Message ID | 20180704124923.32483-6-drjones@redhat.com |
---|---|
State | New |
Headers | show |
Series | hw/arm/virt: Introduce cpu topology support | expand |
On Wed, Jul 04, 2018 at 02:49:22PM +0200, Andrew Jones wrote: > The ACPI PPTT table supports topology descriptions for ACPI > guests. Note, while a DT boot Linux guest with a non-flat CPU > topology will see socket and core IDs being sequential integers > starting from zero, e.g. with -smp 4,sockets=2,cores=2,threads=1 > > a DT boot produces > > cpu: 0 package_id: 0 core_id: 0 > cpu: 1 package_id: 0 core_id: 1 > cpu: 2 package_id: 1 core_id: 0 > cpu: 3 package_id: 1 core_id: 1 > > an ACPI boot produces > > cpu: 0 package_id: 36 core_id: 0 > cpu: 1 package_id: 36 core_id: 1 > cpu: 2 package_id: 96 core_id: 2 > cpu: 3 package_id: 96 core_id: 3 > > This is due to several reasons: > > 1) DT cpu nodes do not have an equivalent field to what the PPTT > ACPI Processor ID must be, i.e. something equal to the MADT CPU > UID or equal to the UID of an ACPI processor container. In both > ACPI cases those are platform dependant IDs assigned by the > vendor. > > 2) While QEMU is the vendor for a guest, if the topology specifies > SMT (> 1 thread), then, with ACPI, it is impossible to assign a > core-id the same value as a package-id, thus it is not possible > to have package-id=0 and core-id=0. This is because package and > core containers must be in the same ACPI namespace and therefore > must have unique UIDs. > > 3) ACPI processor containers are not required for PPTT tables to > be used and, due to the limitations of which IDs are selected > described above in (2), they are not helpful for QEMU, so we > don't build them with this patch. In the absence of them, Linux > assigns its own unique IDs. The maintainers have chosen not to use > counters from zero, but rather ACPI table offsets, which explains > why the numbers are so much larger than with DT. > > 4) When there is no SMT (threads=1) the core IDs for ACPI boot guests > match the logical CPU IDs, because these IDs must be equal to the > MADT CPU UID (as no processor containers are present), and QEMU > uses the logical CPU ID for these MADT IDs. > > Cc: Igor Mammedov <imammedo@redhat.com> > Cc: Shannon Zhao <zhaoshenglong@huawei.com> > Cc: "Michael S. Tsirkin" <mst@redhat.com> > Signed-off-by: Andrew Jones <drjones@redhat.com> > --- > hw/acpi/aml-build.c | 50 +++++++++++++++++++++++++++++++++++++ > hw/arm/virt-acpi-build.c | 5 ++++ > include/hw/acpi/aml-build.h | 2 ++ > 3 files changed, 57 insertions(+) > > diff --git a/hw/acpi/aml-build.c b/hw/acpi/aml-build.c > index 1e43cd736de9..37e8f5182ae9 100644 > --- a/hw/acpi/aml-build.c > +++ b/hw/acpi/aml-build.c > @@ -24,6 +24,7 @@ > #include "hw/acpi/aml-build.h" > #include "qemu/bswap.h" > #include "qemu/bitops.h" > +#include "sysemu/cpus.h" > #include "sysemu/numa.h" > > static GArray *build_alloc_array(void) > @@ -1679,6 +1680,55 @@ void build_slit(GArray *table_data, BIOSLinker *linker) > table_data->len - slit_start, 1, NULL, NULL); > } > > +/* > + * ACPI 6.2 Processor Properties Topology Table (PPTT) > + */ > +static void build_cpu_hierarchy(GArray *tbl, uint32_t flags, > + uint32_t parent, uint32_t id) > +{ > + build_append_byte(tbl, 0); /* Type 0 - processor */ > + build_append_byte(tbl, 20); /* Length, no private resources */ > + build_append_int_noprefix(tbl, 0, 2); /* Reserved */ > + build_append_int_noprefix(tbl, flags, 4); > + build_append_int_noprefix(tbl, parent, 4); > + build_append_int_noprefix(tbl, id, 4); > + build_append_int_noprefix(tbl, 0, 4); /* Num private resources */ > +} > + > +void build_pptt(GArray *table_data, BIOSLinker *linker, int possible_cpus) > +{ > + int pptt_start = table_data->len; > + int uid = 0, cpus = 0, socket; > + > + acpi_data_push(table_data, sizeof(AcpiTableHeader)); > + > + for (socket = 0; cpus < possible_cpus; socket++) { > + uint32_t socket_offset = table_data->len - pptt_start; > + int core; > + > + build_cpu_hierarchy(table_data, 1, 0, socket); > + > + for (core = 0; core < smp_cores; core++) { > + uint32_t core_offset = table_data->len - pptt_start; > + int thread; > + > + if (smp_threads > 1) { > + build_cpu_hierarchy(table_data, 0, socket_offset, core); > + for (thread = 0; thread < smp_threads; thread++) { > + build_cpu_hierarchy(table_data, 2, core_offset, uid++); > + } > + } else { > + build_cpu_hierarchy(table_data, 2, socket_offset, uid++); It just occurred to me that the uid++'s above should be something like uid[i++]'s, where uid[] is an argument to the function instead. Otherwise this general aml builder function isn't written generally enough. > + } > + } > + cpus += smp_cores * smp_threads; > + } > + > + build_header(linker, table_data, > + (void *)(table_data->data + pptt_start), "PPTT", > + table_data->len - pptt_start, 1, NULL, NULL); > +} > + > /* build rev1/rev3/rev5.1 FADT */ > void build_fadt(GArray *tbl, BIOSLinker *linker, const AcpiFadtData *f, > const char *oem_id, const char *oem_table_id) > diff --git a/hw/arm/virt-acpi-build.c b/hw/arm/virt-acpi-build.c > index 1d1fc824da6f..aa77e1f018d9 100644 > --- a/hw/arm/virt-acpi-build.c > +++ b/hw/arm/virt-acpi-build.c > @@ -832,6 +832,11 @@ void virt_acpi_build(VirtMachineState *vms, AcpiBuildTables *tables) > acpi_add_table(table_offsets, tables_blob); > build_madt(tables_blob, tables->linker, vms); > > + if (!vmc->ignore_cpu_topology) { > + acpi_add_table(table_offsets, tables_blob); > + build_pptt(tables_blob, tables->linker, possible_cpus(vms)); > + } > + > acpi_add_table(table_offsets, tables_blob); > build_gtdt(tables_blob, tables->linker, vms); > > diff --git a/include/hw/acpi/aml-build.h b/include/hw/acpi/aml-build.h > index 6c36903c0a5d..2b0fde6bd417 100644 > --- a/include/hw/acpi/aml-build.h > +++ b/include/hw/acpi/aml-build.h > @@ -414,6 +414,8 @@ void build_srat_memory(AcpiSratMemoryAffinity *numamem, uint64_t base, > > void build_slit(GArray *table_data, BIOSLinker *linker); > > +void build_pptt(GArray *table_data, BIOSLinker *linker, int possible_cpus); > + > void build_fadt(GArray *tbl, BIOSLinker *linker, const AcpiFadtData *f, > const char *oem_id, const char *oem_table_id); > #endif > -- > 2.17.1 >
On Wed, 4 Jul 2018 14:49:22 +0200 Andrew Jones <drjones@redhat.com> wrote: > The ACPI PPTT table supports topology descriptions for ACPI > guests. Note, while a DT boot Linux guest with a non-flat CPU > topology will see socket and core IDs being sequential integers > starting from zero, e.g. with -smp 4,sockets=2,cores=2,threads=1 > > a DT boot produces > > cpu: 0 package_id: 0 core_id: 0 > cpu: 1 package_id: 0 core_id: 1 > cpu: 2 package_id: 1 core_id: 0 > cpu: 3 package_id: 1 core_id: 1 > > an ACPI boot produces > > cpu: 0 package_id: 36 core_id: 0 > cpu: 1 package_id: 36 core_id: 1 > cpu: 2 package_id: 96 core_id: 2 > cpu: 3 package_id: 96 core_id: 3 > > This is due to several reasons: > > 1) DT cpu nodes do not have an equivalent field to what the PPTT > ACPI Processor ID must be, i.e. something equal to the MADT CPU > UID or equal to the UID of an ACPI processor container. In both > ACPI cases those are platform dependant IDs assigned by the > vendor. > > 2) While QEMU is the vendor for a guest, if the topology specifies > SMT (> 1 thread), then, with ACPI, it is impossible to assign a > core-id the same value as a package-id, thus it is not possible > to have package-id=0 and core-id=0. This is because package and > core containers must be in the same ACPI namespace and therefore > must have unique UIDs. > > 3) ACPI processor containers are not required for PPTT tables to > be used and, due to the limitations of which IDs are selected > described above in (2), they are not helpful for QEMU, so we > don't build them with this patch. In the absence of them, Linux > assigns its own unique IDs. The maintainers have chosen not to use > counters from zero, but rather ACPI table offsets, which explains > why the numbers are so much larger than with DT. > > 4) When there is no SMT (threads=1) the core IDs for ACPI boot guests > match the logical CPU IDs, because these IDs must be equal to the > MADT CPU UID (as no processor containers are present), and QEMU > uses the logical CPU ID for these MADT IDs. > > Cc: Igor Mammedov <imammedo@redhat.com> > Cc: Shannon Zhao <zhaoshenglong@huawei.com> > Cc: "Michael S. Tsirkin" <mst@redhat.com> > Signed-off-by: Andrew Jones <drjones@redhat.com> > --- > hw/acpi/aml-build.c | 50 +++++++++++++++++++++++++++++++++++++ > hw/arm/virt-acpi-build.c | 5 ++++ > include/hw/acpi/aml-build.h | 2 ++ > 3 files changed, 57 insertions(+) > > diff --git a/hw/acpi/aml-build.c b/hw/acpi/aml-build.c > index 1e43cd736de9..37e8f5182ae9 100644 > --- a/hw/acpi/aml-build.c > +++ b/hw/acpi/aml-build.c > @@ -24,6 +24,7 @@ > #include "hw/acpi/aml-build.h" > #include "qemu/bswap.h" > #include "qemu/bitops.h" > +#include "sysemu/cpus.h" > #include "sysemu/numa.h" > > static GArray *build_alloc_array(void) > @@ -1679,6 +1680,55 @@ void build_slit(GArray *table_data, BIOSLinker *linker) > table_data->len - slit_start, 1, NULL, NULL); > } > > +/* > + * ACPI 6.2 Processor Properties Topology Table (PPTT) ACPI 6.2: 5.2.29.1 Processor hierarchy node structure (Type 0) > + */ > +static void build_cpu_hierarchy(GArray *tbl, uint32_t flags, > + uint32_t parent, uint32_t id) build_processor_hierarchy_node() > +{ > + build_append_byte(tbl, 0); /* Type 0 - processor */ > + build_append_byte(tbl, 20); /* Length, no private resources */ > + build_append_int_noprefix(tbl, 0, 2); /* Reserved */ > + build_append_int_noprefix(tbl, flags, 4); just being pedantic, even for obvious fields add a comment that matches field name in spec, like: build_append_int_noprefix(tbl, flags, 4); /* Parent */ > + build_append_int_noprefix(tbl, parent, 4); > + build_append_int_noprefix(tbl, id, 4); > + build_append_int_noprefix(tbl, 0, 4); /* Num private resources */ put comment on new line if it doesn't fit into 80limit but use full field name from spec > +} > + > +void build_pptt(GArray *table_data, BIOSLinker *linker, int possible_cpus) > +{ > + int pptt_start = table_data->len; > + int uid = 0, cpus = 0, socket; > + > + acpi_data_push(table_data, sizeof(AcpiTableHeader)); > + > + for (socket = 0; cpus < possible_cpus; socket++) { probably should use possible_cpus->cpus[] here to iterate over cpus and maybe socket_id... from there as well > + uint32_t socket_offset = table_data->len - pptt_start; > + int core; > + > + build_cpu_hierarchy(table_data, 1, 0, socket); build_cpu_hierarchy(table_data, ACPI_PROC_HEIRARHY_PACKAGE, 0 /* no parent */ , socket); > + > + for (core = 0; core < smp_cores; core++) { > + uint32_t core_offset = table_data->len - pptt_start; > + int thread; > + > + if (smp_threads > 1) { > + build_cpu_hierarchy(table_data, 0, socket_offset, core); > + for (thread = 0; thread < smp_threads; thread++) { maybe set/use core_id and thread_id from possible_cpus instead of making up ID numbers here? > + build_cpu_hierarchy(table_data, 2, core_offset, uid++); > + } > + } else { > + build_cpu_hierarchy(table_data, 2, socket_offset, uid++); > + } > + } > + cpus += smp_cores * smp_threads; > + } > + > + build_header(linker, table_data, > + (void *)(table_data->data + pptt_start), "PPTT", > + table_data->len - pptt_start, 1, NULL, NULL); > +} > + > /* build rev1/rev3/rev5.1 FADT */ > void build_fadt(GArray *tbl, BIOSLinker *linker, const AcpiFadtData *f, > const char *oem_id, const char *oem_table_id) > diff --git a/hw/arm/virt-acpi-build.c b/hw/arm/virt-acpi-build.c > index 1d1fc824da6f..aa77e1f018d9 100644 > --- a/hw/arm/virt-acpi-build.c > +++ b/hw/arm/virt-acpi-build.c > @@ -832,6 +832,11 @@ void virt_acpi_build(VirtMachineState *vms, AcpiBuildTables *tables) > acpi_add_table(table_offsets, tables_blob); > build_madt(tables_blob, tables->linker, vms); > > + if (!vmc->ignore_cpu_topology) { > + acpi_add_table(table_offsets, tables_blob); > + build_pptt(tables_blob, tables->linker, possible_cpus(vms)); > + } > + > acpi_add_table(table_offsets, tables_blob); > build_gtdt(tables_blob, tables->linker, vms); > > diff --git a/include/hw/acpi/aml-build.h b/include/hw/acpi/aml-build.h > index 6c36903c0a5d..2b0fde6bd417 100644 > --- a/include/hw/acpi/aml-build.h > +++ b/include/hw/acpi/aml-build.h > @@ -414,6 +414,8 @@ void build_srat_memory(AcpiSratMemoryAffinity *numamem, uint64_t base, > > void build_slit(GArray *table_data, BIOSLinker *linker); > > +void build_pptt(GArray *table_data, BIOSLinker *linker, int possible_cpus); > + > void build_fadt(GArray *tbl, BIOSLinker *linker, const AcpiFadtData *f, > const char *oem_id, const char *oem_table_id); > #endif
diff --git a/hw/acpi/aml-build.c b/hw/acpi/aml-build.c index 1e43cd736de9..37e8f5182ae9 100644 --- a/hw/acpi/aml-build.c +++ b/hw/acpi/aml-build.c @@ -24,6 +24,7 @@ #include "hw/acpi/aml-build.h" #include "qemu/bswap.h" #include "qemu/bitops.h" +#include "sysemu/cpus.h" #include "sysemu/numa.h" static GArray *build_alloc_array(void) @@ -1679,6 +1680,55 @@ void build_slit(GArray *table_data, BIOSLinker *linker) table_data->len - slit_start, 1, NULL, NULL); } +/* + * ACPI 6.2 Processor Properties Topology Table (PPTT) + */ +static void build_cpu_hierarchy(GArray *tbl, uint32_t flags, + uint32_t parent, uint32_t id) +{ + build_append_byte(tbl, 0); /* Type 0 - processor */ + build_append_byte(tbl, 20); /* Length, no private resources */ + build_append_int_noprefix(tbl, 0, 2); /* Reserved */ + build_append_int_noprefix(tbl, flags, 4); + build_append_int_noprefix(tbl, parent, 4); + build_append_int_noprefix(tbl, id, 4); + build_append_int_noprefix(tbl, 0, 4); /* Num private resources */ +} + +void build_pptt(GArray *table_data, BIOSLinker *linker, int possible_cpus) +{ + int pptt_start = table_data->len; + int uid = 0, cpus = 0, socket; + + acpi_data_push(table_data, sizeof(AcpiTableHeader)); + + for (socket = 0; cpus < possible_cpus; socket++) { + uint32_t socket_offset = table_data->len - pptt_start; + int core; + + build_cpu_hierarchy(table_data, 1, 0, socket); + + for (core = 0; core < smp_cores; core++) { + uint32_t core_offset = table_data->len - pptt_start; + int thread; + + if (smp_threads > 1) { + build_cpu_hierarchy(table_data, 0, socket_offset, core); + for (thread = 0; thread < smp_threads; thread++) { + build_cpu_hierarchy(table_data, 2, core_offset, uid++); + } + } else { + build_cpu_hierarchy(table_data, 2, socket_offset, uid++); + } + } + cpus += smp_cores * smp_threads; + } + + build_header(linker, table_data, + (void *)(table_data->data + pptt_start), "PPTT", + table_data->len - pptt_start, 1, NULL, NULL); +} + /* build rev1/rev3/rev5.1 FADT */ void build_fadt(GArray *tbl, BIOSLinker *linker, const AcpiFadtData *f, const char *oem_id, const char *oem_table_id) diff --git a/hw/arm/virt-acpi-build.c b/hw/arm/virt-acpi-build.c index 1d1fc824da6f..aa77e1f018d9 100644 --- a/hw/arm/virt-acpi-build.c +++ b/hw/arm/virt-acpi-build.c @@ -832,6 +832,11 @@ void virt_acpi_build(VirtMachineState *vms, AcpiBuildTables *tables) acpi_add_table(table_offsets, tables_blob); build_madt(tables_blob, tables->linker, vms); + if (!vmc->ignore_cpu_topology) { + acpi_add_table(table_offsets, tables_blob); + build_pptt(tables_blob, tables->linker, possible_cpus(vms)); + } + acpi_add_table(table_offsets, tables_blob); build_gtdt(tables_blob, tables->linker, vms); diff --git a/include/hw/acpi/aml-build.h b/include/hw/acpi/aml-build.h index 6c36903c0a5d..2b0fde6bd417 100644 --- a/include/hw/acpi/aml-build.h +++ b/include/hw/acpi/aml-build.h @@ -414,6 +414,8 @@ void build_srat_memory(AcpiSratMemoryAffinity *numamem, uint64_t base, void build_slit(GArray *table_data, BIOSLinker *linker); +void build_pptt(GArray *table_data, BIOSLinker *linker, int possible_cpus); + void build_fadt(GArray *tbl, BIOSLinker *linker, const AcpiFadtData *f, const char *oem_id, const char *oem_table_id); #endif
The ACPI PPTT table supports topology descriptions for ACPI guests. Note, while a DT boot Linux guest with a non-flat CPU topology will see socket and core IDs being sequential integers starting from zero, e.g. with -smp 4,sockets=2,cores=2,threads=1 a DT boot produces cpu: 0 package_id: 0 core_id: 0 cpu: 1 package_id: 0 core_id: 1 cpu: 2 package_id: 1 core_id: 0 cpu: 3 package_id: 1 core_id: 1 an ACPI boot produces cpu: 0 package_id: 36 core_id: 0 cpu: 1 package_id: 36 core_id: 1 cpu: 2 package_id: 96 core_id: 2 cpu: 3 package_id: 96 core_id: 3 This is due to several reasons: 1) DT cpu nodes do not have an equivalent field to what the PPTT ACPI Processor ID must be, i.e. something equal to the MADT CPU UID or equal to the UID of an ACPI processor container. In both ACPI cases those are platform dependant IDs assigned by the vendor. 2) While QEMU is the vendor for a guest, if the topology specifies SMT (> 1 thread), then, with ACPI, it is impossible to assign a core-id the same value as a package-id, thus it is not possible to have package-id=0 and core-id=0. This is because package and core containers must be in the same ACPI namespace and therefore must have unique UIDs. 3) ACPI processor containers are not required for PPTT tables to be used and, due to the limitations of which IDs are selected described above in (2), they are not helpful for QEMU, so we don't build them with this patch. In the absence of them, Linux assigns its own unique IDs. The maintainers have chosen not to use counters from zero, but rather ACPI table offsets, which explains why the numbers are so much larger than with DT. 4) When there is no SMT (threads=1) the core IDs for ACPI boot guests match the logical CPU IDs, because these IDs must be equal to the MADT CPU UID (as no processor containers are present), and QEMU uses the logical CPU ID for these MADT IDs. Cc: Igor Mammedov <imammedo@redhat.com> Cc: Shannon Zhao <zhaoshenglong@huawei.com> Cc: "Michael S. Tsirkin" <mst@redhat.com> Signed-off-by: Andrew Jones <drjones@redhat.com> --- hw/acpi/aml-build.c | 50 +++++++++++++++++++++++++++++++++++++ hw/arm/virt-acpi-build.c | 5 ++++ include/hw/acpi/aml-build.h | 2 ++ 3 files changed, 57 insertions(+)