atikms: RC11.2

git-svn-id: svn://kolibrios.org@2005 a494cfbc-eb01-0410-851d-a64ba20cac60

drivers/include/linux/kernel.h

@@ -213,5 +213,13 @@ static inline void *dev_get_drvdata(struct device *dev)
 
 #define in_dbg_master() (0)
 
+#define HZ 100
+
+#define time_after(a,b)         \
+        (typecheck(unsigned long, a) && \
+        typecheck(unsigned long, b) && \
+        ((long)(b) - (long)(a) < 0))
+
+
 #endif
 

drivers/include/syscall.h

@@ -101,7 +101,7 @@ static inline evhandle_t CreateEvent(kevent_t *ev, u32_t flags)
      :"=A"(evh.raw)
      :"S" (ev), "c"(flags)
      :"memory");
-     __asm__ __volatile__ ("":::"ebx","ecx", "esi", "edi");
+     __asm__ __volatile__ ("":::"ebx","ecx","edx","esi", "edi");
 
      return evh;
 };
@@ -112,16 +112,16 @@ static inline void RaiseEvent(evhandle_t evh, u32_t flags, kevent_t *ev)
      "call *__imp__RaiseEvent"
      ::"a"(evh.handle),"b"(evh.euid),"d"(flags),"S" (ev)
      :"memory");
-     __asm__ __volatile__ ("":::"ebx","ecx", "esi", "edi");
+     __asm__ __volatile__ ("":::"ebx","ecx","edx","esi","edi");
 
 };
 
-static inline void WaitEvent(u32_t handle, u32_t euid)
+static inline void WaitEvent(evhandle_t evh)
 {
      __asm__ __volatile__ (
      "call *__imp__WaitEvent"
-     ::"a"(handle),"b"(euid));
+     ::"a"(evh.handle),"b"(evh.euid));
-     __asm__ __volatile__ ("":::"ecx","edx", "esi");
+     __asm__ __volatile__ ("":::"ebx","ecx","edx","esi","edi");
 };
 
 static inline u32_t GetEvent(kevent_t *ev)

drivers/video/drm/radeon/Makefile

@@ -55,6 +55,7 @@ NAME_SRC=							\
 			radeon_device.c				\
 			evergreen.c				\
 			evergreen_blit_shaders.c		\
+			evergreen_blit_kms.c			\
 			cayman_blit_shaders.c			\
 			radeon_clocks.c				\
 			atom.c					\
@@ -62,6 +63,7 @@ NAME_SRC=							\
 			radeon_agp.c				\
 			radeon_asic.c				\
 			radeon_atombios.c			\
+			radeon_benchmark.c			\
 			radeon_bios.c				\
 			radeon_combios.c			\
 			radeon_connectors.c			\
@@ -88,6 +90,8 @@ NAME_SRC=							\
 			r520.c					\
 			r600.c					\
 			r600_audio.c				\
+			r600_blit_kms.c				\
+			r600_blit_shaders.c			\
 			r600_hdmi.c				\
 			rs400.c					\
 			rs600.c					\

drivers/video/drm/radeon/Makefile.lto

@@ -55,6 +55,7 @@ NAME_SRC=							\
 			radeon_device.c				\
 			evergreen.c				\
 			evergreen_blit_shaders.c		\
+			evergreen_blit_kms.c			\
 			cayman_blit_shaders.c			\
 			radeon_clocks.c				\
 			atom.c					\
@@ -62,6 +63,7 @@ NAME_SRC=							\
 			radeon_agp.c				\
 			radeon_asic.c				\
 			radeon_atombios.c			\
+			radeon_benchmark.c			\
 			radeon_bios.c				\
 			radeon_combios.c			\
 			radeon_connectors.c			\
@@ -88,6 +90,8 @@ NAME_SRC=							\
 			r520.c					\
 			r600.c					\
 			r600_audio.c				\
+			r600_blit_kms.c				\
+			r600_blit_shaders.c			\
 			r600_hdmi.c				\
 			rs400.c					\
 			rs600.c					\

drivers/video/drm/radeon/cmdline.c

@@ -6,6 +6,8 @@
 #include "radeon.h"
 #include "radeon_object.h"
 
+extern int radeon_benchmarking;
+
 static int my_atoi(char **cmd)
 {
     char* p = *cmd;
@@ -74,14 +76,18 @@ void parse_cmdline(char *cmdline, videomode_t *mode, char *log, int *kms)
         {
             switch(*p++)
             {
-                case 'm':
+                case 'b':
-                    p = parse_mode(p, mode);
+                    radeon_benchmarking = 1;
                     break;
 
                 case 'l':
                     p = parse_path(p, log);
                     break;
 
+                case 'm':
+                    p = parse_mode(p, mode);
+                    break;
+
                 case 'n':
                     *kms = 0;
             };

drivers/video/drm/radeon/evergreen.c

@@ -319,12 +319,10 @@ void evergreen_hpd_init(struct radeon_device *rdev)
 			break;
 		}
 	}
-//	if (rdev->irq.installed)
+	if (rdev->irq.installed)
-//		evergreen_irq_set(rdev);
+		evergreen_irq_set(rdev);
 }
 
-#if 0
-
 void evergreen_hpd_fini(struct radeon_device *rdev)
 {
 	struct drm_device *dev = rdev->ddev;
@@ -363,7 +361,6 @@ void evergreen_hpd_fini(struct radeon_device *rdev)
 	}
 }
 
-#endif
 /* watermark setup */
 
 static u32 evergreen_line_buffer_adjust(struct radeon_device *rdev,
@@ -978,17 +975,19 @@ void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *sav
 {
 	save->vga_control[0] = RREG32(D1VGA_CONTROL);
 	save->vga_control[1] = RREG32(D2VGA_CONTROL);
-	save->vga_control[2] = RREG32(EVERGREEN_D3VGA_CONTROL);
-	save->vga_control[3] = RREG32(EVERGREEN_D4VGA_CONTROL);
-	save->vga_control[4] = RREG32(EVERGREEN_D5VGA_CONTROL);
-	save->vga_control[5] = RREG32(EVERGREEN_D6VGA_CONTROL);
 	save->vga_render_control = RREG32(VGA_RENDER_CONTROL);
 	save->vga_hdp_control = RREG32(VGA_HDP_CONTROL);
 	save->crtc_control[0] = RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET);
 	save->crtc_control[1] = RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET);
-	if (!(rdev->flags & RADEON_IS_IGP)) {
+	if (rdev->num_crtc >= 4) {
+		save->vga_control[2] = RREG32(EVERGREEN_D3VGA_CONTROL);
+		save->vga_control[3] = RREG32(EVERGREEN_D4VGA_CONTROL);
 	save->crtc_control[2] = RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET);
 	save->crtc_control[3] = RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET);
+	}
+	if (rdev->num_crtc >= 6) {
+		save->vga_control[4] = RREG32(EVERGREEN_D5VGA_CONTROL);
+		save->vga_control[5] = RREG32(EVERGREEN_D6VGA_CONTROL);
 	save->crtc_control[4] = RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET);
 	save->crtc_control[5] = RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET);
 	}
@@ -997,35 +996,45 @@ void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *sav
 	WREG32(VGA_RENDER_CONTROL, 0);
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC0_REGISTER_OFFSET, 1);
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC1_REGISTER_OFFSET, 1);
-	if (!(rdev->flags & RADEON_IS_IGP)) {
+	if (rdev->num_crtc >= 4) {
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC2_REGISTER_OFFSET, 1);
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC3_REGISTER_OFFSET, 1);
+	}
+	if (rdev->num_crtc >= 6) {
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC4_REGISTER_OFFSET, 1);
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC5_REGISTER_OFFSET, 1);
 	}
 	WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, 0);
 	WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, 0);
-	if (!(rdev->flags & RADEON_IS_IGP)) {
+	if (rdev->num_crtc >= 4) {
 	WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET, 0);
 	WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET, 0);
+	}
+	if (rdev->num_crtc >= 6) {
 	WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET, 0);
 	WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, 0);
 	}
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC0_REGISTER_OFFSET, 0);
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC1_REGISTER_OFFSET, 0);
-	if (!(rdev->flags & RADEON_IS_IGP)) {
+	if (rdev->num_crtc >= 4) {
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC2_REGISTER_OFFSET, 0);
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC3_REGISTER_OFFSET, 0);
+	}
+	if (rdev->num_crtc >= 6) {
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC4_REGISTER_OFFSET, 0);
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC5_REGISTER_OFFSET, 0);
 	}
 
 	WREG32(D1VGA_CONTROL, 0);
 	WREG32(D2VGA_CONTROL, 0);
+	if (rdev->num_crtc >= 4) {
 	WREG32(EVERGREEN_D3VGA_CONTROL, 0);
 	WREG32(EVERGREEN_D4VGA_CONTROL, 0);
+	}
+	if (rdev->num_crtc >= 6) {
 	WREG32(EVERGREEN_D5VGA_CONTROL, 0);
 	WREG32(EVERGREEN_D6VGA_CONTROL, 0);
+	}
 }
 
 void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *save)
@@ -1048,7 +1057,7 @@ void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *s
 	WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS + EVERGREEN_CRTC1_REGISTER_OFFSET,
 	       (u32)rdev->mc.vram_start);
 
-	if (!(rdev->flags & RADEON_IS_IGP)) {
+	if (rdev->num_crtc >= 4) {
 	WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH + EVERGREEN_CRTC2_REGISTER_OFFSET,
 	       upper_32_bits(rdev->mc.vram_start));
 	WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS_HIGH + EVERGREEN_CRTC2_REGISTER_OFFSET,
@@ -1066,7 +1075,8 @@ void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *s
 	       (u32)rdev->mc.vram_start);
 	WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS + EVERGREEN_CRTC3_REGISTER_OFFSET,
 	       (u32)rdev->mc.vram_start);
-
+	}
+	if (rdev->num_crtc >= 6) {
 	WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH + EVERGREEN_CRTC4_REGISTER_OFFSET,
 	       upper_32_bits(rdev->mc.vram_start));
 	WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS_HIGH + EVERGREEN_CRTC4_REGISTER_OFFSET,
@@ -1094,31 +1104,41 @@ void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *s
 	/* Restore video state */
 	WREG32(D1VGA_CONTROL, save->vga_control[0]);
 	WREG32(D2VGA_CONTROL, save->vga_control[1]);
+	if (rdev->num_crtc >= 4) {
 	WREG32(EVERGREEN_D3VGA_CONTROL, save->vga_control[2]);
 	WREG32(EVERGREEN_D4VGA_CONTROL, save->vga_control[3]);
+	}
+	if (rdev->num_crtc >= 6) {
 	WREG32(EVERGREEN_D5VGA_CONTROL, save->vga_control[4]);
 	WREG32(EVERGREEN_D6VGA_CONTROL, save->vga_control[5]);
+	}
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC0_REGISTER_OFFSET, 1);
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC1_REGISTER_OFFSET, 1);
-	if (!(rdev->flags & RADEON_IS_IGP)) {
+	if (rdev->num_crtc >= 4) {
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC2_REGISTER_OFFSET, 1);
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC3_REGISTER_OFFSET, 1);
+	}
+	if (rdev->num_crtc >= 6) {
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC4_REGISTER_OFFSET, 1);
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC5_REGISTER_OFFSET, 1);
 	}
 	WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, save->crtc_control[0]);
 	WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, save->crtc_control[1]);
-	if (!(rdev->flags & RADEON_IS_IGP)) {
+	if (rdev->num_crtc >= 4) {
 	WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET, save->crtc_control[2]);
 	WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET, save->crtc_control[3]);
+	}
+	if (rdev->num_crtc >= 6) {
 	WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET, save->crtc_control[4]);
 	WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, save->crtc_control[5]);
 	}
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC0_REGISTER_OFFSET, 0);
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC1_REGISTER_OFFSET, 0);
-	if (!(rdev->flags & RADEON_IS_IGP)) {
+	if (rdev->num_crtc >= 4) {
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC2_REGISTER_OFFSET, 0);
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC3_REGISTER_OFFSET, 0);
+	}
+	if (rdev->num_crtc >= 6) {
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC4_REGISTER_OFFSET, 0);
 	WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC5_REGISTER_OFFSET, 0);
 	}
@@ -1970,7 +1990,7 @@ static void evergreen_gpu_init(struct radeon_device *rdev)
 			gb_backend_map = 0x66442200;
 			break;
 		case CHIP_JUNIPER:
-			gb_backend_map = 0x00006420;
+			gb_backend_map = 0x00002200;
 			break;
 		default:
 			gb_backend_map =
@@ -2410,18 +2430,22 @@ void evergreen_disable_interrupt_state(struct radeon_device *rdev)
 	WREG32(GRBM_INT_CNTL, 0);
 	WREG32(INT_MASK + EVERGREEN_CRTC0_REGISTER_OFFSET, 0);
 	WREG32(INT_MASK + EVERGREEN_CRTC1_REGISTER_OFFSET, 0);
-	if (!(rdev->flags & RADEON_IS_IGP)) {
+	if (rdev->num_crtc >= 4) {
 	WREG32(INT_MASK + EVERGREEN_CRTC2_REGISTER_OFFSET, 0);
 	WREG32(INT_MASK + EVERGREEN_CRTC3_REGISTER_OFFSET, 0);
+	}
+	if (rdev->num_crtc >= 6) {
 	WREG32(INT_MASK + EVERGREEN_CRTC4_REGISTER_OFFSET, 0);
 	WREG32(INT_MASK + EVERGREEN_CRTC5_REGISTER_OFFSET, 0);
 	}
 
 	WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, 0);
 	WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, 0);
-	if (!(rdev->flags & RADEON_IS_IGP)) {
+	if (rdev->num_crtc >= 4) {
 	WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET, 0);
 	WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET, 0);
+	}
+	if (rdev->num_crtc >= 6) {
 	WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET, 0);
 	WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, 0);
 	}
@@ -2443,6 +2467,526 @@ void evergreen_disable_interrupt_state(struct radeon_device *rdev)
 	WREG32(DC_HPD6_INT_CONTROL, tmp);
 
 }
+
+int evergreen_irq_set(struct radeon_device *rdev)
+{
+	u32 cp_int_cntl = CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE;
+	u32 crtc1 = 0, crtc2 = 0, crtc3 = 0, crtc4 = 0, crtc5 = 0, crtc6 = 0;
+	u32 hpd1, hpd2, hpd3, hpd4, hpd5, hpd6;
+	u32 grbm_int_cntl = 0;
+	u32 grph1 = 0, grph2 = 0, grph3 = 0, grph4 = 0, grph5 = 0, grph6 = 0;
+
+	if (!rdev->irq.installed) {
+		WARN(1, "Can't enable IRQ/MSI because no handler is installed\n");
+		return -EINVAL;
+	}
+	/* don't enable anything if the ih is disabled */
+	if (!rdev->ih.enabled) {
+		r600_disable_interrupts(rdev);
+		/* force the active interrupt state to all disabled */
+		evergreen_disable_interrupt_state(rdev);
+		return 0;
+	}
+
+	hpd1 = RREG32(DC_HPD1_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	hpd2 = RREG32(DC_HPD2_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	hpd3 = RREG32(DC_HPD3_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	hpd4 = RREG32(DC_HPD4_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	hpd5 = RREG32(DC_HPD5_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	hpd6 = RREG32(DC_HPD6_INT_CONTROL) & ~DC_HPDx_INT_EN;
+
+	if (rdev->irq.sw_int) {
+		DRM_DEBUG("evergreen_irq_set: sw int\n");
+		cp_int_cntl |= RB_INT_ENABLE;
+		cp_int_cntl |= TIME_STAMP_INT_ENABLE;
+	}
+	if (rdev->irq.crtc_vblank_int[0] ||
+	    rdev->irq.pflip[0]) {
+		DRM_DEBUG("evergreen_irq_set: vblank 0\n");
+		crtc1 |= VBLANK_INT_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[1] ||
+	    rdev->irq.pflip[1]) {
+		DRM_DEBUG("evergreen_irq_set: vblank 1\n");
+		crtc2 |= VBLANK_INT_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[2] ||
+	    rdev->irq.pflip[2]) {
+		DRM_DEBUG("evergreen_irq_set: vblank 2\n");
+		crtc3 |= VBLANK_INT_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[3] ||
+	    rdev->irq.pflip[3]) {
+		DRM_DEBUG("evergreen_irq_set: vblank 3\n");
+		crtc4 |= VBLANK_INT_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[4] ||
+	    rdev->irq.pflip[4]) {
+		DRM_DEBUG("evergreen_irq_set: vblank 4\n");
+		crtc5 |= VBLANK_INT_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[5] ||
+	    rdev->irq.pflip[5]) {
+		DRM_DEBUG("evergreen_irq_set: vblank 5\n");
+		crtc6 |= VBLANK_INT_MASK;
+	}
+	if (rdev->irq.hpd[0]) {
+		DRM_DEBUG("evergreen_irq_set: hpd 1\n");
+		hpd1 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.hpd[1]) {
+		DRM_DEBUG("evergreen_irq_set: hpd 2\n");
+		hpd2 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.hpd[2]) {
+		DRM_DEBUG("evergreen_irq_set: hpd 3\n");
+		hpd3 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.hpd[3]) {
+		DRM_DEBUG("evergreen_irq_set: hpd 4\n");
+		hpd4 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.hpd[4]) {
+		DRM_DEBUG("evergreen_irq_set: hpd 5\n");
+		hpd5 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.hpd[5]) {
+		DRM_DEBUG("evergreen_irq_set: hpd 6\n");
+		hpd6 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.gui_idle) {
+		DRM_DEBUG("gui idle\n");
+		grbm_int_cntl |= GUI_IDLE_INT_ENABLE;
+	}
+
+	WREG32(CP_INT_CNTL, cp_int_cntl);
+	WREG32(GRBM_INT_CNTL, grbm_int_cntl);
+
+	WREG32(INT_MASK + EVERGREEN_CRTC0_REGISTER_OFFSET, crtc1);
+	WREG32(INT_MASK + EVERGREEN_CRTC1_REGISTER_OFFSET, crtc2);
+	if (rdev->num_crtc >= 4) {
+		WREG32(INT_MASK + EVERGREEN_CRTC2_REGISTER_OFFSET, crtc3);
+		WREG32(INT_MASK + EVERGREEN_CRTC3_REGISTER_OFFSET, crtc4);
+	}
+	if (rdev->num_crtc >= 6) {
+		WREG32(INT_MASK + EVERGREEN_CRTC4_REGISTER_OFFSET, crtc5);
+		WREG32(INT_MASK + EVERGREEN_CRTC5_REGISTER_OFFSET, crtc6);
+	}
+
+	WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, grph1);
+	WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, grph2);
+	if (rdev->num_crtc >= 4) {
+		WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET, grph3);
+		WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET, grph4);
+	}
+	if (rdev->num_crtc >= 6) {
+		WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET, grph5);
+		WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, grph6);
+	}
+
+	WREG32(DC_HPD1_INT_CONTROL, hpd1);
+	WREG32(DC_HPD2_INT_CONTROL, hpd2);
+	WREG32(DC_HPD3_INT_CONTROL, hpd3);
+	WREG32(DC_HPD4_INT_CONTROL, hpd4);
+	WREG32(DC_HPD5_INT_CONTROL, hpd5);
+	WREG32(DC_HPD6_INT_CONTROL, hpd6);
+
+	return 0;
+}
+
+static inline void evergreen_irq_ack(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	rdev->irq.stat_regs.evergreen.disp_int = RREG32(DISP_INTERRUPT_STATUS);
+	rdev->irq.stat_regs.evergreen.disp_int_cont = RREG32(DISP_INTERRUPT_STATUS_CONTINUE);
+	rdev->irq.stat_regs.evergreen.disp_int_cont2 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE2);
+	rdev->irq.stat_regs.evergreen.disp_int_cont3 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE3);
+	rdev->irq.stat_regs.evergreen.disp_int_cont4 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE4);
+	rdev->irq.stat_regs.evergreen.disp_int_cont5 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE5);
+	rdev->irq.stat_regs.evergreen.d1grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET);
+	rdev->irq.stat_regs.evergreen.d2grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET);
+	if (rdev->num_crtc >= 4) {
+		rdev->irq.stat_regs.evergreen.d3grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET);
+		rdev->irq.stat_regs.evergreen.d4grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET);
+	}
+	if (rdev->num_crtc >= 6) {
+		rdev->irq.stat_regs.evergreen.d5grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET);
+		rdev->irq.stat_regs.evergreen.d6grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET);
+	}
+
+	if (rdev->irq.stat_regs.evergreen.d1grph_int & GRPH_PFLIP_INT_OCCURRED)
+		WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR);
+	if (rdev->irq.stat_regs.evergreen.d2grph_int & GRPH_PFLIP_INT_OCCURRED)
+		WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR);
+	if (rdev->irq.stat_regs.evergreen.disp_int & LB_D1_VBLANK_INTERRUPT)
+		WREG32(VBLANK_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET, VBLANK_ACK);
+	if (rdev->irq.stat_regs.evergreen.disp_int & LB_D1_VLINE_INTERRUPT)
+		WREG32(VLINE_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET, VLINE_ACK);
+	if (rdev->irq.stat_regs.evergreen.disp_int_cont & LB_D2_VBLANK_INTERRUPT)
+		WREG32(VBLANK_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET, VBLANK_ACK);
+	if (rdev->irq.stat_regs.evergreen.disp_int_cont & LB_D2_VLINE_INTERRUPT)
+		WREG32(VLINE_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET, VLINE_ACK);
+
+	if (rdev->num_crtc >= 4) {
+		if (rdev->irq.stat_regs.evergreen.d3grph_int & GRPH_PFLIP_INT_OCCURRED)
+			WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR);
+		if (rdev->irq.stat_regs.evergreen.d4grph_int & GRPH_PFLIP_INT_OCCURRED)
+			WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR);
+		if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & LB_D3_VBLANK_INTERRUPT)
+			WREG32(VBLANK_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET, VBLANK_ACK);
+		if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & LB_D3_VLINE_INTERRUPT)
+			WREG32(VLINE_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET, VLINE_ACK);
+		if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & LB_D4_VBLANK_INTERRUPT)
+			WREG32(VBLANK_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET, VBLANK_ACK);
+		if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & LB_D4_VLINE_INTERRUPT)
+			WREG32(VLINE_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET, VLINE_ACK);
+	}
+
+	if (rdev->num_crtc >= 6) {
+		if (rdev->irq.stat_regs.evergreen.d5grph_int & GRPH_PFLIP_INT_OCCURRED)
+			WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR);
+		if (rdev->irq.stat_regs.evergreen.d6grph_int & GRPH_PFLIP_INT_OCCURRED)
+			WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR);
+		if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & LB_D5_VBLANK_INTERRUPT)
+			WREG32(VBLANK_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET, VBLANK_ACK);
+		if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & LB_D5_VLINE_INTERRUPT)
+			WREG32(VLINE_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET, VLINE_ACK);
+		if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & LB_D6_VBLANK_INTERRUPT)
+			WREG32(VBLANK_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET, VBLANK_ACK);
+		if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & LB_D6_VLINE_INTERRUPT)
+			WREG32(VLINE_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET, VLINE_ACK);
+	}
+
+	if (rdev->irq.stat_regs.evergreen.disp_int & DC_HPD1_INTERRUPT) {
+		tmp = RREG32(DC_HPD1_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD1_INT_CONTROL, tmp);
+	}
+	if (rdev->irq.stat_regs.evergreen.disp_int_cont & DC_HPD2_INTERRUPT) {
+		tmp = RREG32(DC_HPD2_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD2_INT_CONTROL, tmp);
+	}
+	if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & DC_HPD3_INTERRUPT) {
+		tmp = RREG32(DC_HPD3_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD3_INT_CONTROL, tmp);
+	}
+	if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & DC_HPD4_INTERRUPT) {
+		tmp = RREG32(DC_HPD4_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD4_INT_CONTROL, tmp);
+	}
+	if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & DC_HPD5_INTERRUPT) {
+		tmp = RREG32(DC_HPD5_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD5_INT_CONTROL, tmp);
+	}
+	if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & DC_HPD6_INTERRUPT) {
+		tmp = RREG32(DC_HPD5_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD6_INT_CONTROL, tmp);
+	}
+}
+static inline u32 evergreen_get_ih_wptr(struct radeon_device *rdev)
+{
+	u32 wptr, tmp;
+
+	if (rdev->wb.enabled)
+		wptr = le32_to_cpu(rdev->wb.wb[R600_WB_IH_WPTR_OFFSET/4]);
+	else
+		wptr = RREG32(IH_RB_WPTR);
+
+	if (wptr & RB_OVERFLOW) {
+		/* When a ring buffer overflow happen start parsing interrupt
+		 * from the last not overwritten vector (wptr + 16). Hopefully
+		 * this should allow us to catchup.
+		 */
+		dev_warn(rdev->dev, "IH ring buffer overflow (0x%08X, %d, %d)\n",
+			wptr, rdev->ih.rptr, (wptr + 16) + rdev->ih.ptr_mask);
+		rdev->ih.rptr = (wptr + 16) & rdev->ih.ptr_mask;
+		tmp = RREG32(IH_RB_CNTL);
+		tmp |= IH_WPTR_OVERFLOW_CLEAR;
+		WREG32(IH_RB_CNTL, tmp);
+	}
+	return (wptr & rdev->ih.ptr_mask);
+}
+
+int evergreen_irq_process(struct radeon_device *rdev)
+{
+	u32 wptr;
+	u32 rptr;
+	u32 src_id, src_data;
+	u32 ring_index;
+	unsigned long flags;
+	bool queue_hotplug = false;
+
+	if (!rdev->ih.enabled || rdev->shutdown)
+		return IRQ_NONE;
+
+	wptr = evergreen_get_ih_wptr(rdev);
+	rptr = rdev->ih.rptr;
+	DRM_DEBUG("r600_irq_process start: rptr %d, wptr %d\n", rptr, wptr);
+
+	spin_lock_irqsave(&rdev->ih.lock, flags);
+	if (rptr == wptr) {
+		spin_unlock_irqrestore(&rdev->ih.lock, flags);
+		return IRQ_NONE;
+	}
+restart_ih:
+	/* display interrupts */
+	evergreen_irq_ack(rdev);
+
+	rdev->ih.wptr = wptr;
+	while (rptr != wptr) {
+		/* wptr/rptr are in bytes! */
+		ring_index = rptr / 4;
+		src_id =  le32_to_cpu(rdev->ih.ring[ring_index]) & 0xff;
+		src_data = le32_to_cpu(rdev->ih.ring[ring_index + 1]) & 0xfffffff;
+
+		switch (src_id) {
+		case 1: /* D1 vblank/vline */
+			switch (src_data) {
+			case 0: /* D1 vblank */
+				if (rdev->irq.stat_regs.evergreen.disp_int & LB_D1_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[0]) {
+				//		drm_handle_vblank(rdev->ddev, 0);
+						rdev->pm.vblank_sync = true;
+				//		wake_up(&rdev->irq.vblank_queue);
+					}
+				//	if (rdev->irq.pflip[0])
+				//		radeon_crtc_handle_flip(rdev, 0);
+					rdev->irq.stat_regs.evergreen.disp_int &= ~LB_D1_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D1 vblank\n");
+				}
+				break;
+			case 1: /* D1 vline */
+				if (rdev->irq.stat_regs.evergreen.disp_int & LB_D1_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int &= ~LB_D1_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D1 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 2: /* D2 vblank/vline */
+			switch (src_data) {
+			case 0: /* D2 vblank */
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont & LB_D2_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[1]) {
+				//		drm_handle_vblank(rdev->ddev, 1);
+						rdev->pm.vblank_sync = true;
+				//		wake_up(&rdev->irq.vblank_queue);
+					}
+			//		if (rdev->irq.pflip[1])
+			//			radeon_crtc_handle_flip(rdev, 1);
+					rdev->irq.stat_regs.evergreen.disp_int_cont &= ~LB_D2_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D2 vblank\n");
+				}
+				break;
+			case 1: /* D2 vline */
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont & LB_D2_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int_cont &= ~LB_D2_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D2 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 3: /* D3 vblank/vline */
+			switch (src_data) {
+			case 0: /* D3 vblank */
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & LB_D3_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[2]) {
+				//		drm_handle_vblank(rdev->ddev, 2);
+						rdev->pm.vblank_sync = true;
+				//		wake_up(&rdev->irq.vblank_queue);
+					}
+				//	if (rdev->irq.pflip[2])
+				//		radeon_crtc_handle_flip(rdev, 2);
+					rdev->irq.stat_regs.evergreen.disp_int_cont2 &= ~LB_D3_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D3 vblank\n");
+				}
+				break;
+			case 1: /* D3 vline */
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & LB_D3_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int_cont2 &= ~LB_D3_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D3 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 4: /* D4 vblank/vline */
+			switch (src_data) {
+			case 0: /* D4 vblank */
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & LB_D4_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[3]) {
+					//	drm_handle_vblank(rdev->ddev, 3);
+						rdev->pm.vblank_sync = true;
+					//	wake_up(&rdev->irq.vblank_queue);
+					}
+		//			if (rdev->irq.pflip[3])
+		//				radeon_crtc_handle_flip(rdev, 3);
+					rdev->irq.stat_regs.evergreen.disp_int_cont3 &= ~LB_D4_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D4 vblank\n");
+				}
+				break;
+			case 1: /* D4 vline */
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & LB_D4_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int_cont3 &= ~LB_D4_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D4 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 5: /* D5 vblank/vline */
+			switch (src_data) {
+			case 0: /* D5 vblank */
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & LB_D5_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[4]) {
+//						drm_handle_vblank(rdev->ddev, 4);
+						rdev->pm.vblank_sync = true;
+//						wake_up(&rdev->irq.vblank_queue);
+					}
+//					if (rdev->irq.pflip[4])
+//						radeon_crtc_handle_flip(rdev, 4);
+					rdev->irq.stat_regs.evergreen.disp_int_cont4 &= ~LB_D5_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D5 vblank\n");
+				}
+				break;
+			case 1: /* D5 vline */
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & LB_D5_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int_cont4 &= ~LB_D5_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D5 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 6: /* D6 vblank/vline */
+			switch (src_data) {
+			case 0: /* D6 vblank */
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & LB_D6_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[5]) {
+				//		drm_handle_vblank(rdev->ddev, 5);
+						rdev->pm.vblank_sync = true;
+				//		wake_up(&rdev->irq.vblank_queue);
+					}
+			//		if (rdev->irq.pflip[5])
+			//			radeon_crtc_handle_flip(rdev, 5);
+					rdev->irq.stat_regs.evergreen.disp_int_cont5 &= ~LB_D6_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D6 vblank\n");
+				}
+				break;
+			case 1: /* D6 vline */
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & LB_D6_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int_cont5 &= ~LB_D6_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D6 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 42: /* HPD hotplug */
+			switch (src_data) {
+			case 0:
+				if (rdev->irq.stat_regs.evergreen.disp_int & DC_HPD1_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int &= ~DC_HPD1_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD1\n");
+				}
+				break;
+			case 1:
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont & DC_HPD2_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int_cont &= ~DC_HPD2_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD2\n");
+				}
+				break;
+			case 2:
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & DC_HPD3_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int_cont2 &= ~DC_HPD3_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD3\n");
+				}
+				break;
+			case 3:
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & DC_HPD4_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int_cont3 &= ~DC_HPD4_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD4\n");
+				}
+				break;
+			case 4:
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & DC_HPD5_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int_cont4 &= ~DC_HPD5_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD5\n");
+				}
+				break;
+			case 5:
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & DC_HPD6_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int_cont5 &= ~DC_HPD6_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD6\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 176: /* CP_INT in ring buffer */
+		case 177: /* CP_INT in IB1 */
+		case 178: /* CP_INT in IB2 */
+			DRM_DEBUG("IH: CP int: 0x%08x\n", src_data);
+			radeon_fence_process(rdev);
+			break;
+		case 181: /* CP EOP event */
+			DRM_DEBUG("IH: CP EOP\n");
+			radeon_fence_process(rdev);
+			break;
+		case 233: /* GUI IDLE */
+			DRM_DEBUG("IH: GUI idle\n");
+			rdev->pm.gui_idle = true;
+//			wake_up(&rdev->irq.idle_queue);
+			break;
+		default:
+			DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+			break;
+		}
+
+		/* wptr/rptr are in bytes! */
+		rptr += 16;
+		rptr &= rdev->ih.ptr_mask;
+	}
+	/* make sure wptr hasn't changed while processing */
+	wptr = evergreen_get_ih_wptr(rdev);
+	if (wptr != rdev->ih.wptr)
+		goto restart_ih;
+//	if (queue_hotplug)
+//		schedule_work(&rdev->hotplug_work);
+	rdev->ih.rptr = rptr;
+	WREG32(IH_RB_RPTR, rdev->ih.rptr);
+	spin_unlock_irqrestore(&rdev->ih.lock, flags);
+	return IRQ_HANDLED;
+}
+
 static int evergreen_startup(struct radeon_device *rdev)
 {
 	int r;
@@ -2483,7 +3027,7 @@ static int evergreen_startup(struct radeon_device *rdev)
 			return r;
 	}
 	evergreen_gpu_init(rdev);
-#if 0
+
 	r = evergreen_blit_init(rdev);
 	if (r) {
 		evergreen_blit_fini(rdev);
@@ -2495,9 +3039,15 @@ static int evergreen_startup(struct radeon_device *rdev)
 	r = radeon_wb_init(rdev);
 	if (r)
 		return r;
-#endif
 
 	/* Enable IRQ */
+	r = r600_irq_init(rdev);
+	if (r) {
+		DRM_ERROR("radeon: IH init failed (%d).\n", r);
+//		radeon_irq_kms_fini(rdev);
+		return r;
+	}
+	evergreen_irq_set(rdev);
 
     r = radeon_ring_init(rdev, rdev->cp.ring_size);
 	if (r)
@@ -2516,7 +3066,26 @@ static int evergreen_startup(struct radeon_device *rdev)
 
 
 
+int evergreen_copy_blit(struct radeon_device *rdev,
+			uint64_t src_offset, uint64_t dst_offset,
+			unsigned num_pages, struct radeon_fence *fence)
+{
+	int r;
 
+	mutex_lock(&rdev->r600_blit.mutex);
+	rdev->r600_blit.vb_ib = NULL;
+	r = evergreen_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;
+	}
+	evergreen_kms_blit_copy(rdev, src_offset, dst_offset, num_pages * RADEON_GPU_PAGE_SIZE);
+	evergreen_blit_done_copy(rdev, fence);
+	mutex_unlock(&rdev->r600_blit.mutex);
+	return 0;
+}
 
 /* Plan is to move initialization in that function and use
  * helper function so that radeon_device_init pretty much
@@ -2566,6 +3135,9 @@ int evergreen_init(struct radeon_device *rdev)
 	/* Initialize clocks */
 	radeon_get_clock_info(rdev->ddev);
 	/* Fence driver */
+	r = radeon_fence_driver_init(rdev);
+	if (r)
+		return r;
     /* initialize AGP */
 	if (rdev->flags & RADEON_IS_AGP) {
 		r = radeon_agp_init(rdev);
@@ -2581,12 +3153,15 @@ int evergreen_init(struct radeon_device *rdev)
 	if (r)
 		return r;
 
+	r = radeon_irq_kms_init(rdev);
+	if (r)
+		return r;
 
 	rdev->cp.ring_obj = NULL;
 	r600_ring_init(rdev, 1024 * 1024);
 
-//   rdev->ih.ring_obj = NULL;
+	rdev->ih.ring_obj = NULL;
-//   r600_ih_ring_init(rdev, 64 * 1024);
+	r600_ih_ring_init(rdev, 64 * 1024);
 
 	r = r600_pcie_gart_init(rdev);
 	if (r)
@@ -2599,6 +3174,16 @@ int evergreen_init(struct radeon_device *rdev)
 		rdev->accel_working = false;
 	}
 	if (rdev->accel_working) {
+		r = radeon_ib_pool_init(rdev);
+		if (r) {
+			DRM_ERROR("radeon: failed initializing IB pool (%d).\n", r);
+			rdev->accel_working = false;
+		}
+		r = r600_ib_test(rdev);
+		if (r) {
+			DRM_ERROR("radeon: failed testing IB (%d).\n", r);
+			rdev->accel_working = false;
+		}
 	}
 	return 0;
 }

drivers/video/drm/radeon/evergreen_blit_kms.c

@@ -0,0 +1,988 @@
+/*
+ * Copyright 2010 Advanced Micro Devices, 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 (including the next
+ * paragraph) 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) AND/OR ITS SUPPLIERS 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:
+ *     Alex Deucher <alexander.deucher@amd.com>
+ */
+
+#include "drmP.h"
+#include "drm.h"
+#include "radeon_drm.h"
+#include "radeon.h"
+
+#include "evergreend.h"
+#include "evergreen_blit_shaders.h"
+#include "cayman_blit_shaders.h"
+
+#define DI_PT_RECTLIST        0x11
+#define DI_INDEX_SIZE_16_BIT  0x0
+#define DI_SRC_SEL_AUTO_INDEX 0x2
+
+#define FMT_8                 0x1
+#define FMT_5_6_5             0x8
+#define FMT_8_8_8_8           0x1a
+#define COLOR_8               0x1
+#define COLOR_5_6_5           0x8
+#define COLOR_8_8_8_8         0x1a
+
+/* emits 17 */
+static void
+set_render_target(struct radeon_device *rdev, int format,
+		  int w, int h, u64 gpu_addr)
+{
+	u32 cb_color_info;
+	int pitch, slice;
+
+	h = ALIGN(h, 8);
+	if (h < 8)
+		h = 8;
+
+	cb_color_info = ((format << 2) | (1 << 24) | (1 << 8));
+	pitch = (w / 8) - 1;
+	slice = ((w * h) / 64) - 1;
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 15));
+	radeon_ring_write(rdev, (CB_COLOR0_BASE - PACKET3_SET_CONTEXT_REG_START) >> 2);
+	radeon_ring_write(rdev, gpu_addr >> 8);
+	radeon_ring_write(rdev, pitch);
+	radeon_ring_write(rdev, slice);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, cb_color_info);
+	radeon_ring_write(rdev, (1 << 4));
+	radeon_ring_write(rdev, (w - 1) | ((h - 1) << 16));
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+}
+
+/* emits 5dw */
+static void
+cp_set_surface_sync(struct radeon_device *rdev,
+		    u32 sync_type, u32 size,
+		    u64 mc_addr)
+{
+	u32 cp_coher_size;
+
+	if (size == 0xffffffff)
+		cp_coher_size = 0xffffffff;
+	else
+		cp_coher_size = ((size + 255) >> 8);
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SURFACE_SYNC, 3));
+	radeon_ring_write(rdev, sync_type);
+	radeon_ring_write(rdev, cp_coher_size);
+	radeon_ring_write(rdev, mc_addr >> 8);
+	radeon_ring_write(rdev, 10); /* poll interval */
+}
+
+/* emits 11dw + 1 surface sync = 16dw */
+static void
+set_shaders(struct radeon_device *rdev)
+{
+	u64 gpu_addr;
+
+	/* VS */
+	gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.vs_offset;
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 3));
+	radeon_ring_write(rdev, (SQ_PGM_START_VS - PACKET3_SET_CONTEXT_REG_START) >> 2);
+	radeon_ring_write(rdev, gpu_addr >> 8);
+	radeon_ring_write(rdev, 2);
+	radeon_ring_write(rdev, 0);
+
+	/* PS */
+	gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.ps_offset;
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 4));
+	radeon_ring_write(rdev, (SQ_PGM_START_PS - PACKET3_SET_CONTEXT_REG_START) >> 2);
+	radeon_ring_write(rdev, gpu_addr >> 8);
+	radeon_ring_write(rdev, 1);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 2);
+
+	gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.vs_offset;
+	cp_set_surface_sync(rdev, PACKET3_SH_ACTION_ENA, 512, gpu_addr);
+}
+
+/* emits 10 + 1 sync (5) = 15 */
+static void
+set_vtx_resource(struct radeon_device *rdev, u64 gpu_addr)
+{
+	u32 sq_vtx_constant_word2, sq_vtx_constant_word3;
+
+	/* high addr, stride */
+	sq_vtx_constant_word2 = ((upper_32_bits(gpu_addr) & 0xff) | (16 << 8));
+#ifdef __BIG_ENDIAN
+	sq_vtx_constant_word2 |= (2 << 30);
+#endif
+	/* xyzw swizzles */
+	sq_vtx_constant_word3 = (0 << 3) | (1 << 6) | (2 << 9) | (3 << 12);
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_RESOURCE, 8));
+	radeon_ring_write(rdev, 0x580);
+	radeon_ring_write(rdev, gpu_addr & 0xffffffff);
+	radeon_ring_write(rdev, 48 - 1); /* size */
+	radeon_ring_write(rdev, sq_vtx_constant_word2);
+	radeon_ring_write(rdev, sq_vtx_constant_word3);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, SQ_TEX_VTX_VALID_BUFFER << 30);
+
+	if ((rdev->family == CHIP_CEDAR) ||
+	    (rdev->family == CHIP_PALM) ||
+	    (rdev->family == CHIP_SUMO) ||
+	    (rdev->family == CHIP_SUMO2) ||
+	    (rdev->family == CHIP_CAICOS))
+		cp_set_surface_sync(rdev,
+				    PACKET3_TC_ACTION_ENA, 48, gpu_addr);
+	else
+		cp_set_surface_sync(rdev,
+				    PACKET3_VC_ACTION_ENA, 48, gpu_addr);
+
+}
+
+/* emits 10 */
+static void
+set_tex_resource(struct radeon_device *rdev,
+		 int format, int w, int h, int pitch,
+		 u64 gpu_addr)
+{
+	u32 sq_tex_resource_word0, sq_tex_resource_word1;
+	u32 sq_tex_resource_word4, sq_tex_resource_word7;
+
+	if (h < 1)
+		h = 1;
+
+	sq_tex_resource_word0 = (1 << 0); /* 2D */
+	sq_tex_resource_word0 |= ((((pitch >> 3) - 1) << 6) |
+				  ((w - 1) << 18));
+	sq_tex_resource_word1 = ((h - 1) << 0) | (1 << 28);
+	/* xyzw swizzles */
+	sq_tex_resource_word4 = (0 << 16) | (1 << 19) | (2 << 22) | (3 << 25);
+
+	sq_tex_resource_word7 = format | (SQ_TEX_VTX_VALID_TEXTURE << 30);
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_RESOURCE, 8));
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, sq_tex_resource_word0);
+	radeon_ring_write(rdev, sq_tex_resource_word1);
+	radeon_ring_write(rdev, gpu_addr >> 8);
+	radeon_ring_write(rdev, gpu_addr >> 8);
+	radeon_ring_write(rdev, sq_tex_resource_word4);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, sq_tex_resource_word7);
+}
+
+/* emits 12 */
+static void
+set_scissors(struct radeon_device *rdev, int x1, int y1,
+	     int x2, int y2)
+{
+	/* workaround some hw bugs */
+	if (x2 == 0)
+		x1 = 1;
+	if (y2 == 0)
+		y1 = 1;
+	if (rdev->family == CHIP_CAYMAN) {
+		if ((x2 == 1) && (y2 == 1))
+			x2 = 2;
+	}
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
+	radeon_ring_write(rdev, (PA_SC_SCREEN_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_START) >> 2);
+	radeon_ring_write(rdev, (x1 << 0) | (y1 << 16));
+	radeon_ring_write(rdev, (x2 << 0) | (y2 << 16));
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
+	radeon_ring_write(rdev, (PA_SC_GENERIC_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_START) >> 2);
+	radeon_ring_write(rdev, (x1 << 0) | (y1 << 16) | (1 << 31));
+	radeon_ring_write(rdev, (x2 << 0) | (y2 << 16));
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
+	radeon_ring_write(rdev, (PA_SC_WINDOW_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_START) >> 2);
+	radeon_ring_write(rdev, (x1 << 0) | (y1 << 16) | (1 << 31));
+	radeon_ring_write(rdev, (x2 << 0) | (y2 << 16));
+}
+
+/* emits 10 */
+static void
+draw_auto(struct radeon_device *rdev)
+{
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1));
+	radeon_ring_write(rdev, (VGT_PRIMITIVE_TYPE - PACKET3_SET_CONFIG_REG_START) >> 2);
+	radeon_ring_write(rdev, DI_PT_RECTLIST);
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_INDEX_TYPE, 0));
+	radeon_ring_write(rdev,
+#ifdef __BIG_ENDIAN
+			  (2 << 2) |
+#endif
+			  DI_INDEX_SIZE_16_BIT);
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_NUM_INSTANCES, 0));
+	radeon_ring_write(rdev, 1);
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_DRAW_INDEX_AUTO, 1));
+	radeon_ring_write(rdev, 3);
+	radeon_ring_write(rdev, DI_SRC_SEL_AUTO_INDEX);
+
+}
+
+/* emits 39 */
+static void
+set_default_state(struct radeon_device *rdev)
+{
+	u32 sq_config, sq_gpr_resource_mgmt_1, sq_gpr_resource_mgmt_2, sq_gpr_resource_mgmt_3;
+	u32 sq_thread_resource_mgmt, sq_thread_resource_mgmt_2;
+	u32 sq_stack_resource_mgmt_1, sq_stack_resource_mgmt_2, sq_stack_resource_mgmt_3;
+	int num_ps_gprs, num_vs_gprs, num_temp_gprs;
+	int num_gs_gprs, num_es_gprs, num_hs_gprs, num_ls_gprs;
+	int num_ps_threads, num_vs_threads, num_gs_threads, num_es_threads;
+	int num_hs_threads, num_ls_threads;
+	int num_ps_stack_entries, num_vs_stack_entries, num_gs_stack_entries, num_es_stack_entries;
+	int num_hs_stack_entries, num_ls_stack_entries;
+	u64 gpu_addr;
+	int dwords;
+
+	/* set clear context state */
+	radeon_ring_write(rdev, PACKET3(PACKET3_CLEAR_STATE, 0));
+	radeon_ring_write(rdev, 0);
+
+	if (rdev->family < CHIP_CAYMAN) {
+		switch (rdev->family) {
+		case CHIP_CEDAR:
+		default:
+			num_ps_gprs = 93;
+			num_vs_gprs = 46;
+			num_temp_gprs = 4;
+			num_gs_gprs = 31;
+			num_es_gprs = 31;
+			num_hs_gprs = 23;
+			num_ls_gprs = 23;
+			num_ps_threads = 96;
+			num_vs_threads = 16;
+			num_gs_threads = 16;
+			num_es_threads = 16;
+			num_hs_threads = 16;
+			num_ls_threads = 16;
+			num_ps_stack_entries = 42;
+			num_vs_stack_entries = 42;
+			num_gs_stack_entries = 42;
+			num_es_stack_entries = 42;
+			num_hs_stack_entries = 42;
+			num_ls_stack_entries = 42;
+			break;
+		case CHIP_REDWOOD:
+			num_ps_gprs = 93;
+			num_vs_gprs = 46;
+			num_temp_gprs = 4;
+			num_gs_gprs = 31;
+			num_es_gprs = 31;
+			num_hs_gprs = 23;
+			num_ls_gprs = 23;
+			num_ps_threads = 128;
+			num_vs_threads = 20;
+			num_gs_threads = 20;
+			num_es_threads = 20;
+			num_hs_threads = 20;
+			num_ls_threads = 20;
+			num_ps_stack_entries = 42;
+			num_vs_stack_entries = 42;
+			num_gs_stack_entries = 42;
+			num_es_stack_entries = 42;
+			num_hs_stack_entries = 42;
+			num_ls_stack_entries = 42;
+			break;
+		case CHIP_JUNIPER:
+			num_ps_gprs = 93;
+			num_vs_gprs = 46;
+			num_temp_gprs = 4;
+			num_gs_gprs = 31;
+			num_es_gprs = 31;
+			num_hs_gprs = 23;
+			num_ls_gprs = 23;
+			num_ps_threads = 128;
+			num_vs_threads = 20;
+			num_gs_threads = 20;
+			num_es_threads = 20;
+			num_hs_threads = 20;
+			num_ls_threads = 20;
+			num_ps_stack_entries = 85;
+			num_vs_stack_entries = 85;
+			num_gs_stack_entries = 85;
+			num_es_stack_entries = 85;
+			num_hs_stack_entries = 85;
+			num_ls_stack_entries = 85;
+			break;
+		case CHIP_CYPRESS:
+		case CHIP_HEMLOCK:
+			num_ps_gprs = 93;
+			num_vs_gprs = 46;
+			num_temp_gprs = 4;
+			num_gs_gprs = 31;
+			num_es_gprs = 31;
+			num_hs_gprs = 23;
+			num_ls_gprs = 23;
+			num_ps_threads = 128;
+			num_vs_threads = 20;
+			num_gs_threads = 20;
+			num_es_threads = 20;
+			num_hs_threads = 20;
+			num_ls_threads = 20;
+			num_ps_stack_entries = 85;
+			num_vs_stack_entries = 85;
+			num_gs_stack_entries = 85;
+			num_es_stack_entries = 85;
+			num_hs_stack_entries = 85;
+			num_ls_stack_entries = 85;
+			break;
+		case CHIP_PALM:
+			num_ps_gprs = 93;
+			num_vs_gprs = 46;
+			num_temp_gprs = 4;
+			num_gs_gprs = 31;
+			num_es_gprs = 31;
+			num_hs_gprs = 23;
+			num_ls_gprs = 23;
+			num_ps_threads = 96;
+			num_vs_threads = 16;
+			num_gs_threads = 16;
+			num_es_threads = 16;
+			num_hs_threads = 16;
+			num_ls_threads = 16;
+			num_ps_stack_entries = 42;
+			num_vs_stack_entries = 42;
+			num_gs_stack_entries = 42;
+			num_es_stack_entries = 42;
+			num_hs_stack_entries = 42;
+			num_ls_stack_entries = 42;
+			break;
+		case CHIP_SUMO:
+			num_ps_gprs = 93;
+			num_vs_gprs = 46;
+			num_temp_gprs = 4;
+			num_gs_gprs = 31;
+			num_es_gprs = 31;
+			num_hs_gprs = 23;
+			num_ls_gprs = 23;
+			num_ps_threads = 96;
+			num_vs_threads = 25;
+			num_gs_threads = 25;
+			num_es_threads = 25;
+			num_hs_threads = 25;
+			num_ls_threads = 25;
+			num_ps_stack_entries = 42;
+			num_vs_stack_entries = 42;
+			num_gs_stack_entries = 42;
+			num_es_stack_entries = 42;
+			num_hs_stack_entries = 42;
+			num_ls_stack_entries = 42;
+			break;
+		case CHIP_SUMO2:
+			num_ps_gprs = 93;
+			num_vs_gprs = 46;
+			num_temp_gprs = 4;
+			num_gs_gprs = 31;
+			num_es_gprs = 31;
+			num_hs_gprs = 23;
+			num_ls_gprs = 23;
+			num_ps_threads = 96;
+			num_vs_threads = 25;
+			num_gs_threads = 25;
+			num_es_threads = 25;
+			num_hs_threads = 25;
+			num_ls_threads = 25;
+			num_ps_stack_entries = 85;
+			num_vs_stack_entries = 85;
+			num_gs_stack_entries = 85;
+			num_es_stack_entries = 85;
+			num_hs_stack_entries = 85;
+			num_ls_stack_entries = 85;
+			break;
+		case CHIP_BARTS:
+			num_ps_gprs = 93;
+			num_vs_gprs = 46;
+			num_temp_gprs = 4;
+			num_gs_gprs = 31;
+			num_es_gprs = 31;
+			num_hs_gprs = 23;
+			num_ls_gprs = 23;
+			num_ps_threads = 128;
+			num_vs_threads = 20;
+			num_gs_threads = 20;
+			num_es_threads = 20;
+			num_hs_threads = 20;
+			num_ls_threads = 20;
+			num_ps_stack_entries = 85;
+			num_vs_stack_entries = 85;
+			num_gs_stack_entries = 85;
+			num_es_stack_entries = 85;
+			num_hs_stack_entries = 85;
+			num_ls_stack_entries = 85;
+			break;
+		case CHIP_TURKS:
+			num_ps_gprs = 93;
+			num_vs_gprs = 46;
+			num_temp_gprs = 4;
+			num_gs_gprs = 31;
+			num_es_gprs = 31;
+			num_hs_gprs = 23;
+			num_ls_gprs = 23;
+			num_ps_threads = 128;
+			num_vs_threads = 20;
+			num_gs_threads = 20;
+			num_es_threads = 20;
+			num_hs_threads = 20;
+			num_ls_threads = 20;
+			num_ps_stack_entries = 42;
+			num_vs_stack_entries = 42;
+			num_gs_stack_entries = 42;
+			num_es_stack_entries = 42;
+			num_hs_stack_entries = 42;
+			num_ls_stack_entries = 42;
+			break;
+		case CHIP_CAICOS:
+			num_ps_gprs = 93;
+			num_vs_gprs = 46;
+			num_temp_gprs = 4;
+			num_gs_gprs = 31;
+			num_es_gprs = 31;
+			num_hs_gprs = 23;
+			num_ls_gprs = 23;
+			num_ps_threads = 128;
+			num_vs_threads = 10;
+			num_gs_threads = 10;
+			num_es_threads = 10;
+			num_hs_threads = 10;
+			num_ls_threads = 10;
+			num_ps_stack_entries = 42;
+			num_vs_stack_entries = 42;
+			num_gs_stack_entries = 42;
+			num_es_stack_entries = 42;
+			num_hs_stack_entries = 42;
+			num_ls_stack_entries = 42;
+			break;
+		}
+
+		if ((rdev->family == CHIP_CEDAR) ||
+		    (rdev->family == CHIP_PALM) ||
+		    (rdev->family == CHIP_SUMO) ||
+		    (rdev->family == CHIP_SUMO2) ||
+		    (rdev->family == CHIP_CAICOS))
+			sq_config = 0;
+		else
+			sq_config = VC_ENABLE;
+
+		sq_config |= (EXPORT_SRC_C |
+			      CS_PRIO(0) |
+			      LS_PRIO(0) |
+			      HS_PRIO(0) |
+			      PS_PRIO(0) |
+			      VS_PRIO(1) |
+			      GS_PRIO(2) |
+			      ES_PRIO(3));
+
+		sq_gpr_resource_mgmt_1 = (NUM_PS_GPRS(num_ps_gprs) |
+					  NUM_VS_GPRS(num_vs_gprs) |
+					  NUM_CLAUSE_TEMP_GPRS(num_temp_gprs));
+		sq_gpr_resource_mgmt_2 = (NUM_GS_GPRS(num_gs_gprs) |
+					  NUM_ES_GPRS(num_es_gprs));
+		sq_gpr_resource_mgmt_3 = (NUM_HS_GPRS(num_hs_gprs) |
+					  NUM_LS_GPRS(num_ls_gprs));
+		sq_thread_resource_mgmt = (NUM_PS_THREADS(num_ps_threads) |
+					   NUM_VS_THREADS(num_vs_threads) |
+					   NUM_GS_THREADS(num_gs_threads) |
+					   NUM_ES_THREADS(num_es_threads));
+		sq_thread_resource_mgmt_2 = (NUM_HS_THREADS(num_hs_threads) |
+					     NUM_LS_THREADS(num_ls_threads));
+		sq_stack_resource_mgmt_1 = (NUM_PS_STACK_ENTRIES(num_ps_stack_entries) |
+					    NUM_VS_STACK_ENTRIES(num_vs_stack_entries));
+		sq_stack_resource_mgmt_2 = (NUM_GS_STACK_ENTRIES(num_gs_stack_entries) |
+					    NUM_ES_STACK_ENTRIES(num_es_stack_entries));
+		sq_stack_resource_mgmt_3 = (NUM_HS_STACK_ENTRIES(num_hs_stack_entries) |
+					    NUM_LS_STACK_ENTRIES(num_ls_stack_entries));
+
+		/* disable dyn gprs */
+		radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1));
+		radeon_ring_write(rdev, (SQ_DYN_GPR_CNTL_PS_FLUSH_REQ - PACKET3_SET_CONFIG_REG_START) >> 2);
+		radeon_ring_write(rdev, 0);
+
+		/* setup LDS */
+		radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1));
+		radeon_ring_write(rdev, (SQ_LDS_RESOURCE_MGMT - PACKET3_SET_CONFIG_REG_START) >> 2);
+		radeon_ring_write(rdev, 0x10001000);
+
+		/* SQ config */
+		radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 11));
+		radeon_ring_write(rdev, (SQ_CONFIG - PACKET3_SET_CONFIG_REG_START) >> 2);
+		radeon_ring_write(rdev, sq_config);
+		radeon_ring_write(rdev, sq_gpr_resource_mgmt_1);
+		radeon_ring_write(rdev, sq_gpr_resource_mgmt_2);
+		radeon_ring_write(rdev, sq_gpr_resource_mgmt_3);
+		radeon_ring_write(rdev, 0);
+		radeon_ring_write(rdev, 0);
+		radeon_ring_write(rdev, sq_thread_resource_mgmt);
+		radeon_ring_write(rdev, sq_thread_resource_mgmt_2);
+		radeon_ring_write(rdev, sq_stack_resource_mgmt_1);
+		radeon_ring_write(rdev, sq_stack_resource_mgmt_2);
+		radeon_ring_write(rdev, sq_stack_resource_mgmt_3);
+	}
+
+	/* CONTEXT_CONTROL */
+	radeon_ring_write(rdev, 0xc0012800);
+	radeon_ring_write(rdev, 0x80000000);
+	radeon_ring_write(rdev, 0x80000000);
+
+	/* SQ_VTX_BASE_VTX_LOC */
+	radeon_ring_write(rdev, 0xc0026f00);
+	radeon_ring_write(rdev, 0x00000000);
+	radeon_ring_write(rdev, 0x00000000);
+	radeon_ring_write(rdev, 0x00000000);
+
+	/* SET_SAMPLER */
+	radeon_ring_write(rdev, 0xc0036e00);
+	radeon_ring_write(rdev, 0x00000000);
+	radeon_ring_write(rdev, 0x00000012);
+	radeon_ring_write(rdev, 0x00000000);
+	radeon_ring_write(rdev, 0x00000000);
+
+	/* set to DX10/11 mode */
+	radeon_ring_write(rdev, PACKET3(PACKET3_MODE_CONTROL, 0));
+	radeon_ring_write(rdev, 1);
+
+	/* emit an IB pointing at default state */
+	dwords = ALIGN(rdev->r600_blit.state_len, 0x10);
+	gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.state_offset;
+	radeon_ring_write(rdev, PACKET3(PACKET3_INDIRECT_BUFFER, 2));
+	radeon_ring_write(rdev, gpu_addr & 0xFFFFFFFC);
+	radeon_ring_write(rdev, upper_32_bits(gpu_addr) & 0xFF);
+	radeon_ring_write(rdev, dwords);
+
+}
+
+static inline uint32_t i2f(uint32_t input)
+{
+	u32 result, i, exponent, fraction;
+
+	if ((input & 0x3fff) == 0)
+		result = 0; /* 0 is a special case */
+	else {
+		exponent = 140; /* exponent biased by 127; */
+		fraction = (input & 0x3fff) << 10; /* cheat and only
+						      handle numbers below 2^^15 */
+		for (i = 0; i < 14; i++) {
+			if (fraction & 0x800000)
+				break;
+			else {
+				fraction = fraction << 1; /* keep
+							     shifting left until top bit = 1 */
+				exponent = exponent - 1;
+			}
+		}
+		result = exponent << 23 | (fraction & 0x7fffff); /* mask
+								    off top bit; assumed 1 */
+	}
+	return result;
+}
+
+int evergreen_blit_init(struct radeon_device *rdev)
+{
+	u32 obj_size;
+	int i, r, dwords;
+	void *ptr;
+	u32 packet2s[16];
+	int num_packet2s = 0;
+
+	/* pin copy shader into vram if already initialized */
+	if (rdev->r600_blit.shader_obj)
+		goto done;
+
+	mutex_init(&rdev->r600_blit.mutex);
+	rdev->r600_blit.state_offset = 0;
+
+	if (rdev->family < CHIP_CAYMAN)
+		rdev->r600_blit.state_len = evergreen_default_size;
+	else
+		rdev->r600_blit.state_len = cayman_default_size;
+
+	dwords = rdev->r600_blit.state_len;
+	while (dwords & 0xf) {
+		packet2s[num_packet2s++] = cpu_to_le32(PACKET2(0));
+		dwords++;
+	}
+
+	obj_size = dwords * 4;
+	obj_size = ALIGN(obj_size, 256);
+
+	rdev->r600_blit.vs_offset = obj_size;
+	if (rdev->family < CHIP_CAYMAN)
+		obj_size += evergreen_vs_size * 4;
+	else
+		obj_size += cayman_vs_size * 4;
+	obj_size = ALIGN(obj_size, 256);
+
+	rdev->r600_blit.ps_offset = obj_size;
+	if (rdev->family < CHIP_CAYMAN)
+		obj_size += evergreen_ps_size * 4;
+	else
+		obj_size += cayman_ps_size * 4;
+	obj_size = ALIGN(obj_size, 256);
+
+	r = radeon_bo_create(rdev, obj_size, PAGE_SIZE, true, RADEON_GEM_DOMAIN_VRAM,
+				&rdev->r600_blit.shader_obj);
+	if (r) {
+		DRM_ERROR("evergreen failed to allocate shader\n");
+		return r;
+	}
+
+	DRM_DEBUG("evergreen blit allocated bo %08x vs %08x ps %08x\n",
+		  obj_size,
+		  rdev->r600_blit.vs_offset, rdev->r600_blit.ps_offset);
+
+	r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false);
+	if (unlikely(r != 0))
+		return r;
+	r = radeon_bo_kmap(rdev->r600_blit.shader_obj, &ptr);
+	if (r) {
+		DRM_ERROR("failed to map blit object %d\n", r);
+		return r;
+	}
+
+	if (rdev->family < CHIP_CAYMAN) {
+        memcpy(ptr + rdev->r600_blit.state_offset,
+			    evergreen_default_state, rdev->r600_blit.state_len * 4);
+
+		if (num_packet2s)
+            memcpy(ptr + rdev->r600_blit.state_offset + (rdev->r600_blit.state_len * 4),
+				    packet2s, num_packet2s * 4);
+		for (i = 0; i < evergreen_vs_size; i++)
+			*(u32 *)((unsigned long)ptr + rdev->r600_blit.vs_offset + i * 4) = cpu_to_le32(evergreen_vs[i]);
+		for (i = 0; i < evergreen_ps_size; i++)
+			*(u32 *)((unsigned long)ptr + rdev->r600_blit.ps_offset + i * 4) = cpu_to_le32(evergreen_ps[i]);
+	} else {
+        memcpy(ptr + rdev->r600_blit.state_offset,
+			    cayman_default_state, rdev->r600_blit.state_len * 4);
+
+		if (num_packet2s)
+            memcpy(ptr + rdev->r600_blit.state_offset + (rdev->r600_blit.state_len * 4),
+				    packet2s, num_packet2s * 4);
+		for (i = 0; i < cayman_vs_size; i++)
+			*(u32 *)((unsigned long)ptr + rdev->r600_blit.vs_offset + i * 4) = cpu_to_le32(cayman_vs[i]);
+		for (i = 0; i < cayman_ps_size; i++)
+			*(u32 *)((unsigned long)ptr + rdev->r600_blit.ps_offset + i * 4) = cpu_to_le32(cayman_ps[i]);
+	}
+	radeon_bo_kunmap(rdev->r600_blit.shader_obj);
+	radeon_bo_unreserve(rdev->r600_blit.shader_obj);
+
+done:
+	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;
+	}
+//   radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size);
+	return 0;
+}
+
+void evergreen_blit_fini(struct radeon_device *rdev)
+{
+	int r;
+
+//   radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size);
+	if (rdev->r600_blit.shader_obj == NULL)
+		return;
+	/* If we can't reserve the bo, unref should be enough to destroy
+	 * it when it becomes idle.
+	 */
+	r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false);
+	if (!r) {
+		radeon_bo_unpin(rdev->r600_blit.shader_obj);
+		radeon_bo_unreserve(rdev->r600_blit.shader_obj);
+	}
+	radeon_bo_unref(&rdev->r600_blit.shader_obj);
+}
+
+static int evergreen_vb_ib_get(struct radeon_device *rdev)
+{
+	int r;
+	r = radeon_ib_get(rdev, &rdev->r600_blit.vb_ib);
+	if (r) {
+		DRM_ERROR("failed to get IB for vertex buffer\n");
+		return r;
+	}
+
+	rdev->r600_blit.vb_total = 64*1024;
+	rdev->r600_blit.vb_used = 0;
+	return 0;
+}
+
+static void evergreen_vb_ib_put(struct radeon_device *rdev)
+{
+	radeon_fence_emit(rdev, rdev->r600_blit.vb_ib->fence);
+	radeon_ib_free(rdev, &rdev->r600_blit.vb_ib);
+}
+
+int evergreen_blit_prepare_copy(struct radeon_device *rdev, int size_bytes)
+{
+	int r;
+	int ring_size, line_size;
+	int max_size;
+	/* loops of emits + fence emit possible */
+	int dwords_per_loop = 74, num_loops;
+
+	r = evergreen_vb_ib_get(rdev);
+	if (r)
+		return r;
+
+	/* 8 bpp vs 32 bpp for xfer unit */
+	if (size_bytes & 3)
+		line_size = 8192;
+	else
+		line_size = 8192 * 4;
+
+	max_size = 8192 * line_size;
+
+	/* major loops cover the max size transfer */
+	num_loops = ((size_bytes + max_size) / max_size);
+	/* minor loops cover the extra non aligned bits */
+	num_loops += ((size_bytes % line_size) ? 1 : 0);
+	/* calculate number of loops correctly */
+	ring_size = num_loops * dwords_per_loop;
+	/* set default  + shaders */
+	ring_size += 55; /* shaders + def state */
+	ring_size += 10; /* fence emit for VB IB */
+	ring_size += 5; /* done copy */
+	ring_size += 10; /* fence emit for done copy */
+	r = radeon_ring_lock(rdev, ring_size);
+	if (r)
+		return r;
+
+	set_default_state(rdev); /* 36 */
+	set_shaders(rdev); /* 16 */
+	return 0;
+}
+
+void evergreen_blit_done_copy(struct radeon_device *rdev, struct radeon_fence *fence)
+{
+	int r;
+
+	if (rdev->r600_blit.vb_ib)
+		evergreen_vb_ib_put(rdev);
+
+	if (fence)
+		r = radeon_fence_emit(rdev, fence);
+
+	radeon_ring_unlock_commit(rdev);
+}
+
+void evergreen_kms_blit_copy(struct radeon_device *rdev,
+			     u64 src_gpu_addr, u64 dst_gpu_addr,
+			     int size_bytes)
+{
+	int max_bytes;
+	u64 vb_gpu_addr;
+	u32 *vb;
+
+	DRM_DEBUG("emitting copy %16llx %16llx %d %d\n", src_gpu_addr, dst_gpu_addr,
+		  size_bytes, rdev->r600_blit.vb_used);
+	vb = (u32 *)(rdev->r600_blit.vb_ib->ptr + rdev->r600_blit.vb_used);
+	if ((size_bytes & 3) || (src_gpu_addr & 3) || (dst_gpu_addr & 3)) {
+		max_bytes = 8192;
+
+		while (size_bytes) {
+			int cur_size = size_bytes;
+			int src_x = src_gpu_addr & 255;
+			int dst_x = dst_gpu_addr & 255;
+			int h = 1;
+			src_gpu_addr = src_gpu_addr & ~255ULL;
+			dst_gpu_addr = dst_gpu_addr & ~255ULL;
+
+			if (!src_x && !dst_x) {
+				h = (cur_size / max_bytes);
+				if (h > 8192)
+					h = 8192;
+				if (h == 0)
+					h = 1;
+				else
+					cur_size = max_bytes;
+			} else {
+				if (cur_size > max_bytes)
+					cur_size = max_bytes;
+				if (cur_size > (max_bytes - dst_x))
+					cur_size = (max_bytes - dst_x);
+				if (cur_size > (max_bytes - src_x))
+					cur_size = (max_bytes - src_x);
+			}
+
+			if ((rdev->r600_blit.vb_used + 48) > rdev->r600_blit.vb_total) {
+            //   WARN_ON(1);
+			}
+
+			vb[0] = i2f(dst_x);
+			vb[1] = 0;
+			vb[2] = i2f(src_x);
+			vb[3] = 0;
+
+			vb[4] = i2f(dst_x);
+			vb[5] = i2f(h);
+			vb[6] = i2f(src_x);
+			vb[7] = i2f(h);
+
+			vb[8] = i2f(dst_x + cur_size);
+			vb[9] = i2f(h);
+			vb[10] = i2f(src_x + cur_size);
+			vb[11] = i2f(h);
+
+			/* src 10 */
+			set_tex_resource(rdev, FMT_8,
+					 src_x + cur_size, h, src_x + cur_size,
+					 src_gpu_addr);
+
+			/* 5 */
+			cp_set_surface_sync(rdev,
+					    PACKET3_TC_ACTION_ENA, (src_x + cur_size * h), src_gpu_addr);
+
+
+			/* dst 17 */
+			set_render_target(rdev, COLOR_8,
+					  dst_x + cur_size, h,
+					  dst_gpu_addr);
+
+			/* scissors 12 */
+			set_scissors(rdev, dst_x, 0, dst_x + cur_size, h);
+
+			/* 15 */
+			vb_gpu_addr = rdev->r600_blit.vb_ib->gpu_addr + rdev->r600_blit.vb_used;
+			set_vtx_resource(rdev, vb_gpu_addr);
+
+			/* draw 10 */
+			draw_auto(rdev);
+
+			/* 5 */
+			cp_set_surface_sync(rdev,
+					    PACKET3_CB_ACTION_ENA | PACKET3_CB0_DEST_BASE_ENA,
+					    cur_size * h, dst_gpu_addr);
+
+			vb += 12;
+			rdev->r600_blit.vb_used += 12 * 4;
+
+			src_gpu_addr += cur_size * h;
+			dst_gpu_addr += cur_size * h;
+			size_bytes -= cur_size * h;
+		}
+	} else {
+		max_bytes = 8192 * 4;
+
+		while (size_bytes) {
+			int cur_size = size_bytes;
+			int src_x = (src_gpu_addr & 255);
+			int dst_x = (dst_gpu_addr & 255);
+			int h = 1;
+			src_gpu_addr = src_gpu_addr & ~255ULL;
+			dst_gpu_addr = dst_gpu_addr & ~255ULL;
+
+			if (!src_x && !dst_x) {
+				h = (cur_size / max_bytes);
+				if (h > 8192)
+					h = 8192;
+				if (h == 0)
+					h = 1;
+				else
+					cur_size = max_bytes;
+			} else {
+				if (cur_size > max_bytes)
+					cur_size = max_bytes;
+				if (cur_size > (max_bytes - dst_x))
+					cur_size = (max_bytes - dst_x);
+				if (cur_size > (max_bytes - src_x))
+					cur_size = (max_bytes - src_x);
+			}
+
+			if ((rdev->r600_blit.vb_used + 48) > rdev->r600_blit.vb_total) {
+           //        WARN_ON(1);
+			}
+
+			vb[0] = i2f(dst_x / 4);
+			vb[1] = 0;
+			vb[2] = i2f(src_x / 4);
+			vb[3] = 0;
+
+			vb[4] = i2f(dst_x / 4);
+			vb[5] = i2f(h);
+			vb[6] = i2f(src_x / 4);
+			vb[7] = i2f(h);
+
+			vb[8] = i2f((dst_x + cur_size) / 4);
+			vb[9] = i2f(h);
+			vb[10] = i2f((src_x + cur_size) / 4);
+			vb[11] = i2f(h);
+
+			/* src 10 */
+			set_tex_resource(rdev, FMT_8_8_8_8,
+					 (src_x + cur_size) / 4,
+					 h, (src_x + cur_size) / 4,
+					 src_gpu_addr);
+			/* 5 */
+			cp_set_surface_sync(rdev,
+					    PACKET3_TC_ACTION_ENA, (src_x + cur_size * h), src_gpu_addr);
+
+			/* dst 17 */
+			set_render_target(rdev, COLOR_8_8_8_8,
+					  (dst_x + cur_size) / 4, h,
+					  dst_gpu_addr);
+
+			/* scissors 12  */
+			set_scissors(rdev, (dst_x / 4), 0, (dst_x + cur_size / 4), h);
+
+			/* Vertex buffer setup 15 */
+			vb_gpu_addr = rdev->r600_blit.vb_ib->gpu_addr + rdev->r600_blit.vb_used;
+			set_vtx_resource(rdev, vb_gpu_addr);
+
+			/* draw 10 */
+			draw_auto(rdev);
+
+			/* 5 */
+			cp_set_surface_sync(rdev,
+					    PACKET3_CB_ACTION_ENA | PACKET3_CB0_DEST_BASE_ENA,
+					    cur_size * h, dst_gpu_addr);
+
+			/* 74 ring dwords per loop */
+			vb += 12;
+			rdev->r600_blit.vb_used += 12 * 4;
+
+			src_gpu_addr += cur_size * h;
+			dst_gpu_addr += cur_size * h;
+			size_bytes -= cur_size * h;
+		}
+	}
+}
+

drivers/video/drm/radeon/ni.c

@@ -1376,10 +1376,26 @@ static int cayman_startup(struct radeon_device *rdev)
 		return r;
 	cayman_gpu_init(rdev);
 
+	r = evergreen_blit_init(rdev);
+	if (r) {
+//		evergreen_blit_fini(rdev);
+		rdev->asic->copy = NULL;
+		dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r);
+	}
 
 	/* allocate wb buffer */
+	r = radeon_wb_init(rdev);
+	if (r)
+		return r;
 
 	/* Enable IRQ */
+	r = r600_irq_init(rdev);
+	if (r) {
+		DRM_ERROR("radeon: IH init failed (%d).\n", r);
+//		radeon_irq_kms_fini(rdev);
+		return r;
+	}
+	evergreen_irq_set(rdev);
 
 	r = radeon_ring_init(rdev, rdev->cp.ring_size);
 	if (r)
@@ -1442,6 +1458,9 @@ int cayman_init(struct radeon_device *rdev)
 	/* Initialize clocks */
 	radeon_get_clock_info(rdev->ddev);
 	/* Fence driver */
+	r = radeon_fence_driver_init(rdev);
+	if (r)
+		return r;
 	/* initialize memory controller */
 	r = evergreen_mc_init(rdev);
 	if (r)
@@ -1451,10 +1470,15 @@ int cayman_init(struct radeon_device *rdev)
 	if (r)
 		return r;
 
+	r = radeon_irq_kms_init(rdev);
+	if (r)
+		return r;
 
 	rdev->cp.ring_obj = NULL;
 	r600_ring_init(rdev, 1024 * 1024);
 
+	rdev->ih.ring_obj = NULL;
+	r600_ih_ring_init(rdev, 64 * 1024);
 
 	r = r600_pcie_gart_init(rdev);
 	if (r)
@@ -1467,6 +1491,16 @@ int cayman_init(struct radeon_device *rdev)
 		rdev->accel_working = false;
 	}
 	if (rdev->accel_working) {
+		r = radeon_ib_pool_init(rdev);
+		if (r) {
+			DRM_ERROR("radeon: failed initializing IB pool (%d).\n", r);
+			rdev->accel_working = false;
+		}
+		r = r600_ib_test(rdev);
+		if (r) {
+			DRM_ERROR("radeon: failed testing IB (%d).\n", r);
+			rdev->accel_working = false;
+		}
 	}
 
 	/* Don't start up if the MC ucode is missing.

drivers/video/drm/radeon/r100.c

@@ -151,17 +151,17 @@ void r100_hpd_init(struct radeon_device *rdev)
 		struct radeon_connector *radeon_connector = to_radeon_connector(connector);
 		switch (radeon_connector->hpd.hpd) {
 		case RADEON_HPD_1:
-//           rdev->irq.hpd[0] = true;
+			rdev->irq.hpd[0] = true;
 			break;
 		case RADEON_HPD_2:
-//           rdev->irq.hpd[1] = true;
+			rdev->irq.hpd[1] = true;
 			break;
 		default:
 			break;
 		}
 	}
-//   if (rdev->irq.installed)
+	if (rdev->irq.installed)
-//   r100_irq_set(rdev);
+		r100_irq_set(rdev);
 }
 
 void r100_hpd_fini(struct radeon_device *rdev)
@@ -173,10 +173,10 @@ void r100_hpd_fini(struct radeon_device *rdev)
 		struct radeon_connector *radeon_connector = to_radeon_connector(connector);
 		switch (radeon_connector->hpd.hpd) {
 		case RADEON_HPD_1:
-//           rdev->irq.hpd[0] = false;
+			rdev->irq.hpd[0] = false;
 			break;
 		case RADEON_HPD_2:
-//           rdev->irq.hpd[1] = false;
+			rdev->irq.hpd[1] = false;
 			break;
 		default:
 			break;
@@ -269,6 +269,38 @@ void r100_pci_gart_fini(struct radeon_device *rdev)
 	radeon_gart_table_ram_free(rdev);
 }
 
+int r100_irq_set(struct radeon_device *rdev)
+{
+	uint32_t tmp = 0;
+
+	if (!rdev->irq.installed) {
+		WARN(1, "Can't enable IRQ/MSI because no handler is installed\n");
+		WREG32(R_000040_GEN_INT_CNTL, 0);
+		return -EINVAL;
+	}
+	if (rdev->irq.sw_int) {
+		tmp |= RADEON_SW_INT_ENABLE;
+	}
+	if (rdev->irq.gui_idle) {
+		tmp |= RADEON_GUI_IDLE_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[0] ||
+	    rdev->irq.pflip[0]) {
+		tmp |= RADEON_CRTC_VBLANK_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[1] ||
+	    rdev->irq.pflip[1]) {
+		tmp |= RADEON_CRTC2_VBLANK_MASK;
+	}
+	if (rdev->irq.hpd[0]) {
+		tmp |= RADEON_FP_DETECT_MASK;
+	}
+	if (rdev->irq.hpd[1]) {
+		tmp |= RADEON_FP2_DETECT_MASK;
+	}
+	WREG32(RADEON_GEN_INT_CNTL, tmp);
+	return 0;
+}
 
 void r100_irq_disable(struct radeon_device *rdev)
 {
@@ -281,7 +313,6 @@ void r100_irq_disable(struct radeon_device *rdev)
 	WREG32(R_000044_GEN_INT_STATUS, tmp);
 }
 
-#if 0
 static inline uint32_t r100_irq_ack(struct radeon_device *rdev)
 {
 	uint32_t irqs = RREG32(RADEON_GEN_INT_STATUS);
@@ -301,8 +332,82 @@ static inline uint32_t r100_irq_ack(struct radeon_device *rdev)
 	return irqs & irq_mask;
 }
 
-#endif
+int r100_irq_process(struct radeon_device *rdev)
+{
+	uint32_t status, msi_rearm;
+	bool queue_hotplug = false;
 
+	/* reset gui idle ack.  the status bit is broken */
+	rdev->irq.gui_idle_acked = false;
+
+	status = r100_irq_ack(rdev);
+	if (!status) {
+		return IRQ_NONE;
+	}
+	if (rdev->shutdown) {
+		return IRQ_NONE;
+	}
+	while (status) {
+		/* SW interrupt */
+		if (status & RADEON_SW_INT_TEST) {
+			radeon_fence_process(rdev);
+		}
+		/* gui idle interrupt */
+		if (status & RADEON_GUI_IDLE_STAT) {
+			rdev->irq.gui_idle_acked = true;
+			rdev->pm.gui_idle = true;
+//			wake_up(&rdev->irq.idle_queue);
+		}
+		/* Vertical blank interrupts */
+		if (status & RADEON_CRTC_VBLANK_STAT) {
+			if (rdev->irq.crtc_vblank_int[0]) {
+//				drm_handle_vblank(rdev->ddev, 0);
+				rdev->pm.vblank_sync = true;
+//				wake_up(&rdev->irq.vblank_queue);
+			}
+//			if (rdev->irq.pflip[0])
+//				radeon_crtc_handle_flip(rdev, 0);
+		}
+		if (status & RADEON_CRTC2_VBLANK_STAT) {
+			if (rdev->irq.crtc_vblank_int[1]) {
+//				drm_handle_vblank(rdev->ddev, 1);
+				rdev->pm.vblank_sync = true;
+//				wake_up(&rdev->irq.vblank_queue);
+			}
+//			if (rdev->irq.pflip[1])
+//				radeon_crtc_handle_flip(rdev, 1);
+		}
+		if (status & RADEON_FP_DETECT_STAT) {
+			queue_hotplug = true;
+			DRM_DEBUG("HPD1\n");
+		}
+		if (status & RADEON_FP2_DETECT_STAT) {
+			queue_hotplug = true;
+			DRM_DEBUG("HPD2\n");
+		}
+		status = r100_irq_ack(rdev);
+	}
+	/* reset gui idle ack.  the status bit is broken */
+	rdev->irq.gui_idle_acked = false;
+//	if (queue_hotplug)
+//		schedule_work(&rdev->hotplug_work);
+	if (rdev->msi_enabled) {
+		switch (rdev->family) {
+		case CHIP_RS400:
+		case CHIP_RS480:
+			msi_rearm = RREG32(RADEON_AIC_CNTL) & ~RS400_MSI_REARM;
+			WREG32(RADEON_AIC_CNTL, msi_rearm);
+			WREG32(RADEON_AIC_CNTL, msi_rearm | RS400_MSI_REARM);
+			break;
+		default:
+			msi_rearm = RREG32(RADEON_MSI_REARM_EN) & ~RV370_MSI_REARM_EN;
+			WREG32(RADEON_MSI_REARM_EN, msi_rearm);
+			WREG32(RADEON_MSI_REARM_EN, msi_rearm | RV370_MSI_REARM_EN);
+			break;
+		}
+	}
+	return IRQ_HANDLED;
+}
 
 u32 r100_get_vblank_counter(struct radeon_device *rdev, int crtc)
 {
@@ -338,8 +443,6 @@ void r100_fence_ring_emit(struct radeon_device *rdev,
 	radeon_ring_write(rdev, RADEON_SW_INT_FIRE);
 }
 
-#if 0
-
 int r100_copy_blit(struct radeon_device *rdev,
 		   uint64_t src_offset,
 		   uint64_t dst_offset,
@@ -413,9 +516,6 @@ int r100_copy_blit(struct radeon_device *rdev,
 	return r;
 }
 
-#endif
-
-
 static int r100_cp_wait_for_idle(struct radeon_device *rdev)
 {
 	unsigned i;
@@ -1617,7 +1717,7 @@ int r100_mc_wait_for_idle(struct radeon_device *rdev)
 void r100_gpu_lockup_update(struct r100_gpu_lockup *lockup, struct radeon_cp *cp)
 {
 	lockup->last_cp_rptr = cp->rptr;
-    lockup->last_jiffies = 0; //jiffies;
+    lockup->last_jiffies = GetTimerTicks();
 }
 
 /**
@@ -1645,18 +1745,17 @@ bool r100_gpu_cp_is_lockup(struct radeon_device *rdev, struct r100_gpu_lockup *l
 {
 	unsigned long cjiffies, elapsed;
 
-#if 0
+	cjiffies = GetTimerTicks();
-	cjiffies = jiffies;
 	if (!time_after(cjiffies, lockup->last_jiffies)) {
 		/* likely a wrap around */
 		lockup->last_cp_rptr = cp->rptr;
-		lockup->last_jiffies = jiffies;
+		lockup->last_jiffies = GetTimerTicks();
 		return false;
 	}
 	if (cp->rptr != lockup->last_cp_rptr) {
 		/* CP is still working no lockup */
 		lockup->last_cp_rptr = cp->rptr;
-		lockup->last_jiffies = jiffies;
+		lockup->last_jiffies = GetTimerTicks();
 		return false;
 	}
 	elapsed = jiffies_to_msecs(cjiffies - lockup->last_jiffies);
@@ -1664,8 +1763,6 @@ bool r100_gpu_cp_is_lockup(struct radeon_device *rdev, struct r100_gpu_lockup *l
 		dev_err(rdev->dev, "GPU lockup CP stall for more than %lumsec\n", elapsed);
 		return true;
 	}
-#endif
-
 	/* give a chance to the GPU ... */
 	return false;
 }
@@ -3195,8 +3292,6 @@ int r100_ring_test(struct radeon_device *rdev)
 	return r;
 }
 
-#if 0
-
 void r100_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
 {
 	radeon_ring_write(rdev, PACKET0(RADEON_CP_IB_BASE, 1));
@@ -3283,7 +3378,6 @@ int r100_ib_init(struct radeon_device *rdev)
 	}
 	return 0;
 }
-#endif
 
 void r100_mc_stop(struct radeon_device *rdev, struct r100_mc_save *save)
 {
@@ -3436,8 +3530,14 @@ static int r100_startup(struct radeon_device *rdev)
 		if (r)
 			return r;
 	}
+
+	/* allocate wb buffer */
+	r = radeon_wb_init(rdev);
+	if (r)
+		return r;
+
 	/* Enable IRQ */
-//   r100_irq_set(rdev);
+	r100_irq_set(rdev);
 	rdev->config.r100.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL);
 	/* 1M ring buffer */
    r = r100_cp_init(rdev, 1024 * 1024);
@@ -3445,11 +3545,11 @@ static int r100_startup(struct radeon_device *rdev)
 		dev_err(rdev->dev, "failed initializing CP (%d).\n", r);
        return r;
    }
-//   r = r100_ib_init(rdev);
+	r = r100_ib_init(rdev);
-//   if (r) {
+	if (r) {
-//       dev_err(rdev->dev, "failled initializing IB (%d).\n", r);
+		dev_err(rdev->dev, "failed initializing IB (%d).\n", r);
-//       return r;
+		return r;
-//   }
+	}
 	return 0;
 }
 
@@ -3530,12 +3630,12 @@ int r100_init(struct radeon_device *rdev)
 	/* initialize VRAM */
 	r100_mc_init(rdev);
 	/* Fence driver */
-//	r = radeon_fence_driver_init(rdev);
+	r = radeon_fence_driver_init(rdev);
-//	if (r)
+	if (r)
-//		return r;
+		return r;
-//	r = radeon_irq_kms_init(rdev);
+	r = radeon_irq_kms_init(rdev);
-//	if (r)
+	if (r)
-//		return r;
+		return r;
 	/* Memory manager */
 	r = radeon_bo_init(rdev);
 	if (r)

drivers/video/drm/radeon/r200.c

@@ -82,6 +82,7 @@ static int r200_get_vtx_size_0(uint32_t vtx_fmt_0)
 		vtx_size += 3;
 	return vtx_size;
 }
+#endif
 
 int r200_copy_dma(struct radeon_device *rdev,
 		  uint64_t src_offset,
@@ -126,7 +127,7 @@ int r200_copy_dma(struct radeon_device *rdev,
 	radeon_ring_unlock_commit(rdev);
 	return r;
 }
-
+#if 0
 
 static int r200_get_vtx_size_1(uint32_t vtx_fmt_1)
 {

drivers/video/drm/radeon/r300.c

@@ -1395,9 +1395,13 @@ static int r300_startup(struct radeon_device *rdev)
 			return r;
 	}
 
+	/* allocate wb buffer */
+	r = radeon_wb_init(rdev);
+	if (r)
+		return r;
 
 	/* Enable IRQ */
-//	r100_irq_set(rdev);
+	r100_irq_set(rdev);
 	rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL);
 	/* 1M ring buffer */
     r = r100_cp_init(rdev, 1024 * 1024);
@@ -1405,11 +1409,11 @@ static int r300_startup(struct radeon_device *rdev)
 		dev_err(rdev->dev, "failed initializing CP (%d).\n", r);
        return r;
     }
-//   r = r100_ib_init(rdev);
+	r = r100_ib_init(rdev);
-//   if (r) {
+	if (r) {
-//       dev_err(rdev->dev, "failled initializing IB (%d).\n", r);
+		dev_err(rdev->dev, "failed initializing IB (%d).\n", r);
-//       return r;
+		return r;
-//   }
+	}
 	return 0;
 }
 
@@ -1467,12 +1471,12 @@ int r300_init(struct radeon_device *rdev)
 	/* initialize memory controller */
 	r300_mc_init(rdev);
 	/* Fence driver */
-//	r = radeon_fence_driver_init(rdev);
+	r = radeon_fence_driver_init(rdev);
-//	if (r)
+	if (r)
-//		return r;
+		return r;
-//	r = radeon_irq_kms_init(rdev);
+	r = radeon_irq_kms_init(rdev);
-//	if (r)
+	if (r)
-//		return r;
+		return r;
 	/* Memory manager */
 	r = radeon_bo_init(rdev);
 	if (r)

drivers/video/drm/radeon/r420.c

@@ -209,7 +209,14 @@ static int r420_startup(struct radeon_device *rdev)
 			return r;
 	}
 	r420_pipes_init(rdev);
+
+	/* allocate wb buffer */
+	r = radeon_wb_init(rdev);
+	if (r)
+		return r;
+
 	/* Enable IRQ */
+	r100_irq_set(rdev);
 	rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL);
 	/* 1M ring buffer */
 	r = r100_cp_init(rdev, 1024 * 1024);
@@ -218,6 +225,11 @@ static int r420_startup(struct radeon_device *rdev)
 		return r;
 	}
 	r420_cp_errata_init(rdev);
+	r = r100_ib_init(rdev);
+	if (r) {
+		dev_err(rdev->dev, "failed initializing IB (%d).\n", r);
+		return r;
+	}
 	return 0;
 }
 
@@ -302,7 +314,14 @@ int r420_init(struct radeon_device *rdev)
 	r300_mc_init(rdev);
 	r420_debugfs(rdev);
 	/* Fence driver */
-
+	r = radeon_fence_driver_init(rdev);
+	if (r) {
+		return r;
+	}
+	r = radeon_irq_kms_init(rdev);
+	if (r) {
+		return r;
+	}
 	/* Memory manager */
 	r = radeon_bo_init(rdev);
 	if (r) {

drivers/video/drm/radeon/r520.c

@@ -181,13 +181,25 @@ static int r520_startup(struct radeon_device *rdev)
 		if (r)
 			return r;
 	}
+
+	/* allocate wb buffer */
+	r = radeon_wb_init(rdev);
+	if (r)
+		return r;
+
 	/* Enable IRQ */
+	rs600_irq_set(rdev);
 	rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL);
 	/* 1M ring buffer */
     r = r100_cp_init(rdev, 1024 * 1024);
     if (r) {
 		dev_err(rdev->dev, "failed initializing CP (%d).\n", r);
         return r;
+    }
+	r = r100_ib_init(rdev);
+	if (r) {
+		dev_err(rdev->dev, "failed initializing IB (%d).\n", r);
+		return r;
 	}
 	return 0;
 }
@@ -246,6 +258,12 @@ int r520_init(struct radeon_device *rdev)
 	r520_mc_init(rdev);
 	rv515_debugfs(rdev);
 	/* Fence driver */
+	r = radeon_fence_driver_init(rdev);
+	if (r)
+		return r;
+	r = radeon_irq_kms_init(rdev);
+	if (r)
+		return r;
 	/* Memory manager */
 	r = radeon_bo_init(rdev);
 	if (r)

drivers/video/drm/radeon/r600.c

@@ -1871,6 +1871,26 @@ void r600_fence_ring_emit(struct radeon_device *rdev,
 	}
 }
 
+int r600_copy_blit(struct radeon_device *rdev,
+		   uint64_t src_offset, uint64_t dst_offset,
+		   unsigned num_pages, struct radeon_fence *fence)
+{
+	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;
+}
 
 int r600_set_surface_reg(struct radeon_device *rdev, int reg,
 			 uint32_t tiling_flags, uint32_t pitch,
@@ -1909,6 +1929,26 @@ int r600_startup(struct radeon_device *rdev)
 			return r;
 	}
 	r600_gpu_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);
+	}
+
+	/* allocate wb buffer */
+	r = radeon_wb_init(rdev);
+	if (r)
+		return r;
+
+	/* Enable IRQ */
+	r = r600_irq_init(rdev);
+	if (r) {
+		DRM_ERROR("radeon: IH init failed (%d).\n", r);
+//		radeon_irq_kms_fini(rdev);
+		return r;
+	}
+	r600_irq_set(rdev);
 
 	r = radeon_ring_init(rdev, rdev->cp.ring_size);
 	if (r)
@@ -2028,19 +2068,19 @@ int r600_init(struct radeon_device *rdev)
 		rdev->accel_working = false;
 	}
 	if (rdev->accel_working) {
-//		r = radeon_ib_pool_init(rdev);
+		r = radeon_ib_pool_init(rdev);
-//		if (r) {
+		if (r) {
-//			DRM_ERROR("radeon: failled initializing IB pool (%d).\n", r);
+			dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
-//			rdev->accel_working = false;
+			rdev->accel_working = false;
-//		}
+		} else {
-//		r = r600_ib_test(rdev);
+			r = r600_ib_test(rdev);
-//		if (r) {
+			if (r) {
-//			DRM_ERROR("radeon: failled testing IB (%d).\n", r);
+				dev_err(rdev->dev, "IB test failed (%d).\n", r);
-//			rdev->accel_working = false;
+				rdev->accel_working = false;
-//		}
 			}
-	if (r)
+	}
-		return r; /* TODO error handling */
+	}
+
 	return 0;
 }
 
@@ -2423,8 +2463,6 @@ int r600_irq_set(struct radeon_device *rdev)
 	u32 hdmi1, hdmi2;
 	u32 d1grph = 0, d2grph = 0;
 
-    ENTER();
-
 	if (!rdev->irq.installed) {
 		WARN(1, "Can't enable IRQ/MSI because no handler is installed\n");
 		return -EINVAL;
@@ -2530,8 +2568,6 @@ int r600_irq_set(struct radeon_device *rdev)
 		WREG32(DC_HOT_PLUG_DETECT3_INT_CONTROL, hpd3);
 	}
 
-    LEAVE();
-
 	return 0;
 }
 
@@ -2825,11 +2861,11 @@ restart_ih:
 		case 177: /* CP_INT in IB1 */
 		case 178: /* CP_INT in IB2 */
 			DRM_DEBUG("IH: CP int: 0x%08x\n", src_data);
-//           radeon_fence_process(rdev);
+            radeon_fence_process(rdev);
 			break;
 		case 181: /* CP EOP event */
 			DRM_DEBUG("IH: CP EOP\n");
-//           radeon_fence_process(rdev);
+			radeon_fence_process(rdev);
 			break;
 		case 233: /* GUI IDLE */
 			DRM_DEBUG("IH: GUI idle\n");

drivers/video/drm/radeon/r600_audio.c

@@ -26,6 +26,7 @@
 #include "drmP.h"
 #include "radeon.h"
 #include "radeon_reg.h"
+#include "radeon_asic.h"
 #include "atom.h"
 
 #define AUDIO_TIMER_INTERVALL 100 /* 1/10 sekund should be enough */

drivers/video/drm/radeon/r600_blit_kms.c

@@ -0,0 +1,845 @@
+/*
+ * Copyright 2009 Advanced Micro Devices, Inc.
+ * Copyright 2009 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 (including the next
+ * paragraph) 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) AND/OR ITS SUPPLIERS 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.
+ *
+ */
+
+#include "drmP.h"
+#include "drm.h"
+#include "radeon_drm.h"
+#include "radeon.h"
+
+#include "r600d.h"
+#include "r600_blit_shaders.h"
+
+#define DI_PT_RECTLIST        0x11
+#define DI_INDEX_SIZE_16_BIT  0x0
+#define DI_SRC_SEL_AUTO_INDEX 0x2
+
+#define FMT_8                 0x1
+#define FMT_5_6_5             0x8
+#define FMT_8_8_8_8           0x1a
+#define COLOR_8               0x1
+#define COLOR_5_6_5           0x8
+#define COLOR_8_8_8_8         0x1a
+
+/* emits 21 on rv770+, 23 on r600 */
+static void
+set_render_target(struct radeon_device *rdev, int format,
+		  int w, int h, u64 gpu_addr)
+{
+	u32 cb_color_info;
+	int pitch, slice;
+
+	h = ALIGN(h, 8);
+	if (h < 8)
+		h = 8;
+
+	cb_color_info = ((format << 2) | (1 << 27) | (1 << 8));
+	pitch = (w / 8) - 1;
+	slice = ((w * h) / 64) - 1;
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (CB_COLOR0_BASE - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, gpu_addr >> 8);
+
+	if (rdev->family > CHIP_R600 && rdev->family < CHIP_RV770) {
+		radeon_ring_write(rdev, PACKET3(PACKET3_SURFACE_BASE_UPDATE, 0));
+		radeon_ring_write(rdev, 2 << 0);
+	}
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (CB_COLOR0_SIZE - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, (pitch << 0) | (slice << 10));
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (CB_COLOR0_VIEW - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, 0);
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (CB_COLOR0_INFO - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, cb_color_info);
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (CB_COLOR0_TILE - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, 0);
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (CB_COLOR0_FRAG - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, 0);
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (CB_COLOR0_MASK - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, 0);
+}
+
+/* emits 5dw */
+static void
+cp_set_surface_sync(struct radeon_device *rdev,
+		    u32 sync_type, u32 size,
+		    u64 mc_addr)
+{
+	u32 cp_coher_size;
+
+	if (size == 0xffffffff)
+		cp_coher_size = 0xffffffff;
+	else
+		cp_coher_size = ((size + 255) >> 8);
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SURFACE_SYNC, 3));
+	radeon_ring_write(rdev, sync_type);
+	radeon_ring_write(rdev, cp_coher_size);
+	radeon_ring_write(rdev, mc_addr >> 8);
+	radeon_ring_write(rdev, 10); /* poll interval */
+}
+
+/* emits 21dw + 1 surface sync = 26dw */
+static void
+set_shaders(struct radeon_device *rdev)
+{
+	u64 gpu_addr;
+	u32 sq_pgm_resources;
+
+	/* setup shader regs */
+	sq_pgm_resources = (1 << 0);
+
+	/* VS */
+	gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.vs_offset;
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (SQ_PGM_START_VS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, gpu_addr >> 8);
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (SQ_PGM_RESOURCES_VS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, sq_pgm_resources);
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (SQ_PGM_CF_OFFSET_VS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, 0);
+
+	/* PS */
+	gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.ps_offset;
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (SQ_PGM_START_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, gpu_addr >> 8);
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (SQ_PGM_RESOURCES_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, sq_pgm_resources | (1 << 28));
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (SQ_PGM_EXPORTS_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, 2);
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (SQ_PGM_CF_OFFSET_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, 0);
+
+	gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.vs_offset;
+	cp_set_surface_sync(rdev, PACKET3_SH_ACTION_ENA, 512, gpu_addr);
+}
+
+/* emits 9 + 1 sync (5) = 14*/
+static void
+set_vtx_resource(struct radeon_device *rdev, u64 gpu_addr)
+{
+	u32 sq_vtx_constant_word2;
+
+	sq_vtx_constant_word2 = ((upper_32_bits(gpu_addr) & 0xff) | (16 << 8));
+#ifdef __BIG_ENDIAN
+	sq_vtx_constant_word2 |= (2 << 30);
+#endif
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_RESOURCE, 7));
+	radeon_ring_write(rdev, 0x460);
+	radeon_ring_write(rdev, gpu_addr & 0xffffffff);
+	radeon_ring_write(rdev, 48 - 1);
+	radeon_ring_write(rdev, sq_vtx_constant_word2);
+	radeon_ring_write(rdev, 1 << 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, SQ_TEX_VTX_VALID_BUFFER << 30);
+
+	if ((rdev->family == CHIP_RV610) ||
+	    (rdev->family == CHIP_RV620) ||
+	    (rdev->family == CHIP_RS780) ||
+	    (rdev->family == CHIP_RS880) ||
+	    (rdev->family == CHIP_RV710))
+		cp_set_surface_sync(rdev,
+				    PACKET3_TC_ACTION_ENA, 48, gpu_addr);
+	else
+		cp_set_surface_sync(rdev,
+				    PACKET3_VC_ACTION_ENA, 48, gpu_addr);
+}
+
+/* emits 9 */
+static void
+set_tex_resource(struct radeon_device *rdev,
+		 int format, int w, int h, int pitch,
+		 u64 gpu_addr)
+{
+	uint32_t sq_tex_resource_word0, sq_tex_resource_word1, sq_tex_resource_word4;
+
+	if (h < 1)
+		h = 1;
+
+	sq_tex_resource_word0 = (1 << 0) | (1 << 3);
+	sq_tex_resource_word0 |= ((((pitch >> 3) - 1) << 8) |
+				  ((w - 1) << 19));
+
+	sq_tex_resource_word1 = (format << 26);
+	sq_tex_resource_word1 |= ((h - 1) << 0);
+
+	sq_tex_resource_word4 = ((1 << 14) |
+				 (0 << 16) |
+				 (1 << 19) |
+				 (2 << 22) |
+				 (3 << 25));
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_RESOURCE, 7));
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, sq_tex_resource_word0);
+	radeon_ring_write(rdev, sq_tex_resource_word1);
+	radeon_ring_write(rdev, gpu_addr >> 8);
+	radeon_ring_write(rdev, gpu_addr >> 8);
+	radeon_ring_write(rdev, sq_tex_resource_word4);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, SQ_TEX_VTX_VALID_TEXTURE << 30);
+}
+
+/* emits 12 */
+static void
+set_scissors(struct radeon_device *rdev, int x1, int y1,
+	     int x2, int y2)
+{
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
+	radeon_ring_write(rdev, (PA_SC_SCREEN_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, (x1 << 0) | (y1 << 16));
+	radeon_ring_write(rdev, (x2 << 0) | (y2 << 16));
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
+	radeon_ring_write(rdev, (PA_SC_GENERIC_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, (x1 << 0) | (y1 << 16) | (1 << 31));
+	radeon_ring_write(rdev, (x2 << 0) | (y2 << 16));
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
+	radeon_ring_write(rdev, (PA_SC_WINDOW_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, (x1 << 0) | (y1 << 16) | (1 << 31));
+	radeon_ring_write(rdev, (x2 << 0) | (y2 << 16));
+}
+
+/* emits 10 */
+static void
+draw_auto(struct radeon_device *rdev)
+{
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1));
+	radeon_ring_write(rdev, (VGT_PRIMITIVE_TYPE - PACKET3_SET_CONFIG_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, DI_PT_RECTLIST);
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_INDEX_TYPE, 0));
+	radeon_ring_write(rdev,
+#ifdef __BIG_ENDIAN
+			  (2 << 2) |
+#endif
+			  DI_INDEX_SIZE_16_BIT);
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_NUM_INSTANCES, 0));
+	radeon_ring_write(rdev, 1);
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_DRAW_INDEX_AUTO, 1));
+	radeon_ring_write(rdev, 3);
+	radeon_ring_write(rdev, DI_SRC_SEL_AUTO_INDEX);
+
+}
+
+/* emits 14 */
+static void
+set_default_state(struct radeon_device *rdev)
+{
+	u32 sq_config, sq_gpr_resource_mgmt_1, sq_gpr_resource_mgmt_2;
+	u32 sq_thread_resource_mgmt, sq_stack_resource_mgmt_1, sq_stack_resource_mgmt_2;
+	int num_ps_gprs, num_vs_gprs, num_temp_gprs, num_gs_gprs, num_es_gprs;
+	int num_ps_threads, num_vs_threads, num_gs_threads, num_es_threads;
+	int num_ps_stack_entries, num_vs_stack_entries, num_gs_stack_entries, num_es_stack_entries;
+	u64 gpu_addr;
+	int dwords;
+
+	switch (rdev->family) {
+	case CHIP_R600:
+		num_ps_gprs = 192;
+		num_vs_gprs = 56;
+		num_temp_gprs = 4;
+		num_gs_gprs = 0;
+		num_es_gprs = 0;
+		num_ps_threads = 136;
+		num_vs_threads = 48;
+		num_gs_threads = 4;
+		num_es_threads = 4;
+		num_ps_stack_entries = 128;
+		num_vs_stack_entries = 128;
+		num_gs_stack_entries = 0;
+		num_es_stack_entries = 0;
+		break;
+	case CHIP_RV630:
+	case CHIP_RV635:
+		num_ps_gprs = 84;
+		num_vs_gprs = 36;
+		num_temp_gprs = 4;
+		num_gs_gprs = 0;
+		num_es_gprs = 0;
+		num_ps_threads = 144;
+		num_vs_threads = 40;
+		num_gs_threads = 4;
+		num_es_threads = 4;
+		num_ps_stack_entries = 40;
+		num_vs_stack_entries = 40;
+		num_gs_stack_entries = 32;
+		num_es_stack_entries = 16;
+		break;
+	case CHIP_RV610:
+	case CHIP_RV620:
+	case CHIP_RS780:
+	case CHIP_RS880:
+	default:
+		num_ps_gprs = 84;
+		num_vs_gprs = 36;
+		num_temp_gprs = 4;
+		num_gs_gprs = 0;
+		num_es_gprs = 0;
+		num_ps_threads = 136;
+		num_vs_threads = 48;
+		num_gs_threads = 4;
+		num_es_threads = 4;
+		num_ps_stack_entries = 40;
+		num_vs_stack_entries = 40;
+		num_gs_stack_entries = 32;
+		num_es_stack_entries = 16;
+		break;
+	case CHIP_RV670:
+		num_ps_gprs = 144;
+		num_vs_gprs = 40;
+		num_temp_gprs = 4;
+		num_gs_gprs = 0;
+		num_es_gprs = 0;
+		num_ps_threads = 136;
+		num_vs_threads = 48;
+		num_gs_threads = 4;
+		num_es_threads = 4;
+		num_ps_stack_entries = 40;
+		num_vs_stack_entries = 40;
+		num_gs_stack_entries = 32;
+		num_es_stack_entries = 16;
+		break;
+	case CHIP_RV770:
+		num_ps_gprs = 192;
+		num_vs_gprs = 56;
+		num_temp_gprs = 4;
+		num_gs_gprs = 0;
+		num_es_gprs = 0;
+		num_ps_threads = 188;
+		num_vs_threads = 60;
+		num_gs_threads = 0;
+		num_es_threads = 0;
+		num_ps_stack_entries = 256;
+		num_vs_stack_entries = 256;
+		num_gs_stack_entries = 0;
+		num_es_stack_entries = 0;
+		break;
+	case CHIP_RV730:
+	case CHIP_RV740:
+		num_ps_gprs = 84;
+		num_vs_gprs = 36;
+		num_temp_gprs = 4;
+		num_gs_gprs = 0;
+		num_es_gprs = 0;
+		num_ps_threads = 188;
+		num_vs_threads = 60;
+		num_gs_threads = 0;
+		num_es_threads = 0;
+		num_ps_stack_entries = 128;
+		num_vs_stack_entries = 128;
+		num_gs_stack_entries = 0;
+		num_es_stack_entries = 0;
+		break;
+	case CHIP_RV710:
+		num_ps_gprs = 192;
+		num_vs_gprs = 56;
+		num_temp_gprs = 4;
+		num_gs_gprs = 0;
+		num_es_gprs = 0;
+		num_ps_threads = 144;
+		num_vs_threads = 48;
+		num_gs_threads = 0;
+		num_es_threads = 0;
+		num_ps_stack_entries = 128;
+		num_vs_stack_entries = 128;
+		num_gs_stack_entries = 0;
+		num_es_stack_entries = 0;
+		break;
+	}
+
+	if ((rdev->family == CHIP_RV610) ||
+	    (rdev->family == CHIP_RV620) ||
+	    (rdev->family == CHIP_RS780) ||
+	    (rdev->family == CHIP_RS880) ||
+	    (rdev->family == CHIP_RV710))
+		sq_config = 0;
+	else
+		sq_config = VC_ENABLE;
+
+	sq_config |= (DX9_CONSTS |
+		      ALU_INST_PREFER_VECTOR |
+		      PS_PRIO(0) |
+		      VS_PRIO(1) |
+		      GS_PRIO(2) |
+		      ES_PRIO(3));
+
+	sq_gpr_resource_mgmt_1 = (NUM_PS_GPRS(num_ps_gprs) |
+				  NUM_VS_GPRS(num_vs_gprs) |
+				  NUM_CLAUSE_TEMP_GPRS(num_temp_gprs));
+	sq_gpr_resource_mgmt_2 = (NUM_GS_GPRS(num_gs_gprs) |
+				  NUM_ES_GPRS(num_es_gprs));
+	sq_thread_resource_mgmt = (NUM_PS_THREADS(num_ps_threads) |
+				   NUM_VS_THREADS(num_vs_threads) |
+				   NUM_GS_THREADS(num_gs_threads) |
+				   NUM_ES_THREADS(num_es_threads));
+	sq_stack_resource_mgmt_1 = (NUM_PS_STACK_ENTRIES(num_ps_stack_entries) |
+				    NUM_VS_STACK_ENTRIES(num_vs_stack_entries));
+	sq_stack_resource_mgmt_2 = (NUM_GS_STACK_ENTRIES(num_gs_stack_entries) |
+				    NUM_ES_STACK_ENTRIES(num_es_stack_entries));
+
+	/* emit an IB pointing at default state */
+	dwords = ALIGN(rdev->r600_blit.state_len, 0x10);
+	gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.state_offset;
+	radeon_ring_write(rdev, PACKET3(PACKET3_INDIRECT_BUFFER, 2));
+	radeon_ring_write(rdev,
+#ifdef __BIG_ENDIAN
+			  (2 << 0) |
+#endif
+			  (gpu_addr & 0xFFFFFFFC));
+	radeon_ring_write(rdev, upper_32_bits(gpu_addr) & 0xFF);
+	radeon_ring_write(rdev, dwords);
+
+	/* SQ config */
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 6));
+	radeon_ring_write(rdev, (SQ_CONFIG - PACKET3_SET_CONFIG_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, sq_config);
+	radeon_ring_write(rdev, sq_gpr_resource_mgmt_1);
+	radeon_ring_write(rdev, sq_gpr_resource_mgmt_2);
+	radeon_ring_write(rdev, sq_thread_resource_mgmt);
+	radeon_ring_write(rdev, sq_stack_resource_mgmt_1);
+	radeon_ring_write(rdev, sq_stack_resource_mgmt_2);
+}
+
+static inline uint32_t i2f(uint32_t input)
+{
+	u32 result, i, exponent, fraction;
+
+	if ((input & 0x3fff) == 0)
+		result = 0; /* 0 is a special case */
+	else {
+		exponent = 140; /* exponent biased by 127; */
+		fraction = (input & 0x3fff) << 10; /* cheat and only
+						      handle numbers below 2^^15 */
+		for (i = 0; i < 14; i++) {
+			if (fraction & 0x800000)
+				break;
+			else {
+				fraction = fraction << 1; /* keep
+							     shifting left until top bit = 1 */
+				exponent = exponent - 1;
+			}
+		}
+		result = exponent << 23 | (fraction & 0x7fffff); /* mask
+								    off top bit; assumed 1 */
+	}
+	return result;
+}
+
+int r600_blit_init(struct radeon_device *rdev)
+{
+	u32 obj_size;
+	int i, r, dwords;
+	void *ptr;
+	u32 packet2s[16];
+	int num_packet2s = 0;
+
+    ENTER();
+
+	/* pin copy shader into vram if already initialized */
+	if (rdev->r600_blit.shader_obj)
+		goto done;
+
+	mutex_init(&rdev->r600_blit.mutex);
+	rdev->r600_blit.state_offset = 0;
+
+	if (rdev->family >= CHIP_RV770)
+		rdev->r600_blit.state_len = r7xx_default_size;
+	else
+		rdev->r600_blit.state_len = r6xx_default_size;
+
+	dwords = rdev->r600_blit.state_len;
+	while (dwords & 0xf) {
+		packet2s[num_packet2s++] = cpu_to_le32(PACKET2(0));
+		dwords++;
+	}
+
+	obj_size = dwords * 4;
+	obj_size = ALIGN(obj_size, 256);
+
+	rdev->r600_blit.vs_offset = obj_size;
+	obj_size += r6xx_vs_size * 4;
+	obj_size = ALIGN(obj_size, 256);
+
+	rdev->r600_blit.ps_offset = obj_size;
+	obj_size += r6xx_ps_size * 4;
+	obj_size = ALIGN(obj_size, 256);
+
+	r = radeon_bo_create(rdev, obj_size, PAGE_SIZE, true, RADEON_GEM_DOMAIN_VRAM,
+				&rdev->r600_blit.shader_obj);
+	if (r) {
+		DRM_ERROR("r600 failed to allocate shader\n");
+		return r;
+	}
+
+	DRM_DEBUG("r6xx blit allocated bo %08x vs %08x ps %08x\n",
+		  obj_size,
+		  rdev->r600_blit.vs_offset, rdev->r600_blit.ps_offset);
+
+	r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false);
+	if (unlikely(r != 0))
+		return r;
+	r = radeon_bo_kmap(rdev->r600_blit.shader_obj, &ptr);
+	if (r) {
+		DRM_ERROR("failed to map blit object %d\n", r);
+		return r;
+	}
+	if (rdev->family >= CHIP_RV770)
+        memcpy(ptr + rdev->r600_blit.state_offset,
+			    r7xx_default_state, rdev->r600_blit.state_len * 4);
+	else
+        memcpy(ptr + rdev->r600_blit.state_offset,
+			    r6xx_default_state, rdev->r600_blit.state_len * 4);
+	if (num_packet2s)
+        memcpy(ptr + rdev->r600_blit.state_offset + (rdev->r600_blit.state_len * 4),
+			    packet2s, num_packet2s * 4);
+	for (i = 0; i < r6xx_vs_size; i++)
+		*(u32 *)((unsigned long)ptr + rdev->r600_blit.vs_offset + i * 4) = cpu_to_le32(r6xx_vs[i]);
+	for (i = 0; i < r6xx_ps_size; i++)
+		*(u32 *)((unsigned long)ptr + rdev->r600_blit.ps_offset + i * 4) = cpu_to_le32(r6xx_ps[i]);
+	radeon_bo_kunmap(rdev->r600_blit.shader_obj);
+	radeon_bo_unreserve(rdev->r600_blit.shader_obj);
+
+done:
+	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;
+	}
+//   radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size);
+
+    LEAVE();
+
+	return 0;
+}
+
+void r600_blit_fini(struct radeon_device *rdev)
+{
+	int r;
+
+//   radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size);
+	if (rdev->r600_blit.shader_obj == NULL)
+		return;
+	/* If we can't reserve the bo, unref should be enough to destroy
+	 * it when it becomes idle.
+	 */
+	r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false);
+	if (!r) {
+		radeon_bo_unpin(rdev->r600_blit.shader_obj);
+		radeon_bo_unreserve(rdev->r600_blit.shader_obj);
+	}
+	radeon_bo_unref(&rdev->r600_blit.shader_obj);
+}
+
+static int r600_vb_ib_get(struct radeon_device *rdev)
+{
+	int r;
+	r = radeon_ib_get(rdev, &rdev->r600_blit.vb_ib);
+	if (r) {
+		DRM_ERROR("failed to get IB for vertex buffer\n");
+		return r;
+	}
+
+	rdev->r600_blit.vb_total = 64*1024;
+	rdev->r600_blit.vb_used = 0;
+	return 0;
+}
+
+static void r600_vb_ib_put(struct radeon_device *rdev)
+{
+	radeon_fence_emit(rdev, rdev->r600_blit.vb_ib->fence);
+	radeon_ib_free(rdev, &rdev->r600_blit.vb_ib);
+}
+
+int r600_blit_prepare_copy(struct radeon_device *rdev, int size_bytes)
+{
+	int r;
+	int ring_size, line_size;
+	int max_size;
+	/* loops of emits 64 + fence emit possible */
+	int dwords_per_loop = 76, num_loops;
+
+	r = r600_vb_ib_get(rdev);
+	if (r)
+		return r;
+
+	/* set_render_target emits 2 extra dwords on rv6xx */
+	if (rdev->family > CHIP_R600 && rdev->family < CHIP_RV770)
+		dwords_per_loop += 2;
+
+	/* 8 bpp vs 32 bpp for xfer unit */
+	if (size_bytes & 3)
+		line_size = 8192;
+	else
+		line_size = 8192*4;
+
+	max_size = 8192 * line_size;
+
+	/* major loops cover the max size transfer */
+	num_loops = ((size_bytes + max_size) / max_size);
+	/* minor loops cover the extra non aligned bits */
+	num_loops += ((size_bytes % line_size) ? 1 : 0);
+	/* calculate number of loops correctly */
+	ring_size = num_loops * dwords_per_loop;
+	/* set default  + shaders */
+	ring_size += 40; /* shaders + def state */
+	ring_size += 10; /* fence emit for VB IB */
+	ring_size += 5; /* done copy */
+	ring_size += 10; /* fence emit for done copy */
+	r = radeon_ring_lock(rdev, ring_size);
+	if (r)
+		return r;
+
+	set_default_state(rdev); /* 14 */
+	set_shaders(rdev); /* 26 */
+	return 0;
+}
+
+void r600_blit_done_copy(struct radeon_device *rdev, struct radeon_fence *fence)
+{
+	int r;
+
+	if (rdev->r600_blit.vb_ib)
+		r600_vb_ib_put(rdev);
+
+	if (fence)
+		r = radeon_fence_emit(rdev, fence);
+
+	radeon_ring_unlock_commit(rdev);
+}
+
+void r600_kms_blit_copy(struct radeon_device *rdev,
+			u64 src_gpu_addr, u64 dst_gpu_addr,
+			int size_bytes)
+{
+	int max_bytes;
+	u64 vb_gpu_addr;
+	u32 *vb;
+
+	DRM_DEBUG("emitting copy %16llx %16llx %d %d\n", src_gpu_addr, dst_gpu_addr,
+		  size_bytes, rdev->r600_blit.vb_used);
+	vb = (u32 *)(rdev->r600_blit.vb_ib->ptr + rdev->r600_blit.vb_used);
+	if ((size_bytes & 3) || (src_gpu_addr & 3) || (dst_gpu_addr & 3)) {
+		max_bytes = 8192;
+
+		while (size_bytes) {
+			int cur_size = size_bytes;
+			int src_x = src_gpu_addr & 255;
+			int dst_x = dst_gpu_addr & 255;
+			int h = 1;
+			src_gpu_addr = src_gpu_addr & ~255ULL;
+			dst_gpu_addr = dst_gpu_addr & ~255ULL;
+
+			if (!src_x && !dst_x) {
+				h = (cur_size / max_bytes);
+				if (h > 8192)
+					h = 8192;
+				if (h == 0)
+					h = 1;
+				else
+					cur_size = max_bytes;
+			} else {
+				if (cur_size > max_bytes)
+					cur_size = max_bytes;
+				if (cur_size > (max_bytes - dst_x))
+					cur_size = (max_bytes - dst_x);
+				if (cur_size > (max_bytes - src_x))
+					cur_size = (max_bytes - src_x);
+			}
+
+			if ((rdev->r600_blit.vb_used + 48) > rdev->r600_blit.vb_total) {
+             //  WARN_ON(1);
+			}
+
+			vb[0] = i2f(dst_x);
+			vb[1] = 0;
+			vb[2] = i2f(src_x);
+			vb[3] = 0;
+
+			vb[4] = i2f(dst_x);
+			vb[5] = i2f(h);
+			vb[6] = i2f(src_x);
+			vb[7] = i2f(h);
+
+			vb[8] = i2f(dst_x + cur_size);
+			vb[9] = i2f(h);
+			vb[10] = i2f(src_x + cur_size);
+			vb[11] = i2f(h);
+
+			/* src 9 */
+			set_tex_resource(rdev, FMT_8,
+					 src_x + cur_size, h, src_x + cur_size,
+					 src_gpu_addr);
+
+			/* 5 */
+			cp_set_surface_sync(rdev,
+					    PACKET3_TC_ACTION_ENA, (src_x + cur_size * h), src_gpu_addr);
+
+			/* dst 23 */
+			set_render_target(rdev, COLOR_8,
+					  dst_x + cur_size, h,
+					  dst_gpu_addr);
+
+			/* scissors 12 */
+			set_scissors(rdev, dst_x, 0, dst_x + cur_size, h);
+
+			/* 14 */
+			vb_gpu_addr = rdev->r600_blit.vb_ib->gpu_addr + rdev->r600_blit.vb_used;
+			set_vtx_resource(rdev, vb_gpu_addr);
+
+			/* draw 10 */
+			draw_auto(rdev);
+
+			/* 5 */
+			cp_set_surface_sync(rdev,
+					    PACKET3_CB_ACTION_ENA | PACKET3_CB0_DEST_BASE_ENA,
+					    cur_size * h, dst_gpu_addr);
+
+			vb += 12;
+			rdev->r600_blit.vb_used += 12 * 4;
+
+			src_gpu_addr += cur_size * h;
+			dst_gpu_addr += cur_size * h;
+			size_bytes -= cur_size * h;
+		}
+	} else {
+		max_bytes = 8192 * 4;
+
+		while (size_bytes) {
+			int cur_size = size_bytes;
+			int src_x = (src_gpu_addr & 255);
+			int dst_x = (dst_gpu_addr & 255);
+			int h = 1;
+			src_gpu_addr = src_gpu_addr & ~255ULL;
+			dst_gpu_addr = dst_gpu_addr & ~255ULL;
+
+			if (!src_x && !dst_x) {
+				h = (cur_size / max_bytes);
+				if (h > 8192)
+					h = 8192;
+				if (h == 0)
+					h = 1;
+				else
+					cur_size = max_bytes;
+			} else {
+				if (cur_size > max_bytes)
+					cur_size = max_bytes;
+				if (cur_size > (max_bytes - dst_x))
+					cur_size = (max_bytes - dst_x);
+				if (cur_size > (max_bytes - src_x))
+					cur_size = (max_bytes - src_x);
+			}
+
+			if ((rdev->r600_blit.vb_used + 48) > rdev->r600_blit.vb_total) {
+              // WARN_ON(1);
+			}
+
+			vb[0] = i2f(dst_x / 4);
+			vb[1] = 0;
+			vb[2] = i2f(src_x / 4);
+			vb[3] = 0;
+
+			vb[4] = i2f(dst_x / 4);
+			vb[5] = i2f(h);
+			vb[6] = i2f(src_x / 4);
+			vb[7] = i2f(h);
+
+			vb[8] = i2f((dst_x + cur_size) / 4);
+			vb[9] = i2f(h);
+			vb[10] = i2f((src_x + cur_size) / 4);
+			vb[11] = i2f(h);
+
+			/* src 9 */
+			set_tex_resource(rdev, FMT_8_8_8_8,
+					 (src_x + cur_size) / 4,
+					 h, (src_x + cur_size) / 4,
+					 src_gpu_addr);
+			/* 5 */
+			cp_set_surface_sync(rdev,
+					    PACKET3_TC_ACTION_ENA, (src_x + cur_size * h), src_gpu_addr);
+
+			/* dst 23 */
+			set_render_target(rdev, COLOR_8_8_8_8,
+					  (dst_x + cur_size) / 4, h,
+					  dst_gpu_addr);
+
+			/* scissors 12  */
+			set_scissors(rdev, (dst_x / 4), 0, (dst_x + cur_size / 4), h);
+
+			/* Vertex buffer setup 14 */
+			vb_gpu_addr = rdev->r600_blit.vb_ib->gpu_addr + rdev->r600_blit.vb_used;
+			set_vtx_resource(rdev, vb_gpu_addr);
+
+			/* draw 10 */
+			draw_auto(rdev);
+
+			/* 5 */
+			cp_set_surface_sync(rdev,
+					    PACKET3_CB_ACTION_ENA | PACKET3_CB0_DEST_BASE_ENA,
+					    cur_size * h, dst_gpu_addr);
+
+			/* 78 ring dwords per loop */
+			vb += 12;
+			rdev->r600_blit.vb_used += 12 * 4;
+
+			src_gpu_addr += cur_size * h;
+			dst_gpu_addr += cur_size * h;
+			size_bytes -= cur_size * h;
+		}
+	}
+}
+

drivers/video/drm/radeon/r600_blit_shaders.c

@@ -0,0 +1,710 @@
+/*
+ * Copyright 2009 Advanced Micro Devices, 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 (including the next
+ * paragraph) 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) AND/OR ITS SUPPLIERS 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:
+ *     Alex Deucher <alexander.deucher@amd.com>
+ */
+
+#include <linux/types.h>
+#include <linux/kernel.h>
+
+/*
+ * R6xx+ cards need to use the 3D engine to blit data which requires
+ * quite a bit of hw state setup.  Rather than pull the whole 3D driver
+ * (which normally generates the 3D state) into the DRM, we opt to use
+ * statically generated state tables.  The regsiter state and shaders
+ * were hand generated to support blitting functionality.  See the 3D
+ * driver or documentation for descriptions of the registers and
+ * shader instructions.
+ */
+
+const u32 r6xx_default_state[] =
+{
+	0xc0002400, /* START_3D_CMDBUF */
+	0x00000000,
+
+	0xc0012800, /* CONTEXT_CONTROL */
+	0x80000000,
+	0x80000000,
+
+	0xc0016800,
+	0x00000010,
+	0x00008000, /* WAIT_UNTIL */
+
+	0xc0016800,
+	0x00000542,
+	0x07000003, /* TA_CNTL_AUX */
+
+	0xc0016800,
+	0x000005c5,
+	0x00000000, /* VC_ENHANCE */
+
+	0xc0016800,
+	0x00000363,
+	0x00000000, /* SQ_DYN_GPR_CNTL_PS_FLUSH_REQ */
+
+	0xc0016800,
+	0x0000060c,
+	0x82000000, /* DB_DEBUG */
+
+	0xc0016800,
+	0x0000060e,
+	0x01020204, /* DB_WATERMARKS */
+
+	0xc0026f00,
+	0x00000000,
+	0x00000000, /* SQ_VTX_BASE_VTX_LOC */
+	0x00000000, /* SQ_VTX_START_INST_LOC */
+
+	0xc0096900,
+	0x0000022a,
+	0x00000000, /* SQ_ESGS_RING_ITEMSIZE */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+
+	0xc0016900,
+	0x00000004,
+	0x00000000, /* DB_DEPTH_INFO */
+
+	0xc0026900,
+	0x0000000a,
+	0x00000000, /* DB_STENCIL_CLEAR */
+	0x00000000, /* DB_DEPTH_CLEAR */
+
+	0xc0016900,
+	0x00000200,
+	0x00000000, /* DB_DEPTH_CONTROL */
+
+	0xc0026900,
+	0x00000343,
+	0x00000060, /* DB_RENDER_CONTROL */
+	0x00000040, /* DB_RENDER_OVERRIDE */
+
+	0xc0016900,
+	0x00000351,
+	0x0000aa00, /* DB_ALPHA_TO_MASK */
+
+	0xc00f6900,
+	0x00000100,
+	0x00000800, /* VGT_MAX_VTX_INDX */
+	0x00000000, /* VGT_MIN_VTX_INDX */
+	0x00000000, /* VGT_INDX_OFFSET */
+	0x00000000, /* VGT_MULTI_PRIM_IB_RESET_INDX */
+	0x00000000, /* SX_ALPHA_TEST_CONTROL */
+	0x00000000, /* CB_BLEND_RED */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000, /* CB_FOG_RED */
+	0x00000000,
+	0x00000000,
+	0x00000000, /* DB_STENCILREFMASK */
+	0x00000000, /* DB_STENCILREFMASK_BF */
+	0x00000000, /* SX_ALPHA_REF */
+
+	0xc0046900,
+	0x0000030c,
+	0x01000000, /* CB_CLRCMP_CNTL */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+
+	0xc0046900,
+	0x00000048,
+	0x3f800000, /* CB_CLEAR_RED */
+	0x00000000,
+	0x3f800000,
+	0x3f800000,
+
+	0xc0016900,
+	0x00000080,
+	0x00000000, /* PA_SC_WINDOW_OFFSET */
+
+	0xc00a6900,
+	0x00000083,
+	0x0000ffff, /* PA_SC_CLIP_RECT_RULE */
+	0x00000000, /* PA_SC_CLIPRECT_0_TL */
+	0x20002000,
+	0x00000000,
+	0x20002000,
+	0x00000000,
+	0x20002000,
+	0x00000000,
+	0x20002000,
+	0x00000000, /* PA_SC_EDGERULE */
+
+	0xc0406900,
+	0x00000094,
+	0x80000000, /* PA_SC_VPORT_SCISSOR_0_TL */
+	0x20002000, /* PA_SC_VPORT_SCISSOR_0_BR */
+	0x80000000, /* PA_SC_VPORT_SCISSOR_1_TL */
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x00000000, /* PA_SC_VPORT_ZMIN_0 */
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+
+	0xc0026900,
+	0x00000292,
+	0x00000000, /* PA_SC_MPASS_PS_CNTL */
+	0x00004010, /* PA_SC_MODE_CNTL */
+
+	0xc0096900,
+	0x00000300,
+	0x00000000, /* PA_SC_LINE_CNTL */
+	0x00000000, /* PA_SC_AA_CONFIG */
+	0x0000002d, /* PA_SU_VTX_CNTL */
+	0x3f800000, /* PA_CL_GB_VERT_CLIP_ADJ */
+	0x3f800000,
+	0x3f800000,
+	0x3f800000,
+	0x00000000, /* PA_SC_SAMPLE_LOCS_MCTX */
+	0x00000000,
+
+	0xc0016900,
+	0x00000312,
+	0xffffffff, /* PA_SC_AA_MASK */
+
+	0xc0066900,
+	0x0000037e,
+	0x00000000, /* PA_SU_POLY_OFFSET_DB_FMT_CNTL */
+	0x00000000, /* PA_SU_POLY_OFFSET_CLAMP */
+	0x00000000, /* PA_SU_POLY_OFFSET_FRONT_SCALE */
+	0x00000000, /* PA_SU_POLY_OFFSET_FRONT_OFFSET */
+	0x00000000, /* PA_SU_POLY_OFFSET_BACK_SCALE */
+	0x00000000, /* PA_SU_POLY_OFFSET_BACK_OFFSET */
+
+	0xc0046900,
+	0x000001b6,
+	0x00000000, /* SPI_INPUT_Z */
+	0x00000000, /* SPI_FOG_CNTL */
+	0x00000000, /* SPI_FOG_FUNC_SCALE */
+	0x00000000, /* SPI_FOG_FUNC_BIAS */
+
+	0xc0016900,
+	0x00000225,
+	0x00000000, /* SQ_PGM_START_FS */
+
+	0xc0016900,
+	0x00000229,
+	0x00000000, /* SQ_PGM_RESOURCES_FS */
+
+	0xc0016900,
+	0x00000237,
+	0x00000000, /* SQ_PGM_CF_OFFSET_FS */
+
+	0xc0026900,
+	0x000002a8,
+	0x00000000, /* VGT_INSTANCE_STEP_RATE_0 */
+	0x00000000, /* VGT_INSTANCE_STEP_RATE_1 */
+
+	0xc0116900,
+	0x00000280,
+	0x00000000, /* PA_SU_POINT_SIZE */
+	0x00000000, /* PA_SU_POINT_MINMAX */
+	0x00000008, /* PA_SU_LINE_CNTL */
+	0x00000000, /* PA_SC_LINE_STIPPLE */
+	0x00000000, /* VGT_OUTPUT_PATH_CNTL */
+	0x00000000, /* VGT_HOS_CNTL */
+	0x00000000, /* VGT_HOS_MAX_TESS_LEVEL */
+	0x00000000, /* VGT_HOS_MIN_TESS_LEVEL */
+	0x00000000, /* VGT_HOS_REUSE_DEPTH */
+	0x00000000, /* VGT_GROUP_PRIM_TYPE */
+	0x00000000, /* VGT_GROUP_FIRST_DECR */
+	0x00000000, /* VGT_GROUP_DECR */
+	0x00000000, /* VGT_GROUP_VECT_0_CNTL */
+	0x00000000, /* VGT_GROUP_VECT_1_CNTL */
+	0x00000000, /* VGT_GROUP_VECT_0_FMT_CNTL */
+	0x00000000, /* VGT_GROUP_VECT_1_FMT_CNTL */
+	0x00000000, /* VGT_GS_MODE */
+
+	0xc0016900,
+	0x000002a1,
+	0x00000000, /* VGT_PRIMITIVEID_EN */
+
+	0xc0016900,
+	0x000002a5,
+	0x00000000, /* VGT_MULTI_PRIM_ID_RESET_EN */
+
+	0xc0036900,
+	0x000002ac,
+	0x00000000, /* VGT_STRMOUT_EN */
+	0x00000000, /* VGT_REUSE_OFF */
+	0x00000000, /* VGT_VTX_CNT_EN */
+
+	0xc0016900,
+	0x000002c8,
+	0x00000000, /* VGT_STRMOUT_BUFFER_EN */
+
+	0xc0076900,
+	0x00000202,
+	0x00cc0000, /* CB_COLOR_CONTROL */
+	0x00000210, /* DB_SHADER_CNTL */
+	0x00010000, /* PA_CL_CLIP_CNTL */
+	0x00000244, /* PA_SU_SC_MODE_CNTL */
+	0x00000100, /* PA_CL_VTE_CNTL */
+	0x00000000, /* PA_CL_VS_OUT_CNTL */
+	0x00000000, /* PA_CL_NANINF_CNTL */
+
+	0xc0026900,
+	0x0000008e,
+	0x0000000f, /* CB_TARGET_MASK */
+	0x0000000f, /* CB_SHADER_MASK */
+
+	0xc0016900,
+	0x000001e8,
+	0x00000001, /* CB_SHADER_CONTROL */
+
+	0xc0016900,
+	0x00000185,
+	0x00000000, /* SPI_VS_OUT_ID_0 */
+
+	0xc0016900,
+	0x00000191,
+	0x00000b00, /* SPI_PS_INPUT_CNTL_0 */
+
+	0xc0056900,
+	0x000001b1,
+	0x00000000, /* SPI_VS_OUT_CONFIG */
+	0x00000000, /* SPI_THREAD_GROUPING */
+	0x00000001, /* SPI_PS_IN_CONTROL_0 */
+	0x00000000, /* SPI_PS_IN_CONTROL_1 */
+	0x00000000, /* SPI_INTERP_CONTROL_0 */
+
+	0xc0036e00, /* SET_SAMPLER */
+	0x00000000,
+	0x00000012,
+	0x00000000,
+	0x00000000,
+};
+
+const u32 r7xx_default_state[] =
+{
+	0xc0012800, /* CONTEXT_CONTROL */
+	0x80000000,
+	0x80000000,
+
+	0xc0016800,
+	0x00000010,
+	0x00008000, /* WAIT_UNTIL */
+
+	0xc0016800,
+	0x00000542,
+	0x07000002, /* TA_CNTL_AUX */
+
+	0xc0016800,
+	0x000005c5,
+	0x00000000, /* VC_ENHANCE */
+
+	0xc0016800,
+	0x00000363,
+	0x00004000, /* SQ_DYN_GPR_CNTL_PS_FLUSH_REQ */
+
+	0xc0016800,
+	0x0000060c,
+	0x00000000, /* DB_DEBUG */
+
+	0xc0016800,
+	0x0000060e,
+	0x00420204, /* DB_WATERMARKS */
+
+	0xc0026f00,
+	0x00000000,
+	0x00000000, /* SQ_VTX_BASE_VTX_LOC */
+	0x00000000, /* SQ_VTX_START_INST_LOC */
+
+	0xc0096900,
+	0x0000022a,
+	0x00000000, /* SQ_ESGS_RING_ITEMSIZE */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+
+	0xc0016900,
+	0x00000004,
+	0x00000000, /* DB_DEPTH_INFO */
+
+	0xc0026900,
+	0x0000000a,
+	0x00000000, /* DB_STENCIL_CLEAR */
+	0x00000000, /* DB_DEPTH_CLEAR */
+
+	0xc0016900,
+	0x00000200,
+	0x00000000, /* DB_DEPTH_CONTROL */
+
+	0xc0026900,
+	0x00000343,
+	0x00000060, /* DB_RENDER_CONTROL */
+	0x00000000, /* DB_RENDER_OVERRIDE */
+
+	0xc0016900,
+	0x00000351,
+	0x0000aa00, /* DB_ALPHA_TO_MASK */
+
+	0xc0096900,
+	0x00000100,
+	0x00000800, /* VGT_MAX_VTX_INDX */
+	0x00000000, /* VGT_MIN_VTX_INDX */
+	0x00000000, /* VGT_INDX_OFFSET */
+	0x00000000, /* VGT_MULTI_PRIM_IB_RESET_INDX */
+	0x00000000, /* SX_ALPHA_TEST_CONTROL */
+	0x00000000, /* CB_BLEND_RED */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+
+	0xc0036900,
+	0x0000010c,
+	0x00000000, /* DB_STENCILREFMASK */
+	0x00000000, /* DB_STENCILREFMASK_BF */
+	0x00000000, /* SX_ALPHA_REF */
+
+	0xc0046900,
+	0x0000030c, /* CB_CLRCMP_CNTL */
+	0x01000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+
+	0xc0016900,
+	0x00000080,
+	0x00000000, /* PA_SC_WINDOW_OFFSET */
+
+	0xc00a6900,
+	0x00000083,
+	0x0000ffff, /* PA_SC_CLIP_RECT_RULE */
+	0x00000000, /* PA_SC_CLIPRECT_0_TL */
+	0x20002000,
+	0x00000000,
+	0x20002000,
+	0x00000000,
+	0x20002000,
+	0x00000000,
+	0x20002000,
+	0xaaaaaaaa, /* PA_SC_EDGERULE */
+
+	0xc0406900,
+	0x00000094,
+	0x80000000, /* PA_SC_VPORT_SCISSOR_0_TL */
+	0x20002000, /* PA_SC_VPORT_SCISSOR_0_BR */
+	0x80000000, /* PA_SC_VPORT_SCISSOR_1_TL */
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x00000000, /* PA_SC_VPORT_ZMIN_0 */
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+	0x00000000,
+	0x3f800000,
+
+	0xc0026900,
+	0x00000292,
+	0x00000000, /* PA_SC_MPASS_PS_CNTL */
+	0x00514000, /* PA_SC_MODE_CNTL */
+
+	0xc0096900,
+	0x00000300,
+	0x00000000, /* PA_SC_LINE_CNTL */
+	0x00000000, /* PA_SC_AA_CONFIG */
+	0x0000002d, /* PA_SU_VTX_CNTL */
+	0x3f800000, /* PA_CL_GB_VERT_CLIP_ADJ */
+	0x3f800000,
+	0x3f800000,
+	0x3f800000,
+	0x00000000, /* PA_SC_SAMPLE_LOCS_MCTX */
+	0x00000000,
+
+	0xc0016900,
+	0x00000312,
+	0xffffffff, /* PA_SC_AA_MASK */
+
+	0xc0066900,
+	0x0000037e,
+	0x00000000, /* PA_SU_POLY_OFFSET_DB_FMT_CNTL */
+	0x00000000, /* PA_SU_POLY_OFFSET_CLAMP */
+	0x00000000, /* PA_SU_POLY_OFFSET_FRONT_SCALE */
+	0x00000000, /* PA_SU_POLY_OFFSET_FRONT_OFFSET */
+	0x00000000, /* PA_SU_POLY_OFFSET_BACK_SCALE */
+	0x00000000, /* PA_SU_POLY_OFFSET_BACK_OFFSET */
+
+	0xc0046900,
+	0x000001b6,
+	0x00000000, /* SPI_INPUT_Z */
+	0x00000000, /* SPI_FOG_CNTL */
+	0x00000000, /* SPI_FOG_FUNC_SCALE */
+	0x00000000, /* SPI_FOG_FUNC_BIAS */
+
+	0xc0016900,
+	0x00000225,
+	0x00000000, /* SQ_PGM_START_FS */
+
+	0xc0016900,
+	0x00000229,
+	0x00000000, /* SQ_PGM_RESOURCES_FS */
+
+	0xc0016900,
+	0x00000237,
+	0x00000000, /* SQ_PGM_CF_OFFSET_FS */
+
+	0xc0026900,
+	0x000002a8,
+	0x00000000, /* VGT_INSTANCE_STEP_RATE_0 */
+	0x00000000, /* VGT_INSTANCE_STEP_RATE_1 */
+
+	0xc0116900,
+	0x00000280,
+	0x00000000, /* PA_SU_POINT_SIZE */
+	0x00000000, /* PA_SU_POINT_MINMAX */
+	0x00000008, /* PA_SU_LINE_CNTL */
+	0x00000000, /* PA_SC_LINE_STIPPLE */
+	0x00000000, /* VGT_OUTPUT_PATH_CNTL */
+	0x00000000, /* VGT_HOS_CNTL */
+	0x00000000, /* VGT_HOS_MAX_TESS_LEVEL */
+	0x00000000, /* VGT_HOS_MIN_TESS_LEVEL */
+	0x00000000, /* VGT_HOS_REUSE_DEPTH */
+	0x00000000, /* VGT_GROUP_PRIM_TYPE */
+	0x00000000, /* VGT_GROUP_FIRST_DECR */
+	0x00000000, /* VGT_GROUP_DECR */
+	0x00000000, /* VGT_GROUP_VECT_0_CNTL */
+	0x00000000, /* VGT_GROUP_VECT_1_CNTL */
+	0x00000000, /* VGT_GROUP_VECT_0_FMT_CNTL */
+	0x00000000, /* VGT_GROUP_VECT_1_FMT_CNTL */
+	0x00000000, /* VGT_GS_MODE */
+
+	0xc0016900,
+	0x000002a1,
+	0x00000000, /* VGT_PRIMITIVEID_EN */
+
+	0xc0016900,
+	0x000002a5,
+	0x00000000, /* VGT_MULTI_PRIM_ID_RESET_EN */
+
+	0xc0036900,
+	0x000002ac,
+	0x00000000, /* VGT_STRMOUT_EN */
+	0x00000000, /* VGT_REUSE_OFF */
+	0x00000000, /* VGT_VTX_CNT_EN */
+
+	0xc0016900,
+	0x000002c8,
+	0x00000000, /* VGT_STRMOUT_BUFFER_EN */
+
+	0xc0076900,
+	0x00000202,
+	0x00cc0000, /* CB_COLOR_CONTROL */
+	0x00000210, /* DB_SHADER_CNTL */
+	0x00010000, /* PA_CL_CLIP_CNTL */
+	0x00000244, /* PA_SU_SC_MODE_CNTL */
+	0x00000100, /* PA_CL_VTE_CNTL */
+	0x00000000, /* PA_CL_VS_OUT_CNTL */
+	0x00000000, /* PA_CL_NANINF_CNTL */
+
+	0xc0026900,
+	0x0000008e,
+	0x0000000f, /* CB_TARGET_MASK */
+	0x0000000f, /* CB_SHADER_MASK */
+
+	0xc0016900,
+	0x000001e8,
+	0x00000001, /* CB_SHADER_CONTROL */
+
+	0xc0016900,
+	0x00000185,
+	0x00000000, /* SPI_VS_OUT_ID_0 */
+
+	0xc0016900,
+	0x00000191,
+	0x00000b00, /* SPI_PS_INPUT_CNTL_0 */
+
+	0xc0056900,
+	0x000001b1,
+	0x00000000, /* SPI_VS_OUT_CONFIG */
+	0x00000001, /* SPI_THREAD_GROUPING */
+	0x00000001, /* SPI_PS_IN_CONTROL_0 */
+	0x00000000, /* SPI_PS_IN_CONTROL_1 */
+	0x00000000, /* SPI_INTERP_CONTROL_0 */
+
+	0xc0036e00, /* SET_SAMPLER */
+	0x00000000,
+	0x00000012,
+	0x00000000,
+	0x00000000,
+};
+
+/* same for r6xx/r7xx */
+const u32 r6xx_vs[] =
+{
+	0x00000004,
+	0x81000000,
+	0x0000203c,
+	0x94000b08,
+	0x00004000,
+	0x14200b1a,
+	0x00000000,
+	0x00000000,
+	0x3c000000,
+	0x68cd1000,
+#ifdef __BIG_ENDIAN
+	0x000a0000,
+#else
+	0x00080000,
+#endif
+	0x00000000,
+};
+
+const u32 r6xx_ps[] =
+{
+	0x00000002,
+	0x80800000,
+	0x00000000,
+	0x94200688,
+	0x00000010,
+	0x000d1000,
+	0xb0800000,
+	0x00000000,
+};
+
+const u32 r6xx_ps_size = ARRAY_SIZE(r6xx_ps);
+const u32 r6xx_vs_size = ARRAY_SIZE(r6xx_vs);
+const u32 r6xx_default_size = ARRAY_SIZE(r6xx_default_state);
+const u32 r7xx_default_size = ARRAY_SIZE(r7xx_default_state);

drivers/video/drm/radeon/r600_blit_shaders.h

@@ -0,0 +1,38 @@
+/*
+ * Copyright 2009 Advanced Micro Devices, Inc.
+ * Copyright 2009 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 (including the next
+ * paragraph) 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) AND/OR ITS SUPPLIERS 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.
+ *
+ */
+
+#ifndef R600_BLIT_SHADERS_H
+#define R600_BLIT_SHADERS_H
+
+extern const u32 r6xx_ps[];
+extern const u32 r6xx_vs[];
+extern const u32 r7xx_default_state[];
+extern const u32 r6xx_default_state[];
+
+
+extern const u32 r6xx_ps_size, r6xx_vs_size;
+extern const u32 r6xx_default_size, r7xx_default_size;
+
+#endif

drivers/video/drm/radeon/radeon.h

@@ -295,6 +295,7 @@ struct radeon_fence {
 	uint32_t			seq;
 	bool				emited;
 	bool				signaled;
+    evhandle_t              evnt;
 };
 
 int radeon_fence_driver_init(struct radeon_device *rdev);

drivers/video/drm/radeon/radeon_asic.c

@@ -134,21 +134,22 @@ static struct radeon_asic r100_asic = {
 //	.suspend = &r100_suspend,
 //	.resume = &r100_resume,
 //	.vga_set_state = &r100_vga_set_state,
+	.gpu_is_lockup = &r100_gpu_is_lockup,
 	.asic_reset = &r100_asic_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,
+	.ring_ib_execute = &r100_ring_ib_execute,
-//	.irq_set = &r100_irq_set,
+	.irq_set = &r100_irq_set,
-//	.irq_process = &r100_irq_process,
+	.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_blit = &r100_copy_blit,
-//	.copy_dma = NULL,
+	.copy_dma = NULL,
-//	.copy = &r100_copy_blit,
+	.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,
@@ -172,21 +173,22 @@ static struct radeon_asic r200_asic = {
 //	.suspend = &r100_suspend,
 //	.resume = &r100_resume,
 //	.vga_set_state = &r100_vga_set_state,
+	.gpu_is_lockup = &r100_gpu_is_lockup,
 	.asic_reset = &r100_asic_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,
+	.ring_ib_execute = &r100_ring_ib_execute,
-//	.irq_set = &r100_irq_set,
+	.irq_set = &r100_irq_set,
-//	.irq_process = &r100_irq_process,
+	.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_blit = &r100_copy_blit,
-//	.copy_dma = NULL,
+	.copy_dma = &r200_copy_dma,
-//	.copy = &r100_copy_blit,
+	.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,
@@ -215,15 +217,15 @@ static struct radeon_asic r300_asic = {
 	.cp_commit = &r100_cp_commit,
 	.ring_start = &r300_ring_start,
 	.ring_test = &r100_ring_test,
-//	.ring_ib_execute = &r100_ring_ib_execute,
+	.ring_ib_execute = &r100_ring_ib_execute,
-//	.irq_set = &r100_irq_set,
+	.irq_set = &r100_irq_set,
-//	.irq_process = &r100_irq_process,
+	.irq_process = &r100_irq_process,
 //	.get_vblank_counter = &r100_get_vblank_counter,
 	.fence_ring_emit = &r300_fence_ring_emit,
 //	.cs_parse = &r300_cs_parse,
-//	.copy_blit = &r100_copy_blit,
+	.copy_blit = &r100_copy_blit,
-//	.copy_dma = &r300_copy_dma,
+	.copy_dma = &r200_copy_dma,
-//	.copy = &r100_copy_blit,
+	.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,
@@ -253,15 +255,15 @@ static struct radeon_asic r300_asic_pcie = {
 	.cp_commit = &r100_cp_commit,
 	.ring_start = &r300_ring_start,
 	.ring_test = &r100_ring_test,
-//	.ring_ib_execute = &r100_ring_ib_execute,
+	.ring_ib_execute = &r100_ring_ib_execute,
-//	.irq_set = &r100_irq_set,
+	.irq_set = &r100_irq_set,
-//	.irq_process = &r100_irq_process,
+	.irq_process = &r100_irq_process,
 //	.get_vblank_counter = &r100_get_vblank_counter,
 	.fence_ring_emit = &r300_fence_ring_emit,
 //	.cs_parse = &r300_cs_parse,
-//	.copy_blit = &r100_copy_blit,
+	.copy_blit = &r100_copy_blit,
-//	.copy_dma = &r300_copy_dma,
+	.copy_dma = &r200_copy_dma,
-//	.copy = &r100_copy_blit,
+	.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,
@@ -290,15 +292,15 @@ static struct radeon_asic r420_asic = {
 	.cp_commit = &r100_cp_commit,
 	.ring_start = &r300_ring_start,
 	.ring_test = &r100_ring_test,
-//	.ring_ib_execute = &r100_ring_ib_execute,
+	.ring_ib_execute = &r100_ring_ib_execute,
-//	.irq_set = &r100_irq_set,
+	.irq_set = &r100_irq_set,
-//	.irq_process = &r100_irq_process,
+	.irq_process = &r100_irq_process,
 //	.get_vblank_counter = &r100_get_vblank_counter,
 	.fence_ring_emit = &r300_fence_ring_emit,
 //	.cs_parse = &r300_cs_parse,
-//	.copy_blit = &r100_copy_blit,
+	.copy_blit = &r100_copy_blit,
-//	.copy_dma = &r300_copy_dma,
+	.copy_dma = &r200_copy_dma,
-//	.copy = &r100_copy_blit,
+	.copy = &r100_copy_blit,
 	.get_engine_clock = &radeon_atom_get_engine_clock,
 	.set_engine_clock = &radeon_atom_set_engine_clock,
 	.get_memory_clock = &radeon_atom_get_memory_clock,
@@ -328,15 +330,15 @@ static struct radeon_asic rs400_asic = {
 	.cp_commit = &r100_cp_commit,
 	.ring_start = &r300_ring_start,
 	.ring_test = &r100_ring_test,
-//	.ring_ib_execute = &r100_ring_ib_execute,
+	.ring_ib_execute = &r100_ring_ib_execute,
-//	.irq_set = &r100_irq_set,
+	.irq_set = &r100_irq_set,
-//	.irq_process = &r100_irq_process,
+	.irq_process = &r100_irq_process,
 //	.get_vblank_counter = &r100_get_vblank_counter,
 	.fence_ring_emit = &r300_fence_ring_emit,
 //	.cs_parse = &r300_cs_parse,
-//	.copy_blit = &r100_copy_blit,
+	.copy_blit = &r100_copy_blit,
-//	.copy_dma = &r300_copy_dma,
+	.copy_dma = &r200_copy_dma,
-//	.copy = &r100_copy_blit,
+	.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,
@@ -366,15 +368,15 @@ static struct radeon_asic rs600_asic = {
 	.cp_commit = &r100_cp_commit,
 	.ring_start = &r300_ring_start,
 	.ring_test = &r100_ring_test,
-//	.ring_ib_execute = &r100_ring_ib_execute,
+	.ring_ib_execute = &r100_ring_ib_execute,
-//	.irq_set = &rs600_irq_set,
+	.irq_set = &rs600_irq_set,
-//	.irq_process = &rs600_irq_process,
+	.irq_process = &rs600_irq_process,
 //	.get_vblank_counter = &rs600_get_vblank_counter,
 	.fence_ring_emit = &r300_fence_ring_emit,
 //   .cs_parse = &r300_cs_parse,
-//   .copy_blit = &r100_copy_blit,
+    .copy_blit = &r100_copy_blit,
-//   .copy_dma = &r300_copy_dma,
+	.copy_dma = &r200_copy_dma,
-//   .copy = &r100_copy_blit,
+    .copy = &r100_copy_blit,
 	.get_engine_clock = &radeon_atom_get_engine_clock,
 	.set_engine_clock = &radeon_atom_set_engine_clock,
 	.get_memory_clock = &radeon_atom_get_memory_clock,
@@ -404,15 +406,15 @@ static struct radeon_asic rs690_asic = {
 	.cp_commit = &r100_cp_commit,
 	.ring_start = &r300_ring_start,
 	.ring_test = &r100_ring_test,
-//	.ring_ib_execute = &r100_ring_ib_execute,
+	.ring_ib_execute = &r100_ring_ib_execute,
-//	.irq_set = &rs600_irq_set,
+	.irq_set = &rs600_irq_set,
-//	.irq_process = &rs600_irq_process,
+	.irq_process = &rs600_irq_process,
 //	.get_vblank_counter = &rs600_get_vblank_counter,
 	.fence_ring_emit = &r300_fence_ring_emit,
 //	.cs_parse = &r300_cs_parse,
-//	.copy_blit = &r100_copy_blit,
+	.copy_blit = &r100_copy_blit,
-//	.copy_dma = &r300_copy_dma,
+	.copy_dma = &r200_copy_dma,
-//	.copy = &r300_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,
@@ -442,15 +444,15 @@ static struct radeon_asic rv515_asic = {
 	.cp_commit = &r100_cp_commit,
 	.ring_start = &rv515_ring_start,
 	.ring_test = &r100_ring_test,
-//	.ring_ib_execute = &r100_ring_ib_execute,
+	.ring_ib_execute = &r100_ring_ib_execute,
-//	.irq_set = &rs600_irq_set,
+	.irq_set = &rs600_irq_set,
-//	.irq_process = &rs600_irq_process,
+	.irq_process = &rs600_irq_process,
 //	.get_vblank_counter = &rs600_get_vblank_counter,
 	.fence_ring_emit = &r300_fence_ring_emit,
 //	.cs_parse = &r300_cs_parse,
-//	.copy_blit = &r100_copy_blit,
+	.copy_blit = &r100_copy_blit,
-//	.copy_dma = &r300_copy_dma,
+	.copy_dma = &r200_copy_dma,
-//	.copy = &r100_copy_blit,
+	.copy = &r100_copy_blit,
 	.get_engine_clock = &radeon_atom_get_engine_clock,
 	.set_engine_clock = &radeon_atom_set_engine_clock,
 	.get_memory_clock = &radeon_atom_get_memory_clock,
@@ -480,15 +482,15 @@ static struct radeon_asic r520_asic = {
 	.cp_commit = &r100_cp_commit,
 	.ring_start = &rv515_ring_start,
 	.ring_test = &r100_ring_test,
-//	.ring_ib_execute = &r100_ring_ib_execute,
+	.ring_ib_execute = &r100_ring_ib_execute,
-//	.irq_set = &rs600_irq_set,
+	.irq_set = &rs600_irq_set,
-//	.irq_process = &rs600_irq_process,
+	.irq_process = &rs600_irq_process,
 //	.get_vblank_counter = &rs600_get_vblank_counter,
 	.fence_ring_emit = &r300_fence_ring_emit,
 //	.cs_parse = &r300_cs_parse,
-//	.copy_blit = &r100_copy_blit,
+	.copy_blit = &r100_copy_blit,
-//	.copy_dma = &r300_copy_dma,
+	.copy_dma = &r200_copy_dma,
-//	.copy = &r100_copy_blit,
+	.copy = &r100_copy_blit,
 	.get_engine_clock = &radeon_atom_get_engine_clock,
 	.set_engine_clock = &radeon_atom_set_engine_clock,
 	.get_memory_clock = &radeon_atom_get_memory_clock,
@@ -517,20 +519,20 @@ static struct radeon_asic r600_asic = {
 	.gart_tlb_flush = &r600_pcie_gart_tlb_flush,
 	.gart_set_page = &rs600_gart_set_page,
 	.ring_test = &r600_ring_test,
-//	.ring_ib_execute = &r600_ring_ib_execute,
+	.ring_ib_execute = &r600_ring_ib_execute,
 	.irq_set = &r600_irq_set,
 	.irq_process = &r600_irq_process,
 	.fence_ring_emit = &r600_fence_ring_emit,
 //	.cs_parse = &r600_cs_parse,
-//	.copy_blit = &r600_copy_blit,
+	.copy_blit = &r600_copy_blit,
-//	.copy_dma = &r600_copy_blit,
+	.copy_dma = &r600_copy_blit,
-//	.copy = &r600_copy_blit,
+	.copy = &r600_copy_blit,
 	.get_engine_clock = &radeon_atom_get_engine_clock,
 	.set_engine_clock = &radeon_atom_set_engine_clock,
 	.get_memory_clock = &radeon_atom_get_memory_clock,
 	.set_memory_clock = &radeon_atom_set_memory_clock,
-	.get_pcie_lanes = &rv370_get_pcie_lanes,
+	.get_pcie_lanes = &r600_get_pcie_lanes,
-	.set_pcie_lanes = NULL,
+	.set_pcie_lanes = &r600_set_pcie_lanes,
 	.set_clock_gating = NULL,
 	.set_surface_reg = r600_set_surface_reg,
 	.clear_surface_reg = r600_clear_surface_reg,
@@ -554,14 +556,14 @@ static struct radeon_asic rs780_asic = {
 	.gart_tlb_flush = &r600_pcie_gart_tlb_flush,
 	.gart_set_page = &rs600_gart_set_page,
 	.ring_test = &r600_ring_test,
-//	.ring_ib_execute = &r600_ring_ib_execute,
+	.ring_ib_execute = &r600_ring_ib_execute,
 	.irq_set = &r600_irq_set,
 	.irq_process = &r600_irq_process,
 	.fence_ring_emit = &r600_fence_ring_emit,
 //	.cs_parse = &r600_cs_parse,
-//	.copy_blit = &r600_copy_blit,
+	.copy_blit = &r600_copy_blit,
-//	.copy_dma = &r600_copy_blit,
+	.copy_dma = &r600_copy_blit,
-//	.copy = &r600_copy_blit,
+	.copy = &r600_copy_blit,
 	.get_engine_clock = &radeon_atom_get_engine_clock,
 	.set_engine_clock = &radeon_atom_set_engine_clock,
 	.get_memory_clock = NULL,
@@ -594,15 +596,15 @@ static struct radeon_asic rv770_asic = {
 	.irq_process = &r600_irq_process,
 	.fence_ring_emit = &r600_fence_ring_emit,
 //	.cs_parse = &r600_cs_parse,
-//	.copy_blit = &r600_copy_blit,
+	.copy_blit = &r600_copy_blit,
-//	.copy_dma = &r600_copy_blit,
+	.copy_dma = &r600_copy_blit,
-//	.copy = &r600_copy_blit,
+	.copy = &r600_copy_blit,
 	.get_engine_clock = &radeon_atom_get_engine_clock,
 	.set_engine_clock = &radeon_atom_set_engine_clock,
 	.get_memory_clock = &radeon_atom_get_memory_clock,
 	.set_memory_clock = &radeon_atom_set_memory_clock,
-	.get_pcie_lanes = &rv370_get_pcie_lanes,
+	.get_pcie_lanes = &r600_get_pcie_lanes,
-	.set_pcie_lanes = NULL,
+	.set_pcie_lanes = &r600_set_pcie_lanes,
 	.set_clock_gating = &radeon_atom_set_clock_gating,
 	.set_surface_reg = r600_set_surface_reg,
 	.clear_surface_reg = r600_clear_surface_reg,
@@ -624,14 +626,14 @@ static struct radeon_asic evergreen_asic = {
 	.gart_tlb_flush = &evergreen_pcie_gart_tlb_flush,
 	.gart_set_page = &rs600_gart_set_page,
 	.ring_test = &r600_ring_test,
-//	.ring_ib_execute = &r600_ring_ib_execute,
+	.ring_ib_execute = &evergreen_ring_ib_execute,
-//	.irq_set = &r600_irq_set,
+	.irq_set = &evergreen_irq_set,
-//	.irq_process = &r600_irq_process,
+	.irq_process = &evergreen_irq_process,
 	.fence_ring_emit = &r600_fence_ring_emit,
-//	.cs_parse = &r600_cs_parse,
+//	.cs_parse = &evergreen_cs_parse,
-//	.copy_blit = &r600_copy_blit,
+	.copy_blit = &evergreen_copy_blit,
-//	.copy_dma = &r600_copy_blit,
+	.copy_dma = &evergreen_copy_blit,
-//	.copy = &r600_copy_blit,
+	.copy = &evergreen_copy_blit,
 	.get_engine_clock = &radeon_atom_get_engine_clock,
 	.set_engine_clock = &radeon_atom_set_engine_clock,
 	.get_memory_clock = &radeon_atom_get_memory_clock,
@@ -642,6 +644,10 @@ static struct radeon_asic evergreen_asic = {
 	.set_surface_reg = r600_set_surface_reg,
 	.clear_surface_reg = r600_clear_surface_reg,
 	.bandwidth_update = &evergreen_bandwidth_update,
+	.hpd_init = &evergreen_hpd_init,
+	.hpd_fini = &evergreen_hpd_fini,
+	.hpd_sense = &evergreen_hpd_sense,
+	.hpd_set_polarity = &evergreen_hpd_set_polarity,
 
 };
 
@@ -656,14 +662,14 @@ static struct radeon_asic sumo_asic = {
 	.gart_tlb_flush = &evergreen_pcie_gart_tlb_flush,
 	.gart_set_page = &rs600_gart_set_page,
 	.ring_test = &r600_ring_test,
-//	.ring_ib_execute = &r600_ring_ib_execute,
+	.ring_ib_execute = &evergreen_ring_ib_execute,
-//	.irq_set = &r600_irq_set,
+	.irq_set = &evergreen_irq_set,
-//	.irq_process = &r600_irq_process,
+	.irq_process = &evergreen_irq_process,
 	.fence_ring_emit = &r600_fence_ring_emit,
 //	.cs_parse = &r600_cs_parse,
-//	.copy_blit = &r600_copy_blit,
+	.copy_blit = &evergreen_copy_blit,
-//	.copy_dma = &r600_copy_blit,
+	.copy_dma = &evergreen_copy_blit,
-//	.copy = &r600_copy_blit,
+	.copy = &evergreen_copy_blit,
 	.get_engine_clock = &radeon_atom_get_engine_clock,
 	.set_engine_clock = &radeon_atom_set_engine_clock,
 	.get_memory_clock = NULL,
@@ -674,9 +680,12 @@ static struct radeon_asic sumo_asic = {
 	.set_surface_reg = r600_set_surface_reg,
 	.clear_surface_reg = r600_clear_surface_reg,
 	.bandwidth_update = &evergreen_bandwidth_update,
+	.hpd_init = &evergreen_hpd_init,
+	.hpd_fini = &evergreen_hpd_fini,
+	.hpd_sense = &evergreen_hpd_sense,
+	.hpd_set_polarity = &evergreen_hpd_set_polarity,
 };
 
-
 static struct radeon_asic btc_asic = {
 	.init = &evergreen_init,
 //	.fini = &evergreen_fini,
@@ -688,14 +697,14 @@ static struct radeon_asic btc_asic = {
 	.gart_tlb_flush = &evergreen_pcie_gart_tlb_flush,
 	.gart_set_page = &rs600_gart_set_page,
 	.ring_test = &r600_ring_test,
-//	.ring_ib_execute = &r600_ring_ib_execute,
+	.ring_ib_execute = &evergreen_ring_ib_execute,
-//	.irq_set = &r600_irq_set,
+	.irq_set = &evergreen_irq_set,
-//	.irq_process = &r600_irq_process,
+	.irq_process = &evergreen_irq_process,
 	.fence_ring_emit = &r600_fence_ring_emit,
-//	.cs_parse = &r600_cs_parse,
+//	.cs_parse = &evergreen_cs_parse,
-//	.copy_blit = &r600_copy_blit,
+	.copy_blit = &evergreen_copy_blit,
-//	.copy_dma = &r600_copy_blit,
+	.copy_dma = &evergreen_copy_blit,
-//	.copy = &r600_copy_blit,
+	.copy = &evergreen_copy_blit,
 	.get_engine_clock = &radeon_atom_get_engine_clock,
 	.set_engine_clock = &radeon_atom_set_engine_clock,
 	.get_memory_clock = &radeon_atom_get_memory_clock,
@@ -707,7 +716,9 @@ static struct radeon_asic btc_asic = {
 	.clear_surface_reg = r600_clear_surface_reg,
 	.bandwidth_update = &evergreen_bandwidth_update,
 	.hpd_init = &evergreen_hpd_init,
+	.hpd_fini = &evergreen_hpd_fini,
 	.hpd_sense = &evergreen_hpd_sense,
+	.hpd_set_polarity = &evergreen_hpd_set_polarity,
 };
 
 static struct radeon_asic cayman_asic = {
@@ -721,14 +732,14 @@ static struct radeon_asic cayman_asic = {
 	.gart_tlb_flush = &cayman_pcie_gart_tlb_flush,
 	.gart_set_page = &rs600_gart_set_page,
 	.ring_test = &r600_ring_test,
-//	.ring_ib_execute = &r600_ring_ib_execute,
+	.ring_ib_execute = &evergreen_ring_ib_execute,
-//	.irq_set = &r600_irq_set,
+	.irq_set = &evergreen_irq_set,
-//	.irq_process = &r600_irq_process,
+	.irq_process = &evergreen_irq_process,
 	.fence_ring_emit = &r600_fence_ring_emit,
-//	.cs_parse = &r600_cs_parse,
+//	.cs_parse = &evergreen_cs_parse,
-//	.copy_blit = &r600_copy_blit,
+	.copy_blit = &evergreen_copy_blit,
-//	.copy_dma = &r600_copy_blit,
+	.copy_dma = &evergreen_copy_blit,
-//	.copy = &r600_copy_blit,
+	.copy = &evergreen_copy_blit,
 	.get_engine_clock = &radeon_atom_get_engine_clock,
 	.set_engine_clock = &radeon_atom_set_engine_clock,
 	.get_memory_clock = &radeon_atom_get_memory_clock,
@@ -739,6 +750,10 @@ static struct radeon_asic cayman_asic = {
 	.set_surface_reg = r600_set_surface_reg,
 	.clear_surface_reg = r600_clear_surface_reg,
 	.bandwidth_update = &evergreen_bandwidth_update,
+	.hpd_init = &evergreen_hpd_init,
+	.hpd_fini = &evergreen_hpd_fini,
+	.hpd_sense = &evergreen_hpd_sense,
+	.hpd_set_polarity = &evergreen_hpd_set_polarity,
 };
 
 int radeon_asic_init(struct radeon_device *rdev)

drivers/video/drm/radeon/radeon_benchmark.c

@@ -0,0 +1,185 @@
+/*
+ * Copyright 2009 Jerome Glisse.
+ *
+ * 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: Jerome Glisse
+ */
+#include <drm/drmP.h>
+#include <drm/radeon_drm.h>
+#include "radeon_reg.h"
+#include "radeon.h"
+
+unsigned int inline jiffies_to_msecs(const unsigned long j)
+{
+    return (10 * j);
+};
+
+
+void radeon_benchmark_move(struct radeon_device *rdev, unsigned bsize,
+			   unsigned sdomain, unsigned ddomain)
+{
+	struct radeon_bo *dobj = NULL;
+	struct radeon_bo *sobj = NULL;
+	struct radeon_fence *fence = NULL;
+	uint64_t saddr, daddr;
+	unsigned long start_jiffies;
+	unsigned long end_jiffies;
+	unsigned long time;
+	unsigned i, n, size;
+	int r;
+
+    ENTER();
+
+	size = bsize;
+    n = 2; //1024;
+
+    dbgprintf("source domain %x\n", sdomain);
+
+	r = radeon_bo_create(rdev, size, PAGE_SIZE, true, sdomain, &sobj);
+	if (r) {
+		goto out_cleanup;
+	}
+	r = radeon_bo_reserve(sobj, false);
+	if (unlikely(r != 0))
+		goto out_cleanup;
+	r = radeon_bo_pin(sobj, sdomain, &saddr);
+//   radeon_bo_unreserve(sobj);
+	if (r) {
+		goto out_cleanup;
+	}
+
+    dbgprintf("destination domain %x\n", ddomain);
+
+	r = radeon_bo_create(rdev, size, PAGE_SIZE, true, ddomain, &dobj);
+	if (r) {
+		goto out_cleanup;
+	}
+	r = radeon_bo_reserve(dobj, false);
+	if (unlikely(r != 0))
+		goto out_cleanup;
+	r = radeon_bo_pin(dobj, ddomain, &daddr);
+//   radeon_bo_unreserve(dobj);
+	if (r) {
+		goto out_cleanup;
+	}
+    dbgprintf("done\n");
+
+	/* r100 doesn't have dma engine so skip the test */
+	if (rdev->asic->copy_dma) {
+
+        dbgprintf("copy dma\n");
+
+        start_jiffies = GetTimerTicks();
+		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);
+		}
+        end_jiffies = GetTimerTicks();
+		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 = GetTimerTicks();
+	for (i = 0; i < n; i++) {
+		r = radeon_fence_create(rdev, &fence);
+		if (r) {
+			goto out_cleanup;
+		}
+		r = radeon_copy_blit(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);
+	}
+    end_jiffies = GetTimerTicks();
+	time = end_jiffies - start_jiffies;
+	time = jiffies_to_msecs(time);
+	if (time > 0) {
+		i = ((n * size) >> 10) / time;
+		printk(KERN_INFO "radeon: blit %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);
+	}
+out_cleanup:
+
+    dbgprintf("cleanup\n");
+
+	if (sobj) {
+		r = radeon_bo_reserve(sobj, false);
+		if (likely(r == 0)) {
+			radeon_bo_unpin(sobj);
+			radeon_bo_unreserve(sobj);
+		}
+		radeon_bo_unref(&sobj);
+	}
+	if (dobj) {
+		r = radeon_bo_reserve(dobj, false);
+		if (likely(r == 0)) {
+			radeon_bo_unpin(dobj);
+			radeon_bo_unreserve(dobj);
+		}
+		radeon_bo_unref(&dobj);
+	}
+	if (fence) {
+		radeon_fence_unref(&fence);
+	}
+	if (r) {
+		printk(KERN_WARNING "Error while benchmarking BO move.\n");
+	}
+
+    LEAVE();
+
+}
+
+void radeon_benchmark(struct radeon_device *rdev)
+{
+   radeon_benchmark_move(rdev, 8192*4096, RADEON_GEM_DOMAIN_GTT,
+                 RADEON_GEM_DOMAIN_VRAM);
+   radeon_benchmark_move(rdev, 8192*4096, RADEON_GEM_DOMAIN_VRAM,
+                 RADEON_GEM_DOMAIN_GTT);
+   radeon_benchmark_move(rdev, 8192*4096, RADEON_GEM_DOMAIN_VRAM,
+                 RADEON_GEM_DOMAIN_VRAM);
+}

drivers/video/drm/radeon/radeon_bios.c

@@ -332,7 +332,7 @@ static bool avivo_read_disabled_bios(struct radeon_device *rdev)
 
 	seprom_cntl1 = RREG32(RADEON_SEPROM_CNTL1);
 	viph_control = RREG32(RADEON_VIPH_CONTROL);
-	bus_cntl = RREG32(RADEON_BUS_CNTL);
+	bus_cntl = RREG32(RV370_BUS_CNTL);
 	d1vga_control = RREG32(AVIVO_D1VGA_CONTROL);
 	d2vga_control = RREG32(AVIVO_D2VGA_CONTROL);
 	vga_render_control = RREG32(AVIVO_VGA_RENDER_CONTROL);
@@ -351,7 +351,7 @@ static bool avivo_read_disabled_bios(struct radeon_device *rdev)
 	WREG32(RADEON_VIPH_CONTROL, (viph_control & ~RADEON_VIPH_EN));
 
 	/* enable the rom */
-	WREG32(RADEON_BUS_CNTL, (bus_cntl & ~RADEON_BUS_BIOS_DIS_ROM));
+	WREG32(RV370_BUS_CNTL, (bus_cntl & ~RV370_BUS_BIOS_DIS_ROM));
 
 	/* Disable VGA mode */
 	WREG32(AVIVO_D1VGA_CONTROL,
@@ -368,7 +368,7 @@ static bool avivo_read_disabled_bios(struct radeon_device *rdev)
 	/* restore regs */
 	WREG32(RADEON_SEPROM_CNTL1, seprom_cntl1);
 	WREG32(RADEON_VIPH_CONTROL, viph_control);
-	WREG32(RADEON_BUS_CNTL, bus_cntl);
+	WREG32(RV370_BUS_CNTL, bus_cntl);
 	WREG32(AVIVO_D1VGA_CONTROL, d1vga_control);
 	WREG32(AVIVO_D2VGA_CONTROL, d2vga_control);
 	WREG32(AVIVO_VGA_RENDER_CONTROL, vga_render_control);
@@ -391,6 +391,9 @@ static bool legacy_read_disabled_bios(struct radeon_device *rdev)
 
 	seprom_cntl1 = RREG32(RADEON_SEPROM_CNTL1);
 	viph_control = RREG32(RADEON_VIPH_CONTROL);
+	if (rdev->flags & RADEON_IS_PCIE)
+		bus_cntl = RREG32(RV370_BUS_CNTL);
+	else
 	bus_cntl = RREG32(RADEON_BUS_CNTL);
 	crtc_gen_cntl = RREG32(RADEON_CRTC_GEN_CNTL);
 	crtc2_gen_cntl = 0;
@@ -413,6 +416,9 @@ static bool legacy_read_disabled_bios(struct radeon_device *rdev)
 	WREG32(RADEON_VIPH_CONTROL, (viph_control & ~RADEON_VIPH_EN));
 
 	/* enable the rom */
+	if (rdev->flags & RADEON_IS_PCIE)
+		WREG32(RV370_BUS_CNTL, (bus_cntl & ~RV370_BUS_BIOS_DIS_ROM));
+	else
 	WREG32(RADEON_BUS_CNTL, (bus_cntl & ~RADEON_BUS_BIOS_DIS_ROM));
 
 	/* Turn off mem requests and CRTC for both controllers */
@@ -440,6 +446,9 @@ static bool legacy_read_disabled_bios(struct radeon_device *rdev)
 	/* restore regs */
 	WREG32(RADEON_SEPROM_CNTL1, seprom_cntl1);
 	WREG32(RADEON_VIPH_CONTROL, viph_control);
+	if (rdev->flags & RADEON_IS_PCIE)
+		WREG32(RV370_BUS_CNTL, bus_cntl);
+	else
 	WREG32(RADEON_BUS_CNTL, bus_cntl);
 	WREG32(RADEON_CRTC_GEN_CNTL, crtc_gen_cntl);
 	if (!(rdev->flags & RADEON_SINGLE_CRTC)) {

drivers/video/drm/radeon/radeon_connectors.c

@@ -52,6 +52,12 @@ void radeon_connector_hotplug(struct drm_connector *connector)
 	struct radeon_device *rdev = dev->dev_private;
 	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
 
+	/* bail if the connector does not have hpd pin, e.g.,
+	 * VGA, TV, etc.
+	 */
+	if (radeon_connector->hpd.hpd == RADEON_HPD_NONE)
+		return;
+
 		radeon_hpd_set_polarity(rdev, radeon_connector->hpd.hpd);
 
 	/* powering up/down the eDP panel generates hpd events which

drivers/video/drm/radeon/radeon_device.c

@@ -716,12 +716,12 @@ int radeon_device_init(struct radeon_device *rdev,
     mutex_init(&rdev->ib_pool.mutex);
     mutex_init(&rdev->cp.mutex);
 	mutex_init(&rdev->dc_hw_i2c_mutex);
-//   if (rdev->family >= CHIP_R600)
+	if (rdev->family >= CHIP_R600)
-//       spin_lock_init(&rdev->ih.lock);
+		spin_lock_init(&rdev->ih.lock);
 	mutex_init(&rdev->gem.mutex);
 	mutex_init(&rdev->pm.mutex);
 	mutex_init(&rdev->vram_mutex);
- //   rwlock_init(&rdev->fence_drv.lock);
+	rwlock_init(&rdev->fence_drv.lock);
 	INIT_LIST_HEAD(&rdev->gem.objects);
 
 	/* Set asic functions */
@@ -793,9 +793,9 @@ int radeon_device_init(struct radeon_device *rdev,
 //	if (radeon_testing) {
 //		radeon_test_moves(rdev);
 //    }
-//	if (radeon_benchmarking) {
+   if (radeon_benchmarking) {
-//		radeon_benchmark(rdev);
+       radeon_benchmark(rdev);
-//    }
+    }
 	return 0;
 }
 
@@ -1071,12 +1071,6 @@ u32_t drvEntry(int action, char *cmdline)
 
     rdev = rdisplay->ddev->dev_private;
 
-//    if( (rdev->asic == &r600_asic) ||
-//        (rdev->asic == &rv770_asic))
-//        r600_2D_test(rdev);
-//    else if (rdev->asic != &evergreen_asic)
-//        r100_2D_test(rdev);
-
     err = RegService("DISPLAY", display_handler);
 
     if( err != 0)

drivers/video/drm/radeon/radeon_fence.c

@@ -101,14 +101,13 @@ static bool radeon_fence_poll_locked(struct radeon_device *rdev)
 	bool wake = false;
 	unsigned long cjiffies;
 
-#if 0
 	seq = radeon_fence_read(rdev);
 	if (seq != rdev->fence_drv.last_seq) {
 		rdev->fence_drv.last_seq = seq;
-		rdev->fence_drv.last_jiffies = jiffies;
+        rdev->fence_drv.last_jiffies = GetTimerTicks();
-//       rdev->fence_drv.last_timeout = RADEON_FENCE_JIFFIES_TIMEOUT;
+        rdev->fence_drv.last_timeout = RADEON_FENCE_JIFFIES_TIMEOUT;
 	} else {
-		cjiffies = jiffies;
+        cjiffies = GetTimerTicks();
 		if (time_after(cjiffies, rdev->fence_drv.last_jiffies)) {
 			cjiffies -= rdev->fence_drv.last_jiffies;
 			if (time_after(rdev->fence_drv.last_timeout, cjiffies)) {
@@ -138,32 +137,23 @@ static bool radeon_fence_poll_locked(struct radeon_device *rdev)
 	}
 	/* all fence previous to this one are considered as signaled */
 	if (n) {
+        kevent_t event;
+        event.code = -1;
 		i = n;
 		do {
 			n = i->prev;
 			list_move_tail(i, &rdev->fence_drv.signaled);
 			fence = list_entry(i, struct radeon_fence, list);
 			fence->signaled = true;
+//            dbgprintf("fence %x done\n", fence);
+            RaiseEvent(fence->evnt, 0, &event);
 			i = n;
 		} while (i != &rdev->fence_drv.emited);
 		wake = true;
 	}
-#endif
 	return wake;
 }
 
-static void radeon_fence_destroy(struct kref *kref)
-{
-	unsigned long irq_flags;
-        struct radeon_fence *fence;
-
-	fence = container_of(kref, struct radeon_fence, kref);
-	write_lock_irqsave(&fence->rdev->fence_drv.lock, irq_flags);
-	list_del(&fence->list);
-	fence->emited = false;
-	write_unlock_irqrestore(&fence->rdev->fence_drv.lock, irq_flags);
-	kfree(fence);
-}
 
 int radeon_fence_create(struct radeon_device *rdev, struct radeon_fence **fence)
 {
@@ -173,6 +163,8 @@ int radeon_fence_create(struct radeon_device *rdev, struct radeon_fence **fence)
 	if ((*fence) == NULL) {
 		return -ENOMEM;
 	}
+
+    (*fence)->evnt = CreateEvent(NULL, MANUAL_DESTROY);
 //	kref_init(&((*fence)->kref));
 	(*fence)->rdev = rdev;
 	(*fence)->emited = false;
@@ -231,8 +223,6 @@ int radeon_fence_wait(struct radeon_fence *fence, bool intr)
 	if (radeon_fence_signaled(fence)) {
 		return 0;
 	}
-
-#if 0
 	timeout = rdev->fence_drv.last_timeout;
 retry:
 	/* save current sequence used to check for GPU lockup */
@@ -240,16 +230,22 @@ retry:
 //   trace_radeon_fence_wait_begin(rdev->ddev, seq);
 	if (intr) {
 		radeon_irq_kms_sw_irq_get(rdev);
-		r = wait_event_interruptible_timeout(rdev->fence_drv.queue,
+//       r = wait_event_interruptible_timeout(rdev->fence_drv.queue,
-				radeon_fence_signaled(fence), timeout);
+//               radeon_fence_signaled(fence), timeout);
+
+        WaitEvent(fence->evnt);
+
 		radeon_irq_kms_sw_irq_put(rdev);
 		if (unlikely(r < 0)) {
 			return r;
 		}
 	} else {
 		radeon_irq_kms_sw_irq_get(rdev);
-		r = wait_event_timeout(rdev->fence_drv.queue,
+//       r = wait_event_timeout(rdev->fence_drv.queue,
-			 radeon_fence_signaled(fence), timeout);
+//            radeon_fence_signaled(fence), timeout);
+
+        WaitEvent(fence->evnt);
+
 		radeon_irq_kms_sw_irq_put(rdev);
 	}
 //   trace_radeon_fence_wait_end(rdev->ddev, seq);
@@ -272,20 +268,21 @@ retry:
 			 * as signaled for now
 			 */
 			rdev->gpu_lockup = true;
-			r = radeon_gpu_reset(rdev);
+//           r = radeon_gpu_reset(rdev);
-			if (r)
+//           if (r)
-				return r;
+//               return r;
-			radeon_fence_write(rdev, fence->seq);
+            return true;
-			rdev->gpu_lockup = false;
+
+//           radeon_fence_write(rdev, fence->seq);
+//           rdev->gpu_lockup = false;
 			}
-//       timeout = RADEON_FENCE_JIFFIES_TIMEOUT;
+        timeout = RADEON_FENCE_JIFFIES_TIMEOUT;
 		write_lock_irqsave(&rdev->fence_drv.lock, irq_flags);
-//       rdev->fence_drv.last_timeout = RADEON_FENCE_JIFFIES_TIMEOUT;
+        rdev->fence_drv.last_timeout = RADEON_FENCE_JIFFIES_TIMEOUT;
-//       rdev->fence_drv.last_jiffies = jiffies;
+        rdev->fence_drv.last_jiffies = GetTimerTicks();
 		write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags);
 		goto retry;
 	}
-#endif
 	return 0;
 }
 
@@ -346,12 +343,20 @@ struct radeon_fence *radeon_fence_ref(struct radeon_fence *fence)
 
 void radeon_fence_unref(struct radeon_fence **fence)
 {
+    unsigned long irq_flags;
 	struct radeon_fence *tmp = *fence;
 
 	*fence = NULL;
+
+    if(tmp)
+    {
+        write_lock_irqsave(&tmp->rdev->fence_drv.lock, irq_flags);
+        list_del(&tmp->list);
+        tmp->emited = false;
+        write_unlock_irqrestore(&tmp->rdev->fence_drv.lock, irq_flags);
+    };
 }
 
-#if 0
 void radeon_fence_process(struct radeon_device *rdev)
 {
 	unsigned long irq_flags;
@@ -360,13 +365,8 @@ void radeon_fence_process(struct radeon_device *rdev)
 	write_lock_irqsave(&rdev->fence_drv.lock, irq_flags);
 	wake = radeon_fence_poll_locked(rdev);
 	write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags);
-	if (wake) {
-		wake_up_all(&rdev->fence_drv.queue);
-	}
 }
 
-#endif
-
 int radeon_fence_driver_init(struct radeon_device *rdev)
 {
 	unsigned long irq_flags;

drivers/video/drm/radeon/radeon_gart.c

@@ -184,11 +184,8 @@ int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
     uint64_t page_base;
     int i, j;
 
-    ENTER();
+//    dbgprintf("offset %x pages %d list %x\n",
-
+//               offset, pages, pagelist);
-    dbgprintf("offset %x pages %x list %x\n",
-               offset, pages, pagelist);
-
     if (!rdev->gart.ready) {
 		WARN(1, "trying to bind memory to unitialized GART !\n");
         return -EINVAL;

drivers/video/drm/radeon/radeon_irq_kms.c

@@ -88,7 +88,7 @@ void radeon_driver_irq_uninstall_kms(struct drm_device *dev)
 
 void irq_handler_kms()
 {
-    dbgprintf("%s\n",__FUNCTION__);
+//    dbgprintf("%s\n",__FUNCTION__);
     radeon_irq_process(main_device);
 }
 

drivers/video/drm/radeon/radeon_object_kos.c

@@ -136,6 +136,8 @@ int radeon_bo_create(struct radeon_device *rdev,
 
     num_pages = (size + PAGE_SIZE - 1) >> PAGE_SHIFT;
 
+    size = num_pages << PAGE_SHIFT;
+
     if (num_pages == 0) {
         dbgprintf("Illegal buffer object size.\n");
         return -EINVAL;

drivers/video/drm/radeon/radeon_reg.h

@@ -300,6 +300,8 @@
 #       define RADEON_BUS_READ_BURST         (1 << 30)
 #define RADEON_BUS_CNTL1                    0x0034
 #       define RADEON_BUS_WAIT_ON_LOCK_EN    (1 << 4)
+#define RV370_BUS_CNTL                      0x004c
+#       define RV370_BUS_BIOS_DIS_ROM        (1 << 2)
 /* rv370/rv380, rv410, r423/r430/r480, r5xx */
 #define RADEON_MSI_REARM_EN		    0x0160
 #	define RV370_MSI_REARM_EN	     (1 << 0)

drivers/video/drm/radeon/radeon_ring.c

@@ -71,20 +71,20 @@ int radeon_ib_get(struct radeon_device *rdev, struct radeon_ib **ib)
 	}
 	rdev->ib_pool.head_id = (nib->idx + 1) & (RADEON_IB_POOL_SIZE - 1);
 	nib->free = false;
-//   if (nib->fence) {
+	if (nib->fence) {
-//       mutex_unlock(&rdev->ib_pool.mutex);
+		mutex_unlock(&rdev->ib_pool.mutex);
-//   r = radeon_fence_wait(nib->fence, false);
+		r = radeon_fence_wait(nib->fence, false);
-//   if (r) {
+		if (r) {
-//           dev_err(rdev->dev, "error waiting fence of IB(%u:0x%016lX:%u)\n",
+			dev_err(rdev->dev, "error waiting fence of IB(%u:0x%016lX:%u)\n",
-//               nib->idx, (unsigned long)nib->gpu_addr, nib->length_dw);
+				nib->idx, (unsigned long)nib->gpu_addr, nib->length_dw);
-//           mutex_lock(&rdev->ib_pool.mutex);
+			mutex_lock(&rdev->ib_pool.mutex);
-//           nib->free = true;
+			nib->free = true;
-//           mutex_unlock(&rdev->ib_pool.mutex);
+			mutex_unlock(&rdev->ib_pool.mutex);
-//           radeon_fence_unref(&fence);
+            radeon_fence_unref(&fence);
-//           return r;
+			return r;
-//       }
+		}
-//       mutex_lock(&rdev->ib_pool.mutex);
+		mutex_lock(&rdev->ib_pool.mutex);
-//   }
+	}
     radeon_fence_unref(&nib->fence);
 	nib->fence = fence;
 	nib->length_dw = 0;

drivers/video/drm/radeon/rs400.c

@@ -400,8 +400,14 @@ static int rs400_startup(struct radeon_device *rdev)
 	r = rs400_gart_enable(rdev);
 	if (r)
 		return r;
+
+	/* allocate wb buffer */
+	r = radeon_wb_init(rdev);
+	if (r)
+		return r;
+
 	/* Enable IRQ */
-//	r100_irq_set(rdev);
+	r100_irq_set(rdev);
 	rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL);
 	/* 1M ring buffer */
    r = r100_cp_init(rdev, 1024 * 1024);
@@ -409,11 +415,11 @@ static int rs400_startup(struct radeon_device *rdev)
 		dev_err(rdev->dev, "failed initializing CP (%d).\n", r);
        return r;
    }
-//	r = r100_ib_init(rdev);
+	r = r100_ib_init(rdev);
-//	if (r) {
+	if (r) {
-//		dev_err(rdev->dev, "failled initializing IB (%d).\n", r);
+		dev_err(rdev->dev, "failed initializing IB (%d).\n", r);
-//		return r;
+		return r;
-//	}
+	}
 	return 0;
 }
 
@@ -462,12 +468,12 @@ int rs400_init(struct radeon_device *rdev)
 	/* initialize memory controller */
 	rs400_mc_init(rdev);
 	/* Fence driver */
-//	r = radeon_fence_driver_init(rdev);
+	r = radeon_fence_driver_init(rdev);
-//	if (r)
+	if (r)
-//		return r;
+		return r;
-//	r = radeon_irq_kms_init(rdev);
+	r = radeon_irq_kms_init(rdev);
-//	if (r)
+	if (r)
-//		return r;
+		return r;
 	/* Memory manager */
 	r = radeon_bo_init(rdev);
 	if (r)

drivers/video/drm/radeon/rs600.c

@@ -108,19 +108,19 @@ void rs600_hpd_init(struct radeon_device *rdev)
 		case RADEON_HPD_1:
 			WREG32(R_007D00_DC_HOT_PLUG_DETECT1_CONTROL,
 			       S_007D00_DC_HOT_PLUG_DETECT1_EN(1));
-//           rdev->irq.hpd[0] = true;
+			rdev->irq.hpd[0] = true;
 			break;
 		case RADEON_HPD_2:
 			WREG32(R_007D10_DC_HOT_PLUG_DETECT2_CONTROL,
 			       S_007D10_DC_HOT_PLUG_DETECT2_EN(1));
-//           rdev->irq.hpd[1] = true;
+			rdev->irq.hpd[1] = true;
 			break;
 		default:
 			break;
 		}
 	}
-//   if (rdev->irq.installed)
+	if (rdev->irq.installed)
-//   rs600_irq_set(rdev);
+		rs600_irq_set(rdev);
 }
 
 void rs600_hpd_fini(struct radeon_device *rdev)
@@ -134,12 +134,12 @@ void rs600_hpd_fini(struct radeon_device *rdev)
 		case RADEON_HPD_1:
 			WREG32(R_007D00_DC_HOT_PLUG_DETECT1_CONTROL,
 			       S_007D00_DC_HOT_PLUG_DETECT1_EN(0));
-//           rdev->irq.hpd[0] = false;
+			rdev->irq.hpd[0] = false;
 			break;
 		case RADEON_HPD_2:
 			WREG32(R_007D10_DC_HOT_PLUG_DETECT2_CONTROL,
 			       S_007D10_DC_HOT_PLUG_DETECT2_EN(0));
-//           rdev->irq.hpd[1] = false;
+			rdev->irq.hpd[1] = false;
 			break;
 		default:
 			break;
@@ -257,7 +257,7 @@ int rs600_gart_init(struct radeon_device *rdev)
 	return radeon_gart_table_vram_alloc(rdev);
 }
 
-int rs600_gart_enable(struct radeon_device *rdev)
+static int rs600_gart_enable(struct radeon_device *rdev)
 {
 	u32 tmp;
 	int r, i;
@@ -271,8 +271,8 @@ int rs600_gart_enable(struct radeon_device *rdev)
 		return r;
 	radeon_gart_restore(rdev);
 	/* Enable bus master */
-	tmp = RREG32(R_00004C_BUS_CNTL) & C_00004C_BUS_MASTER_DIS;
+	tmp = RREG32(RADEON_BUS_CNTL) & ~RS600_BUS_MASTER_DIS;
-	WREG32(R_00004C_BUS_CNTL, tmp);
+	WREG32(RADEON_BUS_CNTL, tmp);
 	/* FIXME: setup default page */
 	WREG32_MC(R_000100_MC_PT0_CNTL,
 		 (S_000100_EFFECTIVE_L2_CACHE_SIZE(6) |
@@ -365,8 +365,6 @@ int rs600_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr)
 	return 0;
 }
 
-#if 0
-
 int rs600_irq_set(struct radeon_device *rdev)
 {
 	uint32_t tmp = 0;
@@ -459,7 +457,80 @@ void rs600_irq_disable(struct radeon_device *rdev)
 	rs600_irq_ack(rdev);
 }
 
-#endif
+int rs600_irq_process(struct radeon_device *rdev)
+{
+	u32 status, msi_rearm;
+	bool queue_hotplug = false;
+
+	/* reset gui idle ack.  the status bit is broken */
+	rdev->irq.gui_idle_acked = false;
+
+	status = rs600_irq_ack(rdev);
+	if (!status && !rdev->irq.stat_regs.r500.disp_int) {
+		return IRQ_NONE;
+	}
+	while (status || rdev->irq.stat_regs.r500.disp_int) {
+		/* SW interrupt */
+		if (G_000044_SW_INT(status)) {
+			radeon_fence_process(rdev);
+		}
+		/* GUI idle */
+		if (G_000040_GUI_IDLE(status)) {
+			rdev->irq.gui_idle_acked = true;
+			rdev->pm.gui_idle = true;
+//			wake_up(&rdev->irq.idle_queue);
+		}
+		/* Vertical blank interrupts */
+		if (G_007EDC_LB_D1_VBLANK_INTERRUPT(rdev->irq.stat_regs.r500.disp_int)) {
+			if (rdev->irq.crtc_vblank_int[0]) {
+//				drm_handle_vblank(rdev->ddev, 0);
+				rdev->pm.vblank_sync = true;
+//				wake_up(&rdev->irq.vblank_queue);
+			}
+//			if (rdev->irq.pflip[0])
+//				radeon_crtc_handle_flip(rdev, 0);
+		}
+		if (G_007EDC_LB_D2_VBLANK_INTERRUPT(rdev->irq.stat_regs.r500.disp_int)) {
+			if (rdev->irq.crtc_vblank_int[1]) {
+//				drm_handle_vblank(rdev->ddev, 1);
+				rdev->pm.vblank_sync = true;
+//				wake_up(&rdev->irq.vblank_queue);
+			}
+//			if (rdev->irq.pflip[1])
+//				radeon_crtc_handle_flip(rdev, 1);
+		}
+		if (G_007EDC_DC_HOT_PLUG_DETECT1_INTERRUPT(rdev->irq.stat_regs.r500.disp_int)) {
+			queue_hotplug = true;
+			DRM_DEBUG("HPD1\n");
+		}
+		if (G_007EDC_DC_HOT_PLUG_DETECT2_INTERRUPT(rdev->irq.stat_regs.r500.disp_int)) {
+			queue_hotplug = true;
+			DRM_DEBUG("HPD2\n");
+		}
+		status = rs600_irq_ack(rdev);
+	}
+	/* reset gui idle ack.  the status bit is broken */
+	rdev->irq.gui_idle_acked = false;
+//	if (queue_hotplug)
+//		schedule_work(&rdev->hotplug_work);
+	if (rdev->msi_enabled) {
+		switch (rdev->family) {
+		case CHIP_RS600:
+		case CHIP_RS690:
+		case CHIP_RS740:
+			msi_rearm = RREG32(RADEON_BUS_CNTL) & ~RS600_MSI_REARM;
+			WREG32(RADEON_BUS_CNTL, msi_rearm);
+			WREG32(RADEON_BUS_CNTL, msi_rearm | RS600_MSI_REARM);
+			break;
+		default:
+			msi_rearm = RREG32(RADEON_MSI_REARM_EN) & ~RV370_MSI_REARM_EN;
+			WREG32(RADEON_MSI_REARM_EN, msi_rearm);
+			WREG32(RADEON_MSI_REARM_EN, msi_rearm | RV370_MSI_REARM_EN);
+			break;
+		}
+	}
+	return IRQ_HANDLED;
+}
 
 u32 rs600_get_vblank_counter(struct radeon_device *rdev, int crtc)
 {
@@ -602,8 +673,14 @@ static int rs600_startup(struct radeon_device *rdev)
 	r = rs600_gart_enable(rdev);
 	if (r)
 	return r;
+
+	/* allocate wb buffer */
+	r = radeon_wb_init(rdev);
+	if (r)
+		return r;
+
 	/* Enable IRQ */
-//	rs600_irq_set(rdev);
+	rs600_irq_set(rdev);
 	rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL);
 	/* 1M ring buffer */
 	r = r100_cp_init(rdev, 1024 * 1024);
@@ -611,11 +688,11 @@ static int rs600_startup(struct radeon_device *rdev)
 		dev_err(rdev->dev, "failed initializing CP (%d).\n", r);
 		return r;
 	}
-//	r = r100_ib_init(rdev);
+	r = r100_ib_init(rdev);
-//	if (r) {
+	if (r) {
-//		dev_err(rdev->dev, "failled initializing IB (%d).\n", r);
+		dev_err(rdev->dev, "failed initializing IB (%d).\n", r);
-//		return r;
+		return r;
-//	}
+	}
 	return 0;
 }
 
@@ -663,12 +740,12 @@ int rs600_init(struct radeon_device *rdev)
 	rs600_mc_init(rdev);
 	rs600_debugfs(rdev);
 	/* Fence driver */
-//	r = radeon_fence_driver_init(rdev);
+	r = radeon_fence_driver_init(rdev);
-//	if (r)
+	if (r)
-//		return r;
+		return r;
-//	r = radeon_irq_kms_init(rdev);
+	r = radeon_irq_kms_init(rdev);
-//	if (r)
+	if (r)
-//		return r;
+		return r;
 	/* Memory manager */
 	r = radeon_bo_init(rdev);
 	if (r)

drivers/video/drm/radeon/rs690.c

@@ -615,8 +615,14 @@ static int rs690_startup(struct radeon_device *rdev)
 	r = rs400_gart_enable(rdev);
 	if (r)
 		return r;
+
+	/* allocate wb buffer */
+	r = radeon_wb_init(rdev);
+	if (r)
+		return r;
+
 	/* Enable IRQ */
-//	rs600_irq_set(rdev);
+	rs600_irq_set(rdev);
 	rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL);
 	/* 1M ring buffer */
 	r = r100_cp_init(rdev, 1024 * 1024);
@@ -624,11 +630,11 @@ static int rs690_startup(struct radeon_device *rdev)
 		dev_err(rdev->dev, "failed initializing CP (%d).\n", r);
 		return r;
 	}
-//	r = r100_ib_init(rdev);
+	r = r100_ib_init(rdev);
-//	if (r) {
+	if (r) {
-//		dev_err(rdev->dev, "failled initializing IB (%d).\n", r);
+		dev_err(rdev->dev, "failed initializing IB (%d).\n", r);
-//		return r;
+		return r;
-//	}
+	}
 	return 0;
 }
 
@@ -678,12 +684,12 @@ int rs690_init(struct radeon_device *rdev)
 	rs690_mc_init(rdev);
 	rv515_debugfs(rdev);
 	/* Fence driver */
-//	r = radeon_fence_driver_init(rdev);
+	r = radeon_fence_driver_init(rdev);
-//	if (r)
+	if (r)
-//		return r;
+		return r;
-//	r = radeon_irq_kms_init(rdev);
+	r = radeon_irq_kms_init(rdev);
-//	if (r)
+	if (r)
-//		return r;
+		return r;
 	/* Memory manager */
 	r = radeon_bo_init(rdev);
 	if (r)

drivers/video/drm/radeon/rv515.c

@@ -386,8 +386,14 @@ static int rv515_startup(struct radeon_device *rdev)
 		if (r)
 			return r;
 	}
+
+	/* allocate wb buffer */
+	r = radeon_wb_init(rdev);
+	if (r)
+		return r;
+
 	/* Enable IRQ */
-//	rs600_irq_set(rdev);
+	rs600_irq_set(rdev);
 	rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL);
 	/* 1M ring buffer */
 	r = r100_cp_init(rdev, 1024 * 1024);
@@ -395,11 +401,11 @@ static int rv515_startup(struct radeon_device *rdev)
 		dev_err(rdev->dev, "failed initializing CP (%d).\n", r);
 		return r;
 	}
-//	r = r100_ib_init(rdev);
+	r = r100_ib_init(rdev);
-//	if (r) {
+	if (r) {
-//		dev_err(rdev->dev, "failled initializing IB (%d).\n", r);
+		dev_err(rdev->dev, "failed initializing IB (%d).\n", r);
-//		return r;
+		return r;
-//	}
+	}
 	return 0;
 }
 
@@ -457,12 +463,12 @@ int rv515_init(struct radeon_device *rdev)
 	rv515_mc_init(rdev);
 	rv515_debugfs(rdev);
 	/* Fence driver */
-//   r = radeon_fence_driver_init(rdev);
+	r = radeon_fence_driver_init(rdev);
-//   if (r)
+	if (r)
-//       return r;
+		return r;
-//	r = radeon_irq_kms_init(rdev);
+	r = radeon_irq_kms_init(rdev);
-//	if (r)
+	if (r)
-//		return r;
+		return r;
 	/* Memory manager */
 	r = radeon_bo_init(rdev);
 	if (r)

drivers/video/drm/radeon/rv770.c

@@ -1088,6 +1088,13 @@ static int rv770_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 	rv770_gpu_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);
+	}
+
 	/* allocate wb buffer */
 	r = radeon_wb_init(rdev);
 	if (r)