| Message ID | 1378195705-5143-1-git-send-email-aik@ozlabs.ru |
|---|---|
| State | New |
| Headers | show |
On Tue, Sep 03, 2013 at 06:08:25PM +1000, Alexey Kardashevskiy wrote: > On PPC64 systems MSI Messages are translated to system IRQ in a PCI > host bridge. This is already supported for emulated MSI/MSIX but > not for irqfd where the current QEMU allocates IRQ numbers from > irqchip and maps MSIMessages to IRQ in the host kernel. > > This adds a new direct mapping flag which tells > the kvm_irqchip_add_msi_route() function that a new VIRQ > should not be allocated, instead the value from MSIMessage::data > should be used. It is up to the platform code to make sure that > this contains a valid IRQ number as sPAPR does in spapr_pci.c. > > Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Fine with me Acked-by: Michael S. Tsirkin <mst@redhat.com> > --- > > The patch does not enable this mapping for any platform in this patch > as it is going be done for spapr only on a separate patch which is not > ready to go as it depends on the in-kernel XICS-KVM patchset which is not > in upstream yet. > > --- > Changes: > v6: > * simplified to a single global flag and putting an IRQ number > in MSIMessage::data > > 2013/08/07 v5: > * pci_bus_map_msi now has default behaviour which is to call > kvm_irqchip_add_msi_route > * kvm_irqchip_release_virq fixed not crash when there is no routes > --- > include/sysemu/kvm.h | 9 +++++++++ > kvm-all.c | 13 +++++++++++++ > kvm-stub.c | 1 + > 3 files changed, 23 insertions(+) > > diff --git a/include/sysemu/kvm.h b/include/sysemu/kvm.h > index 8e76685..0e9ef38 100644 > --- a/include/sysemu/kvm.h > +++ b/include/sysemu/kvm.h > @@ -46,6 +46,7 @@ extern bool kvm_halt_in_kernel_allowed; > extern bool kvm_irqfds_allowed; > extern bool kvm_msi_via_irqfd_allowed; > extern bool kvm_gsi_routing_allowed; > +extern bool kvm_gsi_direct_mapping; > extern bool kvm_readonly_mem_allowed; > > #if defined CONFIG_KVM || !defined NEED_CPU_H > @@ -108,6 +109,13 @@ extern bool kvm_readonly_mem_allowed; > #define kvm_gsi_routing_enabled() (kvm_gsi_routing_allowed) > > /** > + * kvm_gsi_direct_mapping: > + * > + * Returns: true if GSI direct mapping is enabled. > + */ > +#define kvm_gsi_direct_mapping() (kvm_gsi_direct_mapping) > + > +/** > * kvm_readonly_mem_enabled: > * > * Returns: true if KVM readonly memory is enabled (ie the kernel > @@ -123,6 +131,7 @@ extern bool kvm_readonly_mem_allowed; > #define kvm_irqfds_enabled() (false) > #define kvm_msi_via_irqfd_enabled() (false) > #define kvm_gsi_routing_allowed() (false) > +#define kvm_gsi_direct_mapping() (false) > #define kvm_readonly_mem_enabled() (false) > #endif > > diff --git a/kvm-all.c b/kvm-all.c > index 875e32e..17fb865 100644 > --- a/kvm-all.c > +++ b/kvm-all.c > @@ -111,6 +111,7 @@ bool kvm_halt_in_kernel_allowed; > bool kvm_irqfds_allowed; > bool kvm_msi_via_irqfd_allowed; > bool kvm_gsi_routing_allowed; > +bool kvm_gsi_direct_mapping; > bool kvm_allowed; > bool kvm_readonly_mem_allowed; > > @@ -1069,6 +1070,10 @@ void kvm_irqchip_release_virq(KVMState *s, int virq) > struct kvm_irq_routing_entry *e; > int i; > > + if (kvm_gsi_direct_mapping()) { > + return; > + } > + > for (i = 0; i < s->irq_routes->nr; i++) { > e = &s->irq_routes->entries[i]; > if (e->gsi == virq) { > @@ -1190,6 +1195,10 @@ int kvm_irqchip_add_msi_route(KVMState *s, MSIMessage msg) > struct kvm_irq_routing_entry kroute = {}; > int virq; > > + if (kvm_gsi_direct_mapping()) { > + return msg.data & 0xffff; > + } > + > if (!kvm_gsi_routing_enabled()) { > return -ENOSYS; > } > @@ -1216,6 +1225,10 @@ int kvm_irqchip_update_msi_route(KVMState *s, int virq, MSIMessage msg) > { > struct kvm_irq_routing_entry kroute = {}; > > + if (kvm_gsi_direct_mapping()) { > + return 0; > + } > + > if (!kvm_irqchip_in_kernel()) { > return -ENOSYS; > } > diff --git a/kvm-stub.c b/kvm-stub.c > index 548f471..e979f76 100644 > --- a/kvm-stub.c > +++ b/kvm-stub.c > @@ -25,6 +25,7 @@ bool kvm_async_interrupts_allowed; > bool kvm_irqfds_allowed; > bool kvm_msi_via_irqfd_allowed; > bool kvm_gsi_routing_allowed; > +bool kvm_gsi_direct_mapping; > bool kvm_allowed; > bool kvm_readonly_mem_allowed; > > -- > 1.8.4.rc4
On 09/03/2013 10:17 AM, Michael S. Tsirkin wrote: > On Tue, Sep 03, 2013 at 06:08:25PM +1000, Alexey Kardashevskiy wrote: >> On PPC64 systems MSI Messages are translated to system IRQ in a PCI >> host bridge. This is already supported for emulated MSI/MSIX but >> not for irqfd where the current QEMU allocates IRQ numbers from >> irqchip and maps MSIMessages to IRQ in the host kernel. >> >> This adds a new direct mapping flag which tells >> the kvm_irqchip_add_msi_route() function that a new VIRQ >> should not be allocated, instead the value from MSIMessage::data >> should be used. It is up to the platform code to make sure that >> this contains a valid IRQ number as sPAPR does in spapr_pci.c. >> >> Signed-off-by: Alexey Kardashevskiy<aik@ozlabs.ru> > Fine with me > > Acked-by: Michael S. Tsirkin<mst@redhat.com> Thanks, applied to ppc-next. Alex
Il 03/09/2013 10:17, Michael S. Tsirkin ha scritto: >> On PPC64 systems MSI Messages are translated to system IRQ in a PCI >> > host bridge. This is already supported for emulated MSI/MSIX but >> > not for irqfd where the current QEMU allocates IRQ numbers from >> > irqchip and maps MSIMessages to IRQ in the host kernel. >> > >> > This adds a new direct mapping flag which tells >> > the kvm_irqchip_add_msi_route() function that a new VIRQ >> > should not be allocated, instead the value from MSIMessage::data >> > should be used. It is up to the platform code to make sure that >> > this contains a valid IRQ number as sPAPR does in spapr_pci.c. >> > >> > Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> > Fine with me > > Acked-by: Michael S. Tsirkin <mst@redhat.com> > Alex, are you picking this one? Paolo
On 05.09.2013, at 18:01, Paolo Bonzini wrote: > Il 03/09/2013 10:17, Michael S. Tsirkin ha scritto: >>> On PPC64 systems MSI Messages are translated to system IRQ in a PCI >>>> host bridge. This is already supported for emulated MSI/MSIX but >>>> not for irqfd where the current QEMU allocates IRQ numbers from >>>> irqchip and maps MSIMessages to IRQ in the host kernel. >>>> >>>> This adds a new direct mapping flag which tells >>>> the kvm_irqchip_add_msi_route() function that a new VIRQ >>>> should not be allocated, instead the value from MSIMessage::data >>>> should be used. It is up to the platform code to make sure that >>>> this contains a valid IRQ number as sPAPR does in spapr_pci.c. >>>> >>>> Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> >> Fine with me >> >> Acked-by: Michael S. Tsirkin <mst@redhat.com> >> > > Alex, are you picking this one? I wouldn't mind if you pull it in. It mostly touches generic kvm code after all. Alex
Il 03/09/2013 10:08, Alexey Kardashevskiy ha scritto: > On PPC64 systems MSI Messages are translated to system IRQ in a PCI > host bridge. This is already supported for emulated MSI/MSIX but > not for irqfd where the current QEMU allocates IRQ numbers from > irqchip and maps MSIMessages to IRQ in the host kernel. > > This adds a new direct mapping flag which tells > the kvm_irqchip_add_msi_route() function that a new VIRQ > should not be allocated, instead the value from MSIMessage::data > should be used. It is up to the platform code to make sure that > this contains a valid IRQ number as sPAPR does in spapr_pci.c. > > Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> > > --- > > The patch does not enable this mapping for any platform in this patch > as it is going be done for spapr only on a separate patch which is not > ready to go as it depends on the in-kernel XICS-KVM patchset which is not > in upstream yet. > > --- > Changes: > v6: > * simplified to a single global flag and putting an IRQ number > in MSIMessage::data > > 2013/08/07 v5: > * pci_bus_map_msi now has default behaviour which is to call > kvm_irqchip_add_msi_route > * kvm_irqchip_release_virq fixed not crash when there is no routes > --- > include/sysemu/kvm.h | 9 +++++++++ > kvm-all.c | 13 +++++++++++++ > kvm-stub.c | 1 + > 3 files changed, 23 insertions(+) > > diff --git a/include/sysemu/kvm.h b/include/sysemu/kvm.h > index 8e76685..0e9ef38 100644 > --- a/include/sysemu/kvm.h > +++ b/include/sysemu/kvm.h > @@ -46,6 +46,7 @@ extern bool kvm_halt_in_kernel_allowed; > extern bool kvm_irqfds_allowed; > extern bool kvm_msi_via_irqfd_allowed; > extern bool kvm_gsi_routing_allowed; > +extern bool kvm_gsi_direct_mapping; > extern bool kvm_readonly_mem_allowed; > > #if defined CONFIG_KVM || !defined NEED_CPU_H > @@ -108,6 +109,13 @@ extern bool kvm_readonly_mem_allowed; > #define kvm_gsi_routing_enabled() (kvm_gsi_routing_allowed) > > /** > + * kvm_gsi_direct_mapping: > + * > + * Returns: true if GSI direct mapping is enabled. > + */ > +#define kvm_gsi_direct_mapping() (kvm_gsi_direct_mapping) > + > +/** > * kvm_readonly_mem_enabled: > * > * Returns: true if KVM readonly memory is enabled (ie the kernel > @@ -123,6 +131,7 @@ extern bool kvm_readonly_mem_allowed; > #define kvm_irqfds_enabled() (false) > #define kvm_msi_via_irqfd_enabled() (false) > #define kvm_gsi_routing_allowed() (false) > +#define kvm_gsi_direct_mapping() (false) > #define kvm_readonly_mem_enabled() (false) > #endif > > diff --git a/kvm-all.c b/kvm-all.c > index 875e32e..17fb865 100644 > --- a/kvm-all.c > +++ b/kvm-all.c > @@ -111,6 +111,7 @@ bool kvm_halt_in_kernel_allowed; > bool kvm_irqfds_allowed; > bool kvm_msi_via_irqfd_allowed; > bool kvm_gsi_routing_allowed; > +bool kvm_gsi_direct_mapping; > bool kvm_allowed; > bool kvm_readonly_mem_allowed; > > @@ -1069,6 +1070,10 @@ void kvm_irqchip_release_virq(KVMState *s, int virq) > struct kvm_irq_routing_entry *e; > int i; > > + if (kvm_gsi_direct_mapping()) { > + return; > + } > + > for (i = 0; i < s->irq_routes->nr; i++) { > e = &s->irq_routes->entries[i]; > if (e->gsi == virq) { > @@ -1190,6 +1195,10 @@ int kvm_irqchip_add_msi_route(KVMState *s, MSIMessage msg) > struct kvm_irq_routing_entry kroute = {}; > int virq; > > + if (kvm_gsi_direct_mapping()) { > + return msg.data & 0xffff; > + } > + > if (!kvm_gsi_routing_enabled()) { > return -ENOSYS; > } > @@ -1216,6 +1225,10 @@ int kvm_irqchip_update_msi_route(KVMState *s, int virq, MSIMessage msg) > { > struct kvm_irq_routing_entry kroute = {}; > > + if (kvm_gsi_direct_mapping()) { > + return 0; > + } > + > if (!kvm_irqchip_in_kernel()) { > return -ENOSYS; > } > diff --git a/kvm-stub.c b/kvm-stub.c > index 548f471..e979f76 100644 > --- a/kvm-stub.c > +++ b/kvm-stub.c > @@ -25,6 +25,7 @@ bool kvm_async_interrupts_allowed; > bool kvm_irqfds_allowed; > bool kvm_msi_via_irqfd_allowed; > bool kvm_gsi_routing_allowed; > +bool kvm_gsi_direct_mapping; > bool kvm_allowed; > bool kvm_readonly_mem_allowed; > > Applied to uq/master, thanks. Paolo
diff --git a/include/sysemu/kvm.h b/include/sysemu/kvm.h index 8e76685..0e9ef38 100644 --- a/include/sysemu/kvm.h +++ b/include/sysemu/kvm.h @@ -46,6 +46,7 @@ extern bool kvm_halt_in_kernel_allowed; extern bool kvm_irqfds_allowed; extern bool kvm_msi_via_irqfd_allowed; extern bool kvm_gsi_routing_allowed; +extern bool kvm_gsi_direct_mapping; extern bool kvm_readonly_mem_allowed; #if defined CONFIG_KVM || !defined NEED_CPU_H @@ -108,6 +109,13 @@ extern bool kvm_readonly_mem_allowed; #define kvm_gsi_routing_enabled() (kvm_gsi_routing_allowed) /** + * kvm_gsi_direct_mapping: + * + * Returns: true if GSI direct mapping is enabled. + */ +#define kvm_gsi_direct_mapping() (kvm_gsi_direct_mapping) + +/** * kvm_readonly_mem_enabled: * * Returns: true if KVM readonly memory is enabled (ie the kernel @@ -123,6 +131,7 @@ extern bool kvm_readonly_mem_allowed; #define kvm_irqfds_enabled() (false) #define kvm_msi_via_irqfd_enabled() (false) #define kvm_gsi_routing_allowed() (false) +#define kvm_gsi_direct_mapping() (false) #define kvm_readonly_mem_enabled() (false) #endif diff --git a/kvm-all.c b/kvm-all.c index 875e32e..17fb865 100644 --- a/kvm-all.c +++ b/kvm-all.c @@ -111,6 +111,7 @@ bool kvm_halt_in_kernel_allowed; bool kvm_irqfds_allowed; bool kvm_msi_via_irqfd_allowed; bool kvm_gsi_routing_allowed; +bool kvm_gsi_direct_mapping; bool kvm_allowed; bool kvm_readonly_mem_allowed; @@ -1069,6 +1070,10 @@ void kvm_irqchip_release_virq(KVMState *s, int virq) struct kvm_irq_routing_entry *e; int i; + if (kvm_gsi_direct_mapping()) { + return; + } + for (i = 0; i < s->irq_routes->nr; i++) { e = &s->irq_routes->entries[i]; if (e->gsi == virq) { @@ -1190,6 +1195,10 @@ int kvm_irqchip_add_msi_route(KVMState *s, MSIMessage msg) struct kvm_irq_routing_entry kroute = {}; int virq; + if (kvm_gsi_direct_mapping()) { + return msg.data & 0xffff; + } + if (!kvm_gsi_routing_enabled()) { return -ENOSYS; } @@ -1216,6 +1225,10 @@ int kvm_irqchip_update_msi_route(KVMState *s, int virq, MSIMessage msg) { struct kvm_irq_routing_entry kroute = {}; + if (kvm_gsi_direct_mapping()) { + return 0; + } + if (!kvm_irqchip_in_kernel()) { return -ENOSYS; } diff --git a/kvm-stub.c b/kvm-stub.c index 548f471..e979f76 100644 --- a/kvm-stub.c +++ b/kvm-stub.c @@ -25,6 +25,7 @@ bool kvm_async_interrupts_allowed; bool kvm_irqfds_allowed; bool kvm_msi_via_irqfd_allowed; bool kvm_gsi_routing_allowed; +bool kvm_gsi_direct_mapping; bool kvm_allowed; bool kvm_readonly_mem_allowed;
On PPC64 systems MSI Messages are translated to system IRQ in a PCI host bridge. This is already supported for emulated MSI/MSIX but not for irqfd where the current QEMU allocates IRQ numbers from irqchip and maps MSIMessages to IRQ in the host kernel. This adds a new direct mapping flag which tells the kvm_irqchip_add_msi_route() function that a new VIRQ should not be allocated, instead the value from MSIMessage::data should be used. It is up to the platform code to make sure that this contains a valid IRQ number as sPAPR does in spapr_pci.c. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> --- The patch does not enable this mapping for any platform in this patch as it is going be done for spapr only on a separate patch which is not ready to go as it depends on the in-kernel XICS-KVM patchset which is not in upstream yet. --- Changes: v6: * simplified to a single global flag and putting an IRQ number in MSIMessage::data 2013/08/07 v5: * pci_bus_map_msi now has default behaviour which is to call kvm_irqchip_add_msi_route * kvm_irqchip_release_virq fixed not crash when there is no routes --- include/sysemu/kvm.h | 9 +++++++++ kvm-all.c | 13 +++++++++++++ kvm-stub.c | 1 + 3 files changed, 23 insertions(+)