Merge branch 'drm-nouveau-next' of git://anongit.freedesktop.org/git/nouveau/linux-2.6 into drm-core-next

Ben was distracted:

"Apologies for being really late with this, feel free to bash me in the
future so I remember on time!

Overview:
- improvements to reclocking (especially memory) on nva3+
- kepler accel support (if you have blob ucode)
- better inter-channel synchronisation on nv84+
- async ttm buffer moves on nv84+ (earlier cards don't have a non-PGRAPH
engine that's useful)"

* 'drm-nouveau-next' of git://anongit.freedesktop.org/git/nouveau/linux-2.6: (60 commits)
  drm/nouveau/nvd9: Fix GPIO initialisation sequence.
  drm/nouveau: Unregister switcheroo client on exit
  drm/nouveau: Check dsm on switcheroo unregister
  drm/nouveau: fix a minor annoyance in an output string
  drm/nouveau: turn a BUG into a WARN
  drm/nv50: decode PGRAPH DATA_ERROR = 0x24
  drm/nouveau/disp: fix dithering not being enabled on some eDP macbooks
  drm/nvd9/copy: initialise copy engine, seems to work like nvc0
  drm/nvc0/ttm: use copy engines for async buffer moves
  drm/nva3/ttm: use copy engine for async buffer moves
  drm/nv98/ttm: add in a (disabled) crypto engine buffer copy method
  drm/nv84/ttm: use crypto engine for async buffer copies
  drm/nouveau/ttm: untangle code to support accelerated buffer moves
  drm/nouveau/fbcon: use fence for sync, rather than notifier
  drm/nv98/crypt: non-stub implementation of the engine hooks
  drm/nouveau/fifo: turn all fifo modules into engine modules
  drm/nv50/graph: remove ability to do interrupt-driven context switching
  drm/nv50: remove manual context unload on context destruction
  drm/nv50: remove execution engine context saves on suspend
  drm/nv50/fifo: use hardware channel kickoff functionality
  ...

76 files changed, 9520 insertions, 3026 deletions

diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index 01f13351a47..fe5267d06ab 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -16,10 +16,13 @@ nouveau-y := nouveau_drv.o nouveau_state.o nouveau_channel.o nouveau_mem.o \
              nv04_mc.o nv40_mc.o nv50_mc.o \
              nv04_fb.o nv10_fb.o nv20_fb.o nv30_fb.o nv40_fb.o \
              nv50_fb.o nvc0_fb.o \
-             nv04_fifo.o nv10_fifo.o nv40_fifo.o nv50_fifo.o nvc0_fifo.o \
+             nv04_fifo.o nv10_fifo.o nv17_fifo.o nv40_fifo.o nv50_fifo.o \
+             nv84_fifo.o nvc0_fifo.o nve0_fifo.o \
+             nv04_fence.o nv10_fence.o nv84_fence.o nvc0_fence.o \
+             nv04_software.o nv50_software.o nvc0_software.o \
              nv04_graph.o nv10_graph.o nv20_graph.o \
-             nv40_graph.o nv50_graph.o nvc0_graph.o \
-             nv40_grctx.o nv50_grctx.o nvc0_grctx.o \
+             nv40_graph.o nv50_graph.o nvc0_graph.o nve0_graph.o \
+             nv40_grctx.o nv50_grctx.o nvc0_grctx.o nve0_grctx.o \
              nv84_crypt.o nv98_crypt.o \
              nva3_copy.o nvc0_copy.o \
              nv31_mpeg.o nv50_mpeg.o \
diff --git a/drivers/gpu/drm/nouveau/nouveau_acpi.c b/drivers/gpu/drm/nouveau/nouveau_acpi.c
index 284bd25d5d2..fc841e87b34 100644
--- a/drivers/gpu/drm/nouveau/nouveau_acpi.c
+++ b/drivers/gpu/drm/nouveau/nouveau_acpi.c
@@ -338,7 +338,8 @@ void nouveau_switcheroo_optimus_dsm(void)
 
 void nouveau_unregister_dsm_handler(void)
 {
-	vga_switcheroo_unregister_handler();
+	if (nouveau_dsm_priv.optimus_detected || nouveau_dsm_priv.dsm_detected)
+		vga_switcheroo_unregister_handler();
 }
 
 /* retrieve the ROM in 4k blocks */
diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.c b/drivers/gpu/drm/nouveau/nouveau_bios.c
index 0be4a815e70..2f11e16a81a 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bios.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bios.c
@@ -30,6 +30,7 @@
 #include "nouveau_gpio.h"
 
 #include <linux/io-mapping.h>
+#include <linux/firmware.h>
 
 /* these defines are made up */
 #define NV_CIO_CRE_44_HEADA 0x0
@@ -195,35 +196,24 @@ static void
 bios_shadow_acpi(struct nvbios *bios)
 {
 	struct pci_dev *pdev = bios->dev->pdev;
-	int ptr, len, ret;
-	u8 data[3];
+	int cnt = 65536 / ROM_BIOS_PAGE;
+	int ret;
 
 	if (!nouveau_acpi_rom_supported(pdev))
 		return;
 
-	ret = nouveau_acpi_get_bios_chunk(data, 0, sizeof(data));
-	if (ret != sizeof(data))
-		return;
-
-	bios->length = min(data[2] * 512, 65536);
-	bios->data = kmalloc(bios->length, GFP_KERNEL);
+	bios->data = kmalloc(cnt * ROM_BIOS_PAGE, GFP_KERNEL);
 	if (!bios->data)
 		return;
 
-	len = bios->length;
-	ptr = 0;
-	while (len) {
-		int size = (len > ROM_BIOS_PAGE) ? ROM_BIOS_PAGE : len;
-
-		ret = nouveau_acpi_get_bios_chunk(bios->data, ptr, size);
-		if (ret != size) {
-			kfree(bios->data);
-			bios->data = NULL;
+	bios->length = 0;
+	while (cnt--) {
+		ret = nouveau_acpi_get_bios_chunk(bios->data, bios->length,
+						  ROM_BIOS_PAGE);
+		if (ret != ROM_BIOS_PAGE)
 			return;
-		}
 
-		len -= size;
-		ptr += size;
+		bios->length += ROM_BIOS_PAGE;
 	}
 }
 
@@ -249,8 +239,12 @@ bios_shadow(struct drm_device *dev)
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nvbios *bios = &dev_priv->vbios;
 	struct methods *mthd, *best;
+	const struct firmware *fw;
+	char fname[32];
+	int ret;
 
 	if (nouveau_vbios) {
+		/* try to match one of the built-in methods */
 		mthd = shadow_methods;
 		do {
 			if (strcasecmp(nouveau_vbios, mthd->desc))
@@ -263,6 +257,22 @@ bios_shadow(struct drm_device *dev)
 				return true;
 		} while ((++mthd)->shadow);
 
+		/* attempt to load firmware image */
+		snprintf(fname, sizeof(fname), "nouveau/%s", nouveau_vbios);
+		ret = request_firmware(&fw, fname, &dev->pdev->dev);
+		if (ret == 0) {
+			bios->length = fw->size;
+			bios->data   = kmemdup(fw->data, fw->size, GFP_KERNEL);
+			release_firmware(fw);
+
+			NV_INFO(dev, "VBIOS image: %s\n", nouveau_vbios);
+			if (score_vbios(bios, 1))
+				return true;
+
+			kfree(bios->data);
+			bios->data = NULL;
+		}
+
 		NV_ERROR(dev, "VBIOS source \'%s\' invalid\n", nouveau_vbios);
 	}
 
@@ -273,6 +283,7 @@ bios_shadow(struct drm_device *dev)
 		mthd->score = score_vbios(bios, mthd->rw);
 		mthd->size = bios->length;
 		mthd->data = bios->data;
+		bios->data = NULL;
 	} while (mthd->score != 3 && (++mthd)->shadow);
 
 	mthd = shadow_methods;
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 4435e115b92..7f80ed52356 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -35,6 +35,8 @@
 #include "nouveau_dma.h"
 #include "nouveau_mm.h"
 #include "nouveau_vm.h"
+#include "nouveau_fence.h"
+#include "nouveau_ramht.h"
 
 #include <linux/log2.h>
 #include <linux/slab.h>
@@ -478,7 +480,7 @@ nouveau_bo_move_accel_cleanup(struct nouveau_channel *chan,
 	struct nouveau_fence *fence = NULL;
 	int ret;
 
-	ret = nouveau_fence_new(chan, &fence, true);
+	ret = nouveau_fence_new(chan, &fence);
 	if (ret)
 		return ret;
 
@@ -489,6 +491,76 @@ nouveau_bo_move_accel_cleanup(struct nouveau_channel *chan,
 }
 
 static int
+nve0_bo_move_copy(struct nouveau_channel *chan, struct ttm_buffer_object *bo,
+		  struct ttm_mem_reg *old_mem, struct ttm_mem_reg *new_mem)
+{
+	struct nouveau_mem *node = old_mem->mm_node;
+	int ret = RING_SPACE(chan, 10);
+	if (ret == 0) {
+		BEGIN_NVC0(chan, NvSubCopy, 0x0400, 8);
+		OUT_RING  (chan, upper_32_bits(node->vma[0].offset));
+		OUT_RING  (chan, lower_32_bits(node->vma[0].offset));
+		OUT_RING  (chan, upper_32_bits(node->vma[1].offset));
+		OUT_RING  (chan, lower_32_bits(node->vma[1].offset));
+		OUT_RING  (chan, PAGE_SIZE);
+		OUT_RING  (chan, PAGE_SIZE);
+		OUT_RING  (chan, PAGE_SIZE);
+		OUT_RING  (chan, new_mem->num_pages);
+		BEGIN_IMC0(chan, NvSubCopy, 0x0300, 0x0386);
+	}
+	return ret;
+}
+
+static int
+nvc0_bo_move_init(struct nouveau_channel *chan, u32 handle)
+{
+	int ret = RING_SPACE(chan, 2);
+	if (ret == 0) {
+		BEGIN_NVC0(chan, NvSubCopy, 0x0000, 1);
+		OUT_RING  (chan, handle);
+	}
+	return ret;
+}
+
+static int
+nvc0_bo_move_copy(struct nouveau_channel *chan, struct ttm_buffer_object *bo,
+		  struct ttm_mem_reg *old_mem, struct ttm_mem_reg *new_mem)
+{
+	struct nouveau_mem *node = old_mem->mm_node;
+	u64 src_offset = node->vma[0].offset;
+	u64 dst_offset = node->vma[1].offset;
+	u32 page_count = new_mem->num_pages;
+	int ret;
+
+	page_count = new_mem->num_pages;
+	while (page_count) {
+		int line_count = (page_count > 8191) ? 8191 : page_count;
+
+		ret = RING_SPACE(chan, 11);
+		if (ret)
+			return ret;
+
+		BEGIN_NVC0(chan, NvSubCopy, 0x030c, 8);
+		OUT_RING  (chan, upper_32_bits(src_offset));
+		OUT_RING  (chan, lower_32_bits(src_offset));
+		OUT_RING  (chan, upper_32_bits(dst_offset));
+		OUT_RING  (chan, lower_32_bits(dst_offset));
+		OUT_RING  (chan, PAGE_SIZE);
+		OUT_RING  (chan, PAGE_SIZE);
+		OUT_RING  (chan, PAGE_SIZE);
+		OUT_RING  (chan, line_count);
+		BEGIN_NVC0(chan, NvSubCopy, 0x0300, 1);
+		OUT_RING  (chan, 0x00000110);
+
+		page_count -= line_count;
+		src_offset += (PAGE_SIZE * line_count);
+		dst_offset += (PAGE_SIZE * line_count);
+	}
+
+	return 0;
+}
+
+static int
 nvc0_bo_move_m2mf(struct nouveau_channel *chan, struct ttm_buffer_object *bo,
 		  struct ttm_mem_reg *old_mem, struct ttm_mem_reg *new_mem)
 {
@@ -506,17 +578,17 @@ nvc0_bo_move_m2mf(struct nouveau_channel *chan, struct ttm_buffer_object *bo,
 		if (ret)
 			return ret;
 
-		BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0238, 2);
+		BEGIN_NVC0(chan, NvSubCopy, 0x0238, 2);
 		OUT_RING  (chan, upper_32_bits(dst_offset));
 		OUT_RING  (chan, lower_32_bits(dst_offset));
-		BEGIN_NVC0(chan, 2, NvSubM2MF, 0x030c, 6);
+		BEGIN_NVC0(chan, NvSubCopy, 0x030c, 6);
 		OUT_RING  (chan, upper_32_bits(src_offset));
 		OUT_RING  (chan, lower_32_bits(src_offset));
 		OUT_RING  (chan, PAGE_SIZE); /* src_pitch */
 		OUT_RING  (chan, PAGE_SIZE); /* dst_pitch */
 		OUT_RING  (chan, PAGE_SIZE); /* line_length */
 		OUT_RING  (chan, line_count);
-		BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0300, 1);
+		BEGIN_NVC0(chan, NvSubCopy, 0x0300, 1);
 		OUT_RING  (chan, 0x00100110);
 
 		page_count -= line_count;
@@ -528,6 +600,102 @@ nvc0_bo_move_m2mf(struct nouveau_channel *chan, struct ttm_buffer_object *bo,
 }
 
 static int
+nva3_bo_move_copy(struct nouveau_channel *chan, struct ttm_buffer_object *bo,
+		  struct ttm_mem_reg *old_mem, struct ttm_mem_reg *new_mem)
+{
+	struct nouveau_mem *node = old_mem->mm_node;
+	u64 src_offset = node->vma[0].offset;
+	u64 dst_offset = node->vma[1].offset;
+	u32 page_count = new_mem->num_pages;
+	int ret;
+
+	page_count = new_mem->num_pages;
+	while (page_count) {
+		int line_count = (page_count > 8191) ? 8191 : page_count;
+
+		ret = RING_SPACE(chan, 11);
+		if (ret)
+			return ret;
+
+		BEGIN_NV04(chan, NvSubCopy, 0x030c, 8);
+		OUT_RING  (chan, upper_32_bits(src_offset));
+		OUT_RING  (chan, lower_32_bits(src_offset));
+		OUT_RING  (chan, upper_32_bits(dst_offset));
+		OUT_RING  (chan, lower_32_bits(dst_offset));
+		OUT_RING  (chan, PAGE_SIZE);
+		OUT_RING  (chan, PAGE_SIZE);
+		OUT_RING  (chan, PAGE_SIZE);
+		OUT_RING  (chan, line_count);
+		BEGIN_NV04(chan, NvSubCopy, 0x0300, 1);
+		OUT_RING  (chan, 0x00000110);
+
+		page_count -= line_count;
+		src_offset += (PAGE_SIZE * line_count);
+		dst_offset += (PAGE_SIZE * line_count);
+	}
+
+	return 0;
+}
+
+static int
+nv98_bo_move_exec(struct nouveau_channel *chan, struct ttm_buffer_object *bo,
+		  struct ttm_mem_reg *old_mem, struct ttm_mem_reg *new_mem)
+{
+	struct nouveau_mem *node = old_mem->mm_node;
+	int ret = RING_SPACE(chan, 7);
+	if (ret == 0) {
+		BEGIN_NV04(chan, NvSubCopy, 0x0320, 6);
+		OUT_RING  (chan, upper_32_bits(node->vma[0].offset));
+		OUT_RING  (chan, lower_32_bits(node->vma[0].offset));
+		OUT_RING  (chan, upper_32_bits(node->vma[1].offset));
+		OUT_RING  (chan, lower_32_bits(node->vma[1].offset));
+		OUT_RING  (chan, 0x00000000 /* COPY */);
+		OUT_RING  (chan, new_mem->num_pages << PAGE_SHIFT);
+	}
+	return ret;
+}
+
+static int
+nv84_bo_move_exec(struct nouveau_channel *chan, struct ttm_buffer_object *bo,
+		  struct ttm_mem_reg *old_mem, struct ttm_mem_reg *new_mem)
+{
+	struct nouveau_mem *node = old_mem->mm_node;
+	int ret = RING_SPACE(chan, 7);
+	if (ret == 0) {
+		BEGIN_NV04(chan, NvSubCopy, 0x0304, 6);
+		OUT_RING  (chan, new_mem->num_pages << PAGE_SHIFT);
+		OUT_RING  (chan, upper_32_bits(node->vma[0].offset));
+		OUT_RING  (chan, lower_32_bits(node->vma[0].offset));
+		OUT_RING  (chan, upper_32_bits(node->vma[1].offset));
+		OUT_RING  (chan, lower_32_bits(node->vma[1].offset));
+		OUT_RING  (chan, 0x00000000 /* MODE_COPY, QUERY_NONE */);
+	}
+	return ret;
+}
+
+static int
+nv50_bo_move_init(struct nouveau_channel *chan, u32 handle)
+{
+	int ret = nouveau_notifier_alloc(chan, NvNotify0, 32, 0xfe0, 0x1000,
+					 &chan->m2mf_ntfy);
+	if (ret == 0) {
+		ret = RING_SPACE(chan, 6);
+		if (ret == 0) {
+			BEGIN_NV04(chan, NvSubCopy, 0x0000, 1);
+			OUT_RING  (chan, handle);
+			BEGIN_NV04(chan, NvSubCopy, 0x0180, 3);
+			OUT_RING  (chan, NvNotify0);
+			OUT_RING  (chan, NvDmaFB);
+			OUT_RING  (chan, NvDmaFB);
+		} else {
+			nouveau_ramht_remove(chan, NvNotify0);
+		}
+	}
+
+	return ret;
+}
+
+static int
 nv50_bo_move_m2mf(struct nouveau_channel *chan, struct ttm_buffer_object *bo,
 		  struct ttm_mem_reg *old_mem, struct ttm_mem_reg *new_mem)
 {
@@ -551,7 +719,7 @@ nv50_bo_move_m2mf(struct nouveau_channel *chan, struct ttm_buffer_object *bo,
 			if (ret)
 				return ret;
 
-			BEGIN_RING(chan, NvSubM2MF, 0x0200, 7);
+			BEGIN_NV04(chan, NvSubCopy, 0x0200, 7);
 			OUT_RING  (chan, 0);
 			OUT_RING  (chan, 0);
 			OUT_RING  (chan, stride);
@@ -564,7 +732,7 @@ nv50_bo_move_m2mf(struct nouveau_channel *chan, struct ttm_buffer_object *bo,
 			if (ret)
 				return ret;
 
-			BEGIN_RING(chan, NvSubM2MF, 0x0200, 1);
+			BEGIN_NV04(chan, NvSubCopy, 0x0200, 1);
 			OUT_RING  (chan, 1);
 		}
 		if (old_mem->mem_type == TTM_PL_VRAM &&
@@ -573,7 +741,7 @@ nv50_bo_move_m2mf(struct nouveau_channel *chan, struct ttm_buffer_object *bo,
 			if (ret)
 				return ret;
 
-			BEGIN_RING(chan, NvSubM2MF, 0x021c, 7);
+			BEGIN_NV04(chan, NvSubCopy, 0x021c, 7);
 			OUT_RING  (chan, 0);
 			OUT_RING  (chan, 0);
 			OUT_RING  (chan, stride);
@@ -586,7 +754,7 @@ nv50_bo_move_m2mf(struct nouveau_channel *chan, struct ttm_buffer_object *bo,
 			if (ret)
 				return ret;
 
-			BEGIN_RING(chan, NvSubM2MF, 0x021c, 1);
+			BEGIN_NV04(chan, NvSubCopy, 0x021c, 1);
 			OUT_RING  (chan, 1);
 		}
 
@@ -594,10 +762,10 @@ nv50_bo_move_m2mf(struct nouveau_channel *chan, struct ttm_buffer_object *bo,
 		if (ret)
 			return ret;
 
-		BEGIN_RING(chan, NvSubM2MF, 0x0238, 2);
+		BEGIN_NV04(chan, NvSubCopy, 0x0238, 2);
 		OUT_RING  (chan, upper_32_bits(src_offset));
 		OUT_RING  (chan, upper_32_bits(dst_offset));
-		BEGIN_RING(chan, NvSubM2MF, 0x030c, 8);
+		BEGIN_NV04(chan, NvSubCopy, 0x030c, 8);
 		OUT_RING  (chan, lower_32_bits(src_offset));
 		OUT_RING  (chan, lower_32_bits(dst_offset));
 		OUT_RING  (chan, stride);
@@ -606,7 +774,7 @@ nv50_bo_move_m2mf(struct nouveau_channel *chan, struct ttm_buffer_object *bo,
 		OUT_RING  (chan, height);
 		OUT_RING  (chan, 0x00000101);
 		OUT_RING  (chan, 0x00000000);
-		BEGIN_RING(chan, NvSubM2MF, NV_MEMORY_TO_MEMORY_FORMAT_NOP, 1);
+		BEGIN_NV04(chan, NvSubCopy, NV_MEMORY_TO_MEMORY_FORMAT_NOP, 1);