Patchwork kvm: Check if smp_cpus exceeds max cpus supported by kvm

login
register
mail settings
Submitter dunrong huang
Date July 30, 2012, 6:22 p.m.
Message ID <1343672558-8560-1-git-send-email-riegamaths@gmail.com>
Download mbox | patch
Permalink /patch/174073/
State New
Headers show

Comments

dunrong huang - July 30, 2012, 6:22 p.m.
From: Dunrong Huang <riegamaths@gmail.com>

Add a helper function for fetching max cpus supported by kvm.

Make QEMU exit with an error message if smp_cpus exceeds limit
of VCPU count retrieved by invoking this helper function.

Signed-off-by: Dunrong Huang <riegamaths@gmail.com>
---
 kvm-all.c |   25 +++++++++++++++++++++++++
 1 files changed, 25 insertions(+), 0 deletions(-)
Stefan Hajnoczi - July 31, 2012, 9:51 a.m.
On Mon, Jul 30, 2012 at 7:22 PM,  <riegamaths@gmail.com> wrote:
> From: Dunrong Huang <riegamaths@gmail.com>
>
> Add a helper function for fetching max cpus supported by kvm.
>
> Make QEMU exit with an error message if smp_cpus exceeds limit
> of VCPU count retrieved by invoking this helper function.
>
> Signed-off-by: Dunrong Huang <riegamaths@gmail.com>
> ---
>  kvm-all.c |   25 +++++++++++++++++++++++++
>  1 files changed, 25 insertions(+), 0 deletions(-)
>
> diff --git a/kvm-all.c b/kvm-all.c
> index 2148b20..8cb4b92 100644
> --- a/kvm-all.c
> +++ b/kvm-all.c
> @@ -1207,6 +1207,23 @@ static int kvm_irqchip_create(KVMState *s)
>      return 0;
>  }
>
> +static int kvm_max_vcpus(KVMState *s)
> +{
> +    int max_vcpus = 4;
> +    int ret;
> +    ret = kvm_check_extension(s, KVM_CAP_MAX_VCPUS);
> +    if (ret) {
> +        max_vcpus = ret;
> +    } else {
> +       ret = kvm_check_extension(s, KVM_CAP_NR_VCPUS);
> +        if (ret) {
> +            max_vcpus = ret;
> +        }
> +   }

The indentation is off here.  It should be 4 spaces.

Otherwise looks fine.

Stefan
Peter Maydell - July 31, 2012, 10:14 a.m.
On 30 July 2012 19:22,  <riegamaths@gmail.com> wrote:
> +static int kvm_max_vcpus(KVMState *s)
> +{
> +    int max_vcpus = 4;
> +    int ret;
> +    ret = kvm_check_extension(s, KVM_CAP_MAX_VCPUS);
> +    if (ret) {
> +        max_vcpus = ret;
> +    } else {
> +       ret = kvm_check_extension(s, KVM_CAP_NR_VCPUS);
> +        if (ret) {
> +            max_vcpus = ret;
> +        }
> +   }
> +
> +    return max_vcpus;
> +}

A small thing, but I think having code flow like:

    /* Find number of supported CPUs using the recommended
     * procedure from the kernel API documentation to cope with
     * older kernels that may be missing capabilities.
     */
    ret = kvm_check_extension(s, KVM_CAP_MAX_VCPUS);
    if (ret) {
        return ret;
    }
    ret = kvm_check_extension(s, KVM_CAP_NR_VCPUS);
    if (ret) {
        return ret;
    }
    return 4;

would be clearer.

(also I think a comment helps suggest that 4 isn't a magic
number we made up ourselves :-))

-- PMM

Patch

diff --git a/kvm-all.c b/kvm-all.c
index 2148b20..8cb4b92 100644
--- a/kvm-all.c
+++ b/kvm-all.c
@@ -1207,6 +1207,23 @@  static int kvm_irqchip_create(KVMState *s)
     return 0;
 }
 
+static int kvm_max_vcpus(KVMState *s)
+{
+    int max_vcpus = 4;
+    int ret;
+    ret = kvm_check_extension(s, KVM_CAP_MAX_VCPUS);
+    if (ret) {
+        max_vcpus = ret;
+    } else {
+       ret = kvm_check_extension(s, KVM_CAP_NR_VCPUS);
+        if (ret) {
+            max_vcpus = ret;
+        }
+   }
+
+    return max_vcpus;
+}
+
 int kvm_init(void)
 {
     static const char upgrade_note[] =
@@ -1216,6 +1233,7 @@  int kvm_init(void)
     const KVMCapabilityInfo *missing_cap;
     int ret;
     int i;
+    int max_vcpus;
 
     s = g_malloc0(sizeof(KVMState));
 
@@ -1256,6 +1274,13 @@  int kvm_init(void)
         goto err;
     }
 
+    max_vcpus = kvm_max_vcpus(s);
+    if (smp_cpus > max_vcpus) {
+        fprintf(stderr, "Number of SMP cpus requested (%d) exceeds max cpus "
+                "supported by KVM (%d)\n", smp_cpus, max_vcpus);
+        exit(1);
+    }
+
     s->vmfd = kvm_ioctl(s, KVM_CREATE_VM, 0);
     if (s->vmfd < 0) {
 #ifdef TARGET_S390X