On Fri, Nov 29, 2019 at 03:23:36PM +0100, Vitaly Kuznetsov wrote:
> Peter Xu <peterx@xxxxxxxxxx> writes:
>
> > Previously we were using either APIC_DEST_PHYSICAL|APIC_DEST_LOGICAL
> > or 0|1 to fill in kvm_lapic_irq.dest_mode, and it's done in an adhoc
> > way. It's fine imho only because in most cases when we check against
> > dest_mode it's against APIC_DEST_PHYSICAL (which equals to 0).
> > However, that's not consistent, majorly because APIC_DEST_LOGICAL does
> > not equals to 1, so if one day we check irq.dest_mode against
> > APIC_DEST_LOGICAL we'll probably always get a false returned.
> >
> > This patch replaces the 0/1 settings of irq.dest_mode with the macros
> > to make them consistent.
> >
> > CC: Paolo Bonzini <pbonzini@xxxxxxxxxx>
> > CC: Sean Christopherson <sean.j.christopherson@xxxxxxxxx>
> > CC: Vitaly Kuznetsov <vkuznets@xxxxxxxxxx>
> > CC: Nitesh Narayan Lal <nitesh@xxxxxxxxxx>
> > Signed-off-by: Peter Xu <peterx@xxxxxxxxxx>
> > ---
> > arch/x86/kvm/ioapic.c | 9 ++++++---
> > arch/x86/kvm/irq_comm.c | 7 ++++---
> > arch/x86/kvm/x86.c | 2 +-
> > 3 files changed, 11 insertions(+), 7 deletions(-)
> >
> > diff --git a/arch/x86/kvm/ioapic.c b/arch/x86/kvm/ioapic.c
> > index 9fd2dd89a1c5..1e091637d5d5 100644
> > --- a/arch/x86/kvm/ioapic.c
> > +++ b/arch/x86/kvm/ioapic.c
> > @@ -331,7 +331,8 @@ static void ioapic_write_indirect(struct kvm_ioapic *ioapic, u32 val)
> > irq.vector = e->fields.vector;
> > irq.delivery_mode = e->fields.delivery_mode << 8;
> > irq.dest_id = e->fields.dest_id;
> > - irq.dest_mode = e->fields.dest_mode;
> > + irq.dest_mode = e->fields.dest_mode ?
> > + APIC_DEST_LOGICAL : APIC_DEST_PHYSICAL;
> > bitmap_zero(&vcpu_bitmap, 16);
> > kvm_bitmap_or_dest_vcpus(ioapic->kvm, &irq,
> > &vcpu_bitmap);
> > @@ -343,7 +344,8 @@ static void ioapic_write_indirect(struct kvm_ioapic *ioapic, u32 val)
> > * keep ioapic_handled_vectors synchronized.
> > */
> > irq.dest_id = old_dest_id;
> > - irq.dest_mode = old_dest_mode;
> > + irq.dest_mode = old_dest_mode ?
> > + APIC_DEST_LOGICAL : APIC_DEST_PHYSICAL;
> > kvm_bitmap_or_dest_vcpus(ioapic->kvm, &irq,
> > &vcpu_bitmap);
> > }
> > @@ -369,7 +371,8 @@ static int ioapic_service(struct kvm_ioapic *ioapic, int irq, bool line_status)
> >
> > irqe.dest_id = entry->fields.dest_id;
> > irqe.vector = entry->fields.vector;
> > - irqe.dest_mode = entry->fields.dest_mode;
> > + irqe.dest_mode = entry->fields.dest_mode ?
> > + APIC_DEST_LOGICAL : APIC_DEST_PHYSICAL;
> > irqe.trig_mode = entry->fields.trig_mode;
> > irqe.delivery_mode = entry->fields.delivery_mode << 8;
> > irqe.level = 1;
> > diff --git a/arch/x86/kvm/irq_comm.c b/arch/x86/kvm/irq_comm.c
> > index 8ecd48d31800..673b6afd6dbf 100644
> > --- a/arch/x86/kvm/irq_comm.c
> > +++ b/arch/x86/kvm/irq_comm.c
> > @@ -52,8 +52,8 @@ int kvm_irq_delivery_to_apic(struct kvm *kvm, struct kvm_lapic *src,
> > unsigned long dest_vcpu_bitmap[BITS_TO_LONGS(KVM_MAX_VCPUS)];
> > unsigned int dest_vcpus = 0;
> >
> > - if (irq->dest_mode == 0 && irq->dest_id == 0xff &&
> > - kvm_lowest_prio_delivery(irq)) {
> > + if (irq->dest_mode == APIC_DEST_PHYSICAL &&
> > + irq->dest_id == 0xff && kvm_lowest_prio_delivery(irq)) {
> > printk(KERN_INFO "kvm: apic: phys broadcast and lowest prio\n");
> > irq->delivery_mode = APIC_DM_FIXED;
> > }
> > @@ -114,7 +114,8 @@ void kvm_set_msi_irq(struct kvm *kvm, struct kvm_kernel_irq_routing_entry *e,
> > irq->dest_id |= MSI_ADDR_EXT_DEST_ID(e->msi.address_hi);
> > irq->vector = (e->msi.data &
> > MSI_DATA_VECTOR_MASK) >> MSI_DATA_VECTOR_SHIFT;
> > - irq->dest_mode = (1 << MSI_ADDR_DEST_MODE_SHIFT) & e->msi.address_lo;
> > + irq->dest_mode = (1 << MSI_ADDR_DEST_MODE_SHIFT) & e->msi.address_lo ?
> > + APIC_DEST_LOGICAL : APIC_DEST_PHYSICAL;
> > irq->trig_mode = (1 << MSI_DATA_TRIGGER_SHIFT) & e->msi.data;
> > irq->delivery_mode = e->msi.data & 0x700;
> > irq->msi_redir_hint = ((e->msi.address_lo
> > diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> > index 3ed167e039e5..3b00d662dc14 100644
> > --- a/arch/x86/kvm/x86.c
> > +++ b/arch/x86/kvm/x86.c
> > @@ -7356,7 +7356,7 @@ static void kvm_pv_kick_cpu_op(struct kvm *kvm, unsigned long flags, int apicid)
> > struct kvm_lapic_irq lapic_irq;
> >
> > lapic_irq.shorthand = 0;
> > - lapic_irq.dest_mode = 0;
> > + lapic_irq.dest_mode = APIC_DEST_PHYSICAL;
> > lapic_irq.level = 0;
> > lapic_irq.dest_id = apicid;
> > lapic_irq.msi_redir_hint = false;
>
> dest_mode is being passed to kvm_apic_match_dest() where we do:
>
> case APIC_DEST_NOSHORT:
> if (dest_mode == APIC_DEST_PHYSICAL)
> return kvm_apic_match_physical_addr(target, mda);
> else
> return kvm_apic_match_logical_addr(target, mda);
>
> I'd suggest we fix this too then (and BUG() in case it's neither).
That's true. Let me add another patch to fix it (probably easier by
changing the type of dest_mode param in kvm_apic_match_dest from int
to bool).
Thanks,
--
Peter Xu
