drm/nvc0/gr: fix some bugs in grctx generation

Most serious is for chips with only 1 TPC, we'd get stuck in an infinite
loop.  The fix here will slightly change the setup for all other chipsets
too, but, it shouldn't matter too much, and this all needs figuring out
and likely redone anyway.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
diff --git a/drivers/gpu/drm/nouveau/nvc0_grctx.c b/drivers/gpu/drm/nouveau/nvc0_grctx.c
index dd0e6a7..96b0b93 100644
--- a/drivers/gpu/drm/nouveau/nvc0_grctx.c
+++ b/drivers/gpu/drm/nouveau/nvc0_grctx.c
@@ -1812,6 +1812,7 @@
 		/* calculate first set of magics */
 		memcpy(tpnr, priv->tp_nr, sizeof(priv->tp_nr));
 
+		gpc = -1;
 		for (tp = 0; tp < priv->tp_total; tp++) {
 			do {
 				gpc = (gpc + 1) % priv->gpc_nr;
@@ -1861,30 +1862,26 @@
 
 	if (1) {
 		u32 tp_mask = 0, tp_set = 0;
-		u8  tpnr[GPC_MAX];
+		u8  tpnr[GPC_MAX], a, b;
 
 		memcpy(tpnr, priv->tp_nr, sizeof(priv->tp_nr));
 		for (gpc = 0; gpc < priv->gpc_nr; gpc++)
 			tp_mask |= ((1 << priv->tp_nr[gpc]) - 1) << (gpc * 8);
 
-		gpc = -1;
-		for (i = 0, gpc = -1; i < 32; i++) {
-			int ltp = i * (priv->tp_total - 1) / 32;
+		for (i = 0, gpc = -1, b = -1; i < 32; i++) {
+			a = (i * (priv->tp_total - 1)) / 32;
+			if (a != b) {
+				b = a;
+				do {
+					gpc = (gpc + 1) % priv->gpc_nr;
+				} while (!tpnr[gpc]);
+				tp = priv->tp_nr[gpc] - tpnr[gpc]--;
 
-			do {
-				gpc = (gpc + 1) % priv->gpc_nr;
-			} while (!tpnr[gpc]);
-			tp = priv->tp_nr[gpc] - tpnr[gpc]--;
+				tp_set |= 1 << ((gpc * 8) + tp);
+			}
 
-			tp_set |= 1 << ((gpc * 8) + tp);
-
-			do {
-				nv_wr32(dev, 0x406800 + (i * 0x20), tp_set);
-				tp_set ^= tp_mask;
-				nv_wr32(dev, 0x406c00 + (i * 0x20), tp_set);
-				tp_set ^= tp_mask;
-			} while (ltp == (++i * (priv->tp_total - 1) / 32));
-			i--;
+			nv_wr32(dev, 0x406800 + (i * 0x20), tp_set);
+			nv_wr32(dev, 0x406c00 + (i * 0x20), tp_set ^ tp_mask);
 		}
 	}