libreboot
/
lbmk
1
0
Fork 0
Code Packages Projects Releases Activity
lbmk/config/coreboot/default/patches/0042-mb-dell-optiplex_9020-...

603 lines
17 KiB
Diff
Raw Normal View History

9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
From c58e0fea2a4e591e5ecd8a1f376c3b3af0fbb306 Mon Sep 17 00:00:00 2001
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
From: Mate Kukri <kukri.mate@gmail.com>
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
Date: Thu, 18 Apr 2024 20:28:45 +0100
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
Subject: [PATCH 1/1] mb/dell/optiplex_9020: Implement late HWM initialization
There are 4 different chassis types specified by vendor firmware, each
with a slightly different HWM configuration.
The chassis type to use is determined at runtime by reading a set of
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
4 PCH GPIOs: 70, 38, 17, and 1.
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
Additionally vendor firmware also provides an option to run the fans at
full speed. This is substituted with a coreboot nvram option in this
implementation.
This was tested to make fan control work on my OptiPlex 7020 SFF.
NOTE: This is superficially similar to the OptiPlex 9010's SCH5545
however the OptiPlex 9020's SCH5555 does not use externally
programmed EC firmware.
Change-Id: Ibdccd3fc7364e03e84ca606592928410624eed43
Signed-off-by: Mate Kukri <kukri.mate@gmail.com>
---
src/mainboard/dell/optiplex_9020/Makefile.inc | 3 +-
src/mainboard/dell/optiplex_9020/bootblock.c | 25 +-
src/mainboard/dell/optiplex_9020/cmos.default | 1 +
src/mainboard/dell/optiplex_9020/cmos.layout | 5 +-
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
src/mainboard/dell/optiplex_9020/mainboard.c | 387 ++++++++++++++++++
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
src/mainboard/dell/optiplex_9020/sch5555_ec.c | 54 +++
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
src/mainboard/dell/optiplex_9020/sch5555_ec.h | 10 +
7 files changed, 463 insertions(+), 22 deletions(-)
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
create mode 100644 src/mainboard/dell/optiplex_9020/sch5555_ec.c
create mode 100644 src/mainboard/dell/optiplex_9020/sch5555_ec.h
diff --git a/src/mainboard/dell/optiplex_9020/Makefile.inc b/src/mainboard/dell/optiplex_9020/Makefile.inc
index 6ca2f2afaa..08e2e53577 100644
--- a/src/mainboard/dell/optiplex_9020/Makefile.inc
+++ b/src/mainboard/dell/optiplex_9020/Makefile.inc
@@ -2,4 +2,5 @@
romstage-y += gpio.c
ramstage-$(CONFIG_MAINBOARD_USE_LIBGFXINIT) += gma-mainboard.ads
-bootblock-y += bootblock.c
+ramstage-y += sch5555_ec.c
+bootblock-y += bootblock.c sch5555_ec.c
diff --git a/src/mainboard/dell/optiplex_9020/bootblock.c b/src/mainboard/dell/optiplex_9020/bootblock.c
index 2837cf9cf1..e5e759273e 100644
--- a/src/mainboard/dell/optiplex_9020/bootblock.c
+++ b/src/mainboard/dell/optiplex_9020/bootblock.c
@@ -4,29 +4,14 @@
#include <device/pnp_ops.h>
#include <superio/smsc/sch555x/sch555x.h>
#include <southbridge/intel/lynxpoint/pch.h>
-
-static void ec_write(uint8_t addr1, uint16_t addr2, uint8_t val)
-{
- // Clear EC-to-Host mailbox
- uint8_t tmp = inb(SCH555x_EMI_IOBASE + 1);
- outb(tmp, SCH555x_EMI_IOBASE + 1);
-
- // Send address and value to the EC
- sch555x_emi_write16(0, (addr1 * 2) | 0x101);
- sch555x_emi_write32(4, val | (addr2 << 16));
-
- // Wait for acknowledgement message from EC
- outb(1, SCH555x_EMI_IOBASE);
- size_t timeout = 0;
- do {} while (++timeout < 0xfff && (inb(SCH555x_EMI_IOBASE + 1) & 1) == 0);
-}
+#include "sch5555_ec.h"
struct ec_init_entry {
uint16_t addr;
uint8_t val;
};
-static void ec_init(void)
+static void bootblock_ec_init(void)
{
/*
* Tables from CORE_PEI
@@ -108,9 +93,9 @@ void mainboard_config_superio(void)
outb(0x01, SCH555x_RUNTIME_IOBASE + SCH555x_RUNTIME_UNK1);
outb(0x0f, SCH555x_RUNTIME_IOBASE + SCH555x_RUNTIME_LED);
- // Magic EC init
- ec_init();
+ // Perform bootblock EC initialization
+ bootblock_ec_init();
- // Magic EC init is needed for UART1 initialization to work
+ // Bootblock EC initialization is required for UART1 to work
sch555x_enable_serial(SERIAL_DEV, CONFIG_TTYS0_BASE);
}
diff --git a/src/mainboard/dell/optiplex_9020/cmos.default b/src/mainboard/dell/optiplex_9020/cmos.default
index 7bccc80e51..1909abcb9f 100644
--- a/src/mainboard/dell/optiplex_9020/cmos.default
+++ b/src/mainboard/dell/optiplex_9020/cmos.default
@@ -3,3 +3,4 @@ debug_level=Debug
nmi=Disable
power_on_after_fail=Disable
iommu=Disable
+fan_full_speed=Disable
diff --git a/src/mainboard/dell/optiplex_9020/cmos.layout b/src/mainboard/dell/optiplex_9020/cmos.layout
index 72ff9c4bee..4a1496a878 100644
--- a/src/mainboard/dell/optiplex_9020/cmos.layout
+++ b/src/mainboard/dell/optiplex_9020/cmos.layout
@@ -22,7 +22,10 @@ entries
409 2 e 5 power_on_after_fail
# turn iommu on or off
-412 1 e 6 iommu
+411 1 e 6 iommu
+
+# coreboot config options: EC
+412 1 e 1 fan_full_speed
# coreboot config options: check sums
984 16 h 0 check_sum
diff --git a/src/mainboard/dell/optiplex_9020/mainboard.c b/src/mainboard/dell/optiplex_9020/mainboard.c
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
index c834fea5d3..0b7829c736 100644
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
--- a/src/mainboard/dell/optiplex_9020/mainboard.c
+++ b/src/mainboard/dell/optiplex_9020/mainboard.c
@@ -1,7 +1,12 @@
/* SPDX-License-Identifier: GPL-2.0-only */
+#include <bootstate.h>
+#include <cpu/x86/msr.h>
#include <device/device.h>
#include <drivers/intel/gma/int15.h>
+#include <option.h>
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
+#include <southbridge/intel/common/gpio.h>
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
+#include "sch5555_ec.h"
static void mainboard_enable(struct device *dev)
{
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
@@ -13,3 +18,385 @@ static void mainboard_enable(struct device *dev)
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
struct chip_operations mainboard_ops = {
.enable_dev = mainboard_enable,
};
+
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
+#define HWM_TAB_ADD_TEMP_TARGET 1
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
+#define HWM_TAB_PKG_POWER_ANY 0xffff
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
+#define CHASSIS_TYPE_UNKNOWN 0xff
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
+
+struct hwm_tab_entry {
+ uint16_t addr;
+ uint8_t val;
+ uint8_t flags;
+ uint16_t pkg_power;
+};
+
+struct hwm_tab_entry HWM_TAB3[] = {
+ { 0x005, 0x33, 0, 0xffff },
+ { 0x018, 0x2f, 0, 0xffff },
+ { 0x019, 0x2f, 0, 0xffff },
+ { 0x01a, 0x2f, 0, 0xffff },
+ { 0x080, 0x00, 0, 0xffff },
+ { 0x081, 0x00, 0, 0xffff },
+ { 0x083, 0xbb, 0, 0xffff },
+ { 0x085, 0x8a, 0, 0x0010 },
+ { 0x086, 0x4c, 0, 0x0010 },
+ { 0x08a, 0x66, 0, 0x0010 },
+ { 0x08b, 0x5b, 0, 0x0010 },
+ { 0x090, 0x65, 0, 0xffff },
+ { 0x091, 0x70, 0, 0xffff },
+ { 0x092, 0x86, 0, 0xffff },
+ { 0x096, 0xa4, 0, 0xffff },
+ { 0x097, 0xa4, 0, 0xffff },
+ { 0x098, 0xa4, 0, 0xffff },
+ { 0x09b, 0xa4, 0, 0xffff },
+ { 0x0a0, 0x0e, 0, 0xffff },
+ { 0x0a1, 0x0e, 0, 0xffff },
+ { 0x0ae, 0x7c, 0, 0xffff },
+ { 0x0af, 0x86, 0, 0xffff },
+ { 0x0b0, 0x9a, 0, 0xffff },
+ { 0x0b3, 0x9a, 0, 0xffff },
+ { 0x0b6, 0x08, 0, 0xffff },
+ { 0x0b7, 0x08, 0, 0xffff },
+ { 0x0ea, 0x64, 0, 0x0020 },
+ { 0x0ea, 0x5c, 0, 0x0010 },
+ { 0x0ef, 0xff, 0, 0xffff },
+ { 0x0f8, 0x15, 0, 0xffff },
+ { 0x0f9, 0x00, 0, 0xffff },
+ { 0x0f0, 0x30, 0, 0xffff },
+ { 0x0fd, 0x01, 0, 0xffff },
+ { 0x1a1, 0x00, 0, 0xffff },
+ { 0x1a2, 0x00, 0, 0xffff },
+ { 0x1b1, 0x08, 0, 0xffff },
+ { 0x1be, 0x99, 0, 0xffff },
+ { 0x280, 0xa0, 0, 0x0010 },
+ { 0x281, 0x0f, 0, 0x0010 },
+ { 0x282, 0x03, 0, 0xffff },
+ { 0x283, 0x0a, 0, 0xffff },
+ { 0x284, 0x80, 0, 0xffff },
+ { 0x285, 0x03, 0, 0xffff },
+ { 0x288, 0x68, 0, 0x0010 },
+ { 0x289, 0x10, 0, 0x0010 },
+ { 0x28a, 0x03, 0, 0xffff },
+ { 0x28b, 0x0a, 0, 0xffff },
+ { 0x28c, 0x80, 0, 0xffff },
+ { 0x28d, 0x03, 0, 0xffff },
+};
+
+struct hwm_tab_entry HWM_TAB4[] = {
+ { 0x005, 0x33, 0, 0xffff },
+ { 0x018, 0x2f, 0, 0xffff },
+ { 0x019, 0x2f, 0, 0xffff },
+ { 0x01a, 0x2f, 0, 0xffff },
+ { 0x080, 0x00, 0, 0xffff },
+ { 0x081, 0x00, 0, 0xffff },
+ { 0x083, 0xbb, 0, 0xffff },
+ { 0x085, 0x99, 0, 0x0020 },
+ { 0x085, 0xad, 0, 0x0010 },
+ { 0x086, 0x1c, 0, 0xffff },
+ { 0x08a, 0x39, 0, 0x0020 },
+ { 0x08a, 0x41, 0, 0x0010 },
+ { 0x08b, 0x76, 0, 0x0020 },
+ { 0x08b, 0x8b, 0, 0x0010 },
+ { 0x090, 0x5e, 0, 0xffff },
+ { 0x091, 0x5e, 0, 0xffff },
+ { 0x092, 0x86, 0, 0xffff },
+ { 0x096, 0xa4, 0, 0xffff },
+ { 0x097, 0xa4, 0, 0xffff },
+ { 0x098, 0xa4, 0, 0xffff },
+ { 0x09b, 0xa4, 0, 0xffff },
+ { 0x0a0, 0x0a, 0, 0xffff },
+ { 0x0a1, 0x0a, 0, 0xffff },
+ { 0x0ae, 0x7c, 0, 0xffff },
+ { 0x0af, 0x7c, 0, 0xffff },
+ { 0x0b0, 0x9a, 0, 0xffff },
+ { 0x0b3, 0x7c, 0, 0xffff },
+ { 0x0b6, 0x08, 0, 0xffff },
+ { 0x0b7, 0x08, 0, 0xffff },
+ { 0x0ea, 0x64, 0, 0x0020 },
+ { 0x0ea, 0x5c, 0, 0x0010 },
+ { 0x0ef, 0xff, 0, 0xffff },
+ { 0x0f8, 0x15, 0, 0xffff },
+ { 0x0f9, 0x00, 0, 0xffff },
+ { 0x0f0, 0x30, 0, 0xffff },
+ { 0x0fd, 0x01, 0, 0xffff },
+ { 0x1a1, 0x00, 0, 0xffff },
+ { 0x1a2, 0x00, 0, 0xffff },
+ { 0x1b1, 0x08, 0, 0xffff },
+ { 0x1be, 0x90, 0, 0xffff },
+ { 0x280, 0x94, 0, 0x0020 },
+ { 0x281, 0x11, 0, 0x0020 },
+ { 0x280, 0x94, 0, 0x0010 },
+ { 0x281, 0x11, 0, 0x0010 },
+ { 0x282, 0x03, 0, 0xffff },
+ { 0x283, 0x0a, 0, 0xffff },
+ { 0x284, 0x80, 0, 0xffff },
+ { 0x285, 0x03, 0, 0xffff },
+ { 0x288, 0x28, 0, 0x0020 },
+ { 0x289, 0x0a, 0, 0x0020 },
+ { 0x288, 0x28, 0, 0x0010 },
+ { 0x289, 0x0a, 0, 0x0010 },
+ { 0x28a, 0x03, 0, 0xffff },
+ { 0x28b, 0x0a, 0, 0xffff },
+ { 0x28c, 0x80, 0, 0xffff },
+ { 0x28d, 0x03, 0, 0xffff },
+};
+
+struct hwm_tab_entry HWM_TAB5[] = {
+ { 0x005, 0x33, 0, 0xffff },
+ { 0x018, 0x2f, 0, 0xffff },
+ { 0x019, 0x2f, 0, 0xffff },
+ { 0x01a, 0x2f, 0, 0xffff },
+ { 0x080, 0x00, 0, 0xffff },
+ { 0x081, 0x00, 0, 0xffff },
+ { 0x083, 0xbb, 0, 0xffff },
+ { 0x085, 0x66, 0, 0x0020 },
+ { 0x085, 0x5d, 0, 0x0010 },
+ { 0x086, 0x1c, 0, 0xffff },
+ { 0x08a, 0x39, 0, 0x0020 },
+ { 0x08a, 0x41, 0, 0x0010 },
+ { 0x08b, 0x76, 0, 0x0020 },
+ { 0x08b, 0x80, 0, 0x0010 },
+ { 0x090, 0x5d, 0, 0x0020 },
+ { 0x090, 0x5e, 0, 0x0010 },
+ { 0x091, 0x5e, 0, 0xffff },
+ { 0x092, 0x86, 0, 0xffff },
+ { 0x096, 0xa4, 0, 0xffff },
+ { 0x097, 0xa4, 0, 0xffff },
+ { 0x098, 0xa3, 0, 0x0020 },
+ { 0x098, 0xa4, 0, 0x0010 },
+ { 0x09b, 0xa4, 0, 0xffff },
+ { 0x0a0, 0x08, 0, 0xffff },
+ { 0x0a1, 0x0a, 0, 0xffff },
+ { 0x0ae, 0x7c, 0, 0xffff },
+ { 0x0af, 0x7c, 0, 0xffff },
+ { 0x0b0, 0x9a, 0, 0xffff },
+ { 0x0b3, 0x7c, 0, 0xffff },
+ { 0x0b6, 0x08, 0, 0xffff },
+ { 0x0b7, 0x08, 0, 0xffff },
+ { 0x0ea, 0x64, 0, 0x0020 },
+ { 0x0ea, 0x5c, 0, 0x0010 },
+ { 0x0ef, 0xff, 0, 0xffff },
+ { 0x0f8, 0x15, 0, 0xffff },
+ { 0x0f9, 0x00, 0, 0xffff },
+ { 0x0f0, 0x30, 0, 0xffff },
+ { 0x0fd, 0x01, 0, 0xffff },
+ { 0x1a1, 0x00, 0, 0xffff },
+ { 0x1a2, 0x00, 0, 0xffff },
+ { 0x1b1, 0x08, 0, 0xffff },
+ { 0x1be, 0x98, 0, 0x0020 },
+ { 0x1be, 0x90, 0, 0x0010 },
+ { 0x280, 0x94, 0, 0x0020 },
+ { 0x281, 0x11, 0, 0x0020 },
+ { 0x280, 0x94, 0, 0x0010 },
+ { 0x281, 0x11, 0, 0x0010 },
+ { 0x282, 0x03, 0, 0xffff },
+ { 0x283, 0x0a, 0, 0xffff },
+ { 0x284, 0x80, 0, 0xffff },
+ { 0x285, 0x03, 0, 0xffff },
+ { 0x288, 0x28, 0, 0x0020 },
+ { 0x289, 0x0a, 0, 0x0020 },
+ { 0x288, 0x28, 0, 0x0010 },
+ { 0x289, 0x0a, 0, 0x0010 },
+ { 0x28a, 0x03, 0, 0xffff },
+ { 0x28b, 0x0a, 0, 0xffff },
+ { 0x28c, 0x80, 0, 0xffff },
+ { 0x28d, 0x03, 0, 0xffff },
+};
+
+struct hwm_tab_entry HWM_TAB6[] = {
+ { 0x005, 0x33, 0, 0xffff },
+ { 0x018, 0x2f, 0, 0xffff },
+ { 0x019, 0x2f, 0, 0xffff },
+ { 0x01a, 0x2f, 0, 0xffff },
+ { 0x080, 0x00, 0, 0xffff },
+ { 0x081, 0x00, 0, 0xffff },
+ { 0x083, 0xbb, 0, 0xffff },
+ { 0x085, 0x98, 0, 0xffff },
+ { 0x086, 0x3c, 0, 0xffff },
+ { 0x08a, 0x39, 0, 0x0020 },
+ { 0x08a, 0x3d, 0, 0x0010 },
+ { 0x08b, 0x44, 0, 0x0020 },
+ { 0x08b, 0x51, 0, 0x0010 },
+ { 0x090, 0x61, 0, 0xffff },
+ { 0x091, 0x6d, 0, 0xffff },
+ { 0x092, 0x86, 0, 0xffff },
+ { 0x096, 0xa4, 0, 0xffff },
+ { 0x097, 0xa4, 0, 0xffff },
+ { 0x098, 0x9f, 0, 0x0020 },
+ { 0x098, 0xa4, 0, 0x0010 },
+ { 0x09b, 0xa4, 0, 0xffff },
+ { 0x0a0, 0x0e, 0, 0xffff },
+ { 0x0a1, 0x0e, 0, 0xffff },
+ { 0x0ae, 0x7c, 0, 0xffff },
+ { 0x0af, 0x7c, 0, 0xffff },
+ { 0x0b0, 0x9b, 0, 0x0020 },
+ { 0x0b0, 0x98, 0, 0x0010 },
+ { 0x0b3, 0x9a, 0, 0xffff },
+ { 0x0b6, 0x08, 0, 0xffff },
+ { 0x0b7, 0x08, 0, 0xffff },
+ { 0x0ea, 0x64, 0, 0x0020 },
+ { 0x0ea, 0x5c, 0, 0x0010 },
+ { 0x0ef, 0xff, 0, 0xffff },
+ { 0x0f8, 0x15, 0, 0xffff },
+ { 0x0f9, 0x00, 0, 0xffff },
+ { 0x0f0, 0x30, 0, 0xffff },
+ { 0x0fd, 0x01, 0, 0xffff },
+ { 0x1a1, 0x00, 0, 0xffff },
+ { 0x1a2, 0x00, 0, 0xffff },
+ { 0x1b1, 0x08, 0, 0xffff },
+ { 0x1be, 0x9a, 0, 0x0020 },
+ { 0x1be, 0x96, 0, 0x0010 },
+ { 0x280, 0x94, 0, 0x0020 },
+ { 0x281, 0x11, 0, 0x0020 },
+ { 0x280, 0x94, 0, 0x0010 },
+ { 0x281, 0x11, 0, 0x0010 },
+ { 0x282, 0x03, 0, 0xffff },
+ { 0x283, 0x0a, 0, 0xffff },
+ { 0x284, 0x80, 0, 0xffff },
+ { 0x285, 0x03, 0, 0xffff },
+ { 0x288, 0x94, 0, 0x0020 },
+ { 0x289, 0x11, 0, 0x0020 },
+ { 0x288, 0x94, 0, 0x0010 },
+ { 0x289, 0x11, 0, 0x0010 },
+ { 0x28a, 0x03, 0, 0xffff },
+ { 0x28b, 0x0a, 0, 0xffff },
+ { 0x28c, 0x80, 0, 0xffff },
+ { 0x28d, 0x03, 0, 0xffff },
+};
+
+static uint8_t get_chassis_type(void)
+{
+ uint8_t gpio_chassis_type;
+
+ // Read chassis type from GPIO
+ gpio_chassis_type = get_gpio(70) << 3 | get_gpio(38) << 2 |
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
+ get_gpio(17) << 1 | get_gpio(1);
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
+
+ printk(BIOS_DEBUG, "GPIO chassis type = %#x\n", gpio_chassis_type);
+
+ // Turn it into internal chassis index
+ switch (gpio_chassis_type) {
+ case 0x08:
+ case 0x0a:
+ return 4;
+ case 0x0b:
+ return 3;
+ case 0x0c:
+ return 5;
+ case 0x0d: // SFF
+ case 0x0e:
+ case 0x0f:
+ return 6;
+ default:
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
+ return CHASSIS_TYPE_UNKNOWN;
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
+ }
+
+}
+
+static uint8_t get_temp_target(void)
+{
+ uint8_t val = rdmsr(0x1a2).lo >> 8 & 0xff;
+ if (!val)
+ val = 20;
+ return 0x95 - val;
+}
+
+static uint16_t get_pkg_power(void)
+{
+ uint8_t rapl_power_unit = rdmsr(0x606).lo & 0xf;
+ if (rapl_power_unit)
+ rapl_power_unit = 2 << (rapl_power_unit - 1);
+ uint16_t pkg_power_info = rdmsr(0x614).lo & 0x7fff;
+ if (pkg_power_info / rapl_power_unit > 0x41)
+ return 32;
+ else
+ return 16;
+}
+
+static void apply_hwm_tab(struct hwm_tab_entry *arr, size_t size)
+{
+ uint8_t temp_target = get_temp_target();
+ uint16_t pkg_power = get_pkg_power();
+
+ printk(BIOS_DEBUG, "Temp target = %#x\n", temp_target);
+ printk(BIOS_DEBUG, "Package power = %#x\n", pkg_power);
+
+ for (size_t i = 0; i < size; ++i) {
+ // Skip entry if it doesn't apply for this package power
+ if (arr[i].pkg_power != pkg_power &&
+ arr[i].pkg_power != HWM_TAB_PKG_POWER_ANY)
+ continue;
+
+ uint8_t val = arr[i].val;
+
+ // Add temp target to value if requested (current tables never do)
+ if (arr[i].flags & HWM_TAB_ADD_TEMP_TARGET)
+ val += temp_target;
+
+ // Perform write
+ ec_write(1, arr[i].addr, val);
+
+ }
+}
+
+static void sch5555_ec_hwm_init(void *arg)
+{
+ uint8_t chassis_type, saved_2fc;
+
+ printk(BIOS_DEBUG, "OptiPlex 9020 late HWM init\n");
+
+ saved_2fc = ec_read(1, 0x2fc);
+ ec_write(1, 0x2fc, 0xa0);
+ ec_write(1, 0x2fd, 0x32);
+
+ chassis_type = get_chassis_type();
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
+
+ if (chassis_type != CHASSIS_TYPE_UNKNOWN) {
+ printk(BIOS_DEBUG, "Chassis type = %#x\n", chassis_type);
+ } else {
+ printk(BIOS_DEBUG, "WARNING: Unknown chassis type\n");
+ }
+
+ // Apply HWM table based on chassis type
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
+ switch (chassis_type) {
+ case 3:
+ apply_hwm_tab(HWM_TAB3, ARRAY_SIZE(HWM_TAB3));
+ break;
+ case 4:
+ apply_hwm_tab(HWM_TAB4, ARRAY_SIZE(HWM_TAB4));
+ break;
+ case 5:
+ apply_hwm_tab(HWM_TAB5, ARRAY_SIZE(HWM_TAB5));
+ break;
+ case 6:
+ apply_hwm_tab(HWM_TAB6, ARRAY_SIZE(HWM_TAB6));
+ break;
+ }
+
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
+ // NOTE: vendor firmware applies these when "max core address" > 2
+ // i think this is always the case
+ ec_write(1, 0x9e, 0x30);
+ ec_write(1, 0xeb, ec_read(1, 0xea));
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
+
+ ec_write(1, 0x2fc, saved_2fc);
+
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
+ // Apply full speed fan config if requested or if the chassis type is unknown
+ if (chassis_type == CHASSIS_TYPE_UNKNOWN || get_uint_option("fan_full_speed", 0)) {
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
+ printk(BIOS_DEBUG, "Setting full fan speed\n");
+ ec_write(1, 0x80, 0x60 | ec_read(1, 0x80));
+ ec_write(1, 0x81, 0x60 | ec_read(1, 0x81));
+ }
+
+ ec_read(1, 0xb8);
+
+ if ((chassis_type == 4 || chassis_type == 5) && ec_read(1, 0x26) == 0) {
+ ec_write(1, 0xa0, ec_read(1, 0xa0) & 0xfb);
+ ec_write(1, 0xa1, ec_read(1, 0xa1) & 0xfb);
+ ec_write(1, 0xa2, ec_read(1, 0xa2) & 0xfb);
+ ec_write(1, 0x8a, 0x99);
+ ec_write(1, 0x8b, 0x47);
+ ec_write(1, 0x8c, 0x91);
+ }
+}
+
+BOOT_STATE_INIT_ENTRY(BS_POST_DEVICE, BS_ON_EXIT, sch5555_ec_hwm_init, NULL);
diff --git a/src/mainboard/dell/optiplex_9020/sch5555_ec.c b/src/mainboard/dell/optiplex_9020/sch5555_ec.c
new file mode 100644
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
index 0000000000..a1067ac063
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
--- /dev/null
+++ b/src/mainboard/dell/optiplex_9020/sch5555_ec.c
@@ -0,0 +1,54 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
+#include <arch/io.h>
+#include <device/pnp_ops.h>
+#include <superio/smsc/sch555x/sch555x.h>
+#include "sch5555_ec.h"
+
+uint8_t ec_read(uint8_t addr1, uint16_t addr2)
+{
+ // clear ec-to-host mailbox
+ uint8_t tmp = inb(SCH555x_EMI_IOBASE + 1);
+ outb(tmp, SCH555x_EMI_IOBASE + 1);
+
+ // send address
+ outw(0 | 0x8001, SCH555x_EMI_IOBASE + 2);
+ outw((addr1 * 2) | 0x100, SCH555x_EMI_IOBASE + 4);
+
+ outw(4 | 0x8002, SCH555x_EMI_IOBASE + 2);
+ outl(addr2 << 16, SCH555x_EMI_IOBASE + 4);
+
+ // send message to ec
+ outb(1, SCH555x_EMI_IOBASE);
+
+ // wait for ack
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
+ for (size_t retry = 0; retry < 0xfff; ++retry)
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
+ if (inb(SCH555x_EMI_IOBASE + 1) & 1)
+ break;
+
+ // read result
+ outw(4 | 0x8000, SCH555x_EMI_IOBASE + 2);
+ return inb(SCH555x_EMI_IOBASE + 4);
+}
+
+void ec_write(uint8_t addr1, uint16_t addr2, uint8_t val)
+{
+ // clear ec-to-host mailbox
+ uint8_t tmp = inb(SCH555x_EMI_IOBASE + 1);
+ outb(tmp, SCH555x_EMI_IOBASE + 1);
+
+ // send address and value
+ outw(0 | 0x8001, SCH555x_EMI_IOBASE + 2);
+ outw((addr1 * 2) | 0x101, SCH555x_EMI_IOBASE + 4);
+
+ outw(4 | 0x8002, SCH555x_EMI_IOBASE + 2);
+ outl(val | (addr2 << 16), SCH555x_EMI_IOBASE + 4);
+
+ // send message to ec
+ outb(1, SCH555x_EMI_IOBASE);
+
+ // wait for ack
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
+ for (size_t retry = 0; retry < 0xfff; ++retry)
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
+ if (inb(SCH555x_EMI_IOBASE + 1) & 1)
+ break;
+}
diff --git a/src/mainboard/dell/optiplex_9020/sch5555_ec.h b/src/mainboard/dell/optiplex_9020/sch5555_ec.h
new file mode 100644
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
index 0000000000..7e399e8e74
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
--- /dev/null
+++ b/src/mainboard/dell/optiplex_9020/sch5555_ec.h
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
@@ -0,0 +1,10 @@
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
+/* SPDX-License-Identifier: GPL-2.0-only */
+
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
+#ifndef __SCH5555_EC_H__
+#define __SCH5555_EC_H__
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
+
+uint8_t ec_read(uint8_t addr1, uint16_t addr2);
+
+void ec_write(uint8_t addr1, uint16_t addr2, uint8_t val);
9020 sff/mt: fix bad gpio read on hwm patch sff happened to work, but mt would not boot with the patch, because it called die() on unknown chassis type, and the gpio happened to have a bad value in the old patch, because it wasn't reading the right gpio. i tested the fix on the old patch, but then decided to use mate's new patch because instead of calling die(), it simply boots with fan control disabled (max fan speed in that case), if this happens again. mt and sff have both been tested with this new version of the patch. both of them boot, and they both have proper fan control. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-21 20:55:57 +00:00
+
+#endif
add mate's patch for 9020 sff/mt fan controls see: https://review.coreboot.org/c/coreboot/+/81529 what i've merged is patchset 4. i had to rebase it slightly, because the libreboot version has the iommu toggle on cmos configs, which are files that mate's patch also changes, leading to merge conflict. Signed-off-by: Leah Rowe <leah@libreboot.org>
2024-04-07 02:45:32 +00:00
--
2.39.2