diff options
Diffstat (limited to 'board/xes/xpedite537x/ddr.c')
| -rw-r--r-- | board/xes/xpedite537x/ddr.c | 234 |
1 files changed, 0 insertions, 234 deletions
diff --git a/board/xes/xpedite537x/ddr.c b/board/xes/xpedite537x/ddr.c deleted file mode 100644 index f55102a072e..00000000000 --- a/board/xes/xpedite537x/ddr.c +++ /dev/null @@ -1,234 +0,0 @@ -// SPDX-License-Identifier: GPL-2.0+ -/* - * Copyright 2008 Extreme Engineering Solutions, Inc. - * Copyright 2008 Freescale Semiconductor, Inc. - */ - -#include <common.h> -#include <i2c.h> -#include <log.h> - -#include <fsl_ddr_sdram.h> -#include <fsl_ddr_dimm_params.h> - -void get_spd(ddr2_spd_eeprom_t *spd, u8 i2c_address) -{ - i2c_read(i2c_address, SPD_EEPROM_OFFSET, 2, (uchar *)spd, - sizeof(ddr2_spd_eeprom_t)); -} - -/* - * There are four board-specific SDRAM timing parameters which must be - * calculated based on the particular PCB artwork. These are: - * 1.) CPO (Read Capture Delay) - * - TIMING_CFG_2 register - * Source: Calculation based on board trace lengths and - * chip-specific internal delays. - * 2.) WR_DATA_DELAY (Write Command to Data Strobe Delay) - * - TIMING_CFG_2 register - * Source: Calculation based on board trace lengths. - * Unless clock and DQ lanes are very different - * lengths (>2"), this should be set to the nominal value - * of 1/2 clock delay. - * 3.) CLK_ADJUST (Clock and Addr/Cmd alignment control) - * - DDR_SDRAM_CLK_CNTL register - * Source: Signal Integrity Simulations - * 4.) 2T Timing on Addr/Ctl - * - TIMING_CFG_2 register - * Source: Signal Integrity Simulations - * Usually only needed with heavy load/very high speed (>DDR2-800) - * - * ====== XPedite5370 DDR2-600 read delay calculations ====== - * - * See Freescale's App Note AN2583 as refrence. This document also - * contains the chip-specific delays for 8548E, 8572, etc. - * - * For MPC8572E - * Minimum chip delay (Ch 0): 1.372ns - * Maximum chip delay (Ch 0): 2.914ns - * Minimum chip delay (Ch 1): 1.220ns - * Maximum chip delay (Ch 1): 2.595ns - * - * CLK adjust = 5 (from simulations) = 5/8* 3.33ns = 2080ps - * - * Minimum delay calc (Ch 0): - * clock prop - dram skew + min dqs prop delay + clk_adjust + min chip dly - * 2.3" * 180 - 400ps + 1.9" * 180 + 2080ps + 1372ps - * = 3808ps - * = 3.808ns - * - * Maximum delay calc (Ch 0): - * clock prop + dram skew + max dqs prop delay + clk_adjust + max chip dly - * 2.3" * 180 + 400ps + 2.4" * 180 + 2080ps + 2914ps - * = 6240ps - * = 6.240ns - * - * Minimum delay calc (Ch 1): - * clock prop - dram skew + min dqs prop delay + clk_adjust + min chip dly - * 1.46" * 180- 400ps + 0.7" * 180 + 2080ps + 1220ps - * = 3288ps - * = 3.288ns - * - * Maximum delay calc (Ch 1): - * clock prop + dram skew + max dqs prop delay + clk_adjust + min chip dly - * 1.46" * 180+ 400ps + 1.1" * 180 + 2080ps + 2595ps - * = 5536ps - * = 5.536ns - * - * Ch.0: 3.808ns to 6.240ns additional delay needed (pick 5ns as target) - * This is 1.5 clock cycles, pick CPO = READ_LAT + 3/2 (0x8) - * Ch.1: 3.288ns to 5.536ns additional delay needed (pick 4.4ns as target) - * This is 1.32 clock cycles, pick CPO = READ_LAT + 5/4 (0x7) - * - * - * ====== XPedite5370 DDR2-800 read delay calculations ====== - * - * See Freescale's App Note AN2583 as refrence. This document also - * contains the chip-specific delays for 8548E, 8572, etc. - * - * For MPC8572E - * Minimum chip delay (Ch 0): 1.372ns - * Maximum chip delay (Ch 0): 2.914ns - * Minimum chip delay (Ch 1): 1.220ns - * Maximum chip delay (Ch 1): 2.595ns - * - * CLK adjust = 5 (from simulations) = 5/8* 2.5ns = 1563ps - * - * Minimum delay calc (Ch 0): - * clock prop - dram skew + min dqs prop delay + clk_adjust + min chip dly - * 2.3" * 180 - 350ps + 1.9" * 180 + 1563ps + 1372ps - * = 3341ps - * = 3.341ns - * - * Maximum delay calc (Ch 0): - * clock prop + dram skew + max dqs prop delay + clk_adjust + max chip dly - * 2.3" * 180 + 350ps + 2.4" * 180 + 1563ps + 2914ps - * = 5673ps - * = 5.673ns - * - * Minimum delay calc (Ch 1): - * clock prop - dram skew + min dqs prop delay + clk_adjust + min chip dly - * 1.46" * 180- 350ps + 0.7" * 180 + 1563ps + 1220ps - * = 2822ps - * = 2.822ns - * - * Maximum delay calc (Ch 1): - * clock prop + dram skew + max dqs prop delay + clk_adjust + min chip dly - * 1.46" * 180+ 350ps + 1.1" * 180 + 1563ps + 2595ps - * = 4968ps - * = 4.968ns - * - * Ch.0: 3.341ns to 5.673ns additional delay needed (pick 4.5ns as target) - * This is 1.8 clock cycles, pick CPO = READ_LAT + 7/4 (0x9) - * Ch.1: 2.822ns to 4.968ns additional delay needed (pick 3.9ns as target) - * This is 1.56 clock cycles, pick CPO = READ_LAT + 3/2 (0x8) - * - * Write latency (WR_DATA_DELAY) is calculated by doing the following: - * - * The DDR SDRAM specification requires DQS be received no sooner than - * 75% of an SDRAM clock period—and no later than 125% of a clock - * period—from the capturing clock edge of the command/address at the - * SDRAM. - * - * Based on the above tracelengths, the following are calculated: - * Ch. 0 8572 to DRAM propagation (DQ lanes) : 1.9" * 180 = 0.342ns - * Ch. 0 8572 to DRAM propagation (CLKs) : 2.3" * 180 = 0.414ns - * Ch. 1 8572 to DRAM propagation (DQ lanes) : 0.7" * 180 = 0.126ns - * Ch. 1 8572 to DRAM propagation (CLKs ) : 1.47" * 180 = 0.264ns - * - * Difference in arrival time CLK vs. DQS: - * Ch. 0 0.072ns - * Ch. 1 0.138ns - * - * Both of these values are much less than 25% of the clock - * period at DDR2-600 or DDR2-800, so no additional delay is needed over - * the 1/2 cycle which normally aligns the first DQS transition - * exactly WL (CAS latency minus one cycle) after the CAS strobe. - * See Figure 9-53 in MPC8572E manual: "1/2 delay" in Freescale's - * terminology corresponds to exactly one clock period delay after - * the CAS strobe. (due to the fact that the "delay" is referenced - * from the *falling* edge of the CLK, just after the rising edge - * which the CAS strobe is latched on. - */ - -typedef struct board_memctl_options { - uint16_t datarate_mhz_low; - uint16_t datarate_mhz_high; - uint8_t clk_adjust; - uint8_t cpo_override; - uint8_t write_data_delay; -} board_memctl_options_t; - -static struct board_memctl_options bopts_ctrl[][2] = { - { - /* Controller 0 */ - { - /* DDR2 600/667 */ - .datarate_mhz_low = 500, - .datarate_mhz_high = 750, - .clk_adjust = 5, - .cpo_override = 8, - .write_data_delay = 2, - }, - { - /* DDR2 800 */ - .datarate_mhz_low = 750, - .datarate_mhz_high = 850, - .clk_adjust = 5, - .cpo_override = 9, - .write_data_delay = 2, - }, - }, - { - /* Controller 1 */ - { - /* DDR2 600/667 */ - .datarate_mhz_low = 500, - .datarate_mhz_high = 750, - .clk_adjust = 5, - .cpo_override = 7, - .write_data_delay = 2, - }, - { - /* DDR2 800 */ - .datarate_mhz_low = 750, - .datarate_mhz_high = 850, - .clk_adjust = 5, - .cpo_override = 8, - .write_data_delay = 2, - }, - }, -}; - -void fsl_ddr_board_options(memctl_options_t *popts, - dimm_params_t *pdimm, - unsigned int ctrl_num) -{ - struct board_memctl_options *bopts = bopts_ctrl[ctrl_num]; - sys_info_t sysinfo; - int i; - unsigned int datarate; - - get_sys_info(&sysinfo); - datarate = sysinfo.freq_ddrbus / 1000 / 1000; - - for (i = 0; i < ARRAY_SIZE(bopts_ctrl[ctrl_num]); i++) { - if ((bopts[i].datarate_mhz_low <= datarate) && - (bopts[i].datarate_mhz_high >= datarate)) { - debug("controller %d:\n", ctrl_num); - debug(" clk_adjust = %d\n", bopts[i].clk_adjust); - debug(" cpo = %d\n", bopts[i].cpo_override); - debug(" write_data_delay = %d\n", - bopts[i].write_data_delay); - popts->clk_adjust = bopts[i].clk_adjust; - popts->cpo_override = bopts[i].cpo_override; - popts->write_data_delay = bopts[i].write_data_delay; - } - } - - /* - * Factors to consider for half-strength driver enable: - * - number of DIMMs installed - */ - popts->half_strength_driver_enable = 0; -} |