[U-Boot] [UBOOT PATCH v5 2/3] x86: Add ACPI table support to QEMU
Saket Sinha
saket.sinha89 at gmail.com
Fri Aug 21 06:24:39 CEST 2015
Hi Bin,
Please find my response inline.
On Tue, Aug 18, 2015 at 12:36 PM, Bin Meng <bmeng.cn at gmail.com> wrote:
> Hi Saket,
>
> On Tue, Aug 18, 2015 at 3:29 AM, Saket Sinha <saket.sinha89 at gmail.com> wrote:
>> This patch mainly adds ACPI support to QEMU.
>> Verified by booting Linux kernel on QEMU i440FX and Q35.
>>
>> Signed-off-by: Saket Sinha <saket.sinha89 at gmail.com>
>> ---
>>
>> arch/x86/cpu/qemu/Makefile | 1 +
>> arch/x86/cpu/qemu/acpi.c | 179 +++++++++++++++++++++++++++++++++++++++++++++
>> 2 files changed, 180 insertions(+)
>> create mode 100644 arch/x86/cpu/qemu/acpi.c
>>
>> diff --git a/arch/x86/cpu/qemu/Makefile b/arch/x86/cpu/qemu/Makefile
>> index 9a66b16..8c3884c 100644
>> --- a/arch/x86/cpu/qemu/Makefile
>> +++ b/arch/x86/cpu/qemu/Makefile
>> @@ -8,4 +8,5 @@ ifndef CONFIG_EFI_STUB
>> obj-y += car.o dram.o
>> endif
>> obj-y += qemu.o
>> +obj-$(CONFIG_GENERATE_ACPI_TABLE) += acpi.o
>> obj-$(CONFIG_PCI) += pci.o
>> diff --git a/arch/x86/cpu/qemu/acpi.c b/arch/x86/cpu/qemu/acpi.c
>> new file mode 100644
>> index 0000000..7c981d0
>> --- /dev/null
>> +++ b/arch/x86/cpu/qemu/acpi.c
>> @@ -0,0 +1,179 @@
>> +/*
>> + * Copyright (C) 2015, Saket Sinha <saket.sinha89 at gmail.com>
>> + *
>> + * SPDX-License-Identifier: GPL-2.0+
>> + */
>> +
>> +#include <common.h>
>> +#include <asm/acpi_table.h>
>> +#include <asm/ioapic.h>
>> +#include <asm/tables.h>
>> +
>> +void acpi_create_fadt(struct acpi_fadt *fadt, struct acpi_facs *facs,
>> + void *dsdt)
>> +{
>> + acpi_header_t *header = &(fadt->header);
>> + u16 pmbase;
>> +
>> + pci_dev_t bdf = PCI_BDF(0, 0x1f, 0);
>> + pci_read_config_word(bdf, 0x40, &pmbase);
>> +
>> + /*
>> + * TODO(saket.sinha89 at gmail.com): wrong value
>> + * of pmbase by above function. Harcoding it to
>> + * correct value. Since no PCI register is
>> + * programmed Power Management Interface is
>> + * not working
>> + */
>
> Given you already know the root cause here (PMBASE is not programmed),
> can you program this register in the QEMU codes to make it work?
Please elaborate on this. Do you want me to hack QEMU code to verify
that it works ?
> you can check arch/x86/cpu/quark/Kconfig to use common names for these
> ACPI register blocks, like the one used in quark. (I remember I
> mentioned this comment before)
>
This is not clear. I know these registers are defined in
arch/x86/cpu/quark/Kconfig where configurable values can be assigned
to it.
But here I am trying to create fadt table structure which have these
registers for sure but I am trying to fill those values after reading
them from the hardware( in our case QEMU)
>> +
>> + pmbase = 0x0600;
>> +
>> + memset((void *)fadt, 0, sizeof(struct acpi_fadt));
>> + memcpy(header->signature, "FACP", 4);
>> + header->length = sizeof(struct acpi_fadt);
>> + header->revision = 3;
>> + memcpy(header->oem_id, OEM_ID, 6);
>> + memcpy(header->oem_table_id, ACPI_TABLE_CREATOR, 8);
>> + memcpy(header->asl_compiler_id, ASLC, 4);
>> + header->asl_compiler_revision = 0;
>> +
>> + fadt->firmware_ctrl = (unsigned long) facs;
>> + fadt->dsdt = (unsigned long) dsdt;
>> + fadt->model = 0x00;
>> + fadt->preferred_pm_profile = PM_MOBILE;
>> + fadt->sci_int = 0x9;
>> + fadt->smi_cmd = 0;
>> + fadt->acpi_enable = 0;
>> + fadt->acpi_disable = 0;
>> + fadt->s4bios_req = 0x0;
>> + fadt->pstate_cnt = 0;
>> + fadt->pm1a_evt_blk = pmbase;
>> + fadt->pm1b_evt_blk = 0x0;
>> + fadt->pm1a_cnt_blk = pmbase + 0x4;
>> + fadt->pm1b_cnt_blk = 0x0;
>> + fadt->pm2_cnt_blk = pmbase + 0x50;
>> + fadt->pm_tmr_blk = pmbase + 0x8;
>> + fadt->gpe0_blk = pmbase + 0x20;
>> + fadt->gpe1_blk = 0;
>> + fadt->pm1_evt_len = 4;
>> + /*
>> + * Upper word is reserved and
>> + * Linux complains about 32 bit
>> + */
>> + fadt->pm1_cnt_len = 2;
>> + fadt->pm2_cnt_len = 1;
>> + fadt->pm_tmr_len = 4;
>> + fadt->gpe0_blk_len = 16;
>> + fadt->gpe1_blk_len = 0;
>> + fadt->gpe1_base = 0;
>> + fadt->cst_cnt = 0;
>> + fadt->p_lvl2_lat = 1;
>> + fadt->p_lvl3_lat = 0x39;
>> + fadt->flush_size = 0;
>> + fadt->flush_stride = 0;
>> + fadt->duty_offset = 1;
>> + fadt->duty_width = 3;
>> + fadt->day_alrm = 0xd;
>> + fadt->mon_alrm = 0x00;
>> + fadt->century = 0x32;
>> + fadt->iapc_boot_arch = 0x00;
>> + fadt->flags = ACPI_FADT_WBINVD | ACPI_FADT_C1_SUPPORTED |
>> + ACPI_FADT_SLEEP_BUTTON | ACPI_FADT_S4_RTC_WAKE |
>> + ACPI_FADT_DOCKING_SUPPORTED | ACPI_FADT_RESET_REGISTER |
>> + ACPI_FADT_PLATFORM_CLOCK;
>> + fadt->reset_reg.space_id = ACPI_ADDRESS_SPACE_IO;
>> + fadt->reset_reg.bit_width = 8;
>> + fadt->reset_reg.bit_offset = 0;
>> + fadt->reset_reg.resv = 0;
>> + fadt->reset_reg.addrl = 0xcf9;
>> + fadt->reset_reg.addrh = 0;
>> + fadt->reset_value = 0x06;
>> + /*
>> + * Set X_FIRMWARE_CTRL only if FACS is
>> + * above 4GB. If X_FIRMWARE_CTRL is set,
>> + * then FIRMWARE_CTRL must be zero
>> + */
>> + fadt->x_firmware_ctl_l = 0;
>> + fadt->x_firmware_ctl_h = 0;
>> + fadt->x_dsdt_l = (unsigned long)dsdt;
>> + fadt->x_dsdt_h = 0;
>> + fadt->x_pm1a_evt_blk.space_id = 1;
>> + fadt->x_pm1a_evt_blk.bit_width = 32;
>> + fadt->x_pm1a_evt_blk.bit_offset = 0;
>> + fadt->x_pm1a_evt_blk.resv = 0;
>> + fadt->x_pm1a_evt_blk.addrl = pmbase;
>> + fadt->x_pm1a_evt_blk.addrh = 0x0;
>> + fadt->x_pm1b_evt_blk.space_id = 0;
>> + fadt->x_pm1b_evt_blk.bit_width = 0;
>> + fadt->x_pm1b_evt_blk.bit_offset = 0;
>> + fadt->x_pm1b_evt_blk.resv = 0;
>> + fadt->x_pm1b_evt_blk.addrl = 0x0;
>> + fadt->x_pm1b_evt_blk.addrh = 0x0;
>> + fadt->x_pm1a_cnt_blk.space_id = 1;
>> + /*
>> + * Upper word is reserved and
>> + * Linux complains about 32 bit
>> + */
>> + fadt->x_pm1a_cnt_blk.bit_width = 16;
>> + fadt->x_pm1a_cnt_blk.bit_offset = 0;
>> + fadt->x_pm1a_cnt_blk.resv = 0;
>> + fadt->x_pm1a_cnt_blk.addrl = pmbase + 0x4;
>> + fadt->x_pm1a_cnt_blk.addrh = 0x0;
>> + fadt->x_pm1b_cnt_blk.space_id = 0;
>> + fadt->x_pm1b_cnt_blk.bit_width = 0;
>> + fadt->x_pm1b_cnt_blk.bit_offset = 0;
>> + fadt->x_pm1b_cnt_blk.resv = 0;
>> + fadt->x_pm1b_cnt_blk.addrl = 0x0;
>> + fadt->x_pm1b_cnt_blk.addrh = 0x0;
>> + fadt->x_pm2_cnt_blk.space_id = 1;
>> + fadt->x_pm2_cnt_blk.bit_width = 8;
>> + fadt->x_pm2_cnt_blk.bit_offset = 0;
>> + fadt->x_pm2_cnt_blk.resv = 0;
>> + fadt->x_pm2_cnt_blk.addrl = pmbase + 0x50;
>> + fadt->x_pm2_cnt_blk.addrh = 0x0;
>> + fadt->x_pm_tmr_blk.space_id = 1;
>> + fadt->x_pm_tmr_blk.bit_width = 32;
>> + fadt->x_pm_tmr_blk.bit_offset = 0;
>> + fadt->x_pm_tmr_blk.resv = 0;
>> + fadt->x_pm_tmr_blk.addrl = pmbase + 0x8;
>> + fadt->x_pm_tmr_blk.addrh = 0x0;
>> + fadt->x_gpe0_blk.space_id = 1;
>> + fadt->x_gpe0_blk.bit_width = 128;
>> + fadt->x_gpe0_blk.bit_offset = 0;
>> + fadt->x_gpe0_blk.resv = 0;
>> + fadt->x_gpe0_blk.addrl = pmbase + 0x20;
>> + fadt->x_gpe0_blk.addrh = 0x0;
>> + fadt->x_gpe1_blk.space_id = 0;
>> + fadt->x_gpe1_blk.bit_width = 0;
>> + fadt->x_gpe1_blk.bit_offset = 0;
>> + fadt->x_gpe1_blk.resv = 0;
>> + fadt->x_gpe1_blk.addrl = 0x0;
>> + fadt->x_gpe1_blk.addrh = 0x0;
>> +
>> + header->checksum =
>> + table_compute_checksum((void *)fadt, header->length);
>> +}
>> +
>> +unsigned long acpi_fill_madt(unsigned long current)
>> +{
>> + /* create all subtables for processors */
>> + current = acpi_create_madt_lapics(current);
>> +
>> + /*
>> + * TODO(saket.sinha89 at gmail.com): get these
>> + * IRQ values from device tree
>> + */
>> + current += acpi_create_madt_ioapic
>> + ((struct acpi_madt_ioapic *)current, 2, IO_APIC_ADDR, 0);
>> + current += acpi_create_madt_irqoverride
>> + ((struct acpi_madt_irqoverride *)current, 0, 0, 2, 0);
>> + current += acpi_create_madt_irqoverride
>> + ((struct acpi_madt_irqoverride *)current, 0, 9, 9, 0xD);
>> + current += acpi_create_madt_irqoverride
>> + ((struct acpi_madt_irqoverride *)current, 0, 0xB, 0xB, 0xD);
>
> Nits: please use lower case hex numbers.
>
Addressed in patchset series v6.
>> + acpi_create_madt_lapic_nmi
>> + ((struct acpi_madt_lapic_nmi *)current, 0, 0, 0);
>> +
>> + return current;
>> +}
>> +
>> --
>
> Regards,
> Bin
Regards,
Saket Sinha
More information about the U-Boot
mailing list