* [PATCH v2 0/2] add qemu_irq/gsi hashtable @ 2015-04-23 15:08 ` Eric Auger 0 siblings, 0 replies; 10+ messages in thread From: Eric Auger @ 2015-04-23 15:08 UTC (permalink / raw) To: eric.auger, eric.auger, qemu-devel, peter.maydell, agraf, pbonzini Cc: alex.williamson, kvmarm, patches The VFIO platform needs to setup irqfd but it does not know the gsi associated to the sysbus device qemu_irq. With PCI there is a PCIINTxRoute bus lookup mechanism that enables to retrieve the gsi from the PCI host controller/bridge pin. The conversion is implemented by the PCI host controller. With platform devices, this conversion function is implemented by the interrupt controller. This series implements a hash table in kvm-all that stores the mapping between qemu_irq and gsi. A setter is introduced for the interrupt controller to insert mappings. Also functions are added to setup/tear down irqfd directly from a qemu_irq. Those functions now can be used by the VFIO platform device to setup VFIO signaling. v1 -> v2: - series renamed from "irq: add get_gsi callback" - follow implementation suggested by Paolo (hope I fully understood it) Eric Auger (2): kvm-all.c: add qemu_irq/gsi hash table and utility routines intc: arm_gic_kvm: set the qemu_irq/gsi mapping hw/intc/arm_gic_kvm.c | 5 +++++ include/sysemu/kvm.h | 6 ++++++ kvm-all.c | 35 +++++++++++++++++++++++++++++++++++ 3 files changed, 46 insertions(+) -- 1.8.3.2 ^ permalink raw reply [flat|nested] 10+ messages in thread
* [Qemu-devel] [PATCH v2 0/2] add qemu_irq/gsi hashtable @ 2015-04-23 15:08 ` Eric Auger 0 siblings, 0 replies; 10+ messages in thread From: Eric Auger @ 2015-04-23 15:08 UTC (permalink / raw) To: eric.auger, eric.auger, qemu-devel, peter.maydell, agraf, pbonzini Cc: alex.williamson, kvmarm, christoffer.dall, patches The VFIO platform needs to setup irqfd but it does not know the gsi associated to the sysbus device qemu_irq. With PCI there is a PCIINTxRoute bus lookup mechanism that enables to retrieve the gsi from the PCI host controller/bridge pin. The conversion is implemented by the PCI host controller. With platform devices, this conversion function is implemented by the interrupt controller. This series implements a hash table in kvm-all that stores the mapping between qemu_irq and gsi. A setter is introduced for the interrupt controller to insert mappings. Also functions are added to setup/tear down irqfd directly from a qemu_irq. Those functions now can be used by the VFIO platform device to setup VFIO signaling. v1 -> v2: - series renamed from "irq: add get_gsi callback" - follow implementation suggested by Paolo (hope I fully understood it) Eric Auger (2): kvm-all.c: add qemu_irq/gsi hash table and utility routines intc: arm_gic_kvm: set the qemu_irq/gsi mapping hw/intc/arm_gic_kvm.c | 5 +++++ include/sysemu/kvm.h | 6 ++++++ kvm-all.c | 35 +++++++++++++++++++++++++++++++++++ 3 files changed, 46 insertions(+) -- 1.8.3.2 ^ permalink raw reply [flat|nested] 10+ messages in thread
* [PATCH v2 1/2] kvm-all.c: add qemu_irq/gsi hash table and utility routines 2015-04-23 15:08 ` [Qemu-devel] " Eric Auger @ 2015-04-23 15:08 ` Eric Auger -1 siblings, 0 replies; 10+ messages in thread From: Eric Auger @ 2015-04-23 15:08 UTC (permalink / raw) To: eric.auger, eric.auger, qemu-devel, peter.maydell, agraf, pbonzini Cc: alex.williamson, kvmarm, patches VFIO platform device needs to setup irqfd but it does not know the gsi corresponding to the device qemu_irq. This series proposes to store a hash table in kvm_state using the qemu_irq as key and the gsi as a value. kvm_irqchip_set_qemuirq_gsi allows to insert such a pair. The interrupt controller is supposed to use it. kvm_irqchip_[add, remove]_qemuirq_irqfd_notifier allows to setup/tear down irqfd directly from the qemu_irq. Signed-off-by: Eric Auger <eric.auger@linaro.org> --- include/sysemu/kvm.h | 6 ++++++ kvm-all.c | 35 +++++++++++++++++++++++++++++++++++ 2 files changed, 41 insertions(+) diff --git a/include/sysemu/kvm.h b/include/sysemu/kvm.h index 197e6c0..911aaa8 100644 --- a/include/sysemu/kvm.h +++ b/include/sysemu/kvm.h @@ -18,6 +18,7 @@ #include "config-host.h" #include "qemu/queue.h" #include "qom/cpu.h" +#include "hw/irq.h" #ifdef CONFIG_KVM #include <linux/kvm.h> @@ -416,6 +417,11 @@ int kvm_irqchip_add_adapter_route(KVMState *s, AdapterInfo *adapter); int kvm_irqchip_add_irqfd_notifier(KVMState *s, EventNotifier *n, EventNotifier *rn, int virq); int kvm_irqchip_remove_irqfd_notifier(KVMState *s, EventNotifier *n, int virq); +int kvm_irqchip_add_qemuirq_irqfd_notifier(KVMState *s, EventNotifier *n, + EventNotifier *rn, qemu_irq irq); +int kvm_irqchip_remove_qemuirq_irqfd_notifier(KVMState *s, EventNotifier *n, + qemu_irq irq); +void kvm_irqchip_set_qemuirq_gsi(KVMState *s, qemu_irq irq, int gsi); void kvm_pc_gsi_handler(void *opaque, int n, int level); void kvm_pc_setup_irq_routing(bool pci_enabled); void kvm_init_irq_routing(KVMState *s); diff --git a/kvm-all.c b/kvm-all.c index dd44f8c..ec9f4ac 100644 --- a/kvm-all.c +++ b/kvm-all.c @@ -37,6 +37,7 @@ #include "exec/address-spaces.h" #include "qemu/event_notifier.h" #include "trace.h" +#include "hw/irq.h" #include "hw/boards.h" @@ -99,6 +100,7 @@ struct KVMState * unsigned, and treating them as signed here can break things */ unsigned irq_set_ioctl; unsigned int sigmask_len; + GHashTable *gsimap; #ifdef KVM_CAP_IRQ_ROUTING struct kvm_irq_routing *irq_routes; int nr_allocated_irq_routes; @@ -1381,6 +1383,37 @@ int kvm_irqchip_remove_irqfd_notifier(KVMState *s, EventNotifier *n, int virq) false); } +int kvm_irqchip_add_qemuirq_irqfd_notifier(KVMState *s, EventNotifier *n, + EventNotifier *rn, qemu_irq irq) +{ + gpointer key, gsi; + gboolean found = g_hash_table_lookup_extended(s->gsimap, irq, &key, &gsi); + + if (!found) { + return -ENXIO; + } else { + return kvm_irqchip_add_irqfd_notifier(s, n, rn, GPOINTER_TO_UINT(gsi)); + } +} + +int kvm_irqchip_remove_qemuirq_irqfd_notifier(KVMState *s, EventNotifier *n, + qemu_irq irq) +{ + gpointer key, gsi; + gboolean found = g_hash_table_lookup_extended(s->gsimap, irq, &key, &gsi); + + if (!found) { + return -ENXIO; + } else { + return kvm_irqchip_remove_irqfd_notifier(s, n, GPOINTER_TO_INT(gsi)); + } +} + +void kvm_irqchip_set_qemuirq_gsi(KVMState *s, qemu_irq irq, int gsi) +{ + g_hash_table_insert(s->gsimap, irq, GINT_TO_POINTER(gsi)); +} + static int kvm_irqchip_create(MachineState *machine, KVMState *s) { int ret; @@ -1413,6 +1446,8 @@ static int kvm_irqchip_create(MachineState *machine, KVMState *s) kvm_init_irq_routing(s); + s->gsimap = g_hash_table_new(g_direct_hash, g_direct_equal); + return 0; } -- 1.8.3.2 ^ permalink raw reply related [flat|nested] 10+ messages in thread
* [Qemu-devel] [PATCH v2 1/2] kvm-all.c: add qemu_irq/gsi hash table and utility routines @ 2015-04-23 15:08 ` Eric Auger 0 siblings, 0 replies; 10+ messages in thread From: Eric Auger @ 2015-04-23 15:08 UTC (permalink / raw) To: eric.auger, eric.auger, qemu-devel, peter.maydell, agraf, pbonzini Cc: alex.williamson, kvmarm, christoffer.dall, patches VFIO platform device needs to setup irqfd but it does not know the gsi corresponding to the device qemu_irq. This series proposes to store a hash table in kvm_state using the qemu_irq as key and the gsi as a value. kvm_irqchip_set_qemuirq_gsi allows to insert such a pair. The interrupt controller is supposed to use it. kvm_irqchip_[add, remove]_qemuirq_irqfd_notifier allows to setup/tear down irqfd directly from the qemu_irq. Signed-off-by: Eric Auger <eric.auger@linaro.org> --- include/sysemu/kvm.h | 6 ++++++ kvm-all.c | 35 +++++++++++++++++++++++++++++++++++ 2 files changed, 41 insertions(+) diff --git a/include/sysemu/kvm.h b/include/sysemu/kvm.h index 197e6c0..911aaa8 100644 --- a/include/sysemu/kvm.h +++ b/include/sysemu/kvm.h @@ -18,6 +18,7 @@ #include "config-host.h" #include "qemu/queue.h" #include "qom/cpu.h" +#include "hw/irq.h" #ifdef CONFIG_KVM #include <linux/kvm.h> @@ -416,6 +417,11 @@ int kvm_irqchip_add_adapter_route(KVMState *s, AdapterInfo *adapter); int kvm_irqchip_add_irqfd_notifier(KVMState *s, EventNotifier *n, EventNotifier *rn, int virq); int kvm_irqchip_remove_irqfd_notifier(KVMState *s, EventNotifier *n, int virq); +int kvm_irqchip_add_qemuirq_irqfd_notifier(KVMState *s, EventNotifier *n, + EventNotifier *rn, qemu_irq irq); +int kvm_irqchip_remove_qemuirq_irqfd_notifier(KVMState *s, EventNotifier *n, + qemu_irq irq); +void kvm_irqchip_set_qemuirq_gsi(KVMState *s, qemu_irq irq, int gsi); void kvm_pc_gsi_handler(void *opaque, int n, int level); void kvm_pc_setup_irq_routing(bool pci_enabled); void kvm_init_irq_routing(KVMState *s); diff --git a/kvm-all.c b/kvm-all.c index dd44f8c..ec9f4ac 100644 --- a/kvm-all.c +++ b/kvm-all.c @@ -37,6 +37,7 @@ #include "exec/address-spaces.h" #include "qemu/event_notifier.h" #include "trace.h" +#include "hw/irq.h" #include "hw/boards.h" @@ -99,6 +100,7 @@ struct KVMState * unsigned, and treating them as signed here can break things */ unsigned irq_set_ioctl; unsigned int sigmask_len; + GHashTable *gsimap; #ifdef KVM_CAP_IRQ_ROUTING struct kvm_irq_routing *irq_routes; int nr_allocated_irq_routes; @@ -1381,6 +1383,37 @@ int kvm_irqchip_remove_irqfd_notifier(KVMState *s, EventNotifier *n, int virq) false); } +int kvm_irqchip_add_qemuirq_irqfd_notifier(KVMState *s, EventNotifier *n, + EventNotifier *rn, qemu_irq irq) +{ + gpointer key, gsi; + gboolean found = g_hash_table_lookup_extended(s->gsimap, irq, &key, &gsi); + + if (!found) { + return -ENXIO; + } else { + return kvm_irqchip_add_irqfd_notifier(s, n, rn, GPOINTER_TO_UINT(gsi)); + } +} + +int kvm_irqchip_remove_qemuirq_irqfd_notifier(KVMState *s, EventNotifier *n, + qemu_irq irq) +{ + gpointer key, gsi; + gboolean found = g_hash_table_lookup_extended(s->gsimap, irq, &key, &gsi); + + if (!found) { + return -ENXIO; + } else { + return kvm_irqchip_remove_irqfd_notifier(s, n, GPOINTER_TO_INT(gsi)); + } +} + +void kvm_irqchip_set_qemuirq_gsi(KVMState *s, qemu_irq irq, int gsi) +{ + g_hash_table_insert(s->gsimap, irq, GINT_TO_POINTER(gsi)); +} + static int kvm_irqchip_create(MachineState *machine, KVMState *s) { int ret; @@ -1413,6 +1446,8 @@ static int kvm_irqchip_create(MachineState *machine, KVMState *s) kvm_init_irq_routing(s); + s->gsimap = g_hash_table_new(g_direct_hash, g_direct_equal); + return 0; } -- 1.8.3.2 ^ permalink raw reply related [flat|nested] 10+ messages in thread
* Re: [PATCH v2 1/2] kvm-all.c: add qemu_irq/gsi hash table and utility routines 2015-04-23 15:08 ` [Qemu-devel] " Eric Auger @ 2015-04-23 15:22 ` Paolo Bonzini -1 siblings, 0 replies; 10+ messages in thread From: Paolo Bonzini @ 2015-04-23 15:22 UTC (permalink / raw) To: Eric Auger, eric.auger, qemu-devel, peter.maydell, agraf Cc: alex.williamson, kvmarm, christoffer.dall, patches On 23/04/2015 17:08, Eric Auger wrote: > int kvm_irqchip_add_irqfd_notifier(KVMState *s, EventNotifier *n, > EventNotifier *rn, int virq); > int kvm_irqchip_remove_irqfd_notifier(KVMState *s, EventNotifier *n, int virq); > +int kvm_irqchip_add_qemuirq_irqfd_notifier(KVMState *s, EventNotifier *n, > + EventNotifier *rn, qemu_irq irq); > +int kvm_irqchip_remove_qemuirq_irqfd_notifier(KVMState *s, EventNotifier *n, > + qemu_irq irq); Perhaps rename the existing ones to kvm_irqchip_add_irqfd_notifier_gsi kvm_irqchip_remove_irqfd_notifier_gsi and change the new ones to kvm_irqchip_add_irqfd_notifier and kvm_irqchip_remove_irqfd_notifier? Okay with that change; include it in the vfio-platform series and I'll ack it. Paolo ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [Qemu-devel] [PATCH v2 1/2] kvm-all.c: add qemu_irq/gsi hash table and utility routines @ 2015-04-23 15:22 ` Paolo Bonzini 0 siblings, 0 replies; 10+ messages in thread From: Paolo Bonzini @ 2015-04-23 15:22 UTC (permalink / raw) To: Eric Auger, eric.auger, qemu-devel, peter.maydell, agraf Cc: alex.williamson, kvmarm, christoffer.dall, patches On 23/04/2015 17:08, Eric Auger wrote: > int kvm_irqchip_add_irqfd_notifier(KVMState *s, EventNotifier *n, > EventNotifier *rn, int virq); > int kvm_irqchip_remove_irqfd_notifier(KVMState *s, EventNotifier *n, int virq); > +int kvm_irqchip_add_qemuirq_irqfd_notifier(KVMState *s, EventNotifier *n, > + EventNotifier *rn, qemu_irq irq); > +int kvm_irqchip_remove_qemuirq_irqfd_notifier(KVMState *s, EventNotifier *n, > + qemu_irq irq); Perhaps rename the existing ones to kvm_irqchip_add_irqfd_notifier_gsi kvm_irqchip_remove_irqfd_notifier_gsi and change the new ones to kvm_irqchip_add_irqfd_notifier and kvm_irqchip_remove_irqfd_notifier? Okay with that change; include it in the vfio-platform series and I'll ack it. Paolo ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH v2 1/2] kvm-all.c: add qemu_irq/gsi hash table and utility routines 2015-04-23 15:22 ` [Qemu-devel] " Paolo Bonzini @ 2015-04-23 15:21 ` Eric Auger -1 siblings, 0 replies; 10+ messages in thread From: Eric Auger @ 2015-04-23 15:21 UTC (permalink / raw) To: Paolo Bonzini, eric.auger, qemu-devel, peter.maydell, agraf Cc: alex.williamson, kvmarm, patches On 04/23/2015 05:22 PM, Paolo Bonzini wrote: > > > On 23/04/2015 17:08, Eric Auger wrote: >> int kvm_irqchip_add_irqfd_notifier(KVMState *s, EventNotifier *n, >> EventNotifier *rn, int virq); >> int kvm_irqchip_remove_irqfd_notifier(KVMState *s, EventNotifier *n, int virq); >> +int kvm_irqchip_add_qemuirq_irqfd_notifier(KVMState *s, EventNotifier *n, >> + EventNotifier *rn, qemu_irq irq); >> +int kvm_irqchip_remove_qemuirq_irqfd_notifier(KVMState *s, EventNotifier *n, >> + qemu_irq irq); > > > Perhaps rename the existing ones to > > kvm_irqchip_add_irqfd_notifier_gsi > kvm_irqchip_remove_irqfd_notifier_gsi > > and change the new ones to kvm_irqchip_add_irqfd_notifier and > kvm_irqchip_remove_irqfd_notifier? OK thanks! Eric > > Okay with that change; include it in the vfio-platform series and I'll > ack it. > > Paolo > ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [Qemu-devel] [PATCH v2 1/2] kvm-all.c: add qemu_irq/gsi hash table and utility routines @ 2015-04-23 15:21 ` Eric Auger 0 siblings, 0 replies; 10+ messages in thread From: Eric Auger @ 2015-04-23 15:21 UTC (permalink / raw) To: Paolo Bonzini, eric.auger, qemu-devel, peter.maydell, agraf Cc: alex.williamson, kvmarm, christoffer.dall, patches On 04/23/2015 05:22 PM, Paolo Bonzini wrote: > > > On 23/04/2015 17:08, Eric Auger wrote: >> int kvm_irqchip_add_irqfd_notifier(KVMState *s, EventNotifier *n, >> EventNotifier *rn, int virq); >> int kvm_irqchip_remove_irqfd_notifier(KVMState *s, EventNotifier *n, int virq); >> +int kvm_irqchip_add_qemuirq_irqfd_notifier(KVMState *s, EventNotifier *n, >> + EventNotifier *rn, qemu_irq irq); >> +int kvm_irqchip_remove_qemuirq_irqfd_notifier(KVMState *s, EventNotifier *n, >> + qemu_irq irq); > > > Perhaps rename the existing ones to > > kvm_irqchip_add_irqfd_notifier_gsi > kvm_irqchip_remove_irqfd_notifier_gsi > > and change the new ones to kvm_irqchip_add_irqfd_notifier and > kvm_irqchip_remove_irqfd_notifier? OK thanks! Eric > > Okay with that change; include it in the vfio-platform series and I'll > ack it. > > Paolo > ^ permalink raw reply [flat|nested] 10+ messages in thread
* [PATCH v2 2/2] intc: arm_gic_kvm: set the qemu_irq/gsi mapping 2015-04-23 15:08 ` [Qemu-devel] " Eric Auger @ 2015-04-23 15:08 ` Eric Auger -1 siblings, 0 replies; 10+ messages in thread From: Eric Auger @ 2015-04-23 15:08 UTC (permalink / raw) To: eric.auger, eric.auger, qemu-devel, peter.maydell, agraf, pbonzini Cc: alex.williamson, kvmarm, patches The arm_gic_kvm now calls kvm_irqchip_set_qemuirq_gsi to build the hash table storing qemu_irq/gsi mappings. From that point on irqfd can be setup directly from the qemu_irq using kvm_irqchip_add_qemuirq_irqfd_notifier. Signed-off-by: Eric Auger <eric.auger@linaro.org> --- hw/intc/arm_gic_kvm.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/hw/intc/arm_gic_kvm.c b/hw/intc/arm_gic_kvm.c index e1952ad..28506e3 100644 --- a/hw/intc/arm_gic_kvm.c +++ b/hw/intc/arm_gic_kvm.c @@ -554,6 +554,11 @@ static void kvm_arm_gic_realize(DeviceState *dev, Error **errp) */ i += (GIC_INTERNAL * s->num_cpu); qdev_init_gpio_in(dev, kvm_arm_gic_set_irq, i); + + for (i = 0; i < s->num_irq - GIC_INTERNAL; i++) { + qemu_irq irq = qdev_get_gpio_in(dev, i); + kvm_irqchip_set_qemuirq_gsi(kvm_state, irq, i); + } /* We never use our outbound IRQ lines but provide them so that * we maintain the same interface as the non-KVM GIC. */ -- 1.8.3.2 ^ permalink raw reply related [flat|nested] 10+ messages in thread
* [Qemu-devel] [PATCH v2 2/2] intc: arm_gic_kvm: set the qemu_irq/gsi mapping @ 2015-04-23 15:08 ` Eric Auger 0 siblings, 0 replies; 10+ messages in thread From: Eric Auger @ 2015-04-23 15:08 UTC (permalink / raw) To: eric.auger, eric.auger, qemu-devel, peter.maydell, agraf, pbonzini Cc: alex.williamson, kvmarm, christoffer.dall, patches The arm_gic_kvm now calls kvm_irqchip_set_qemuirq_gsi to build the hash table storing qemu_irq/gsi mappings. From that point on irqfd can be setup directly from the qemu_irq using kvm_irqchip_add_qemuirq_irqfd_notifier. Signed-off-by: Eric Auger <eric.auger@linaro.org> --- hw/intc/arm_gic_kvm.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/hw/intc/arm_gic_kvm.c b/hw/intc/arm_gic_kvm.c index e1952ad..28506e3 100644 --- a/hw/intc/arm_gic_kvm.c +++ b/hw/intc/arm_gic_kvm.c @@ -554,6 +554,11 @@ static void kvm_arm_gic_realize(DeviceState *dev, Error **errp) */ i += (GIC_INTERNAL * s->num_cpu); qdev_init_gpio_in(dev, kvm_arm_gic_set_irq, i); + + for (i = 0; i < s->num_irq - GIC_INTERNAL; i++) { + qemu_irq irq = qdev_get_gpio_in(dev, i); + kvm_irqchip_set_qemuirq_gsi(kvm_state, irq, i); + } /* We never use our outbound IRQ lines but provide them so that * we maintain the same interface as the non-KVM GIC. */ -- 1.8.3.2 ^ permalink raw reply related [flat|nested] 10+ messages in thread
end of thread, other threads:[~2015-04-23 15:24 UTC | newest] Thread overview: 10+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2015-04-23 15:08 [PATCH v2 0/2] add qemu_irq/gsi hashtable Eric Auger 2015-04-23 15:08 ` [Qemu-devel] " Eric Auger 2015-04-23 15:08 ` [PATCH v2 1/2] kvm-all.c: add qemu_irq/gsi hash table and utility routines Eric Auger 2015-04-23 15:08 ` [Qemu-devel] " Eric Auger 2015-04-23 15:22 ` Paolo Bonzini 2015-04-23 15:22 ` [Qemu-devel] " Paolo Bonzini 2015-04-23 15:21 ` Eric Auger 2015-04-23 15:21 ` [Qemu-devel] " Eric Auger 2015-04-23 15:08 ` [PATCH v2 2/2] intc: arm_gic_kvm: set the qemu_irq/gsi mapping Eric Auger 2015-04-23 15:08 ` [Qemu-devel] " Eric Auger
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.