 };
 
+u64
+nvc0_fence_crtc(struct nouveau_channel *chan, int crtc)
+{
+	struct nvc0_fence_chan *fctx = chan->fence;
+	return fctx->dispc_vma[crtc].offset;
+}
+
 static int
 nvc0_fence_emit(struct nouveau_fence *fence)
 {
@@ -94,9 +103,25 @@ nvc0_fence_read(struct nouveau_channel *chan)
 static void
 nvc0_fence_context_del(struct nouveau_channel *chan)
 {
-	struct drm_nouveau_private *dev_priv = chan->dev->dev_private;
+	struct drm_device *dev = chan->dev;
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nvc0_fence_priv *priv = dev_priv->fence.func;
 	struct nvc0_fence_chan *fctx = chan->fence;
+	int i;
+
+	if (dev_priv->card_type >= NV_D0) {
+		for (i = 0; i < dev->mode_config.num_crtc; i++) {
+			struct nouveau_bo *bo = nvd0_display_crtc_sema(dev, i);
+			nouveau_bo_vma_del(bo, &fctx->dispc_vma[i]);
+		}
+	} else
+	if (dev_priv->card_type >= NV_50) {
+		struct nv50_display *disp = nv50_display(dev);
+		for (i = 0; i < dev->mode_config.num_crtc; i++) {
+			struct nv50_display_crtc *dispc = &disp->crtc[i];
+			nouveau_bo_vma_del(dispc->sem.bo, &fctx->dispc_vma[i]);
+		}
+	}
 
 	nouveau_bo_vma_del(priv->bo, &fctx->vma);
 	nouveau_fence_context_del(&fctx->base);
@@ -107,10 +132,11 @@ nvc0_fence_context_del(struct nouveau_channel *chan)
 static int
 nvc0_fence_context_new(struct nouveau_channel *chan)
 {
-	struct drm_nouveau_private *dev_priv = chan->dev->dev_private;
+	struct drm_device *dev = chan->dev;
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nvc0_fence_priv *priv = dev_priv->fence.func;
 	struct nvc0_fence_chan *fctx;
-	int ret;
+	int ret, i;
 
 	fctx = chan->fence = kzalloc(sizeof(*fctx), GFP_KERNEL);
 	if (!fctx)
@@ -122,6 +148,17 @@ nvc0_fence_context_new(struct nouveau_channel *chan)
 	if (ret)
 		nvc0_fence_context_del(chan);
 
+	/* map display semaphore buffers into channel's vm */
+	for (i = 0; !ret && i < dev->mode_config.num_crtc; i++) {
+		struct nouveau_bo *bo;
+		if (dev_priv->card_type >= NV_D0)
+			bo = nvd0_display_crtc_sema(dev, i);
+		else
+			bo = nv50_display(dev)->crtc[i].sem.bo;
+
+		ret = nouveau_bo_vma_add(bo, chan->vm, &fctx->dispc_vma[i]);
+	}
+
 	nouveau_bo_wr32(priv->bo, chan->id * 16/4, 0x00000000);
 	return ret;
 }
diff --git a/drivers/gpu/drm/nouveau/nvc0_software.c b/drivers/gpu/drm/nouveau/nvc0_software.c
index 22527682ec5..eaaa5768f4f 100644
--- a/drivers/gpu/drm/nouveau/nvc0_software.c
+++ b/drivers/gpu/drm/nouveau/nvc0_software.c
@@ -36,70 +36,26 @@ struct nvc0_software_priv {
 
 struct nvc0_software_chan {
 	struct nouveau_software_chan base;
-	struct nouveau_vma dispc_vma[4];
 };
 
-u64
-nvc0_software_crtc(struct nouveau_channel *chan, int crtc)
-{
-	struct nvc0_software_chan *pch = chan->engctx[NVOBJ_ENGINE_SW];
-	return pch->dispc_vma[crtc].offset;
-}
-
 static int
 nvc0_software_context_new(struct nouveau_channel *chan, int engine)
 {
-	struct drm_device *dev = chan->dev;
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nvc0_software_priv *psw = nv_engine(dev, NVOBJ_ENGINE_SW);
 	struct nvc0_software_chan *pch;
-	int ret = 0, i;
 
 	pch = kzalloc(sizeof(*pch), GFP_KERNEL);
 	if (!pch)
 		return -ENOMEM;
 
-	nouveau_software_context_new(&pch->base);
+	nouveau_software_context_new(chan, &pch->base);
 	chan->engctx[engine] = pch;
-
-	/* map display semaphore buffers into channel's vm */
-	for (i = 0; !ret && i < dev->mode_config.num_crtc; i++) {
-		struct nouveau_bo *bo;
-		if (dev_priv->card_type >= NV_D0)
-			bo = nvd0_display_crtc_sema(dev, i);
-		else
-			bo = nv50_display(dev)->crtc[i].sem.bo;
-
-		ret = nouveau_bo_vma_add(bo, chan->vm, &pch->dispc_vma[i]);
-	}
-
-	if (ret)
-		psw->base.base.context_del(chan, engine);
-	return ret;
+	return 0;
 }
 
 static void
 nvc0_software_context_del(struct nouveau_channel *chan, int engine)
 {
-	struct drm_device *dev = chan->dev;
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nvc0_software_chan *pch = chan->engctx[engine];
-	int i;
-
-	if (dev_priv->card_type >= NV_D0) {
-		for (i = 0; i < dev->mode_config.num_crtc; i++) {
-			struct nouveau_bo *bo = nvd0_display_crtc_sema(dev, i);
-			nouveau_bo_vma_del(bo, &pch->dispc_vma[i]);
-		}
-	} else
-	if (dev_priv->card_type >= NV_50) {
-		struct nv50_display *disp = nv50_display(dev);
-		for (i = 0; i < dev->mode_config.num_crtc; i++) {
-			struct nv50_display_crtc *dispc = &disp->crtc[i];
-			nouveau_bo_vma_del(dispc->sem.bo, &pch->dispc_vma[i]);
-		}
-	}
-
 	chan->engctx[engine] = NULL;
 	kfree(pch);
 }
diff --git a/drivers/gpu/drm/nouveau/nvd0_display.c b/drivers/gpu/drm/nouveau/nvd0_display.c
index 715359ef921..2da4927b5e0 100644
--- a/drivers/gpu/drm/nouveau/nvd0_display.c
+++ b/drivers/gpu/drm/nouveau/nvd0_display.c
@@ -33,7 +33,7 @@
 #include "nouveau_crtc.h"
 #include "nouveau_dma.h"
 #include "nouveau_fb.h"
-#include "nouveau_software.h"
+#include "nouveau_fence.h"
 #include "nv50_display.h"
 
 #define EVO_DMA_NR 9
@@ -300,7 +300,7 @@ nvd0_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 			return ret;
 
 
-		offset  = nvc0_software_crtc(chan, nv_crtc->index);
+		offset  = nvc0_fence_crtc(chan, nv_crtc->index);
 		offset += evo->sem.offset;
 
 		BEGIN_NVC0(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4);