Add support for UVD and VCE status

amdgpu.c

@@ -26,6 +26,16 @@ static int getgrbm_amdgpu(uint32_t *out) {
 					0xffffffff, 0, out);
 }
 
+static int getsrbm_amdgpu(uint32_t *out) {
+	return amdgpu_read_mm_registers(amdgpu_dev, SRBM_STATUS / 4, 1,
+					0xffffffff, 0, out);
+}
+
+static int getsrbm2_amdgpu(uint32_t *out) {
+	return amdgpu_read_mm_registers(amdgpu_dev, SRBM_STATUS2 / 4, 1,
+					0xffffffff, 0, out);
+}
+
 static int getvram_amdgpu(uint64_t *out) {
 	return amdgpu_query_info(amdgpu_dev, AMDGPU_INFO_VRAM_USAGE,
 				sizeof(uint64_t), out);
@@ -59,9 +69,11 @@ void init_amdgpu(int fd) {
 	if (amdgpu_device_initialize(fd, &drm_major, &drm_minor, &amdgpu_dev))
 		return;
 
-	if (!(ret = getgrbm_amdgpu(&out32)))
+	if (!(ret = getgrbm_amdgpu(&out32))) {
 		getgrbm = getgrbm_amdgpu;
-	else
+		getsrbm = getsrbm_amdgpu;
+		getsrbm2 = getsrbm2_amdgpu;
+	} else
 		drmError(ret, _("Failed to get GPU usage"));
 
 #ifdef HAS_AMDGPU_QUERY_SENSOR_INFO

detect.c

@@ -28,9 +28,12 @@ uint64_t gttsize;
 unsigned int sclk_max = 0; // kilohertz
 unsigned int mclk_max = 0; // kilohertz
 const void *area;
+static const void *srbm_area;
 
 // function pointers to the right backend
 int (*getgrbm)(uint32_t *out);
+int (*getsrbm)(uint32_t *out);
+int (*getsrbm2)(uint32_t *out);
 int (*getvram)(uint64_t *out);
 int (*getgtt)(uint64_t *out);
 int (*getsclk)(uint32_t *out);
@@ -75,6 +78,16 @@ static int getgrbm_pci(uint32_t *out) {
 	return 0;
 }
 
+static int getsrbm_pci(uint32_t *out) {
+	*out = *(uint32_t *)((const char *) srbm_area + SRBM_STATUS);
+	return 0;
+}
+
+static int getsrbm2_pci(uint32_t *out) {
+	*out = *(uint32_t *)((const char *) srbm_area + SRBM_STATUS2);
+	return 0;
+}
+
 static void open_pci(struct pci_device *gpu_device) {
 	// Warning: gpu_device is a copy of a freed struct. Do not access any pointers!
 	int reg = 2;
@@ -91,12 +104,18 @@ static void open_pci(struct pci_device *gpu_device) {
 	area = mmap(NULL, MMAP_SIZE, PROT_READ, MAP_PRIVATE, mem,
 			gpu_device->regions[reg].base_addr + 0x8000);
 	if (area == MAP_FAILED) die(_("mmap failed"));
+	srbm_area = mmap(NULL, SRBM_MMAP_SIZE, PROT_READ, MAP_PRIVATE, mem,
+			gpu_device->regions[reg].base_addr);
+	if (srbm_area == MAP_FAILED) die(_("mmap failed"));
 
 	getgrbm = getgrbm_pci;
+	getsrbm = getsrbm_pci;
+	getsrbm2 = getsrbm2_pci;
 }
 
 static void cleanup_pci() {
 	munmap((void *) area, MMAP_SIZE);
+	munmap((void *) srbm_area, SRBM_MMAP_SIZE);
 }
 
 static int init_drm(int drm_fd) {
@@ -247,6 +266,7 @@ void init_pci(const char *path, short *bus, unsigned int *device_id, const unsig
 	short device_bus = -1;
 	int err = 1;
 	getgrbm = getsclk = getmclk = getuint32_null;
+	getsrbm = getsrbm2 = getuint32_null;
 	getvram = getgtt = getuint64_null;
 
 	if (path) {
@@ -339,6 +359,8 @@ void initbits(int fam) {
 	bits.cr = (1U << 27);
 	bits.cb = (1U << 30);
 	bits.gui = (1U << 31);
+	bits.uvd = 0;
+	bits.vce0 = 0;
 
 	// R600 has a different texture bit, and only R600 has the TC, CR, SMX bits
 	if (fam < RV770) {
@@ -348,4 +370,11 @@ void initbits(int fam) {
 		bits.cr = 0;
 		bits.smx = 0;
 	}
+
+	if (fam >= RV610 && fam < VEGAM) {
+		bits.uvd = (1U << 19);
+		if (fam >= CAYMAN) {
+			bits.vce0 = (1U << 7);
+		}
+	}
 }

dump.c

@@ -96,6 +96,8 @@ void dumpdata(const unsigned int ticks, const char file[], const unsigned int li
 		float db = 100 * results->db * k;
 		float cr = 100 * results->cr * k;
 		float cb = 100 * results->cb * k;
+		float uvd = 100 * results->uvd * k;
+		float vce0 = 100 * results->vce0 * k;
 		float vram = 100.0f * results->vram / vramsize;
 		float vrammb = results->vram / 1024.0f / 1024.0f;
 		float gtt = 100.0f * results->gtt / gttsize;
@@ -127,6 +129,10 @@ void dumpdata(const unsigned int ticks, const char file[], const unsigned int li
 		fprintf(f, "pa %.2f%%, ", pa);
 		fprintf(f, "db %.2f%%, ", db);
 		fprintf(f, "cb %.2f%%", cb);
+		if (bits.uvd)
+			fprintf(f, ", uvd %.2f%%", uvd);
+		if (bits.vce0)
+			fprintf(f, ", vce0 %.2f%%", vce0);
 
 		if (bits.vram)
 			fprintf(f, ", vram %.2f%% %.2fmb", vram, vrammb);

include/radeontop.h

@@ -36,7 +36,10 @@
 
 enum {
 	GRBM_STATUS = 0x8010,
+	SRBM_STATUS = 0xe50,
+	SRBM_STATUS2 = 0xe4c,
 	MMAP_SIZE = 0x14,
+	SRBM_MMAP_SIZE = 0xe54,
 	VENDOR_AMD = 0x1002
 };
 
@@ -53,6 +56,8 @@ void initbits(int fam);
 void cleanup();
 
 extern int (*getgrbm)(uint32_t *out);
+extern int (*getsrbm)(uint32_t *out);
+extern int (*getsrbm2)(uint32_t *out);
 extern int (*getvram)(uint64_t *out);
 extern int (*getgtt)(uint64_t *out);
 extern int (*getsclk)(uint32_t *out);
@@ -151,6 +156,8 @@ struct bits_t {
 	unsigned int db;
 	unsigned int cb;
 	unsigned int cr;
+	unsigned int uvd;
+	unsigned int vce0;
 	uint64_t vram;
 	uint64_t gtt;
 	unsigned int sclk;

radeon.c

@@ -50,6 +50,16 @@ static int getgrbm_radeon(uint32_t *out) {
 	return radeon_get_drm_value(drm_fd, RADEON_INFO_READ_REG, out);
 }
 
+static int getsrbm_radeon(uint32_t *out) {
+	*out = SRBM_STATUS;
+	return radeon_get_drm_value(drm_fd, RADEON_INFO_READ_REG, out);
+}
+
+static int getsrbm2_radeon(uint32_t *out) {
+	*out = SRBM_STATUS2;
+	return radeon_get_drm_value(drm_fd, RADEON_INFO_READ_REG, out);
+}
+
 static int getsclk_radeon(uint32_t *out) {
 	return radeon_get_drm_value(drm_fd, RADEON_INFO_CURRENT_GPU_SCLK, out);
 }
@@ -83,9 +93,11 @@ void init_radeon(int fd, int drm_major, int drm_minor) {
 
 #ifdef RADEON_INFO_READ_REG
 	if (DRM_ATLEAST_VERSION(2, 42)) {
-		if (!(ret = getgrbm_radeon(&out32)))
+		if (!(ret = getgrbm_radeon(&out32))) {
 			getgrbm = getgrbm_radeon;
-		else
+			getsrbm = getsrbm_radeon;
+			getsrbm2 = getsrbm2_radeon;
+		} else
 			drmError(ret, _("Failed to get GPU usage"));
 
 		if ((ret = radeon_get_drm_value(drm_fd, RADEON_INFO_MAX_SCLK,

ticks.c

@@ -41,6 +41,10 @@ static void *collector(void *arg) {
 	while (1) {
 		unsigned int stat;
 		getgrbm(&stat);
+		unsigned int uvd;
+		getsrbm(&uvd);
+		unsigned int vce;
+		getsrbm2(&vce);
 
 		memset(&history[cur], 0, sizeof(struct bits_t));
 
@@ -58,6 +62,8 @@ static void *collector(void *arg) {
 		if (stat & bits.db) history[cur].db = 1;
 		if (stat & bits.cr) history[cur].cr = 1;
 		if (stat & bits.cb) history[cur].cb = 1;
+		if (uvd & bits.uvd) history[cur].uvd = 1;
+		if (vce & bits.vce0) history[cur].vce0 = 1;
 		getsclk(&history[cur].sclk);
 		getmclk(&history[cur].mclk);
 
@@ -86,6 +92,8 @@ static void *collector(void *arg) {
 				res[curres].db += history[i].db;
 				res[curres].cb += history[i].cb;
 				res[curres].cr += history[i].cr;
+				res[curres].uvd += history[i].uvd;
+				res[curres].vce0 += history[i].vce0;
 				res[curres].mclk += history[i].mclk;
 				res[curres].sclk += history[i].sclk;
 			}

ui.c

@@ -109,7 +109,7 @@ void present(const unsigned int ticks, const char card[], unsigned int color,
 		init_pair(5, COLOR_YELLOW, COLOR_BLACK);
 	}
 
-	const unsigned int bigh = 23;
+	const unsigned int bigh = 26;
 
 	// Screen dimensions. (Re)calculated only when resize is non-zero.
 	unsigned int h = 1, w = 1, hw = 1;
@@ -150,6 +150,8 @@ void present(const unsigned int ticks, const char card[], unsigned int color,
 		float db = 100 * results->db * k;
 		float cr = 100 * results->cr * k;
 		float cb = 100 * results->cb * k;
+		float uvd = 100 * results->uvd * k;
+		float vce0 = 100 * results->vce0 * k;
 		float vram = 100.0f * results->vram / vramsize;
 		float vrammb = results->vram / 1024.0f / 1024.0f;
 		float vramsizemb = vramsize / 1024.0f / 1024.0f;
@@ -243,6 +245,18 @@ void present(const unsigned int ticks, const char card[], unsigned int color,
 		}
 		if (color) attroff(COLOR_PAIR(5));
 
+		if (bits.uvd) {
+			percentage(start, w, uvd);
+			printright(start++, hw, _("UVD %6.2f%%"), uvd);
+		}
+		if (bits.vce0) {
+			percentage(start, w, vce0);
+			printright(start++, hw, _("VCE %6.2f%%"), vce0);
+		}
+
+		// Enough height?
+		if(h > bigh) start++;
+
 		if (bits.vram || bits.gtt) {
 			// Enough height?
 			if (h > bigh) start++;