Hi Alireza,
On 2/20/26 11:11, Alireza Sanaee wrote:
Add cache topology to PPTT table.
Signed-off-by: Alireza Sanaee <[email protected]>
---
hw/acpi/aml-build.c | 198 +++++++++++++++++++++++++++++++++++++--
hw/arm/virt-acpi-build.c | 8 +-
include/hw/acpi/cpu.h | 10 ++
3 files changed, 207 insertions(+), 9 deletions(-)
diff --git a/hw/acpi/aml-build.c b/hw/acpi/aml-build.c
index b75aa2336e..a8de45036c 100644
--- a/hw/acpi/aml-build.c
+++ b/hw/acpi/aml-build.c
@@ -32,6 +32,7 @@
#include "hw/pci/pci_bus.h"
#include "hw/pci/pci_bridge.h"
#include "qemu/cutils.h"
+#include "hw/core/cpu.h"
static GArray *build_alloc_array(void)
{
@@ -2148,6 +2149,102 @@ void build_spcr(GArray *table_data, BIOSLinker *linker,
}
acpi_table_end(linker, &table);
}
+
+static void build_cache_nodes(GArray *tbl, CPUCoreCaches *cache,
+ uint32_t next_offset)
+{
+ const uint8_t node_length = 24;
+ int start_len = tbl->len;
+ int val;
+
+ build_append_byte(tbl, 1); /* Type 1 - cache */
+ build_append_byte(tbl, node_length); /* Length */
+ build_append_int_noprefix(tbl, 0, 2); /* Reserved */
+ build_append_int_noprefix(tbl, 0x7f, 4); /* Flags */
+ build_append_int_noprefix(tbl, next_offset, 4); /* Offset of next cache up
*/
+ build_append_int_noprefix(tbl, cache->size, 4); /* Size */
+ build_append_int_noprefix(tbl, cache->sets, 4); /* Number of sets */
+ build_append_byte(tbl, cache->associativity); /* Associativity */
+ val = 0x3;
+ switch (cache->type) {
+ case INSTRUCTION_CACHE:
+ val |= (1 << 2); /* Instruction Cache */
+ break;
+ case DATA_CACHE:
+ val |= (0 << 2); /* Data Cache */
+ break;
+ case UNIFIED_CACHE:
+ val |= (3 << 2); /* Unified */
+ break;
+ }
+ build_append_byte(tbl, val); /* Attributes */
+ build_append_int_noprefix(tbl, cache->linesize, 2); /* Line size */
+ g_assert(tbl->len == start_len + node_length);
+}
Please document the ACPI spec version on top of build_cache_nodes() and
align the comment about the fields. See
build_processor_hierarchy_node(), as an example.
I thought "Cache ID" field was missing and length was wrong (28), but
then I realized that it optional/non-existent on older spec versions.
+/*
+ * builds caches from the top level (`level_high` parameter) to the bottom
+ * level (`level_low` parameter). It searches for caches found in
+ * systems' registers, and fills up the table. Then it updates the
+ * `data_offset` and `instr_offset` parameters with the offset of the data
+ * and instruction caches of the lowest level, respectively.
+ */
Please, start the comment with upper case. But, more important, I don't
think this comment is much helpful. I'd like to read more about the
meaning and need of 'data_offset' and 'instr_offset' in the context of
the ACPI table.
+static bool build_caches(GArray *table_data, uint32_t pptt_start,
+ int num_caches, CPUCoreCaches *caches,
+ uint8_t level_high, /* Inclusive */
+ uint8_t level_low, /* Inclusive */
+ uint32_t *data_offset,
+ uint32_t *instr_offset)
+{
+ uint32_t next_level_offset_data = 0, next_level_offset_instruction = 0;
+ uint32_t this_offset, next_offset = 0;
+ int c, level;
+ bool found_cache = false;
+
+ /* Walk caches from top to bottom */
+ for (level = level_high; level >= level_low; level--) {
+ for (c = 0; c < num_caches; c++) {
+ if (caches[c].level != level) {
+ continue;
+ }
+
+ /* Assume only unified above l1 for now */
+ this_offset = table_data->len - pptt_start;
Is 'this_offset' an invariant that can be hoisted out of the loops?
+ switch (caches[c].type) {
+ case INSTRUCTION_CACHE:
+ next_offset = next_level_offset_instruction;
+ break;
+ case DATA_CACHE:
+ next_offset = next_level_offset_data;
+ break;
+ case UNIFIED_CACHE:
+ /* Either is fine here */
+ next_offset = next_level_offset_instruction;
+ break;
+ }
+ build_cache_nodes(table_data, &caches[c], next_offset);
+ switch (caches[c].type) {
+ case INSTRUCTION_CACHE:
+ next_level_offset_instruction = this_offset;
+ break;
+ case DATA_CACHE:
+ next_level_offset_data = this_offset;
+ break;
+ case UNIFIED_CACHE:
+ next_level_offset_instruction = this_offset;
+ next_level_offset_data = this_offset;
+ break;
+ }
+ *data_offset = next_level_offset_data;
+ *instr_offset = next_level_offset_instruction;
+
+ found_cache = true;
+ }
+ }
+
+ return found_cache;
+}
+
/*
* ACPI spec, Revision 6.3
* 5.2.29 Processor Properties Topology Table (PPTT)
@@ -2158,11 +2255,32 @@ void build_pptt(GArray *table_data, BIOSLinker *linker,
MachineState *ms,
{
MachineClass *mc = MACHINE_GET_CLASS(ms);
CPUArchIdList *cpus = ms->possible_cpus;
- int64_t socket_id = -1, cluster_id = -1, core_id = -1;
- uint32_t socket_offset = 0, cluster_offset = 0, core_offset = 0;
+ uint32_t core_data_offset = 0;
+ uint32_t core_instr_offset = 0;
+ uint32_t cluster_instr_offset = 0;
+ uint32_t cluster_data_offset = 0;
+ uint32_t node_data_offset = 0;
+ uint32_t node_instr_offset = 0;
+ int top_node = 3;
+ int top_cluster = 3;
+ int top_core = 3;
+ int bottom_node = 3;
+ int bottom_cluster = 3;
+ int bottom_core = 3;
+ int64_t socket_id = -1;
+ int64_t cluster_id = -1;
+ int64_t core_id = -1;
+ uint32_t socket_offset = 0;
+ uint32_t cluster_offset = 0;
+ uint32_t core_offset = 0;
uint32_t pptt_start = table_data->len;
uint32_t root_offset;
int n;
+ uint32_t priv_rsrc[2];
+ uint32_t num_priv = 0;
+ bool cache_available;
+ bool llevel;
+
AcpiTable table = { .sig = "PPTT", .rev = 2,
.oem_id = oem_id, .oem_table_id = oem_table_id };
@@ -2192,11 +2310,30 @@ void build_pptt(GArray *table_data, BIOSLinker *linker, MachineState *ms,
socket_id = cpus->cpus[n].props.socket_id;
cluster_id = -1;
core_id = -1;
+ bottom_node = top_node;
+ num_priv = 0;
+ cache_available = machine_defines_cache_at_topo_level(
+ ms, CPU_TOPOLOGY_LEVEL_SOCKET);
+ llevel = machine_find_lowest_level_cache_at_topo_level(
+ ms, &bottom_node, CPU_TOPOLOGY_LEVEL_SOCKET);
See my comment in patch 3/8 about machine_defines_cache_at_topo_level()
and machine_find_lowest_level_cache_at_topo_level() similar semantics.
+ if (cache_available && llevel) {
+ build_caches(table_data, pptt_start, num_caches, caches,
+ top_node, bottom_node, &node_data_offset,
+ &node_instr_offset);
+ priv_rsrc[0] = node_instr_offset;
+ priv_rsrc[1] = node_data_offset;
+ if (node_instr_offset || node_data_offset) {
+ num_priv = node_instr_offset == node_data_offset ? 1 : 2;
+ }
+
+ top_cluster = bottom_node - 1;
+ }
+
socket_offset = table_data->len - pptt_start;
build_processor_hierarchy_node(table_data,
(1 << 0) | /* Physical package */
(1 << 4), /* Identical Implementation */
- root_offset, socket_id, NULL, 0);
+ root_offset, socket_id, priv_rsrc, num_priv);
}
if (mc->smp_props.clusters_supported && mc->smp_props.has_clusters) {
@@ -2204,21 +2341,68 @@ void build_pptt(GArray *table_data, BIOSLinker *linker,
MachineState *ms,
assert(cpus->cpus[n].props.cluster_id > cluster_id);
cluster_id = cpus->cpus[n].props.cluster_id;
core_id = -1;
+ bottom_cluster = top_cluster;
+ num_priv = 0;
+ cache_available = machine_defines_cache_at_topo_level(
+ ms, CPU_TOPOLOGY_LEVEL_CLUSTER);
+ llevel = machine_find_lowest_level_cache_at_topo_level(
+ ms, &bottom_cluster, CPU_TOPOLOGY_LEVEL_CLUSTER);
+
+ if (cache_available && llevel) {
+ build_caches(table_data, pptt_start, num_caches, caches,
+ top_cluster, bottom_cluster,
+ &cluster_data_offset, &cluster_instr_offset);
+ priv_rsrc[0] = cluster_instr_offset;
+ priv_rsrc[1] = cluster_data_offset;
+ if (cluster_instr_offset || cluster_data_offset) {
+ num_priv =
+ cluster_instr_offset == cluster_data_offset ? 1 :
2;
+ }
+ top_core = bottom_cluster - 1;
+ } else if (top_cluster == bottom_node - 1) {
+ /* socket cache but no cluster cache */
+ top_core = bottom_node - 1;
+ }
+
cluster_offset = table_data->len - pptt_start;
build_processor_hierarchy_node(table_data,
(0 << 0) | /* Not a physical package */
(1 << 4), /* Identical Implementation */
- socket_offset, cluster_id, NULL, 0);
+ socket_offset, cluster_id, priv_rsrc, num_priv);
}
} else {
+ if (machine_defines_cache_at_topo_level(
+ ms, CPU_TOPOLOGY_LEVEL_CLUSTER)) {
+ error_setg(&error_fatal, "Not clusters found for the cache");
+ return;
+ }
+
cluster_offset = socket_offset;
+ top_core = bottom_node - 1; /* there is no cluster */
+ }
+
+ if (cpus->cpus[n].props.core_id != core_id) {
+ bottom_core = top_core;
+ num_priv = 0;
+ cache_available = machine_defines_cache_at_topo_level(
+ ms, CPU_TOPOLOGY_LEVEL_CORE);
+ llevel = machine_find_lowest_level_cache_at_topo_level(
+ ms, &bottom_core, CPU_TOPOLOGY_LEVEL_CORE);
+ if (cache_available && llevel) {
+ build_caches(table_data, pptt_start, num_caches, caches,
+ top_core, bottom_core, &core_data_offset,
+ &core_instr_offset);
+ priv_rsrc[0] = core_instr_offset;
+ priv_rsrc[1] = core_data_offset;
+ num_priv = core_instr_offset == core_data_offset ? 1 : 2;
+ }
}
if (ms->smp.threads == 1) {
build_processor_hierarchy_node(table_data,
(1 << 1) | /* ACPI Processor ID valid */
- (1 << 3), /* Node is a Leaf */
- cluster_offset, n, NULL, 0);
+ (1 << 3), /* Node is a Leaf */
+ cluster_offset, n, priv_rsrc, num_priv);
} else {
if (cpus->cpus[n].props.core_id != core_id) {
assert(cpus->cpus[n].props.core_id > core_id);
@@ -2227,7 +2411,7 @@ void build_pptt(GArray *table_data, BIOSLinker *linker,
MachineState *ms,
build_processor_hierarchy_node(table_data,
(0 << 0) | /* Not a physical package */
(1 << 4), /* Identical Implementation */
- cluster_offset, core_id, NULL, 0);
+ cluster_offset, core_id, priv_rsrc, num_priv);
}
build_processor_hierarchy_node(table_data,
diff --git a/hw/arm/virt-acpi-build.c b/hw/arm/virt-acpi-build.c
index a655d1e42b..a6ad8cba1b 100644
--- a/hw/arm/virt-acpi-build.c
+++ b/hw/arm/virt-acpi-build.c
@@ -1255,6 +1255,10 @@ void virt_acpi_build(VirtMachineState *vms,
AcpiBuildTables *tables)
unsigned dsdt, xsdt;
GArray *tables_blob = tables->table_data;
MachineState *ms = MACHINE(vms);
+ CPUCoreCaches caches[CPU_MAX_CACHES];
+ unsigned int num_caches;
+
+ num_caches = virt_get_caches(vms, caches);
table_offsets = g_array_new(false, true /* clear */,
sizeof(uint32_t));
@@ -1276,8 +1280,8 @@ void virt_acpi_build(VirtMachineState *vms,
AcpiBuildTables *tables)
if (!vmc->no_cpu_topology) {
acpi_add_table(table_offsets, tables_blob);
- build_pptt(tables_blob, tables->linker, ms,
- vms->oem_id, vms->oem_table_id, 0, NULL);
+ build_pptt(tables_blob, tables->linker, ms, vms->oem_id,
+ vms->oem_table_id, num_caches, caches);
}
acpi_add_table(table_offsets, tables_blob);
diff --git a/include/hw/acpi/cpu.h b/include/hw/acpi/cpu.h
index 2809dd8a91..04c821d2b9 100644
--- a/include/hw/acpi/cpu.h
+++ b/include/hw/acpi/cpu.h
@@ -69,6 +69,16 @@ void build_cpus_aml(Aml *table, MachineState *machine,
CPUHotplugFeatures opts,
void acpi_cpu_ospm_status(CPUHotplugState *cpu_st, ACPIOSTInfoList ***list);
+struct CPUPPTTCaches {
+ enum CacheType type;
+ uint32_t sets;
+ uint32_t size;
+ uint32_t level;
+ uint16_t linesize;
+ uint8_t attributes; /* write policy: 0x0 write back, 0x1 write through */
+ uint8_t associativity;
+};
+
extern const VMStateDescription vmstate_cpu_hotplug;
#define VMSTATE_CPU_HOTPLUG(cpuhp, state) \
VMSTATE_STRUCT(cpuhp, state, 1, \
Otherwise, the ACPI table generation looks good to me.
Cheers,
Gustavo
