diff mbox

[2.1,02/28] NUMA: check if the total numa memory size is equal to ram_size

Message ID 1393941656-29068-3-git-send-email-pbonzini@redhat.com
State New
Headers show

Commit Message

Paolo Bonzini March 4, 2014, 2 p.m. UTC
From: Wanlong Gao <gaowanlong@cn.fujitsu.com>

If the total number of the assigned numa nodes memory is not
equal to the assigned ram size, it will write the wrong data
to ACPI talb, then the guest will ignore the wrong ACPI table
and recognize all memory to one node. It's buggy, we should
check it to ensure that we write the right data to ACPI table.

Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 numa.c | 11 +++++++++++
 1 file changed, 11 insertions(+)

Comments

Eric Blake March 4, 2014, 5 p.m. UTC | #1
On 03/04/2014 07:00 AM, Paolo Bonzini wrote:
> From: Wanlong Gao <gaowanlong@cn.fujitsu.com>
> 
> If the total number of the assigned numa nodes memory is not
> equal to the assigned ram size, it will write the wrong data
> to ACPI talb, then the guest will ignore the wrong ACPI table

s/talb/table/

> and recognize all memory to one node. It's buggy, we should
> check it to ensure that we write the right data to ACPI table.
> 
> Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
> Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---
>  numa.c | 11 +++++++++++
>  1 file changed, 11 insertions(+)
> 

> +        if (numa_total != ram_size) {
> +            fprintf(stderr, "qemu: numa nodes total memory size "
> +                            "should equal to ram_size\n");

Is it worth also printing numa_total or ram_size values in this error
message?
Paolo Bonzini March 4, 2014, 5:19 p.m. UTC | #2
Il 04/03/2014 18:00, Eric Blake ha scritto:
> On 03/04/2014 07:00 AM, Paolo Bonzini wrote:
>> From: Wanlong Gao <gaowanlong@cn.fujitsu.com>
>>
>> If the total number of the assigned numa nodes memory is not
>> equal to the assigned ram size, it will write the wrong data
>> to ACPI talb, then the guest will ignore the wrong ACPI table
>
> s/talb/table/
>
>> and recognize all memory to one node. It's buggy, we should
>> check it to ensure that we write the right data to ACPI table.
>>
>> Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
>> Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
>> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
>> ---
>>  numa.c | 11 +++++++++++
>>  1 file changed, 11 insertions(+)
>>
>
>> +        if (numa_total != ram_size) {
>> +            fprintf(stderr, "qemu: numa nodes total memory size "
>> +                            "should equal to ram_size\n");
>
> Is it worth also printing numa_total or ram_size values in this error
> message?

Good idea.

Paolo
diff mbox

Patch

diff --git a/numa.c b/numa.c
index 395c14f..8ba66f1 100644
--- a/numa.c
+++ b/numa.c
@@ -127,6 +127,7 @@  void numa_add(const char *optarg)
 void set_numa_nodes(void)
 {
     if (nb_numa_nodes > 0) {
+        uint64_t numa_total;
         int i;
 
         if (nb_numa_nodes > MAX_NODES) {
@@ -154,6 +155,16 @@  void set_numa_nodes(void)
             node_mem[i] = ram_size - usedmem;
         }
 
+        numa_total = 0;
+        for (i = 0; i < nb_numa_nodes; i++) {
+            numa_total += node_mem[i];
+        }
+        if (numa_total != ram_size) {
+            fprintf(stderr, "qemu: numa nodes total memory size "
+                            "should equal to ram_size\n");
+            exit(1);
+        }
+
         for (i = 0; i < nb_numa_nodes; i++) {
             if (!bitmap_empty(node_cpumask[i], MAX_CPUMASK_BITS)) {
                 break;