[coreboot-gerrit] New patch to review for coreboot: google/gru: Determine Board ID based on the input voltage of ADC1
Patrick Georgi (pgeorgi@google.com)
gerrit at coreboot.org
Sat May 7 08:30:41 CEST 2016
Patrick Georgi (pgeorgi at google.com) just uploaded a new patch set to gerrit, which you can find at https://review.coreboot.org/14722
-gerrit
commit 410a40c1aca39d9ea32d8c3d7ffeead22807a9de
Author: Vadim Bendebury <vbendeb at chromium.org>
Date: Fri Apr 22 07:03:43 2016 -0700
google/gru: Determine Board ID based on the input voltage of ADC1
The Board ID on the Gru family of boards is determined by reading the
voltage from a resistor divider, each hardware revision is supposed to
have a unique resistor ratio, which allows to distinctly tell between
different Board ID.
While the long time approach to mapping resistor ratios (and voltages)
into Board ID remains under discussion, we know for sure the values
for Proto 1 and Proto 2. Let's just use them for now.
Since Board ID can be queried multiple times during boot, ideally it
should be read once and placed in the coreboot table to be available
to all coreboot stages. For now we just cache it so that at least
during the same stage the ADC has to run only once.
BRANCH=None
BUG=chrome-os-partner:51537
TEST=verified that the voltage reading on Proto 1 is as expected, and
Board ID 0 is reported.
Change-Id: I94bc7fc235dae4155feb6ca35b5ef0ab20c3ec9c
Signed-off-by: Patrick Georgi <pgeorgi at chromium.org>
Original-Commit-Id: bb4064d0af8174b6ae247cdad9378b7f4e5f22ba
Original-Change-Id: I105ea97f8862b5707b582904c6f2e3e9406a0f07
Original-Signed-off-by: Vadim Bendebury <vbendeb at chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/340428
Original-Reviewed-by: Patrick Georgi <patrick at georgi-clan.de>
---
src/mainboard/google/gru/Makefile.inc | 3 ---
src/mainboard/google/gru/boardid.c | 37 +++++++++++++++++++++++++++++++++++
2 files changed, 37 insertions(+), 3 deletions(-)
diff --git a/src/mainboard/google/gru/Makefile.inc b/src/mainboard/google/gru/Makefile.inc
index da61513..a209d1b 100644
--- a/src/mainboard/google/gru/Makefile.inc
+++ b/src/mainboard/google/gru/Makefile.inc
@@ -13,16 +13,13 @@
## GNU General Public License for more details.
##
-bootblock-y += boardid.c
bootblock-y += bootblock.c
bootblock-y += chromeos.c
bootblock-y += memlayout.ld
bootblock-y += reset.c
-verstage-y += boardid.c
verstage-y += chromeos.c
verstage-y += memlayout.ld
-verstage-y += memlayout.ld
verstage-y += reset.c
romstage-y += boardid.c
diff --git a/src/mainboard/google/gru/boardid.c b/src/mainboard/google/gru/boardid.c
index f62fe97..5f030ad 100644
--- a/src/mainboard/google/gru/boardid.c
+++ b/src/mainboard/google/gru/boardid.c
@@ -16,9 +16,46 @@
#include <boardid.h>
#include <console/console.h>
#include <stdlib.h>
+#include <soc/saradc.h>
+
+/*
+ * This matches two Kevin prototypes, needs to be sorted out with HW engs to
+ * have more regular mapping between the voltage and board ID.
+ */
+static const int board_id_readings[] = { 42, 120 };
+
+/*
+ * The ADC produces a 10 bit value, the resistor accuracy is 1%, let's leave
+ * 2% room for error on both sides, total variation would be 4%, which is
+ * approximately 40 points with a 10 bit ADC. The hardware specification
+ * guarantees valid readings to be at least 64 bits (2^6) apart.
+ */
+#define ACCEPTABLE_DELTA (int)(1024 * .02)
uint8_t board_id(void)
{
+ static int cached_board_id = -1;
+ int i;
+ int adc_reading;
+
+ if (cached_board_id != -1)
+ return cached_board_id;
+
+ adc_reading = get_saradc_value(1);
+ for (i = 0; i < ARRAY_SIZE(board_id_readings); i++) {
+ int delta = board_id_readings[i] - adc_reading;
+
+ if ((delta < ACCEPTABLE_DELTA) && (delta > -ACCEPTABLE_DELTA)) {
+ printk(BIOS_DEBUG, "ADC reading %d, "
+ "expected value %d board ID %d\n",
+ adc_reading, delta + adc_reading, i);
+ cached_board_id = i;
+ return i;
+ }
+ }
+
+ printk(BIOS_ERR, "Unmatched ADC reading of %d, using Board ID of 0\n",
+ adc_reading);
return 0;
}
More information about the coreboot-gerrit
mailing list