Re: [PATCH] hw/i386/acpi-build: Get NUMA information from struct NumaState

[Liu, Jingqi]

On 8/5/2021 5:26 PM, Igor Mammedov wrote:
On Tue,  3 Aug 2021 14:30:05 +0800
Jingqi Liu <jingqi....@intel.com> wrote:

The NUMA information in PCMachineState is copied from MachineState.
We get this information uniformly from struct NumaState in MachineState.
Is there a another reason behind this patch?

As cleanup it's not complete, why do you keep
PCMachineState::numa_nodes & co around?
I'd suggest to remove it completely and use data from
MachineState everywhere.

Thanks for your confirmation.
I think so too.
We can get NUMA information completely from MachineState::numa_state.

Actually, MachineState::numa_state is used everywhere except for 
pc_guest_info_init(), which just copies NUMA information to 
PCMachineState::numa_nodes and PCMachineState::node_mem.

I'll remove them completely.

Thanks,
Jingqi

Signed-off-by: Jingqi Liu <jingqi....@intel.com>
---
  hw/i386/acpi-build.c | 12 +++++++-----
  1 file changed, 7 insertions(+), 5 deletions(-)

diff --git a/hw/i386/acpi-build.c b/hw/i386/acpi-build.c
index 17836149fe..e3c9ad011e 100644
--- a/hw/i386/acpi-build.c
+++ b/hw/i386/acpi-build.c
@@ -1902,6 +1902,8 @@ build_srat(GArray *table_data, BIOSLinker *linker, 
MachineState *machine)
      X86MachineState *x86ms = X86_MACHINE(machine);
      const CPUArchIdList *apic_ids = mc->possible_cpu_arch_ids(machine);
      PCMachineState *pcms = PC_MACHINE(machine);
+    int nb_numa_nodes = machine->numa_state->num_nodes;
+    NodeInfo *numa_info = machine->numa_state->nodes;
      ram_addr_t hotplugabble_address_space_size =
          object_property_get_int(OBJECT(pcms), PC_MACHINE_DEVMEM_REGION_SIZE,
                                  NULL);
@@ -1945,9 +1947,9 @@ build_srat(GArray *table_data, BIOSLinker *linker, 
MachineState *machine)
      next_base = 0;
      numa_start = table_data->len;
  
-    for (i = 1; i < pcms->numa_nodes + 1; ++i) {
+    for (i = 1; i < nb_numa_nodes + 1; ++i) {
          mem_base = next_base;
-        mem_len = pcms->node_mem[i - 1];
+        mem_len = numa_info[i - 1].node_mem;
          next_base = mem_base + mem_len;
  
          /* Cut out the 640K hole */
@@ -1995,7 +1997,7 @@ build_srat(GArray *table_data, BIOSLinker *linker, 
MachineState *machine)
      }
  
      slots = (table_data->len - numa_start) / sizeof *numamem;
-    for (; slots < pcms->numa_nodes + 2; slots++) {
+    for (; slots < nb_numa_nodes + 2; slots++) {
          numamem = acpi_data_push(table_data, sizeof *numamem);
          build_srat_memory(numamem, 0, 0, 0, MEM_AFFINITY_NOFLAGS);
      }
@@ -2011,7 +2013,7 @@ build_srat(GArray *table_data, BIOSLinker *linker, 
MachineState *machine)
      if (hotplugabble_address_space_size) {
          numamem = acpi_data_push(table_data, sizeof *numamem);
          build_srat_memory(numamem, machine->device_memory->base,
-                          hotplugabble_address_space_size, pcms->numa_nodes - 
1,
+                          hotplugabble_address_space_size, nb_numa_nodes - 1,
                            MEM_AFFINITY_HOTPLUGGABLE | MEM_AFFINITY_ENABLED);
      }
  
@@ -2513,7 +2515,7 @@ void acpi_build(AcpiBuildTables *tables, MachineState *machine)
          }
      }
  #endif
-    if (pcms->numa_nodes) {
+    if (machine->numa_state->num_nodes) {
          acpi_add_table(table_offsets, tables_blob);
          build_srat(tables_blob, tables->linker, machine);
          if (machine->numa_state->have_numa_distance) {