Skip to content
Snippets Groups Projects
sequencer.c 120 KiB
Newer Older
/*
 * Copyright Altera Corporation (C) 2012-2015
 *
 * SPDX-License-Identifier:    BSD-3-Clause
 */

#include <common.h>
#include <asm/io.h>
#include <asm/arch/sdram.h>
#include "sequencer.h"
#include "sequencer_auto.h"
#include "sequencer_auto_ac_init.h"
#include "sequencer_auto_inst_init.h"
#include "sequencer_defines.h"

static void scc_mgr_load_dqs_for_write_group(uint32_t write_group);

static struct socfpga_sdr_rw_load_manager *sdr_rw_load_mgr_regs =
	(struct socfpga_sdr_rw_load_manager *)(BASE_RW_MGR + 0x800);

static struct socfpga_sdr_rw_load_jump_manager *sdr_rw_load_jump_mgr_regs =
	(struct socfpga_sdr_rw_load_jump_manager *)(BASE_RW_MGR + 0xC00);

static struct socfpga_sdr_reg_file *sdr_reg_file =
	(struct socfpga_sdr_reg_file *)(BASE_REG_FILE);

static struct socfpga_sdr_scc_mgr *sdr_scc_mgr =
	(struct socfpga_sdr_scc_mgr *)(BASE_SCC_MGR + 0x0E00);

static struct socfpga_phy_mgr_cmd *phy_mgr_cmd =
	(struct socfpga_phy_mgr_cmd *)(BASE_PHY_MGR);

static struct socfpga_phy_mgr_cfg *phy_mgr_cfg =
	(struct socfpga_phy_mgr_cfg *)(BASE_PHY_MGR + 0x4000);

static struct socfpga_data_mgr *data_mgr =
	(struct socfpga_data_mgr *)(BASE_DATA_MGR);

#define DELTA_D		1
#define MGR_SELECT_MASK		0xf8000

/*
 * In order to reduce ROM size, most of the selectable calibration steps are
 * decided at compile time based on the user's calibration mode selection,
 * as captured by the STATIC_CALIB_STEPS selection below.
 *
 * However, to support simulation-time selection of fast simulation mode, where
 * we skip everything except the bare minimum, we need a few of the steps to
 * be dynamic.  In those cases, we either use the DYNAMIC_CALIB_STEPS for the
 * check, which is based on the rtl-supplied value, or we dynamically compute
 * the value to use based on the dynamically-chosen calibration mode
 */

#define DLEVEL 0
#define STATIC_IN_RTL_SIM 0
#define STATIC_SKIP_DELAY_LOOPS 0

#define STATIC_CALIB_STEPS (STATIC_IN_RTL_SIM | CALIB_SKIP_FULL_TEST | \
	STATIC_SKIP_DELAY_LOOPS)

/* calibration steps requested by the rtl */
uint16_t dyn_calib_steps;

/*
 * To make CALIB_SKIP_DELAY_LOOPS a dynamic conditional option
 * instead of static, we use boolean logic to select between
 * non-skip and skip values
 *
 * The mask is set to include all bits when not-skipping, but is
 * zero when skipping
 */

uint16_t skip_delay_mask;	/* mask off bits when skipping/not-skipping */

#define SKIP_DELAY_LOOP_VALUE_OR_ZERO(non_skip_value) \
	((non_skip_value) & skip_delay_mask)

struct gbl_type *gbl;
struct param_type *param;
uint32_t curr_shadow_reg;

static uint32_t rw_mgr_mem_calibrate_write_test(uint32_t rank_bgn,
	uint32_t write_group, uint32_t use_dm,
	uint32_t all_correct, uint32_t *bit_chk, uint32_t all_ranks);

static u32 sdr_get_addr(u32 *base)
{
	u32 addr = (u32)base & MGR_SELECT_MASK;

	switch (addr) {
	case BASE_PHY_MGR:
		addr = (((u32)base >> 8) & (1 << 6)) | ((u32)base & 0x3f) |
			SDR_PHYGRP_PHYMGRGRP_ADDRESS;
		break;
	case BASE_RW_MGR:
		addr = ((u32)base & 0x1fff) | SDR_PHYGRP_RWMGRGRP_ADDRESS;
		break;
	case BASE_DATA_MGR:
		addr = ((u32)base & 0x7ff) | SDR_PHYGRP_DATAMGRGRP_ADDRESS;
		break;
	case BASE_SCC_MGR:
		addr = ((u32)base & 0xfff) | SDR_PHYGRP_SCCGRP_ADDRESS;
		break;
	case BASE_REG_FILE:
		addr = ((u32)base & 0x7ff) | SDR_PHYGRP_REGFILEGRP_ADDRESS;
		break;
	case BASE_MMR:
		addr = ((u32)base & 0xfff) | SDR_CTRLGRP_ADDRESS;
		break;
	default:
		return -1;
	}

	return addr;
}

static void set_failing_group_stage(uint32_t group, uint32_t stage,
	uint32_t substage)
{
	/*
	 * Only set the global stage if there was not been any other
	 * failing group
	 */
	if (gbl->error_stage == CAL_STAGE_NIL)	{
		gbl->error_substage = substage;
		gbl->error_stage = stage;
		gbl->error_group = group;
	}
}

static void reg_file_set_group(uint32_t set_group)
{
	u32 addr = sdr_get_addr(&sdr_reg_file->cur_stage);

	/* Read the current group and stage */
	uint32_t cur_stage_group = readl(SOCFPGA_SDR_ADDRESS + addr);

	/* Clear the group */
	cur_stage_group &= 0x0000FFFF;

	/* Set the group */
	cur_stage_group |= (set_group << 16);

	/* Write the data back */
	writel(cur_stage_group, SOCFPGA_SDR_ADDRESS + addr);
}

static void reg_file_set_stage(uint32_t set_stage)
{
	u32 addr = sdr_get_addr(&sdr_reg_file->cur_stage);
	/* Read the current group and stage */
	uint32_t cur_stage_group = readl(SOCFPGA_SDR_ADDRESS + addr);

	/* Clear the stage and substage */
	cur_stage_group &= 0xFFFF0000;

	/* Set the stage */
	cur_stage_group |= (set_stage & 0x000000FF);

	/* Write the data back */
	writel(cur_stage_group, SOCFPGA_SDR_ADDRESS + addr);
}

static void reg_file_set_sub_stage(uint32_t set_sub_stage)
{
	u32 addr = sdr_get_addr(&sdr_reg_file->cur_stage);
	/* Read the current group and stage */
	uint32_t cur_stage_group = readl(SOCFPGA_SDR_ADDRESS + addr);

	/* Clear the substage */
	cur_stage_group &= 0xFFFF00FF;

	/* Set the sub stage */
	cur_stage_group |= ((set_sub_stage << 8) & 0x0000FF00);

	/* Write the data back */
	writel(cur_stage_group, SOCFPGA_SDR_ADDRESS + addr);
}

static void initialize(void)
{
	u32 addr = sdr_get_addr(&phy_mgr_cfg->mux_sel);

	debug("%s:%d\n", __func__, __LINE__);
	/* USER calibration has control over path to memory */
	/*
	 * In Hard PHY this is a 2-bit control:
	 * 0: AFI Mux Select
	 * 1: DDIO Mux Select
	 */
	writel(0x3, SOCFPGA_SDR_ADDRESS + addr);

	/* USER memory clock is not stable we begin initialization  */
	addr = sdr_get_addr(&phy_mgr_cfg->reset_mem_stbl);
	writel(0, SOCFPGA_SDR_ADDRESS + addr);

	/* USER calibration status all set to zero */
	addr = sdr_get_addr(&phy_mgr_cfg->cal_status);
	writel(0, SOCFPGA_SDR_ADDRESS + addr);

201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200
	addr = sdr_get_addr(&phy_mgr_cfg->cal_debug_info);
	writel(0, SOCFPGA_SDR_ADDRESS + addr);

	if ((dyn_calib_steps & CALIB_SKIP_ALL) != CALIB_SKIP_ALL) {
		param->read_correct_mask_vg  = ((uint32_t)1 <<
			(RW_MGR_MEM_DQ_PER_READ_DQS /
			RW_MGR_MEM_VIRTUAL_GROUPS_PER_READ_DQS)) - 1;
		param->write_correct_mask_vg = ((uint32_t)1 <<
			(RW_MGR_MEM_DQ_PER_READ_DQS /
			RW_MGR_MEM_VIRTUAL_GROUPS_PER_READ_DQS)) - 1;
		param->read_correct_mask     = ((uint32_t)1 <<
			RW_MGR_MEM_DQ_PER_READ_DQS) - 1;
		param->write_correct_mask    = ((uint32_t)1 <<
			RW_MGR_MEM_DQ_PER_WRITE_DQS) - 1;
		param->dm_correct_mask       = ((uint32_t)1 <<
			(RW_MGR_MEM_DATA_WIDTH / RW_MGR_MEM_DATA_MASK_WIDTH))
			- 1;
	}
}

static void set_rank_and_odt_mask(uint32_t rank, uint32_t odt_mode)
{
	uint32_t odt_mask_0 = 0;
	uint32_t odt_mask_1 = 0;
	uint32_t cs_and_odt_mask;
	uint32_t addr;

	if (odt_mode == RW_MGR_ODT_MODE_READ_WRITE) {
		if (RW_MGR_MEM_NUMBER_OF_RANKS == 1) {
			/*
			 * 1 Rank
			 * Read: ODT = 0
			 * Write: ODT = 1
			 */
			odt_mask_0 = 0x0;
			odt_mask_1 = 0x1;
		} else if (RW_MGR_MEM_NUMBER_OF_RANKS == 2) {
			/* 2 Ranks */
			if (RW_MGR_MEM_NUMBER_OF_CS_PER_DIMM == 1) {
				/* - Dual-Slot , Single-Rank
				 * (1 chip-select per DIMM)
				 * OR
				 * - RDIMM, 4 total CS (2 CS per DIMM)
				 * means 2 DIMM
				 * Since MEM_NUMBER_OF_RANKS is 2 they are
				 * both single rank
				 * with 2 CS each (special for RDIMM)
				 * Read: Turn on ODT on the opposite rank
				 * Write: Turn on ODT on all ranks
				 */
				odt_mask_0 = 0x3 & ~(1 << rank);
				odt_mask_1 = 0x3;
			} else {
				/*
				 * USER - Single-Slot , Dual-rank DIMMs
				 * (2 chip-selects per DIMM)
				 * USER Read: Turn on ODT off on all ranks
				 * USER Write: Turn on ODT on active rank
				 */
				odt_mask_0 = 0x0;
				odt_mask_1 = 0x3 & (1 << rank);
			}
				} else {
			/* 4 Ranks
			 * Read:
			 * ----------+-----------------------+
			 *           |                       |
			 *           |         ODT           |
			 * Read From +-----------------------+
			 *   Rank    |  3  |  2  |  1  |  0  |
			 * ----------+-----+-----+-----+-----+
			 *     0     |  0  |  1  |  0  |  0  |
			 *     1     |  1  |  0  |  0  |  0  |
			 *     2     |  0  |  0  |  0  |  1  |
			 *     3     |  0  |  0  |  1  |  0  |
			 * ----------+-----+-----+-----+-----+
			 *
			 * Write:
			 * ----------+-----------------------+
			 *           |                       |
			 *           |         ODT           |
			 * Write To  +-----------------------+
			 *   Rank    |  3  |  2  |  1  |  0  |
			 * ----------+-----+-----+-----+-----+
			 *     0     |  0  |  1  |  0  |  1  |
			 *     1     |  1  |  0  |  1  |  0  |
			 *     2     |  0  |  1  |  0  |  1  |
			 *     3     |  1  |  0  |  1  |  0  |
			 * ----------+-----+-----+-----+-----+
			 */
			switch (rank) {
			case 0:
				odt_mask_0 = 0x4;
				odt_mask_1 = 0x5;
				break;
			case 1:
				odt_mask_0 = 0x8;
				odt_mask_1 = 0xA;
				break;
			case 2:
				odt_mask_0 = 0x1;
				odt_mask_1 = 0x5;
				break;
			case 3:
				odt_mask_0 = 0x2;
				odt_mask_1 = 0xA;
				break;
			}
		}
	} else {
		odt_mask_0 = 0x0;
		odt_mask_1 = 0x0;
	}

	cs_and_odt_mask =
		(0xFF & ~(1 << rank)) |
		((0xFF & odt_mask_0) << 8) |
		((0xFF & odt_mask_1) << 16);
	addr = sdr_get_addr((u32 *)RW_MGR_SET_CS_AND_ODT_MASK);
	writel(cs_and_odt_mask, SOCFPGA_SDR_ADDRESS + addr);
}

static void scc_mgr_initialize(void)
{
	u32 addr = sdr_get_addr((u32 *)SCC_MGR_HHP_RFILE);

	/*
	 * Clear register file for HPS
	 * 16 (2^4) is the size of the full register file in the scc mgr:
	 *	RFILE_DEPTH = log2(MEM_DQ_PER_DQS + 1 + MEM_DM_PER_DQS +
	 * MEM_IF_READ_DQS_WIDTH - 1) + 1;
	 */
	uint32_t i;
	for (i = 0; i < 16; i++) {
		debug_cond(DLEVEL == 1, "%s:%d: Clearing SCC RFILE index %u",
			   __func__, __LINE__, i);
		writel(0, SOCFPGA_SDR_ADDRESS + addr + (i << 2));
	}
}

static void scc_mgr_set_dqs_bus_in_delay(uint32_t read_group,
						uint32_t delay)
{
	u32 addr = sdr_get_addr((u32 *)SCC_MGR_DQS_IN_DELAY);

	/* Load the setting in the SCC manager */
	writel(delay, SOCFPGA_SDR_ADDRESS + addr + (read_group << 2));
}

static void scc_mgr_set_dqs_io_in_delay(uint32_t write_group,
	uint32_t delay)
{
	u32 addr = sdr_get_addr((u32 *)SCC_MGR_IO_IN_DELAY);

	writel(delay, SOCFPGA_SDR_ADDRESS + addr + (RW_MGR_MEM_DQ_PER_WRITE_DQS << 2));
}

static void scc_mgr_set_dqs_en_phase(uint32_t read_group, uint32_t phase)
{
	u32 addr = sdr_get_addr((u32 *)SCC_MGR_DQS_EN_PHASE);

	/* Load the setting in the SCC manager */
	writel(phase, SOCFPGA_SDR_ADDRESS + addr + (read_group << 2));
}

static void scc_mgr_set_dqs_en_phase_all_ranks(uint32_t read_group,
					       uint32_t phase)
{
	uint32_t r;
	uint32_t update_scan_chains;
	uint32_t addr;

	for (r = 0; r < RW_MGR_MEM_NUMBER_OF_RANKS;
	     r += NUM_RANKS_PER_SHADOW_REG) {
		/*
		 * USER although the h/w doesn't support different phases per
		 * shadow register, for simplicity our scc manager modeling
		 * keeps different phase settings per shadow reg, and it's
		 * important for us to keep them in sync to match h/w.
		 * for efficiency, the scan chain update should occur only
		 * once to sr0.
		 */
		update_scan_chains = (r == 0) ? 1 : 0;

		scc_mgr_set_dqs_en_phase(read_group, phase);

		if (update_scan_chains) {
			addr = sdr_get_addr(&sdr_scc_mgr->dqs_ena);
			writel(read_group, SOCFPGA_SDR_ADDRESS + addr);

			addr = sdr_get_addr(&sdr_scc_mgr->update);
			writel(0, SOCFPGA_SDR_ADDRESS + addr);
		}
	}
}

static void scc_mgr_set_dqdqs_output_phase(uint32_t write_group,
						  uint32_t phase)
{
	u32 addr = sdr_get_addr((u32 *)SCC_MGR_DQDQS_OUT_PHASE);

	/* Load the setting in the SCC manager */
	writel(phase, SOCFPGA_SDR_ADDRESS + addr + (write_group << 2));
}

static void scc_mgr_set_dqdqs_output_phase_all_ranks(uint32_t write_group,
						     uint32_t phase)
{
	uint32_t r;
	uint32_t update_scan_chains;
	uint32_t addr;

	for (r = 0; r < RW_MGR_MEM_NUMBER_OF_RANKS;
	     r += NUM_RANKS_PER_SHADOW_REG) {
		/*
		 * USER although the h/w doesn't support different phases per
		 * shadow register, for simplicity our scc manager modeling
		 * keeps different phase settings per shadow reg, and it's
		 * important for us to keep them in sync to match h/w.
		 * for efficiency, the scan chain update should occur only
		 * once to sr0.
		 */
		update_scan_chains = (r == 0) ? 1 : 0;

		scc_mgr_set_dqdqs_output_phase(write_group, phase);

		if (update_scan_chains) {
			addr = sdr_get_addr(&sdr_scc_mgr->dqs_ena);
			writel(write_group, SOCFPGA_SDR_ADDRESS + addr);

			addr = sdr_get_addr(&sdr_scc_mgr->update);
			writel(0, SOCFPGA_SDR_ADDRESS + addr);
		}
	}
}

static void scc_mgr_set_dqs_en_delay(uint32_t read_group, uint32_t delay)
{
	uint32_t addr = sdr_get_addr((u32 *)SCC_MGR_DQS_EN_DELAY);

	/* Load the setting in the SCC manager */
	writel(delay + IO_DQS_EN_DELAY_OFFSET, SOCFPGA_SDR_ADDRESS + addr +
	       (read_group << 2));
}

static void scc_mgr_set_dqs_en_delay_all_ranks(uint32_t read_group,
					       uint32_t delay)
{
	uint32_t r;
	uint32_t addr;

	for (r = 0; r < RW_MGR_MEM_NUMBER_OF_RANKS;
		r += NUM_RANKS_PER_SHADOW_REG) {
		scc_mgr_set_dqs_en_delay(read_group, delay);

		addr = sdr_get_addr(&sdr_scc_mgr->dqs_ena);
		writel(read_group, SOCFPGA_SDR_ADDRESS + addr);
		/*
		 * In shadow register mode, the T11 settings are stored in
		 * registers in the core, which are updated by the DQS_ENA
		 * signals. Not issuing the SCC_MGR_UPD command allows us to
		 * save lots of rank switching overhead, by calling
		 * select_shadow_regs_for_update with update_scan_chains
		 * set to 0.
		 */
		addr = sdr_get_addr(&sdr_scc_mgr->update);
		writel(0, SOCFPGA_SDR_ADDRESS + addr);
	}
	/*
	 * In shadow register mode, the T11 settings are stored in
	 * registers in the core, which are updated by the DQS_ENA
	 * signals. Not issuing the SCC_MGR_UPD command allows us to
	 * save lots of rank switching overhead, by calling
	 * select_shadow_regs_for_update with update_scan_chains
	 * set to 0.
	 */
	addr = sdr_get_addr(&sdr_scc_mgr->update);
	writel(0, SOCFPGA_SDR_ADDRESS + addr);
}

static void scc_mgr_set_oct_out1_delay(uint32_t write_group, uint32_t delay)
{
	uint32_t read_group;
	uint32_t addr = sdr_get_addr((u32 *)SCC_MGR_OCT_OUT1_DELAY);

	/*
	 * Load the setting in the SCC manager
	 * Although OCT affects only write data, the OCT delay is controlled
	 * by the DQS logic block which is instantiated once per read group.
	 * For protocols where a write group consists of multiple read groups,
	 * the setting must be set multiple times.
	 */
	for (read_group = write_group * RW_MGR_MEM_IF_READ_DQS_WIDTH /
	     RW_MGR_MEM_IF_WRITE_DQS_WIDTH;
	     read_group < (write_group + 1) * RW_MGR_MEM_IF_READ_DQS_WIDTH /
	     RW_MGR_MEM_IF_WRITE_DQS_WIDTH; ++read_group)
		writel(delay, SOCFPGA_SDR_ADDRESS + addr + (read_group << 2));
}

static void scc_mgr_set_dq_out1_delay(uint32_t write_group,
				      uint32_t dq_in_group, uint32_t delay)
{
	uint32_t addr = sdr_get_addr((u32 *)SCC_MGR_IO_OUT1_DELAY);

	/* Load the setting in the SCC manager */
	writel(delay, SOCFPGA_SDR_ADDRESS + addr + (dq_in_group << 2));
}

static void scc_mgr_set_dq_in_delay(uint32_t write_group,
	uint32_t dq_in_group, uint32_t delay)
{
	uint32_t addr = sdr_get_addr((u32 *)SCC_MGR_IO_IN_DELAY);

	/* Load the setting in the SCC manager */
	writel(delay, SOCFPGA_SDR_ADDRESS + addr + (dq_in_group << 2));
}

static void scc_mgr_set_hhp_extras(void)
{
	/*
	 * Load the fixed setting in the SCC manager
	 * bits: 0:0 = 1'b1   - dqs bypass
	 * bits: 1:1 = 1'b1   - dq bypass
	 * bits: 4:2 = 3'b001   - rfifo_mode
	 * bits: 6:5 = 2'b01  - rfifo clock_select
	 * bits: 7:7 = 1'b0  - separate gating from ungating setting
	 * bits: 8:8 = 1'b0  - separate OE from Output delay setting
	 */
	uint32_t value = (0<<8) | (0<<7) | (1<<5) | (1<<2) | (1<<1) | (1<<0);
	uint32_t addr = sdr_get_addr((u32 *)SCC_MGR_HHP_GLOBALS);

	writel(value, SOCFPGA_SDR_ADDRESS + addr + SCC_MGR_HHP_EXTRAS_OFFSET);
}

static void scc_mgr_set_dqs_out1_delay(uint32_t write_group,
					      uint32_t delay)
{
	uint32_t addr = sdr_get_addr((u32 *)SCC_MGR_IO_OUT1_DELAY);

	/* Load the setting in the SCC manager */
	writel(delay, SOCFPGA_SDR_ADDRESS + addr + (RW_MGR_MEM_DQ_PER_WRITE_DQS << 2));
}

static void scc_mgr_set_dm_out1_delay(uint32_t write_group,
					     uint32_t dm, uint32_t delay)
{
	uint32_t addr = sdr_get_addr((u32 *)SCC_MGR_IO_OUT1_DELAY);

	/* Load the setting in the SCC manager */
	writel(delay, SOCFPGA_SDR_ADDRESS + addr +
		((RW_MGR_MEM_DQ_PER_WRITE_DQS + 1 + dm) << 2));
}

/*
 * USER Zero all DQS config
 * TODO: maybe rename to scc_mgr_zero_dqs_config (or something)
 */
static void scc_mgr_zero_all(void)
{
	uint32_t i, r;
	uint32_t addr;

	/*
	 * USER Zero all DQS config settings, across all groups and all
	 * shadow registers
	 */
	for (r = 0; r < RW_MGR_MEM_NUMBER_OF_RANKS; r +=
	     NUM_RANKS_PER_SHADOW_REG) {
		for (i = 0; i < RW_MGR_MEM_IF_READ_DQS_WIDTH; i++) {
			/*
			 * The phases actually don't exist on a per-rank basis,
			 * but there's no harm updating them several times, so
			 * let's keep the code simple.
			 */
			scc_mgr_set_dqs_bus_in_delay(i, IO_DQS_IN_RESERVE);
			scc_mgr_set_dqs_en_phase(i, 0);
			scc_mgr_set_dqs_en_delay(i, 0);
		}

		for (i = 0; i < RW_MGR_MEM_IF_WRITE_DQS_WIDTH; i++) {
			scc_mgr_set_dqdqs_output_phase(i, 0);
			/* av/cv don't have out2 */
			scc_mgr_set_oct_out1_delay(i, IO_DQS_OUT_RESERVE);
		}
	}

	/* multicast to all DQS group enables */
	addr = sdr_get_addr(&sdr_scc_mgr->dqs_ena);
	writel(0xff, SOCFPGA_SDR_ADDRESS + addr);

	addr = sdr_get_addr(&sdr_scc_mgr->update);
	writel(0, SOCFPGA_SDR_ADDRESS + addr);
}

static void scc_set_bypass_mode(uint32_t write_group, uint32_t mode)
{
	uint32_t addr;
	/* mode = 0 : Do NOT bypass - Half Rate Mode */
	/* mode = 1 : Bypass - Full Rate Mode */

	/* only need to set once for all groups, pins, dq, dqs, dm */
	if (write_group == 0) {
		debug_cond(DLEVEL == 1, "%s:%d Setting HHP Extras\n", __func__,
			   __LINE__);
		scc_mgr_set_hhp_extras();
		debug_cond(DLEVEL == 1, "%s:%d Done Setting HHP Extras\n",
			  __func__, __LINE__);
	}
	/* multicast to all DQ enables */
	addr = sdr_get_addr(&sdr_scc_mgr->dq_ena);
	writel(0xff, SOCFPGA_SDR_ADDRESS + addr);

	addr = sdr_get_addr(&sdr_scc_mgr->dm_ena);
	writel(0xff, SOCFPGA_SDR_ADDRESS + addr);

	/* update current DQS IO enable */
	addr = sdr_get_addr(&sdr_scc_mgr->dqs_io_ena);
	writel(0, SOCFPGA_SDR_ADDRESS + addr);

	/* update the DQS logic */
	addr = sdr_get_addr(&sdr_scc_mgr->dqs_ena);
	writel(write_group, SOCFPGA_SDR_ADDRESS + addr);

	/* hit update */
	addr = sdr_get_addr(&sdr_scc_mgr->update);
	writel(0, SOCFPGA_SDR_ADDRESS + addr);
}

static void scc_mgr_zero_group(uint32_t write_group, uint32_t test_begin,
			       int32_t out_only)
{
	uint32_t i, r;
	uint32_t addr;

	for (r = 0; r < RW_MGR_MEM_NUMBER_OF_RANKS; r +=
		NUM_RANKS_PER_SHADOW_REG) {
		/* Zero all DQ config settings */
		for (i = 0; i < RW_MGR_MEM_DQ_PER_WRITE_DQS; i++) {
			scc_mgr_set_dq_out1_delay(write_group, i, 0);
			if (!out_only)
				scc_mgr_set_dq_in_delay(write_group, i, 0);
		}

		/* multicast to all DQ enables */
		addr = sdr_get_addr(&sdr_scc_mgr->dq_ena);
		writel(0xff, SOCFPGA_SDR_ADDRESS + addr);

		/* Zero all DM config settings */
		for (i = 0; i < RW_MGR_NUM_DM_PER_WRITE_GROUP; i++) {
			scc_mgr_set_dm_out1_delay(write_group, i, 0);
		}

		/* multicast to all DM enables */
		addr = sdr_get_addr(&sdr_scc_mgr->dm_ena);
		writel(0xff, SOCFPGA_SDR_ADDRESS + addr);

		/* zero all DQS io settings */
		if (!out_only)
			scc_mgr_set_dqs_io_in_delay(write_group, 0);
		/* av/cv don't have out2 */
		scc_mgr_set_dqs_out1_delay(write_group, IO_DQS_OUT_RESERVE);
		scc_mgr_set_oct_out1_delay(write_group, IO_DQS_OUT_RESERVE);
		scc_mgr_load_dqs_for_write_group(write_group);

		/* multicast to all DQS IO enables (only 1) */
		addr = sdr_get_addr(&sdr_scc_mgr->dqs_io_ena);
		writel(0, SOCFPGA_SDR_ADDRESS + addr);

		/* hit update to zero everything */
		addr = sdr_get_addr(&sdr_scc_mgr->update);
		writel(0, SOCFPGA_SDR_ADDRESS + addr);
	}
}

/* load up dqs config settings */
static void scc_mgr_load_dqs(uint32_t dqs)
{
	uint32_t addr = sdr_get_addr(&sdr_scc_mgr->dqs_ena);

	writel(dqs, SOCFPGA_SDR_ADDRESS + addr);
}

static void scc_mgr_load_dqs_for_write_group(uint32_t write_group)
{
	uint32_t read_group;
	uint32_t addr = sdr_get_addr(&sdr_scc_mgr->dqs_ena);
	/*
	 * Although OCT affects only write data, the OCT delay is controlled
	 * by the DQS logic block which is instantiated once per read group.
	 * For protocols where a write group consists of multiple read groups,
	 * the setting must be scanned multiple times.
	 */
	for (read_group = write_group * RW_MGR_MEM_IF_READ_DQS_WIDTH /
	     RW_MGR_MEM_IF_WRITE_DQS_WIDTH;
	     read_group < (write_group + 1) * RW_MGR_MEM_IF_READ_DQS_WIDTH /
	     RW_MGR_MEM_IF_WRITE_DQS_WIDTH; ++read_group)
		writel(read_group, SOCFPGA_SDR_ADDRESS + addr);
}

/* load up dqs io config settings */
static void scc_mgr_load_dqs_io(void)
{
	uint32_t addr = sdr_get_addr(&sdr_scc_mgr->dqs_io_ena);

	writel(0, SOCFPGA_SDR_ADDRESS + addr);
}

/* load up dq config settings */
static void scc_mgr_load_dq(uint32_t dq_in_group)
{
	uint32_t addr = sdr_get_addr(&sdr_scc_mgr->dq_ena);

	writel(dq_in_group, SOCFPGA_SDR_ADDRESS + addr);
}

/* load up dm config settings */
static void scc_mgr_load_dm(uint32_t dm)
{
	uint32_t addr = sdr_get_addr(&sdr_scc_mgr->dm_ena);

	writel(dm, SOCFPGA_SDR_ADDRESS + addr);
}

/*
 * apply and load a particular input delay for the DQ pins in a group
 * group_bgn is the index of the first dq pin (in the write group)
 */
static void scc_mgr_apply_group_dq_in_delay(uint32_t write_group,
					    uint32_t group_bgn, uint32_t delay)
{
	uint32_t i, p;

	for (i = 0, p = group_bgn; i < RW_MGR_MEM_DQ_PER_READ_DQS; i++, p++) {
		scc_mgr_set_dq_in_delay(write_group, p, delay);
		scc_mgr_load_dq(p);
	}
}

/* apply and load a particular output delay for the DQ pins in a group */
static void scc_mgr_apply_group_dq_out1_delay(uint32_t write_group,
					      uint32_t group_bgn,
					      uint32_t delay1)
{
	uint32_t i, p;

	for (i = 0, p = group_bgn; i < RW_MGR_MEM_DQ_PER_WRITE_DQS; i++, p++) {
		scc_mgr_set_dq_out1_delay(write_group, i, delay1);
		scc_mgr_load_dq(i);
	}
}

/* apply and load a particular output delay for the DM pins in a group */
static void scc_mgr_apply_group_dm_out1_delay(uint32_t write_group,
					      uint32_t delay1)
{
	uint32_t i;

	for (i = 0; i < RW_MGR_NUM_DM_PER_WRITE_GROUP; i++) {
		scc_mgr_set_dm_out1_delay(write_group, i, delay1);
		scc_mgr_load_dm(i);
	}
}


/* apply and load delay on both DQS and OCT out1 */
static void scc_mgr_apply_group_dqs_io_and_oct_out1(uint32_t write_group,
						    uint32_t delay)
{
	scc_mgr_set_dqs_out1_delay(write_group, delay);
	scc_mgr_load_dqs_io();

	scc_mgr_set_oct_out1_delay(write_group, delay);
	scc_mgr_load_dqs_for_write_group(write_group);
}

/* apply a delay to the entire output side: DQ, DM, DQS, OCT */
static void scc_mgr_apply_group_all_out_delay_add(uint32_t write_group,
						  uint32_t group_bgn,
						  uint32_t delay)
{
	uint32_t i, p, new_delay;

	/* dq shift */
	for (i = 0, p = group_bgn; i < RW_MGR_MEM_DQ_PER_WRITE_DQS; i++, p++) {
		new_delay = READ_SCC_DQ_OUT2_DELAY;
		new_delay += delay;

		if (new_delay > IO_IO_OUT2_DELAY_MAX) {
			debug_cond(DLEVEL == 1, "%s:%d (%u, %u, %u) DQ[%u,%u]:\
				   %u > %lu => %lu", __func__, __LINE__,
				   write_group, group_bgn, delay, i, p, new_delay,
				   (long unsigned int)IO_IO_OUT2_DELAY_MAX,
				   (long unsigned int)IO_IO_OUT2_DELAY_MAX);
			new_delay = IO_IO_OUT2_DELAY_MAX;
		}

		scc_mgr_load_dq(i);
	}

	/* dm shift */
	for (i = 0; i < RW_MGR_NUM_DM_PER_WRITE_GROUP; i++) {
		new_delay = READ_SCC_DM_IO_OUT2_DELAY;
		new_delay += delay;

		if (new_delay > IO_IO_OUT2_DELAY_MAX) {
			debug_cond(DLEVEL == 1, "%s:%d (%u, %u, %u) DM[%u]:\
				   %u > %lu => %lu\n",  __func__, __LINE__,
				   write_group, group_bgn, delay, i, new_delay,
				   (long unsigned int)IO_IO_OUT2_DELAY_MAX,
				   (long unsigned int)IO_IO_OUT2_DELAY_MAX);
			new_delay = IO_IO_OUT2_DELAY_MAX;
		}

		scc_mgr_load_dm(i);
	}

	/* dqs shift */
	new_delay = READ_SCC_DQS_IO_OUT2_DELAY;
	new_delay += delay;

	if (new_delay > IO_IO_OUT2_DELAY_MAX) {
		debug_cond(DLEVEL == 1, "%s:%d (%u, %u, %u) DQS: %u > %d => %d;"
			   " adding %u to OUT1\n", __func__, __LINE__,
			   write_group, group_bgn, delay, new_delay,
			   IO_IO_OUT2_DELAY_MAX, IO_IO_OUT2_DELAY_MAX,
			   new_delay - IO_IO_OUT2_DELAY_MAX);
		scc_mgr_set_dqs_out1_delay(write_group, new_delay -
					   IO_IO_OUT2_DELAY_MAX);
		new_delay = IO_IO_OUT2_DELAY_MAX;
	}

	scc_mgr_load_dqs_io();

	/* oct shift */
	new_delay = READ_SCC_OCT_OUT2_DELAY;
	new_delay += delay;

	if (new_delay > IO_IO_OUT2_DELAY_MAX) {
		debug_cond(DLEVEL == 1, "%s:%d (%u, %u, %u) DQS: %u > %d => %d;"
			   " adding %u to OUT1\n", __func__, __LINE__,
			   write_group, group_bgn, delay, new_delay,
			   IO_IO_OUT2_DELAY_MAX, IO_IO_OUT2_DELAY_MAX,
			   new_delay - IO_IO_OUT2_DELAY_MAX);
		scc_mgr_set_oct_out1_delay(write_group, new_delay -
					   IO_IO_OUT2_DELAY_MAX);
		new_delay = IO_IO_OUT2_DELAY_MAX;
	}

	scc_mgr_load_dqs_for_write_group(write_group);
}

/*
 * USER apply a delay to the entire output side (DQ, DM, DQS, OCT)
 * and to all ranks
 */
static void scc_mgr_apply_group_all_out_delay_add_all_ranks(
	uint32_t write_group, uint32_t group_bgn, uint32_t delay)
{
	uint32_t r;
	uint32_t addr = sdr_get_addr(&sdr_scc_mgr->update);

	for (r = 0; r < RW_MGR_MEM_NUMBER_OF_RANKS;
		r += NUM_RANKS_PER_SHADOW_REG) {
		scc_mgr_apply_group_all_out_delay_add(write_group,
						      group_bgn, delay);
		writel(0, SOCFPGA_SDR_ADDRESS + addr);
	}
}

/* optimization used to recover some slots in ddr3 inst_rom */
/* could be applied to other protocols if we wanted to */
static void set_jump_as_return(void)
{
	uint32_t addr = sdr_get_addr(&sdr_rw_load_mgr_regs->load_cntr0);

	/*
	 * to save space, we replace return with jump to special shared
	 * RETURN instruction so we set the counter to large value so that
	 * we always jump
	 */
	writel(0xff, SOCFPGA_SDR_ADDRESS + addr);
	addr = sdr_get_addr(&sdr_rw_load_jump_mgr_regs->load_jump_add0);
	writel(RW_MGR_RETURN, SOCFPGA_SDR_ADDRESS + addr);
}

/*
 * should always use constants as argument to ensure all computations are
 * performed at compile time
 */
static void delay_for_n_mem_clocks(const uint32_t clocks)
{
	uint32_t afi_clocks;
	uint8_t inner = 0;
	uint8_t outer = 0;
	uint16_t c_loop = 0;
	uint32_t addr;

	debug("%s:%d: clocks=%u ... start\n", __func__, __LINE__, clocks);


	afi_clocks = (clocks + AFI_RATE_RATIO-1) / AFI_RATE_RATIO;
	/* scale (rounding up) to get afi clocks */

	/*
	 * Note, we don't bother accounting for being off a little bit
	 * because of a few extra instructions in outer loops
	 * Note, the loops have a test at the end, and do the test before
	 * the decrement, and so always perform the loop
	 * 1 time more than the counter value
	 */
	if (afi_clocks == 0) {
		;
	} else if (afi_clocks <= 0x100) {
		inner = afi_clocks-1;
		outer = 0;
		c_loop = 0;
	} else if (afi_clocks <= 0x10000) {
		inner = 0xff;
		outer = (afi_clocks-1) >> 8;
		c_loop = 0;
	} else {
		inner = 0xff;
		outer = 0xff;
		c_loop = (afi_clocks-1) >> 16;
	}

	/*
	 * rom instructions are structured as follows:
	 *
	 *    IDLE_LOOP2: jnz cntr0, TARGET_A
	 *    IDLE_LOOP1: jnz cntr1, TARGET_B
	 *                return
	 *
	 * so, when doing nested loops, TARGET_A is set to IDLE_LOOP2, and
	 * TARGET_B is set to IDLE_LOOP2 as well
	 *
	 * if we have no outer loop, though, then we can use IDLE_LOOP1 only,
	 * and set TARGET_B to IDLE_LOOP1 and we skip IDLE_LOOP2 entirely
	 *
	 * a little confusing, but it helps save precious space in the inst_rom
	 * and sequencer rom and keeps the delays more accurate and reduces
	 * overhead
	 */
	if (afi_clocks <= 0x100) {
		addr = sdr_get_addr(&sdr_rw_load_mgr_regs->load_cntr1);
		writel(SKIP_DELAY_LOOP_VALUE_OR_ZERO(inner), SOCFPGA_SDR_ADDRESS + addr);

		addr = sdr_get_addr(&sdr_rw_load_jump_mgr_regs->load_jump_add1);
		writel(RW_MGR_IDLE_LOOP1, SOCFPGA_SDR_ADDRESS + addr);

		addr = sdr_get_addr((u32 *)RW_MGR_RUN_SINGLE_GROUP);
		writel(RW_MGR_IDLE_LOOP1, SOCFPGA_SDR_ADDRESS + addr);
	} else {
		addr = sdr_get_addr(&sdr_rw_load_mgr_regs->load_cntr0);
		writel(SKIP_DELAY_LOOP_VALUE_OR_ZERO(inner), SOCFPGA_SDR_ADDRESS + addr);

		addr = sdr_get_addr(&sdr_rw_load_mgr_regs->load_cntr1);
		writel(SKIP_DELAY_LOOP_VALUE_OR_ZERO(outer), SOCFPGA_SDR_ADDRESS + addr);

		addr = sdr_get_addr(&sdr_rw_load_jump_mgr_regs->load_jump_add0);
		writel(RW_MGR_IDLE_LOOP2, SOCFPGA_SDR_ADDRESS + addr);

		addr = sdr_get_addr(&sdr_rw_load_jump_mgr_regs->load_jump_add1);
		writel(RW_MGR_IDLE_LOOP2, SOCFPGA_SDR_ADDRESS + addr);

		/* hack to get around compiler not being smart enough */
		if (afi_clocks <= 0x10000) {
			/* only need to run once */
			addr = sdr_get_addr((u32 *)RW_MGR_RUN_SINGLE_GROUP);
			writel(RW_MGR_IDLE_LOOP2, SOCFPGA_SDR_ADDRESS + addr);
		} else {
			do {
				addr = sdr_get_addr((u32 *)RW_MGR_RUN_SINGLE_GROUP);
				writel(RW_MGR_IDLE_LOOP2, SOCFPGA_SDR_ADDRESS + addr);
			} while (c_loop-- != 0);
		}
	}
	debug("%s:%d clocks=%u ... end\n", __func__, __LINE__, clocks);
}

static void rw_mgr_mem_initialize(void)
{
	uint32_t r;
	uint32_t addr;

	debug("%s:%d\n", __func__, __LINE__);

	/* The reset / cke part of initialization is broadcasted to all ranks */
	addr = sdr_get_addr((u32 *)RW_MGR_SET_CS_AND_ODT_MASK);
	writel(RW_MGR_RANK_ALL, SOCFPGA_SDR_ADDRESS + addr);

	/*
	 * Here's how you load register for a loop
	 * Counters are located @ 0x800
	 * Jump address are located @ 0xC00
	 * For both, registers 0 to 3 are selected using bits 3 and 2, like
	 * in 0x800, 0x804, 0x808, 0x80C and 0xC00, 0xC04, 0xC08, 0xC0C
	 * I know this ain't pretty, but Avalon bus throws away the 2 least
	 * significant bits
	 */

	/* start with memory RESET activated */

	/* tINIT = 200us */

	/*
	 * 200us @ 266MHz (3.75 ns) ~ 54000 clock cycles
	 * If a and b are the number of iteration in 2 nested loops
	 * it takes the following number of cycles to complete the operation:
	 * number_of_cycles = ((2 + n) * a + 2) * b
	 * where n is the number of instruction in the inner loop
	 * One possible solution is n = 0 , a = 256 , b = 106 => a = FF,
	 * b = 6A
	 */

	/* Load counters */
	addr = sdr_get_addr(&sdr_rw_load_mgr_regs->load_cntr0);
	writel(SKIP_DELAY_LOOP_VALUE_OR_ZERO(SEQ_TINIT_CNTR0_VAL),
	       SOCFPGA_SDR_ADDRESS + addr);
	addr = sdr_get_addr(&sdr_rw_load_mgr_regs->load_cntr1);
	writel(SKIP_DELAY_LOOP_VALUE_OR_ZERO(SEQ_TINIT_CNTR1_VAL),
	       SOCFPGA_SDR_ADDRESS + addr);
	addr = sdr_get_addr(&sdr_rw_load_mgr_regs->load_cntr2);
	writel(SKIP_DELAY_LOOP_VALUE_OR_ZERO(SEQ_TINIT_CNTR2_VAL),
	       SOCFPGA_SDR_ADDRESS + addr);

	/* Load jump address */
	addr = sdr_get_addr(&sdr_rw_load_jump_mgr_regs->load_jump_add0);
	writel(RW_MGR_INIT_RESET_0_CKE_0, SOCFPGA_SDR_ADDRESS + addr);

	addr = sdr_get_addr(&sdr_rw_load_jump_mgr_regs->load_jump_add1);
	writel(RW_MGR_INIT_RESET_0_CKE_0, SOCFPGA_SDR_ADDRESS + addr);

	addr = sdr_get_addr(&sdr_rw_load_jump_mgr_regs->load_jump_add2);
	writel(RW_MGR_INIT_RESET_0_CKE_0, SOCFPGA_SDR_ADDRESS + addr);

	/* Execute count instruction */
	addr = sdr_get_addr((u32 *)RW_MGR_RUN_SINGLE_GROUP);
	writel(RW_MGR_INIT_RESET_0_CKE_0, SOCFPGA_SDR_ADDRESS + addr);

	/* indicate that memory is stable */
	addr = sdr_get_addr(&phy_mgr_cfg->reset_mem_stbl);
	writel(1, SOCFPGA_SDR_ADDRESS + addr);

	/*
	 * transition the RESET to high
	 * Wait for 500us
	 */

	/*
	 * 500us @ 266MHz (3.75 ns) ~ 134000 clock cycles
	 * If a and b are the number of iteration in 2 nested loops
	 * it takes the following number of cycles to complete the operation
	 * number_of_cycles = ((2 + n) * a + 2) * b
	 * where n is the number of instruction in the inner loop
	 * One possible solution is n = 2 , a = 131 , b = 256 => a = 83,
	 * b = FF
	 */

	/* Load counters */
	addr = sdr_get_addr(&sdr_rw_load_mgr_regs->load_cntr0);
	writel(SKIP_DELAY_LOOP_VALUE_OR_ZERO(SEQ_TRESET_CNTR0_VAL),
	       SOCFPGA_SDR_ADDRESS + addr);
	addr = sdr_get_addr(&sdr_rw_load_mgr_regs->load_cntr1);
	writel(SKIP_DELAY_LOOP_VALUE_OR_ZERO(SEQ_TRESET_CNTR1_VAL),
	       SOCFPGA_SDR_ADDRESS + addr);
	addr = sdr_get_addr(&sdr_rw_load_mgr_regs->load_cntr2);
	writel(SKIP_DELAY_LOOP_VALUE_OR_ZERO(SEQ_TRESET_CNTR2_VAL),
	       SOCFPGA_SDR_ADDRESS + addr);

	/* Load jump address */
	addr = sdr_get_addr(&sdr_rw_load_jump_mgr_regs->load_jump_add0);
	writel(RW_MGR_INIT_RESET_1_CKE_0, SOCFPGA_SDR_ADDRESS + addr);
	addr = sdr_get_addr(&sdr_rw_load_jump_mgr_regs->load_jump_add1);
	writel(RW_MGR_INIT_RESET_1_CKE_0, SOCFPGA_SDR_ADDRESS + addr);
	addr = sdr_get_addr(&sdr_rw_load_jump_mgr_regs->load_jump_add2);
	writel(RW_MGR_INIT_RESET_1_CKE_0, SOCFPGA_SDR_ADDRESS + addr);

	addr = sdr_get_addr((u32 *)RW_MGR_RUN_SINGLE_GROUP);
	writel(RW_MGR_INIT_RESET_1_CKE_0, SOCFPGA_SDR_ADDRESS + addr);

	/* bring up clock enable */

	/* tXRP < 250 ck cycles */
	delay_for_n_mem_clocks(250);

	for (r = 0; r < RW_MGR_MEM_NUMBER_OF_RANKS; r++) {
		if (param->skip_ranks[r]) {
			/* request to skip the rank */
			continue;
		}

		/* set rank */
		set_rank_and_odt_mask(r, RW_MGR_ODT_MODE_OFF);

		/*
		 * USER Use Mirror-ed commands for odd ranks if address
		 * mirrorring is on
		 */
		if ((RW_MGR_MEM_ADDRESS_MIRRORING >> r) & 0x1) {
			set_jump_as_return();
			addr = sdr_get_addr((u32 *)RW_MGR_RUN_SINGLE_GROUP);
			writel(RW_MGR_MRS2_MIRR, SOCFPGA_SDR_ADDRESS + addr);
			delay_for_n_mem_clocks(4);
			set_jump_as_return();
			writel(RW_MGR_MRS3_MIRR, SOCFPGA_SDR_ADDRESS + addr);
			delay_for_n_mem_clocks(4);
			set_jump_as_return();
			writel(RW_MGR_MRS1_MIRR, SOCFPGA_SDR_ADDRESS + addr);
			delay_for_n_mem_clocks(4);
			set_jump_as_return();
			writel(RW_MGR_MRS0_DLL_RESET_MIRR, SOCFPGA_SDR_ADDRESS + addr);
		} else {
			set_jump_as_return();
			addr = sdr_get_addr((u32 *)RW_MGR_RUN_SINGLE_GROUP);
			writel(RW_MGR_MRS2, SOCFPGA_SDR_ADDRESS + addr);
			delay_for_n_mem_clocks(4);
			set_jump_as_return();
			writel(RW_MGR_MRS3, SOCFPGA_SDR_ADDRESS + addr);
			delay_for_n_mem_clocks(4);
			set_jump_as_return();
			writel(RW_MGR_MRS1, SOCFPGA_SDR_ADDRESS + addr);
			set_jump_as_return();
			writel(RW_MGR_MRS0_DLL_RESET, SOCFPGA_SDR_ADDRESS + addr);
		}
		set_jump_as_return();
		addr = sdr_get_addr((u32 *)RW_MGR_RUN_SINGLE_GROUP);
		writel(RW_MGR_ZQCL, SOCFPGA_SDR_ADDRESS + addr);

		/* tZQinit = tDLLK = 512 ck cycles */
		delay_for_n_mem_clocks(512);
	}
}

/*
 * At the end of calibration we have to program the user settings in, and
 * USER  hand off the memory to the user.
 */
static void rw_mgr_mem_handoff(void)
{
	uint32_t r;
	uint32_t addr;

	debug("%s:%d\n", __func__, __LINE__);
	for (r = 0; r < RW_MGR_MEM_NUMBER_OF_RANKS; r++) {
		if (param->skip_ranks[r])
			/* request to skip the rank */
			continue;
		/* set rank */
		set_rank_and_odt_mask(r, RW_MGR_ODT_MODE_OFF);

		/* precharge all banks ... */
		addr = sdr_get_addr((u32 *)RW_MGR_RUN_SINGLE_GROUP);
		writel(RW_MGR_PRECHARGE_ALL, SOCFPGA_SDR_ADDRESS + addr);

		/* load up MR settings specified by user */

		/*
		 * Use Mirror-ed commands for odd ranks if address
		 * mirrorring is on
		 */
		addr = sdr_get_addr((u32 *)RW_MGR_RUN_SINGLE_GROUP);
		if ((RW_MGR_MEM_ADDRESS_MIRRORING >> r) & 0x1) {
			set_jump_as_return();
			writel(RW_MGR_MRS2_MIRR, SOCFPGA_SDR_ADDRESS + addr);
			delay_for_n_mem_clocks(4);
			set_jump_as_return();
			writel(RW_MGR_MRS3_MIRR, SOCFPGA_SDR_ADDRESS + addr);
			delay_for_n_mem_clocks(4);
			set_jump_as_return();
			writel(RW_MGR_MRS1_MIRR, SOCFPGA_SDR_ADDRESS + addr);
			delay_for_n_mem_clocks(4);
			set_jump_as_return();
			writel(RW_MGR_MRS0_USER_MIRR, SOCFPGA_SDR_ADDRESS + addr);
		} else {
			set_jump_as_return();
			writel(RW_MGR_MRS2, SOCFPGA_SDR_ADDRESS + addr);
			delay_for_n_mem_clocks(4);
			set_jump_as_return();
			writel(RW_MGR_MRS3, SOCFPGA_SDR_ADDRESS + addr);
			delay_for_n_mem_clocks(4);
			set_jump_as_return();
			writel(RW_MGR_MRS1, SOCFPGA_SDR_ADDRESS + addr);
			delay_for_n_mem_clocks(4);
			set_jump_as_return();
			writel(RW_MGR_MRS0_USER, SOCFPGA_SDR_ADDRESS + addr);
		}
		/*
		 * USER  need to wait tMOD (12CK or 15ns) time before issuing
		 * other commands, but we will have plenty of NIOS cycles before
		 * actual handoff so its okay.
		 */
	}
}

/*
 * performs a guaranteed read on the patterns we are going to use during a
 * read test to ensure memory works
 */
static uint32_t rw_mgr_mem_calibrate_read_test_patterns(uint32_t rank_bgn,
	uint32_t group, uint32_t num_tries, uint32_t *bit_chk,
	uint32_t all_ranks)
{
	uint32_t r, vg;
	uint32_t correct_mask_vg;
	uint32_t tmp_bit_chk;
	uint32_t rank_end = all_ranks ? RW_MGR_MEM_NUMBER_OF_RANKS :
		(rank_bgn + NUM_RANKS_PER_SHADOW_REG);
	uint32_t addr;
	uint32_t base_rw_mgr;

	*bit_chk = param->read_correct_mask;
	correct_mask_vg = param->read_correct_mask_vg;

	for (r = rank_bgn; r < rank_end; r++) {
		if (param->skip_ranks[r])
			/* request to skip the rank */
			continue;

		/* set rank */
		set_rank_and_odt_mask(r, RW_MGR_ODT_MODE_READ_WRITE);

		/* Load up a constant bursts of read commands */
		addr = sdr_get_addr(&sdr_rw_load_mgr_regs->load_cntr0);
		writel(0x20, SOCFPGA_SDR_ADDRESS + addr);
		addr = sdr_get_addr(&sdr_rw_load_jump_mgr_regs->load_jump_add0);
		writel(RW_MGR_GUARANTEED_READ, SOCFPGA_SDR_ADDRESS + addr);

		addr = sdr_get_addr(&sdr_rw_load_mgr_regs->load_cntr1);
		writel(0x20, SOCFPGA_SDR_ADDRESS + addr);
		addr = sdr_get_addr(&sdr_rw_load_jump_mgr_regs->load_jump_add1);
		writel(RW_MGR_GUARANTEED_READ_CONT, SOCFPGA_SDR_ADDRESS + addr);

		tmp_bit_chk = 0;
		for (vg = RW_MGR_MEM_VIRTUAL_GROUPS_PER_READ_DQS-1; ; vg--) {
			/* reset the fifos to get pointers to known state */

			addr = sdr_get_addr(&phy_mgr_cmd->fifo_reset);
			writel(0, SOCFPGA_SDR_ADDRESS + addr);
			addr = sdr_get_addr((u32 *)RW_MGR_RESET_READ_DATAPATH);
			writel(0, SOCFPGA_SDR_ADDRESS + addr);

			tmp_bit_chk = tmp_bit_chk << (RW_MGR_MEM_DQ_PER_READ_DQS
				/ RW_MGR_MEM_VIRTUAL_GROUPS_PER_READ_DQS);

			addr = sdr_get_addr((u32 *)RW_MGR_RUN_SINGLE_GROUP);
			writel(RW_MGR_GUARANTEED_READ, SOCFPGA_SDR_ADDRESS + addr +
			       ((group * RW_MGR_MEM_VIRTUAL_GROUPS_PER_READ_DQS +
				vg) << 2));

			addr = sdr_get_addr((u32 *)BASE_RW_MGR);
			base_rw_mgr = readl(SOCFPGA_SDR_ADDRESS + addr);
			tmp_bit_chk = tmp_bit_chk | (correct_mask_vg & (~base_rw_mgr));

			if (vg == 0)
				break;
		}
		*bit_chk &= tmp_bit_chk;
	}

	addr = sdr_get_addr((u32 *)RW_MGR_RUN_SINGLE_GROUP);
	writel(RW_MGR_CLEAR_DQS_ENABLE, SOCFPGA_SDR_ADDRESS + addr + (group << 2));

	set_rank_and_odt_mask(0, RW_MGR_ODT_MODE_OFF);
	debug_cond(DLEVEL == 1, "%s:%d test_load_patterns(%u,ALL) => (%u == %u) =>\
		   %lu\n", __func__, __LINE__, group, *bit_chk, param->read_correct_mask,
		   (long unsigned int)(*bit_chk == param->read_correct_mask));
	return *bit_chk == param->read_correct_mask;
}

static uint32_t rw_mgr_mem_calibrate_read_test_patterns_all_ranks
	(uint32_t group, uint32_t num_tries, uint32_t *bit_chk)
{
	return rw_mgr_mem_calibrate_read_test_patterns(0, group,
		num_tries, bit_chk, 1);
}

/* load up the patterns we are going to use during a read test */
static void rw_mgr_mem_calibrate_read_load_patterns(uint32_t rank_bgn,
	uint32_t all_ranks)
{
	uint32_t r;
	uint32_t addr;
	uint32_t rank_end = all_ranks ? RW_MGR_MEM_NUMBER_OF_RANKS :
		(rank_bgn + NUM_RANKS_PER_SHADOW_REG);

	debug("%s:%d\n", __func__, __LINE__);
	for (r = rank_bgn; r < rank_end; r++) {
		if (param->skip_ranks[r])
			/* request to skip the rank */
			continue;

		/* set rank */
		set_rank_and_odt_mask(r, RW_MGR_ODT_MODE_READ_WRITE);

		/* Load up a constant bursts */
		addr = sdr_get_addr(&sdr_rw_load_mgr_regs->load_cntr0);
		writel(0x20, SOCFPGA_SDR_ADDRESS + addr);

		addr = sdr_get_addr(&sdr_rw_load_jump_mgr_regs->load_jump_add0);
		writel(RW_MGR_GUARANTEED_WRITE_WAIT0, SOCFPGA_SDR_ADDRESS + addr);

		addr = sdr_get_addr(&sdr_rw_load_mgr_regs->load_cntr1);
		writel(0x20, SOCFPGA_SDR_ADDRESS + addr);

		addr = sdr_get_addr(&sdr_rw_load_jump_mgr_regs->load_jump_add1);
		writel(RW_MGR_GUARANTEED_WRITE_WAIT1, SOCFPGA_SDR_ADDRESS + addr);

		addr = sdr_get_addr(&sdr_rw_load_mgr_regs->load_cntr2);
		writel(0x04, SOCFPGA_SDR_ADDRESS + addr);

		addr = sdr_get_addr(&sdr_rw_load_jump_mgr_regs->load_jump_add2);
		writel(RW_MGR_GUARANTEED_WRITE_WAIT2, SOCFPGA_SDR_ADDRESS + addr);

		addr = sdr_get_addr(&sdr_rw_load_mgr_regs->load_cntr3);
		writel(0x04, SOCFPGA_SDR_ADDRESS + addr);

		addr = sdr_get_addr(&sdr_rw_load_jump_mgr_regs->load_jump_add3);
		writel(RW_MGR_GUARANTEED_WRITE_WAIT3, SOCFPGA_SDR_ADDRESS + addr);

		addr = sdr_get_addr((u32 *)RW_MGR_RUN_SINGLE_GROUP);
		writel(RW_MGR_GUARANTEED_WRITE, SOCFPGA_SDR_ADDRESS + addr);
	}

	set_rank_and_odt_mask(0, RW_MGR_ODT_MODE_OFF);
}

/*
 * try a read and see if it returns correct data back. has dummy reads
 * inserted into the mix used to align dqs enable. has more thorough checks
 * than the regular read test.
 */
static uint32_t rw_mgr_mem_calibrate_read_test(uint32_t rank_bgn, uint32_t group,
	uint32_t num_tries, uint32_t all_correct, uint32_t *bit_chk,
	uint32_t all_groups, uint32_t all_ranks)
{
	uint32_t r, vg;
	uint32_t correct_mask_vg;
	uint32_t tmp_bit_chk;
	uint32_t rank_end = all_ranks ? RW_MGR_MEM_NUMBER_OF_RANKS :
		(rank_bgn + NUM_RANKS_PER_SHADOW_REG);
	uint32_t addr;
	uint32_t base_rw_mgr;

	*bit_chk = param->read_correct_mask;
	correct_mask_vg = param->read_correct_mask_vg;

	uint32_t quick_read_mode = (((STATIC_CALIB_STEPS) &
		CALIB_SKIP_DELAY_SWEEPS) && ENABLE_SUPER_QUICK_CALIBRATION);

	for (r = rank_bgn; r < rank_end; r++) {
		if (param->skip_ranks[r])
			/* request to skip the rank */
			continue;

		/* set rank */
		set_rank_and_odt_mask(r, RW_MGR_ODT_MODE_READ_WRITE);

		addr = sdr_get_addr(&sdr_rw_load_mgr_regs->load_cntr1);
		writel(0x10, SOCFPGA_SDR_ADDRESS + addr);

		addr = sdr_get_addr(&sdr_rw_load_jump_mgr_regs->load_jump_add1);
		writel(RW_MGR_READ_B2B_WAIT1, SOCFPGA_SDR_ADDRESS + addr);

		addr = sdr_get_addr(&sdr_rw_load_mgr_regs->load_cntr2);
		writel(0x10, SOCFPGA_SDR_ADDRESS + addr);
		addr = sdr_get_addr(&sdr_rw_load_jump_mgr_regs->load_jump_add2);
		writel(RW_MGR_READ_B2B_WAIT2, SOCFPGA_SDR_ADDRESS + addr);

		addr = sdr_get_addr(&sdr_rw_load_mgr_regs->load_cntr0);
		if (quick_read_mode)
			writel(0x1, SOCFPGA_SDR_ADDRESS + addr);
			/* need at least two (1+1) reads to capture failures */
		else if (all_groups)
			writel(0x06, SOCFPGA_SDR_ADDRESS + addr);
		else
			writel(0x32, SOCFPGA_SDR_ADDRESS + addr);

		addr = sdr_get_addr(&sdr_rw_load_jump_mgr_regs->load_jump_add0);
		writel(RW_MGR_READ_B2B, SOCFPGA_SDR_ADDRESS + addr);
		addr = sdr_get_addr(&sdr_rw_load_mgr_regs->load_cntr3);
		if (all_groups)
			writel(RW_MGR_MEM_IF_READ_DQS_WIDTH *
			       RW_MGR_MEM_VIRTUAL_GROUPS_PER_READ_DQS - 1,
			       SOCFPGA_SDR_ADDRESS + addr);
		else
			writel(0x0, SOCFPGA_SDR_ADDRESS + addr);

		addr = sdr_get_addr(&sdr_rw_load_jump_mgr_regs->load_jump_add3);
		writel(RW_MGR_READ_B2B, SOCFPGA_SDR_ADDRESS + addr);

		tmp_bit_chk = 0;
		for (vg = RW_MGR_MEM_VIRTUAL_GROUPS_PER_READ_DQS-1; ; vg--) {
			/* reset the fifos to get pointers to known state */
			addr = sdr_get_addr(&phy_mgr_cmd->fifo_reset);
			writel(0, SOCFPGA_SDR_ADDRESS + addr);
			addr = sdr_get_addr((u32 *)RW_MGR_RESET_READ_DATAPATH);
			writel(0, SOCFPGA_SDR_ADDRESS + addr);

			tmp_bit_chk = tmp_bit_chk << (RW_MGR_MEM_DQ_PER_READ_DQS
				/ RW_MGR_MEM_VIRTUAL_GROUPS_PER_READ_DQS);

			addr = sdr_get_addr((u32 *)(all_groups ? RW_MGR_RUN_ALL_GROUPS :
					    RW_MGR_RUN_SINGLE_GROUP));
			writel(RW_MGR_READ_B2B, SOCFPGA_SDR_ADDRESS + addr +
			       ((group * RW_MGR_MEM_VIRTUAL_GROUPS_PER_READ_DQS +
			       vg) << 2));

			addr = sdr_get_addr((u32 *)BASE_RW_MGR);
			base_rw_mgr = readl(SOCFPGA_SDR_ADDRESS + addr);
			tmp_bit_chk = tmp_bit_chk | (correct_mask_vg & ~(base_rw_mgr));

			if (vg == 0)
				break;
		}
		*bit_chk &= tmp_bit_chk;
	}

	addr = sdr_get_addr((u32 *)RW_MGR_RUN_SINGLE_GROUP);
	writel(RW_MGR_CLEAR_DQS_ENABLE, SOCFPGA_SDR_ADDRESS + addr + (group << 2));

	if (all_correct) {
		set_rank_and_odt_mask(0, RW_MGR_ODT_MODE_OFF);
		debug_cond(DLEVEL == 2, "%s:%d read_test(%u,ALL,%u) =>\
			   (%u == %u) => %lu", __func__, __LINE__, group,
			   all_groups, *bit_chk, param->read_correct_mask,
			   (long unsigned int)(*bit_chk ==
			   param->read_correct_mask));
		return *bit_chk == param->read_correct_mask;
	} else	{
		set_rank_and_odt_mask(0, RW_MGR_ODT_MODE_OFF);
		debug_cond(DLEVEL == 2, "%s:%d read_test(%u,ONE,%u) =>\
			   (%u != %lu) => %lu\n", __func__, __LINE__,
			   group, all_groups, *bit_chk, (long unsigned int)0,
			   (long unsigned int)(*bit_chk != 0x00));
		return *bit_chk != 0x00;
	}
}

static uint32_t rw_mgr_mem_calibrate_read_test_all_ranks(uint32_t group,
	uint32_t num_tries, uint32_t all_correct, uint32_t *bit_chk,
	uint32_t all_groups)
{
	return rw_mgr_mem_calibrate_read_test(0, group, num_tries, all_correct,
					      bit_chk, all_groups, 1);
}

static void rw_mgr_incr_vfifo(uint32_t grp, uint32_t *v)
{
	uint32_t addr = sdr_get_addr(&phy_mgr_cmd->inc_vfifo_hard_phy);

	writel(grp, SOCFPGA_SDR_ADDRESS + addr);
	(*v)++;
}

static void rw_mgr_decr_vfifo(uint32_t grp, uint32_t *v)
{
	uint32_t i;

	for (i = 0; i < VFIFO_SIZE-1; i++)
		rw_mgr_incr_vfifo(grp, v);
}

static int find_vfifo_read(uint32_t grp, uint32_t *bit_chk)
{
	uint32_t  v;
	uint32_t fail_cnt = 0;
	uint32_t test_status;

	for (v = 0; v < VFIFO_SIZE; ) {
		debug_cond(DLEVEL == 2, "%s:%d find_dqs_en_phase: vfifo %u\n",
			   __func__, __LINE__, v);
		test_status = rw_mgr_mem_calibrate_read_test_all_ranks
			(grp, 1, PASS_ONE_BIT, bit_chk, 0);
		if (!test_status) {
			fail_cnt++;

			if (fail_cnt == 2)
				break;
		}

		/* fiddle with FIFO */
		rw_mgr_incr_vfifo(grp, &v);
	}

	if (v >= VFIFO_SIZE) {
		/* no failing read found!! Something must have gone wrong */
		debug_cond(DLEVEL == 2, "%s:%d find_dqs_en_phase: vfifo failed\n",
			   __func__, __LINE__);
		return 0;
	} else {
		return v;
	}
}

static int find_working_phase(uint32_t *grp, uint32_t *bit_chk,
			      uint32_t dtaps_per_ptap, uint32_t *work_bgn,
			      uint32_t *v, uint32_t *d, uint32_t *p,
			      uint32_t *i, uint32_t *max_working_cnt)
{
	uint32_t found_begin = 0;
	uint32_t tmp_delay = 0;
	uint32_t test_status;

	for (*d = 0; *d <= dtaps_per_ptap; (*d)++, tmp_delay +=
		IO_DELAY_PER_DQS_EN_DCHAIN_TAP) {
		*work_bgn = tmp_delay;
		scc_mgr_set_dqs_en_delay_all_ranks(*grp, *d);

		for (*i = 0; *i < VFIFO_SIZE; (*i)++) {
			for (*p = 0; *p <= IO_DQS_EN_PHASE_MAX; (*p)++, *work_bgn +=
				IO_DELAY_PER_OPA_TAP) {
				scc_mgr_set_dqs_en_phase_all_ranks(*grp, *p);

				test_status =
				rw_mgr_mem_calibrate_read_test_all_ranks
				(*grp, 1, PASS_ONE_BIT, bit_chk, 0);

				if (test_status) {
					*max_working_cnt = 1;
					found_begin = 1;
					break;
				}
			}

			if (found_begin)
				break;

			if (*p > IO_DQS_EN_PHASE_MAX)
				/* fiddle with FIFO */
				rw_mgr_incr_vfifo(*grp, v);
		}

		if (found_begin)
			break;
	}

	if (*i >= VFIFO_SIZE) {
		/* cannot find working solution */
		debug_cond(DLEVEL == 2, "%s:%d find_dqs_en_phase: no vfifo/\
			   ptap/dtap\n", __func__, __LINE__);
		return 0;
	} else {
		return 1;
	}
}

static void sdr_backup_phase(uint32_t *grp, uint32_t *bit_chk,
			     uint32_t *work_bgn, uint32_t *v, uint32_t *d,
			     uint32_t *p, uint32_t *max_working_cnt)
{
	uint32_t found_begin = 0;
	uint32_t tmp_delay;

	/* Special case code for backing up a phase */
	if (*p == 0) {
		*p = IO_DQS_EN_PHASE_MAX;
		rw_mgr_decr_vfifo(*grp, v);
	} else {
		(*p)--;
	}
	tmp_delay = *work_bgn - IO_DELAY_PER_OPA_TAP;
	scc_mgr_set_dqs_en_phase_all_ranks(*grp, *p);

	for (*d = 0; *d <= IO_DQS_EN_DELAY_MAX && tmp_delay < *work_bgn;
		(*d)++, tmp_delay += IO_DELAY_PER_DQS_EN_DCHAIN_TAP) {
		scc_mgr_set_dqs_en_delay_all_ranks(*grp, *d);

		if (rw_mgr_mem_calibrate_read_test_all_ranks(*grp, 1,
							     PASS_ONE_BIT,
							     bit_chk, 0)) {
			found_begin = 1;
			*work_bgn = tmp_delay;
			break;
		}
	}

	/* We have found a working dtap before the ptap found above */
	if (found_begin == 1)
		(*max_working_cnt)++;

	/*
	 * Restore VFIFO to old state before we decremented it
	 * (if needed).
	 */
	(*p)++;
	if (*p > IO_DQS_EN_PHASE_MAX) {
		*p = 0;
		rw_mgr_incr_vfifo(*grp, v);
	}

	scc_mgr_set_dqs_en_delay_all_ranks(*grp, 0);
}

static int sdr_nonworking_phase(uint32_t *grp, uint32_t *bit_chk,
			     uint32_t *work_bgn, uint32_t *v, uint32_t *d,
			     uint32_t *p, uint32_t *i, uint32_t *max_working_cnt,
			     uint32_t *work_end)
{
	uint32_t found_end = 0;

	(*p)++;
	*work_end += IO_DELAY_PER_OPA_TAP;
	if (*p > IO_DQS_EN_PHASE_MAX) {
		/* fiddle with FIFO */
		*p = 0;
		rw_mgr_incr_vfifo(*grp, v);
	}

	for (; *i < VFIFO_SIZE + 1; (*i)++) {
		for (; *p <= IO_DQS_EN_PHASE_MAX; (*p)++, *work_end
			+= IO_DELAY_PER_OPA_TAP) {
			scc_mgr_set_dqs_en_phase_all_ranks(*grp, *p);

			if (!rw_mgr_mem_calibrate_read_test_all_ranks
				(*grp, 1, PASS_ONE_BIT, bit_chk, 0)) {
				found_end = 1;
				break;
			} else {
				(*max_working_cnt)++;
			}
		}

		if (found_end)
			break;

		if (*p > IO_DQS_EN_PHASE_MAX) {
			/* fiddle with FIFO */
			rw_mgr_incr_vfifo(*grp, v);
			*p = 0;
		}
	}

	if (*i >= VFIFO_SIZE + 1) {
		/* cannot see edge of failing read */
		debug_cond(DLEVEL == 2, "%s:%d sdr_nonworking_phase: end:\
			   failed\n", __func__, __LINE__);
		return 0;
	} else {
		return 1;
	}
}

static int sdr_find_window_centre(uint32_t *grp, uint32_t *bit_chk,
				  uint32_t *work_bgn, uint32_t *v, uint32_t *d,
				  uint32_t *p, uint32_t *work_mid,
				  uint32_t *work_end)
{
	int i;
	int tmp_delay = 0;

	*work_mid = (*work_bgn + *work_end) / 2;

	debug_cond(DLEVEL == 2, "work_bgn=%d work_end=%d work_mid=%d\n",
		   *work_bgn, *work_end, *work_mid);
	/* Get the middle delay to be less than a VFIFO delay */
	for (*p = 0; *p <= IO_DQS_EN_PHASE_MAX;
		(*p)++, tmp_delay += IO_DELAY_PER_OPA_TAP)
		;
	debug_cond(DLEVEL == 2, "vfifo ptap delay %d\n", tmp_delay);
	while (*work_mid > tmp_delay)
		*work_mid -= tmp_delay;
	debug_cond(DLEVEL == 2, "new work_mid %d\n", *work_mid);

	tmp_delay = 0;
	for (*p = 0; *p <= IO_DQS_EN_PHASE_MAX && tmp_delay < *work_mid;
		(*p)++, tmp_delay += IO_DELAY_PER_OPA_TAP)
		;
	tmp_delay -= IO_DELAY_PER_OPA_TAP;
	debug_cond(DLEVEL == 2, "new p %d, tmp_delay=%d\n", (*p) - 1, tmp_delay);
	for (*d = 0; *d <= IO_DQS_EN_DELAY_MAX && tmp_delay < *work_mid; (*d)++,
		tmp_delay += IO_DELAY_PER_DQS_EN_DCHAIN_TAP)
		;
	debug_cond(DLEVEL == 2, "new d %d, tmp_delay=%d\n", *d, tmp_delay);

	scc_mgr_set_dqs_en_phase_all_ranks(*grp, (*p) - 1);
	scc_mgr_set_dqs_en_delay_all_ranks(*grp, *d);

	/*
	 * push vfifo until we can successfully calibrate. We can do this
	 * because the largest possible margin in 1 VFIFO cycle.
	 */
	for (i = 0; i < VFIFO_SIZE; i++) {
		debug_cond(DLEVEL == 2, "find_dqs_en_phase: center: vfifo=%u\n",
			   *v);
		if (rw_mgr_mem_calibrate_read_test_all_ranks(*grp, 1,
							     PASS_ONE_BIT,
							     bit_chk, 0)) {
			break;
		}

		/* fiddle with FIFO */
		rw_mgr_incr_vfifo(*grp, v);
	}

	if (i >= VFIFO_SIZE) {
		debug_cond(DLEVEL == 2, "%s:%d find_dqs_en_phase: center: \
			   failed\n", __func__, __LINE__);
		return 0;
	} else {
		return 1;
	}
}

/* find a good dqs enable to use */
static uint32_t rw_mgr_mem_calibrate_vfifo_find_dqs_en_phase(uint32_t grp)
{
	uint32_t v, d, p, i;
	uint32_t max_working_cnt;
	uint32_t bit_chk;
	uint32_t dtaps_per_ptap;
	uint32_t work_bgn, work_mid, work_end;
	uint32_t found_passing_read, found_failing_read, initial_failing_dtap;
	uint32_t addr;

	debug("%s:%d %u\n", __func__, __LINE__, grp);

	reg_file_set_sub_stage(CAL_SUBSTAGE_VFIFO_CENTER);

	scc_mgr_set_dqs_en_delay_all_ranks(grp, 0);
	scc_mgr_set_dqs_en_phase_all_ranks(grp, 0);

	/* ************************************************************** */
	/* * Step 0 : Determine number of delay taps for each phase tap * */
	dtaps_per_ptap = IO_DELAY_PER_OPA_TAP/IO_DELAY_PER_DQS_EN_DCHAIN_TAP;

	/* ********************************************************* */
	/* * Step 1 : First push vfifo until we get a failing read * */
	v = find_vfifo_read(grp, &bit_chk);

	max_working_cnt = 0;

	/* ******************************************************** */
	/* * step 2: find first working phase, increment in ptaps * */
	work_bgn = 0;
	if (find_working_phase(&grp, &bit_chk, dtaps_per_ptap, &work_bgn, &v, &d,
				&p, &i, &max_working_cnt) == 0)
		return 0;

	work_end = work_bgn;

	/*
	 * If d is 0 then the working window covers a phase tap and
	 * we can follow the old procedure otherwise, we've found the beginning,
	 * and we need to increment the dtaps until we find the end.
	 */
	if (d == 0) {
		/* ********************************************************* */
		/* * step 3a: if we have room, back off by one and
		increment in dtaps * */

		sdr_backup_phase(&grp, &bit_chk, &work_bgn, &v, &d, &p,
				 &max_working_cnt);

		/* ********************************************************* */
		/* * step 4a: go forward from working phase to non working
		phase, increment in ptaps * */
		if (sdr_nonworking_phase(&grp, &bit_chk, &work_bgn, &v, &d, &p,
					 &i, &max_working_cnt, &work_end) == 0)
			return 0;

		/* ********************************************************* */
		/* * step 5a:  back off one from last, increment in dtaps  * */

		/* Special case code for backing up a phase */
		if (p == 0) {
			p = IO_DQS_EN_PHASE_MAX;
			rw_mgr_decr_vfifo(grp, &v);
		} else {
			p = p - 1;
		}

		work_end -= IO_DELAY_PER_OPA_TAP;
		scc_mgr_set_dqs_en_phase_all_ranks(grp, p);

		/* * The actual increment of dtaps is done outside of
		the if/else loop to share code */
		d = 0;

		debug_cond(DLEVEL == 2, "%s:%d find_dqs_en_phase: v/p: \
			   vfifo=%u ptap=%u\n", __func__, __LINE__,
			   v, p);
	} else {
		/* ******************************************************* */
		/* * step 3-5b:  Find the right edge of the window using
		delay taps   * */
		debug_cond(DLEVEL == 2, "%s:%d find_dqs_en_phase:vfifo=%u \
			   ptap=%u dtap=%u bgn=%u\n", __func__, __LINE__,
			   v, p, d, work_bgn);

		work_end = work_bgn;

		/* * The actual increment of dtaps is done outside of the
		if/else loop to share code */

		/* Only here to counterbalance a subtract later on which is
		not needed if this branch of the algorithm is taken */
		max_working_cnt++;
	}

	/* The dtap increment to find the failing edge is done here */
	for (; d <= IO_DQS_EN_DELAY_MAX; d++, work_end +=
		IO_DELAY_PER_DQS_EN_DCHAIN_TAP) {
			debug_cond(DLEVEL == 2, "%s:%d find_dqs_en_phase: \
				   end-2: dtap=%u\n", __func__, __LINE__, d);
			scc_mgr_set_dqs_en_delay_all_ranks(grp, d);

			if (!rw_mgr_mem_calibrate_read_test_all_ranks(grp, 1,
								      PASS_ONE_BIT,
								      &bit_chk, 0)) {
				break;
			}
	}

	/* Go back to working dtap */
	if (d != 0)
		work_end -= IO_DELAY_PER_DQS_EN_DCHAIN_TAP;

	debug_cond(DLEVEL == 2, "%s:%d find_dqs_en_phase: v/p/d: vfifo=%u \
		   ptap=%u dtap=%u end=%u\n", __func__, __LINE__,
		   v, p, d-1, work_end);

	if (work_end < work_bgn) {
		/* nil range */
		debug_cond(DLEVEL == 2, "%s:%d find_dqs_en_phase: end-2: \
			   failed\n", __func__, __LINE__);
		return 0;
	}

	debug_cond(DLEVEL == 2, "%s:%d find_dqs_en_phase: found range [%u,%u]\n",
		   __func__, __LINE__, work_bgn, work_end);

	/* *************************************************************** */
	/*
	 * * We need to calculate the number of dtaps that equal a ptap
	 * * To do that we'll back up a ptap and re-find the edge of the
	 * * window using dtaps
	 */

	debug_cond(DLEVEL == 2, "%s:%d find_dqs_en_phase: calculate dtaps_per_ptap \
		   for tracking\n", __func__, __LINE__);

	/* Special case code for backing up a phase */
	if (p == 0) {
		p = IO_DQS_EN_PHASE_MAX;
		rw_mgr_decr_vfifo(grp, &v);
		debug_cond(DLEVEL == 2, "%s:%d find_dqs_en_phase: backedup \
			   cycle/phase: v=%u p=%u\n", __func__, __LINE__,
			   v, p);
	} else {
		p = p - 1;
		debug_cond(DLEVEL == 2, "%s:%d find_dqs_en_phase: backedup \
			   phase only: v=%u p=%u", __func__, __LINE__,
			   v, p);
	}

	scc_mgr_set_dqs_en_phase_all_ranks(grp, p);

	/*
	 * Increase dtap until we first see a passing read (in case the
	 * window is smaller than a ptap),
	 * and then a failing read to mark the edge of the window again
	 */

	/* Find a passing read */
	debug_cond(DLEVEL == 2, "%s:%d find_dqs_en_phase: find passing read\n",
		   __func__, __LINE__);
	found_passing_read = 0;
	found_failing_read = 0;
	initial_failing_dtap = d;
	for (; d <= IO_DQS_EN_DELAY_MAX; d++) {
		debug_cond(DLEVEL == 2, "%s:%d find_dqs_en_phase: testing \
			   read d=%u\n", __func__, __LINE__, d);
		scc_mgr_set_dqs_en_delay_all_ranks(grp, d);

		if (rw_mgr_mem_calibrate_read_test_all_ranks(grp, 1,
							     PASS_ONE_BIT,
							     &bit_chk, 0)) {
			found_passing_read = 1;
			break;
		}
	}

	if (found_passing_read) {
		/* Find a failing read */
		debug_cond(DLEVEL == 2, "%s:%d find_dqs_en_phase: find failing \
			   read\n", __func__, __LINE__);
		for (d = d + 1; d <= IO_DQS_EN_DELAY_MAX; d++) {
			debug_cond(DLEVEL == 2, "%s:%d find_dqs_en_phase: \
				   testing read d=%u\n", __func__, __LINE__, d);
			scc_mgr_set_dqs_en_delay_all_ranks(grp, d);

			if (!rw_mgr_mem_calibrate_read_test_all_ranks
				(grp, 1, PASS_ONE_BIT, &bit_chk, 0)) {
				found_failing_read = 1;
				break;
			}
		}
	} else {
		debug_cond(DLEVEL == 1, "%s:%d find_dqs_en_phase: failed to \
			   calculate dtaps", __func__, __LINE__);
		debug_cond(DLEVEL == 1, "per ptap. Fall back on static value\n");
	}

	/*
	 * The dynamically calculated dtaps_per_ptap is only valid if we
	 * found a passing/failing read. If we didn't, it means d hit the max
	 * (IO_DQS_EN_DELAY_MAX). Otherwise, dtaps_per_ptap retains its
	 * statically calculated value.
	 */
	if (found_passing_read && found_failing_read)
		dtaps_per_ptap = d - initial_failing_dtap;

	addr = sdr_get_addr(&sdr_reg_file->dtaps_per_ptap);
	writel(dtaps_per_ptap, SOCFPGA_SDR_ADDRESS + addr);
	debug_cond(DLEVEL == 2, "%s:%d find_dqs_en_phase: dtaps_per_ptap=%u \
		   - %u = %u",  __func__, __LINE__, d,
		   initial_failing_dtap, dtaps_per_ptap);

	/* ******************************************** */
	/* * step 6:  Find the centre of the window   * */
	if (sdr_find_window_centre(&grp, &bit_chk, &work_bgn, &v, &d, &p,
				   &work_mid, &work_end) == 0)
		return 0;

	debug_cond(DLEVEL == 2, "%s:%d find_dqs_en_phase: center found: \
		   vfifo=%u ptap=%u dtap=%u\n", __func__, __LINE__,
		   v, p-1, d);
	return 1;
}

/*
 * Try rw_mgr_mem_calibrate_vfifo_find_dqs_en_phase across different
 * dq_in_delay values
 */
static uint32_t
rw_mgr_mem_calibrate_vfifo_find_dqs_en_phase_sweep_dq_in_delay
(uint32_t write_group, uint32_t read_group, uint32_t test_bgn)
{
	uint32_t found;
	uint32_t i;
	uint32_t p;
	uint32_t d;
	uint32_t r;
	uint32_t addr;

	const uint32_t delay_step = IO_IO_IN_DELAY_MAX /
		(RW_MGR_MEM_DQ_PER_READ_DQS-1);
		/* we start at zero, so have one less dq to devide among */

	debug("%s:%d (%u,%u,%u)", __func__, __LINE__, write_group, read_group,
	      test_bgn);

	/* try different dq_in_delays since the dq path is shorter than dqs */

	for (r = 0; r < RW_MGR_MEM_NUMBER_OF_RANKS;
	     r += NUM_RANKS_PER_SHADOW_REG) {
		for (i = 0, p = test_bgn, d = 0; i < RW_MGR_MEM_DQ_PER_READ_DQS;
			i++, p++, d += delay_step) {
			debug_cond(DLEVEL == 1, "%s:%d rw_mgr_mem_calibrate_\
				   vfifo_find_dqs_", __func__, __LINE__);
			debug_cond(DLEVEL == 1, "en_phase_sweep_dq_in_delay: g=%u/%u ",
			       write_group, read_group);
			debug_cond(DLEVEL == 1, "r=%u, i=%u p=%u d=%u\n", r, i , p, d);
			scc_mgr_set_dq_in_delay(write_group, p, d);
			scc_mgr_load_dq(p);
		}
		addr = sdr_get_addr(&sdr_scc_mgr->update);
		writel(0, SOCFPGA_SDR_ADDRESS + addr);
	}

	found = rw_mgr_mem_calibrate_vfifo_find_dqs_en_phase(read_group);

	debug_cond(DLEVEL == 1, "%s:%d rw_mgr_mem_calibrate_vfifo_find_dqs_\
		   en_phase_sweep_dq", __func__, __LINE__);
	debug_cond(DLEVEL == 1, "_in_delay: g=%u/%u found=%u; Reseting delay \
		   chain to zero\n", write_group, read_group, found);

	for (r = 0; r < RW_MGR_MEM_NUMBER_OF_RANKS;
	     r += NUM_RANKS_PER_SHADOW_REG) {
		for (i = 0, p = test_bgn; i < RW_MGR_MEM_DQ_PER_READ_DQS;
			i++, p++) {
			scc_mgr_set_dq_in_delay(write_group, p, 0);
			scc_mgr_load_dq(p);
		}
		addr = sdr_get_addr(&sdr_scc_mgr->update);
		writel(0, SOCFPGA_SDR_ADDRESS + addr);
	}

	return found;
}

/* per-bit deskew DQ and center */
static uint32_t rw_mgr_mem_calibrate_vfifo_center(uint32_t rank_bgn,
	uint32_t write_group, uint32_t read_group, uint32_t test_bgn,
	uint32_t use_read_test, uint32_t update_fom)
{
	uint32_t i, p, d, min_index;
	/*
	 * Store these as signed since there are comparisons with
	 * signed numbers.
	 */
	uint32_t bit_chk;
	uint32_t sticky_bit_chk;
	int32_t left_edge[RW_MGR_MEM_DQ_PER_READ_DQS];
	int32_t right_edge[RW_MGR_MEM_DQ_PER_READ_DQS];
	int32_t final_dq[RW_MGR_MEM_DQ_PER_READ_DQS];
	int32_t mid;
	int32_t orig_mid_min, mid_min;
	int32_t new_dqs, start_dqs, start_dqs_en, shift_dq, final_dqs,
		final_dqs_en;
	int32_t dq_margin, dqs_margin;
	uint32_t stop;
	uint32_t temp_dq_in_delay1, temp_dq_in_delay2;
	uint32_t addr;

	debug("%s:%d: %u %u", __func__, __LINE__, read_group, test_bgn);

	addr = sdr_get_addr((u32 *)SCC_MGR_DQS_IN_DELAY);
	start_dqs = readl(SOCFPGA_SDR_ADDRESS + addr + (read_group << 2));
	if (IO_SHIFT_DQS_EN_WHEN_SHIFT_DQS)
		start_dqs_en = readl(SOCFPGA_SDR_ADDRESS + addr + ((read_group << 2)
				     - IO_DQS_EN_DELAY_OFFSET));

	/* set the left and right edge of each bit to an illegal value */
	/* use (IO_IO_IN_DELAY_MAX + 1) as an illegal value */
	sticky_bit_chk = 0;
	for (i = 0; i < RW_MGR_MEM_DQ_PER_READ_DQS; i++) {
		left_edge[i]  = IO_IO_IN_DELAY_MAX + 1;
		right_edge[i] = IO_IO_IN_DELAY_MAX + 1;
	}

	addr = sdr_get_addr(&sdr_scc_mgr->update);
	/* Search for the left edge of the window for each bit */
	for (d = 0; d <= IO_IO_IN_DELAY_MAX; d++) {
		scc_mgr_apply_group_dq_in_delay(write_group, test_bgn, d);

		writel(0, SOCFPGA_SDR_ADDRESS + addr);

		/*
		 * Stop searching when the read test doesn't pass AND when
		 * we've seen a passing read on every bit.
		 */
		if (use_read_test) {
			stop = !rw_mgr_mem_calibrate_read_test(rank_bgn,
				read_group, NUM_READ_PB_TESTS, PASS_ONE_BIT,
				&bit_chk, 0, 0);
		} else {
			rw_mgr_mem_calibrate_write_test(rank_bgn, write_group,
							0, PASS_ONE_BIT,
							&bit_chk, 0);
			bit_chk = bit_chk >> (RW_MGR_MEM_DQ_PER_READ_DQS *
				(read_group - (write_group *
					RW_MGR_MEM_IF_READ_DQS_WIDTH /
					RW_MGR_MEM_IF_WRITE_DQS_WIDTH)));
			stop = (bit_chk == 0);
		}
		sticky_bit_chk = sticky_bit_chk | bit_chk;
		stop = stop && (sticky_bit_chk == param->read_correct_mask);
		debug_cond(DLEVEL == 2, "%s:%d vfifo_center(left): dtap=%u => %u == %u \
			   && %u", __func__, __LINE__, d,
			   sticky_bit_chk,
			param->read_correct_mask, stop);

		if (stop == 1) {
			break;
		} else {
			for (i = 0; i < RW_MGR_MEM_DQ_PER_READ_DQS; i++) {
				if (bit_chk & 1) {
					/* Remember a passing test as the
					left_edge */
					left_edge[i] = d;
				} else {
					/* If a left edge has not been seen yet,
					then a future passing test will mark
					this edge as the right edge */
					if (left_edge[i] ==
						IO_IO_IN_DELAY_MAX + 1) {
						right_edge[i] = -(d + 1);
					}
				}
				bit_chk = bit_chk >> 1;
			}
		}
	}

	/* Reset DQ delay chains to 0 */
	scc_mgr_apply_group_dq_in_delay(write_group, test_bgn, 0);
	sticky_bit_chk = 0;
	for (i = RW_MGR_MEM_DQ_PER_READ_DQS - 1;; i--) {
		debug_cond(DLEVEL == 2, "%s:%d vfifo_center: left_edge[%u]: \
			   %d right_edge[%u]: %d\n", __func__, __LINE__,
			   i, left_edge[i], i, right_edge[i]);

		/*
		 * Check for cases where we haven't found the left edge,
		 * which makes our assignment of the the right edge invalid.
		 * Reset it to the illegal value.
		 */
		if ((left_edge[i] == IO_IO_IN_DELAY_MAX + 1) && (
			right_edge[i] != IO_IO_IN_DELAY_MAX + 1)) {
			right_edge[i] = IO_IO_IN_DELAY_MAX + 1;
			debug_cond(DLEVEL == 2, "%s:%d vfifo_center: reset \
				   right_edge[%u]: %d\n", __func__, __LINE__,
				   i, right_edge[i]);
		}

		/*
		 * Reset sticky bit (except for bits where we have seen
		 * both the left and right edge).
		 */
		sticky_bit_chk = sticky_bit_chk << 1;
		if ((left_edge[i] != IO_IO_IN_DELAY_MAX + 1) &&
		    (right_edge[i] != IO_IO_IN_DELAY_MAX + 1)) {
			sticky_bit_chk = sticky_bit_chk | 1;
		}

		if (i == 0)
			break;
	}

	addr = sdr_get_addr(&sdr_scc_mgr->update);
	/* Search for the right edge of the window for each bit */
	for (d = 0; d <= IO_DQS_IN_DELAY_MAX - start_dqs; d++) {
		scc_mgr_set_dqs_bus_in_delay(read_group, d + start_dqs);
		if (IO_SHIFT_DQS_EN_WHEN_SHIFT_DQS) {
			uint32_t delay = d + start_dqs_en;
			if (delay > IO_DQS_EN_DELAY_MAX)
				delay = IO_DQS_EN_DELAY_MAX;
			scc_mgr_set_dqs_en_delay(read_group, delay);
		}
		scc_mgr_load_dqs(read_group);

		writel(0, SOCFPGA_SDR_ADDRESS + addr);

		/*
		 * Stop searching when the read test doesn't pass AND when
		 * we've seen a passing read on every bit.
		 */
		if (use_read_test) {
			stop = !rw_mgr_mem_calibrate_read_test(rank_bgn,
				read_group, NUM_READ_PB_TESTS, PASS_ONE_BIT,
				&bit_chk, 0, 0);
		} else {
			rw_mgr_mem_calibrate_write_test(rank_bgn, write_group,
							0, PASS_ONE_BIT,
							&bit_chk, 0);
			bit_chk = bit_chk >> (RW_MGR_MEM_DQ_PER_READ_DQS *
				(read_group - (write_group *
					RW_MGR_MEM_IF_READ_DQS_WIDTH /
					RW_MGR_MEM_IF_WRITE_DQS_WIDTH)));
			stop = (bit_chk == 0);
		}
		sticky_bit_chk = sticky_bit_chk | bit_chk;
		stop = stop && (sticky_bit_chk == param->read_correct_mask);

		debug_cond(DLEVEL == 2, "%s:%d vfifo_center(right): dtap=%u => %u == \
			   %u && %u", __func__, __LINE__, d,
			   sticky_bit_chk, param->read_correct_mask, stop);

		if (stop == 1) {
			break;
		} else {
			for (i = 0; i < RW_MGR_MEM_DQ_PER_READ_DQS; i++) {
				if (bit_chk & 1) {
					/* Remember a passing test as
					the right_edge */
					right_edge[i] = d;
				} else {
					if (d != 0) {
						/* If a right edge has not been
						seen yet, then a future passing
						test will mark this edge as the
						left edge */
						if (right_edge[i] ==
						IO_IO_IN_DELAY_MAX + 1) {
							left_edge[i] = -(d + 1);
						}
					} else {
						/* d = 0 failed, but it passed
						when testing the left edge,
						so it must be marginal,
						set it to -1 */
						if (right_edge[i] ==
							IO_IO_IN_DELAY_MAX + 1 &&
							left_edge[i] !=
							IO_IO_IN_DELAY_MAX
							+ 1) {
							right_edge[i] = -1;
						}
						/* If a right edge has not been
						seen yet, then a future passing
						test will mark this edge as the
						left edge */
						else if (right_edge[i] ==
							IO_IO_IN_DELAY_MAX +
							1) {
							left_edge[i] = -(d + 1);
						}
					}
Loading
Loading full blame...