Hi David,
Thanks for the review.
> From: David Hildenbrand <[email protected]>
> Sent: Monday, October 9, 2023 1:21 PM
> To: Salil Mehta <[email protected]>; [email protected];
> [email protected]
> Cc: [email protected]; [email protected]; Jonathan Cameron
> <[email protected]>; [email protected];
> [email protected]; [email protected];
> [email protected]; [email protected]; [email protected];
> [email protected]; [email protected]; [email protected];
> [email protected]; [email protected]; [email protected]; [email protected];
> [email protected]; [email protected];
> [email protected]; [email protected];
> [email protected]; [email protected];
> [email protected]; [email protected]; zhukeqian
> <[email protected]>; wangxiongfeng (C) <[email protected]>;
> wangyanan (Y) <[email protected]>; [email protected];
> [email protected]; [email protected]; Linuxarm <[email protected]>
> Subject: Re: [PATCH V3 01/10] accel/kvm: Extract common KVM vCPU
> {creation,parking} code
>
> On 09.10.23 13:28, Salil Mehta wrote:
> > KVM vCPU creation is done once during the initialization of the VM when Qemu
> > thread is spawned. This is common to all the architectures.
> >
> > Hot-unplug of vCPU results in destruction of the vCPU object in QOM but the
> > corresponding KVM vCPU object in the Host KVM is not destroyed and its
> > representative KVM vCPU object/context in Qemu is parked.
> >
> > Refactor common logic so that some APIs could be reused by vCPU Hotplug
> > code.
> >
> > Signed-off-by: Salil Mehta <[email protected]>
>
> [...]
>
> >
> > int kvm_init_vcpu(CPUState *cpu, Error **errp)
> > @@ -395,19 +434,14 @@ int kvm_init_vcpu(CPUState *cpu, Error **errp)
> >
> > trace_kvm_init_vcpu(cpu->cpu_index, kvm_arch_vcpu_id(cpu));
> >
> > - ret = kvm_get_vcpu(s, kvm_arch_vcpu_id(cpu));
> > + ret = kvm_create_vcpu(cpu);
> > if (ret < 0) {
> > - error_setg_errno(errp, -ret, "kvm_init_vcpu: kvm_get_vcpu failed
> > (%lu)",
> > + error_setg_errno(errp, -ret,
> > + "kvm_init_vcpu: kvm_create_vcpu failed (%lu)",
>
> Unrelated change.
It is related. I think you missed kvm_get_vcpu -> kvm_create_vcpu change
in the string.
> > kvm_arch_vcpu_id(cpu));
> > goto err;
> > }
> >
> > - cpu->kvm_fd = ret;
> > - cpu->kvm_state = s;
> > - cpu->vcpu_dirty = true;
> > - cpu->dirty_pages = 0;
> > - cpu->throttle_us_per_full = 0;
> > -
> > mmap_size = kvm_ioctl(s, KVM_GET_VCPU_MMAP_SIZE, 0);
> > if (mmap_size < 0) {
> > ret = mmap_size;
> > diff --git a/accel/kvm/trace-events b/accel/kvm/trace-events
> > index 399aaeb0ec..08e2dc253f 100644
> > --- a/accel/kvm/trace-events
> > +++ b/accel/kvm/trace-events
> > @@ -9,6 +9,10 @@ kvm_device_ioctl(int fd, int type, void *arg) "dev fd %d,
> > type 0x%x, arg %p"
> > kvm_failed_reg_get(uint64_t id, const char *msg) "Warning: Unable to
> > retrieve ONEREG %" PRIu64 " from KVM: %s"
> > kvm_failed_reg_set(uint64_t id, const char *msg) "Warning: Unable to set
> > ONEREG %" PRIu64 " to KVM: %s"
> > kvm_init_vcpu(int cpu_index, unsigned long arch_cpu_id) "index: %d id:
> > %lu"
> > +kvm_create_vcpu(int cpu_index, unsigned long arch_cpu_id) "creating KVM
> > cpu: cpu_index: %d arch vcpu-id: %lu"
> > +kvm_get_vcpu(unsigned long arch_cpu_id) "unparking KVM vcpu: arch vcpu-id:
> > %lu"
> > +kvm_destroy_vcpu(int cpu_index, unsigned long arch_cpu_id) "destroy vcpu:
> > cpu_index: %d arch vcpu-id: %lu"
> > +kvm_park_vcpu(int cpu_index, unsigned long arch_cpu_id) "parking KVM vcpu:
> > cpu_index: %d arch vcpu-id: %lu"
>
> It's a bit confusing that there is now
>
> 1) create (create new or return parked)
> 2) destroy (cleanup + park)
> 3) park (park only)
>
> Why would one use 2) instead of 3) or the other way around? But I
> suspect that kvm_destroy_vcpu() is only supposed to be a KVM-internal
> helper ...
kvm_destroy_vcpu is more than just parking:
1. Arch destroy vcpu
2. Unmap cpu->kvm_run
3. Parking logic
To support virtual CPU Hotplug on ARM platforms we pre-create all
the KVM vCPUs but their corresponding Qemu threads are not spawned
(and hence cpu->kvm_run is not mapped). Unplugged vCPUs remains
parked in the list. Hence, only step-3 is required.
https://lore.kernel.org/qemu-devel/[email protected]/
When a virtual CPU is plugged. QOM CPU object is realized and
corresponding thread is spawned. kvm_init_vcpu ends up in unaprking
the KVM vCPU, mapping of cpu->kvm_run and kvm_arch_init_vcpu.
When a virtul CPU is un-plugged, reverse of step-1, 2 and 3 is
required during un-realization of QOM CPU object. We do not destroy
vCPU inside the KVM.
> > kvm_irqchip_commit_routes(void) ""
> > kvm_irqchip_add_msi_route(char *name, int vector, int virq) "dev %s
> > vector %d virq %d"
> > kvm_irqchip_update_msi_route(int virq) "Updating MSI route virq=%d"
> > diff --git a/include/sysemu/kvm.h b/include/sysemu/kvm.h
> > index ee9025f8e9..57bd8f8fd6 100644
> > --- a/include/sysemu/kvm.h
> > +++ b/include/sysemu/kvm.h
> > @@ -464,6 +464,20 @@ void kvm_set_sigmask_len(KVMState *s, unsigned int
> > sigmask_len);
> >
> > int kvm_physical_memory_addr_from_host(KVMState *s, void *ram_addr,
> > hwaddr *phys_addr);
> > +/**
> > + * kvm_create_vcpu - Gets a parked KVM vCPU or creates a KVM vCPU
> > + * @cpu: QOM CPUState object for which KVM vCPU has to be fetched/created.
> > + *
> > + * @returns: 0 when success, errno (<0) when failed.
> > + */
> > +int kvm_create_vcpu(CPUState *cpu);
> > +/**
> > + * kvm_park_vcpu - Gets a parked KVM vCPU if it exists
>
>
> ^ I suspect that description is wrong.
Good catch. I think manual merge error while copying the change.
Will fix.
Thanks
Salil.
