From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alexander Graf Subject: Re: [Android-virt] [PATCH v4] KVM: Factor out kvm_vcpu_kick to arch-generic code Date: Tue, 13 Mar 2012 21:48:03 +0100 Message-ID: <3E82EACC-5632-4910-A7E6-A294CB4B999F@suse.de> References: <20120210014013.16869.20143.stgit@ubuntu> <1605556A-F833-4B38-8BEF-BCB584F2C8D2@suse.de> Mime-Version: 1.0 (Apple Message framework v1257) Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7BIT Cc: KVM devel mailing list , Marcelo Tosatti , kvm-ppc , Paul Mackerras , android-virt , Scott Wood , tech@virtualopensystems.com To: Christoffer Dall Return-path: In-Reply-To: Sender: kvm-ppc-owner@vger.kernel.org List-Id: kvm.vger.kernel.org On 13.03.2012, at 21:47, Christoffer Dall wrote: > On Tue, Mar 13, 2012 at 4:34 PM, Alexander Graf wrote: >> >> On 08.03.2012, at 22:44, Christoffer Dall wrote: >> >>> Any news on the status of this? >>> >>> On Thu, Feb 9, 2012 at 8:45 PM, Alexander Graf wrote: >>>> >>>> On 10.02.2012, at 02:40, Christoffer Dall wrote: >>>> >>>>> The kvm_vcpu_kick function performs roughly the same funcitonality on >>>>> most all architectures, so we shouldn't have separate copies. >>>>> >>>>> PowerPC keeps a pointer to interchanging waitqueues on the vcpu_arch >>>>> structure and to accomodate this special need a >>>>> __KVM_HAVE_ARCH_VCPU_GET_WQ define and accompanying function >>>>> kvm_arch_vcpu_wq have been defined. For all other architectures this >>>>> is a generic inline that just returns &vcpu->wq; >>>>> >>>>> This patch applies to d1c28f7568a74faacec896ee4f84afbffd20e5ab on >>>>> git://git.kernel.org/pub/scm/virt/kvm/kvm.git. >>>>> >>>>> Changes since v3: >>>>> - Doesn't try to generalize vcpu->mode across all architectures and >>>>> instead calls kvm_arch_vcpu_should_kick, which is properly defined >>>>> on x86 and ia64, but other architectures simply return 1 as to maintain >>>>> status quo. >>>>> >>>>> Changes since v2: >>>>> - Restore arch-specific vcpu->cpu assignment to arch-specific code >>>>> >>>>> Changes since v1: >>>>> - Abstact CPU mode check into arch-specific function >>>>> - Remove redundant vcpu->cpu assignment >>>>> >>>>> Signed-off-by: Christoffer Dall >>>>> --- >>>>> arch/ia64/include/asm/kvm_host.h | 1 + >>>>> arch/ia64/kvm/kvm-ia64.c | 20 +++++--------------- >>>>> arch/powerpc/include/asm/kvm_host.h | 6 ++++++ >>>>> arch/powerpc/kvm/powerpc.c | 21 ++++++--------------- >>>>> arch/s390/kvm/kvm-s390.c | 8 ++++++++ >>>>> arch/x86/kvm/x86.c | 16 ++-------------- >>>>> include/linux/kvm_host.h | 9 +++++++++ >>>>> virt/kvm/kvm_main.c | 22 ++++++++++++++++++++++ >>>>> 8 files changed, 59 insertions(+), 44 deletions(-) >>>>> >>>>> diff --git a/arch/ia64/include/asm/kvm_host.h b/arch/ia64/include/asm/kvm_host.h >>>>> index 2689ee5..06a5e91 100644 >>>>> --- a/arch/ia64/include/asm/kvm_host.h >>>>> +++ b/arch/ia64/include/asm/kvm_host.h >>>>> @@ -365,6 +365,7 @@ struct thash_cb { >>>>> }; >>>>> >>>>> struct kvm_vcpu_stat { >>>>> + u32 halt_wakeup; >>>>> }; >>>>> >>>>> struct kvm_vcpu_arch { >>>>> diff --git a/arch/ia64/kvm/kvm-ia64.c b/arch/ia64/kvm/kvm-ia64.c >>>>> index 8ca7261..96dceaa 100644 >>>>> --- a/arch/ia64/kvm/kvm-ia64.c >>>>> +++ b/arch/ia64/kvm/kvm-ia64.c >>>>> @@ -1857,21 +1857,6 @@ void kvm_arch_hardware_unsetup(void) >>>>> { >>>>> } >>>>> >>>>> -void kvm_vcpu_kick(struct kvm_vcpu *vcpu) >>>>> -{ >>>>> - int me; >>>>> - int cpu = vcpu->cpu; >>>>> - >>>>> - if (waitqueue_active(&vcpu->wq)) >>>>> - wake_up_interruptible(&vcpu->wq); >>>>> - >>>>> - me = get_cpu(); >>>>> - if (cpu != me && (unsigned) cpu < nr_cpu_ids && cpu_online(cpu)) >>>>> - if (!test_and_set_bit(KVM_REQ_KICK, &vcpu->requests)) >>>>> - smp_send_reschedule(cpu); >>>>> - put_cpu(); >>>>> -} >>>>> - >>>>> int kvm_apic_set_irq(struct kvm_vcpu *vcpu, struct kvm_lapic_irq *irq) >>>>> { >>>>> return __apic_accept_irq(vcpu, irq->vector); >>>>> @@ -1941,6 +1926,11 @@ int kvm_arch_vcpu_runnable(struct kvm_vcpu *vcpu) >>>>> (kvm_highest_pending_irq(vcpu) != -1); >>>>> } >>>>> >>>>> +int kvm_arch_vcpu_should_kick(struct kvm_vcpu *vcpu) >>>>> +{ >>>>> + return (!test_and_set_bit(KVM_REQ_KICK, &vcpu->requests)); >>>>> +} >>>>> + >>>>> int kvm_arch_vcpu_ioctl_get_mpstate(struct kvm_vcpu *vcpu, >>>>> struct kvm_mp_state *mp_state) >>>>> { >>>>> diff --git a/arch/powerpc/include/asm/kvm_host.h b/arch/powerpc/include/asm/kvm_host.h >>>>> index 1843d5d..d089969 100644 >>>>> --- a/arch/powerpc/include/asm/kvm_host.h >>>>> +++ b/arch/powerpc/include/asm/kvm_host.h >>>>> @@ -495,4 +495,10 @@ struct kvm_vcpu_arch { >>>>> #define KVM_MMIO_REG_QPR 0x0040 >>>>> #define KVM_MMIO_REG_FQPR 0x0060 >>>>> >>>>> +#define __KVM_HAVE_ARCH_VCPU_GET_WQ 1 >>>>> +static inline wait_queue_head *kvm_arch_vcpu_wq(struct kvm_vcpu *vcpu) >>>>> +{ >>>>> + return vcpu->arch.wqp; >> >> Sorry for the late notification - I didn't compile test it. >> >> That code breaks for 2 reasons: >> >> 1) It's wait_queue_head_t, not wait_queue_head >> 2) struct kvm_vcpu isn't defined here yet, only arch is. Maybe we could pass in &vcpu->arch instead of vcpu? Or move the function somewhere else? >> >> > > Alex, would you mind fixing it up? I don't have powerpc cross > compilation set up and would like to avoid getting into that just > these days? Yeah, trying to get my head around all the different layers of #include's ;). It's not exactly easy to find the right spot to put this in. Alex From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alexander Graf Date: Tue, 13 Mar 2012 20:48:03 +0000 Subject: Re: [Android-virt] [PATCH v4] KVM: Factor out kvm_vcpu_kick to arch-generic code Message-Id: <3E82EACC-5632-4910-A7E6-A294CB4B999F@suse.de> List-Id: References: <20120210014013.16869.20143.stgit@ubuntu> <1605556A-F833-4B38-8BEF-BCB584F2C8D2@suse.de> In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Christoffer Dall Cc: KVM devel mailing list , Marcelo Tosatti , kvm-ppc , Paul Mackerras , android-virt , Scott Wood , tech@virtualopensystems.com On 13.03.2012, at 21:47, Christoffer Dall wrote: > On Tue, Mar 13, 2012 at 4:34 PM, Alexander Graf wrote: >> >> On 08.03.2012, at 22:44, Christoffer Dall wrote: >> >>> Any news on the status of this? >>> >>> On Thu, Feb 9, 2012 at 8:45 PM, Alexander Graf wrote: >>>> >>>> On 10.02.2012, at 02:40, Christoffer Dall wrote: >>>> >>>>> The kvm_vcpu_kick function performs roughly the same funcitonality on >>>>> most all architectures, so we shouldn't have separate copies. >>>>> >>>>> PowerPC keeps a pointer to interchanging waitqueues on the vcpu_arch >>>>> structure and to accomodate this special need a >>>>> __KVM_HAVE_ARCH_VCPU_GET_WQ define and accompanying function >>>>> kvm_arch_vcpu_wq have been defined. For all other architectures this >>>>> is a generic inline that just returns &vcpu->wq; >>>>> >>>>> This patch applies to d1c28f7568a74faacec896ee4f84afbffd20e5ab on >>>>> git://git.kernel.org/pub/scm/virt/kvm/kvm.git. >>>>> >>>>> Changes since v3: >>>>> - Doesn't try to generalize vcpu->mode across all architectures and >>>>> instead calls kvm_arch_vcpu_should_kick, which is properly defined >>>>> on x86 and ia64, but other architectures simply return 1 as to maintain >>>>> status quo. >>>>> >>>>> Changes since v2: >>>>> - Restore arch-specific vcpu->cpu assignment to arch-specific code >>>>> >>>>> Changes since v1: >>>>> - Abstact CPU mode check into arch-specific function >>>>> - Remove redundant vcpu->cpu assignment >>>>> >>>>> Signed-off-by: Christoffer Dall >>>>> --- >>>>> arch/ia64/include/asm/kvm_host.h | 1 + >>>>> arch/ia64/kvm/kvm-ia64.c | 20 +++++--------------- >>>>> arch/powerpc/include/asm/kvm_host.h | 6 ++++++ >>>>> arch/powerpc/kvm/powerpc.c | 21 ++++++--------------- >>>>> arch/s390/kvm/kvm-s390.c | 8 ++++++++ >>>>> arch/x86/kvm/x86.c | 16 ++-------------- >>>>> include/linux/kvm_host.h | 9 +++++++++ >>>>> virt/kvm/kvm_main.c | 22 ++++++++++++++++++++++ >>>>> 8 files changed, 59 insertions(+), 44 deletions(-) >>>>> >>>>> diff --git a/arch/ia64/include/asm/kvm_host.h b/arch/ia64/include/asm/kvm_host.h >>>>> index 2689ee5..06a5e91 100644 >>>>> --- a/arch/ia64/include/asm/kvm_host.h >>>>> +++ b/arch/ia64/include/asm/kvm_host.h >>>>> @@ -365,6 +365,7 @@ struct thash_cb { >>>>> }; >>>>> >>>>> struct kvm_vcpu_stat { >>>>> + u32 halt_wakeup; >>>>> }; >>>>> >>>>> struct kvm_vcpu_arch { >>>>> diff --git a/arch/ia64/kvm/kvm-ia64.c b/arch/ia64/kvm/kvm-ia64.c >>>>> index 8ca7261..96dceaa 100644 >>>>> --- a/arch/ia64/kvm/kvm-ia64.c >>>>> +++ b/arch/ia64/kvm/kvm-ia64.c >>>>> @@ -1857,21 +1857,6 @@ void kvm_arch_hardware_unsetup(void) >>>>> { >>>>> } >>>>> >>>>> -void kvm_vcpu_kick(struct kvm_vcpu *vcpu) >>>>> -{ >>>>> - int me; >>>>> - int cpu = vcpu->cpu; >>>>> - >>>>> - if (waitqueue_active(&vcpu->wq)) >>>>> - wake_up_interruptible(&vcpu->wq); >>>>> - >>>>> - me = get_cpu(); >>>>> - if (cpu != me && (unsigned) cpu < nr_cpu_ids && cpu_online(cpu)) >>>>> - if (!test_and_set_bit(KVM_REQ_KICK, &vcpu->requests)) >>>>> - smp_send_reschedule(cpu); >>>>> - put_cpu(); >>>>> -} >>>>> - >>>>> int kvm_apic_set_irq(struct kvm_vcpu *vcpu, struct kvm_lapic_irq *irq) >>>>> { >>>>> return __apic_accept_irq(vcpu, irq->vector); >>>>> @@ -1941,6 +1926,11 @@ int kvm_arch_vcpu_runnable(struct kvm_vcpu *vcpu) >>>>> (kvm_highest_pending_irq(vcpu) != -1); >>>>> } >>>>> >>>>> +int kvm_arch_vcpu_should_kick(struct kvm_vcpu *vcpu) >>>>> +{ >>>>> + return (!test_and_set_bit(KVM_REQ_KICK, &vcpu->requests)); >>>>> +} >>>>> + >>>>> int kvm_arch_vcpu_ioctl_get_mpstate(struct kvm_vcpu *vcpu, >>>>> struct kvm_mp_state *mp_state) >>>>> { >>>>> diff --git a/arch/powerpc/include/asm/kvm_host.h b/arch/powerpc/include/asm/kvm_host.h >>>>> index 1843d5d..d089969 100644 >>>>> --- a/arch/powerpc/include/asm/kvm_host.h >>>>> +++ b/arch/powerpc/include/asm/kvm_host.h >>>>> @@ -495,4 +495,10 @@ struct kvm_vcpu_arch { >>>>> #define KVM_MMIO_REG_QPR 0x0040 >>>>> #define KVM_MMIO_REG_FQPR 0x0060 >>>>> >>>>> +#define __KVM_HAVE_ARCH_VCPU_GET_WQ 1 >>>>> +static inline wait_queue_head *kvm_arch_vcpu_wq(struct kvm_vcpu *vcpu) >>>>> +{ >>>>> + return vcpu->arch.wqp; >> >> Sorry for the late notification - I didn't compile test it. >> >> That code breaks for 2 reasons: >> >> 1) It's wait_queue_head_t, not wait_queue_head >> 2) struct kvm_vcpu isn't defined here yet, only arch is. Maybe we could pass in &vcpu->arch instead of vcpu? Or move the function somewhere else? >> >> > > Alex, would you mind fixing it up? I don't have powerpc cross > compilation set up and would like to avoid getting into that just > these days? Yeah, trying to get my head around all the different layers of #include's ;). It's not exactly easy to find the right spot to put this in. Alex