| Message ID | 1525854869-13975-1-git-send-email-jingqi.liu@intel.com |
|---|---|
| State | New |
| Headers | show |
| Series | [v1,1/7] hmat acpi: Build Memory Subsystem Address Range Structre(s) in ACPI HMAT | expand |
On 05/09/2018 03:34 AM, Liu Jingqi wrote: In the subject line: s/Structre/Structure/ Long subject line; read 'git shortlog -30' to get a feel for more typical subjects, and try to keep it at 60 characters or less (you still want legible subjects in an 80-column window even when git adds indentation and/or abbreviated commit id prefixes) > HMAT is defined in ACPI 6.2: 5.2.27 Heterogeneous Memory Attribute Table (HMAT). > The specification references below link: > http://www.uefi.org/sites/default/files/resources/ACPI_6_2.pdf Also, your mail was sent without any In-Reply-To: or References: headers, meaning it was not threaded to a 0/7 cover letter, but instead 7 different top-level threads. That makes it harder to track replies to your thread.
> -----Original Message----- > From: Eric Blake [mailto:eblake@redhat.com] > Sent: Wednesday, May 9, 2018 9:13 PM > To: Liu, Jingqi <jingqi.liu@intel.com>; pbonzini@redhat.com; rth@twiddle.net; > ehabkost@redhat.com > Cc: imammedo@redhat.com; qemu-devel@nongnu.org; mst@redhat.com > Subject: Re: [Qemu-devel] [PATCH v1 1/7] hmat acpi: Build Memory Subsystem > Address Range Structre(s) in ACPI HMAT > > On 05/09/2018 03:34 AM, Liu Jingqi wrote: > > In the subject line: s/Structre/Structure/ > > Long subject line; read 'git shortlog -30' to get a feel for more typical subjects, > and try to keep it at 60 characters or less (you still want legible subjects in an 80- > column window even when git adds indentation and/or abbreviated commit id > prefixes) > > > HMAT is defined in ACPI 6.2: 5.2.27 Heterogeneous Memory Attribute Table > (HMAT). > > The specification references below link: > > http://www.uefi.org/sites/default/files/resources/ACPI_6_2.pdf > > Also, your mail was sent without any In-Reply-To: or References: > headers, meaning it was not threaded to a 0/7 cover letter, but instead > 7 different top-level threads. That makes it harder to track replies to your > thread. > > -- > Eric Blake, Principal Software Engineer > Red Hat, Inc. +1-919-301-3266 > Virtualization: qemu.org | libvirt.org Hi Eric, Thanks for your reviewing. I will correct these issues. Jingqi Liu
On Wed, May 09, 2018 at 04:34:29PM +0800, Liu Jingqi wrote: > HMAT is defined in ACPI 6.2: 5.2.27 Heterogeneous Memory Attribute Table (HMAT). > The specification references below link: > http://www.uefi.org/sites/default/files/resources/ACPI_6_2.pdf > > It describes the memory attributes, such as memory side cache > attributes and bandwidth and latency details, related to the > System Physical Address (SPA) Memory Ranges. The software is > expected to use this information as hint for optimization. > > This structure describes the System Physical Address(SPA) range > occupied by memory subsystem and its associativity with processor > proximity domain as well as hint for memory usage. > [...] > +/* > + * The Proximity Domain of System Physical Address ranges defined > + * in the HMAT, NFIT and SRAT tables shall match each other. > + */ > +static void hmat_build_spa(GArray *table_data, PCMachineState *pcms) > +{ > + GSList *device_list = NULL; > + AcpiHmatSpaRange *hmat_spa; > + uint64_t mem_base, next_base, mem_len; > + int node; > + > + next_base = 0; > + for (node = 0; node < nb_numa_nodes; node++) { > + mem_len = numa_info[node].node_mem; > + if (!mem_len) { > + continue; > + } > + > + mem_base = next_base; > + next_base = mem_base + mem_len; > + > + /* Cut out the 640K hole */ > + if (mem_base <= HOLE_640K_START && > + next_base > HOLE_640K_START) { > + mem_len -= next_base - HOLE_640K_START; > + if (mem_len > 0) { > + hmat_spa = acpi_data_push(table_data, sizeof(*hmat_spa)); > + hmat_build_spa_info(hmat_spa, mem_base, mem_len, node); > + } > + > + /* Check for the rare case: 640K < RAM < 1M */ > + if (next_base <= HOLE_640K_END) { > + next_base = HOLE_640K_END; > + continue; > + } > + mem_base = HOLE_640K_END; > + mem_len = next_base - HOLE_640K_END; > + } > + > + /* Cut out the ACPI_PCI hole */ > + if (mem_base <= pcms->below_4g_mem_size && > + next_base > pcms->below_4g_mem_size) { > + mem_len -= next_base - pcms->below_4g_mem_size; > + if (mem_len > 0) { > + hmat_spa = acpi_data_push(table_data, sizeof(*hmat_spa)); > + hmat_build_spa_info(hmat_spa, mem_base, mem_len, node); > + } > + mem_base = 1ULL << 32; > + mem_len = next_base - pcms->below_4g_mem_size; > + next_base = mem_base + mem_len; > + } This duplicates very complex logic that already exists in build_srat(). We need to make the existing logic reusable. > + hmat_spa = acpi_data_push(table_data, sizeof(*hmat_spa)); > + hmat_build_spa_info(hmat_spa, mem_base, mem_len, node); > + } > + > + /* Build HMAT SPA structures for PC-DIMM devices. */ > + object_child_foreach(qdev_get_machine(), pc_dimm_device_list, &device_list); > + > + for (; device_list; device_list = device_list->next) { > + PCDIMMDevice *dimm = device_list->data; > + mem_base = object_property_get_uint(OBJECT(dimm), PC_DIMM_ADDR_PROP, > + NULL); > + mem_len = object_property_get_uint(OBJECT(dimm), PC_DIMM_SIZE_PROP, > + NULL); > + node = object_property_get_uint(OBJECT(dimm), PC_DIMM_NODE_PROP, NULL); > + > + hmat_spa = acpi_data_push(table_data, sizeof(*hmat_spa)); > + hmat_build_spa_info(hmat_spa, mem_base, mem_len, node); > + } > +}
> -----Original Message----- > From: Eduardo Habkost [mailto:ehabkost@redhat.com] > Sent: Friday, May 11, 2018 10:41 PM > To: Liu, Jingqi <jingqi.liu@intel.com> > Cc: pbonzini@redhat.com; rth@twiddle.net; mst@redhat.com; > imammedo@redhat.com; marcel.apfelbaum@gmail.com; qemu- > devel@nongnu.org > Subject: Re: [PATCH v1 1/7] hmat acpi: Build Memory Subsystem Address Range > Structre(s) in ACPI HMAT > > On Wed, May 09, 2018 at 04:34:29PM +0800, Liu Jingqi wrote: > > HMAT is defined in ACPI 6.2: 5.2.27 Heterogeneous Memory Attribute Table > (HMAT). > > The specification references below link: > > http://www.uefi.org/sites/default/files/resources/ACPI_6_2.pdf > > > > It describes the memory attributes, such as memory side cache > > attributes and bandwidth and latency details, related to the System > > Physical Address (SPA) Memory Ranges. The software is expected to use > > this information as hint for optimization. > > > > This structure describes the System Physical Address(SPA) range > > occupied by memory subsystem and its associativity with processor > > proximity domain as well as hint for memory usage. > > > [...] > > +/* > > + * The Proximity Domain of System Physical Address ranges defined > > + * in the HMAT, NFIT and SRAT tables shall match each other. > > + */ > > +static void hmat_build_spa(GArray *table_data, PCMachineState *pcms) > > +{ > > + GSList *device_list = NULL; > > + AcpiHmatSpaRange *hmat_spa; > > + uint64_t mem_base, next_base, mem_len; > > + int node; > > + > > + next_base = 0; > > + for (node = 0; node < nb_numa_nodes; node++) { > > + mem_len = numa_info[node].node_mem; > > + if (!mem_len) { > > + continue; > > + } > > + > > + mem_base = next_base; > > + next_base = mem_base + mem_len; > > + > > + /* Cut out the 640K hole */ > > + if (mem_base <= HOLE_640K_START && > > + next_base > HOLE_640K_START) { > > + mem_len -= next_base - HOLE_640K_START; > > + if (mem_len > 0) { > > + hmat_spa = acpi_data_push(table_data, sizeof(*hmat_spa)); > > + hmat_build_spa_info(hmat_spa, mem_base, mem_len, node); > > + } > > + > > + /* Check for the rare case: 640K < RAM < 1M */ > > + if (next_base <= HOLE_640K_END) { > > + next_base = HOLE_640K_END; > > + continue; > > + } > > + mem_base = HOLE_640K_END; > > + mem_len = next_base - HOLE_640K_END; > > + } > > + > > + /* Cut out the ACPI_PCI hole */ > > + if (mem_base <= pcms->below_4g_mem_size && > > + next_base > pcms->below_4g_mem_size) { > > + mem_len -= next_base - pcms->below_4g_mem_size; > > + if (mem_len > 0) { > > + hmat_spa = acpi_data_push(table_data, sizeof(*hmat_spa)); > > + hmat_build_spa_info(hmat_spa, mem_base, mem_len, node); > > + } > > + mem_base = 1ULL << 32; > > + mem_len = next_base - pcms->below_4g_mem_size; > > + next_base = mem_base + mem_len; > > + } > > This duplicates very complex logic that already exists in build_srat(). We need to > make the existing logic reusable. > Thanks Eduardo. Agree. We need to make them reusable. HMAT connects memory address ranges to proximity domains, which must be consistent with the affinity attributes defined in other tables, such as SRAT and NFIT. Jingqi > > > + hmat_spa = acpi_data_push(table_data, sizeof(*hmat_spa)); > > + hmat_build_spa_info(hmat_spa, mem_base, mem_len, node); > > + } > > + > > + /* Build HMAT SPA structures for PC-DIMM devices. */ > > + object_child_foreach(qdev_get_machine(), pc_dimm_device_list, > > + &device_list); > > + > > + for (; device_list; device_list = device_list->next) { > > + PCDIMMDevice *dimm = device_list->data; > > + mem_base = object_property_get_uint(OBJECT(dimm), > PC_DIMM_ADDR_PROP, > > + NULL); > > + mem_len = object_property_get_uint(OBJECT(dimm), > PC_DIMM_SIZE_PROP, > > + NULL); > > + node = object_property_get_uint(OBJECT(dimm), > > + PC_DIMM_NODE_PROP, NULL); > > + > > + hmat_spa = acpi_data_push(table_data, sizeof(*hmat_spa)); > > + hmat_build_spa_info(hmat_spa, mem_base, mem_len, node); > > + } > > +} > > -- > Eduardo
On Wed, 9 May 2018 16:34:29 +0800 Liu Jingqi <jingqi.liu@intel.com> wrote: > HMAT is defined in ACPI 6.2: 5.2.27 Heterogeneous Memory Attribute Table (HMAT). > The specification references below link: > http://www.uefi.org/sites/default/files/resources/ACPI_6_2.pdf > > It describes the memory attributes, such as memory side cache > attributes and bandwidth and latency details, related to the > System Physical Address (SPA) Memory Ranges. The software is > expected to use this information as hint for optimization. > > This structure describes the System Physical Address(SPA) range > occupied by memory subsystem and its associativity with processor > proximity domain as well as hint for memory usage. > > Signed-off-by: Liu Jingqi <jingqi.liu@intel.com> > --- > default-configs/x86_64-softmmu.mak | 1 + > hw/acpi/Makefile.objs | 1 + > hw/acpi/hmat.c | 174 +++++++++++++++++++++++++++++++++++++ > hw/acpi/hmat.h | 75 ++++++++++++++++ > hw/i386/acpi-build.c | 3 + > 5 files changed, 254 insertions(+) > create mode 100644 hw/acpi/hmat.c > create mode 100644 hw/acpi/hmat.h > > diff --git a/default-configs/x86_64-softmmu.mak b/default-configs/x86_64-softmmu.mak > index 0390b43..3b4a37d 100644 > --- a/default-configs/x86_64-softmmu.mak > +++ b/default-configs/x86_64-softmmu.mak > @@ -66,3 +66,4 @@ CONFIG_I2C=y > CONFIG_SEV=$(CONFIG_KVM) > CONFIG_VTD=y > CONFIG_AMD_IOMMU=y > +CONFIG_ACPI_HMAT=y > diff --git a/hw/acpi/Makefile.objs b/hw/acpi/Makefile.objs > index 11c35bc..21889fd 100644 > --- a/hw/acpi/Makefile.objs > +++ b/hw/acpi/Makefile.objs > @@ -6,6 +6,7 @@ common-obj-$(CONFIG_ACPI_MEMORY_HOTPLUG) += memory_hotplug.o > common-obj-$(CONFIG_ACPI_CPU_HOTPLUG) += cpu.o > common-obj-$(CONFIG_ACPI_NVDIMM) += nvdimm.o > common-obj-$(CONFIG_ACPI_VMGENID) += vmgenid.o > +common-obj-$(CONFIG_ACPI_HMAT) += hmat.o > common-obj-$(call lnot,$(CONFIG_ACPI_X86)) += acpi-stub.o > > common-obj-y += acpi_interface.o > diff --git a/hw/acpi/hmat.c b/hw/acpi/hmat.c > new file mode 100644 > index 0000000..bca1fbb > --- /dev/null > +++ b/hw/acpi/hmat.c > @@ -0,0 +1,174 @@ > +/* > + * HMAT ACPI Implementation > + * > + * Copyright(C) 2018 Intel Corporation. > + * > + * Author: > + * Liu jingqi <jingqi.liu@linux.intel.com> > + * > + * HMAT is defined in ACPI 6.2. > + * > + * This library is free software; you can redistribute it and/or > + * modify it under the terms of the GNU Lesser General Public > + * License as published by the Free Software Foundation; either > + * version 2 of the License, or (at your option) any later version. > + * > + * This library is distributed in the hope that it will be useful, > + * but WITHOUT ANY WARRANTY; without even the implied warranty of > + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU > + * Lesser General Public License for more details. > + * > + * You should have received a copy of the GNU Lesser General Public > + * License along with this library; if not, see <http://www.gnu.org/licenses/> > + */ > + > +#include "unistd.h" > +#include "fcntl.h" > +#include "qemu/osdep.h" > +#include "sysemu/numa.h" > +#include "hw/i386/pc.h" > +#include "hw/acpi/acpi.h" > +#include "hw/acpi/hmat.h" > +#include "hw/acpi/aml-build.h" > +#include "hw/nvram/fw_cfg.h" > +#include "hw/acpi/bios-linker-loader.h" > + > +#define HOLE_640K_START (640 * 1024) > +#define HOLE_640K_END (1024 * 1024) > + > +uint32_t initiator_pxm[MAX_NODES], target_pxm[MAX_NODES]; > +uint32_t num_initiator = 0, num_target = 0; > + > +static void hmat_build_spa_info(AcpiHmatSpaRange *spa, > + uint64_t base, uint64_t length, int node) > +{ > + int i; > + > + spa->type = ACPI_HMAT_SPA; > + spa->length = sizeof(*spa); > + spa->spa_base = base; > + spa->spa_length = length; all of above will break on big-endian host, that's one of the reasons we prefer new code/tables to use build_append_foo() API. Pls rewrite it using preferred build_append_int_noprefix() API, build_amd_iommu() can serve as an example. As comments above fields use exact field names from spec tables so one could easily match spec vs code. > + spa->flags = 0; > + > + for (i = 0; i < num_initiator; i++) { > + if (initiator_pxm[i] == node) { > + spa->proc_proximity = node; > + spa->flags |= HMAT_SPA_PROC_VALID; > + break; > + } > + } > + > + for (i = 0; i < num_target; i++) { > + if (target_pxm[i] == node) { > + spa->mem_proximity = node; > + spa->flags |= HMAT_SPA_MEM_VALID; > + break; > + } > + } > +} > + > +static int pc_dimm_device_list(Object *obj, void *opaque) > +{ > + GSList **list = opaque; > + > + if (object_dynamic_cast(obj, TYPE_PC_DIMM)) { > + *list = g_slist_append(*list, DEVICE(obj)); > + } > + > + object_child_foreach(obj, pc_dimm_device_list, opaque); > + return 0; > +} > + > +/* > + * The Proximity Domain of System Physical Address ranges defined > + * in the HMAT, NFIT and SRAT tables shall match each other. > + */ > +static void hmat_build_spa(GArray *table_data, PCMachineState *pcms) > +{ > + GSList *device_list = NULL; > + AcpiHmatSpaRange *hmat_spa; > + uint64_t mem_base, next_base, mem_len; > + int node; > + > + next_base = 0; > + for (node = 0; node < nb_numa_nodes; node++) { > + mem_len = numa_info[node].node_mem; > + if (!mem_len) { > + continue; > + } > + > + mem_base = next_base; > + next_base = mem_base + mem_len; > + > + /* Cut out the 640K hole */ > + if (mem_base <= HOLE_640K_START && > + next_base > HOLE_640K_START) { > + mem_len -= next_base - HOLE_640K_START; > + if (mem_len > 0) { > + hmat_spa = acpi_data_push(table_data, sizeof(*hmat_spa)); > + hmat_build_spa_info(hmat_spa, mem_base, mem_len, node); > + } > + > + /* Check for the rare case: 640K < RAM < 1M */ > + if (next_base <= HOLE_640K_END) { > + next_base = HOLE_640K_END; > + continue; > + } > + mem_base = HOLE_640K_END; > + mem_len = next_base - HOLE_640K_END; > + } > + > + /* Cut out the ACPI_PCI hole */ > + if (mem_base <= pcms->below_4g_mem_size && > + next_base > pcms->below_4g_mem_size) { > + mem_len -= next_base - pcms->below_4g_mem_size; > + if (mem_len > 0) { > + hmat_spa = acpi_data_push(table_data, sizeof(*hmat_spa)); > + hmat_build_spa_info(hmat_spa, mem_base, mem_len, node); > + } > + mem_base = 1ULL << 32; > + mem_len = next_base - pcms->below_4g_mem_size; > + next_base = mem_base + mem_len; > + } > + hmat_spa = acpi_data_push(table_data, sizeof(*hmat_spa)); > + hmat_build_spa_info(hmat_spa, mem_base, mem_len, node); > + } > + > + /* Build HMAT SPA structures for PC-DIMM devices. */ > + object_child_foreach(qdev_get_machine(), pc_dimm_device_list, &device_list); > + > + for (; device_list; device_list = device_list->next) { > + PCDIMMDevice *dimm = device_list->data; > + mem_base = object_property_get_uint(OBJECT(dimm), PC_DIMM_ADDR_PROP, > + NULL); > + mem_len = object_property_get_uint(OBJECT(dimm), PC_DIMM_SIZE_PROP, > + NULL); > + node = object_property_get_uint(OBJECT(dimm), PC_DIMM_NODE_PROP, NULL); > + > + hmat_spa = acpi_data_push(table_data, sizeof(*hmat_spa)); > + hmat_build_spa_info(hmat_spa, mem_base, mem_len, node); > + } > +} > + > +static void hmat_build_hma(GArray *hma, PCMachineState *pcms) > +{ > + /* Build HMAT Memory Subsystem Address Range. */ > + hmat_build_spa(hma, pcms); > +} > + > +void hmat_build_acpi(GArray *table_data, BIOSLinker *linker, > + MachineState *machine) > +{ > + PCMachineState *pcms = PC_MACHINE(machine); > + uint64_t hmat_start, hmat_len; > + > + hmat_start = table_data->len; > + acpi_data_push(table_data, sizeof(AcpiHmat)); > + > + hmat_build_hma(table_data, pcms); > + hmat_len = table_data->len - hmat_start; > + > + build_header(linker, table_data, > + (void *)(table_data->data + hmat_start), > + "HMAT", hmat_len, 1, NULL, NULL); > +} > diff --git a/hw/acpi/hmat.h b/hw/acpi/hmat.h > new file mode 100644 > index 0000000..841c698 > --- /dev/null > +++ b/hw/acpi/hmat.h > @@ -0,0 +1,75 @@ > +/* > + * HMAT ACPI Implementation Header > + * > + * Copyright(C) 2018 Intel Corporation. > + * > + * Author: > + * Liu jingqi <jingqi.liu@linux.intel.com> > + * > + * HMAT is defined in ACPI 6.2. > + * > + * This library is free software; you can redistribute it and/or > + * modify it under the terms of the GNU Lesser General Public > + * License as published by the Free Software Foundation; either > + * version 2 of the License, or (at your option) any later version. > + * > + * This library is distributed in the hope that it will be useful, > + * but WITHOUT ANY WARRANTY; without even the implied warranty of > + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU > + * Lesser General Public License for more details. > + * > + * You should have received a copy of the GNU Lesser General Public > + * License along with this library; if not, see <http://www.gnu.org/licenses/> > + */ > + > +#ifndef HMAT_H > +#define HMAT_H > + > +#include "qemu/osdep.h" > +#include "hw/acpi/acpi-defs.h" > +#include "hw/acpi/acpi.h" > +#include "hw/acpi/bios-linker-loader.h" > +#include "hw/acpi/aml-build.h" > + > +#define ACPI_HMAT_SPA 0 > + > +/* ACPI HMAT sub-structure header */ > +#define ACPI_HMAT_SUB_HEADER_DEF \ > + uint16_t type; \ > + uint16_t reserved0; \ > + uint32_t length; > + > +/* the values of AcpiHmatSpaRange flag */ > +enum { > + HMAT_SPA_PROC_VALID = 0x1, > + HMAT_SPA_MEM_VALID = 0x2, > + HMAT_SPA_RESERVATION_HINT = 0x4, > +}; > + > +/* > + * HMAT (Heterogeneous Memory Attributes Table) > + */ > +struct AcpiHmat { > + ACPI_TABLE_HEADER_DEF > + uint32_t reserved; > +} QEMU_PACKED; > +typedef struct AcpiHmat AcpiHmat; > + > +struct AcpiHmatSpaRange { > + ACPI_HMAT_SUB_HEADER_DEF > + uint16_t flags; > + uint16_t reserved1; > + uint32_t proc_proximity; > + uint32_t mem_proximity; > + uint32_t reserved2; > + uint64_t spa_base; > + uint64_t spa_length; > +} QEMU_PACKED; > +typedef struct AcpiHmatSpaRange AcpiHmatSpaRange; > + > +extern uint32_t initiator_pxm[MAX_NODES], target_pxm[MAX_NODES]; > +extern uint32_t num_initiator, num_target; > +void hmat_build_acpi(GArray *table_data, BIOSLinker *linker, > + MachineState *machine); > + > +#endif > diff --git a/hw/i386/acpi-build.c b/hw/i386/acpi-build.c > index 9bc6d97..0f9fbe4 100644 > --- a/hw/i386/acpi-build.c > +++ b/hw/i386/acpi-build.c > @@ -64,6 +64,7 @@ > #include "hw/i386/intel_iommu.h" > > #include "hw/acpi/ipmi.h" > +#include "hw/acpi/hmat.h" > > /* These are used to size the ACPI tables for -M pc-i440fx-1.7 and > * -M pc-i440fx-2.0. Even if the actual amount of AML generated grows > @@ -2713,6 +2714,8 @@ void acpi_build(AcpiBuildTables *tables, MachineState *machine) > acpi_add_table(table_offsets, tables_blob); > build_slit(tables_blob, tables->linker); > } > + acpi_add_table(table_offsets, tables_blob); > + hmat_build_acpi(tables_blob, tables->linker, machine); > } > if (acpi_get_mcfg(&mcfg)) { > acpi_add_table(table_offsets, tables_blob);
On Wed, 9 May 2018 16:34:29 +0800 Liu Jingqi <jingqi.liu@intel.com> wrote: series submission note Could you use cover letter when sending multipatch series please? It should describe whole series impact and provides anchor point for thread. Look for 'git --cover-letter' in git's man page.
On Tue, May 15, 2018 10:36 PM, Igor Mammedov <imammedo@redhat.com> wrote: > On Wed, 9 May 2018 16:34:29 +0800 > Liu Jingqi <jingqi.liu@intel.com> wrote: > > > HMAT is defined in ACPI 6.2: 5.2.27 Heterogeneous Memory Attribute Table > (HMAT). > > The specification references below link: > > http://www.uefi.org/sites/default/files/resources/ACPI_6_2.pdf > > > > It describes the memory attributes, such as memory side cache > > attributes and bandwidth and latency details, related to the System > > Physical Address (SPA) Memory Ranges. The software is expected to use > > this information as hint for optimization. > > > > This structure describes the System Physical Address(SPA) range > > occupied by memory subsystem and its associativity with processor > > proximity domain as well as hint for memory usage. > > > > Signed-off-by: Liu Jingqi <jingqi.liu@intel.com> > > --- > > default-configs/x86_64-softmmu.mak | 1 + > > hw/acpi/Makefile.objs | 1 + > > hw/acpi/hmat.c | 174 > +++++++++++++++++++++++++++++++++++++ > > hw/acpi/hmat.h | 75 ++++++++++++++++ > > hw/i386/acpi-build.c | 3 + > > 5 files changed, 254 insertions(+) > > create mode 100644 hw/acpi/hmat.c > > create mode 100644 hw/acpi/hmat.h > > > > diff --git a/default-configs/x86_64-softmmu.mak > > b/default-configs/x86_64-softmmu.mak > > index 0390b43..3b4a37d 100644 > > --- a/default-configs/x86_64-softmmu.mak > > +++ b/default-configs/x86_64-softmmu.mak > > @@ -66,3 +66,4 @@ CONFIG_I2C=y > > CONFIG_SEV=$(CONFIG_KVM) > > CONFIG_VTD=y > > CONFIG_AMD_IOMMU=y > > +CONFIG_ACPI_HMAT=y > > diff --git a/hw/acpi/Makefile.objs b/hw/acpi/Makefile.objs index > > 11c35bc..21889fd 100644 > > --- a/hw/acpi/Makefile.objs > > +++ b/hw/acpi/Makefile.objs > > @@ -6,6 +6,7 @@ common-obj-$(CONFIG_ACPI_MEMORY_HOTPLUG) += > > memory_hotplug.o > > common-obj-$(CONFIG_ACPI_CPU_HOTPLUG) += cpu.o > > common-obj-$(CONFIG_ACPI_NVDIMM) += nvdimm.o > > common-obj-$(CONFIG_ACPI_VMGENID) += vmgenid.o > > +common-obj-$(CONFIG_ACPI_HMAT) += hmat.o > > common-obj-$(call lnot,$(CONFIG_ACPI_X86)) += acpi-stub.o > > > > common-obj-y += acpi_interface.o > > diff --git a/hw/acpi/hmat.c b/hw/acpi/hmat.c new file mode 100644 > > index 0000000..bca1fbb > > --- /dev/null > > +++ b/hw/acpi/hmat.c > > @@ -0,0 +1,174 @@ > > +/* > > + * HMAT ACPI Implementation > > + * > > + * Copyright(C) 2018 Intel Corporation. > > + * > > + * Author: > > + * Liu jingqi <jingqi.liu@linux.intel.com> > > + * > > + * HMAT is defined in ACPI 6.2. > > + * > > + * This library is free software; you can redistribute it and/or > > + * modify it under the terms of the GNU Lesser General Public > > + * License as published by the Free Software Foundation; either > > + * version 2 of the License, or (at your option) any later version. > > + * > > + * This library is distributed in the hope that it will be useful, > > + * but WITHOUT ANY WARRANTY; without even the implied warranty of > > + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU > > + * Lesser General Public License for more details. > > + * > > + * You should have received a copy of the GNU Lesser General Public > > + * License along with this library; if not, see > > +<http://www.gnu.org/licenses/> */ > > + > > +#include "unistd.h" > > +#include "fcntl.h" > > +#include "qemu/osdep.h" > > +#include "sysemu/numa.h" > > +#include "hw/i386/pc.h" > > +#include "hw/acpi/acpi.h" > > +#include "hw/acpi/hmat.h" > > +#include "hw/acpi/aml-build.h" > > +#include "hw/nvram/fw_cfg.h" > > +#include "hw/acpi/bios-linker-loader.h" > > + > > +#define HOLE_640K_START (640 * 1024) > > +#define HOLE_640K_END (1024 * 1024) > > + > > +uint32_t initiator_pxm[MAX_NODES], target_pxm[MAX_NODES]; uint32_t > > +num_initiator = 0, num_target = 0; > > + > > +static void hmat_build_spa_info(AcpiHmatSpaRange *spa, > > + uint64_t base, uint64_t length, int > > +node) { > > + int i; > > + > > + spa->type = ACPI_HMAT_SPA; > > + spa->length = sizeof(*spa); > > + spa->spa_base = base; > > + spa->spa_length = length; > all of above will break on big-endian host, that's one of the reasons we prefer > new code/tables to use build_append_foo() API. > > Pls rewrite it using preferred build_append_int_noprefix() API, > build_amd_iommu() can serve as an example. > > As comments above fields use exact field names from spec tables so one could > easily match spec vs code. > Thanks for your reviewing. I will use preferred build_append_int_noprefix() API in next version. Jingqi Liu
diff --git a/default-configs/x86_64-softmmu.mak b/default-configs/x86_64-softmmu.mak index 0390b43..3b4a37d 100644 --- a/default-configs/x86_64-softmmu.mak +++ b/default-configs/x86_64-softmmu.mak @@ -66,3 +66,4 @@ CONFIG_I2C=y CONFIG_SEV=$(CONFIG_KVM) CONFIG_VTD=y CONFIG_AMD_IOMMU=y +CONFIG_ACPI_HMAT=y diff --git a/hw/acpi/Makefile.objs b/hw/acpi/Makefile.objs index 11c35bc..21889fd 100644 --- a/hw/acpi/Makefile.objs +++ b/hw/acpi/Makefile.objs @@ -6,6 +6,7 @@ common-obj-$(CONFIG_ACPI_MEMORY_HOTPLUG) += memory_hotplug.o common-obj-$(CONFIG_ACPI_CPU_HOTPLUG) += cpu.o common-obj-$(CONFIG_ACPI_NVDIMM) += nvdimm.o common-obj-$(CONFIG_ACPI_VMGENID) += vmgenid.o +common-obj-$(CONFIG_ACPI_HMAT) += hmat.o common-obj-$(call lnot,$(CONFIG_ACPI_X86)) += acpi-stub.o common-obj-y += acpi_interface.o diff --git a/hw/acpi/hmat.c b/hw/acpi/hmat.c new file mode 100644 index 0000000..bca1fbb --- /dev/null +++ b/hw/acpi/hmat.c @@ -0,0 +1,174 @@ +/* + * HMAT ACPI Implementation + * + * Copyright(C) 2018 Intel Corporation. + * + * Author: + * Liu jingqi <jingqi.liu@linux.intel.com> + * + * HMAT is defined in ACPI 6.2. + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library; if not, see <http://www.gnu.org/licenses/> + */ + +#include "unistd.h" +#include "fcntl.h" +#include "qemu/osdep.h" +#include "sysemu/numa.h" +#include "hw/i386/pc.h" +#include "hw/acpi/acpi.h" +#include "hw/acpi/hmat.h" +#include "hw/acpi/aml-build.h" +#include "hw/nvram/fw_cfg.h" +#include "hw/acpi/bios-linker-loader.h" + +#define HOLE_640K_START (640 * 1024) +#define HOLE_640K_END (1024 * 1024) + +uint32_t initiator_pxm[MAX_NODES], target_pxm[MAX_NODES]; +uint32_t num_initiator = 0, num_target = 0; + +static void hmat_build_spa_info(AcpiHmatSpaRange *spa, + uint64_t base, uint64_t length, int node) +{ + int i; + + spa->type = ACPI_HMAT_SPA; + spa->length = sizeof(*spa); + spa->spa_base = base; + spa->spa_length = length; + spa->flags = 0; + + for (i = 0; i < num_initiator; i++) { + if (initiator_pxm[i] == node) { + spa->proc_proximity = node; + spa->flags |= HMAT_SPA_PROC_VALID; + break; + } + } + + for (i = 0; i < num_target; i++) { + if (target_pxm[i] == node) { + spa->mem_proximity = node; + spa->flags |= HMAT_SPA_MEM_VALID; + break; + } + } +} + +static int pc_dimm_device_list(Object *obj, void *opaque) +{ + GSList **list = opaque; + + if (object_dynamic_cast(obj, TYPE_PC_DIMM)) { + *list = g_slist_append(*list, DEVICE(obj)); + } + + object_child_foreach(obj, pc_dimm_device_list, opaque); + return 0; +} + +/* + * The Proximity Domain of System Physical Address ranges defined + * in the HMAT, NFIT and SRAT tables shall match each other. + */ +static void hmat_build_spa(GArray *table_data, PCMachineState *pcms) +{ + GSList *device_list = NULL; + AcpiHmatSpaRange *hmat_spa; + uint64_t mem_base, next_base, mem_len; + int node; + + next_base = 0; + for (node = 0; node < nb_numa_nodes; node++) { + mem_len = numa_info[node].node_mem; + if (!mem_len) { + continue; + } + + mem_base = next_base; + next_base = mem_base + mem_len; + + /* Cut out the 640K hole */ + if (mem_base <= HOLE_640K_START && + next_base > HOLE_640K_START) { + mem_len -= next_base - HOLE_640K_START; + if (mem_len > 0) { + hmat_spa = acpi_data_push(table_data, sizeof(*hmat_spa)); + hmat_build_spa_info(hmat_spa, mem_base, mem_len, node); + } + + /* Check for the rare case: 640K < RAM < 1M */ + if (next_base <= HOLE_640K_END) { + next_base = HOLE_640K_END; + continue; + } + mem_base = HOLE_640K_END; + mem_len = next_base - HOLE_640K_END; + } + + /* Cut out the ACPI_PCI hole */ + if (mem_base <= pcms->below_4g_mem_size && + next_base > pcms->below_4g_mem_size) { + mem_len -= next_base - pcms->below_4g_mem_size; + if (mem_len > 0) { + hmat_spa = acpi_data_push(table_data, sizeof(*hmat_spa)); + hmat_build_spa_info(hmat_spa, mem_base, mem_len, node); + } + mem_base = 1ULL << 32; + mem_len = next_base - pcms->below_4g_mem_size; + next_base = mem_base + mem_len; + } + hmat_spa = acpi_data_push(table_data, sizeof(*hmat_spa)); + hmat_build_spa_info(hmat_spa, mem_base, mem_len, node); + } + + /* Build HMAT SPA structures for PC-DIMM devices. */ + object_child_foreach(qdev_get_machine(), pc_dimm_device_list, &device_list); + + for (; device_list; device_list = device_list->next) { + PCDIMMDevice *dimm = device_list->data; + mem_base = object_property_get_uint(OBJECT(dimm), PC_DIMM_ADDR_PROP, + NULL); + mem_len = object_property_get_uint(OBJECT(dimm), PC_DIMM_SIZE_PROP, + NULL); + node = object_property_get_uint(OBJECT(dimm), PC_DIMM_NODE_PROP, NULL); + + hmat_spa = acpi_data_push(table_data, sizeof(*hmat_spa)); + hmat_build_spa_info(hmat_spa, mem_base, mem_len, node); + } +} + +static void hmat_build_hma(GArray *hma, PCMachineState *pcms) +{ + /* Build HMAT Memory Subsystem Address Range. */ + hmat_build_spa(hma, pcms); +} + +void hmat_build_acpi(GArray *table_data, BIOSLinker *linker, + MachineState *machine) +{ + PCMachineState *pcms = PC_MACHINE(machine); + uint64_t hmat_start, hmat_len; + + hmat_start = table_data->len; + acpi_data_push(table_data, sizeof(AcpiHmat)); + + hmat_build_hma(table_data, pcms); + hmat_len = table_data->len - hmat_start; + + build_header(linker, table_data, + (void *)(table_data->data + hmat_start), + "HMAT", hmat_len, 1, NULL, NULL); +} diff --git a/hw/acpi/hmat.h b/hw/acpi/hmat.h new file mode 100644 index 0000000..841c698 --- /dev/null +++ b/hw/acpi/hmat.h @@ -0,0 +1,75 @@ +/* + * HMAT ACPI Implementation Header + * + * Copyright(C) 2018 Intel Corporation. + * + * Author: + * Liu jingqi <jingqi.liu@linux.intel.com> + * + * HMAT is defined in ACPI 6.2. + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library; if not, see <http://www.gnu.org/licenses/> + */ + +#ifndef HMAT_H +#define HMAT_H + +#include "qemu/osdep.h" +#include "hw/acpi/acpi-defs.h" +#include "hw/acpi/acpi.h" +#include "hw/acpi/bios-linker-loader.h" +#include "hw/acpi/aml-build.h" + +#define ACPI_HMAT_SPA 0 + +/* ACPI HMAT sub-structure header */ +#define ACPI_HMAT_SUB_HEADER_DEF \ + uint16_t type; \ + uint16_t reserved0; \ + uint32_t length; + +/* the values of AcpiHmatSpaRange flag */ +enum { + HMAT_SPA_PROC_VALID = 0x1, + HMAT_SPA_MEM_VALID = 0x2, + HMAT_SPA_RESERVATION_HINT = 0x4, +}; + +/* + * HMAT (Heterogeneous Memory Attributes Table) + */ +struct AcpiHmat { + ACPI_TABLE_HEADER_DEF + uint32_t reserved; +} QEMU_PACKED; +typedef struct AcpiHmat AcpiHmat; + +struct AcpiHmatSpaRange { + ACPI_HMAT_SUB_HEADER_DEF + uint16_t flags; + uint16_t reserved1; + uint32_t proc_proximity; + uint32_t mem_proximity; + uint32_t reserved2; + uint64_t spa_base; + uint64_t spa_length; +} QEMU_PACKED; +typedef struct AcpiHmatSpaRange AcpiHmatSpaRange; + +extern uint32_t initiator_pxm[MAX_NODES], target_pxm[MAX_NODES]; +extern uint32_t num_initiator, num_target; +void hmat_build_acpi(GArray *table_data, BIOSLinker *linker, + MachineState *machine); + +#endif diff --git a/hw/i386/acpi-build.c b/hw/i386/acpi-build.c index 9bc6d97..0f9fbe4 100644 --- a/hw/i386/acpi-build.c +++ b/hw/i386/acpi-build.c @@ -64,6 +64,7 @@ #include "hw/i386/intel_iommu.h" #include "hw/acpi/ipmi.h" +#include "hw/acpi/hmat.h" /* These are used to size the ACPI tables for -M pc-i440fx-1.7 and * -M pc-i440fx-2.0. Even if the actual amount of AML generated grows @@ -2713,6 +2714,8 @@ void acpi_build(AcpiBuildTables *tables, MachineState *machine) acpi_add_table(table_offsets, tables_blob); build_slit(tables_blob, tables->linker); } + acpi_add_table(table_offsets, tables_blob); + hmat_build_acpi(tables_blob, tables->linker, machine); } if (acpi_get_mcfg(&mcfg)) { acpi_add_table(table_offsets, tables_blob);
HMAT is defined in ACPI 6.2: 5.2.27 Heterogeneous Memory Attribute Table (HMAT). The specification references below link: http://www.uefi.org/sites/default/files/resources/ACPI_6_2.pdf It describes the memory attributes, such as memory side cache attributes and bandwidth and latency details, related to the System Physical Address (SPA) Memory Ranges. The software is expected to use this information as hint for optimization. This structure describes the System Physical Address(SPA) range occupied by memory subsystem and its associativity with processor proximity domain as well as hint for memory usage. Signed-off-by: Liu Jingqi <jingqi.liu@intel.com> --- default-configs/x86_64-softmmu.mak | 1 + hw/acpi/Makefile.objs | 1 + hw/acpi/hmat.c | 174 +++++++++++++++++++++++++++++++++++++ hw/acpi/hmat.h | 75 ++++++++++++++++ hw/i386/acpi-build.c | 3 + 5 files changed, 254 insertions(+) create mode 100644 hw/acpi/hmat.c create mode 100644 hw/acpi/hmat.h