Re: [PATCH 4/4] KVM: x86: enable configurable MAX_VCPU

From: "Radim Krčmář" <rkrcmar@redhat.com>
To: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
	Christoffer Dall <cdall@linaro.org>,
	Marc Zyngier <marc.zyngier@arm.com>,
	Paolo Bonzini <pbonzini@redhat.com>,
	Cornelia Huck <cornelia.huck@de.ibm.com>,
	James Hogan <james.hogan@imgtec.com>,
	Paul Mackerras <paulus@ozlabs.org>,
	Alexander Graf <agraf@suse.com>
Subject: Re: [PATCH 4/4] KVM: x86: enable configurable MAX_VCPU
Date: Mon, 24 Apr 2017 19:00:10 +0200	[thread overview]
Message-ID: <20170424170009.GB5713@potion> (raw)
In-Reply-To: <cbecee82-8198-58b0-35ec-e513317195ce@de.ibm.com>

2017-04-19 10:08+0200, Christian Borntraeger:
> On 04/13/2017 10:19 PM, Radim Krčmář wrote:
>> The only user of KVM_MAX_VCPU is switched to kvm->max_vcpu.
>> 
>> The limit could have been INT_MAX as it makes no difference, but there
>> is no point in making it bigger than KVM_MAX_VCPU_ID.
>> 
>> Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>
>> ---
>>  arch/x86/include/asm/kvm_host.h | 1 +
>>  arch/x86/kvm/irq_comm.c         | 4 ++--
>>  2 files changed, 3 insertions(+), 2 deletions(-)
>> 
>> diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
>> index 2cc5ec7cc6f5..eeeb88eedabf 100644
>> --- a/arch/x86/include/asm/kvm_host.h
>> +++ b/arch/x86/include/asm/kvm_host.h
>> @@ -38,6 +38,7 @@
>>  #define KVM_MAX_VCPUS 288
>>  #define KVM_SOFT_MAX_VCPUS 240
>>  #define KVM_MAX_VCPU_ID 1023
>> +#define KVM_CONFIGURABLE_MAX_VCPUS KVM_MAX_VCPU_ID
>>  #define KVM_USER_MEM_SLOTS 509
>>  /* memory slots that are not exposed to userspace */
>>  #define KVM_PRIVATE_MEM_SLOTS 3
>> diff --git a/arch/x86/kvm/irq_comm.c b/arch/x86/kvm/irq_comm.c
>> index 4517a4c2ac3a..a7baeb44539a 100644
>> --- a/arch/x86/kvm/irq_comm.c
>> +++ b/arch/x86/kvm/irq_comm.c
>> @@ -60,7 +60,7 @@ int kvm_irq_delivery_to_apic(struct kvm *kvm, struct kvm_lapic *src,
>>  {
>>  	int i, r = -1;
>>  	struct kvm_vcpu *vcpu, *lowest = NULL;
>> -	unsigned long dest_vcpu_bitmap[BITS_TO_LONGS(KVM_MAX_VCPUS)];
>> +	unsigned long dest_vcpu_bitmap[BITS_TO_LONGS(kvm->max_vcpus)];
> 
> Doesnt that allow unlimited stack usage?

Right, it doesn't scale.

kvm->max_vcpus is at most KVM_CONFIGURABLE_MAX_VCPUS, but that will be
bigger than the stack.  I'll allocate the bitmap in v2.

Thanks.