Merge commit 'v2.6.33' into perf/core

Merge reason:
	__percpu annotations need the corresponding sparse address
space definition upstream.

Conflicts:
	tools/perf/util/probe-event.c (trivial)

33 files changed, 449 insertions, 346 deletions

diff --git a/drivers/gpu/drm/radeon/Kconfig b/drivers/gpu/drm/radeon/Kconfig
index 5982321be4d5..1c02d23f6fcc 100644
--- a/drivers/gpu/drm/radeon/Kconfig
+++ b/drivers/gpu/drm/radeon/Kconfig
@@ -1,10 +1,14 @@
 config DRM_RADEON_KMS
-	bool "Enable modesetting on radeon by default"
+	bool "Enable modesetting on radeon by default - NEW DRIVER"
 	depends on DRM_RADEON
 	help
-	  Choose this option if you want kernel modesetting enabled by default,
-	  and you have a new enough userspace to support this. Running old
-	  userspaces with this enabled will cause pain.
+	  Choose this option if you want kernel modesetting enabled by default.
+
+	  This is a completely new driver. It's only part of the existing drm
+	  for compatibility reasons. It requires an entirely different graphics
+	  stack above it and works very differently from the old drm stack.
+	  i.e. don't enable this unless you know what you are doing it may
+	  cause issues or bugs compared to the previous userspace driver stack.
 
 	  When kernel modesetting is enabled the IOCTL of radeon/drm
 	  driver are considered as invalid and an error message is printed
diff --git a/drivers/gpu/drm/radeon/atom.c b/drivers/gpu/drm/radeon/atom.c
index e3b44562d265..7f152f66f196 100644
--- a/drivers/gpu/drm/radeon/atom.c
+++ b/drivers/gpu/drm/radeon/atom.c
@@ -24,6 +24,7 @@
 
 #include <linux/module.h>
 #include <linux/sched.h>
+#include <asm/unaligned.h>
 
 #define ATOM_DEBUG
 
@@ -212,7 +213,9 @@ static uint32_t atom_get_src_int(atom_exec_context *ctx, uint8_t attr,
 	case ATOM_ARG_PS:
 		idx = U8(*ptr);
 		(*ptr)++;
-		val = le32_to_cpu(ctx->ps[idx]);
+		/* get_unaligned_le32 avoids unaligned accesses from atombios
+		 * tables, noticed on a DEC Alpha. */
+		val = get_unaligned_le32((u32 *)&ctx->ps[idx]);
 		if (print)
 			DEBUG("PS[0x%02X,0x%04X]", idx, val);
 		break;
@@ -640,7 +643,7 @@ static void atom_op_delay(atom_exec_context *ctx, int *ptr, int arg)
 	uint8_t count = U8((*ptr)++);
 	SDEBUG("   count: %d\n", count);
 	if (arg == ATOM_UNIT_MICROSEC)
-		schedule_timeout_uninterruptible(usecs_to_jiffies(count));
+		udelay(count);
 	else
 		schedule_timeout_uninterruptible(msecs_to_jiffies(count));
 }
diff --git a/drivers/gpu/drm/radeon/atombios_dp.c b/drivers/gpu/drm/radeon/atombios_dp.c
index 3eb0ca5b3d73..99915a682d59 100644
--- a/drivers/gpu/drm/radeon/atombios_dp.c
+++ b/drivers/gpu/drm/radeon/atombios_dp.c
@@ -332,11 +332,13 @@ bool radeon_process_aux_ch(struct radeon_i2c_chan *chan, u8 *req_bytes,
 	PROCESS_AUX_CHANNEL_TRANSACTION_PS_ALLOCATION args;
 	int index = GetIndexIntoMasterTable(COMMAND, ProcessAuxChannelTransaction);
 	unsigned char *base;
+	int retry_count = 0;
 
 	memset(&args, 0, sizeof(args));
 
 	base = (unsigned char *)rdev->mode_info.atom_context->scratch;
 
+retry:
 	memcpy(base, req_bytes, num_bytes);
 
 	args.lpAuxRequest = 0;
@@ -347,10 +349,12 @@ bool radeon_process_aux_ch(struct radeon_i2c_chan *chan, u8 *req_bytes,
 
 	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
 
-	if (args.ucReplyStatus) {
-		DRM_DEBUG("failed to get auxch %02x%02x %02x %02x 0x%02x %02x\n",
+	if (args.ucReplyStatus && !args.ucDataOutLen) {
+		if (args.ucReplyStatus == 0x20 && retry_count++ < 10)
+			goto retry;
+		DRM_DEBUG("failed to get auxch %02x%02x %02x %02x 0x%02x %02x after %d retries\n",
 			  req_bytes[1], req_bytes[0], req_bytes[2], req_bytes[3],
-			  chan->rec.i2c_id, args.ucReplyStatus);
+			  chan->rec.i2c_id, args.ucReplyStatus, retry_count);
 		return false;
 	}
 
@@ -468,7 +472,7 @@ void radeon_dp_set_link_config(struct drm_connector *connector,
 	struct radeon_connector *radeon_connector;
 	struct radeon_connector_atom_dig *dig_connector;
 
-	if ((connector->connector_type != DRM_MODE_CONNECTOR_DisplayPort) ||
+	if ((connector->connector_type != DRM_MODE_CONNECTOR_DisplayPort) &&
 	    (connector->connector_type != DRM_MODE_CONNECTOR_eDP))
 		return;
 
@@ -583,7 +587,7 @@ void dp_link_train(struct drm_encoder *encoder,
 	u8 train_set[4];
 	int i;
 
-	if ((connector->connector_type != DRM_MODE_CONNECTOR_DisplayPort) ||
+	if ((connector->connector_type != DRM_MODE_CONNECTOR_DisplayPort) &&
 	    (connector->connector_type != DRM_MODE_CONNECTOR_eDP))
 		return;
 
@@ -596,21 +600,14 @@ void dp_link_train(struct drm_encoder *encoder,
 		return;
 	dig_connector = radeon_connector->con_priv;
 
-	if (ASIC_IS_DCE32(rdev)) {
-		if (dig->dig_block)
-			enc_id |= ATOM_DP_CONFIG_DIG2_ENCODER;
-		else
-			enc_id |= ATOM_DP_CONFIG_DIG1_ENCODER;
-		if (dig_connector->linkb)
-			enc_id |= ATOM_DP_CONFIG_LINK_B;
-		else
-			enc_id |= ATOM_DP_CONFIG_LINK_A;
-	} else {
-		if (dig_connector->linkb)
-			enc_id |= ATOM_DP_CONFIG_DIG2_ENCODER | ATOM_DP_CONFIG_LINK_B;
-		else
-			enc_id |= ATOM_DP_CONFIG_DIG1_ENCODER | ATOM_DP_CONFIG_LINK_A;
-	}
+	if (dig->dig_encoder)
+		enc_id |= ATOM_DP_CONFIG_DIG2_ENCODER;
+	else
+		enc_id |= ATOM_DP_CONFIG_DIG1_ENCODER;
+	if (dig_connector->linkb)
+		enc_id |= ATOM_DP_CONFIG_LINK_B;
+	else
+		enc_id |= ATOM_DP_CONFIG_LINK_A;
 
 	memset(link_configuration, 0, DP_LINK_CONFIGURATION_SIZE);
 	if (dig_connector->dp_clock == 270000)
diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index 11c9a3fe6810..c0d4650cdb79 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -354,11 +354,17 @@ u32 r100_get_vblank_counter(struct radeon_device *rdev, int crtc)
 		return RREG32(RADEON_CRTC2_CRNT_FRAME);
 }
 
+/* Who ever call radeon_fence_emit should call ring_lock and ask
+ * for enough space (today caller are ib schedule and buffer move) */
 void r100_fence_ring_emit(struct radeon_device *rdev,
 			  struct radeon_fence *fence)
 {
-	/* Who ever call radeon_fence_emit should call ring_lock and ask
-	 * for enough space (today caller are ib schedule and buffer move) */
+	/* We have to make sure that caches are flushed before
+	 * CPU might read something from VRAM. */
+	radeon_ring_write(rdev, PACKET0(RADEON_RB3D_DSTCACHE_CTLSTAT, 0));
+	radeon_ring_write(rdev, RADEON_RB3D_DC_FLUSH_ALL);
+	radeon_ring_write(rdev, PACKET0(RADEON_RB3D_ZCACHE_CTLSTAT, 0));
+	radeon_ring_write(rdev, RADEON_RB3D_ZC_FLUSH_ALL);
 	/* Wait until IDLE & CLEAN */
 	radeon_ring_write(rdev, PACKET0(0x1720, 0));
 	radeon_ring_write(rdev, (1 << 16) | (1 << 17));
@@ -3369,7 +3375,6 @@ int r100_suspend(struct radeon_device *rdev)
 
 void r100_fini(struct radeon_device *rdev)
 {
-	r100_suspend(rdev);
 	r100_cp_fini(rdev);
 	r100_wb_fini(rdev);
 	r100_ib_fini(rdev);
@@ -3481,13 +3486,12 @@ int r100_init(struct radeon_device *rdev)
 	if (r) {
 		/* Somethings want wront with the accel init stop accel */
 		dev_err(rdev->dev, "Disabling GPU acceleration\n");
-		r100_suspend(rdev);
 		r100_cp_fini(rdev);
 		r100_wb_fini(rdev);
 		r100_ib_fini(rdev);
+		radeon_irq_kms_fini(rdev);
 		if (rdev->flags & RADEON_IS_PCI)
 			r100_pci_gart_fini(rdev);
-		radeon_irq_kms_fini(rdev);
 		rdev->accel_working = false;
 	}
 	return 0;
diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c
index 0051d11b907c..43b55a030b4d 100644
--- a/drivers/gpu/drm/radeon/r300.c
+++ b/drivers/gpu/drm/radeon/r300.c
@@ -506,11 +506,14 @@ void r300_vram_info(struct radeon_device *rdev)
 
 	/* DDR for all card after R300 & IGP */
 	rdev->mc.vram_is_ddr = true;
+
 	tmp = RREG32(RADEON_MEM_CNTL);
-	if (tmp & R300_MEM_NUM_CHANNELS_MASK) {
-		rdev->mc.vram_width = 128;
-	} else {
-		rdev->mc.vram_width = 64;
+	tmp &= R300_MEM_NUM_CHANNELS_MASK;
+	switch (tmp) {
+	case 0: rdev->mc.vram_width = 64; break;
+	case 1: rdev->mc.vram_width = 128; break;
+	case 2: rdev->mc.vram_width = 256; break;
+	default:  rdev->mc.vram_width = 128; break;
 	}
 
 	r100_vram_init_sizes(rdev);
@@ -1327,7 +1330,6 @@ int r300_suspend(struct radeon_device *rdev)
 
 void r300_fini(struct radeon_device *rdev)
 {
-	r300_suspend(rdev);
 	r100_cp_fini(rdev);
 	r100_wb_fini(rdev);
 	r100_ib_fini(rdev);
@@ -1418,15 +1420,15 @@ int r300_init(struct radeon_device *rdev)
 	if (r) {
 		/* Somethings want wront with the accel init stop accel */
 		dev_err(rdev->dev, "Disabling GPU acceleration\n");
-		r300_suspend(rdev);
 		r100_cp_fini(rdev);
 		r100_wb_fini(rdev);
 		r100_ib_fini(rdev);
+		radeon_irq_kms_fini(rdev);
 		if (rdev->flags & RADEON_IS_PCIE)
 			rv370_pcie_gart_fini(rdev);
 		if (rdev->flags & RADEON_IS_PCI)
 			r100_pci_gart_fini(rdev);
-		radeon_irq_kms_fini(rdev);
+		radeon_agp_fini(rdev);
 		rdev->accel_working = false;
 	}
 	return 0;
diff --git a/drivers/gpu/drm/radeon/r420.c b/drivers/gpu/drm/radeon/r420.c
index 4526faaacca8..d9373246c97f 100644
--- a/drivers/gpu/drm/radeon/r420.c
+++ b/drivers/gpu/drm/radeon/r420.c
@@ -389,16 +389,15 @@ int r420_init(struct radeon_device *rdev)
 	if (r) {
 		/* Somethings want wront with the accel init stop accel */
 		dev_err(rdev->dev, "Disabling GPU acceleration\n");
-		r420_suspend(rdev);
 		r100_cp_fini(rdev);
 		r100_wb_fini(rdev);
 		r100_ib_fini(rdev);
+		radeon_irq_kms_fini(rdev);
 		if (rdev->flags & RADEON_IS_PCIE)
 			rv370_pcie_gart_fini(rdev);
 		if (rdev->flags & RADEON_IS_PCI)
 			r100_pci_gart_fini(rdev);
 		radeon_agp_fini(rdev);
-		radeon_irq_kms_fini(rdev);
 		rdev->accel_working = false;
 	}
 	return 0;
diff --git a/drivers/gpu/drm/radeon/r520.c b/drivers/gpu/drm/radeon/r520.c
index 9a189072f2b9..ddf5731eba0d 100644
--- a/drivers/gpu/drm/radeon/r520.c
+++ b/drivers/gpu/drm/radeon/r520.c
@@ -294,13 +294,12 @@ int r520_init(struct radeon_device *rdev)
 	if (r) {
 		/* Somethings want wront with the accel init stop accel */
 		dev_err(rdev->dev, "Disabling GPU acceleration\n");
-		rv515_suspend(rdev);
 		r100_cp_fini(rdev);
 		r100_wb_fini(rdev);
 		r100_ib_fini(rdev);
+		radeon_irq_kms_fini(rdev);
 		rv370_pcie_gart_fini(rdev);
 		radeon_agp_fini(rdev);
-		radeon_irq_kms_fini(rdev);
 		rdev->accel_working = false;
 	}
 	return 0;
diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index da9aa3c31bcf..2ffcf5a03551 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -1654,6 +1654,12 @@ void r600_ring_init(struct radeon_device *rdev, unsigned ring_size)
 	rdev->cp.align_mask = 16 - 1;
 }
 
+void r600_cp_fini(struct radeon_device *rdev)
+{
+	r600_cp_stop(rdev);
+	radeon_ring_fini(rdev);
+}
+
 
 /*
  * GPU scratch registers helpers function.
@@ -1788,23 +1794,24 @@ void r600_fence_ring_emit(struct radeon_device *rdev,
 	radeon_ring_write(rdev, RB_INT_STAT);
 }
 
-int r600_copy_dma(struct radeon_device *rdev,
-		  uint64_t src_offset,
-		  uint64_t dst_offset,
-		  unsigned num_pages,
-		  struct radeon_fence *fence)
-{
-	/* FIXME: implement */
-	return 0;
-}
-
 int r600_copy_blit(struct radeon_device *rdev,
 		   uint64_t src_offset, uint64_t dst_offset,
 		   unsigned num_pages, struct radeon_fence *fence)
 {
-	r600_blit_prepare_copy(rdev, num_pages * RADEON_GPU_PAGE_SIZE);
+	int r;
+
+	mutex_lock(&rdev->r600_blit.mutex);
+	rdev->r600_blit.vb_ib = NULL;
+	r = r600_blit_prepare_copy(rdev, num_pages * RADEON_GPU_PAGE_SIZE);
+	if (r) {
+		if (rdev->r600_blit.vb_ib)
+			radeon_ib_free(rdev, &rdev->r600_blit.vb_ib);
+		mutex_unlock(&rdev->r600_blit.mutex);
+		return r;
+	}
 	r600_kms_blit_copy(rdev, src_offset, dst_offset, num_pages * RADEON_GPU_PAGE_SIZE);
 	r600_blit_done_copy(rdev, fence);
+	mutex_unlock(&rdev->r600_blit.mutex);
 	return 0;
 }
 
@@ -1860,26 +1867,25 @@ int r600_startup(struct radeon_device *rdev)
 			return r;
 	}
 	r600_gpu_init(rdev);
-
-	if (!rdev->r600_blit.shader_obj) {
-		r = r600_blit_init(rdev);
+	r = r600_blit_init(rdev);
+	if (r) {
+		r600_blit_fini(rdev);
+		rdev->asic->copy = NULL;
+		dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r);
+	}
+	/* pin copy shader into vram */
+	if (rdev->r600_blit.shader_obj) {
+		r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false);
+		if (unlikely(r != 0))
+			return r;
+		r = radeon_bo_pin(rdev->r600_blit.shader_obj, RADEON_GEM_DOMAIN_VRAM,
+				&rdev->r600_blit.shader_gpu_addr);
+		radeon_bo_unreserve(rdev->r600_blit.shader_obj);
 		if (r) {
-			DRM_ERROR("radeon: failed blitter (%d).\n", r);
+			dev_err(rdev->dev, "(%d) pin blit object failed\n", r);
 			return r;
 		}
 	}
-
-	r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false);
-	if (unlikely(r != 0))
-		return r;
-	r = radeon_bo_pin(rdev->r600_blit.shader_obj, RADEON_GEM_DOMAIN_VRAM,
-			&rdev->r600_blit.shader_gpu_addr);
-	radeon_bo_unreserve(rdev->r600_blit.shader_obj);
-	if (r) {
-		dev_err(rdev->dev, "(%d) pin blit object failed\n", r);
-		return r;
-	}
-
 	/* Enable IRQ */
 	r = r600_irq_init(rdev);
 	if (r) {
@@ -1944,6 +1950,13 @@ int r600_resume(struct radeon_device *rdev)
 		DRM_ERROR("radeon: failled testing IB (%d).\n", r);
 		return r;
 	}
+
+	r = r600_audio_init(rdev);
+	if (r) {
+		DRM_ERROR("radeon: audio resume failed\n");
+		return r;
+	}
+
 	return r;
 }
 
@@ -1951,6 +1964,7 @@ int r600_suspend(struct radeon_device *rdev)
 {
 	int r;
 
+	r600_audio_fini(rdev);
 	/* FIXME: we should wait for ring to be empty */
 	r600_cp_stop(rdev);
 	rdev->cp.ready = false;
@@ -2055,9 +2069,11 @@ int r600_init(struct radeon_device *rdev)
 	rdev->accel_working = true;
 	r = r600_startup(rdev);
 	if (r) {
-		r600_suspend(rdev);
+		dev_err(rdev->dev, "disabling GPU acceleration\n");
+		r600_cp_fini(rdev);
 		r600_wb_fini(rdev);
-		radeon_ring_fini(rdev);
+		r600_irq_fini(rdev);
+		radeon_irq_kms_fini(rdev);
 		r600_pcie_gart_fini(rdev);
 		rdev->accel_working = false;
 	}
@@ -2083,20 +2099,17 @@ int r600_init(struct radeon_device *rdev)
 
 void r600_fini(struct radeon_device *rdev)
 {
-	/* Suspend operations */
-	r600_suspend(rdev);
-
 	r600_audio_fini(rdev);
 	r600_blit_fini(rdev);
+	r600_cp_fini(rdev);
+	r600_wb_fini(rdev);
 	r600_irq_fini(rdev);
 	radeon_irq_kms_fini(rdev);
-	radeon_ring_fini(rdev);
-	r600_wb_fini(rdev);
 	r600_pcie_gart_fini(rdev);
+	radeon_agp_fini(rdev);
 	radeon_gem_fini(rdev);
 	radeon_fence_driver_fini(rdev);
 	radeon_clocks_fini(rdev);
-	radeon_agp_fini(rdev);
 	radeon_bo_fini(rdev);
 	radeon_atombios_fini(rdev);
 	kfree(rdev->bios);
@@ -2900,3 +2913,18 @@ int r600_debugfs_mc_info_init(struct radeon_device *rdev)
 	return 0;
 #endif
 }
+
+/**
+ * r600_ioctl_wait_idle - flush host path cache on wait idle ioctl
+ * rdev: radeon device structure
+ * bo: buffer object struct which userspace is waiting for idle
+ *
+ * Some R6XX/R7XX doesn't seems to take into account HDP flush performed
+ * through ring buffer, this leads to corruption in rendering, see
+ * http://bugzilla.kernel.org/show_bug.cgi?id=15186 to avoid this we
+ * directly perform HDP flush by writing register through MMIO.
+ */
+void r600_ioctl_wait_idle(struct radeon_device *rdev, struct radeon_bo *bo)
+{
+	WREG32(R_005480_HDP_MEM_COHERENCY_FLUSH_CNTL, 0x1);
+}
diff --git a/drivers/gpu/drm/radeon/r600_audio.c b/drivers/gpu/drm/radeon/r600_audio.c
index 99e2c3891a7d..0dcb6904c4ff 100644
--- a/drivers/gpu/drm/radeon/r600_audio.c
+++ b/drivers/gpu/drm/radeon/r600_audio.c
@@ -35,7 +35,7 @@
  */
 static int r600_audio_chipset_supported(struct radeon_device *rdev)
 {
-	return rdev->family >= CHIP_R600
+	return (rdev->family >= CHIP_R600 && rdev->family < CHIP_RV710)
 		|| rdev->family == CHIP_RS600
 		|| rdev->family == CHIP_RS690
 		|| rdev->family == CHIP_RS740;
@@ -261,7 +261,6 @@ void r600_audio_fini(struct radeon_device *rdev)
 	if (!r600_audio_chipset_supported(rdev))
 		return;
 
-	WREG32_P(R600_AUDIO_ENABLE, 0x0, ~0x81000000);
-
 	del_timer(&rdev->audio_timer);
+	WREG32_P(R600_AUDIO_ENABLE, 0x0, ~0x81000000);
 }
diff --git a/drivers/gpu/drm/radeon/r600_blit_kms.c b/drivers/gpu/drm/radeon/r600_blit_kms.c
index 2bedce477a97..446b765ac72a 100644
--- a/drivers/gpu/drm/radeon/r600_blit_kms.c
+++ b/drivers/gpu/drm/radeon/r600_blit_kms.c
@@ -449,6 +449,7 @@ int r600_blit_init(struct radeon_device *rdev)
 	u32 packet2s[16];
 	int num_packet2s = 0;
 
+	mutex_init(&rdev->r600_blit.mutex);
 	rdev->r600_blit.state_offset = 0;
 
 	if (rdev->family >= CHIP_RV770)
@@ -542,9 +543,6 @@ int r600_vb_ib_get(struct radeon_device *rdev)
 void r600_vb_ib_put(struct radeon_device *rdev)
 {
 	radeon_fence_emit(rdev, rdev->r600_blit.vb_ib->fence);
-	mutex_lock(&rdev->ib_pool.mutex);
-	list_add_tail(&rdev->r600_blit.vb_ib->list, &rdev->ib_pool.scheduled_ibs);
-	mutex_unlock(&rdev->ib_pool.mutex);
 	radeon_ib_free(rdev, &rdev->r600_blit.vb_ib);
 }
 
@@ -557,7 +555,8 @@ int r600_blit_prepare_copy(struct radeon_device *rdev, int size_bytes)
 	int dwords_per_loop = 76, num_loops;
 
 	r = r600_vb_ib_get(rdev);
-	WARN_ON(r);
+	if (r)
+		return r;
 
 	/* set_render_target emits 2 extra dwords on rv6xx */
 	if (rdev->family > CHIP_R600 && rdev->family < CHIP_RV770)
@@ -583,7 +582,8 @@ int r600_blit_prepare_copy(struct radeon_device *rdev, int size_bytes)
 	ring_size += 5; /* done copy */
 	ring_size += 7; /* fence emit for done copy */
 	r = radeon_ring_lock(rdev, ring_size);
-	WARN_ON(r);
+	if (r)
+		return r;
 
 	set_default_state(rdev); /* 14 */
 	set_shaders(rdev); /* 26 */
diff --git a/drivers/gpu/drm/radeon/r600_cp.c b/drivers/gpu/drm/radeon/r600_cp.c
index 6d5a711c2e91..75bcf35a0931 100644
--- a/drivers/gpu/drm/radeon/r600_cp.c
+++ b/drivers/gpu/drm/radeon/r600_cp.c
@@ -1428,9 +1428,12 @@ static void r700_gfx_init(struct drm_device *dev,
 
 	gb_tiling_config |= R600_BANK_SWAPS(1);
 
-	backend_map = r700_get_tile_pipe_to_backend_map(dev_priv->r600_max_tile_pipes,
-							dev_priv->r600_max_backends,
-							(0xff << dev_priv->r600_max_backends) & 0xff);
+	if ((dev_priv->flags & RADEON_FAMILY_MASK) == CHIP_RV740)
+		backend_map = 0x28;
+	else
+		backend_map = r700_get_tile_pipe_to_backend_map(dev_priv->r600_max_tile_pipes,
+								dev_priv->r600_max_backends,
+								(0xff << dev_priv->r600_max_backends) & 0xff);
 	gb_tiling_config |= R600_BACKEND_MAP(backend_map);
 
 	cc_gc_shader_pipe_config =
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index f7df1a7e4413..c0356bb193e5 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -96,6 +96,7 @@ extern int radeon_audio;
  * symbol;
  */
 #define RADEON_MAX_USEC_TIMEOUT		100000	/* 100 ms */
+/* RADEON_IB_POOL_SIZE must be a power of 2 */
 #define RADEON_IB_POOL_SIZE		16
 #define RADEON_DEBUGFS_MAX_NUM_FILES	32
 #define RADEONFB_CONN_LIMIT		4
@@ -363,11 +364,12 @@ void radeon_irq_kms_sw_irq_put(struct radeon_device *rdev);
  */
 struct radeon_ib {
 	struct list_head	list;
-	unsigned long		idx;
+	unsigned		idx;
 	uint64_t		gpu_addr;
 	struct radeon_fence	*fence;
-	uint32_t	*ptr;
+	uint32_t		*ptr;
 	uint32_t		length_dw;
+	bool			free;
 };
 
 /*
@@ -377,10 +379,9 @@ struct radeon_ib {
 struct radeon_ib_pool {
 	struct mutex		mutex;
 	struct radeon_bo	*robj;
-	struct list_head	scheduled_ibs;
 	struct radeon_ib	ibs[RADEON_IB_POOL_SIZE];
 	bool			ready;
-	DECLARE_BITMAP(alloc_bm, RADEON_IB_POOL_SIZE);
+	unsigned		head_id;
 };
 
 struct radeon_cp {
@@ -416,6 +417,7 @@ struct r600_ih {
 };
 
 struct r600_blit {
+	struct mutex		mutex;
 	struct radeon_bo	*shader_obj;
 	u64 shader_gpu_addr;
 	u32 vs_offset, ps_offset;
@@ -660,6 +662,13 @@ struct radeon_asic {
 	void (*hpd_fini)(struct radeon_device *rdev);
 	bool (*hpd_sense)(struct radeon_device *rdev, enum radeon_hpd_id hpd);
 	void (*hpd_set_polarity)(struct radeon_device *rdev, enum radeon_hpd_id hpd);
+	/* ioctl hw specific callback. Some hw might want to perform special
+	 * operation on specific ioctl. For instance on wait idle some hw
+	 * might want to perform and HDP flush through MMIO as it seems that
+	 * some R6XX/R7XX hw doesn't take HDP flush into account if programmed
+	 * through ring.
+	 */
+	void (*ioctl_wait_idle)(struct radeon_device *rdev, struct radeon_bo *bo);
 };
 
 /*
@@ -1142,6 +1151,7 @@ extern bool r600_card_posted(struct radeon_device *rdev);
 extern void r600_cp_stop(struct radeon_device *rdev);
 extern void r600_ring_init(struct radeon_device *rdev, unsigned ring_size);
 extern int r600_cp_resume(struct radeon_device *rdev);
+extern void r600_cp_fini(struct radeon_device *rdev);
 extern int r600_count_pipe_bits(uint32_t val);
 extern int r600_gart_clear_page(struct radeon_device *rdev, int i);
 extern int r600_mc_wait_for_idle(struct radeon_device *rdev);
diff --git a/drivers/gpu/drm/radeon/radeon_agp.c b/drivers/gpu/drm/radeon/radeon_agp.c
index c9ad7f5cc1ac..c0681a5556dc 100644
--- a/drivers/gpu/drm/radeon/radeon_agp.c
+++ b/drivers/gpu/drm/radeon/radeon_agp.c
@@ -133,13 +133,6 @@ int radeon_agp_init(struct radeon_device *rdev)
 	bool is_v3;
 	int ret;
 
-	if (rdev->ddev->agp->agp_info.aper_size < 32) {
-		dev_warn(rdev->dev, "AGP aperture to small (%dM) "
-			"need at least 32M, disabling AGP\n",
-			rdev->ddev->agp->agp_info.aper_size);
-		return -EINVAL;
-	}
-
 	/* Acquire AGP. */
 	if (!rdev->ddev->agp->acquired) {
 		ret = drm_agp_acquire(rdev->ddev);
@@ -151,9 +144,19 @@ int radeon_agp_init(struct radeon_device *rdev)
 
 	ret = drm_agp_info(rdev->ddev, &info);
 	if (ret) {
+		drm_agp_release(rdev->ddev);
 		DRM_ERROR("Unable to get AGP info: %d\n", ret);
 		return ret;
 	}
+
+	if (rdev->ddev->agp->agp_info.aper_size < 32) {
+		drm_agp_release(rdev->ddev);
+		dev_warn(rdev->dev, "AGP aperture too small (%zuM) "
+			"need at least 32M, disabling AGP\n",
+			rdev->ddev->agp->agp_info.aper_size);
+		return -EINVAL;
+	}
+
 	mode.mode = info.mode;
 	agp_status = (RREG32(RADEON_AGP_STATUS) | RADEON_AGPv3_MODE) & mode.mode;
 	is_v3 = !!(agp_status & RADEON_AGPv3_MODE);
@@ -228,6 +231,7 @@ int radeon_agp_init(struct radeon_device *rdev)
 	ret = drm_agp_enable(rdev->ddev, mode);
 	if (ret) {
 		DRM_ERROR("Unable to enable AGP (mode = 0x%lx)\n", mode.mode);
+		drm_agp_release(rdev->ddev);
 		return ret;
 	}
 
diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h
index f2fbd2e4e9df..05ee1aeac3fd 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.h
+++ b/drivers/gpu/drm/radeon/radeon_asic.h
@@ -117,6 +117,7 @@ static struct radeon_asic r100_asic = {
 	.hpd_fini = &r100_hpd_fini,
 	.hpd_sense = &r100_hpd_sense,
 	.hpd_set_polarity = &r100_hpd_set_polarity,
+	.ioctl_wait_idle = NULL,
 };
 
 
@@ -176,6 +177,7 @@ static struct radeon_asic r300_asic = {
 	.hpd_fini = &r100_hpd_fini,
 	.hpd_sense = &r100_hpd_sense,
 	.hpd_set_polarity = &r100_hpd_set_polarity,
+	.ioctl_wait_idle = NULL,
 };
 
 /*
@@ -219,6 +221,7 @@ static struct radeon_asic r420_asic = {
 	.hpd_fini = &r100_hpd_fini,
 	.hpd_sense = &r100_hpd_sense,
 	.hpd_set_polarity = &r100_hpd_set_polarity,
+	.ioctl_wait_idle = NULL,
 };
 
 
@@ -267,6 +270,7 @@ static struct radeon_asic rs400_asic = {
 	.hpd_fini = &r100_hpd_fini,
 	.hpd_sense = &r100_hpd_sense,
 	.hpd_set_polarity = &r100_hpd_set_polarity,
+	.ioctl_wait_idle = NULL,
 };
 
 
@@ -323,6 +327,7 @@ static struct radeon_asic rs600_asic = {
 	.hpd_fini = &rs600_hpd_fini,
 	.hpd_sense = &rs600_hpd_sense,
 	.hpd_set_polarity = &rs600_hpd_set_polarity,
+	.ioctl_wait_idle = NULL,
 };
 
 
@@ -370,6 +375,7 @@ static struct radeon_asic rs690_asic = {
 	.hpd_fini = &rs600_hpd_fini,
 	.hpd_sense = &rs600_hpd_sense,
 	.hpd_set_polarity = &rs600_hpd_set_polarity,
+	.ioctl_wait_idle = NULL,
 };
 
 
@@ -421,6 +427,7 @@ static struct radeon_asic rv515_asic = {
 	.hpd_fini = &rs600_hpd_fini,
 	.hpd_sense = &rs600_hpd_sense,
 	.hpd_set_polarity = &rs600_hpd_set_polarity,
+	.ioctl_wait_idle = NULL,
 };
 
 
@@ -463,6 +470,7 @@ static struct radeon_asic r520_asic = {
 	.hpd_fini = &rs600_hpd_fini,
 	.hpd_sense = &rs600_hpd_sense,
 	.hpd_set_polarity = &rs600_hpd_set_polarity,
+	.ioctl_wait_idle = NULL,
 };
 
 /*
@@ -504,6 +512,7 @@ void r600_hpd_fini(struct radeon_device *rdev);
 bool r600_hpd_sense(struct radeon_device *rdev, enum radeon_hpd_id hpd);
 void r600_hpd_set_polarity(struct radeon_device *rdev,
 			   enum radeon_hpd_id hpd);
+extern void r600_ioctl_wait_idle(struct radeon_device *rdev, struct radeon_bo *bo);
 
 static struct radeon_asic r600_asic = {
 	.init = &r600_init,
@@ -538,6 +547,7 @@ static struct radeon_asic r600_asic = {
 	.hpd_fini = &r600_hpd_fini,
 	.hpd_sense = &r600_hpd_sense,
 	.hpd_set_polarity = &r600_hpd_set_polarity,
+	.ioctl_wait_idle = r600_ioctl_wait_idle,
 };
 
 /*
@@ -582,6 +592,7 @@ static struct radeon_asic rv770_asic = {
 	.hpd_fini = &r600_hpd_fini,
 	.hpd_sense = &r600_hpd_sense,
 	.hpd_set_polarity = &r600_hpd_set_polarity,
+	.ioctl_wait_idle = r600_ioctl_wait_idle,
 };
 
 #endif
diff --git a/drivers/gpu/drm/radeon/radeon_atombios.c b/drivers/gpu/drm/radeon/radeon_atombios.c
index fa82ca74324e..4d8831548a5f 100644
--- a/drivers/gpu/drm/radeon/radeon_atombios.c
+++ b/drivers/gpu/drm/radeon/radeon_atombios.c
@@ -206,6 +206,15 @@ static bool radeon_atom_apply_quirks(struct drm_device *dev,
 			*connector_type = DRM_MODE_CONNECTOR_DVID;
 	}
 
+	/* Asrock RS600 board lists the DVI port as HDMI */
+	if ((dev->pdev->device == 0x7941) &&
+	    (dev->pdev->subsystem_vendor == 0x1849) &&
+	    (dev->pdev->subsystem_device == 0x7941)) {
+		if ((*connector_type == DRM_MODE_CONNECTOR_HDMIA) &&
+		    (supported_device == ATOM_DEVICE_DFP3_SUPPORT))
+			*connector_type = DRM_MODE_CONNECTOR_DVID;
+	}
+
 	/* a-bit f-i90hd - ciaranm on #radeonhd - this board has no DVI */
 	if ((dev->pdev->device == 0x7941) &&
 	    (dev->pdev->subsystem_vendor == 0x147b) &&
@@ -287,6 +296,15 @@ static bool radeon_atom_apply_quirks(struct drm_device *dev,
 			*connector_type = DRM_MODE_CONNECTOR_DVID;
 	}
 
+	/* XFX Pine Group device rv730 reports no VGA DDC lines
+	 * even though they are wired up to record 0x93
+	 */
+	if ((dev->pdev->device == 0x9498) &&
+	    (dev->pdev->subsystem_vendor == 0x1682) &&
+	    (dev->pdev->subsystem_device == 0x2452)) {
+		struct radeon_device *rdev = dev->dev_private;
+		*i2c_bus = radeon_lookup_i2c_gpio(rdev, 0x93);
+	}
 	return true;
 }
 
diff --git a/drivers/gpu/drm/radeon/radeon_benchmark.c b/drivers/gpu/drm/radeon/radeon_benchmark.c
index 4ddfd4b5bc51..7932dc4d6b90 100644
--- a/drivers/gpu/drm/radeon/radeon_benchmark.c
+++ b/drivers/gpu/drm/radeon/radeon_benchmark.c
@@ -65,31 +65,42 @@ void radeon_benchmark_move(struct radeon_device *rdev, unsigned bsize,
 	if (r) {
 		goto out_cleanup;
 	}
-	start_jiffies = jiffies;
-	for (i = 0; i < n; i++) {
-		r = radeon_fence_create(rdev, &fence);
-		if (r) {
-			goto out_cleanup;
+
+	/* r100 doesn't have dma engine so skip the test */
+	if (rdev->asic->copy_dma) {
+
+		start_jiffies = jiffies;
+		for (i = 0; i < n; i++) {
+			r = radeon_fence_create(rdev, &fence);
+			if (r) {
+				goto out_cleanup;
+			}
+
+			r = radeon_copy_dma(rdev, saddr, daddr,
+					size / RADEON_GPU_PAGE_SIZE, fence);
+
+			if (r) {
+				goto out_cleanup;
+			}
+			r = radeon_fence_wait(fence, false);
+			if (r) {
+				goto out_cleanup;
+			}
+			radeon_fence_unref(&fence);
 		}
-		r = radeon_copy_dma(rdev, saddr, daddr, size / RADEON_GPU_PAGE_SIZE, fence);
-		if (r) {
-			goto out_cleanup;
+		end_jiffies = jiffies;
+		time = end_jiffies - start_jiffies;
+		time = jiffies_to_msecs(time);
+		if (time > 0) {
+			i = ((n * size) >> 10) / time;
+			printk(KERN_INFO "radeon: dma %u bo moves of %ukb from"
+					" %d to %d in %lums (%ukb/ms %ukb/s %uM/s)\n",
+					n, size >> 10,
+					sdomain, ddomain, time,
+					i, i * 1000, (i * 1000) / 1024);
 		}
-		r = radeon_fence_wait(fence, false);
-		if (r) {
-			goto out_cleanup;
-		}
-		radeon_fence_unref(&fence);
-	}
-	end_jiffies = jiffies;
-	time = end_jiffies - start_jiffies;
-	time = jiffies_to_msecs(time);
-	if (time > 0) {
-		i = ((n * size) >> 10) / time;
-		printk(KERN_INFO "radeon: dma %u bo moves of %ukb from %d to %d"
-		       " in %lums (%ukb/ms %ukb/s %uM/s)\n", n, size >> 10,
-		       sdomain, ddomain, time, i, i * 1000, (i * 1000) / 1024);
 	}
+
 	start_jiffies = jiffies;
 	for (i = 0; i < n; i++) {
 		r = radeon_fence_create(rdev, &fence);
diff --git a/drivers/gpu/drm/radeon/radeon_combios.c b/drivers/gpu/drm/radeon/radeon_combios.c
index 579c8920e081..e7b19440102e 100644
--- a/drivers/gpu/drm/radeon/radeon_combios.c
+++ b/drivers/gpu/drm/radeon/radeon_combios.c
@@ -971,8 +971,7 @@ struct radeon_encoder_lvds *radeon_combios_get_lvds_info(struct radeon_encoder
 			 lvds->native_mode.vdisplay);
 
 		lvds->panel_vcc_delay = RBIOS16(lcd_info + 0x2c);
-		if (lvds->panel_vcc_delay > 2000 || lvds->panel_vcc_delay < 0)
-			lvds->panel_vcc_delay = 2000;
+		lvds->panel_vcc_delay = min_t(u16, lvds->panel_vcc_delay, 2000);
 
 		lvds->panel_pwr_delay = RBIOS8(lcd_info + 0x24);
 		lvds->panel_digon_delay = RBIOS16(lcd_info + 0x38) & 0xf;
diff --git a/drivers/gpu/drm/radeon/radeon_connectors.c b/drivers/gpu/drm/radeon/radeon_connectors.c
index 55266416fa47..65f81942f399 100644
--- a/drivers/gpu/drm/radeon/radeon_connectors.c
+++ b/drivers/gpu/drm/radeon/radeon_connectors.c
@@ -580,16 +580,18 @@ static enum drm_connector_status radeon_vga_detect(struct drm_connector *connect
 	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
 	struct drm_encoder *encoder;
 	struct drm_encoder_helper_funcs *encoder_funcs;
-	bool dret;
+	bool dret = false;
 	enum drm_connector_status ret = connector_status_disconnected;
 
 	encoder = radeon_best_single_encoder(connector);
 	if (!encoder)
 		ret = connector_status_disconnected;
 
-	radeon_i2c_do_lock(radeon_connector->ddc_bus, 1);
-	dret = radeon_ddc_probe(radeon_connector);
-	radeon_i2c_do_lock(radeon_connector->ddc_bus, 0);
+	if (radeon_connector->ddc_bus) {
+		radeon_i2c_do_lock(radeon_connector->ddc_bus, 1);
+		dret = radeon_ddc_probe(radeon_connector);
+		radeon_i2c_do_lock(radeon_connector->ddc_bus, 0);
+	}
 	if (dret) {
 		if (radeon_connector->edid) {
 			kfree(radeon_connector->edid);
@@ -740,11 +742,13 @@ static enum drm_connector_status radeon_dvi_detect(struct drm_connector *connect
 	struct drm_mode_object *obj;
 	int i;
 	enum drm_connector_status ret = connector_status_disconnected;
-	bool dret;
+	bool dret = false;
 
-	radeon_i2c_do_lock(radeon_connector->ddc_bus, 1);
-	dret = radeon_ddc_probe(radeon_connector);
-	radeon_i2c_do_lock(radeon_connector->ddc_bus, 0);
+	if (radeon_connector->ddc_bus) {
+		radeon_i2c_do_lock(radeon_connector->ddc_bus, 1);
+		dret = radeon_ddc_probe(radeon_connector);
+		radeon_i2c_do_lock(radeon_connector->ddc_bus, 0);
+	}
 	if (dret) {
 		if (radeon_connector->edid) {
 			kfree(radeon_connector->edid);
@@ -776,7 +780,7 @@ static enum drm_connector_status radeon_dvi_detect(struct drm_connector *connect
 			 * connected and the DVI port disconnected.  If the edid doesn't
 			 * say HDMI, vice versa.
 			 */
-			if (radeon_connector->shared_ddc && connector_status_connected) {
+			if (radeon_connector->shared_ddc && (ret == connector_status_connected)) {
 				struct drm_device *dev = connector->dev;
 				struct drm_connector *list_connector;
 				struct radeon_connector *list_radeon_connector;
@@ -1056,8 +1060,7 @@ radeon_add_atom_connector(struct drm_device *dev,
 			return;
 		}
 		if (radeon_connector->ddc_bus && i2c_bus->valid) {
-			if (memcmp(&radeon_connector->ddc_bus->rec, i2c_bus,
-				    sizeof(struct radeon_i2c_bus_rec)) == 0) {
+			if (radeon_connector->ddc_bus->rec.i2c_id == i2c_bus->i2c_id) {
 				radeon_connector->shared_ddc = true;
 				shared_ddc = true;
 			}
@@ -1343,7 +1346,7 @@ radeon_add_legacy_connector(struct drm_device *dev,
 				radeon_connector->dac_load_detect = false;
 			drm_connector_attach_property(&radeon_connector->base,
 						      rdev->mode_info.load_detect_property,
-						      1);
+						      radeon_connector->dac_load_detect);
 			drm_connector_attach_property(&radeon_connector->base,
 						      rdev->mode_info.tv_std_property,
 						      radeon_combios_get_tv_info(rdev));
diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c
index 1496cb8658ef..e9d085021c1f 100644
--- a/drivers/gpu/drm/radeon/radeon_cs.c
+++ b/drivers/gpu/drm/radeon/radeon_cs.c
@@ -86,7 +86,7 @@ int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 						&p->validated);
 		}
 	}
-	return radeon_bo_list_validate(&p->validated, p->ib->fence);
+	return radeon_bo_list_validate(&p->validated);
 }
 
 int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
@@ -189,12 +189,10 @@ static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error)
 {
 	unsigned i;
 
-	if (error) {
-		radeon_bo_list_unvalidate(&parser->validated,
-						parser->ib->fence);
-	} else {
-		radeon_bo_list_unreserve(&parser->validated);
+	if (!error && parser->ib) {
+		radeon_bo_list_fence(&parser->validated, parser->ib->fence);
 	}
+	radeon_bo_list_unreserve(&parser->validated);
 	for (i = 0; i < parser->nrelocs; i++) {
 		if (parser->relocs[i].gobj) {
 			mutex_lock(&parser->rdev->ddev->struct_mutex);
diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c
index 6a92f994cc26..7e17a362b54b 100644
--- a/drivers/gpu/drm/radeon/radeon_display.c
+++ b/drivers/gpu/drm/radeon/radeon_display.c
@@ -278,7 +278,7 @@ static void radeon_print_display_setup(struct drm_device *dev)
 		DRM_INFO("  %s\n", connector_names[connector->connector_type]);
 		if (radeon_connector->hpd.hpd != RADEON_HPD_NONE)
 			DRM_INFO("  %s\n", hpd_names[radeon_connector->hpd.hpd]);
-		if (radeon_connector->ddc_bus)
+		if (radeon_connector->ddc_bus) {
 			DRM_INFO("  DDC: 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x\n",
 				 radeon_connector->ddc_bus->rec.mask_clk_reg,
 				 radeon_connector->ddc_bus->rec.mask_data_reg,
@@ -288,6 +288,15 @@ static void radeon_print_display_setup(struct drm_device *dev)
 				 radeon_connector->ddc_bus->rec.en_data_reg,
 				 radeon_connector->ddc_bus->rec.y_clk_reg,
 				 radeon_connector->ddc_bus->rec.y_data_reg);
+		} else {
+			if (connector->connector_type == DRM_MODE_CONNECTOR_VGA ||
+			    connector->connector_type == DRM_MODE_CONNECTOR_DVII ||
+			    connector->connector_type == DRM_MODE_CONNECTOR_DVID ||
+			    connector->connector_type == DRM_MODE_CONNECTOR_DVIA ||
+			    connector->connector_type == DRM_MODE_CONNECTOR_HDMIA ||
+			    connector->connector_type == DRM_MODE_CONNECTOR_HDMIB)
+				DRM_INFO("  DDC: no ddc bus - possible BIOS bug - please report to xorg-driver-ati@lists.x.org\n");
+		}
 		DRM_INFO("  Encoders:\n");
 		list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
 			radeon_encoder = to_radeon_encoder(encoder);
diff --git a/drivers/gpu/drm/radeon/radeon_drv.h b/drivers/gpu/drm/radeon/radeon_drv.h
index e13785282a82..c57ad606504d 100644
--- a/drivers/gpu/drm/radeon/radeon_drv.h
+++ b/drivers/gpu/drm/radeon/radeon_drv.h
@@ -106,9 +106,10 @@
  * 1.29- R500 3D cmd buffer support
  * 1.30- Add support for occlusion queries
  * 1.31- Add support for num Z pipes from GET_PARAM
+ * 1.32- fixes for rv740 setup
  */
 #define DRIVER_MAJOR		1
-#define DRIVER_MINOR		31
+#define DRIVER_MINOR		32
 #define DRIVER_PATCHLEVEL	0
 
 enum radeon_cp_microcode_version {
diff --git a/drivers/gpu/drm/radeon/radeon_encoders.c b/drivers/gpu/drm/radeon/radeon_encoders.c
index 82eb551970b9..3c91724457ca 100644
--- a/drivers/gpu/drm/radeon/radeon_encoders.c
+++ b/drivers/gpu/drm/radeon/radeon_encoders.c
@@ -156,6 +156,26 @@ radeon_get_encoder_id(struct drm_device *dev, uint32_t supported_device, uint8_t
 	return ret;
 }
 
+static inline bool radeon_encoder_is_digital(struct drm_encoder *encoder)
+{
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	switch (radeon_encoder->encoder_id) {
+	case ENCODER_OBJECT_ID_INTERNAL_LVDS:
+	case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
+	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
+	case ENCODER_OBJECT_ID_INTERNAL_DVO1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
+	case ENCODER_OBJECT_ID_INTERNAL_DDI:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+		return true;
+	default:
+		return false;
+	}
+}
 void
 radeon_link_encoder_connector(struct drm_device *dev)
 {
@@ -202,7 +222,7 @@ radeon_get_connector_for_encoder(struct drm_encoder *encoder)
 
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 		radeon_connector = to_radeon_connector(connector);
-		if (radeon_encoder->devices & radeon_connector->devices)
+		if (radeon_encoder->active_device & radeon_connector->devices)
 			return connector;
 	}
 	return NULL;
@@ -676,31 +696,11 @@ atombios_dig_encoder_setup(struct drm_encoder *encoder, int action)
 
 	memset(&args, 0, sizeof(args));
 
-	if (ASIC_IS_DCE32(rdev)) {
-		if (dig->dig_block)
-			index = GetIndexIntoMasterTable(COMMAND, DIG2EncoderControl);
-		else
-			index = GetIndexIntoMasterTable(COMMAND, DIG1EncoderControl);
-		num = dig->dig_block + 1;
-	} else {
-		switch (radeon_encoder->encoder_id) {
-		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-			/* XXX doesn't really matter which dig encoder we pick as long as it's
-			 * not already in use
-			 */
-			if (dig_connector->linkb)
-				index = GetIndexIntoMasterTable(COMMAND, DIG2EncoderControl);
-			else
-				index = GetIndexIntoMasterTable(COMMAND, DIG1EncoderControl);
-			num = 1;
-			break;
-		case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
-			/* Only dig2 encoder can drive LVTMA */
-			index = GetIndexIntoMasterTable(COMMAND, DIG2EncoderControl);
-			num = 2;
-			break;
-		}
-	}
+	if (dig->dig_encoder)
+		index = GetIndexIntoMasterTable(COMMAND, DIG2EncoderControl);
+	else
+		index = GetIndexIntoMasterTable(COMMAND, DIG1EncoderControl);
+	num = dig->dig_encoder + 1;
 
 	atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev);
 
@@ -822,7 +822,7 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t
 			args.v1.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
 	}
 	if (ASIC_IS_DCE32(rdev)) {
-		if (dig->dig_block)
+		if (dig->dig_encoder == 1)
 			args.v2.acConfig.ucEncoderSel = 1;
 		if (dig_connector->linkb)
 			args.v2.acConfig.ucLinkSel = 1;
@@ -849,17 +849,16 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t
 				args.v2.acConfig.fCoherentMode = 1;
 		}
 	} else {
+
 		args.v1.ucConfig = ATOM_TRANSMITTER_CONFIG_CLKSRC_PPLL;
 
+		if (dig->dig_encoder)
+			args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_DIG2_ENCODER;
+		else
+			args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_DIG1_ENCODER;
+
 		switch (radeon_encoder->encoder_id) {
 		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-			/* XXX doesn't really matter which dig encoder we pick as long as it's
-			 * not already in use
-			 */
-			if (dig_connector->linkb)
-				args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_DIG2_ENCODER;
-			else
-				args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_DIG1_ENCODER;
 			if (rdev->flags & RADEON_IS_IGP) {
 				if (radeon_encoder->pixel_clock > 165000) {
 					if (dig_connector->igp_lane_info & 0x3)
@@ -878,10 +877,6 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t
 				}
 			}
 			break;
-		case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
-			/* Only dig2 encoder can drive LVTMA */
-			args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_DIG2_ENCODER;
-			break;
 		}
 
 		if (radeon_encoder->pixel_clock > 165000)
@@ -1046,6 +1041,7 @@ atombios_set_encoder_crtc_source(struct drm_encoder *encoder)
 	union crtc_sourc_param args;
 	int index = GetIndexIntoMasterTable(COMMAND, SelectCRTC_Source);
 	uint8_t frev, crev;
+	struct radeon_encoder_atom_dig *dig;
 
 	memset(&args, 0, sizeof(args));
 
@@ -1109,40 +1105,16 @@ atombios_set_encoder_crtc_source(struct drm_encoder *encoder)
 			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
 			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
 			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-				if (ASIC_IS_DCE32(rdev)) {
-					if (radeon_crtc->crtc_id)
-						args.v2.ucEncoderID = ASIC_INT_DIG2_ENCODER_ID;
-					else
-						args.v2.ucEncoderID = ASIC_INT_DIG1_ENCODER_ID;
-				} else {
-					struct drm_connector *connector;
-					struct radeon_connector *radeon_connector;
-					struct radeon_connector_atom_dig *dig_connector;
-
-					connector = radeon_get_connector_for_encoder(encoder);
-					if (!connector)
-						return;
-					radeon_connector = to_radeon_connector(connector);
-					if (!radeon_connector->con_priv)
-						return;
-					dig_connector = radeon_connector->con_priv;
-
-					/* XXX doesn't really matter which dig encoder we pick as long as it's
-					 * not already in use
-					 */
-					if (dig_connector->linkb)
-						args.v2.ucEncoderID = ASIC_INT_DIG2_ENCODER_ID;
-					else
-						args.v2.ucEncoderID = ASIC_INT_DIG1_ENCODER_ID;
-				}
+			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
+				dig = radeon_encoder->enc_priv;
+				if (dig->dig_encoder)
+					args.v2.ucEncoderID = ASIC_INT_DIG2_ENCODER_ID;
+				else
+					args.v2.ucEncoderID = ASIC_INT_DIG1_ENCODER_ID;
 				break;
 			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
 				args.v2.ucEncoderID = ASIC_INT_DVO_ENCODER_ID;
 				break;
-			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
-				/* Only dig2 encoder can drive LVTMA */
-				args.v2.ucEncoderID = ASIC_INT_DIG2_ENCODER_ID;
-				break;
 			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1:
 				if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT))
 					args.v2.ucEncoderID = ASIC_INT_TV_ENCODER_ID;
@@ -1202,6 +1174,47 @@ atombios_apply_encoder_quirks(struct drm_encoder *encoder,
 	}
 }
 
+static int radeon_atom_pick_dig_encoder(struct drm_encoder *encoder)
+{
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc);
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct drm_encoder *test_encoder;
+	struct radeon_encoder_atom_dig *dig;
+	uint32_t dig_enc_in_use = 0;
+	/* on DCE32 and encoder can driver any block so just crtc id */
+	if (ASIC_IS_DCE32(rdev)) {
+		return radeon_crtc->crtc_id;
+	}
+
+	/* on DCE3 - LVTMA can only be driven by DIGB */
+	list_for_each_entry(test_encoder, &dev->mode_config.encoder_list, head) {
+		struct radeon_encoder *radeon_test_encoder;
+
+		if (encoder == test_encoder)
+			continue;
+
+		if (!radeon_encoder_is_digital(test_encoder))
+			continue;
+
+		radeon_test_encoder = to_radeon_encoder(test_encoder);
+		dig = radeon_test_encoder->enc_priv;
+
+		if (dig->dig_encoder >= 0)
+			dig_enc_in_use |= (1 << dig->dig_encoder);
+	}
+
+	if (radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA) {
+		if (dig_enc_in_use & 0x2)
+			DRM_ERROR("LVDS required digital encoder 2 but it was in use - stealing\n");
+		return 1;
+	}
+	if (!(dig_enc_in_use & 1))
+		return 0;
+	return 1;
+}
+
 static void
 radeon_atom_encoder_mode_set(struct drm_encoder *encoder,
 			     struct drm_display_mode *mode,
@@ -1214,12 +1227,9 @@ radeon_atom_encoder_mode_set(struct drm_encoder *encoder,
 
 	if (radeon_encoder->active_device &
 	    (ATOM_DEVICE_DFP_SUPPORT | ATOM_DEVICE_LCD_SUPPORT)) {
-		if (radeon_encoder->enc_priv) {
-			struct radeon_encoder_atom_dig *dig;
-
-			dig = radeon_encoder->enc_priv;
-			dig->dig_block = radeon_crtc->crtc_id;
-		}
+		struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
+		if (dig)
+			dig->dig_encoder = radeon_atom_pick_dig_encoder(encoder);
 	}
 	radeon_encoder->pixel_clock = adjusted_mode->clock;
 
@@ -1379,7 +1389,13 @@ static void radeon_atom_encoder_commit(struct drm_encoder *encoder)
 static void radeon_atom_encoder_disable(struct drm_encoder *encoder)
 {
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct radeon_encoder_atom_dig *dig;
 	radeon_atom_encoder_dpms(encoder, DRM_MODE_DPMS_OFF);
+
+	if (radeon_encoder_is_digital(encoder)) {
+		dig = radeon_encoder->enc_priv;
+		dig->dig_encoder = -1;
+	}
 	radeon_encoder->active_device = 0;
 }
 
@@ -1436,6 +1452,7 @@ radeon_atombios_set_dig_info(struct radeon_encoder *radeon_encoder)
 
 	/* coherent mode by default */
 	dig->coherent_mode = true;
+	dig->dig_encoder = -1;
 
 	return dig;
 }
diff --git a/drivers/gpu/drm/radeon/radeon_fb.c b/drivers/gpu/drm/radeon/radeon_fb.c
index 3ba213d1b06c..d71e346e9ab5 100644
--- a/drivers/gpu/drm/radeon/radeon_fb.c
+++ b/drivers/gpu/drm/radeon/radeon_fb.c
@@ -248,7 +248,7 @@ int radeonfb_create(struct drm_device *dev,
 	if (ret)
 		goto out_unref;
 
-	memset_io(fbptr, 0xff, aligned_size);
+	memset_io(fbptr, 0x0, aligned_size);
 
 	strcpy(info->fix.id, "radeondrmfb");
 
diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c
index 0e1325e18534..db8e9a355a01 100644
--- a/drivers/gpu/drm/radeon/radeon_gem.c
+++ b/drivers/gpu/drm/radeon/radeon_gem.c
@@ -308,6 +308,9 @@ int radeon_gem_wait_idle_ioctl(struct drm_device *dev, void *data,
 	}
 	robj = gobj->driver_private;
 	r = radeon_bo_wait(robj, NULL, false);
+	/* callback hw specific functions if any */
+	if (robj->rdev->asic->ioctl_wait_idle)
+		robj->rdev->asic->ioctl_wait_idle(robj->rdev, robj);
 	mutex_lock(&dev->struct_mutex);
 	drm_gem_object_unreference(gobj);
 	mutex_unlock(&dev->struct_mutex);
diff --git a/drivers/gpu/drm/radeon/radeon_mode.h b/drivers/gpu/drm/radeon/radeon_mode.h
index 96b851f92f4c..e81b2aeb6a8f 100644
--- a/drivers/gpu/drm/radeon/radeon_mode.h
+++ b/drivers/gpu/drm/radeon/radeon_mode.h
@@ -299,7 +299,7 @@ struct radeon_atom_ss {
 struct radeon_encoder_atom_dig {
 	/* atom dig */
 	bool coherent_mode;
-	int dig_block;
+	int dig_encoder; /* -1 disabled, 0 DIGA, 1 DIGB */
 	/* atom lvds */
 	uint32_t lvds_misc;
 	uint16_t panel_pwr_delay;
diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c
index d72a71bff218..f1da370928eb 100644
--- a/drivers/gpu/drm/radeon/radeon_object.c
+++ b/drivers/gpu/drm/radeon/radeon_object.c
@@ -306,11 +306,10 @@ void radeon_bo_list_unreserve(struct list_head *head)
 	}
 }
 
-int radeon_bo_list_validate(struct list_head *head, void *fence)
+int radeon_bo_list_validate(struct list_head *head)
 {
 	struct radeon_bo_list *lobj;
 	struct radeon_bo *bo;
-	struct radeon_fence *old_fence = NULL;
 	int r;
 
 	r = radeon_bo_list_reserve(head);
@@ -334,32 +333,27 @@ int radeon_bo_list_validate(struct list_head *head, void *fence)
 		}
 		lobj->gpu_offset = radeon_bo_gpu_offset(bo);
 		lobj->tiling_flags = bo->tiling_flags;
-		if (fence) {
-			old_fence = (struct radeon_fence *)bo->tbo.sync_obj;
-			bo->tbo.sync_obj = radeon_fence_ref(fence);
-			bo->tbo.sync_obj_arg = NULL;
-		}
-		if (old_fence) {
-			radeon_fence_unref(&old_fence);
-		}
 	}
 	return 0;
 }
 
-void radeon_bo_list_unvalidate(struct list_head *head, void *fence)
+void radeon_bo_list_fence(struct list_head *head, void *fence)
 {
 	struct radeon_bo_list *lobj;
-	struct radeon_fence *old_fence;
-
-	if (fence)
-		list_for_each_entry(lobj, head, list) {
-			old_fence = to_radeon_fence(lobj->bo->tbo.sync_obj);
-			if (old_fence == fence) {
-				lobj->bo->tbo.sync_obj = NULL;
-				radeon_fence_unref(&old_fence);
-			}
+	struct radeon_bo *bo;
+	struct radeon_fence *old_fence = NULL;
+
+	list_for_each_entry(lobj, head, list) {
+		bo = lobj->bo;
+		spin_lock(&bo->tbo.lock);
+		old_fence = (struct radeon_fence *)bo->tbo.sync_obj;
+		bo->tbo.sync_obj = radeon_fence_ref(fence);
+		bo->tbo.sync_obj_arg = NULL;
+		spin_unlock(&bo->tbo.lock);
+		if (old_fence) {
+			radeon_fence_unref(&old_fence);
 		}
-	radeon_bo_list_unreserve(head);
+	}
 }
 
 int radeon_bo_fbdev_mmap(struct radeon_bo *bo,
diff --git a/drivers/gpu/drm/radeon/radeon_object.h b/drivers/gpu/drm/radeon/radeon_object.h
index a02f18011ad1..7ab43de1e244 100644
--- a/drivers/gpu/drm/radeon/radeon_object.h
+++ b/drivers/gpu/drm/radeon/radeon_object.h
@@ -156,8 +156,8 @@ extern void radeon_bo_list_add_object(struct radeon_bo_list *lobj,
 				struct list_head *head);
 extern int radeon_bo_list_reserve(struct list_head *head);
 extern void radeon_bo_list_unreserve(struct list_head *head);
-extern int radeon_bo_list_validate(struct list_head *head, void *fence);
-extern void radeon_bo_list_unvalidate(struct list_head *head, void *fence);
+extern int radeon_bo_list_validate(struct list_head *head);
+extern void radeon_bo_list_fence(struct list_head *head, void *fence);
 extern int radeon_bo_fbdev_mmap(struct radeon_bo *bo,
 				struct vm_area_struct *vma);
 extern int radeon_bo_set_tiling_flags(struct radeon_bo *bo,
diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c
index 4d12b2d17b4d..6579eb4c1f28 100644
--- a/drivers/gpu/drm/radeon/radeon_ring.c
+++ b/drivers/gpu/drm/radeon/radeon_ring.c
@@ -41,68 +41,55 @@ int radeon_ib_get(struct radeon_device *rdev, struct radeon_ib **ib)
 {
 	struct radeon_fence *fence;
 	struct radeon_ib *nib;
-	unsigned long i;
-	int r = 0;
+	int r = 0, i, c;
 
 	*ib = NULL;
 	r = radeon_fence_create(rdev, &fence);
 	if (r) {
-		DRM_ERROR("failed to create fence for new IB\n");
+		dev_err(rdev->dev, "failed to create fence for new IB\n");
 		return r;
 	}
 	mutex_lock(&rdev->ib_pool.mutex);
-	i = find_first_zero_bit(rdev->ib_pool.alloc_bm, RADEON_IB_POOL_SIZE);
-	if (i < RADEON_IB_POOL_SIZE) {
-		set_bit(i, rdev->ib_pool.alloc_bm);
-		rdev->ib_pool.ibs[i].length_dw = 0;
-		*ib = &rdev->ib_pool.ibs[i];
-		mutex_unlock(&rdev->ib_pool.mutex);
-		goto out;
+	for (i = rdev->ib_pool.head_id, c = 0, nib = NULL; c < RADEON_IB_POOL_SIZE; c++, i++) {
+		i &= (RADEON_IB_POOL_SIZE - 1);
+		if (rdev->ib_pool.ibs[i].free) {
+			nib = &rdev->ib_pool.ibs[i];
+			break;
+		}
 	}
-	if (list_empty(&rdev->ib_pool.scheduled_ibs)) {
-		/* we go do nothings here */
+	if (nib == NULL) {
+		/* This should never happen, it means we allocated all
+		 * IB and haven't scheduled one yet, return EBUSY to
+		 * userspace hoping that on ioctl recall we get better
+		 * luck
+		 */
+		dev_err(rdev->dev, "no free indirect buffer !\n");
 		mutex_unlock(&rdev->ib_pool.mutex);
-		DRM_ERROR("all IB allocated none scheduled.\n");
-		r = -EINVAL;
-		goto out;
+		radeon_fence_unref(&fence);
+		return -EBUSY;
 	}
-	/* get the first ib on the scheduled list */
-	nib = list_entry(rdev->ib_pool.scheduled_ibs.next,
-			 struct radeon_ib, list);
-	if (nib->fence == NULL) {
-		/* we go do nothings here */
+	rdev->ib_pool.head_id = (nib->idx + 1) & (RADEON_IB_POOL_SIZE - 1);
+	nib->free = false;
+	if (nib->fence) {
 		mutex_unlock(&rdev->ib_pool.mutex);
-		DRM_ERROR("IB %lu scheduled without a fence.\n", nib->idx);
-		r = -EINVAL;
-		goto out;
-	}
-	mutex_unlock(&rdev->ib_pool.mutex);
-
-	r = radeon_fence_wait(nib->fence, false);
-	if (r) {
-		DRM_ERROR("radeon: IB(%lu:0x%016lX:%u)\n", nib->idx,
-			  (unsigned long)nib->gpu_addr, nib->length_dw);
-		DRM_ERROR("radeon: GPU lockup detected, fail to get a IB\n");
-		goto out;
+		r = radeon_fence_wait(nib->fence, false);
+		if (r) {
+			dev_err(rdev->dev, "error waiting fence of IB(%u:0x%016lX:%u)\n",
+				nib->idx, (unsigned long)nib->gpu_addr, nib->length_dw);
+			mutex_lock(&rdev->ib_pool.mutex);
+			nib->free = true;
+			mutex_unlock(&rdev->ib_pool.mutex);
+			radeon_fence_unref(&fence);
+			return r;
+		}
+		mutex_lock(&rdev->ib_pool.mutex);
 	}
 	radeon_fence_unref(&nib->fence);
-
+	nib->fence = fence;
 	nib->length_dw = 0;
-
-	/* scheduled list is accessed here */
-	mutex_lock(&rdev->ib_pool.mutex);
-	list_del(&nib->list);
-	INIT_LIST_HEAD(&nib->list);
 	mutex_unlock(&rdev->ib_pool.mutex);
-
 	*ib = nib;
-out:
-	if (r) {
-		radeon_fence_unref(&fence);
-	} else {
-		(*ib)->fence = fence;
-	}
-	return r;
+	return 0;
 }
 
 void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib **ib)
@@ -113,19 +100,10 @@ void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib **ib)
 	if (tmp == NULL) {
 		return;
 	}
-	mutex_lock(&rdev->ib_pool.mutex);
-	if (!list_empty(&tmp->list) && !radeon_fence_signaled(tmp->fence)) {
-		/* IB is scheduled & not signaled don't do anythings */
-		mutex_unlock(&rdev->ib_pool.mutex);
-		return;
-	}
-	list_del(&tmp->list);
-	INIT_LIST_HEAD(&tmp->list);
-	if (tmp->fence)
+	if (!tmp->fence->emited)
 		radeon_fence_unref(&tmp->fence);
-
-	tmp->length_dw = 0;
-	clear_bit(tmp->idx, rdev->ib_pool.alloc_bm);
+	mutex_lock(&rdev->ib_pool.mutex);
+	tmp->free = true;
 	mutex_unlock(&rdev->ib_pool.mutex);
 }
 
@@ -135,7 +113,7 @@ int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib)
 
 	if (!ib->length_dw || !rdev->cp.ready) {
 		/* TODO: Nothings in the ib we should report. */
-		DRM_ERROR("radeon: couldn't schedule IB(%lu).\n", ib->idx);
+		DRM_ERROR("radeon: couldn't schedule IB(%u).\n", ib->idx);
 		return -EINVAL;
 	}
 
@@ -148,7 +126,8 @@ int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib)
 	radeon_ring_ib_execute(rdev, ib);
 	radeon_fence_emit(rdev, ib->fence);
 	mutex_lock(&rdev->ib_pool.mutex);
-	list_add_tail(&ib->list, &rdev->ib_pool.scheduled_ibs);
+	/* once scheduled IB is considered free and protected by the fence */
+	ib->free = true;
 	mutex_unlock(&rdev->ib_pool.mutex);
 	radeon_ring_unlock_commit(rdev);
 	return 0;
@@ -164,7 +143,6 @@ int radeon_ib_pool_init(struct radeon_device *rdev)
 	if (rdev->ib_pool.robj)
 		return 0;
 	/* Allocate 1M object buffer */
-	INIT_LIST_HEAD(&rdev->ib_pool.scheduled_ibs);
 	r = radeon_bo_create(rdev, NULL,  RADEON_IB_POOL_SIZE*64*1024,
 				true, RADEON_GEM_DOMAIN_GTT,
 				&rdev->ib_pool.robj);
@@ -195,9 +173,9 @@ int radeon_ib_pool_init(struct radeon_device *rdev)
 		rdev->ib_pool.ibs[i].ptr = ptr + offset;
 		rdev->ib_pool.ibs[i].idx = i;
 		rdev->ib_pool.ibs[i].length_dw = 0;
-		INIT_LIST_HEAD(&rdev->ib_pool.ibs[i].list);
+		rdev->ib_pool.ibs[i].free = true;
 	}
-	bitmap_zero(rdev->ib_pool.alloc_bm, RADEON_IB_POOL_SIZE);
+	rdev->ib_pool.head_id = 0;
 	rdev->ib_pool.ready = true;
 	DRM_INFO("radeon: ib pool ready.\n");
 	if (radeon_debugfs_ib_init(rdev)) {
@@ -214,7 +192,6 @@ void radeon_ib_pool_fini(struct radeon_device *rdev)
 		return;
 	}
 	mutex_lock(&rdev->ib_pool.mutex);
-	bitmap_zero(rdev->ib_pool.alloc_bm, RADEON_IB_POOL_SIZE);
 	if (rdev->ib_pool.robj) {
 		r = radeon_bo_reserve(rdev->ib_pool.robj, false);
 		if (likely(r == 0)) {
@@ -363,7 +340,7 @@ static int radeon_debugfs_ib_info(struct seq_file *m, void *data)
 	if (ib == NULL) {
 		return 0;
 	}
-	seq_printf(m, "IB %04lu\n", ib->idx);
+	seq_printf(m, "IB %04u\n", ib->idx);
 	seq_printf(m, "IB fence %p\n", ib->fence);
 	seq_printf(m, "IB size %05u dwords\n", ib->length_dw);
 	for (i = 0; i < ib->length_dw; i++) {
diff --git a/drivers/gpu/drm/radeon/rs400.c b/drivers/gpu/drm/radeon/rs400.c
index 9f5418983e2a..287fcebfb4e6 100644
--- a/drivers/gpu/drm/radeon/rs400.c
+++ b/drivers/gpu/drm/radeon/rs400.c
@@ -223,15 +223,31 @@ int rs400_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr)
 	return 0;
 }
 
+int rs400_mc_wait_for_idle(struct radeon_device *rdev)
+{
+	unsigned i;
+	uint32_t tmp;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		/* read MC_STATUS */
+		tmp = RREG32(0x0150);
+		if (tmp & (1 << 2)) {
+			return 0;
+		}
+		DRM_UDELAY(1);
+	}
+	return -1;
+}
+
 void rs400_gpu_init(struct radeon_device *rdev)
 {
 	/* FIXME: HDP same place on rs400 ? */
 	r100_hdp_reset(rdev);
 	/* FIXME: is this correct ? */
 	r420_pipes_init(rdev);
-	if (r300_mc_wait_for_idle(rdev)) {
-		printk(KERN_WARNING "Failed to wait MC idle while "
-		       "programming pipes. Bad things might happen.\n");
+	if (rs400_mc_wait_for_idle(rdev)) {
+		printk(KERN_WARNING "rs400: Failed to wait MC idle while "
+		       "programming pipes. Bad things might happen. %08x\n", RREG32(0x150));
 	}
 }
 
@@ -370,8 +386,8 @@ void rs400_mc_program(struct radeon_device *rdev)
 	r100_mc_stop(rdev, &save);
 
 	/* Wait for mc idle */
-	if (r300_mc_wait_for_idle(rdev))
-		dev_warn(rdev->dev, "Wait MC idle timeout before updating MC.\n");
+	if (rs400_mc_wait_for_idle(rdev))
+		dev_warn(rdev->dev, "rs400: Wait MC idle timeout before updating MC.\n");
 	WREG32(R_000148_MC_FB_LOCATION,
 		S_000148_MC_FB_START(rdev->mc.vram_start >> 16) |
 		S_000148_MC_FB_TOP(rdev->mc.vram_end >> 16));
@@ -448,7 +464,6 @@ int rs400_suspend(struct radeon_device *rdev)
 
 void rs400_fini(struct radeon_device *rdev)
 {
-	rs400_suspend(rdev);
 	r100_cp_fini(rdev);
 	r100_wb_fini(rdev);
 	r100_ib_fini(rdev);
@@ -527,7 +542,6 @@ int rs400_init(struct radeon_device *rdev)
 	if (r) {
 		/* Somethings want wront with the accel init stop accel */
 		dev_err(rdev->dev, "Disabling GPU acceleration\n");
-		rs400_suspend(rdev);
 		r100_cp_fini(rdev);
 		r100_wb_fini(rdev);
 		r100_ib_fini(rdev);
diff --git a/drivers/gpu/drm/radeon/rs600.c b/drivers/gpu/drm/radeon/rs600.c
index d5255751e7b3..c3818562a13e 100644
--- a/drivers/gpu/drm/radeon/rs600.c
+++ b/drivers/gpu/drm/radeon/rs600.c
@@ -610,7 +610,6 @@ int rs600_suspend(struct radeon_device *rdev)
 
 void rs600_fini(struct radeon_device *rdev)
 {
-	rs600_suspend(rdev);
 	r100_cp_fini(rdev);
 	r100_wb_fini(rdev);
 	r100_ib_fini(rdev);
@@ -689,7 +688,6 @@ int rs600_init(struct radeon_device *rdev)
 	if (r) {
 		/* Somethings want wront with the accel init stop accel */
 		dev_err(rdev->dev, "Disabling GPU acceleration\n");
-		rs600_suspend(rdev);
 		r100_cp_fini(rdev);
 		r100_wb_fini(rdev);
 		r100_ib_fini(rdev);
diff --git a/drivers/gpu/drm/radeon/rs690.c b/drivers/gpu/drm/radeon/rs690.c
index cd31da913771..06e2771aee5a 100644
--- a/drivers/gpu/drm/radeon/rs690.c
+++ b/drivers/gpu/drm/radeon/rs690.c
@@ -676,7 +676,6 @@ int rs690_suspend(struct radeon_device *rdev)
 
 void rs690_fini(struct radeon_device *rdev)
 {
-	rs690_suspend(rdev);
 	r100_cp_fini(rdev);
 	r100_wb_fini(rdev);
 	r100_ib_fini(rdev);
@@ -756,7 +755,6 @@ int rs690_init(struct radeon_device *rdev)
 	if (r) {
 		/* Somethings want wront with the accel init stop accel */
 		dev_err(rdev->dev, "Disabling GPU acceleration\n");
-		rs690_suspend(rdev);
 		r100_cp_fini(rdev);
 		r100_wb_fini(rdev);
 		r100_ib_fini(rdev);
diff --git a/drivers/gpu/drm/radeon/rv515.c b/drivers/gpu/drm/radeon/rv515.c
index 62756717b044..0e1e6b8632b8 100644
--- a/drivers/gpu/drm/radeon/rv515.c
+++ b/drivers/gpu/drm/radeon/rv515.c
@@ -537,7 +537,6 @@ void rv515_set_safe_registers(struct radeon_device *rdev)
 
 void rv515_fini(struct radeon_device *rdev)
 {
-	rv515_suspend(rdev);
 	r100_cp_fini(rdev);
 	r100_wb_fini(rdev);
 	r100_ib_fini(rdev);
@@ -615,13 +614,12 @@ int rv515_init(struct radeon_device *rdev)
 	if (r) {
 		/* Somethings want wront with the accel init stop accel */
 		dev_err(rdev->dev, "Disabling GPU acceleration\n");
-		rv515_suspend(rdev);
 		r100_cp_fini(rdev);
 		r100_wb_fini(rdev);
 		r100_ib_fini(rdev);
+		radeon_irq_kms_fini(rdev);
 		rv370_pcie_gart_fini(rdev);
 		radeon_agp_fini(rdev);
-		radeon_irq_kms_fini(rdev);
 		rdev->accel_working = false;
 	}
 	return 0;
diff --git a/drivers/gpu/drm/radeon/rv770.c b/drivers/gpu/drm/radeon/rv770.c
index 55f6ffc4e58b..03021674d097 100644
--- a/drivers/gpu/drm/radeon/rv770.c
+++ b/drivers/gpu/drm/radeon/rv770.c
@@ -549,9 +549,12 @@ static void rv770_gpu_init(struct radeon_device *rdev)
 
 	gb_tiling_config |= BANK_SWAPS(1);
 
-	backend_map = r700_get_tile_pipe_to_backend_map(rdev->config.rv770.max_tile_pipes,
-							rdev->config.rv770.max_backends,
-							(0xff << rdev->config.rv770.max_backends) & 0xff);
+	if (rdev->family == CHIP_RV740)
+		backend_map = 0x28;
+	else
+		backend_map = r700_get_tile_pipe_to_backend_map(rdev->config.rv770.max_tile_pipes,
+								rdev->config.rv770.max_backends,
+								(0xff << rdev->config.rv770.max_backends) & 0xff);
 	gb_tiling_config |= BACKEND_MAP(backend_map);
 
 	cc_gc_shader_pipe_config =
@@ -887,26 +890,25 @@ static int rv770_startup(struct radeon_device *rdev)
 			return r;
 	}
 	rv770_gpu_init(rdev);
-
-	if (!rdev->r600_blit.shader_obj) {
-		r = r600_blit_init(rdev);
+	r = r600_blit_init(rdev);
+	if (r) {
+		r600_blit_fini(rdev);
+		rdev->asic->copy = NULL;
+		dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r);
+	}
+	/* pin copy shader into vram */
+	if (rdev->r600_blit.shader_obj) {
+		r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false);
+		if (unlikely(r != 0))
+			return r;
+		r = radeon_bo_pin(rdev->r600_blit.shader_obj, RADEON_GEM_DOMAIN_VRAM,
+				&rdev->r600_blit.shader_gpu_addr);
+		radeon_bo_unreserve(rdev->r600_blit.shader_obj);
 		if (r) {
-			DRM_ERROR("radeon: failed blitter (%d).\n", r);
+			DRM_ERROR("failed to pin blit object %d\n", r);
 			return r;
 		}
 	}
-
-	r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false);
-	if (unlikely(r != 0))
-		return r;
-	r = radeon_bo_pin(rdev->r600_blit.shader_obj, RADEON_GEM_DOMAIN_VRAM,
-			&rdev->r600_blit.shader_gpu_addr);
-	radeon_bo_unreserve(rdev->r600_blit.shader_obj);
-	if (r) {
-		DRM_ERROR("failed to pin blit object %d\n", r);
-		return r;
-	}
-
 	/* Enable IRQ */
 	r = r600_irq_init(rdev);
 	if (r) {
@@ -1066,9 +1068,11 @@ int rv770_init(struct radeon_device *rdev)
 	rdev->accel_working = true;
 	r = rv770_startup(rdev);
 	if (r) {
-		rv770_suspend(rdev);
+		dev_err(rdev->dev, "disabling GPU acceleration\n");
+		r600_cp_fini(rdev);
 		r600_wb_fini(rdev);
-		radeon_ring_fini(rdev);
+		r600_irq_fini(rdev);
+		radeon_irq_kms_fini(rdev);
 		rv770_pcie_gart_fini(rdev);
 		rdev->accel_working = false;
 	}
@@ -1090,13 +1094,11 @@ int rv770_init(struct radeon_device *rdev)
 
 void rv770_fini(struct radeon_device *rdev)
 {
-	rv770_suspend(rdev);
-
 	r600_blit_fini(rdev);
+	r600_cp_fini(rdev);
+	r600_wb_fini(rdev);
 	r600_irq_fini(rdev);
 	radeon_irq_kms_fini(rdev);
-	radeon_ring_fini(rdev);
-	r600_wb_fini(rdev);
 	rv770_pcie_gart_fini(rdev);
 	radeon_gem_fini(rdev);
 	radeon_fence_driver_fini(rdev);