nb/intel/sandybridge: Rework timA minmax code

There's no need to use `struct timA_minmax`, since most cases only care
about the difference between logic delay deltas. The final step does use
the minimum logic delay across all lanes, but it's a special case.

Tested on Asus P8H61-M PRO, still boots.

Change-Id: I1da95520ac915ab003e1a839685cbf5f1970eb6a
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/47604
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
This commit is contained in:
Angel Pons 2020-11-13 23:10:52 +01:00
parent bd04995cdf
commit 12bd8ab2f9

View file

@ -1,5 +1,6 @@
/* SPDX-License-Identifier: GPL-2.0-only */
#include <assert.h>
#include <commonlib/helpers.h>
#include <console/console.h>
#include <string.h>
@ -1225,53 +1226,74 @@ static int discover_402x(ramctr_timing *ctrl, int channel, int slotrank, int *up
return 0;
}
struct timA_minmax {
int timA_min_high, timA_max_high;
};
static void pre_timA_change(ramctr_timing *ctrl, int channel, int slotrank,
struct timA_minmax *mnmx)
static int get_logic_delay_delta(ramctr_timing *ctrl, int channel, int slotrank)
{
int lane;
mnmx->timA_min_high = 7;
mnmx->timA_max_high = 0;
u16 logic_delay_min = 7;
u16 logic_delay_max = 0;
FOR_ALL_LANES {
if (mnmx->timA_min_high >
(ctrl->timings[channel][slotrank].lanes[lane].timA >> 6))
mnmx->timA_min_high =
(ctrl->timings[channel][slotrank].lanes[lane].timA >> 6);
if (mnmx->timA_max_high <
(ctrl->timings[channel][slotrank].lanes[lane].timA >> 6))
mnmx->timA_max_high =
(ctrl->timings[channel][slotrank].lanes[lane].timA >> 6);
const u16 logic_delay = ctrl->timings[channel][slotrank].lanes[lane].timA >> 6;
logic_delay_min = MIN(logic_delay_min, logic_delay);
logic_delay_max = MAX(logic_delay_max, logic_delay);
}
if (logic_delay_max < logic_delay_min) {
printk(BIOS_EMERG, "Logic delay max < min (%u < %u): %d, %d\n",
logic_delay_max, logic_delay_min, channel, slotrank);
}
assert(logic_delay_max >= logic_delay_min);
return logic_delay_max - logic_delay_min;
}
static void post_timA_change(ramctr_timing *ctrl, int channel, int slotrank,
struct timA_minmax *mnmx)
static int align_rt_io_latency(ramctr_timing *ctrl, int channel, int slotrank, int prev)
{
struct timA_minmax post;
int shift_402x = 0;
int latency_offset = 0;
/* Get changed maxima */
pre_timA_change(ctrl, channel, slotrank, &post);
const int post = get_logic_delay_delta(ctrl, channel, slotrank);
if (mnmx->timA_max_high - mnmx->timA_min_high <
post.timA_max_high - post.timA_min_high)
shift_402x = +1;
if (prev < post)
latency_offset = +1;
else if (mnmx->timA_max_high - mnmx->timA_min_high >
post.timA_max_high - post.timA_min_high)
shift_402x = -1;
else if (prev > post)
latency_offset = -1;
else
shift_402x = 0;
latency_offset = 0;
ctrl->timings[channel][slotrank].io_latency += shift_402x;
ctrl->timings[channel][slotrank].roundtrip_latency += shift_402x;
printram("4024 += %d;\n", shift_402x);
printram("4028 += %d;\n", shift_402x);
ctrl->timings[channel][slotrank].io_latency += latency_offset;
ctrl->timings[channel][slotrank].roundtrip_latency += latency_offset;
printram("4024 += %d;\n", latency_offset);
printram("4028 += %d;\n", latency_offset);
return post;
}
static void compute_final_logic_delay(ramctr_timing *ctrl, int channel, int slotrank)
{
u16 logic_delay_min = 7;
int lane;
FOR_ALL_LANES {
const u16 logic_delay = ctrl->timings[channel][slotrank].lanes[lane].timA >> 6;
logic_delay_min = MIN(logic_delay_min, logic_delay);
}
if (logic_delay_min >= 2) {
printk(BIOS_WARNING, "Logic delay %u greater than 1: %d %d\n",
logic_delay_min, channel, slotrank);
}
FOR_ALL_LANES {
ctrl->timings[channel][slotrank].lanes[lane].timA -= logic_delay_min << 6;
}
ctrl->timings[channel][slotrank].io_latency -= logic_delay_min;
printram("4028 -= %d;\n", logic_delay_min);
}
/*
@ -1300,7 +1322,7 @@ int read_training(ramctr_timing *ctrl)
FOR_ALL_CHANNELS FOR_ALL_POPULATED_RANKS {
int all_high, some_high;
int upperA[NUM_LANES];
struct timA_minmax mnmx;
int prev;
wait_for_iosav(channel);
@ -1343,28 +1365,21 @@ int read_training(ramctr_timing *ctrl)
program_timings(ctrl, channel);
pre_timA_change(ctrl, channel, slotrank, &mnmx);
prev = get_logic_delay_delta(ctrl, channel, slotrank);
err = discover_402x(ctrl, channel, slotrank, upperA);
if (err)
return err;
post_timA_change(ctrl, channel, slotrank, &mnmx);
pre_timA_change(ctrl, channel, slotrank, &mnmx);
prev = align_rt_io_latency(ctrl, channel, slotrank, prev);
discover_timA_fine(ctrl, channel, slotrank, upperA);
post_timA_change(ctrl, channel, slotrank, &mnmx);
pre_timA_change(ctrl, channel, slotrank, &mnmx);
prev = align_rt_io_latency(ctrl, channel, slotrank, prev);
FOR_ALL_LANES {
ctrl->timings[channel][slotrank].lanes[lane].timA -=
mnmx.timA_min_high * 0x40;
}
ctrl->timings[channel][slotrank].io_latency -= mnmx.timA_min_high;
printram("4028 -= %d;\n", mnmx.timA_min_high);
compute_final_logic_delay(ctrl, channel, slotrank);
post_timA_change(ctrl, channel, slotrank, &mnmx);
align_rt_io_latency(ctrl, channel, slotrank, prev);
printram("4/8: %d, %d, %x, %x\n", channel, slotrank,
ctrl->timings[channel][slotrank].roundtrip_latency,