drm/radeon: Add asic hook for dma copy to r200 cards.

r200 cards have dma engine which can be used to tranfer data
between vram and system memory.

r300 dma engine registers match r200 dma engine. Enabling
dma copy for r200 is simple as hooking r200 asic to already
existing function r300_copy_dma.

Rename r300_dma_copy to r200_dma_copyto reflect that supports
starts from r200 cards.

v2: Created a new asic object for r200 cards.

Signed-off-by: Pauli Nieminen <suokkos@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

4 files changed, 104 insertions, 59 deletions

diff --git a/drivers/gpu/drm/radeon/r200.c b/drivers/gpu/drm/radeon/r200.c
index ff1e0cd608b..1146c9909c2 100644
--- a/drivers/gpu/drm/radeon/r200.c
+++ b/drivers/gpu/drm/radeon/r200.c
@@ -31,6 +31,7 @@
 #include "radeon_reg.h"
 #include "radeon.h"
 
+#include "r100d.h"
 #include "r200_reg_safe.h"
 
 #include "r100_track.h"
@@ -79,6 +80,51 @@ static int r200_get_vtx_size_0(uint32_t vtx_fmt_0)
 	return vtx_size;
 }
 
+int r200_copy_dma(struct radeon_device *rdev,
+		  uint64_t src_offset,
+		  uint64_t dst_offset,
+		  unsigned num_pages,
+		  struct radeon_fence *fence)
+{
+	uint32_t size;
+	uint32_t cur_size;
+	int i, num_loops;
+	int r = 0;
+
+	/* radeon pitch is /64 */
+	size = num_pages << PAGE_SHIFT;
+	num_loops = DIV_ROUND_UP(size, 0x1FFFFF);
+	r = radeon_ring_lock(rdev, num_loops * 4 + 64);
+	if (r) {
+		DRM_ERROR("radeon: moving bo (%d).\n", r);
+		return r;
+	}
+	/* Must wait for 2D idle & clean before DMA or hangs might happen */
+	radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0));
+	radeon_ring_write(rdev, (1 << 16));
+	for (i = 0; i < num_loops; i++) {
+		cur_size = size;
+		if (cur_size > 0x1FFFFF) {
+			cur_size = 0x1FFFFF;
+		}
+		size -= cur_size;
+		radeon_ring_write(rdev, PACKET0(0x720, 2));
+		radeon_ring_write(rdev, src_offset);
+		radeon_ring_write(rdev, dst_offset);
+		radeon_ring_write(rdev, cur_size | (1 << 31) | (1 << 30));
+		src_offset += cur_size;
+		dst_offset += cur_size;
+	}
+	radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0));
+	radeon_ring_write(rdev, RADEON_WAIT_DMA_GUI_IDLE);
+	if (fence) {
+		r = radeon_fence_emit(rdev, fence);
+	}
+	radeon_ring_unlock_commit(rdev);
+	return r;
+}
+
+
 static int r200_get_vtx_size_1(uint32_t vtx_fmt_1)
 {
 	int vtx_size, i, tex_size;
diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c
index 654aca1cdf0..a4f395226b3 100644
--- a/drivers/gpu/drm/radeon/r300.c
+++ b/drivers/gpu/drm/radeon/r300.c
@@ -201,50 +201,6 @@ void r300_fence_ring_emit(struct radeon_device *rdev,
 	radeon_ring_write(rdev, RADEON_SW_INT_FIRE);
 }
 
-int r300_copy_dma(struct radeon_device *rdev,
-		  uint64_t src_offset,
-		  uint64_t dst_offset,
-		  unsigned num_pages,
-		  struct radeon_fence *fence)
-{
-	uint32_t size;
-	uint32_t cur_size;
-	int i, num_loops;
-	int r = 0;
-
-	/* radeon pitch is /64 */
-	size = num_pages << PAGE_SHIFT;
-	num_loops = DIV_ROUND_UP(size, 0x1FFFFF);
-	r = radeon_ring_lock(rdev, num_loops * 4 + 64);
-	if (r) {
-		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		return r;
-	}
-	/* Must wait for 2D idle & clean before DMA or hangs might happen */
-	radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0 ));
-	radeon_ring_write(rdev, RADEON_WAIT_2D_IDLECLEAN);
-	for (i = 0; i < num_loops; i++) {
-		cur_size = size;
-		if (cur_size > 0x1FFFFF) {
-			cur_size = 0x1FFFFF;
-		}
-		size -= cur_size;
-		radeon_ring_write(rdev, PACKET0(0x720, 2));
-		radeon_ring_write(rdev, src_offset);
-		radeon_ring_write(rdev, dst_offset);
-		radeon_ring_write(rdev, cur_size | (1 << 31) | (1 << 30));
-		src_offset += cur_size;
-		dst_offset += cur_size;
-	}
-	radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0));
-	radeon_ring_write(rdev, RADEON_WAIT_DMA_GUI_IDLE);
-	if (fence) {
-		r = radeon_fence_emit(rdev, fence);
-	}
-	radeon_ring_unlock_commit(rdev);
-	return r;
-}
-
 void r300_ring_start(struct radeon_device *rdev)
 {
 	unsigned gb_tile_config;
diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h
index 9d2a113f78e..b7030d7c039 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.h
+++ b/drivers/gpu/drm/radeon/radeon_asic.h
@@ -43,7 +43,7 @@ void radeon_atom_set_memory_clock(struct radeon_device *rdev, uint32_t mem_clock
 void radeon_atom_set_clock_gating(struct radeon_device *rdev, int enable);
 
 /*
- * r100,rv100,rs100,rv200,rs200,r200,rv250,rs300,rv280
+ * r100,rv100,rs100,rv200,rs200
  */
 extern int r100_init(struct radeon_device *rdev);
 extern void r100_fini(struct radeon_device *rdev);
@@ -121,6 +121,51 @@ static struct radeon_asic r100_asic = {
 	.ioctl_wait_idle = NULL,
 };
 
+/*
+ * r200,rv250,rs300,rv280
+ */
+extern int r200_copy_dma(struct radeon_device *rdev,
+			uint64_t src_offset,
+			uint64_t dst_offset,
+			unsigned num_pages,
+			struct radeon_fence *fence);
+static struct radeon_asic r200_asic = {
+	.init = &r100_init,
+	.fini = &r100_fini,
+	.suspend = &r100_suspend,
+	.resume = &r100_resume,
+	.vga_set_state = &r100_vga_set_state,
+	.gpu_reset = &r100_gpu_reset,
+	.gart_tlb_flush = &r100_pci_gart_tlb_flush,
+	.gart_set_page = &r100_pci_gart_set_page,
+	.cp_commit = &r100_cp_commit,
+	.ring_start = &r100_ring_start,
+	.ring_test = &r100_ring_test,
+	.ring_ib_execute = &r100_ring_ib_execute,
+	.irq_set = &r100_irq_set,
+	.irq_process = &r100_irq_process,
+	.get_vblank_counter = &r100_get_vblank_counter,
+	.fence_ring_emit = &r100_fence_ring_emit,
+	.cs_parse = &r100_cs_parse,
+	.copy_blit = &r100_copy_blit,
+	.copy_dma = &r200_copy_dma,
+	.copy = &r100_copy_blit,
+	.get_engine_clock = &radeon_legacy_get_engine_clock,
+	.set_engine_clock = &radeon_legacy_set_engine_clock,
+	.get_memory_clock = &radeon_legacy_get_memory_clock,
+	.set_memory_clock = NULL,
+	.set_pcie_lanes = NULL,
+	.set_clock_gating = &radeon_legacy_set_clock_gating,
+	.set_surface_reg = r100_set_surface_reg,
+	.clear_surface_reg = r100_clear_surface_reg,
+	.bandwidth_update = &r100_bandwidth_update,
+	.hpd_init = &r100_hpd_init,
+	.hpd_fini = &r100_hpd_fini,
+	.hpd_sense = &r100_hpd_sense,
+	.hpd_set_polarity = &r100_hpd_set_polarity,
+	.ioctl_wait_idle = NULL,
+};
+
 
 /*
  * r300,r350,rv350,rv380
@@ -140,11 +185,7 @@ extern uint32_t rv370_pcie_rreg(struct radeon_device *rdev, uint32_t reg);
 extern void rv370_pcie_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
 extern void rv370_set_pcie_lanes(struct radeon_device *rdev, int lanes);
 extern int rv370_get_pcie_lanes(struct radeon_device *rdev);
-extern int r300_copy_dma(struct radeon_device *rdev,
-			uint64_t src_offset,
-			uint64_t dst_offset,
-			unsigned num_pages,
-			struct radeon_fence *fence);
+
 static struct radeon_asic r300_asic = {
 	.init = &r300_init,
 	.fini = &r300_fini,
@@ -164,7 +205,7 @@ static struct radeon_asic r300_asic = {
 	.fence_ring_emit = &r300_fence_ring_emit,
 	.cs_parse = &r300_cs_parse,
 	.copy_blit = &r100_copy_blit,
-	.copy_dma = &r300_copy_dma,
+	.copy_dma = &r200_copy_dma,
 	.copy = &r100_copy_blit,
 	.get_engine_clock = &radeon_legacy_get_engine_clock,
 	.set_engine_clock = &radeon_legacy_set_engine_clock,
@@ -247,7 +288,7 @@ static struct radeon_asic r420_asic = {
 	.fence_ring_emit = &r300_fence_ring_emit,
 	.cs_parse = &r300_cs_parse,
 	.copy_blit = &r100_copy_blit,
-	.copy_dma = &r300_copy_dma,
+	.copy_dma = &r200_copy_dma,
 	.copy = &r100_copy_blit,
 	.get_engine_clock = &radeon_atom_get_engine_clock,
 	.set_engine_clock = &radeon_atom_set_engine_clock,
@@ -297,7 +338,7 @@ static struct radeon_asic rs400_asic = {
 	.fence_ring_emit = &r300_fence_ring_emit,
 	.cs_parse = &r300_cs_parse,
 	.copy_blit = &r100_copy_blit,
-	.copy_dma = &r300_copy_dma,
+	.copy_dma = &r200_copy_dma,
 	.copy = &r100_copy_blit,
 	.get_engine_clock = &radeon_legacy_get_engine_clock,
 	.set_engine_clock = &radeon_legacy_set_engine_clock,
@@ -357,7 +398,7 @@ static struct radeon_asic rs600_asic = {
 	.fence_ring_emit = &r300_fence_ring_emit,
 	.cs_parse = &r300_cs_parse,
 	.copy_blit = &r100_copy_blit,
-	.copy_dma = &r300_copy_dma,
+	.copy_dma = &r200_copy_dma,
 	.copy = &r100_copy_blit,
 	.get_engine_clock = &radeon_atom_get_engine_clock,
 	.set_engine_clock = &radeon_atom_set_engine_clock,
@@ -404,8 +445,8 @@ static struct radeon_asic rs690_asic = {
 	.fence_ring_emit = &r300_fence_ring_emit,
 	.cs_parse = &r300_cs_parse,
 	.copy_blit = &r100_copy_blit,
-	.copy_dma = &r300_copy_dma,
-	.copy = &r300_copy_dma,
+	.copy_dma = &r200_copy_dma,
+	.copy = &r200_copy_dma,
 	.get_engine_clock = &radeon_atom_get_engine_clock,
 	.set_engine_clock = &radeon_atom_set_engine_clock,
 	.get_memory_clock = &radeon_atom_get_memory_clock,
@@ -457,7 +498,7 @@ static struct radeon_asic rv515_asic = {
 	.fence_ring_emit = &r300_fence_ring_emit,
 	.cs_parse = &r300_cs_parse,
 	.copy_blit = &r100_copy_blit,
-	.copy_dma = &r300_copy_dma,
+	.copy_dma = &r200_copy_dma,
 	.copy = &r100_copy_blit,
 	.get_engine_clock = &radeon_atom_get_engine_clock,
 	.set_engine_clock = &radeon_atom_set_engine_clock,
@@ -501,7 +542,7 @@ static struct radeon_asic r520_asic = {
 	.fence_ring_emit = &r300_fence_ring_emit,
 	.cs_parse = &r300_cs_parse,
 	.copy_blit = &r100_copy_blit,
-	.copy_dma = &r300_copy_dma,
+	.copy_dma = &r200_copy_dma,
 	.copy = &r100_copy_blit,
 	.get_engine_clock = &radeon_atom_get_engine_clock,
 	.set_engine_clock = &radeon_atom_set_engine_clock,
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index 7a30f6955e4..7be3a696846 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -340,11 +340,13 @@ int radeon_asic_init(struct radeon_device *rdev)
 	case CHIP_RS100:
 	case CHIP_RV200:
 	case CHIP_RS200:
+		rdev->asic = &r100_asic;
+		break;
 	case CHIP_R200:
 	case CHIP_RV250:
 	case CHIP_RS300:
 	case CHIP_RV280:
-		rdev->asic = &r100_asic;
+		rdev->asic = &r200_asic;
 		break;
 	case CHIP_R300:
 	case CHIP_R350: