diff --git a/drivers/gpu/drm/radeon/r600_blit.c b/drivers/gpu/drm/radeon/r600_blit.c
index ca5c29f7077..7f1043448d2 100644
--- a/drivers/gpu/drm/radeon/r600_blit.c
+++ b/drivers/gpu/drm/radeon/r600_blit.c
@@ -137,9 +137,9 @@ set_shaders(struct drm_device *dev)
 	ps = (u32 *) ((char *)dev->agp_buffer_map->handle + dev_priv->blit_vb->offset + 256);
 
 	for (i = 0; i < r6xx_vs_size; i++)
-		vs[i] = r6xx_vs[i];
+		vs[i] = cpu_to_le32(r6xx_vs[i]);
 	for (i = 0; i < r6xx_ps_size; i++)
-		ps[i] = r6xx_ps[i];
+		ps[i] = cpu_to_le32(r6xx_ps[i]);
 
 	dev_priv->blit_vb->used = 512;
 
@@ -192,6 +192,9 @@ set_vtx_resource(drm_radeon_private_t *dev_priv, u64 gpu_addr)
 	DRM_DEBUG("\n");
 
 	sq_vtx_constant_word2 = (((gpu_addr >> 32) & 0xff) | (16 << 8));
+#ifdef __BIG_ENDIAN
+	sq_vtx_constant_word2 |= (2 << 30);
+#endif
 
 	BEGIN_RING(9);
 	OUT_RING(CP_PACKET3(R600_IT_SET_RESOURCE, 7));
@@ -291,7 +294,11 @@ draw_auto(drm_radeon_private_t *dev_priv)
 	OUT_RING(DI_PT_RECTLIST);
 
 	OUT_RING(CP_PACKET3(R600_IT_INDEX_TYPE, 0));
+#ifdef __BIG_ENDIAN
+	OUT_RING((2 << 2) | DI_INDEX_SIZE_16_BIT);
+#else
 	OUT_RING(DI_INDEX_SIZE_16_BIT);
+#endif
 
 	OUT_RING(CP_PACKET3(R600_IT_NUM_INSTANCES, 0));
 	OUT_RING(1);
diff --git a/drivers/gpu/drm/radeon/r600_cp.c b/drivers/gpu/drm/radeon/r600_cp.c
index 4f4cd8b286d..c3ab959bdc7 100644
--- a/drivers/gpu/drm/radeon/r600_cp.c
+++ b/drivers/gpu/drm/radeon/r600_cp.c
@@ -396,6 +396,9 @@ static void r600_cp_load_microcode(drm_radeon_private_t *dev_priv)
 	r600_do_cp_stop(dev_priv);
 
 	RADEON_WRITE(R600_CP_RB_CNTL,
+#ifdef __BIG_ENDIAN
+		     R600_BUF_SWAP_32BIT |
+#endif
 		     R600_RB_NO_UPDATE |
 		     R600_RB_BLKSZ(15) |
 		     R600_RB_BUFSZ(3));
@@ -486,9 +489,12 @@ static void r700_cp_load_microcode(drm_radeon_private_t *dev_priv)
 	r600_do_cp_stop(dev_priv);
 
 	RADEON_WRITE(R600_CP_RB_CNTL,
+#ifdef __BIG_ENDIAN
+		     R600_BUF_SWAP_32BIT |
+#endif
 		     R600_RB_NO_UPDATE |
-		     (15 << 8) |
-		     (3 << 0));
+		     R600_RB_BLKSZ(15) |
+		     R600_RB_BUFSZ(3));
 
 	RADEON_WRITE(R600_GRBM_SOFT_RESET, R600_SOFT_RESET_CP);
 	RADEON_READ(R600_GRBM_SOFT_RESET);
@@ -550,8 +556,12 @@ static void r600_test_writeback(drm_radeon_private_t *dev_priv)
 
 	if (!dev_priv->writeback_works) {
 		/* Disable writeback to avoid unnecessary bus master transfer */
-		RADEON_WRITE(R600_CP_RB_CNTL, RADEON_READ(R600_CP_RB_CNTL) |
-			     RADEON_RB_NO_UPDATE);
+		RADEON_WRITE(R600_CP_RB_CNTL,
+#ifdef __BIG_ENDIAN
+			     R600_BUF_SWAP_32BIT |
+#endif
+			     RADEON_READ(R600_CP_RB_CNTL) |
+			     R600_RB_NO_UPDATE);
 		RADEON_WRITE(R600_SCRATCH_UMSK, 0);
 	}
 }
@@ -575,7 +585,11 @@ int r600_do_engine_reset(struct drm_device *dev)
 
 	RADEON_WRITE(R600_CP_RB_WPTR_DELAY, 0);
 	cp_rb_cntl = RADEON_READ(R600_CP_RB_CNTL);
-	RADEON_WRITE(R600_CP_RB_CNTL, R600_RB_RPTR_WR_ENA);
+	RADEON_WRITE(R600_CP_RB_CNTL,
+#ifdef __BIG_ENDIAN
+		     R600_BUF_SWAP_32BIT |
+#endif
+		     R600_RB_RPTR_WR_ENA);
 
 	RADEON_WRITE(R600_CP_RB_RPTR_WR, cp_ptr);
 	RADEON_WRITE(R600_CP_RB_WPTR, cp_ptr);
@@ -1838,7 +1852,10 @@ static void r600_cp_init_ring_buffer(struct drm_device *dev,
 			+ dev_priv->gart_vm_start;
 	}
 	RADEON_WRITE(R600_CP_RB_RPTR_ADDR,
-		     rptr_addr & 0xffffffff);
+#ifdef __BIG_ENDIAN
+		     (2 << 0) |
+#endif
+		     (rptr_addr & 0xfffffffc));
 	RADEON_WRITE(R600_CP_RB_RPTR_ADDR_HI,
 		     upper_32_bits(rptr_addr));
 
@@ -1889,7 +1906,7 @@ static void r600_cp_init_ring_buffer(struct drm_device *dev,
 	{
 		u64 scratch_addr;
 
-		scratch_addr = RADEON_READ(R600_CP_RB_RPTR_ADDR);
+		scratch_addr = RADEON_READ(R600_CP_RB_RPTR_ADDR) & 0xFFFFFFFC;
 		scratch_addr |= ((u64)RADEON_READ(R600_CP_RB_RPTR_ADDR_HI)) << 32;
 		scratch_addr += R600_SCRATCH_REG_OFFSET;
 		scratch_addr >>= 8;
diff --git a/drivers/gpu/drm/radeon/radeon_drv.h b/drivers/gpu/drm/radeon/radeon_drv.h
index 448eba89d1e..5cba46b9779 100644
--- a/drivers/gpu/drm/radeon/radeon_drv.h
+++ b/drivers/gpu/drm/radeon/radeon_drv.h
@@ -1524,6 +1524,7 @@ extern u32 radeon_get_scratch(drm_radeon_private_t *dev_priv, int index);
 #define R600_CP_RB_CNTL                                        0xc104
 #       define R600_RB_BUFSZ(x)                                ((x) << 0)
 #       define R600_RB_BLKSZ(x)                                ((x) << 8)
+#	define R600_BUF_SWAP_32BIT		               (2 << 16)
 #       define R600_RB_NO_UPDATE                               (1 << 27)
 #       define R600_RB_RPTR_WR_ENA                             (1 << 31)
 #define R600_CP_RB_RPTR_WR                                     0xc108