diff options
| author | Linus Torvalds <torvalds@linux-foundation.org> | 2017-05-10 11:29:23 -0700 |
|---|---|---|
| committer | Linus Torvalds <torvalds@linux-foundation.org> | 2017-05-10 11:29:23 -0700 |
| commit | 5ccd414080822d5257c3569f4aeca74f63f4a257 (patch) | |
| tree | 4b567bfc0a8b29dbac9712821062b28337eb3408 | |
| parent | 29250d301b0c75ef142b51eebee6b7403cc79624 (diff) | |
| parent | 36c344f3f1ffc0b1b20abd237b7401dc6687ee8f (diff) | |
| download | linux-5ccd414080822d5257c3569f4aeca74f63f4a257.tar.bz2 | |
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Pull more KVM updates from Paolo Bonzini:
"ARM:
- bugfixes
- moved shared 32-bit/64-bit files to virt/kvm/arm
- support for saving/restoring virtual ITS state to userspace
PPC:
- XIVE (eXternal Interrupt Virtualization Engine) support
x86:
- nVMX improvements, including emulated page modification logging
(PML) which brings nice performance improvements on some workloads"
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (45 commits)
KVM: arm/arm64: vgic-its: Cleanup after failed ITT restore
KVM: arm/arm64: Don't call map_resources when restoring ITS tables
KVM: arm/arm64: Register ITS iodev when setting base address
KVM: arm/arm64: Get rid of its->initialized field
KVM: arm/arm64: Register iodevs when setting redist base and creating VCPUs
KVM: arm/arm64: Slightly rework kvm_vgic_addr
KVM: arm/arm64: Make vgic_v3_check_base more broadly usable
KVM: arm/arm64: Refactor vgic_register_redist_iodevs
KVM: Add kvm_vcpu_get_idx to get vcpu index in kvm->vcpus
nVMX: Advertise PML to L1 hypervisor
nVMX: Implement emulated Page Modification Logging
kvm: x86: Add a hook for arch specific dirty logging emulation
kvm: nVMX: Validate CR3 target count on nested VM-entry
KVM: set no_llseek in stat_fops_per_vm
KVM: arm/arm64: vgic: Rename kvm_vgic_vcpu_init to kvm_vgic_vcpu_enable
KVM: arm/arm64: Clarification and relaxation to ITS save/restore ABI
KVM: arm64: vgic-v3: KVM_DEV_ARM_VGIC_SAVE_PENDING_TABLES
KVM: arm64: vgic-its: Fix pending table sync
KVM: arm64: vgic-its: ITT save and restore
KVM: arm64: vgic-its: Device table save/restore
...
58 files changed, 5318 insertions, 635 deletions
diff --git a/Documentation/virtual/kvm/devices/arm-vgic-its.txt b/Documentation/virtual/kvm/devices/arm-vgic-its.txt index 6081a5b7fc1e..eb06beb75960 100644 --- a/Documentation/virtual/kvm/devices/arm-vgic-its.txt +++ b/Documentation/virtual/kvm/devices/arm-vgic-its.txt @@ -32,7 +32,128 @@ Groups: KVM_DEV_ARM_VGIC_CTRL_INIT request the initialization of the ITS, no additional parameter in kvm_device_attr.addr. + + KVM_DEV_ARM_ITS_SAVE_TABLES + save the ITS table data into guest RAM, at the location provisioned + by the guest in corresponding registers/table entries. + + The layout of the tables in guest memory defines an ABI. The entries + are laid out in little endian format as described in the last paragraph. + + KVM_DEV_ARM_ITS_RESTORE_TABLES + restore the ITS tables from guest RAM to ITS internal structures. + + The GICV3 must be restored before the ITS and all ITS registers but + the GITS_CTLR must be restored before restoring the ITS tables. + + The GITS_IIDR read-only register must also be restored before + calling KVM_DEV_ARM_ITS_RESTORE_TABLES as the IIDR revision field + encodes the ABI revision. + + The expected ordering when restoring the GICv3/ITS is described in section + "ITS Restore Sequence". + Errors: -ENXIO: ITS not properly configured as required prior to setting this attribute -ENOMEM: Memory shortage when allocating ITS internal data + -EINVAL: Inconsistent restored data + -EFAULT: Invalid guest ram access + -EBUSY: One or more VCPUS are running + + KVM_DEV_ARM_VGIC_GRP_ITS_REGS + Attributes: + The attr field of kvm_device_attr encodes the offset of the + ITS register, relative to the ITS control frame base address + (ITS_base). + + kvm_device_attr.addr points to a __u64 value whatever the width + of the addressed register (32/64 bits). 64 bit registers can only + be accessed with full length. + + Writes to read-only registers are ignored by the kernel except for: + - GITS_CREADR. It must be restored otherwise commands in the queue + will be re-executed after restoring CWRITER. GITS_CREADR must be + restored before restoring the GITS_CTLR which is likely to enable the + ITS. Also it must be restored after GITS_CBASER since a write to + GITS_CBASER resets GITS_CREADR. + - GITS_IIDR. The Revision field encodes the table layout ABI revision. + In the future we might implement direct injection of virtual LPIs. + This will require an upgrade of the table layout and an evolution of + the ABI. GITS_IIDR must be restored before calling + KVM_DEV_ARM_ITS_RESTORE_TABLES. + + For other registers, getting or setting a register has the same + effect as reading/writing the register on real hardware. + Errors: + -ENXIO: Offset does not correspond to any supported register + -EFAULT: Invalid user pointer for attr->addr + -EINVAL: Offset is not 64-bit aligned + -EBUSY: one or more VCPUS are running + + ITS Restore Sequence: + ------------------------- + +The following ordering must be followed when restoring the GIC and the ITS: +a) restore all guest memory and create vcpus +b) restore all redistributors +c) provide the its base address + (KVM_DEV_ARM_VGIC_GRP_ADDR) +d) restore the ITS in the following order: + 1. Restore GITS_CBASER + 2. Restore all other GITS_ registers, except GITS_CTLR! + 3. Load the ITS table data (KVM_DEV_ARM_ITS_RESTORE_TABLES) + 4. Restore GITS_CTLR + +Then vcpus can be started. + + ITS Table ABI REV0: + ------------------- + + Revision 0 of the ABI only supports the features of a virtual GICv3, and does + not support a virtual GICv4 with support for direct injection of virtual + interrupts for nested hypervisors. + + The device table and ITT are indexed by the DeviceID and EventID, + respectively. The collection table is not indexed by CollectionID, and the + entries in the collection are listed in no particular order. + All entries are 8 bytes. + + Device Table Entry (DTE): + + bits: | 63| 62 ... 49 | 48 ... 5 | 4 ... 0 | + values: | V | next | ITT_addr | Size | + + where; + - V indicates whether the entry is valid. If not, other fields + are not meaningful. + - next: equals to 0 if this entry is the last one; otherwise it + corresponds to the DeviceID offset to the next DTE, capped by + 2^14 -1. + - ITT_addr matches bits [51:8] of the ITT address (256 Byte aligned). + - Size specifies the supported number of bits for the EventID, + minus one + + Collection Table Entry (CTE): + + bits: | 63| 62 .. 52 | 51 ... 16 | 15 ... 0 | + values: | V | RES0 | RDBase | ICID | + + where: + - V indicates whether the entry is valid. If not, other fields are + not meaningful. + - RES0: reserved field with Should-Be-Zero-or-Preserved behavior. + - RDBase is the PE number (GICR_TYPER.Processor_Number semantic), + - ICID is the collection ID + + Interrupt Translation Entry (ITE): + + bits: | 63 ... 48 | 47 ... 16 | 15 ... 0 | + values: | next | pINTID | ICID | + + where: + - next: equals to 0 if this entry is the last one; otherwise it corresponds + to the EventID offset to the next ITE capped by 2^16 -1. + - pINTID is the physical LPI ID; if zero, it means the entry is not valid + and other fields are not meaningful. + - ICID is the collection ID diff --git a/Documentation/virtual/kvm/devices/arm-vgic-v3.txt b/Documentation/virtual/kvm/devices/arm-vgic-v3.txt index c1a24612c198..9293b45abdb9 100644 --- a/Documentation/virtual/kvm/devices/arm-vgic-v3.txt +++ b/Documentation/virtual/kvm/devices/arm-vgic-v3.txt @@ -167,11 +167,17 @@ Groups: KVM_DEV_ARM_VGIC_CTRL_INIT request the initialization of the VGIC, no additional parameter in kvm_device_attr.addr. + KVM_DEV_ARM_VGIC_SAVE_PENDING_TABLES + save all LPI pending bits into guest RAM pending tables. + + The first kB of the pending table is not altered by this operation. Errors: -ENXIO: VGIC not properly configured as required prior to calling this attribute -ENODEV: no online VCPU -ENOMEM: memory shortage when allocating vgic internal data + -EFAULT: Invalid guest ram access + -EBUSY: One or more VCPUS are running KVM_DEV_ARM_VGIC_GRP_LEVEL_INFO diff --git a/arch/arm/include/uapi/asm/kvm.h b/arch/arm/include/uapi/asm/kvm.h index a88726359e5f..5e3c673fa3f4 100644 --- a/arch/arm/include/uapi/asm/kvm.h +++ b/arch/arm/include/uapi/asm/kvm.h @@ -196,13 +196,17 @@ struct kvm_arch_memory_slot { #define KVM_DEV_ARM_VGIC_GRP_REDIST_REGS 5 #define KVM_DEV_ARM_VGIC_GRP_CPU_SYSREGS 6 #define KVM_DEV_ARM_VGIC_GRP_LEVEL_INFO 7 +#define KVM_DEV_ARM_VGIC_GRP_ITS_REGS 8 #define KVM_DEV_ARM_VGIC_LINE_LEVEL_INFO_SHIFT 10 #define KVM_DEV_ARM_VGIC_LINE_LEVEL_INFO_MASK \ (0x3fffffULL << KVM_DEV_ARM_VGIC_LINE_LEVEL_INFO_SHIFT) #define KVM_DEV_ARM_VGIC_LINE_LEVEL_INTID_MASK 0x3ff #define VGIC_LEVEL_INFO_LINE_LEVEL 0 -#define KVM_DEV_ARM_VGIC_CTRL_INIT 0 +#define KVM_DEV_ARM_VGIC_CTRL_INIT 0 +#define KVM_DEV_ARM_ITS_SAVE_TABLES 1 +#define KVM_DEV_ARM_ITS_RESTORE_TABLES 2 +#define KVM_DEV_ARM_VGIC_SAVE_PENDING_TABLES 3 /* KVM_IRQ_LINE irq field index values */ #define KVM_ARM_IRQ_TYPE_SHIFT 24 diff --git a/arch/arm/kvm/Makefile b/arch/arm/kvm/Makefile index 7b3670c2ae7b..d9beee652d36 100644 --- a/arch/arm/kvm/Makefile +++ b/arch/arm/kvm/Makefile @@ -18,9 +18,12 @@ KVM := ../../../virt/kvm kvm-arm-y = $(KVM)/kvm_main.o $(KVM)/coalesced_mmio.o $(KVM)/eventfd.o $(KVM)/vfio.o obj-$(CONFIG_KVM_ARM_HOST) += hyp/ + obj-y += kvm-arm.o init.o interrupts.o -obj-y += arm.o handle_exit.o guest.o mmu.o emulate.o reset.o -obj-y += coproc.o coproc_a15.o coproc_a7.o mmio.o psci.o perf.o vgic-v3-coproc.o +obj-y += handle_exit.o guest.o emulate.o reset.o +obj-y += coproc.o coproc_a15.o coproc_a7.o vgic-v3-coproc.o +obj-y += $(KVM)/arm/arm.o $(KVM)/arm/mmu.o $(KVM)/arm/mmio.o +obj-y += $(KVM)/arm/psci.o $(KVM)/arm/perf.o obj-y += $(KVM)/arm/aarch32.o obj-y += $(KVM)/arm/vgic/vgic.o diff --git a/arch/arm/kvm/trace.h b/arch/arm/kvm/trace.h index c25a88598eb0..fc0943776db2 100644 --- a/arch/arm/kvm/trace.h +++ b/arch/arm/kvm/trace.h @@ -6,133 +6,6 @@ #undef TRACE_SYSTEM #define TRACE_SYSTEM kvm -/* - * Tracepoints for entry/exit to guest - */ -TRACE_EVENT(kvm_entry, - TP_PROTO(unsigned long vcpu_pc), - TP_ARGS(vcpu_pc), - - TP_STRUCT__entry( - __field( unsigned long, vcpu_pc ) - ), - - TP_fast_assign( - __entry->vcpu_pc = vcpu_pc; - ), - - TP_printk("PC: 0x%08lx", __entry->vcpu_pc) -); - -TRACE_EVENT(kvm_exit, - TP_PROTO(int idx, unsigned int exit_reason, unsigned long vcpu_pc), - TP_ARGS(idx, exit_reason, vcpu_pc), - - TP_STRUCT__entry( - __field( int, idx ) - __field( unsigned int, exit_reason ) - __field( unsigned long, vcpu_pc ) - ), - - TP_fast_assign( - __entry->idx = idx; - __entry->exit_reason = exit_reason; - __entry->vcpu_pc = vcpu_pc; - ), - - TP_printk("%s: HSR_EC: 0x%04x (%s), PC: 0x%08lx", - __print_symbolic(__entry->idx, kvm_arm_exception_type), - __entry->exit_reason, - __print_symbolic(__entry->exit_reason, kvm_arm_exception_class), - __entry->vcpu_pc) -); - -TRACE_EVENT(kvm_guest_fault, - TP_PROTO(unsigned long vcpu_pc, unsigned long hsr, - unsigned long hxfar, - unsigned long long ipa), - TP_ARGS(vcpu_pc, hsr, hxfar, ipa), - - TP_STRUCT__entry( - __field( unsigned long, vcpu_pc ) - __field( unsigned long, hsr ) - __field( unsigned long, hxfar ) - __field( unsigned long long, ipa ) - ), - - TP_fast_assign( - __entry->vcpu_pc = vcpu_pc; - __entry->hsr = hsr; - __entry->hxfar = hxfar; - __entry->ipa = ipa; - ), - - TP_printk("ipa %#llx, hsr %#08lx, hxfar %#08lx, pc %#08lx", - __entry->ipa, __entry->hsr, - __entry->hxfar, __entry->vcpu_pc) -); - -TRACE_EVENT(kvm_access_fault, - TP_PROTO(unsigned long ipa), - TP_ARGS(ipa), - - TP_STRUCT__entry( - __field( unsigned long, ipa ) - ), - - TP_fast_assign( - __entry->ipa = ipa; - ), - - TP_printk("IPA: %lx", __entry->ipa) -); - -TRACE_EVENT(kvm_irq_line, - TP_PROTO(unsigned int type, int vcpu_idx, int irq_num, int level), - TP_ARGS(type, vcpu_idx, irq_num, level), - - TP_STRUCT__entry( - __field( unsigned int, type ) - __field( int, vcpu_idx ) - __field( int, irq_num ) - __field( int, level ) - ), - - TP_fast_assign( - __entry->type = type; - __entry->vcpu_idx = vcpu_idx; - __entry->irq_num = irq_num; - __entry->level = level; - ), - - TP_printk("Inject %s interrupt (%d), vcpu->idx: %d, num: %d, level: %d", - (__entry->type == KVM_ARM_IRQ_TYPE_CPU) ? "CPU" : - (__entry->type == KVM_ARM_IRQ_TYPE_PPI) ? "VGIC PPI" : - (__entry->type == KVM_ARM_IRQ_TYPE_SPI) ? "VGIC SPI" : "UNKNOWN", - __entry->type, __entry->vcpu_idx, __entry->irq_num, __entry->level) -); - -TRACE_EVENT(kvm_mmio_emulate, - TP_PROTO(unsigned long vcpu_pc, unsigned long instr, - unsigned long cpsr), - TP_ARGS(vcpu_pc, instr, cpsr), - - TP_STRUCT__entry( - __field( unsigned long, vcpu_pc ) - __field( unsigned long, instr ) - __field( unsigned long, cpsr ) - ), - - TP_fast_assign( - __entry->vcpu_pc = vcpu_pc; - __entry->instr = instr; - __entry->cpsr = cpsr; - ), - - TP_printk("Emulate MMIO at: 0x%08lx (instr: %08lx, cpsr: %08lx)", - __entry->vcpu_pc, __entry->instr, __entry->cpsr) -); - /* Architecturally implementation defined CP15 register access */ TRACE_EVENT(kvm_emulate_cp15_imp, TP_PROTO(unsigned long Op1, unsigned long Rt1, unsigned long CRn, @@ -181,87 +54,6 @@ TRACE_EVENT(kvm_wfx, __entry->is_wfe ? 'e' : 'i', __entry->vcpu_pc) ); -TRACE_EVENT(kvm_unmap_hva, - TP_PROTO(unsigned long hva), - TP_ARGS(hva), - - TP_STRUCT__entry( - __field( unsigned long, hva ) - ), - - TP_fast_assign( - __entry->hva = hva; - ), - - TP_printk("mmu notifier unmap hva: %#08lx", __entry->hva) -); - -TRACE_EVENT(kvm_unmap_hva_range, - TP_PROTO(unsigned long start, unsigned long end), - TP_ARGS(start, end), - - TP_STRUCT__entry( - __field( unsigned long, start ) - __field( unsigned long, end ) - ), - - TP_fast_assign( - __entry->start = start; - __entry->end = end; - ), - - TP_printk("mmu notifier unmap range: %#08lx -- %#08lx", - __entry->start, __entry->end) -); - -TRACE_EVENT(kvm_set_spte_hva, - TP_PROTO(unsigned long hva), - TP_ARGS(hva), - - TP_STRUCT__entry( - __field( unsigned long, hva ) - ), - - TP_fast_assign( - __entry->hva = hva; - ), - - TP_printk("mmu notifier set pte hva: %#08lx", __entry->hva) -); - -TRACE_EVENT(kvm_age_hva, - TP_PROTO(unsigned long start, unsigned long end), - TP_ARGS(start, end), - - TP_STRUCT__entry( - __field( unsigned long, start ) - __field( unsigned long, end ) - ), - - TP_fast_assign( - __entry->start = start; - __entry->end = end; - ), - - TP_printk("mmu notifier age hva: %#08lx -- %#08lx", - __entry->start, __entry->end) -); - -TRACE_EVENT(kvm_test_age_hva, - TP_PROTO(unsigned long hva), - TP_ARGS(hva), - - TP_STRUCT__entry( - __field( unsigned long, hva ) - ), - - TP_fast_assign( - __entry->hva = hva; - ), - - TP_printk("mmu notifier test age hva: %#08lx", __entry->hva) -); - TRACE_EVENT(kvm_hvc, TP_PROTO(unsigned long vcpu_pc, unsigned long r0, unsigned long imm), TP_ARGS(vcpu_pc, r0, imm), @@ -282,45 +74,6 @@ TRACE_EVENT(kvm_hvc, __entry->vcpu_pc, __entry->r0, __entry->imm) ); -TRACE_EVENT(kvm_set_way_flush, - TP_PROTO(unsigned long vcpu_pc, bool cache), - TP_ARGS(vcpu_pc, cache), - - TP_STRUCT__entry( - __field( unsigned long, vcpu_pc ) - __field( bool, cache ) - ), - - TP_fast_assign( - __entry->vcpu_pc = vcpu_pc; - __entry->cache = cache; - ), - - TP_printk("S/W flush at 0x%016lx (cache %s)", - __entry->vcpu_pc, __entry->cache ? "on" : "off") -); - -TRACE_EVENT(kvm_toggle_cache, - TP_PROTO(unsigned long vcpu_pc, bool was, bool now), - TP_ARGS(vcpu_pc, was, now), - - TP_STRUCT__entry( - __field( unsigned long, vcpu_pc ) - __field( bool, was ) - __field( bool, now ) - ), - - TP_fast_assign( - __entry->vcpu_pc = vcpu_pc; - __entry->was = was; - __entry->now = now; - ), - - TP_printk("VM op at 0x%016lx (cache was %s, now %s)", - __entry->vcpu_pc, __entry->was ? "on" : "off", - __entry->now ? "on" : "off") -); - #endif /* _TRACE_KVM_H */ #undef TRACE_INCLUDE_PATH diff --git a/arch/arm64/include/asm/kvm_emulate.h b/arch/arm64/include/asm/kvm_emulate.h index f5ea0ba70f07..fe39e6841326 100644 --- a/arch/arm64/include/asm/kvm_emulate.h +++ b/arch/arm64/include/asm/kvm_emulate.h @@ -240,6 +240,12 @@ static inline u8 kvm_vcpu_trap_get_fault_type(const struct kvm_vcpu *vcpu) return kvm_vcpu_get_hsr(vcpu) & ESR_ELx_FSC_TYPE; } +static inline int kvm_vcpu_sys_get_rt(struct kvm_vcpu *vcpu) +{ + u32 esr = kvm_vcpu_get_hsr(vcpu); + return (esr & ESR_ELx_SYS64_ISS_RT_MASK) >> ESR_ELx_SYS64_ISS_RT_SHIFT; +} + static inline unsigned long kvm_vcpu_get_mpidr_aff(struct kvm_vcpu *vcpu) { return vcpu_sys_reg(vcpu, MPIDR_EL1) & MPIDR_HWID_BITMASK; diff --git a/arch/arm64/include/uapi/asm/kvm.h b/arch/arm64/include/uapi/asm/kvm.h index 869ee480deed..70eea2ecc663 100644 --- a/arch/arm64/include/uapi/asm/kvm.h +++ b/arch/arm64/include/uapi/asm/kvm.h @@ -216,13 +216,17 @@ struct kvm_arch_memory_slot { #define KVM_DEV_ARM_VGIC_GRP_REDIST_REGS 5 #define KVM_DEV_ARM_VGIC_GRP_CPU_SYSREGS 6 #define KVM_DEV_ARM_VGIC_GRP_LEVEL_INFO 7 +#define KVM_DEV_ARM_VGIC_GRP_ITS_REGS 8 #define KVM_DEV_ARM_VGIC_LINE_LEVEL_INFO_SHIFT 10 #define KVM_DEV_ARM_VGIC_LINE_LEVEL_INFO_MASK \ (0x3fffffULL << KVM_DEV_ARM_VGIC_LINE_LEVEL_INFO_SHIFT) #define KVM_DEV_ARM_VGIC_LINE_LEVEL_INTID_MASK 0x3ff #define VGIC_LEVEL_INFO_LINE_LEVEL 0 -#define KVM_DEV_ARM_VGIC_CTRL_INIT 0 +#define KVM_DEV_ARM_VGIC_CTRL_INIT 0 +#define KVM_DEV_ARM_ITS_SAVE_TABLES 1 +#define KVM_DEV_ARM_ITS_RESTORE_TABLES 2 +#define KVM_DEV_ARM_VGIC_SAVE_PENDING_TABLES 3 /* Device Control API on vcpu fd */ #define KVM_ARM_VCPU_PMU_V3_CTRL 0 diff --git a/arch/arm64/kvm/Makefile b/arch/arm64/kvm/Makefile index afd51bebb9c5..5d9810086c25 100644 --- a/arch/arm64/kvm/Makefile +++ b/arch/arm64/kvm/Makefile @@ -7,14 +7,13 @@ CFLAGS_arm.o := -I. CFLAGS_mmu.o := -I. KVM=../../../virt/kvm -ARM=../../../arch/arm/kvm obj-$(CONFIG_KVM_ARM_HOST) += kvm.o obj-$(CONFIG_KVM_ARM_HOST) += hyp/ kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/kvm_main.o $(KVM)/coalesced_mmio.o $(KVM)/eventfd.o $(KVM)/vfio.o -kvm-$(CONFIG_KVM_ARM_HOST) += $(ARM)/arm.o $(ARM)/mmu.o $(ARM)/mmio.o -kvm-$(CONFIG_KVM_ARM_HOST) += $(ARM)/psci.o $(ARM)/perf.o +kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/arm.o $(KVM)/arm/mmu.o $(KVM)/arm/mmio.o +kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/psci.o $(KVM)/arm/perf.o kvm-$(CONFIG_KVM_ARM_HOST) += inject_fault.o regmap.o kvm-$(CONFIG_KVM_ARM_HOST) += hyp.o hyp-init.o handle_exit.o diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c index efbe9e8e7a78..0fe27024a2e1 100644 --- a/arch/arm64/kvm/sys_regs.c +++ b/arch/arm64/kvm/sys_regs.c @@ -1529,8 +1529,8 @@ static int kvm_handle_cp_64(struct kvm_vcpu *vcpu, { struct sys_reg_params params; u32 hsr = kvm_vcpu_get_hsr(vcpu); - int Rt = (hsr >> 5) & 0xf; - int Rt2 = (hsr >> 10) & 0xf; + int Rt = kvm_vcpu_sys_get_rt(vcpu); + int Rt2 = (hsr >> 10) & 0x1f; params.is_aarch32 = true; params.is_32bit = false; @@ -1586,7 +1586,7 @@ static int kvm_handle_cp_32(struct kvm_vcpu *vcpu, { struct sys_reg_params params; u32 hsr = kvm_vcpu_get_hsr(vcpu); - int Rt = (hsr >> 5) & 0xf; + int Rt = kvm_vcpu_sys_get_rt(vcpu); params.is_aarch32 = true; params.is_32bit = true; @@ -1688,7 +1688,7 @@ int kvm_handle_sys_reg(struct kvm_vcpu *vcpu, struct kvm_run *run) { struct sys_reg_params params; unsigned long esr = kvm_vcpu_get_hsr(vcpu); - int Rt = (esr >> 5) & 0x1f; + int Rt = kvm_vcpu_sys_get_rt(vcpu); int ret; trace_kvm_handle_sys_reg(esr); diff --git a/arch/powerpc/include/asm/kvm_book3s_asm.h b/arch/powerpc/include/asm/kvm_book3s_asm.h index 0593d9479f74..b148496ffe36 100644 --- a/arch/powerpc/include/asm/kvm_book3s_asm.h +++ b/arch/powerpc/include/asm/kvm_book3s_asm.h @@ -111,6 +111,8 @@ struct kvmppc_host_state { struct kvm_vcpu *kvm_vcpu; struct kvmppc_vcore *kvm_vcore; void __iomem *xics_phys; + void __iomem *xive_tima_phys; + void __iomem *xive_tima_virt; u32 saved_xirr; u64 dabr; u64 host_mmcr[7]; /* MMCR 0,1,A, SIAR, SDAR, MMCR2, SIER */ diff --git a/arch/powerpc/include/asm/kvm_host.h b/arch/powerpc/include/asm/kvm_host.h index 77c60826d145..9c51ac4b8f36 100644 --- a/arch/powerpc/include/asm/kvm_host.h +++ b/arch/powerpc/include/asm/kvm_host.h @@ -210,6 +210,12 @@ struct kvmppc_spapr_tce_table { /* XICS components, defined in book3s_xics.c */ struct kvmppc_xics; struct kvmppc_icp; +extern struct kvm_device_ops kvm_xics_ops; + +/* XIVE components, defined in book3s_xive.c */ +struct kvmppc_xive; +struct kvmppc_xive_vcpu; +extern struct kvm_device_ops kvm_xive_ops; struct kvmppc_passthru_irqmap; @@ -298,6 +304,7 @@ struct kvm_arch { #endif #ifdef CONFIG_KVM_XICS struct kvmppc_xics *xics; + struct kvmppc_xive *xive; struct kvmppc_passthru_irqmap *pimap; #endif struct kvmppc_ops *kvm_ops; @@ -427,7 +434,7 @@ struct kvmppc_passthru_irqmap { #define KVMPPC_IRQ_DEFAULT 0 #define KVMPPC_IRQ_MPIC 1 -#define KVMPPC_IRQ_XICS 2 +#define KVMPPC_IRQ_XICS 2 /* Includes a XIVE option */ #define MMIO_HPTE_CACHE_SIZE 4 @@ -454,6 +461,21 @@ struct mmio_hpte_cache { struct openpic; +/* W0 and W1 of a XIVE thread management context */ +union xive_tma_w01 { + struct { + u8 nsr; + u8 cppr; + u8 ipb; + u8 lsmfb; + u8 ack; + u8 inc; + u8 age; + u8 pipr; + }; + __be64 w01; +}; + struct kvm_vcpu_arch { ulong host_stack; u32 host_pid; @@ -714,6 +736,10 @@ struct kvm_vcpu_arch { struct openpic *mpic; /* KVM_IRQ_MPIC */ #ifdef CONFIG_KVM_XICS struct kvmppc_icp *icp; /* XICS presentation controller */ + struct kvmppc_xive_vcpu *xive_vcpu; /* XIVE virtual CPU data */ + __be32 xive_cam_word; /* Cooked W2 in proper endian with valid bit */ + u32 xive_pushed; /* Is the VP pushed on the physical CPU ? */ + union xive_tma_w01 xive_saved_state; /* W0..1 of XIVE thread state */ #endif #ifdef CONFIG_KVM_BOOK3S_HV_POSSIBLE diff --git a/arch/powerpc/include/asm/kvm_ppc.h b/arch/powerpc/include/asm/kvm_ppc.h index 76e940a3c145..e0d88c38602b 100644 --- a/arch/powerpc/include/asm/kvm_ppc.h +++ b/arch/powerpc/include/asm/kvm_ppc.h @@ -240,6 +240,7 @@ int kvm_vcpu_ioctl_interrupt(struct kvm_vcpu *vcpu, struct kvm_interrupt *irq); extern int kvm_vm_ioctl_rtas_define_token(struct kvm *kvm, void __user *argp); extern int kvmppc_rtas_hcall(struct kvm_vcpu *vcpu); extern void kvmppc_rtas_tokens_free(struct kvm *kvm); + extern int kvmppc_xics_set_xive(struct kvm *kvm, u32 irq, u32 server, u32 priority); extern int kvmppc_xics_get_xive(struct kvm *kvm, u32 irq, u32 *server, @@ -428,6 +429,14 @@ static inline void kvmppc_set_xics_phys(int cpu, unsigned long addr) paca[cpu].kvm_hstate.xics_phys = (void __iomem *)addr; } +static inline void kvmppc_set_xive_tima(int cpu, + unsigned long phys_addr, + void __iomem *virt_addr) +{ + paca[cpu].kvm_hstate.xive_tima_phys = (void __iomem *)phys_addr; + paca[cpu].kvm_hstate.xive_tima_virt = virt_addr; +} + static inline u32 kvmppc_get_xics_latch(void) { u32 xirr; @@ -458,6 +467,11 @@ static inline void __init kvm_cma_reserve(void) static inline void kvmppc_set_xics_phys(int cpu, unsigned long addr) {} +static inline void kvmppc_set_xive_tima(int cpu, + unsigned long phys_addr, + void __iomem *virt_addr) +{} + static inline u32 kvmppc_get_xics_latch(void) { return 0; @@ -508,6 +522,10 @@ extern long kvmppc_deliver_irq_passthru(struct kvm_vcpu *vcpu, __be32 xirr, struct kvmppc_irq_map *irq_map, struct kvmppc_passthru_irqmap *pimap, bool *again); + +extern int kvmppc_xics_set_irq(struct kvm *kvm, int irq_source_id, u32 irq, + int level, bool line_status); + extern int h_ipi_redirect; #else static inline struct kvmppc_passthru_irqmap *kvmppc_get_passthru_irqmap( @@ -525,6 +543,60 @@ static inline int kvmppc_xics_hcall(struct kvm_vcpu *vcpu, u32 cmd) { return 0; } #endif +#ifdef CONFIG_KVM_XIVE +/* + * Below the first "xive" is the "eXternal Interrupt Virtualization Engine" + * ie. P9 new interrupt controller, while the second "xive" is the legacy + * "eXternal Interrupt Vector Entry" which is the configuration of an + * interrupt on the "xics" interrupt controller on P8 and earlier. Those + * two function consume or produce a legacy "XIVE" state from the + * new "XIVE" interrupt controller. + */ +extern int kvmppc_xive_set_xive(struct kvm *kvm, u32 irq, u32 server, + u32 priority); +extern int kvmppc_xive_get_xive(struct kvm *kvm, u32 irq, u32 *server, + u32 *priority); +extern int kvmppc_xive_int_on(struct kvm *kvm, u32 irq); +extern int kvmppc_xive_int_off(struct kvm *kvm, u32 irq); +extern void kvmppc_xive_init_module(void); +extern void kvmppc_xive_exit_module(void); + +extern int kvmppc_xive_connect_vcpu(struct kvm_device *dev, + struct kvm_vcpu *vcpu, u32 cpu); +extern void kvmppc_xive_cleanup_vcpu(struct kvm_vcpu *vcpu); +extern int kvmppc_xive_set_mapped(struct kvm *kvm, unsigned long guest_irq, + struct irq_desc *host_desc); +extern int kvmppc_xive_clr_mapped(struct kvm *kvm, unsigned long guest_irq, + struct irq_desc *host_desc); +extern u64 kvmppc_xive_get_icp(struct kvm_vcpu *vcpu); +extern int kvmppc_xive_set_icp(struct kvm_vcpu *vcpu, u64 icpval); + +extern int kvmppc_xive_set_irq(struct kvm *kvm, int irq_source_id, u32 irq, + int level, bool line_status); +#else +static inline int kvmppc_xive_set_xive(struct kvm *kvm, u32 irq, u32 server, + u32 priority) { return -1; } +static inline int kvmppc_xive_get_xive(struct kvm *kvm, u32 irq, u32 *server, + u32 *priority) { return -1; } +static inline int kvmppc_xive_int_on(struct kvm *kvm, u32 irq) { return -1; } +static inline int kvmppc_xive_int_off(struct kvm *kvm, u32 irq) { return -1; } +static inline void kvmppc_xive_init_module(void) { } +static inline void kvmppc_xive_exit_module(void) { } + +static inline int kvmppc_xive_connect_vcpu(struct kvm_device *dev, + struct kvm_vcpu *vcpu, u32 cpu) { return -EBUSY; } +static inline void kvmppc_xive_cleanup_vcpu(struct kvm_vcpu *vcpu) { } +static inline int kvmppc_xive_set_mapped(struct kvm *kvm, unsigned long guest_irq, + struct irq_desc *host_desc) { return -ENODEV; } +static inline int kvmppc_xive_clr_mapped(struct kvm *kvm, unsigned long guest_irq, + struct irq_desc *host_desc) { return -ENODEV; } +static inline u64 kvmppc_xive_get_icp(struct kvm_vcpu *vcpu) { return 0; } +static inline int kvmppc_xive_set_icp(struct kvm_vcpu *vcpu, u64 icpval) { return -ENOENT; } + +static inline int kvmppc_xive_set_irq(struct kvm *kvm, int irq_source_id, u32 irq, + int level, bool line_status) { return -ENODEV; } +#endif /* CONFIG_KVM_XIVE */ + /* * Prototypes for functions called only from assembler code. * Having prototypes reduces sparse errors. @@ -562,6 +634,8 @@ long kvmppc_h_clear_mod(struct kvm_vcpu *vcpu, unsigned long flags, long kvmppc_hpte_hv_fault(struct kvm_vcpu *vcpu, unsigned long addr, unsigned long slb_v, unsigned int status, bool data); unsigned long kvmppc_rm_h_xirr(struct kvm_vcpu *vcpu); +unsigned long kvmppc_rm_h_xirr_x(struct kvm_vcpu *vcpu); +unsigned long kvmppc_rm_h_ipoll(struct kvm_vcpu *vcpu, unsigned long server); int kvmppc_rm_h_ipi(struct kvm_vcpu *vcpu, unsigned long server, unsigned long mfrr); int kvmppc_rm_h_cppr(struct kvm_vcpu *vcpu, unsigned long cppr); diff --git a/arch/powerpc/include/asm/xive.h b/arch/powerpc/include/asm/xive.h index 3cdbeaeac397..c8a822acf962 100644 --- a/arch/powerpc/include/asm/xive.h +++ b/arch/powerpc/include/asm/xive.h @@ -99,7 +99,6 @@ struct xive_q { #define XIVE_ESB_SET_PQ_01 0xd00 #define XIVE_ESB_SET_PQ_10 0xe00 #define XIVE_ESB_SET_PQ_11 0xf00 -#define XIVE_ESB_MASK XIVE_ESB_SET_PQ_01 #define XIVE_ESB_VAL_P 0x2 #define XIVE_ESB_VAL_Q 0x1 @@ -136,11 +135,11 @@ extern int xive_native_configure_queue(u32 vp_id, struct xive_q *q, u8 prio, __be32 *qpage, u32 order, bool can_escalate); extern void xive_native_disable_queue(u32 vp_id, struct xive_q *q, u8 prio); -extern bool __xive_irq_trigger(struct xive_irq_data *xd); -extern bool __xive_irq_retrigger(struct xive_irq_data *xd); -extern void xive_do_source_eoi(u32 hw_irq, struct xive_irq_data *xd); - +extern void xive_native_sync_source(u32 hw_irq); extern bool is_xive_irq(struct irq_chip *chip); +extern int xive_native_enable_vp(u32 vp_id); +extern int xive_native_disable_vp(u32 vp_id); +extern int xive_native_get_vp_info(u32 vp_id, u32 *out_cam_id, u32 *out_chip_id); #else diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c index 439c257dec4a..709e23425317 100644 --- a/arch/powerpc/kernel/asm-offsets.c +++ b/arch/powerpc/kernel/asm-offsets.c @@ -634,6 +634,8 @@ int main(void) HSTATE_FIELD(HSTATE_KVM_VCPU, kvm_vcpu); HSTATE_FIELD(HSTATE_KVM_VCORE, kvm_vcore); HSTATE_FIELD(HSTATE_XICS_PHYS, xics_phys); + HSTATE_FIELD(HSTATE_XIVE_TIMA_PHYS, xive_tima_phys); + HSTATE_FIELD(HSTATE_XIVE_TIMA_VIRT, xive_tima_virt); HSTATE_FIELD(HSTATE_SAVED_XIRR, saved_xirr); HSTATE_FIELD(HSTATE_HOST_IPI, host_ipi); HSTATE_FIELD(HSTATE_PTID, ptid); @@ -719,6 +721,14 @@ int main(void) OFFSET(VCPU_HOST_MAS6, kvm_vcpu, arch.host_mas6); #endif +#ifdef CONFIG_KVM_XICS + DEFINE(VCPU_XIVE_SAVED_STATE, offsetof(struct kvm_vcpu, + arch.xive_saved_state)); + DEFINE(VCPU_XIVE_CAM_WORD, offsetof(struct kvm_vcpu, + arch.xive_cam_word)); + DEFINE(VCPU_XIVE_PUSHED, offsetof(struct kvm_vcpu, arch.xive_pushed)); +#endif + #ifdef CONFIG_KVM_EXIT_TIMING OFFSET(VCPU_TIMING_EXIT_TBU, kvm_vcpu, arch.timing_exit.tv32.tbu); OFFSET(VCPU_TIMING_EXIT_TBL, kvm_vcpu, arch.timing_exit.tv32.tbl); diff --git a/arch/powerpc/kvm/Kconfig b/arch/powerpc/kvm/Kconfig index 65a471de96de..24de532c1736 100644 --- a/arch/powerpc/kvm/Kconfig +++ b/arch/powerpc/kvm/Kconfig @@ -197,6 +197,11 @@ config KVM_XICS Specification) interrupt controller architecture used on IBM POWER (pSeries) servers. +config KVM_XIVE + bool + default y + depends on KVM_XICS && PPC_XIVE_NATIVE && KVM_BOOK3S_HV_POSSIBLE + source drivers/vhost/Kconfig endif # VIRTUALIZATION diff --git a/arch/powerpc/kvm/Makefile b/arch/powerpc/kvm/Makefile index b87ccde2137a..d91a2604c496 100644 --- a/arch/powerpc/kvm/Makefile +++ b/arch/powerpc/kvm/Makefile @@ -74,7 +74,7 @@ kvm-hv-y += \ book3s_64_mmu_radix.o kvm-book3s_64-builtin-xics-objs-$(CONFIG_KVM_XICS) := \ - book3s_hv_rm_xics.o + book3s_hv_rm_xics.o book3s_hv_rm_xive.o ifdef CONFIG_KVM_BOOK3S_HV_POSSIBLE kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HANDLER) += \ @@ -89,6 +89,8 @@ endif kvm-book3s_64-objs-$(CONFIG_KVM_XICS) += \ book3s_xics.o +kvm-book3s_64-objs-$(CONFIG_KVM_XIVE) += book3s_xive.o + kvm-book3s_64-module-objs := \ $(common-objs-y) \ book3s.o \ diff --git a/arch/powerpc/kvm/book3s.c b/arch/powerpc/kvm/book3s.c index 8c4d7e9d27d2..72d977e30952 100644 --- a/arch/powerpc/kvm/book3s.c +++ b/arch/powerpc/kvm/book3s.c @@ -35,6 +35,7 @@ #include <asm/kvm_book3s.h> #include <asm/mmu_context.h> #include <asm/page.h> +#include <asm/xive.h> #include "book3s.h" #include "trace.h" @@ -596,11 +597,14 @@ int kvmppc_get_one_reg(struct kvm_vcpu *vcpu, u64 id, break; #ifdef CONFIG_KVM_XICS case KVM_REG_PPC_ICP_STATE: - if (!vcpu->arch.icp) { + if (!vcpu->arch.icp && !vcpu->arch.xive_vcpu) { r = -ENXIO; break; } - *val = get_reg_val(id, kvmppc_xics_get_icp(vcpu)); + if (xive_enabled()) + *val = get_reg_val(id, kvmppc_xive_get_icp(vcpu)); + else + *val = get_reg_val(id, kvmppc_xics_get_icp(vcpu)); break; #endif /* CONFIG_KVM_XICS */ case KVM_REG_PPC_FSCR: @@ -666,12 +670,14 @@ int kvmppc_set_one_reg(struct kvm_vcpu *vcpu, u64 id, #endif /* CONFIG_VSX */ #ifdef CONFIG_KVM_XICS case KVM_REG_PPC_ICP_STATE: - if (!vcpu->arch.icp) { + if (!vcpu->arch.icp && !vcpu->arch.xive_vcpu) { r = -ENXIO; break; } - r = kvmppc_xics_set_icp(vcpu, - set_reg_val(id, *val)); + if (xive_enabled()) + r = kvmppc_xive_set_icp(vcpu, set_reg_val(id, *val)); + else + r = kvmppc_xics_set_icp(vcpu, set_reg_val(id, *val)); break; #endif /* CONFIG_KVM_XICS */ case KVM_REG_PPC_FSCR: @@ -942,6 +948,50 @@ int kvmppc_book3s_hcall_implemented(struct kvm *kvm, unsigned long hcall) return kvm->arch.kvm_ops->hcall_implemented(hcall); } +#ifdef CONFIG_KVM_XICS +int kvm_set_irq(struct kvm *kvm, int irq_source_id, u32 irq, int level, + bool line_status) +{ + if (xive_enabled()) + return kvmppc_xive_set_irq(kvm, irq_source_id, irq, level, + line_status); + else + return kvmppc_xics_set_irq(kvm, irq_source_id, irq, level, + line_status); +} + +int kvm_arch_set_irq_inatomic(struct kvm_kernel_irq_routing_entry *irq_entry, + struct kvm *kvm, int irq_source_id, + int level, bool line_status) +{ + return kvm_set_irq(kvm, irq_source_id, irq_entry->gsi, + level, line_status); +} +static int kvmppc_book3s_set_irq(struct kvm_kernel_irq_routing_entry *e, + struct kvm *kvm, int irq_source_id, int level, + bool line_status) +{ + return kvm_set_irq(kvm, irq_source_id, e->gsi, level, line_status); +} + +int kvm_irq_map_gsi(struct kvm *kvm, + struct kvm_kernel_irq_routing_entry *entries, int gsi) +{ + entries->gsi = gsi; + entries->type = KVM_IRQ_ROUTING_IRQCHIP; + entries->set = kvmppc_book3s_set_irq; + entries->irqchip.irqchip = 0; + entries->irqchip.pin = gsi; + return 1; +} + +int kvm_irq_map_chip_pin(struct kvm *kvm, unsigned irqchip, unsigned pin) +{ + return pin; +} + +#endif /* CONFIG_KVM_XICS */ + static int kvmppc_book3s_init(void) { int r; @@ -952,12 +1002,25 @@ static int kvmppc_book3s_init(void) #ifdef CONFIG_KVM_BOOK3S_32_HANDLER r = kvmppc_book3s_init_pr(); #endif - return r; +#ifdef CONFIG_KVM_XICS +#ifdef CONFIG_KVM_XIVE + if (xive_enabled()) { + kvmppc_xive_init_module(); + kvm_register_device_ops(&kvm_xive_ops, KVM_DEV_TYPE_XICS); + } else +#endif + kvm_register_device_ops(&kvm_xics_ops, KVM_DEV_TYPE_XICS); +#endif + return r; } static void kvmppc_book3s_exit(void) { +#ifdef CONFIG_KVM_XICS + if (xive_enabled()) + kvmppc_xive_exit_module(); +#endif #ifdef CONFIG_KVM_BOOK3S_32_HANDLER kvmppc_book3s_exit_pr(); #endif diff --git a/arch/powerpc/kvm/book3s_hv.c b/arch/powerpc/kvm/book3s_hv.c index 549dd6070dee..42b7a4fd57d9 100644 --- a/arch/powerpc/kvm/book3s_hv.c +++ b/arch/powerpc/kvm/book3s_hv.c @@ -67,6 +67,7 @@ #include <asm/mmu.h> #include <asm/opal.h> #include <asm/xics.h> +#include <asm/xive.h> #include "book3s.h" @@ -837,6 +838,10 @@ int kvmppc_pseries_do_hcall(struct kvm_vcpu *vcpu) case H_IPOLL: case H_XIRR_X: if (kvmppc_xics_enabled(vcpu)) { + if (xive_enabled()) { + ret = H_NOT_AVAILABLE; + return RESUME_GUEST; + } ret = kvmppc_xics_hcall(vcpu, req); break; } @@ -2947,8 +2952,12 @@ static int kvmppc_vcpu_run_hv(struct kvm_run *run, struct kvm_vcpu *vcpu) r = kvmppc_book3s_hv_page_fault(run, vcpu, vcpu->arch.fault_dar, vcpu->arch.fault_dsisr); srcu_read_unlock(&vcpu->kvm->srcu, srcu_idx); - } else if (r == RESUME_PASSTHROUGH) - r = kvmppc_xics_rm_complete(vcpu, 0); + } else if (r == RESUME_PASSTHROUGH) { + if (WARN_ON(xive_enabled())) + r = H_SUCCESS; + else + r = kvmppc_xics_rm_complete(vcpu, 0); + } } while (is_kvmppc_resume_guest(r)); out: @@ -3400,10 +3409,20 @@ static int kvmppc_core_init_vm_hv(struct kvm *kvm) /* * On POWER9, VPM0 bit is reserved (VPM0=1 behaviour is assumed) * Set HVICE bit to enable hypervisor virtualization interrupts. + * Set HEIC to prevent OS interrupts to go to hypervisor (should + * be unnecessary but better safe than sorry in case we re-enable + * EE in HV mode with this LPCR still set) */ if (cpu_has_feature(CPU_FTR_ARCH_300)) { lpcr &= ~LPCR_VPM0; - lpcr |= LPCR_HVICE; + lpcr |= LPCR_HVICE | LPCR_HEIC; + + /* + * If xive is enabled, we route 0x500 interrupts directly + * to the guest. + */ + if (xive_enabled()) + lpcr |= LPCR_LPES; } /* @@ -3533,7 +3552,7 @@ static int kvmppc_set_passthru_irq(struct kvm *kvm, int host_irq, int guest_gsi) struct kvmppc_irq_map *irq_map; struct kvmppc_passthru_irqmap *pimap; struct irq_chip *chip; - int i; + int i, rc = 0; if (!kvm_irq_bypass) return 1; @@ -3558,10 +3577,10 @@ static int kvmppc_set_passthru_irq(struct kvm *kvm, int host_irq, int guest_gsi) /* * For now, we only support interrupts for which the EOI operation * is an OPAL call followed by a write to XIRR, since that's - * what our real-mode EOI code does. + * what our real-mode EOI code does, or a XIVE interrupt */ chip = irq_data_get_irq_chip(&desc->irq_data); - if (!chip || !is_pnv_opal_msi(chip)) { + if (!chip || !(is_pnv_opal_msi(chip) || is_xive_irq(chip))) { pr_warn("kvmppc_set_passthru_irq_hv: Could not assign IRQ map for (%d,%d)\n", host_irq, guest_gsi); mutex_unlock(&kvm->lock); @@ -3603,7 +3622,12 @@ static int kvmppc_set_passthru_irq(struct kvm *kvm, int host_irq, int guest_gsi) if (i == pimap->n_mapped) pimap->n_mapped++; - kvmppc_xics_set_mapped(kvm, guest_gsi, desc->irq_data.hwirq); + if (xive_enabled()) + rc = kvmppc_xive_set_mapped(kvm, guest_gsi, desc); + else + kvmppc_xics_set_mapped(kvm, guest_gsi, desc->irq_data.hwirq); + if (rc) + irq_map->r_hwirq = 0; mutex_unlock(&kvm->lock); @@ -3614,7 +3638,7 @@ static int kvmppc_clr_passthru_irq(struct kvm *kvm, int host_irq, int guest_gsi) { struct irq_desc *desc; struct kvmppc_passthru_irqmap *pimap; - int i; + int i, rc = 0; if (!kvm_irq_bypass) return 0; @@ -3639,9 +3663,12 @@ static int kvmppc_clr_passthru_irq(struct kvm *kvm, int host_irq, int guest_gsi) return -ENODEV; } - kvmppc_xics_clr_mapped(kvm, guest_gsi, pimap->mapped[i].r_hwirq); + if (xive_enabled()) + rc = kvmppc_xive_clr_mapped(kvm, guest_gsi, pimap->mapped[i].desc); + else + kvmppc_xics_clr_mapped(kvm, guest_gsi, pimap->mapped[i].r_hwirq); - /* invalidate the entry */ + /* invalidate the entry (what do do on error from the above ?) */ pimap->mapped[i].r_hwirq = 0; /* @@ -3650,7 +3677,7 @@ static int kvmppc_clr_passthru_irq(struct kvm *kvm, int host_irq, int guest_gsi) */ unlock: mutex_unlock(&kvm->lock); - return 0; + return rc; } static int kvmppc_irq_bypass_add_producer_hv(struct irq_bypass_consumer *cons, @@ -3928,7 +3955,7 @@ static int kvmppc_book3s_init_hv(void) * indirectly, via OPAL. */ #ifdef CONFIG_SMP - if (!get_paca()->kvm_hstate.xics_phys) { + if (!xive_enabled() && !local_paca->kvm_hstate.xics_phys) { struct device_node *np; np = of_find_compatible_node(NULL, NULL, "ibm,opal-intc"); diff --git a/arch/powerpc/kvm/book3s_hv_builtin.c b/arch/powerpc/kvm/book3s_hv_builtin.c index 9c71c72e65ce..88a65923c649 100644 --- a/arch/powerpc/kvm/book3s_hv_builtin.c +++ b/arch/powerpc/kvm/book3s_hv_builtin.c @@ -32,6 +32,24 @@ #define KVM_CMA_CHUNK_ORDER 18 +#include "book3s_xics.h" +#include "book3s_xive.h" + +/* + * The XIVE module will populate these when it loads + */ +unsigned long (*__xive_vm_h_xirr)(struct kvm_vcpu *vcpu); +unsigned long (*__xive_vm_h_ipoll)(struct kvm_vcpu *vcpu, unsigned long server); +int (*__xive_vm_h_ipi)(struct kvm_vcpu *vcpu, unsigned long server, + unsigned long mfrr); +int (*__xive_vm_h_cppr)(struct kvm_vcpu *vcpu, unsigned long cppr); +int (*__xive_vm_h_eoi)(struct kvm_vcpu *vcpu, unsigned long xirr); +EXPORT_SYMBOL_GPL(__xive_vm_h_xirr); +EXPORT_SYMBOL_GPL(__xive_vm_h_ipoll); +EXPORT_SYMBOL_GPL(__xive_vm_h_ipi); +EXPORT_SYMBOL_GPL(__xive_vm_h_cppr); +EXPORT_SYMBOL_GPL(__xive_vm_h_eoi); + /* * Hash page table alignment on newer cpus(CPU_FTR_ARCH_206) * should be power of 2. @@ -211,6 +229,7 @@ void kvmhv_rm_send_ipi(int cpu) __asm__ __volatile__ (PPC_MSGSND(%0) : : "r" (msg)); return; } + /* On POWER8 for IPIs to threads in the same core, use msgsnd. */ if (cpu_has_feature(CPU_FTR_ARCH_207S) && cpu_first_thread_sibling(cpu) == @@ -407,6 +426,9 @@ static long kvmppc_read_one_intr(bool *again) u8 host_ipi; int64_t rc; + if (xive_enabled()) + return 1; + /* see if a host IPI is pending */ host_ipi = local_paca->kvm_hstate.host_ipi; if (host_ipi) @@ -491,3 +513,84 @@ static long kvmppc_read_one_intr(bool *again) return kvmppc_check_passthru(xisr, xirr, again); } + +#ifdef CONFIG_KVM_XICS +static inline bool is_rm(void) +{ + return !(mfmsr() & MSR_DR); +} + +unsigned long kvmppc_rm_h_xirr(struct kvm_vcpu *vcpu) +{ + if (xive_enabled()) { + if (is_rm()) + return xive_rm_h_xirr(vcpu); + if (unlikely(!__xive_vm_h_xirr)) + return H_NOT_AVAILABLE; + return __xive_vm_h_xirr(vcpu); + } else + return xics_rm_h_xirr(vcpu); +} + +unsigned long kvmppc_rm_h_xirr_x(struct kvm_vcpu *vcpu) +{ + vcpu->arch.gpr[5] = get_tb(); + if (xive_enabled()) { + if (is_rm()) + return xive_rm_h_xirr(vcpu); + if (unlikely(!__xive_vm_h_xirr)) + return H_NOT_AVAILABLE; + return __xive_vm_h_xirr(vcpu); + } else + return xics_rm_h_xirr(vcpu); +} + +unsigned long kvmppc_rm_h_ipoll(struct kvm_vcpu *vcpu, unsigned long server) +{ + if (xive_enabled()) { + if (is_rm()) + return xive_rm_h_ipoll(vcpu, server); + if (unlikely(!__xive_vm_h_ipoll)) + return H_NOT_AVAILABLE; + return __xive_vm_h_ipoll(vcpu, server); + } else + return H_TOO_HARD; +} + +int kvmppc_rm_h_ipi(struct kvm_vcpu *vcpu, unsigned long server, + unsigned long mfrr) +{ + if (xive_enabled()) { + if (is_rm()) + return xive_rm_h_ipi(vcpu, server, mfrr); + if (unlikely(!__xive_vm_h_ipi)) + return H_NOT_AVAILABLE; + return __xive_vm_h_ipi(vcpu, server, mfrr); + } else + return xics_rm_h_ipi(vcpu, server, mfrr); +} + +int kvmppc_rm_h_cppr(struct kvm_vcpu *vcpu, unsigned long cppr) +{ + if (xive_enabled()) { + if (is_rm()) + return xive_rm_h_cppr(vcpu, cppr); + if (unlikely(!__xive_vm_h_cppr)) + return H_NOT_AVAILABLE; + return __xive_vm_h_cppr(vcpu, cppr); + } else + return xics_rm_h_cppr(vcpu, cppr); +} + +int kvmppc_rm_h_eoi(struct kvm_vcpu *vcpu, unsigned long xirr) +{ + if (xive_enabled()) { + if (is_rm()) + return xive_rm_h_eoi(vcpu, xirr); + if (unlikely(!__xive_vm_h_eoi)) + return H_NOT_AVAILABLE; + return __xive_vm_h_eoi(vcpu, xirr); + } else + return xics_rm_h_eoi(vcpu, xirr); +} +#endif /* CONFIG_KVM_XICS */ diff --git a/arch/powerpc/kvm/book3s_hv_rm_xics.c b/arch/powerpc/kvm/book3s_hv_rm_xics.c index ffde4507ddfd..2a862618f072 100644 --- a/arch/powerpc/kvm/book3s_hv_rm_xics.c +++ b/arch/powerpc/kvm/book3s_hv_rm_xics.c @@ -484,7 +484,7 @@ static void icp_rm_down_cppr(struct kvmppc_xics *xics, struct kvmppc_icp *icp, } -unsigned long kvmppc_rm_h_xirr(struct kvm_vcpu *vcpu) +unsigned long xics_rm_h_xirr(struct kvm_vcpu *vcpu) { union kvmppc_icp_state old_state, new_state; struct kvmppc_xics *xics = vcpu->kvm->arch.xics; @@ -522,8 +522,8 @@ unsigned long kvmppc_rm_h_xirr(struct kvm_vcpu *vcpu) return check_too_hard(xics, icp); } -int kvmppc_rm_h_ipi(struct kvm_vcpu *vcpu, unsigned long server, - unsigned long mfrr) +int xics_rm_h_ipi(struct kvm_vcpu *vcpu, unsigned long server, + unsigned long mfrr) { union kvmppc_icp_state old_state, new_state; struct kvmppc_xics *xics = vcpu->kvm->arch.xics; @@ -609,7 +609,7 @@ int kvmppc_rm_h_ipi(struct kvm_vcpu *vcpu, unsigned long server, return check_too_hard(xics, this_icp); } -int kvmppc_rm_h_cppr(struct kvm_vcpu *vcpu, unsigned long cppr) +int xics_rm_h_cppr(struct kvm_vcpu *vcpu, unsigned long cppr) { union kvmppc_icp_state old_state, new_state; struct kvmppc_xics *xics = vcpu->kvm->arch.xics; @@ -729,7 +729,7 @@ static int ics_rm_eoi(struct kvm_vcpu *vcpu, u32 irq) return check_too_hard(xics, icp); } -int kvmppc_rm_h_eoi(struct kvm_vcpu *vcpu, unsigned long xirr) +int xics_rm_h_eoi(struct kvm_vcpu *vcpu, unsigned long xirr) { struct kvmppc_xics *xics = vcpu->kvm->arch.xics; struct kvmppc_icp *icp = vcpu->arch.icp; diff --git a/arch/powerpc/kvm/book3s_hv_rm_xive.c b/arch/powerpc/kvm/book3s_hv_rm_xive.c new file mode 100644 index 000000000000..abf5f01b6eb1 --- /dev/null +++ b/arch/powerpc/kvm/book3s_hv_rm_xive.c @@ -0,0 +1,47 @@ +#include <linux/kernel.h> +#include <linux/kvm_host.h> +#include <linux/err.h> +#include <linux/kernel_stat.h> + +#include <asm/kvm_book3s.h> +#include <asm/kvm_ppc.h> +#include <asm/hvcall.h> +#include <asm/xics.h> +#include <asm/debug.h> +#include <asm/synch.h> +#include <asm/cputhreads.h> +#include <asm/pgtable.h> +#include <asm/ppc-opcode.h> +#include <asm/pnv-pci.h> +#include <asm/opal.h> +#include <asm/smp.h> +#include <asm/asm-prototypes.h> +#include <asm/xive.h> +#include <asm/xive-regs.h> + +#include "book3s_xive.h" + +/* XXX */ +#include <asm/udbg.h> +//#define DBG(fmt...) udbg_printf(fmt) +#define DBG(fmt...) do { } while(0) + +static inline void __iomem *get_tima_phys(void) +{ + return local_paca->kvm_hstate.xive_tima_phys; +} + +#undef XIVE_RUNTIME_CHECKS +#define X_PFX xive_rm_ +#define X_STATIC +#define X_STAT_PFX stat_rm_ +#define __x_tima get_tima_phys() +#define __x_eoi_page(xd) ((void __iomem *)((xd)->eoi_page)) +#define __x_trig_page(xd) ((void __iomem *)((xd)->trig_page)) +#define __x_readb __raw_rm_readb +#define __x_writeb __raw_rm_writeb +#define __x_readw __raw_rm_readw +#define __x_readq __raw_rm_readq +#define __x_writeq __raw_rm_writeq + +#include "book3s_xive_template.c" diff --git a/arch/powerpc/kvm/book3s_hv_rmhandlers.S b/arch/powerpc/kvm/book3s_hv_rmhandlers.S index 7c6477d1840a..bdb3f76ceb6b 100644 --- a/arch/powerpc/kvm/book3s_hv_rmhandlers.S +++ b/arch/powerpc/kvm/book3s_hv_rmhandlers.S @@ -30,6 +30,7 @@ #include <asm/book3s/64/mmu-hash.h> #include <asm/tm.h> #include <asm/opal.h> +#include <asm/xive-regs.h> #define VCPU_GPRS_TM(reg) (((reg) * ULONG_SIZE) + VCPU_GPR_TM) @@ -970,6 +971,23 @@ ALT_FTR_SECTION_END_IFCLR(CPU_FTR_ARCH_300) cmpwi r3, 512 /* 1 microsecond */ blt hdec_soon +#ifdef CONFIG_KVM_XICS + /* We are entering the guest on that thread, push VCPU to XIVE */ + ld r10, HSTATE_XIVE_TIMA_PHYS(r13) + cmpldi cr0, r10, r0 + beq no_xive + ld r11, VCPU_XIVE_SAVED_STATE(r4) + li r9, TM_QW1_OS + stdcix r11,r9,r10 + eieio + lwz r11, VCPU_XIVE_CAM_WORD(r4) + li r9, TM_QW1_OS + TM_WORD2 + stwcix r11,r9,r10 + li r9, 1 + stw r9, VCPU_XIVE_PUSHED(r4) +no_xive: +#endif /* CONFIG_KVM_XICS */ + deliver_guest_interrupt: ld r6, VCPU_CTR(r4) ld r7, VCPU_XER(r4) @@ -1307,6 +1325,42 @@ END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR) blt deliver_guest_interrupt guest_exit_cont: /* r9 = vcpu, r12 = trap, r13 = paca */ +#ifdef CONFIG_KVM_XICS + /* We are exiting, pull the VP from the XIVE */ + lwz r0, VCPU_XIVE_PUSHED(r9) + cmpwi cr0, r0, 0 + beq 1f + li r7, TM_SPC_PULL_OS_CTX + li r6, TM_QW1_OS + mfmsr r0 + andi. r0, r0, MSR_IR /* in real mode? */ + beq 2f + ld r10, HSTATE_XIVE_TIMA_VIRT(r13) + cmpldi cr0, r10, 0 + beq 1f + /* First load to pull the context, we ignore the value */ + lwzx r11, r7, r10 + eieio + /* Second load to recover the context state (Words 0 and 1) */ + ldx r11, r6, r10 + b 3f +2: ld r10, HSTATE_XIVE_TIMA_PHYS(r13) + cmpldi cr0, r10, 0 + beq 1f + /* First load to pull the context, we ignore the value */ + lwzcix r11, r7, r10 + eieio + /* Second load to recover the context state (Words 0 and 1) */ + ldcix r11, r6, r10 +3: std r11, VCPU_XIVE_SAVED_STATE(r9) + /* Fixup some of the state for the next load */ + li r10, 0 + li r0, 0xff + stw r10, VCPU_XIVE_PUSHED(r9) + stb r10, (VCPU_XIVE_SAVED_STATE+3)(r9) + stb r0, (VCPU_XIVE_SAVED_STATE+4)(r9) +1: +#endif /* CONFIG_KVM_XICS */ /* Save more register state */ mfdar r6 mfdsisr r7 @@ -2011,7 +2065,7 @@ hcall_real_table: .long DOTSYM(kvmppc_rm_h_eoi) - hcall_real_table .long DOTSYM(kvmppc_rm_h_cppr) - hcall_real_table .long DOTSYM(kvmppc_rm_h_ipi) - hcall_real_table - .long 0 /* 0x70 - H_IPOLL */ + .long DOTSYM(kvmppc_rm_h_ipoll) - hcall_real_table .long DOTSYM(kvmppc_rm_h_xirr) - hcall_real_table #else .long 0 /* 0x64 - H_EOI */ @@ -2181,7 +2235,11 @@ hcall_real_table: .long 0 /* 0x2f0 */ .long 0 /* 0x2f4 */ .long 0 /* 0x2f8 */ - .long 0 /* 0x2fc */ +#ifdef CONFIG_KVM_XICS + .long DOTSYM(kvmppc_rm_h_xirr_x) - hcall_real_table +#else + .long 0 /* 0x2fc - H_XIRR_X*/ +#endif .long DOTSYM(kvmppc_h_random) - hcall_real_table .globl hcall_real_table_end hcall_real_table_end: diff --git a/arch/powerpc/kvm/book3s_rtas.c b/arch/powerpc/kvm/book3s_rtas.c index 20528701835b..2d3b2b1cc272 100644 --- a/arch/powerpc/kvm/book3s_rtas.c +++ b/arch/powerpc/kvm/book3s_rtas.c @@ -16,6 +16,7 @@ #include <asm/kvm_ppc.h> #include <asm/hvcall.h> #include <asm/rtas.h> +#include <asm/xive.h> #ifdef CONFIG_KVM_XICS static void kvm_rtas_set_xive(struct kvm_vcpu *vcpu, struct rtas_args *args) @@ -32,7 +33,10 @@ static void kvm_rtas_set_xive(struct kvm_vcpu *vcpu, struct rtas_args *args) server = be32_to_cpu(args->args[1]); priority = be32_to_cpu(args->args[2]); - rc = kvmppc_xics_set_xive(vcpu->kvm, irq, server, priority); + if (xive_enabled()) + rc = kvmppc_xive_set_xive(vcpu->kvm, irq, server, priority); + else + rc = kvmppc_xics_set_xive(vcpu->kvm, irq, server, priority); if (rc) rc = -3; out: @@ -52,7 +56,10 @@ static void kvm_rtas_get_xive(struct kvm_vcpu *vcpu, struct rtas_args *args) irq = be32_to_cpu(args->args[0]); server = priority = 0; - rc = kvmppc_xics_get_xive(vcpu->kvm, irq, &server, &priority); + if (xive_enabled()) + rc = kvmppc_xive_get_xive(vcpu->kvm, irq, &server, &priority); + else + rc = kvmppc_xics_get_xive(vcpu->kvm, irq, &server, &priority); if (rc) { rc = -3; goto out; @@ -76,7 +83,10 @@ static void kvm_rtas_int_off(struct kvm_vcpu *vcpu, struct rtas_args *args) irq = be32_to_cpu(args->args[0]); - rc = kvmppc_xics_int_off(vcpu->kvm, irq); + if (xive_enabled()) + rc = kvmppc_xive_int_off(vcpu->kvm, irq); + else + rc = kvmppc_xics_int_off(vcpu->kvm, irq); if (rc) rc = -3; out: @@ -95,7 +105,10 @@ static void kvm_rtas_int_on(struct kvm_vcpu *vcpu, struct rtas_args *args) irq = be32_to_cpu(args->args[0]); - rc = kvmppc_xics_int_on(vcpu->kvm, irq); + if (xive_enabled()) + rc = kvmppc_xive_int_on(vcpu->kvm, irq); + else + rc = kvmppc_xics_int_on(vcpu->kvm, irq); if (rc) rc = -3; out: diff --git a/arch/powerpc/kvm/book3s_xics.c b/arch/powerpc/kvm/book3s_xics.c index 459b72cb617a..d329b2add7e2 100644 --- a/arch/powerpc/kvm/book3s_xics.c +++ b/arch/powerpc/kvm/book3s_xics.c @@ -1306,8 +1306,8 @@ static int xics_set_source(struct kvmppc_xics *xics, long irq, u64 addr) return 0; } -int kvm_set_irq(struct kvm *kvm, int irq_source_id, u32 irq, int level, - bool line_status) +int kvmppc_xics_set_irq(struct kvm *kvm, int irq_source_id, u32 irq, int level, + bool line_status) { struct kvmppc_xics *xics = kvm->arch.xics; @@ -1316,14 +1316,6 @@ int kvm_set_irq(struct kvm *kvm, int irq_source_id, u32 irq, int level, return ics_deliver_irq(xics, irq, level); } -int kvm_arch_set_irq_inatomic(struct kvm_kernel_irq_routing_entry *irq_entry, - struct kvm *kvm, int irq_source_id, - int level, bool line_status) -{ - return kvm_set_irq(kvm, irq_source_id, irq_entry->gsi, - level, line_status); -} - static int xics_set_attr(struct kvm_device *dev, struct kvm_device_attr *attr) { struct kvmppc_xics *xics = dev->private; @@ -1457,29 +1449,6 @@ void kvmppc_xics_free_icp(struct kvm_vcpu *vcpu) vcpu->arch.irq_type = KVMPPC_IRQ_DEFAULT; } -static int xics_set_irq(struct kvm_kernel_irq_routing_entry *e, - struct kvm *kvm, int irq_source_id, int level, - bool line_status) -{ - return kvm_set_irq(kvm, irq_source_id, e->gsi, level, line_status); -} - -int kvm_irq_map_gsi(struct kvm *kvm, - struct kvm_kernel_irq_routing_entry *entries, int gsi) -{ - entries->gsi = gsi; - entries->type = KVM_IRQ_ROUTING_IRQCHIP; - entries->set = xics_set_irq; - entries->irqchip.irqchip = 0; - entries->irqchip.pin = gsi; - return 1; -} - -int kvm_irq_map_chip_pin(struct kvm *kvm, unsigned irqchip, unsigned pin) -{ - return pin; -} - void kvmppc_xics_set_mapped(struct kvm *kvm, unsigned long irq, unsigned long host_irq) { diff --git a/arch/powerpc/kvm/book3s_xics.h b/arch/powerpc/kvm/book3s_xics.h index ec5474cf70c6..453c9e518c19 100644 --- a/arch/powerpc/kvm/book3s_xics.h +++ b/arch/powerpc/kvm/book3s_xics.h @@ -10,6 +10,7 @@ #ifndef _KVM_PPC_BOOK3S_XICS_H #define _KVM_PPC_BOOK3S_XICS_H +#ifdef CONFIG_KVM_XICS /* * We use a two-level tree to store interrupt source information. * There are up to 1024 ICS nodes, each of which can represent @@ -144,5 +145,11 @@ static inline struct kvmppc_ics *kvmppc_xics_find_ics(struct kvmppc_xics *xics, return ics; } +extern unsigned long xics_rm_h_xirr(struct kvm_vcpu *vcpu); +extern int xics_rm_h_ipi(struct kvm_vcpu *vcpu, unsigned long server, + unsigned long mfrr); +extern int xics_rm_h_cppr(struct kvm_vcpu *vcpu, unsigned long cppr); +extern int xics_rm_h_eoi(struct kvm_vcpu *vcpu, unsigned long xirr); +#endif /* CONFIG_KVM_XICS */ #endif /* _KVM_PPC_BOOK3S_XICS_H */ diff --git a/arch/powerpc/kvm/book3s_xive.c b/arch/powerpc/kvm/book3s_xive.c new file mode 100644 index 000000000000..ffe1da95033a --- /dev/null +++ b/arch/powerpc/kvm/book3s_xive.c @@ -0,0 +1,1894 @@ +/* + * Copyright 2017 Benjamin Herrenschmidt, IBM Corporation. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License, version 2, as + * published by the Free Software Foundation. + */ + +#define pr_fmt(fmt) "xive-kvm: " fmt + +#include <linux/kernel.h> +#include <linux/kvm_host.h> +#include <linux/err.h> +#include <linux/gfp.h> +#include <linux/spinlock.h> +#include <linux/delay.h> +#include <linux/percpu.h> +#include <linux/cpumask.h> +#include <asm/uaccess.h> +#include <asm/kvm_book3s.h> +#include <asm/kvm_ppc.h> +#include <asm/hvcall.h> +#include <asm/xics.h> +#include <asm/xive.h> +#include <asm/xive-regs.h> +#include <asm/debug.h> +#include <asm/debugfs.h> +#include <asm/time.h> +#include <asm/opal.h> + +#include <linux/debugfs.h> +#include <linux/seq_file.h> + +#include "book3s_xive.h" + + +/* + * Virtual mode variants of the hcalls for use on radix/radix + * with AIL. They require the VCPU's VP to be "pushed" + * + * We still instanciate them here because we use some of the + * generated utility functions as well in this file. + */ +#define XIVE_RUNTIME_CHECKS +#define X_PFX xive_vm_ +#define X_STATIC static +#define X_STAT_PFX stat_vm_ +#define __x_tima xive_tima +#define __x_eoi_page(xd) ((void __iomem *)((xd)->eoi_mmio)) +#define __x_trig_page(xd) ((void __iomem *)((xd)->trig_mmio)) +#define __x_readb __raw_readb +#define __x_writeb __raw_writeb +#define __x_readw __raw_readw +#define __x_readq __raw_readq +#define __x_writeq __raw_writeq + +#include "book3s_xive_template.c" + +/* + * We leave a gap of a couple of interrupts in the queue to + * account for the IPI and additional safety guard. + */ +#define XIVE_Q_GAP 2 + +/* + * This is a simple trigger for a generic XIVE IRQ. This must + * only be called for interrupts that support a trigger page + */ +static bool xive_irq_trigger(struct xive_irq_data *xd) +{ + /* This should be only for MSIs */ + if (WARN_ON(xd->flags & XIVE_IRQ_FLAG_LSI)) + return false; + + /* Those interrupts should always have a trigger page */ + if (WARN_ON(!xd->trig_mmio)) + return false; + + out_be64(xd->trig_mmio, 0); + + return true; +} + +static irqreturn_t xive_esc_irq(int irq, void *data) +{ + struct kvm_vcpu *vcpu = data; + + /* We use the existing H_PROD mechanism to wake up the target */ + vcpu->arch.prodded = 1; + smp_mb(); + if (vcpu->arch.ceded) + kvmppc_fast_vcpu_kick(vcpu); + + return IRQ_HANDLED; +} + +static int xive_attach_escalation(struct kvm_vcpu *vcpu, u8 prio) +{ + struct kvmppc_xive_vcpu *xc = vcpu->arch.xive_vcpu; + struct xive_q *q = &xc->queues[prio]; + char *name = NULL; + int rc; + + /* Already there ? */ + if (xc->esc_virq[prio]) + return 0; + + /* Hook up the escalation interrupt */ + xc->esc_virq[prio] = irq_create_mapping(NULL, q->esc_irq); + if (!xc->esc_virq[prio]) { + pr_err("Failed to map escalation interrupt for queue %d of VCPU %d\n", + prio, xc->server_num); + return -EIO; + } + + /* + * Future improvement: start with them disabled + * and handle DD2 and later scheme of merged escalation + * interrupts + */ + name = kasprintf(GFP_KERNEL, "kvm-%d-%d-%d", + vcpu->kvm->arch.lpid, xc->server_num, prio); + if (!name) { + pr_err("Failed to allocate escalation irq name for queue %d of VCPU %d\n", + prio, xc->server_num); + rc = -ENOMEM; + goto error; + } + rc = request_irq(xc->esc_virq[prio], xive_esc_irq, + IRQF_NO_THREAD, name, vcpu); + if (rc) { + pr_err("Failed to request escalation interrupt for queue %d of VCPU %d\n", + prio, xc->server_num); + goto error; + } + xc->esc_virq_names[prio] = name; + return 0; +error: + irq_dispose_mapping(xc->esc_virq[prio]); + xc->esc_virq[prio] = 0; + kfree(name); + return rc; +} + +static int xive_provision_queue(struct kvm_vcpu *vcpu, u8 prio) +{ + struct kvmppc_xive_vcpu *xc = vcpu->arch.xive_vcpu; + struct kvmppc_xive *xive = xc->xive; + struct xive_q *q = &xc->queues[prio]; + void *qpage; + int rc; + + if (WARN_ON(q->qpage)) + return 0; + + /* Allocate the queue and retrieve infos on current node for now */ + qpage = (__be32 *)__get_free_pages(GFP_KERNEL, xive->q_page_order); + if (!qpage) { + pr_err("Failed to allocate queue %d for VCPU %d\n", + prio, xc->server_num); + return -ENOMEM;; + } + memset(qpage, 0, 1 << xive->q_order); + + /* + * Reconfigure the queue. This will set q->qpage only once the + * queue is fully configured. This is a requirement for prio 0 + * as we will stop doing EOIs for every IPI as soon as we observe + * qpage being non-NULL, and instead will only EOI when we receive + * corresponding queue 0 entries + */ + rc = xive_native_configure_queue(xc->vp_id, q, prio, qpage, + xive->q_order, true); + if (rc) + pr_err("Failed to configure queue %d for VCPU %d\n", + prio, xc->server_num); + return rc; +} + +/* Called with kvm_lock held */ +static int xive_check_provisioning(struct kvm *kvm, u8 prio) +{ + struct kvmppc_xive *xive = kvm->arch.xive; + struct kvm_vcpu *vcpu; + int i, rc; + + lockdep_assert_held(&kvm->lock); + + /* Already provisioned ? */ + if (xive->qmap & (1 << prio)) + return 0; + + pr_devel("Provisioning prio... %d\n", prio); + + /* Provision each VCPU and enable escalations */ + kvm_for_each_vcpu(i, vcpu, kvm) { + if (!vcpu->arch.xive_vcpu) + continue; + rc = xive_provision_queue(vcpu, prio); + if (rc == 0) + xive_attach_escalation(vcpu, prio); + if (rc) + return rc; + } + + /* Order previous stores and mark it as provisioned */ + mb(); + xive->qmap |= (1 << prio); + return 0; +} + +static void xive_inc_q_pending(struct kvm *kvm, u32 server, u8 prio) +{ + struct kvm_vcpu *vcpu; + struct kvmppc_xive_vcpu *xc; + struct xive_q *q; + + /* Locate target server */ + vcpu = kvmppc_xive_find_server(kvm, server); + if (!vcpu) { + pr_warn("%s: Can't find server %d\n", __func__, server); + return; + } + xc = vcpu->arch.xive_vcpu; + if (WARN_ON(!xc)) + return; + + q = &xc->queues[prio]; + atomic_inc(&q->pending_count); +} + +static int xive_try_pick_queue(struct kvm_vcpu *vcpu, u8 prio) +{ + struct kvmppc_xive_vcpu *xc = vcpu->arch.xive_vcpu; + struct xive_q *q; + u32 max; + + if (WARN_ON(!xc)) + return -ENXIO; + if (!xc->valid) + return -ENXIO; + + q = &xc->queues[prio]; + if (WARN_ON(!q->qpage)) + return -ENXIO; + + /* Calculate max number of interrupts in that queue. */ + max = (q->msk + 1) - XIVE_Q_GAP; + return atomic_add_unless(&q->count, 1, max) ? 0 : -EBUSY; +} + +static int xive_select_target(struct kvm *kvm, u32 *server, u8 prio) +{ + struct kvm_vcpu *vcpu; + int i, rc; + + /* Locate target server */ + vcpu = kvmppc_xive_find_server(kvm, *server); + if (!vcpu) { + pr_devel("Can't find server %d\n", *server); + return -EINVAL; + } + + pr_devel("Finding irq target on 0x%x/%d...\n", *server, prio); + + /* Try pick it */ + rc = xive_try_pick_queue(vcpu, prio); + if (rc == 0) + return rc; + + pr_devel(" .. failed, looking up candidate...\n"); + + /* Failed, pick another VCPU */ + kvm_for_each_vcpu(i, vcpu, kvm) { + if (!vcpu->arch.xive_vcpu) + continue; + rc = xive_try_pick_queue(vcpu, prio); + if (rc == 0) { + *server = vcpu->arch.xive_vcpu->server_num; + pr_devel(" found on 0x%x/%d\n", *server, prio); + return rc; + } + } + pr_devel(" no available target !\n"); + + /* No available target ! */ + return -EBUSY; +} + +static u8 xive_lock_and_mask(struct kvmppc_xive *xive, + struct kvmppc_xive_src_block *sb, + struct kvmppc_xive_irq_state *state) +{ + struct xive_irq_data *xd; + u32 hw_num; + u8 old_prio; + u64 val; + + /* + * Take the lock, set masked, try again if racing + * with H_EOI + */ + for (;;) { + arch_spin_lock(&sb->lock); + old_prio = state->guest_priority; + state->guest_priority = MASKED; + mb(); + if (!state->in_eoi) + break; + state->guest_priority = old_prio; + arch_spin_unlock(&sb->lock); + } + + /* No change ? Bail */ + if (old_prio == MASKED) + return old_prio; + + /* Get the right irq */ + kvmppc_xive_select_irq(state, &hw_num, &xd); + + /* + * If the interrupt is marked as needing masking via + * firmware, we do it here. Firmware masking however + * is "lossy", it won't return the old p and q bits + * and won't set the interrupt to a state where it will + * record queued ones. If this is an issue we should do + * lazy masking instead. + * + * For now, we work around this in unmask by forcing + * an interrupt whenever we unmask a non-LSI via FW + * (if ever). + */ + if (xd->flags & OPAL_XIVE_IRQ_MASK_VIA_FW) { + xive_native_configure_irq(hw_num, + xive->vp_base + state->act_server, + MASKED, state->number); + /* set old_p so we can track if an H_EOI was done */ + state->old_p = true; + state->old_q = false; + } else { + /* Set PQ to 10, return old P and old Q and remember them */ + val = xive_vm_esb_load(xd, XIVE_ESB_SET_PQ_10); + state->old_p = !!(val & 2); + state->old_q = !!(val & 1); + + /* + * Synchronize hardware to sensure the queues are updated + * when masking + */ + xive_native_sync_source(hw_num); + } + + return old_prio; +} + +static void xive_lock_for_unmask(struct kvmppc_xive_src_block *sb, + struct kvmppc_xive_irq_state *state) +{ + /* + * Take the lock try again if racing with H_EOI + */ + for (;;) { + arch_spin_lock(&sb->lock); + if (!state->in_eoi) + break; + arch_spin_unlock(&sb->lock); + } +} + +static void xive_finish_unmask(struct kvmppc_xive *xive, + struct kvmppc_xive_src_block *sb, + struct kvmppc_xive_irq_state *state, + u8 prio) +{ + struct xive_irq_data *xd; + u32 hw_num; + + /* If we aren't changing a thing, move on */ + if (state->guest_priority != MASKED) + goto bail; + + /* Get the right irq */ + kvmppc_xive_select_irq(state, &hw_num, &xd); + + /* + * See command in xive_lock_and_mask() concerning masking + * via firmware. + */ + if (xd->flags & OPAL_XIVE_IRQ_MASK_VIA_FW) { + xive_native_configure_irq(hw_num, + xive->vp_base + state->act_server, + state->act_priority, state->number); + /* If an EOI is needed, do it here */ + if (!state->old_p) + xive_vm_source_eoi(hw_num, xd); + /* If this is not an LSI, force a trigger */ + if (!(xd->flags & OPAL_XIVE_IRQ_LSI)) + xive_irq_trigger(xd); + goto bail; + } + + /* Old Q set, set PQ to 11 */ + if (state->old_q) + xive_vm_esb_load(xd, XIVE_ESB_SET_PQ_11); + + /* + * If not old P, then perform an "effective" EOI, + * on the source. This will handle the cases where + * FW EOI is needed. + */ + if (!state->old_p) + xive_vm_source_eoi(hw_num, xd); + + /* Synchronize ordering and mark unmasked */ + mb(); +bail: + state->guest_priority = prio; +} + +/* + * Target an interrupt to a given server/prio, this will fallback + * to another server if necessary and perform the HW targetting + * updates as needed + * + * NOTE: Must be called with the state lock held + */ +static int xive_target_interrupt(struct kvm *kvm, + struct kvmppc_xive_irq_state *state, + u32 server, u8 prio) +{ + struct kvmppc_xive *xive = kvm->arch.xive; + u32 hw_num; + int rc; + + /* + * This will return a tentative server and actual + * priority. The count for that new target will have + * already been incremented. + */ + rc = xive_select_target(kvm, &server, prio); + + /* + * We failed to find a target ? Not much we can do + * at least until we support the GIQ. + */ + if (rc) + return rc; + + /* + * Increment the old queue pending count if there + * was one so that the old queue count gets adjusted later + * when observed to be empty. + */ + if (state->act_priority != MASKED) + xive_inc_q_pending(kvm, + state->act_server, + state->act_priority); + /* + * Update state and HW + */ + state->act_priority = prio; + state->act_server = server; + + /* Get the right irq */ + kvmppc_xive_select_irq(state, &hw_num, NULL); + + return xive_native_configure_irq(hw_num, + xive->vp_base + server, + prio, state->number); +} + +/* + * Targetting rules: In order to avoid losing track of + * pending interrupts accross mask and unmask, which would + * allow queue overflows, we implement the following rules: + * + * - Unless it was never enabled (or we run out of capacity) + * an interrupt is always targetted at a valid server/queue + * pair even when "masked" by the guest. This pair tends to + * be the last one used but it can be changed under some + * circumstances. That allows us to separate targetting + * from masking, we only handle accounting during (re)targetting, + * this also allows us to let an interrupt drain into its target + * queue after masking, avoiding complex schemes to remove + * interrupts out of remote processor queues. + * + * - When masking, we set PQ to 10 and save the previous value + * of P and Q. + * + * - When unmasking, if saved Q was set, we set PQ to 11 + * otherwise we leave PQ to the HW state which will be either + * 10 if nothing happened or 11 if the interrupt fired while + * masked. Effectively we are OR'ing the previous Q into the + * HW Q. + * + * Then if saved P is clear, we do an effective EOI (Q->P->Trigger) + * which will unmask the interrupt and shoot a new one if Q was + * set. + * + * Otherwise (saved P is set) we leave PQ unchanged (so 10 or 11, + * effectively meaning an H_EOI from the guest is still expected + * for that interrupt). + * + * - If H_EOI occurs while masked, we clear the saved P. + * + * - When changing target, we account on the new target and + * increment a separate "pending" counter on the old one. + * This pending counter will be used to decrement the old + * target's count when its queue has been observed empty. + */ + +int kvmppc_xive_set_xive(struct kvm *kvm, u32 irq, u32 server, + u32 priority) +{ + struct kvmppc_xive *xive = kvm->arch.xive; + struct kvmppc_xive_src_block *sb; + struct kvmppc_xive_irq_state *state; + u8 new_act_prio; + int rc = 0; + u16 idx; + + if (!xive) + return -ENODEV; + + pr_devel("set_xive ! irq 0x%x server 0x%x prio %d\n", + irq, server, priority); + + /* First, check provisioning of queues */ + if (priority != MASKED) + rc = xive_check_provisioning(xive->kvm, + xive_prio_from_guest(priority)); + if (rc) { + pr_devel(" provisioning failure %d !\n", rc); + return rc; + } + + sb = kvmppc_xive_find_source(xive, irq, &idx); + if (!sb) + return -EINVAL; + state = &sb->irq_state[idx]; + + /* + * We first handle masking/unmasking since the locking + * might need to be retried due to EOIs, we'll handle + * targetting changes later. These functions will return + * with the SB lock held. + * + * xive_lock_and_mask() will also set state->guest_priority + * but won't otherwise change other fields of the state. + * + * xive_lock_for_unmask will not actually unmask, this will + * be done later by xive_finish_unmask() once the targetting + * has been done, so we don't try to unmask an interrupt + * that hasn't yet been targetted. + */ + if (priority == MASKED) + xive_lock_and_mask(xive, sb, state); + else + xive_lock_for_unmask(sb, state); + + + /* + * Then we handle targetting. + * + * First calculate a new "actual priority" + */ + new_act_prio = state->act_priority; + if (priority != MASKED) + new_act_prio = xive_prio_from_guest(priority); + + pr_devel(" new_act_prio=%x act_server=%x act_prio=%x\n", + new_act_prio, state->act_server, state->act_priority); + + /* + * Then check if we actually need to change anything, + * + * The condition for re-targetting the interrupt is that + * we have a valid new priority (new_act_prio is not 0xff) + * and either the server or the priority changed. + * + * Note: If act_priority was ff and the new priority is + * also ff, we don't do anything and leave the interrupt + * untargetted. An attempt of doing an int_on on an + * untargetted interrupt will fail. If that is a problem + * we could initialize interrupts with valid default + */ + + if (new_act_prio != MASKED && + (state->act_server != server || + state->act_priority != new_act_prio)) + rc = xive_target_interrupt(kvm, state, server, new_act_prio); + + /* + * Perform the final unmasking of the interrupt source + * if necessary + */ + if (priority != MASKED) + xive_finish_unmask(xive, sb, state, priority); + + /* + * Finally Update saved_priority to match. Only int_on/off + * set this field to a different value. + */ + state->saved_priority = priority; + + arch_spin_unlock(&sb->lock); + return rc; +} + +int kvmppc_xive_get_xive(struct kvm *kvm, u32 irq, u32 *server, + u32 *priority) +{ + struct kvmppc_xive *xive = kvm->arch.xive; + struct kvmppc_xive_src_block *sb; + struct kvmppc_xive_irq_state *state; + u16 idx; + + if (!xive) + return -ENODEV; + + sb = kvmppc_xive_find_source(xive, irq, &idx); + if (!sb) + return -EINVAL; + state = &sb->irq_state[idx]; + arch_spin_lock(&sb->lock); + *server = state->guest_server; + *priority = state->guest_priority; + arch_spin_unlock(&sb->lock); + + return 0; +} + +int kvmppc_xive_int_on(struct kvm *kvm, u32 irq) +{ + struct kvmppc_xive *xive = kvm->arch.xive; + struct kvmppc_xive_src_block *sb; + struct kvmppc_xive_irq_state *state; + u16 idx; + + if (!xive) + return -ENODEV; + + sb = kvmppc_xive_find_source(xive, irq, &idx); + if (!sb) + return -EINVAL; + state = &sb->irq_state[idx]; + + pr_devel("int_on(irq=0x%x)\n", irq); + + /* + * Check if interrupt was not targetted + */ + if (state->act_priority == MASKED) { + pr_devel("int_on on untargetted interrupt\n"); + return -EINVAL; + } + + /* If saved_priority is 0xff, do nothing */ + if (state->saved_priority == MASKED) + return 0; + + /* + * Lock and unmask it. + */ + xive_lock_for_unmask(sb, state); + xive_finish_unmask(xive, sb, state, state->saved_priority); + arch_spin_unlock(&sb->lock); + + return 0; +} + +int kvmppc_xive_int_off(struct kvm *kvm, u32 irq) +{ + struct kvmppc_xive *xive = kvm->arch.xive; + struct kvmppc_xive_src_block *sb; + struct kvmppc_xive_irq_state *state; + u16 idx; + + if (!xive) + return -ENODEV; + + sb = kvmppc_xive_find_source(xive, irq, &idx); + if (!sb) + return -EINVAL; + state = &sb->irq_state[idx]; + + pr_devel("int_off(irq=0x%x)\n", irq); + + /* + * Lock and mask + */ + state->saved_priority = xive_lock_and_mask(xive, sb, state); + arch_spin_unlock(&sb->lock); + + return 0; +} + +static bool xive_restore_pending_irq(struct kvmppc_xive *xive, u32 irq) +{ + struct kvmppc_xive_src_block *sb; + struct kvmppc_xive_irq_state *state; + u16 idx; + + sb = kvmppc_xive_find_source(xive, irq, &idx); + if (!sb) + return false; + state = &sb->irq_state[idx]; + if (!state->valid) + return false; + + /* + * Trigger the IPI. This assumes we never restore a pass-through + * interrupt which should be safe enough + */ + xive_irq_trigger(&state->ipi_data); + + return true; +} + +u64 kvmppc_xive_get_icp(struct kvm_vcpu *vcpu) +{ + struct kvmppc_xive_vcpu *xc = vcpu->arch.xive_vcpu; + + if (!xc) + return 0; + + /* Return the per-cpu state for state saving/migration */ + return (u64)xc->cppr << KVM_REG_PPC_ICP_CPPR_SHIFT | + (u64)xc->mfrr << KVM_REG_PPC_ICP_MFRR_SHIFT; +} + +int kvmppc_xive_set_icp(struct kvm_vcpu *vcpu, u64 icpval) +{ + struct kvmppc_xive_vcpu *xc = vcpu->arch.xive_vcpu; + struct kvmppc_xive *xive = vcpu->kvm->arch.xive; + u8 cppr, mfrr; + u32 xisr; + + if (!xc || !xive) + return -ENOENT; + + /* Grab individual state fields. We don't use pending_pri */ + cppr = icpval >> KVM_REG_PPC_ICP_CPPR_SHIFT; + xisr = (icpval >> KVM_REG_PPC_ICP_XISR_SHIFT) & + KVM_REG_PPC_ICP_XISR_MASK; + mfrr = icpval >> KVM_REG_PPC_ICP_MFRR_SHIFT; + + pr_devel("set_icp vcpu %d cppr=0x%x mfrr=0x%x xisr=0x%x\n", + xc->server_num, cppr, mfrr, xisr); + + /* + * We can't update the state of a "pushed" VCPU, but that + * shouldn't happen. + */ + if (WARN_ON(vcpu->arch.xive_pushed)) + return -EIO; + + /* Update VCPU HW saved state */ + vcpu->arch.xive_saved_state.cppr = cppr; + xc->hw_cppr = xc->cppr = cppr; + + /* + * Update MFRR state. If it's not 0xff, we mark the VCPU as + * having a pending MFRR change, which will re-evaluate the + * target. The VCPU will thus potentially get a spurious + * interrupt but that's not a big deal. + */ + xc->mfrr = mfrr; + if (mfrr < cppr) + xive_irq_trigger(&xc->vp_ipi_data); + + /* + * Now saved XIRR is "interesting". It means there's something in + * the legacy "1 element" queue... for an IPI we simply ignore it, + * as the MFRR restore will handle that. For anything else we need + * to force a resend of the source. + * However the source may not have been setup yet. If that's the + * case, we keep that info and increment a counter in the xive to + * tell subsequent xive_set_source() to go look. + */ + if (xisr > XICS_IPI && !xive_restore_pending_irq(xive, xisr)) { + xc->delayed_irq = xisr; + xive->delayed_irqs++; + pr_devel(" xisr restore delayed\n"); + } + + return 0; +} + +int kvmppc_xive_set_mapped(struct kvm *kvm, unsigned long guest_irq, + struct irq_desc *host_desc) +{ + struct kvmppc_xive *xive = kvm->arch.xive; + struct kvmppc_xive_src_block *sb; + struct kvmppc_xive_irq_state *state; + struct irq_data *host_data = irq_desc_get_irq_data(host_desc); + unsigned int host_irq = irq_desc_get_irq(host_desc); + unsigned int hw_irq = (unsigned int)irqd_to_hwirq(host_data); + u16 idx; + u8 prio; + int rc; + + if (!xive) + return -ENODEV; + + pr_devel("set_mapped girq 0x%lx host HW irq 0x%x...\n",guest_irq, hw_irq); + + sb = kvmppc_xive_find_source(xive, guest_irq, &idx); + if (!sb) + return -EINVAL; + state = &sb->irq_state[idx]; + + /* + * Mark the passed-through interrupt as going to a VCPU, + * this will prevent further EOIs and similar operations + * from the XIVE code. It will also mask the interrupt + * to either PQ=10 or 11 state, the latter if the interrupt + * is pending. This will allow us to unmask or retrigger it + * after routing it to the guest with a simple EOI. + * + * The "state" argument is a "token", all it needs is to be + * non-NULL to switch to passed-through or NULL for the + * other way around. We may not yet have an actual VCPU + * target here and we don't really care. + */ + rc = irq_set_vcpu_affinity(host_irq, state); + if (rc) { + pr_err("Failed to set VCPU affinity for irq %d\n", host_irq); + return rc; + } + + /* + * Mask and read state of IPI. We need to know if its P bit + * is set as that means it's potentially already using a + * queue entry in the target + */ + prio = xive_lock_and_mask(xive, sb, state); + pr_devel(" old IPI prio %02x P:%d Q:%d\n", prio, + state->old_p, state->old_q); + + /* Turn the IPI hard off */ + xive_vm_esb_load(&state->ipi_data, XIVE_ESB_SET_PQ_01); + + /* Grab info about irq */ + state->pt_number = hw_irq; + state->pt_data = irq_data_get_irq_handler_data(host_data); + + /* + * Configure the IRQ to match the existing configuration of + * the IPI if it was already targetted. Otherwise this will + * mask the interrupt in a lossy way (act_priority is 0xff) + * which is fine for a never started interrupt. + */ + xive_native_configure_irq(hw_irq, + xive->vp_base + state->act_server, + state->act_priority, state->number); + + /* + * We do an EOI to enable the interrupt (and retrigger if needed) + * if the guest has the interrupt unmasked and the P bit was *not* + * set in the IPI. If it was set, we know a slot may still be in + * use in the target queue thus we have to wait for a guest + * originated EOI + */ + if (prio != MASKED && !state->old_p) + xive_vm_source_eoi(hw_irq, state->pt_data); + + /* Clear old_p/old_q as they are no longer relevant */ + state->old_p = state->old_q = false; + + /* Restore guest prio (unlocks EOI) */ + mb(); + state->guest_priority = prio; + arch_spin_unlock(&sb->lock); + + return 0; +} +EXPORT_SYMBOL_GPL(kvmppc_xive_set_mapped); + +int kvmppc_xive_clr_mapped(struct kvm *kvm, unsigned long guest_irq, + struct irq_desc *host_desc) +{ + struct kvmppc_xive *xive = kvm->arch.xive; + struct kvmppc_xive_src_block *sb; + struct kvmppc_xive_irq_state *state; + unsigned int host_irq = irq_desc_get_irq(host_desc); + u16 idx; + u8 prio; + int rc; + + if (!xive) + return -ENODEV; + + pr_devel("clr_mapped girq 0x%lx...\n", guest_irq); + + sb = kvmppc_xive_find_source(xive, guest_irq, &idx); + if (!sb) + return -EINVAL; + state = &sb->irq_state[idx]; + + /* + * Mask and read state of IRQ. We need to know if its P bit + * is set as that means it's potentially already using a + * queue entry in the target + */ + prio = xive_lock_and_mask(xive, sb, state); + pr_devel(" old IRQ prio %02x P:%d Q:%d\n", prio, + state->old_p, state->old_q); + + /* + * If old_p is set, the interrupt is pending, we switch it to + * PQ=11. This will force a resend in the host so the interrupt + * isn't lost to whatver host driver may pick it up + */ + if (state->old_p) + xive_vm_esb_load(state->pt_data, XIVE_ESB_SET_PQ_11); + + /* Release the passed-through interrupt to the host */ + rc = irq_set_vcpu_affinity(host_irq, NULL); + if (rc) { + pr_err("Failed to clr VCPU affinity for irq %d\n", host_irq); + return rc; + } + + /* Forget about the IRQ */ + state->pt_number = 0; + state->pt_data = NULL; + + /* Reconfigure the IPI */ + xive_native_configure_irq(state->ipi_number, + xive->vp_base + state->act_server, + state->act_priority, state->number); + + /* + * If old_p is set (we have a queue entry potentially + * occupied) or the interrupt is masked, we set the IPI + * to PQ=10 state. Otherwise we just re-enable it (PQ=00). + */ + if (prio == MASKED || state->old_p) + xive_vm_esb_load(&state->ipi_data, XIVE_ESB_SET_PQ_10); + else + xive_vm_esb_load(&state->ipi_data, XIVE_ESB_SET_PQ_00); + + /* Restore guest prio (unlocks EOI) */ + mb(); + state->guest_priority = prio; + arch_spin_unlock(&sb->lock); + + return 0; +} +EXPORT_SYMBOL_GPL(kvmppc_xive_clr_mapped); + +static void kvmppc_xive_disable_vcpu_interrupts(struct kvm_vcpu *vcpu) +{ + struct kvmppc_xive_vcpu *xc = vcpu->arch.xive_vcpu; + struct kvm *kvm = vcpu->kvm; + struct kvmppc_xive *xive = kvm->arch.xive; + int i, j; + + for (i = 0; i <= xive->max_sbid; i++) { + struct kvmppc_xive_src_block *sb = xive->src_blocks[i]; + + if (!sb) + continue; + for (j = 0; j < KVMPPC_XICS_IRQ_PER_ICS; j++) { + struct kvmppc_xive_irq_state *state = &sb->irq_state[j]; + + if (!state->valid) + continue; + if (state->act_priority == MASKED) + continue; + if (state->act_server != xc->server_num) + continue; + + /* Clean it up */ + arch_spin_lock(&sb->lock); + state->act_priority = MASKED; + xive_vm_esb_load(&state->ipi_data, XIVE_ESB_SET_PQ_01); + xive_native_configure_irq(state->ipi_number, 0, MASKED, 0); + if (state->pt_number) { + xive_vm_esb_load(state->pt_data, XIVE_ESB_SET_PQ_01); + xive_native_configure_irq(state->pt_number, 0, MASKED, 0); + } + arch_spin_unlock(&sb->lock); + } + } +} + +void kvmppc_xive_cleanup_vcpu(struct kvm_vcpu *vcpu) +{ + struct kvmppc_xive_vcpu *xc = vcpu->arch.xive_vcpu; + struct kvmppc_xive *xive = xc->xive; + int i; + + pr_devel("cleanup_vcpu(cpu=%d)\n", xc->server_num); + + /* Ensure no interrupt is still routed to that VP */ + xc->valid = false; + kvmppc_xive_disable_vcpu_interrupts(vcpu); + + /* Mask the VP IPI */ + xive_vm_esb_load(&xc->vp_ipi_data, XIVE_ESB_SET_PQ_01); + + /* Disable the VP */ + xive_native_disable_vp(xc->vp_id); + + /* Free the queues & associated interrupts */ + for (i = 0; i < KVMPPC_XIVE_Q_COUNT; i++) { + struct xive_q *q = &xc->queues[i]; + + /* Free the escalation irq */ + if (xc->esc_virq[i]) { + free_irq(xc->esc_virq[i], vcpu); + irq_dispose_mapping(xc->esc_virq[i]); + kfree(xc->esc_virq_names[i]); + } + /* Free the queue */ + xive_native_disable_queue(xc->vp_id, q, i); + if (q->qpage) { + free_pages((unsigned long)q->qpage, + xive->q_page_order); + q->qpage = NULL; + } + } + + /* Free the IPI */ + if (xc->vp_ipi) { + xive_cleanup_irq_data(&xc->vp_ipi_data); + xive_native_free_irq(xc->vp_ipi); + } + /* Free the VP */ + kfree(xc); +} + +int kvmppc_xive_connect_vcpu(struct kvm_device *dev, + struct kvm_vcpu *vcpu, u32 cpu) +{ + struct kvmppc_xive *xive = dev->private; + struct kvmppc_xive_vcpu *xc; + int i, r = -EBUSY; + + pr_devel("connect_vcpu(cpu=%d)\n", cpu); + + if (dev->ops != &kvm_xive_ops) { + pr_devel("Wrong ops !\n"); + return -EPERM; + } + if (xive->kvm != vcpu->kvm) + return -EPERM; + if (vcpu->arch.irq_type) + return -EBUSY; + if (kvmppc_xive_find_server(vcpu->kvm, cpu)) { + pr_devel("Duplicate !\n"); + return -EEXIST; + } + if (cpu >= KVM_MAX_VCPUS) { + pr_devel("Out of bounds !\n"); + return -EINVAL; + } + xc = kzalloc(sizeof(*xc), GFP_KERNEL); + if (!xc) + return -ENOMEM; + + /* We need to synchronize with queue provisioning */ + mutex_lock(&vcpu->kvm->lock); + vcpu->arch.xive_vcpu = xc; + xc->xive = xive; + xc->vcpu = vcpu; + xc->server_num = cpu; + xc->vp_id = xive->vp_base + cpu; + xc->mfrr = 0xff; + xc->valid = true; + + r = xive_native_get_vp_info(xc->vp_id, &xc->vp_cam, &xc->vp_chip_id); + if (r) + goto bail; + + /* Configure VCPU fields for use by assembly push/pull */ + vcpu->arch.xive_saved_state.w01 = cpu_to_be64(0xff000000); + vcpu->arch.xive_cam_word = cpu_to_be32(xc->vp_cam | TM_QW1W2_VO); + + /* Allocate IPI */ + xc->vp_ipi = xive_native_alloc_irq(); + if (!xc->vp_ipi) { + r = -EIO; + goto bail; + } + pr_devel(" IPI=0x%x\n", xc->vp_ipi); + + r = xive_native_populate_irq_data(xc->vp_ipi, &xc->vp_ipi_data); + if (r) + goto bail; + + /* + * Initialize queues. Initially we set them all for no queueing + * and we enable escalation for queue 0 only which we'll use for + * our mfrr change notifications. If the VCPU is hot-plugged, we + * do handle provisioning however. + */ + for (i = 0; i < KVMPPC_XIVE_Q_COUNT; i++) { + struct xive_q *q = &xc->queues[i]; + + /* Is queue already enabled ? Provision it */ + if (xive->qmap & (1 << i)) { + r = xive_provision_queue(vcpu, i); + if (r == 0) + xive_attach_escalation(vcpu, i); + if (r) + goto bail; + } else { + r = xive_native_configure_queue(xc->vp_id, + q, i, NULL, 0, true); + if (r) { + pr_err("Failed to configure queue %d for VCPU %d\n", + i, cpu); + goto bail; + } + } + } + + /* If not done above, attach priority 0 escalation */ + r = xive_attach_escalation(vcpu, 0); + if (r) + goto bail; + + /* Enable the VP */ + r = xive_native_enable_vp(xc->vp_id); + if (r) + goto bail; + + /* Route the IPI */ + r = xive_native_configure_irq(xc->vp_ipi, xc->vp_id, 0, XICS_IPI); + if (!r) + xive_vm_esb_load(&xc->vp_ipi_data, XIVE_ESB_SET_PQ_00); + +bail: + mutex_unlock(&vcpu->kvm->lock); + if (r) { + kvmppc_xive_cleanup_vcpu(vcpu); + return r; + } + + vcpu->arch.irq_type = KVMPPC_IRQ_XICS; + return 0; +} + +/* + * Scanning of queues before/after migration save + */ +static void xive_pre_save_set_queued(struct kvmppc_xive *xive, u32 irq) +{ + struct kvmppc_xive_src_block *sb; + struct kvmppc_xive_irq_state *state; + u16 idx; + + sb = kvmppc_xive_find_source(xive, irq, &idx); + if (!sb) + return; + + state = &sb->irq_state[idx]; + + /* Some sanity checking */ + if (!state->valid) { + pr_err("invalid irq 0x%x in cpu queue!\n", irq); + return; + } + + /* + * If the interrupt is in a queue it should have P set. + * We warn so that gets reported. A backtrace isn't useful + * so no need to use a WARN_ON. + */ + if (!state->saved_p) + pr_err("Interrupt 0x%x is marked in a queue but P not set !\n", irq); + + /* Set flag */ + state->in_queue = true; +} + +static void xive_pre_save_mask_irq(struct kvmppc_xive *xive, + struct kvmppc_xive_src_block *sb, + u32 irq) +{ + struct kvmppc_xive_irq_state *state = &sb->irq_state[irq]; + + if (!state->valid) + return; + + /* Mask and save state, this will also sync HW queues */ + state->saved_scan_prio = xive_lock_and_mask(xive, sb, state); + + /* Transfer P and Q */ + state->saved_p = state->old_p; + state->saved_q = state->old_q; + + /* Unlock */ + arch_spin_unlock(&sb->lock); +} + +static void xive_pre_save_unmask_irq(struct kvmppc_xive *xive, + struct kvmppc_xive_src_block *sb, + u32 irq) +{ + struct kvmppc_xive_irq_state *state = &sb->irq_state[irq]; + + if (!state->valid) + return; + + /* + * Lock / exclude EOI (not technically necessary if the + * guest isn't running concurrently. If this becomes a + * performance issue we can probably remove the lock. + */ + xive_lock_for_unmask(sb, state); + + /* Restore mask/prio if it wasn't masked */ + if (state->saved_scan_prio != MASKED) + xive_finish_unmask(xive, sb, state, state->saved_scan_prio); + + /* Unlock */ + arch_spin_unlock(&sb->lock); +} + +static void xive_pre_save_queue(struct kvmppc_xive *xive, struct xive_q *q) +{ + u32 idx = q->idx; + u32 toggle = q->toggle; + u32 irq; + + do { + irq = __xive_read_eq(q->qpage, q->msk, &idx, &toggle); + if (irq > XICS_IPI) + xive_pre_save_set_queued(xive, irq); + } while(irq); +} + +static void xive_pre_save_scan(struct kvmppc_xive *xive) +{ + struct kvm_vcpu *vcpu = NULL; + int i, j; + + /* + * See comment in xive_get_source() about how this + * work. Collect a stable state for all interrupts + */ + for (i = 0; i <= xive->max_sbid; i++) { + struct kvmppc_xive_src_block *sb = xive->src_blocks[i]; + if (!sb) + continue; + for (j = 0; j < KVMPPC_XICS_IRQ_PER_ICS; j++) + xive_pre_save_mask_irq(xive, sb, j); + } + + /* Then scan the queues and update the "in_queue" flag */ + kvm_for_each_vcpu(i, vcpu, xive->kvm) { + struct kvmppc_xive_vcpu *xc = vcpu->arch.xive_vcpu; + if (!xc) + continue; + for (j = 0; j < KVMPPC_XIVE_Q_COUNT; j++) { + if (xc->queues[i].qpage) + xive_pre_save_queue(xive, &xc->queues[i]); + } + } + + /* Finally restore interrupt states */ + for (i = 0; i <= xive->max_sbid; i++) { + struct kvmppc_xive_src_block *sb = xive->src_blocks[i]; + if (!sb) + continue; + for (j = 0; j < KVMPPC_XICS_IRQ_PER_ICS; j++) + xive_pre_save_unmask_irq(xive, sb, j); + } +} + +static void xive_post_save_scan(struct kvmppc_xive *xive) +{ + u32 i, j; + + /* Clear all the in_queue flags */ + for (i = 0; i <= xive->max_sbid; i++) { + struct kvmppc_xive_src_block *sb = xive->src_blocks[i]; + if (!sb) + continue; + for (j = 0; j < KVMPPC_XICS_IRQ_PER_ICS; j++) + sb->irq_state[j].in_queue = false; + } + + /* Next get_source() will do a new scan */ + xive->saved_src_count = 0; +} + +/* + * This returns the source configuration and state to user space. + */ +static int xive_get_source(struct kvmppc_xive *xive, long irq, u64 addr) +{ + struct kvmppc_xive_src_block *sb; + struct kvmppc_xive_irq_state *state; + u64 __user *ubufp = (u64 __user *) addr; + u64 val, prio; + u16 idx; + + sb = kvmppc_xive_find_source(xive, irq, &idx); + if (!sb) + return -ENOENT; + + state = &sb->irq_state[idx]; + + if (!state->valid) + return -ENOENT; + + pr_devel("get_source(%ld)...\n", irq); + + /* + * So to properly save the state into something that looks like a + * XICS migration stream we cannot treat interrupts individually. + * + * We need, instead, mask them all (& save their previous PQ state) + * to get a stable state in the HW, then sync them to ensure that + * any interrupt that had already fired hits its queue, and finally + * scan all the queues to collect which interrupts are still present + * in the queues, so we can set the "pending" flag on them and + * they can be resent on restore. + * + * So we do it all when the "first" interrupt gets saved, all the + * state is collected at that point, the rest of xive_get_source() + * will merely collect and convert that state to the expected + * userspace bit mask. + */ + if (xive->saved_src_count == 0) + xive_pre_save_scan(xive); + xive->saved_src_count++; + + /* Convert saved state into something compatible with xics */ + val = state->guest_server; + prio = state->saved_scan_prio; + + if (prio == MASKED) { + val |= KVM_XICS_MASKED; + prio = state->saved_priority; + } + val |= prio << KVM_XICS_PRIORITY_SHIFT; + if (state->lsi) { + val |= KVM_XICS_LEVEL_SENSITIVE; + if (state->saved_p) + val |= KVM_XICS_PENDING; + } else { + if (state->saved_p) + val |= KVM_XICS_PRESENTED; + + if (state->saved_q) + val |= KVM_XICS_QUEUED; + + /* + * We mark it pending (which will attempt a re-delivery) + * if we are in a queue *or* we were masked and had + * Q set which is equivalent to the XICS "masked pending" + * state + */ + if (state->in_queue || (prio == MASKED && state->saved_q)) + val |= KVM_XICS_PENDING; + } + + /* + * If that was the last interrupt saved, reset the + * in_queue flags + */ + if (xive->saved_src_count == xive->src_count) + xive_post_save_scan(xive); + + /* Copy the result to userspace */ + if (put_user(val, ubufp)) + return -EFAULT; + + return 0; +} + +static struct kvmppc_xive_src_block *xive_create_src_block(struct kvmppc_xive *xive, + int irq) +{ + struct kvm *kvm = xive->kvm; + struct kvmppc_xive_src_block *sb; + int i, bid; + + bid = irq >> KVMPPC_XICS_ICS_SHIFT; + + mutex_lock(&kvm->lock); + + /* block already exists - somebody else got here first */ + if (xive->src_blocks[bid]) + goto out; + + /* Create the ICS */ + sb = kzalloc(sizeof(*sb), GFP_KERNEL); + if (!sb) + goto out; + + sb->id = bid; + + for (i = 0; i < KVMPPC_XICS_IRQ_PER_ICS; i++) { + sb->irq_state[i].number = (bid << KVMPPC_XICS_ICS_SHIFT) | i; + sb->irq_state[i].guest_priority = MASKED; + sb->irq_state[i].saved_priority = MASKED; + sb->irq_state[i].act_priority = MASKED; + } + smp_wmb(); + xive->src_blocks[bid] = sb; + + if (bid > xive->max_sbid) + xive->max_sbid = bid; + +out: + mutex_unlock(&kvm->lock); + return xive->src_blocks[bid]; +} + +static bool xive_check_delayed_irq(struct kvmppc_xive *xive, u32 irq) +{ + struct kvm *kvm = xive->kvm; + struct kvm_vcpu *vcpu = NULL; + int i; + + kvm_for_each_vcpu(i, vcpu, kvm) { + struct kvmppc_xive_vcpu *xc = vcpu->arch.xive_vcpu; + + if (!xc) + continue; + + if (xc->delayed_irq == irq) { + xc->delayed_irq = 0; + xive->delayed_irqs--; + return true; + } + } + return false; +} + +static int xive_set_source(struct kvmppc_xive *xive, long irq, u64 addr) +{ + struct kvmppc_xive_src_block *sb; + struct kvmppc_xive_irq_state *state; + u64 __user *ubufp = (u64 __user *) addr; + u16 idx; + u64 val; + u8 act_prio, guest_prio; + u32 server; + int rc = 0; + + if (irq < KVMPPC_XICS_FIRST_IRQ || irq >= KVMPPC_XICS_NR_IRQS) + return -ENOENT; + + pr_devel("set_source(irq=0x%lx)\n", irq); + + /* Find the source */ + sb = kvmppc_xive_find_source(xive, irq, &idx); + if (!sb) { + pr_devel("No source, creating source block...\n"); + sb = xive_create_src_block(xive, irq); + if (!sb) { + pr_devel("Failed to create block...\n"); + return -ENOMEM; + } + } + state = &sb->irq_state[idx]; + + /* Read user passed data */ + if (get_user(val, ubufp)) { + pr_devel("fault getting user info !\n"); + return -EFAULT; + } + + server = val & KVM_XICS_DESTINATION_MASK; + guest_prio = val >> KVM_XICS_PRIORITY_SHIFT; + + pr_devel(" val=0x016%llx (server=0x%x, guest_prio=%d)\n", + val, server, guest_prio); + /* + * If the source doesn't already have an IPI, allocate + * one and get the corresponding data + */ + if (!state->ipi_number) { + state->ipi_number = xive_native_alloc_irq(); + if (state->ipi_number == 0) { + pr_devel("Failed to allocate IPI !\n"); + return -ENOMEM; + } + xive_native_populate_irq_data(state->ipi_number, &state->ipi_data); + pr_devel(" src_ipi=0x%x\n", state->ipi_number); + } + + /* + * We use lock_and_mask() to set us in the right masked + * state. We will override that state from the saved state + * further down, but this will handle the cases of interrupts + * that need FW masking. We set the initial guest_priority to + * 0 before calling it to ensure it actually performs the masking. + */ + state->guest_priority = 0; + xive_lock_and_mask(xive, sb, state); + + /* + * Now, we select a target if we have one. If we don't we + * leave the interrupt untargetted. It means that an interrupt + * can become "untargetted" accross migration if it was masked + * by set_xive() but there is little we can do about it. + */ + + /* First convert prio and mark interrupt as untargetted */ + act_prio = xive_prio_from_guest(guest_prio); + state->act_priority = MASKED; + state->guest_server = server; + + /* + * We need to drop the lock due to the mutex below. Hopefully + * nothing is touching that interrupt yet since it hasn't been + * advertized to a running guest yet + */ + arch_spin_unlock(&sb->lock); + + /* If we have a priority target the interrupt */ + if (act_prio != MASKED) { + /* First, check provisioning of queues */ + mutex_lock(&xive->kvm->lock); + rc = xive_check_provisioning(xive->kvm, act_prio); + mutex_unlock(&xive->kvm->lock); + + /* Target interrupt */ + if (rc == 0) + rc = xive_target_interrupt(xive->kvm, state, + server, act_prio); + /* + * If provisioning or targetting failed, leave it + * alone and masked. It will remain disabled until + * the guest re-targets it. + */ + } + + /* + * Find out if this was a delayed irq stashed in an ICP, + * in which case, treat it as pending + */ + if (xive->delayed_irqs && xive_check_delayed_irq(xive, irq)) { + val |= KVM_XICS_PENDING; + pr_devel(" Found delayed ! forcing PENDING !\n"); + } + + /* Cleanup the SW state */ + state->old_p = false; + state->old_q = false; + state->lsi = false; + state->asserted = false; + + /* Restore LSI state */ + if (val & KVM_XICS_LEVEL_SENSITIVE) { + state->lsi = true; + if (val & KVM_XICS_PENDING) + state->asserted = true; + pr_devel(" LSI ! Asserted=%d\n", state->asserted); + } + + /* + * Restore P and Q. If the interrupt was pending, we + * force both P and Q, which will trigger a resend. + * + * That means that a guest that had both an interrupt + * pending (queued) and Q set will restore with only + * one instance of that interrupt instead of 2, but that + * is perfectly fine as coalescing interrupts that haven't + * been presented yet is always allowed. + */ + if (val & KVM_XICS_PRESENTED || val & KVM_XICS_PENDING) + state->old_p = true; + if (val & KVM_XICS_QUEUED || val & KVM_XICS_PENDING) + state->old_q = true; + + pr_devel(" P=%d, Q=%d\n", state->old_p, state->old_q); + + /* + * If the interrupt was unmasked, update guest priority and + * perform the appropriate state transition and do a + * re-trigger if necessary. + */ + if (val & KVM_XICS_MASKED) { + pr_devel(" masked, saving prio\n"); + state->guest_priority = MASKED; + state->saved_priority = guest_prio; + } else { + pr_devel(" unmasked, restoring to prio %d\n", guest_prio); + xive_finish_unmask(xive, sb, state, guest_prio); + state->saved_priority = guest_prio; + } + + /* Increment the number of valid sources and mark this one valid */ + if (!state->valid) + xive->src_count++; + state->valid = true; + + return 0; +} + +int kvmppc_xive_set_irq(struct kvm *kvm, int irq_source_id, u32 irq, int level, + bool line_status) +{ + struct kvmppc_xive *xive = kvm->arch.xive; + struct kvmppc_xive_src_block *sb; + struct kvmppc_xive_irq_state *state; + u16 idx; + + if (!xive) + return -ENODEV; + + sb = kvmppc_xive_find_source(xive, irq, &idx); + if (!sb) + return -EINVAL; + + /* Perform locklessly .... (we need to do some RCUisms here...) */ + state = &sb->irq_state[idx]; + if (!state->valid) + return -EINVAL; + + /* We don't allow a trigger on a passed-through interrupt */ + if (state->pt_number) + return -EINVAL; + + if ((level == 1 && state->lsi) || level == KVM_INTERRUPT_SET_LEVEL) + state->asserted = 1; + else if (level == 0 || level == KVM_INTERRUPT_UNSET) { + state->asserted = 0; + return 0; + } + + /* Trigger the IPI */ + xive_irq_trigger(&state->ipi_data); + + return 0; +} + +static int xive_set_attr(struct kvm_device *dev, struct kvm_device_attr *attr) +{ + struct kvmppc_xive *xive = dev->private; + + /* We honor the existing XICS ioctl */ + switch (attr->group) { + case KVM_DEV_XICS_GRP_SOURCES: + return xive_set_source(xive, attr->attr, attr->addr); + } + return -ENXIO; +} + +static int xive_get_attr(struct kvm_device *dev, struct kvm_device_attr *attr) +{ + struct kvmppc_xive *xive = dev->private; + + /* We honor the existing XICS ioctl */ + switch (attr->group) { + case KVM_DEV_XICS_GRP_SOURCES: + return xive_get_source(xive, attr->attr, attr->addr); + } + return -ENXIO; +} + +static int xive_has_attr(struct kvm_device *dev, struct kvm_device_attr *attr) +{ + /* We honor the same limits as XICS, at least for now */ + switch (attr->group) { + case KVM_DEV_XICS_GRP_SOURCES: + if (attr->attr >= KVMPPC_XICS_FIRST_IRQ && + attr->attr < KVMPPC_XICS_NR_IRQS) + return 0; + break; + } + return -ENXIO; +} + +static void kvmppc_xive_cleanup_irq(u32 hw_num, struct xive_irq_data *xd) +{ + xive_vm_esb_load(xd, XIVE_ESB_SET_PQ_01); + xive_native_configure_irq(hw_num, 0, MASKED, 0); + xive_cleanup_irq_data(xd); +} + +static void kvmppc_xive_free_sources(struct kvmppc_xive_src_block *sb) +{ + int i; + + for (i = 0; i < KVMPPC_XICS_IRQ_PER_ICS; i++) { + struct kvmppc_xive_irq_state *state = &sb->irq_state[i]; + + if (!state->valid) + continue; + + kvmppc_xive_cleanup_irq(state->ipi_number, &state->ipi_data); + xive_native_free_irq(state->ipi_number); + + /* Pass-through, cleanup too */ + if (state->pt_number) + kvmppc_xive_cleanup_irq(state->pt_number, state->pt_data); + + state->valid = false; + } +} + +static void kvmppc_xive_free(struct kvm_device *dev) +{ + struct kvmppc_xive *xive = dev->private; + struct kvm *kvm = xive->kvm; + int i; + + debugfs_remove(xive->dentry); + + if (kvm) + kvm->arch.xive = NULL; + + /* Mask and free interrupts */ + for (i = 0; i <= xive->max_sbid; i++) { + if (xive->src_blocks[i]) + kvmppc_xive_free_sources(xive->src_blocks[i]); + kfree(xive->src_blocks[i]); + xive->src_blocks[i] = NULL; + } + + if (xive->vp_base != XIVE_INVALID_VP) + xive_native_free_vp_block(xive->vp_base); + + + kfree(xive); + kfree(dev); +} + +static int kvmppc_xive_create(struct kvm_device *dev, u32 type) +{ + struct kvmppc_xive *xive; + struct kvm *kvm = dev->kvm; + int ret = 0; + + pr_devel("Creating xive for partition\n"); + + xive = kzalloc(sizeof(*xive), GFP_KERNEL); + if (!xive) + return -ENOMEM; + + dev->private = xive; + xive->dev = dev; + xive->kvm = kvm; + + /* Already there ? */ + if (kvm->arch.xive) + ret = -EEXIST; + else + kvm->arch.xive = xive; + + /* We use the default queue size set by the host */ + xive->q_order = xive_native_default_eq_shift(); + if (xive->q_order < PAGE_SHIFT) + xive->q_page_order = 0; + else + xive->q_page_order = xive->q_order - PAGE_SHIFT; + + /* Allocate a bunch of VPs */ + xive->vp_base = xive_native_alloc_vp_block(KVM_MAX_VCPUS); + pr_devel("VP_Base=%x\n", xive->vp_base); + + if (xive->vp_base == XIVE_INVALID_VP) + ret = -ENOMEM; + + if (ret) { + kfree(xive); + return ret; + } + + return 0; +} + + +static int xive_debug_show(struct seq_file *m, void *private) +{ + struct kvmppc_xive *xive = m->private; + struct kvm *kvm = xive->kvm; + struct kvm_vcpu *vcpu; + u64 t_rm_h_xirr = 0; + u64 t_rm_h_ipoll = 0; + u64 t_rm_h_cppr = 0; + u64 t_rm_h_eoi = 0; + u64 t_rm_h_ipi = 0; + u64 t_vm_h_xirr = 0; + u64 t_vm_h_ipoll = 0; + u64 t_vm_h_cppr = 0; + u64 t_vm_h_eoi = 0; + u64 t_vm_h_ipi = 0; + unsigned int i; + + if (!kvm) + return 0; + + seq_printf(m, "=========\nVCPU state\n=========\n"); + + kvm_for_each_vcpu(i, vcpu, kvm) { + struct kvmppc_xive_vcpu *xc = vcpu->arch.xive_vcpu; + + if (!xc) + continue; + + seq_printf(m, "cpu server %#x CPPR:%#x HWCPPR:%#x" + " MFRR:%#x PEND:%#x h_xirr: R=%lld V=%lld\n", + xc->server_num, xc->cppr, xc->hw_cppr, + xc->mfrr, xc->pending, + xc->stat_rm_h_xirr, xc->stat_vm_h_xirr); + + t_rm_h_xirr += xc->stat_rm_h_xirr; + t_rm_h_ipoll += xc->stat_rm_h_ipoll; + t_rm_h_cppr += xc->stat_rm_h_cppr; + t_rm_h_eoi += xc->stat_rm_h_eoi; + t_rm_h_ipi += xc->stat_rm_h_ipi; + t_vm_h_xirr += xc->stat_vm_h_xirr; + t_vm_h_ipoll += xc->stat_vm_h_ipoll; + t_vm_h_cppr += xc->stat_vm_h_cppr; + t_vm_h_eoi += xc->stat_vm_h_eoi; + t_vm_h_ipi += xc->stat_vm_h_ipi; + } + + seq_printf(m, "Hcalls totals\n"); + seq_printf(m, " H_XIRR R=%10lld V=%10lld\n", t_rm_h_xirr, t_vm_h_xirr); + seq_printf(m, " H_IPOLL R=%10lld V=%10lld\n", t_rm_h_ipoll, t_vm_h_ipoll); + seq_printf(m, " H_CPPR R=%10lld V=%10lld\n", t_rm_h_cppr, t_vm_h_cppr); + seq_printf(m, " H_EOI R=%10lld V=%10lld\n", t_rm_h_eoi, t_vm_h_eoi); + seq_printf(m, " H_IPI R=%10lld V=%10lld\n", t_rm_h_ipi, t_vm_h_ipi); + + return 0; +} + +static int xive_debug_open(struct inode *inode, struct file *file) +{ + return single_open(file, xive_debug_show, inode->i_private); +} + +static const struct file_operations xive_debug_fops = { + .open = xive_debug_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, +}; + +static void xive_debugfs_init(struct kvmppc_xive *xive) +{ + char *name; + + name = kasprintf(GFP_KERNEL, "kvm-xive-%p", xive); + if (!name) { + pr_err("%s: no memory for name\n", __func__); + return; + } + + xive->dentry = debugfs_create_file(name, S_IRUGO, powerpc_debugfs_root, + xive, &xive_debug_fops); + + pr_debug("%s: created %s\n", __func__, name); + kfree(name); +} + +static void kvmppc_xive_init(struct kvm_device *dev) +{ + struct kvmppc_xive *xive = (struct kvmppc_xive *)dev->private; + + /* Register some debug interfaces */ + xive_debugfs_init(xive); +} + +struct kvm_device_ops kvm_xive_ops = { + .name = "kvm-xive", + .create = kvmppc_xive_create, + .init = kvmppc_xive_init, + .destroy = kvmppc_xive_free, + .set_attr = xive_set_attr, + .get_attr = xive_get_attr, + .has_attr = xive_has_attr, +}; + +void kvmppc_xive_init_module(void) +{ + __xive_vm_h_xirr = xive_vm_h_xirr; + __xive_vm_h_ipoll = xive_vm_h_ipoll; + __xive_vm_h_ipi = xive_vm_h_ipi; + __xive_vm_h_cppr = xive_vm_h_cppr; + __xive_vm_h_eoi = xive_vm_h_eoi; +} + +void kvmppc_xive_exit_module(void) +{ + __xive_vm_h_xirr = NULL; + __xive_vm_h_ipoll = NULL; + __xive_vm_h_ipi = NULL; + __xive_vm_h_cppr = NULL; + __xive_vm_h_eoi = NULL; +} diff --git a/arch/powerpc/kvm/book3s_xive.h b/arch/powerpc/kvm/book3s_xive.h new file mode 100644 index 000000000000..5938f7644dc1 --- /dev/null +++ b/arch/powerpc/kvm/book3s_xive.h @@ -0,0 +1,256 @@ +/* + * Copyright 2017 Benjamin Herrenschmidt, IBM Corporation + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License, version 2, as + * published by the Free Software Foundation. + */ + +#ifndef _KVM_PPC_BOOK3S_XIVE_H +#define _KVM_PPC_BOOK3S_XIVE_H + +#ifdef CONFIG_KVM_XICS +#include "book3s_xics.h" + +/* + * State for one guest irq source. + * + * For each guest source we allocate a HW interrupt in the XIVE + * which we use for all SW triggers. It will be unused for + * pass-through but it's easier to keep around as the same + * guest interrupt can alternatively be emulated or pass-through + * if a physical device is hot unplugged and replaced with an + * emulated one. + * + * This state structure is very similar to the XICS one with + * additional XIVE specific tracking. + */ +struct kvmppc_xive_irq_state { + bool valid; /* Interrupt entry is valid */ + + u32 number; /* Guest IRQ number */ + u32 ipi_number; /* XIVE IPI HW number */ + struct xive_irq_data ipi_data; /* XIVE IPI associated data */ + u32 pt_number; /* XIVE Pass-through number if any */ + struct xive_irq_data *pt_data; /* XIVE Pass-through associated data */ + + /* Targetting as set by guest */ + u32 guest_server; /* Current guest selected target */ + u8 guest_priority; /* Guest set priority */ + u8 saved_priority; /* Saved priority when masking */ + + /* Actual targetting */ + u32 act_server; /* Actual server */ + u8 act_priority; /* Actual priority */ + + /* Various state bits */ + bool in_eoi; /* Synchronize with H_EOI */ + bool old_p; /* P bit state when masking */ + bool old_q; /* Q bit state when masking */ + bool lsi; /* level-sensitive interrupt */ + bool asserted; /* Only for emulated LSI: current state */ + + /* Saved for migration state */ + bool in_queue; + bool saved_p; + bool saved_q; + u8 saved_scan_prio; +}; + +/* Select the "right" interrupt (IPI vs. passthrough) */ +static inline void kvmppc_xive_select_irq(struct kvmppc_xive_irq_state *state, + u32 *out_hw_irq, + struct xive_irq_data **out_xd) +{ + if (state->pt_number) { + if (out_hw_irq) + *out_hw_irq = state->pt_number; + if (out_xd) + *out_xd = state->pt_data; + } else { + if (out_hw_irq) + *out_hw_irq = state->ipi_number; + if (out_xd) + *out_xd = &state->ipi_data; + } +} + +/* + * This corresponds to an "ICS" in XICS terminology, we use it + * as a mean to break up source information into multiple structures. + */ +struct kvmppc_xive_src_block { + arch_spinlock_t lock; + u16 id; + struct kvmppc_xive_irq_state irq_state[KVMPPC_XICS_IRQ_PER_ICS]; +}; + + +struct kvmppc_xive { + struct kvm *kvm; + struct kvm_device *dev; + struct dentry *dentry; + + /* VP block associated with the VM */ + u32 vp_base; + + /* Blocks of sources */ + struct kvmppc_xive_src_block *src_blocks[KVMPPC_XICS_MAX_ICS_ID + 1]; + u32 max_sbid; + + /* + * For state save, we lazily scan the queues on the first interrupt + * being migrated. We don't have a clean way to reset that flags + * so we keep track of the number of valid sources and how many of + * them were migrated so we can reset when all of them have been + * processed. + */ + u32 src_count; + u32 saved_src_count; + + /* + * Some irqs are delayed on restore until the source is created, + * keep track here of how many of them + */ + u32 delayed_irqs; + + /* Which queues (priorities) are in use by the guest */ + u8 qmap; + + /* Queue orders */ + u32 q_order; + u32 q_page_order; + +}; + +#define KVMPPC_XIVE_Q_COUNT 8 + +struct kvmppc_xive_vcpu { + struct kvmppc_xive *xive; + struct kvm_vcpu *vcpu; + bool valid; + + /* Server number. This is the HW CPU ID from a guest perspective */ + u32 server_num; + + /* + * HW VP corresponding to this VCPU. This is the base of the VP + * block plus the server number. + */ + u32 vp_id; + u32 vp_chip_id; + u32 vp_cam; + + /* IPI used for sending ... IPIs */ + u32 vp_ipi; + struct xive_irq_data vp_ipi_data; + + /* Local emulation state */ + uint8_t cppr; /* guest CPPR */ + uint8_t hw_cppr;/* Hardware CPPR */ + uint8_t mfrr; + uint8_t pending; + + /* Each VP has 8 queues though we only provision some */ + struct xive_q queues[KVMPPC_XIVE_Q_COUNT]; + u32 esc_virq[KVMPPC_XIVE_Q_COUNT]; + char *esc_virq_names[KVMPPC_XIVE_Q_COUNT]; + + /* Stash a delayed irq on restore from migration (see set_icp) */ + u32 delayed_irq; + + /* Stats */ + u64 stat_rm_h_xirr; + u64 stat_rm_h_ipoll; + u64 stat_rm_h_cppr; + u64 stat_rm_h_eoi; + u64 stat_rm_h_ipi; + u64 stat_vm_h_xirr; + u64 stat_vm_h_ipoll; + u64 stat_vm_h_cppr; + u64 stat_vm_h_eoi; + u64 stat_vm_h_ipi; +}; + +static inline struct kvm_vcpu *kvmppc_xive_find_server(struct kvm *kvm, u32 nr) +{ + struct kvm_vcpu *vcpu = NULL; + int i; + + kvm_for_each_vcpu(i, vcpu, kvm) { + if (vcpu->arch.xive_vcpu && nr == vcpu->arch.xive_vcpu->server_num) + return vcpu; + } + return NULL; +} + +static inline struct kvmppc_xive_src_block *kvmppc_xive_find_source(struct kvmppc_xive *xive, + u32 irq, u16 *source) +{ + u32 bid = irq >> KVMPPC_XICS_ICS_SHIFT; + u16 src = irq & KVMPPC_XICS_SRC_MASK; + + if (source) + *source = src; + if (bid > KVMPPC_XICS_MAX_ICS_ID) + return NULL; + return xive->src_blocks[bid]; +} + +/* + * Mapping between guest priorities and host priorities + * is as follow. + * + * Guest request for 0...6 are honored. Guest request for anything + * higher results in a priority of 7 being applied. + * + * However, when XIRR is returned via H_XIRR, 7 is translated to 0xb + * in order to match AIX expectations + * + * Similar mapping is done for CPPR values + */ +static inline u8 xive_prio_from_guest(u8 prio) +{ + if (prio == 0xff || prio < 8) + return prio; + return 7; +} + +static inline u8 xive_prio_to_guest(u8 prio) +{ + if (prio == 0xff || prio < 7) + return prio; + return 0xb; +} + +static inline u32 __xive_read_eq(__be32 *qpage, u32 msk, u32 *idx, u32 *toggle) +{ + u32 cur; + + if (!qpage) + return 0; + cur = be32_to_cpup(qpage + *idx); + if ((cur >> 31) == *toggle) + return 0; + *idx = (*idx + 1) & msk; + if (*idx == 0) + (*toggle) ^= 1; + return cur & 0x7fffffff; +} + +extern unsigned long xive_rm_h_xirr(struct kvm_vcpu *vcpu); +extern unsigned long xive_rm_h_ipoll(struct kvm_vcpu *vcpu, unsigned long server); +extern int xive_rm_h_ipi(struct kvm_vcpu *vcpu, unsigned long server, + unsigned long mfrr); +extern int xive_rm_h_cppr(struct kvm_vcpu *vcpu, unsigned long cppr); +extern int xive_rm_h_eoi(struct kvm_vcpu *vcpu, unsigned long xirr); + +extern unsigned long (*__xive_vm_h_xirr)(struct kvm_vcpu *vcpu); +extern unsigned long (*__xive_vm_h_ipoll)(struct kvm_vcpu *vcpu, unsigned long server); +extern int (*__xive_vm_h_ipi)(struct kvm_vcpu *vcpu, unsigned long server, + unsigned long mfrr); +extern int (*__xive_vm_h_cppr)(struct kvm_vcpu *vcpu, unsigned long cppr); +extern int (*__xive_vm_h_eoi)(struct kvm_vcpu *vcpu, unsigned long xirr); + +#endif /* CONFIG_KVM_XICS */ +#endif /* _KVM_PPC_BOOK3S_XICS_H */ diff --git a/arch/powerpc/kvm/book3s_xive_template.c b/arch/powerpc/kvm/book3s_xive_template.c new file mode 100644 index 000000000000..023a31133c37 --- /dev/null +++ b/arch/powerpc/kvm/book3s_xive_template.c @@ -0,0 +1,503 @@ +/* + * Copyright 2017 Benjamin Herrenschmidt, IBM Corporation + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License, version 2, as + * published by the Free Software Foundation. + */ + +/* File to be included by other .c files */ + +#define XGLUE(a,b) a##b +#define GLUE(a,b) XGLUE(a,b) + +static void GLUE(X_PFX,ack_pending)(struct kvmppc_xive_vcpu *xc) +{ + u8 cppr; + u16 ack; + + /* XXX DD1 bug workaround: Check PIPR vs. CPPR first ! */ + + /* Perform the acknowledge OS to register cycle. */ + ack = be16_to_cpu(__x_readw(__x_tima + TM_SPC_ACK_OS_REG)); + + /* Synchronize subsequent queue accesses */ + mb(); + + /* XXX Check grouping level */ + + /* Anything ? */ + if (!((ack >> 8) & TM_QW1_NSR_EO)) + return; + + /* Grab CPPR of the most favored pending interrupt */ + cppr = ack & 0xff; + if (cppr < 8) + xc->pending |= 1 << cppr; + +#ifdef XIVE_RUNTIME_CHECKS + /* Check consistency */ + if (cppr >= xc->hw_cppr) + pr_warn("KVM-XIVE: CPU %d odd ack CPPR, got %d at %d\n", + smp_processor_id(), cppr, xc->hw_cppr); +#endif + + /* + * Update our image of the HW CPPR. We don't yet modify + * xc->cppr, this will be done as we scan for interrupts + * in the queues. + */ + xc->hw_cppr = cppr; +} + +static u8 GLUE(X_PFX,esb_load)(struct xive_irq_data *xd, u32 offset) +{ + u64 val; + + if (xd->flags & XIVE_IRQ_FLAG_SHIFT_BUG) + offset |= offset << 4; + + val =__x_readq(__x_eoi_page(xd) + offset); +#ifdef __LITTLE_ENDIAN__ + val >>= 64-8; +#endif + return (u8)val; +} + + +static void GLUE(X_PFX,source_eoi)(u32 hw_irq, struct xive_irq_data *xd) +{ + /* If the XIVE supports the new "store EOI facility, use it */ + if (xd->flags & XIVE_IRQ_FLAG_STORE_EOI) + __x_writeq(0, __x_eoi_page(xd)); + else if (hw_irq && xd->flags & XIVE_IRQ_FLAG_EOI_FW) { + opal_int_eoi(hw_irq); + } else { + uint64_t eoi_val; + + /* + * Otherwise for EOI, we use the special MMIO that does + * a clear of both P and Q and returns the old Q, + * except for LSIs where we use the "EOI cycle" special + * load. + * + * This allows us to then do a re-trigger if Q was set + * rather than synthetizing an interrupt in software + * + * For LSIs, using the HW EOI cycle works around a problem + * on P9 DD1 PHBs where the other ESB accesses don't work + * properly. + */ + if (xd->flags & XIVE_IRQ_FLAG_LSI) + __x_readq(__x_eoi_page(xd)); + else { + eoi_val = GLUE(X_PFX,esb_load)(xd, XIVE_ESB_SET_PQ_00); + + /* Re-trigger if needed */ + if ((eoi_val & 1) && __x_trig_page(xd)) + __x_writeq(0, __x_trig_page(xd)); + } + } +} + +enum { + scan_fetch, + scan_poll, + scan_eoi, +}; + +static u32 GLUE(X_PFX,scan_interrupts)(struct kvmppc_xive_vcpu *xc, + u8 pending, int scan_type) +{ + u32 hirq = 0; + u8 prio = 0xff; + + /* Find highest pending priority */ + while ((xc->mfrr != 0xff || pending != 0) && hirq == 0) { + struct xive_q *q; + u32 idx, toggle; + __be32 *qpage; + + /* + * If pending is 0 this will return 0xff which is what + * we want + */ + prio = ffs(pending) - 1; + + /* + * If the most favoured prio we found pending is less + * favored (or equal) than a pending IPI, we return + * the IPI instead. + * + * Note: If pending was 0 and mfrr is 0xff, we will + * not spurriously take an IPI because mfrr cannot + * then be smaller than cppr. + */ + if (prio >= xc->mfrr && xc->mfrr < xc->cppr) { + prio = xc->mfrr; + hirq = XICS_IPI; + break; + } + + /* Don't scan past the guest cppr */ + if (prio >= xc->cppr || prio > 7) + break; + + /* Grab queue and pointers */ + q = &xc->queues[prio]; + idx = q->idx; + toggle = q->toggle; + + /* + * Snapshot the queue page. The test further down for EOI + * must use the same "copy" that was used by __xive_read_eq + * since qpage can be set concurrently and we don't want + * to miss an EOI. + */ + qpage = READ_ONCE(q->qpage); + +skip_ipi: + /* + * Try to fetch from the queue. Will return 0 for a + * non-queueing priority (ie, qpage = 0). + */ + hirq = __xive_read_eq(qpage, q->msk, &idx, &toggle); + + /* + * If this was a signal for an MFFR change done by + * H_IPI we skip it. Additionally, if we were fetching + * we EOI it now, thus re-enabling reception of a new + * such signal. + * + * We also need to do that if prio is 0 and we had no + * page for the queue. In this case, we have non-queued + * IPI that needs to be EOId. + * + * This is safe because if we have another pending MFRR + * change that wasn't observed above, the Q bit will have + * been set and another occurrence of the IPI will trigger. + */ + if (hirq == XICS_IPI || (prio == 0 && !qpage)) { + if (scan_type == scan_fetch) + GLUE(X_PFX,source_eoi)(xc->vp_ipi, + &xc->vp_ipi_data); + /* Loop back on same queue with updated idx/toggle */ +#ifdef XIVE_RUNTIME_CHECKS + WARN_ON(hirq && hirq != XICS_IPI); +#endif + if (hirq) + goto skip_ipi; + } + + /* If fetching, update queue pointers */ + if (scan_type == scan_fetch) { + q->idx = idx; + q->toggle = toggle; + } + + /* Something found, stop searching */ + if (hirq) + break; + + /* Clear the pending bit on the now empty queue */ + pending &= ~(1 << prio); + + /* + * Check if the queue count needs adjusting due to + * interrupts being moved away. + */ + if (atomic_read(&q->pending_count)) { + int p = atomic_xchg(&q->pending_count, 0); + if (p) { +#ifdef XIVE_RUNTIME_CHECKS + WARN_ON(p > atomic_read(&q->count)); +#endif + atomic_sub(p, &q->count); + } + } + } + + /* If we are just taking a "peek", do nothing else */ + if (scan_type == scan_poll) + return hirq; + + /* Update the pending bits */ + xc->pending = pending; + + /* + * If this is an EOI that's it, no CPPR adjustment done here, + * all we needed was cleanup the stale pending bits and check + * if there's anything left. + */ + if (scan_type == scan_eoi) + return hirq; + + /* + * If we found an interrupt, adjust what the guest CPPR should + * be as if we had just fetched that interrupt from HW. + */ + if (hirq) + xc->cppr = prio; + /* + * If it was an IPI the HW CPPR might have been lowered too much + * as the HW interrupt we use for IPIs is routed to priority 0. + * + * We re-sync it here. + */ + if (xc->cppr != xc->hw_cppr) { + xc->hw_cppr = xc->cppr; + __x_writeb(xc->cppr, __x_tima + TM_QW1_OS + TM_CPPR); + } + + return hirq; +} + +X_STATIC unsigned long GLUE(X_PFX,h_xirr)(struct kvm_vcpu *vcpu) +{ + struct kvmppc_xive_vcpu *xc = vcpu->arch.xive_vcpu; + u8 old_cppr; + u32 hirq; + + pr_devel("H_XIRR\n"); + + xc->GLUE(X_STAT_PFX,h_xirr)++; + + /* First collect pending bits from HW */ + GLUE(X_PFX,ack_pending)(xc); + + /* + * Cleanup the old-style bits if needed (they may have been + * set by pull or an escalation interrupts). + */ + if (test_bit(BOOK3S_IRQPRIO_EXTERNAL, &vcpu->arch.pending_exceptions)) + clear_bit(BOOK3S_IRQPRIO_EXTERNAL_LEVEL, + &vcpu->arch.pending_exceptions); + + pr_devel(" new pending=0x%02x hw_cppr=%d cppr=%d\n", + xc->pending, xc->hw_cppr, xc->cppr); + + /* Grab previous CPPR and reverse map it */ + old_cppr = xive_prio_to_guest(xc->cppr); + + /* Scan for actual interrupts */ + hirq = GLUE(X_PFX,scan_interrupts)(xc, xc->pending, scan_fetch); + + pr_devel(" got hirq=0x%x hw_cppr=%d cppr=%d\n", + hirq, xc->hw_cppr, xc->cppr); + +#ifdef XIVE_RUNTIME_CHECKS + /* That should never hit */ + if (hirq & 0xff000000) + pr_warn("XIVE: Weird guest interrupt number 0x%08x\n", hirq); +#endif + + /* + * XXX We could check if the interrupt is masked here and + * filter it. If we chose to do so, we would need to do: + * + * if (masked) { + * lock(); + * if (masked) { + * old_Q = true; + * hirq = 0; + * } + * unlock(); + * } + */ + + /* Return interrupt and old CPPR in GPR4 */ + vcpu->arch.gpr[4] = hirq | (old_cppr << 24); + + return H_SUCCESS; +} + +X_STATIC unsigned long GLUE(X_PFX,h_ipoll)(struct kvm_vcpu *vcpu, unsigned long server) +{ + struct kvmppc_xive_vcpu *xc = vcpu->arch.xive_vcpu; + u8 pending = xc->pending; + u32 hirq; + u8 pipr; + + pr_devel("H_IPOLL(server=%ld)\n", server); + + xc->GLUE(X_STAT_PFX,h_ipoll)++; + + /* Grab the target VCPU if not the current one */ + if (xc->server_num != server) { + vcpu = kvmppc_xive_find_server(vcpu->kvm, server); + if (!vcpu) + return H_PARAMETER; + xc = vcpu->arch.xive_vcpu; + + /* Scan all priorities */ + pending = 0xff; + } else { + /* Grab pending interrupt if any */ + pipr = __x_readb(__x_tima + TM_QW1_OS + TM_PIPR); + if (pipr < 8) + pending |= 1 << pipr; + } + + hirq = GLUE(X_PFX,scan_interrupts)(xc, pending, scan_poll); + + /* Return interrupt and old CPPR in GPR4 */ + vcpu->arch.gpr[4] = hirq | (xc->cppr << 24); + + return H_SUCCESS; +} + +static void GLUE(X_PFX,push_pending_to_hw)(struct kvmppc_xive_vcpu *xc) +{ + u8 pending, prio; + + pending = xc->pending; + if (xc->mfrr != 0xff) { + if (xc->mfrr < 8) + pending |= 1 << xc->mfrr; + else + pending |= 0x80; + } + if (!pending) + return; + prio = ffs(pending) - 1; + + __x_writeb(prio, __x_tima + TM_SPC_SET_OS_PENDING); +} + +X_STATIC int GLUE(X_PFX,h_cppr)(struct kvm_vcpu *vcpu, unsigned long cppr) +{ + struct kvmppc_xive_vcpu *xc = vcpu->arch.xive_vcpu; + u8 old_cppr; + + pr_devel("H_CPPR(cppr=%ld)\n", cppr); + + xc->GLUE(X_STAT_PFX,h_cppr)++; + + /* Map CPPR */ + cppr = xive_prio_from_guest(cppr); + + /* Remember old and update SW state */ + old_cppr = xc->cppr; + xc->cppr = cppr; + + /* + * We are masking less, we need to look for pending things + * to deliver and set VP pending bits accordingly to trigger + * a new interrupt otherwise we might miss MFRR changes for + * which we have optimized out sending an IPI signal. + */ + if (cppr > old_cppr) + GLUE(X_PFX,push_pending_to_hw)(xc); + + /* Apply new CPPR */ + xc->hw_cppr = cppr; + __x_writeb(cppr, __x_tima + TM_QW1_OS + TM_CPPR); + + return H_SUCCESS; +} + +X_STATIC int GLUE(X_PFX,h_eoi)(struct kvm_vcpu *vcpu, unsigned long xirr) +{ + struct kvmppc_xive *xive = vcpu->kvm->arch.xive; + struct kvmppc_xive_src_block *sb; + struct kvmppc_xive_irq_state *state; + struct kvmppc_xive_vcpu *xc = vcpu->arch.xive_vcpu; + struct xive_irq_data *xd; + u8 new_cppr = xirr >> 24; + u32 irq = xirr & 0x00ffffff, hw_num; + u16 src; + int rc = 0; + + pr_devel("H_EOI(xirr=%08lx)\n", xirr); + + xc->GLUE(X_STAT_PFX,h_eoi)++; + + xc->cppr = xive_prio_from_guest(new_cppr); + + /* + * IPIs are synthetized from MFRR and thus don't need + * any special EOI handling. The underlying interrupt + * used to signal MFRR changes is EOId when fetched from + * the queue. + */ + if (irq == XICS_IPI || irq == 0) + goto bail; + + /* Find interrupt source */ + sb = kvmppc_xive_find_source(xive, irq, &src); + if (!sb) { + pr_devel(" source not found !\n"); + rc = H_PARAMETER; + goto bail; + } + state = &sb->irq_state[src]; + kvmppc_xive_select_irq(state, &hw_num, &xd); + + state->in_eoi = true; + mb(); + +again: + if (state->guest_priority == MASKED) { + arch_spin_lock(&sb->lock); + if (state->guest_priority != MASKED) { + arch_spin_unlock(&sb->lock); + goto again; + } + pr_devel(" EOI on saved P...\n"); + + /* Clear old_p, that will cause unmask to perform an EOI */ + state->old_p = false; + + arch_spin_unlock(&sb->lock); + } else { + pr_devel(" EOI on source...\n"); + + /* Perform EOI on the source */ + GLUE(X_PFX,source_eoi)(hw_num, xd); + + /* If it's an emulated LSI, check level and resend */ + if (state->lsi && state->asserted) + __x_writeq(0, __x_trig_page(xd)); + + } + + mb(); + state->in_eoi = false; +bail: + + /* Re-evaluate pending IRQs and update HW */ + GLUE(X_PFX,scan_interrupts)(xc, xc->pending, scan_eoi); + GLUE(X_PFX,push_pending_to_hw)(xc); + pr_devel(" after scan pending=%02x\n", xc->pending); + + /* Apply new CPPR */ + xc->hw_cppr = xc->cppr; + __x_writeb(xc->cppr, __x_tima + TM_QW1_OS + TM_CPPR); + + return rc; +} + +X_STATIC int GLUE(X_PFX,h_ipi)(struct kvm_vcpu *vcpu, unsigned long server, + unsigned long mfrr) +{ + struct kvmppc_xive_vcpu *xc = vcpu->arch.xive_vcpu; + + pr_devel("H_IPI(server=%08lx,mfrr=%ld)\n", server, mfrr); + + xc->GLUE(X_STAT_PFX,h_ipi)++; + + /* Find target */ + vcpu = kvmppc_xive_find_server(vcpu->kvm, server); + if (!vcpu) + return H_PARAMETER; + xc = vcpu->arch.xive_vcpu; + + /* Locklessly write over MFRR */ + xc->mfrr = mfrr; + + /* Shoot the IPI if most favored than target cppr */ + if (mfrr < xc->cppr) + __x_writeq(0, __x_trig_page(&xc->vp_ipi_data)); + + return H_SUCCESS; +} diff --git a/arch/powerpc/kvm/irq.h b/arch/powerpc/kvm/irq.h index 5a9a10b90762..3f1be85a83bc 100644 --- a/arch/powerpc/kvm/irq.h +++ b/arch/powerpc/kvm/irq.h @@ -12,6 +12,7 @@ static inline int irqchip_in_kernel(struct kvm *kvm) #endif #ifdef CONFIG_KVM_XICS ret = ret || (kvm->arch.xics != NULL); + ret = ret || (kvm->arch.xive != NULL); #endif smp_rmb(); return ret; diff --git a/arch/powerpc/kvm/powerpc.c b/arch/powerpc/kvm/powerpc.c index 1ee22a910074..f7cf2cd564ef 100644 --- a/arch/powerpc/kvm/powerpc.c +++ b/arch/powerpc/kvm/powerpc.c @@ -38,6 +38,8 @@ #include <asm/irqflags.h> #include <asm/iommu.h> #include <asm/switch_to.h> +#include <asm/xive.h> + #include "timing.h" #include "irq.h" #include "../mm/mmu_decl.h" @@ -697,7 +699,10 @@ void kvm_arch_vcpu_free(struct kvm_vcpu *vcpu) kvmppc_mpic_disconnect_vcpu(vcpu->arch.mpic, vcpu); break; case KVMPPC_IRQ_XICS: - kvmppc_xics_free_icp(vcpu); + if (xive_enabled()) + kvmppc_xive_cleanup_vcpu(vcpu); + else + kvmppc_xics_free_icp(vcpu); break; } @@ -1522,8 +1527,12 @@ static int kvm_vcpu_ioctl_enable_cap(struct kvm_vcpu *vcpu, r = -EPERM; dev = kvm_device_from_filp(f.file); - if (dev) - r = kvmppc_xics_connect_vcpu(dev, vcpu, cap->args[1]); + if (dev) { + if (xive_enabled()) + r = kvmppc_xive_connect_vcpu(dev, vcpu, cap->args[1]); + else + r = kvmppc_xics_connect_vcpu(dev, vcpu, cap->args[1]); + } fdput(f); break; @@ -1547,7 +1556,7 @@ bool kvm_arch_intc_initialized(struct kvm *kvm) return true; #endif #ifdef CONFIG_KVM_XICS - if (kvm->arch.xics) + if (kvm->arch.xics || kvm->arch.xive) return true; #endif return false; diff --git a/arch/powerpc/platforms/powernv/opal.c b/arch/powerpc/platforms/powernv/opal.c index 7925a9d72cca..59684b4af4d1 100644 --- a/arch/powerpc/platforms/powernv/opal.c +++ b/arch/powerpc/platforms/powernv/opal.c @@ -967,3 +967,4 @@ EXPORT_SYMBOL_GPL(opal_leds_set_ind); EXPORT_SYMBOL_GPL(opal_write_oppanel_async); /* Export this for KVM */ EXPORT_SYMBOL_GPL(opal_int_set_mfrr); +EXPORT_SYMBOL_GPL(opal_int_eoi); diff --git a/arch/powerpc/sysdev/xive/common.c b/arch/powerpc/sysdev/xive/common.c index 6a98efb14264..913825086b8d 100644 --- a/arch/powerpc/sysdev/xive/common.c +++ b/arch/powerpc/sysdev/xive/common.c @@ -46,13 +46,15 @@ #endif bool __xive_enabled; +EXPORT_SYMBOL_GPL(__xive_enabled); bool xive_cmdline_disabled; /* We use only one priority for now */ static u8 xive_irq_priority; -/* TIMA */ +/* TIMA exported to KVM */ void __iomem *xive_tima; +EXPORT_SYMBOL_GPL(xive_tima); u32 xive_tima_offset; /* Backend ops */ @@ -345,8 +347,11 @@ static void xive_irq_eoi(struct irq_data *d) DBG_VERBOSE("eoi_irq: irq=%d [0x%lx] pending=%02x\n", d->irq, irqd_to_hwirq(d), xc->pending_prio); - /* EOI the source if it hasn't been disabled */ - if (!irqd_irq_disabled(d)) + /* + * EOI the source if it hasn't been disabled and hasn't + * been passed-through to a KVM guest + */ + if (!irqd_irq_disabled(d) && !irqd_is_forwarded_to_vcpu(d)) xive_do_source_eoi(irqd_to_hwirq(d), xd); /* @@ -689,9 +694,14 @@ static int xive_irq_set_affinity(struct irq_data *d, old_target = xd->target; - rc = xive_ops->configure_irq(hw_irq, - get_hard_smp_processor_id(target), - xive_irq_priority, d->irq); + /* + * Only configure the irq if it's not currently passed-through to + * a KVM guest + */ + if (!irqd_is_forwarded_to_vcpu(d)) + rc = xive_ops->configure_irq(hw_irq, + get_hard_smp_processor_id(target), + xive_irq_priority, d->irq); if (rc < 0) { pr_err("Error %d reconfiguring irq %d\n", rc, d->irq); return rc; @@ -771,6 +781,123 @@ static int xive_irq_retrigger(struct irq_data *d) return 1; } +static int xive_irq_set_vcpu_affinity(struct irq_data *d, void *state) +{ + struct xive_irq_data *xd = irq_data_get_irq_handler_data(d); + unsigned int hw_irq = (unsigned int)irqd_to_hwirq(d); + int rc; + u8 pq; + + /* + * We only support this on interrupts that do not require + * firmware calls for masking and unmasking + */ + if (xd->flags & XIVE_IRQ_FLAG_MASK_FW) + return -EIO; + + /* + * This is called by KVM with state non-NULL for enabling + * pass-through or NULL for disabling it + */ + if (state) { + irqd_set_forwarded_to_vcpu(d); + + /* Set it to PQ=10 state to prevent further sends */ + pq = xive_poke_esb(xd, XIVE_ESB_SET_PQ_10); + + /* No target ? nothing to do */ + if (xd->target == XIVE_INVALID_TARGET) { + /* + * An untargetted interrupt should have been + * also masked at the source + */ + WARN_ON(pq & 2); + + return 0; + } + + /* + * If P was set, adjust state to PQ=11 to indicate + * that a resend is needed for the interrupt to reach + * the guest. Also remember the value of P. + * + * This also tells us that it's in flight to a host queue + * or has already been fetched but hasn't been EOIed yet + * by the host. This it's potentially using up a host + * queue slot. This is important to know because as long + * as this is the case, we must not hard-unmask it when + * "returning" that interrupt to the host. + * + * This saved_p is cleared by the host EOI, when we know + * for sure the queue slot is no longer in use. + */ + if (pq & 2) { + pq = xive_poke_esb(xd, XIVE_ESB_SET_PQ_11); + xd->saved_p = true; + + /* + * Sync the XIVE source HW to ensure the interrupt + * has gone through the EAS before we change its + * target to the guest. That should guarantee us + * that we *will* eventually get an EOI for it on + * the host. Otherwise there would be a small window + * for P to be seen here but the interrupt going + * to the guest queue. + */ + if (xive_ops->sync_source) + xive_ops->sync_source(hw_irq); + } else + xd->saved_p = false; + } else { + irqd_clr_forwarded_to_vcpu(d); + + /* No host target ? hard mask and return */ + if (xd->target == XIVE_INVALID_TARGET) { + xive_do_source_set_mask(xd, true); + return 0; + } + + /* + * Sync the XIVE source HW to ensure the interrupt + * has gone through the EAS before we change its + * target to the host. + */ + if (xive_ops->sync_source) + xive_ops->sync_source(hw_irq); + + /* + * By convention we are called with the interrupt in + * a PQ=10 or PQ=11 state, ie, it won't fire and will + * have latched in Q whether there's a pending HW + * interrupt or not. + * + * First reconfigure the target. + */ + rc = xive_ops->configure_irq(hw_irq, + get_hard_smp_processor_id(xd->target), + xive_irq_priority, d->irq); + if (rc) + return rc; + + /* + * Then if saved_p is not set, effectively re-enable the + * interrupt with an EOI. If it is set, we know there is + * still a message in a host queue somewhere that will be + * EOId eventually. + * + * Note: We don't check irqd_irq_disabled(). Effectively, + * we *will* let the irq get through even if masked if the + * HW is still firing it in order to deal with the whole + * saved_p business properly. If the interrupt triggers + * while masked, the generic code will re-mask it anyway. + */ + if (!xd->saved_p) + xive_do_source_eoi(hw_irq, xd); + + } + return 0; +} + static struct irq_chip xive_irq_chip = { .name = "XIVE-IRQ", .irq_startup = xive_irq_startup, @@ -781,12 +908,14 @@ static struct irq_chip xive_irq_chip = { .irq_set_affinity = xive_irq_set_affinity, .irq_set_type = xive_irq_set_type, .irq_retrigger = xive_irq_retrigger, + .irq_set_vcpu_affinity = xive_irq_set_vcpu_affinity, }; bool is_xive_irq(struct irq_chip *chip) { return chip == &xive_irq_chip; } +EXPORT_SYMBOL_GPL(is_xive_irq); void xive_cleanup_irq_data(struct xive_irq_data *xd) { @@ -801,6 +930,7 @@ void xive_cleanup_irq_data(struct xive_irq_data *xd) xd->trig_mmio = NULL; } } +EXPORT_SYMBOL_GPL(xive_cleanup_irq_data); static int xive_irq_alloc_data(unsigned int virq, irq_hw_number_t hw) { diff --git a/arch/powerpc/sysdev/xive/native.c b/arch/powerpc/sysdev/xive/native.c index 1a726229a427..ab9ecce61ee5 100644 --- a/arch/powerpc/sysdev/xive/native.c +++ b/arch/powerpc/sysdev/xive/native.c @@ -31,6 +31,7 @@ #include <asm/xive.h> #include <asm/xive-regs.h> #include <asm/opal.h> +#include <asm/kvm_ppc.h> #include "xive-internal.h" @@ -95,6 +96,7 @@ int xive_native_populate_irq_data(u32 hw_irq, struct xive_irq_data *data) } return 0; } +EXPORT_SYMBOL_GPL(xive_native_populate_irq_data); int xive_native_configure_irq(u32 hw_irq, u32 target, u8 prio, u32 sw_irq) { @@ -108,6 +110,8 @@ int xive_native_configure_irq(u32 hw_irq, u32 target, u8 prio, u32 sw_irq) } return rc == 0 ? 0 : -ENXIO; } +EXPORT_SYMBOL_GPL(xive_native_configure_irq); + /* This can be called multiple time to change a queue configuration */ int xive_native_configure_queue(u32 vp_id, struct xive_q *q, u8 prio, @@ -172,6 +176,7 @@ int xive_native_configure_queue(u32 vp_id, struct xive_q *q, u8 prio, fail: return rc; } +EXPORT_SYMBOL_GPL(xive_native_configure_queue); static void __xive_native_disable_queue(u32 vp_id, struct xive_q *q, u8 prio) { @@ -192,6 +197,7 @@ void xive_native_disable_queue(u32 vp_id, struct xive_q *q, u8 prio) { __xive_native_disable_queue(vp_id, q, prio); } +EXPORT_SYMBOL_GPL(xive_native_disable_queue); static int xive_native_setup_queue(unsigned int cpu, struct xive_cpu *xc, u8 prio) { @@ -262,6 +268,7 @@ static int xive_native_get_ipi(unsigned int cpu, struct xive_cpu *xc) } return 0; } +#endif /* CONFIG_SMP */ u32 xive_native_alloc_irq(void) { @@ -277,6 +284,7 @@ u32 xive_native_alloc_irq(void) return 0; return rc; } +EXPORT_SYMBOL_GPL(xive_native_alloc_irq); void xive_native_free_irq(u32 irq) { @@ -287,7 +295,9 @@ void xive_native_free_irq(u32 irq) msleep(1); } } +EXPORT_SYMBOL_GPL(xive_native_free_irq); +#ifdef CONFIG_SMP static void xive_native_put_ipi(unsigned int cpu, struct xive_cpu *xc) { s64 rc; @@ -383,7 +393,7 @@ static void xive_native_setup_cpu(unsigned int cpu, struct xive_cpu *xc) return; /* Enable the pool VP */ - vp = xive_pool_vps + get_hard_smp_processor_id(cpu); + vp = xive_pool_vps + cpu; pr_debug("CPU %d setting up pool VP 0x%x\n", cpu, vp); for (;;) { rc = opal_xive_set_vp_info(vp, OPAL_XIVE_VP_ENABLED, 0); @@ -428,7 +438,7 @@ static void xive_native_teardown_cpu(unsigned int cpu, struct xive_cpu *xc) in_be64(xive_tima + TM_SPC_PULL_POOL_CTX); /* Disable it */ - vp = xive_pool_vps + get_hard_smp_processor_id(cpu); + vp = xive_pool_vps + cpu; for (;;) { rc = opal_xive_set_vp_info(vp, 0, 0); if (rc != OPAL_BUSY) @@ -437,10 +447,11 @@ static void xive_native_teardown_cpu(unsigned int cpu, struct xive_cpu *xc) } } -static void xive_native_sync_source(u32 hw_irq) +void xive_native_sync_source(u32 hw_irq) { opal_xive_sync(XIVE_SYNC_EAS, hw_irq); } +EXPORT_SYMBOL_GPL(xive_native_sync_source); static const struct xive_ops xive_native_ops = { .populate_irq_data = xive_native_populate_irq_data, @@ -501,10 +512,24 @@ static bool xive_parse_provisioning(struct device_node *np) return true; } +static void xive_native_setup_pools(void) +{ + /* Allocate a pool big enough */ + pr_debug("XIVE: Allocating VP block for pool size %d\n", nr_cpu_ids); + + xive_pool_vps = xive_native_alloc_vp_block(nr_cpu_ids); + if (WARN_ON(xive_pool_vps == XIVE_INVALID_VP)) + pr_err("XIVE: Failed to allocate pool VP, KVM might not function\n"); + + pr_debug("XIVE: Pool VPs allocated at 0x%x for %d max CPUs\n", + xive_pool_vps, nr_cpu_ids); +} + u32 xive_native_default_eq_shift(void) { return xive_queue_shift; } +EXPORT_SYMBOL_GPL(xive_native_default_eq_shift); bool xive_native_init(void) { @@ -514,7 +539,7 @@ bool xive_native_init(void) struct property *prop; u8 max_prio = 7; const __be32 *p; - u32 val; + u32 val, cpu; s64 rc; if (xive_cmdline_disabled) @@ -550,7 +575,11 @@ bool xive_native_init(void) break; } - /* Grab size of provisioning pages */ + /* Configure Thread Management areas for KVM */ + for_each_possible_cpu(cpu) + kvmppc_set_xive_tima(cpu, r.start, tima); + + /* Grab size of provisionning pages */ xive_parse_provisioning(np); /* Switch the XIVE to exploitation mode */ @@ -560,6 +589,9 @@ bool xive_native_init(void) return false; } + /* Setup some dummy HV pool VPs */ + xive_native_setup_pools(); + /* Initialize XIVE core with our backend */ if (!xive_core_init(&xive_native_ops, tima, TM_QW3_HV_PHYS, max_prio)) { @@ -638,3 +670,47 @@ void xive_native_free_vp_block(u32 vp_base) pr_warn("OPAL error %lld freeing VP block\n", rc); } EXPORT_SYMBOL_GPL(xive_native_free_vp_block); + +int xive_native_enable_vp(u32 vp_id) +{ + s64 rc; + + for (;;) { + rc = opal_xive_set_vp_info(vp_id, OPAL_XIVE_VP_ENABLED, 0); + if (rc != OPAL_BUSY) + break; + msleep(1); + } + return rc ? -EIO : 0; +} +EXPORT_SYMBOL_GPL(xive_native_enable_vp); + +int xive_native_disable_vp(u32 vp_id) +{ + s64 rc; + + for (;;) { + rc = opal_xive_set_vp_info(vp_id, 0, 0); + if (rc != OPAL_BUSY) + break; + msleep(1); + } + return rc ? -EIO : 0; +} +EXPORT_SYMBOL_GPL(xive_native_disable_vp); + +int xive_native_get_vp_info(u32 vp_id, u32 *out_cam_id, u32 *out_chip_id) +{ + __be64 vp_cam_be; + __be32 vp_chip_id_be; + s64 rc; + + rc = opal_xive_get_vp_info(vp_id, NULL, &vp_cam_be, NULL, &vp_chip_id_be); + if (rc) + return -EIO; + *out_cam_id = be64_to_cpu(vp_cam_be) & 0xffffffffu; + *out_chip_id = be32_to_cpu(vp_chip_id_be); + + return 0; +} +EXPORT_SYMBOL_GPL(xive_native_get_vp_info); diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index f5bddf92faba..9c761fea0c98 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -1020,6 +1020,8 @@ struct kvm_x86_ops { void (*enable_log_dirty_pt_masked)(struct kvm *kvm, struct kvm_memory_slot *slot, gfn_t offset, unsigned long mask); + int (*write_log_dirty)(struct kvm_vcpu *vcpu); + /* pmu operations of sub-arch */ const struct kvm_pmu_ops *pmu_ops; diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c index 558676538fca..5d3376f67794 100644 --- a/arch/x86/kvm/mmu.c +++ b/arch/x86/kvm/mmu.c @@ -1498,6 +1498,21 @@ void kvm_arch_mmu_enable_log_dirty_pt_masked(struct kvm *kvm, kvm_mmu_write_protect_pt_masked(kvm, slot, gfn_offset, mask); } +/** + * kvm_arch_write_log_dirty - emulate dirty page logging + * @vcpu: Guest mode vcpu + * + * Emulate arch specific page modification logging for the + * nested hypervisor + */ +int kvm_arch_write_log_dirty(struct kvm_vcpu *vcpu) +{ + if (kvm_x86_ops->write_log_dirty) + return kvm_x86_ops->write_log_dirty(vcpu); + + return 0; +} + bool kvm_mmu_slot_gfn_write_protect(struct kvm *kvm, struct kvm_memory_slot *slot, u64 gfn) { diff --git a/arch/x86/kvm/mmu.h b/arch/x86/kvm/mmu.h index d8ccb32f7308..27975807cc64 100644 --- a/arch/x86/kvm/mmu.h +++ b/arch/x86/kvm/mmu.h @@ -202,4 +202,5 @@ void kvm_mmu_gfn_disallow_lpage(struct kvm_memory_slot *slot, gfn_t gfn); void kvm_mmu_gfn_allow_lpage(struct kvm_memory_slot *slot, gfn_t gfn); bool kvm_mmu_slot_gfn_write_protect(struct kvm *kvm, struct kvm_memory_slot *slot, u64 gfn); +int kvm_arch_write_log_dirty(struct kvm_vcpu *vcpu); #endif diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h index 314d2071b337..56241746abbd 100644 --- a/arch/x86/kvm/paging_tmpl.h +++ b/arch/x86/kvm/paging_tmpl.h @@ -226,6 +226,10 @@ static int FNAME(update_accessed_dirty_bits)(struct kvm_vcpu *vcpu, if (level == walker->level && write_fault && !(pte & PT_GUEST_DIRTY_MASK)) { trace_kvm_mmu_set_dirty_bit(table_gfn, index, sizeof(pte)); +#if PTTYPE == PTTYPE_EPT + if (kvm_arch_write_log_dirty(vcpu)) + return -EINVAL; +#endif pte |= PT_GUEST_DIRTY_MASK; } if (pte == orig_pte) diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index c5fd459c4043..c6f4ad44aa95 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -248,6 +248,7 @@ struct __packed vmcs12 { u64 xss_exit_bitmap; u64 guest_physical_address; u64 vmcs_link_pointer; + u64 pml_address; u64 guest_ia32_debugctl; u64 guest_ia32_pat; u64 guest_ia32_efer; @@ -369,6 +370,7 @@ struct __packed vmcs12 { u16 guest_ldtr_selector; u16 guest_tr_selector; u16 guest_intr_status; + u16 guest_pml_index; u16 host_es_selector; u16 host_cs_selector; u16 host_ss_selector; @@ -407,6 +409,7 @@ struct nested_vmx { /* Has the level1 guest done vmxon? */ bool vmxon; gpa_t vmxon_ptr; + bool pml_full; /* The guest-physical address of the current VMCS L1 keeps for L2 */ gpa_t current_vmptr; @@ -742,6 +745,7 @@ static const unsigned short vmcs_field_to_offset_table[] = { FIELD(GUEST_LDTR_SELECTOR, guest_ldtr_selector), FIELD(GUEST_TR_SELECTOR, guest_tr_selector), FIELD(GUEST_INTR_STATUS, guest_intr_status), + FIELD(GUEST_PML_INDEX, guest_pml_index), FIELD(HOST_ES_SELECTOR, host_es_selector), FIELD(HOST_CS_SELECTOR, host_cs_selector), FIELD(HOST_SS_SELECTOR, host_ss_selector), @@ -767,6 +771,7 @@ static const unsigned short vmcs_field_to_offset_table[] = { FIELD64(XSS_EXIT_BITMAP, xss_exit_bitmap), FIELD64(GUEST_PHYSICAL_ADDRESS, guest_physical_address), FIELD64(VMCS_LINK_POINTER, vmcs_link_pointer), + FIELD64(PML_ADDRESS, pml_address), FIELD64(GUEST_IA32_DEBUGCTL, guest_ia32_debugctl), FIELD64(GUEST_IA32_PAT, guest_ia32_pat), FIELD64(GUEST_IA32_EFER, guest_ia32_efer), @@ -1314,6 +1319,11 @@ static inline bool report_flexpriority(void) return flexpriority_enabled; } +static inline unsigned nested_cpu_vmx_misc_cr3_count(struct kvm_vcpu *vcpu) +{ + return vmx_misc_cr3_count(to_vmx(vcpu)->nested.nested_vmx_misc_low); +} + static inline bool nested_cpu_has(struct vmcs12 *vmcs12, u32 bit) { return vmcs12->cpu_based_vm_exec_control & bit; @@ -1348,6 +1358,11 @@ static inline bool nested_cpu_has_xsaves(struct vmcs12 *vmcs12) vmx_xsaves_supported(); } +static inline bool nested_cpu_has_pml(struct vmcs12 *vmcs12) +{ + return nested_cpu_has2(vmcs12, SECONDARY_EXEC_ENABLE_PML); +} + static inline bool nested_cpu_has_virt_x2apic_mode(struct vmcs12 *vmcs12) { return nested_cpu_has2(vmcs12, SECONDARY_EXEC_VIRTUALIZE_X2APIC_MODE); @@ -2751,8 +2766,11 @@ static void nested_vmx_setup_ctls_msrs(struct vcpu_vmx *vmx) vmx->nested.nested_vmx_ept_caps |= VMX_EPT_EXTENT_GLOBAL_BIT | VMX_EPT_EXTENT_CONTEXT_BIT | VMX_EPT_2MB_PAGE_BIT | VMX_EPT_1GB_PAGE_BIT; - if (enable_ept_ad_bits) + if (enable_ept_ad_bits) { + vmx->nested.nested_vmx_secondary_ctls_high |= + SECONDARY_EXEC_ENABLE_PML; vmx->nested.nested_vmx_ept_caps |= VMX_EPT_AD_BIT; + } } else vmx->nested.nested_vmx_ept_caps = 0; @@ -8114,7 +8132,7 @@ static bool nested_vmx_exit_handled(struct kvm_vcpu *vcpu) case EXIT_REASON_PREEMPTION_TIMER: return false; case EXIT_REASON_PML_FULL: - /* We don't expose PML support to L1. */ + /* We emulate PML support to L1. */ return false; default: return true; @@ -9364,13 +9382,20 @@ static void nested_ept_inject_page_fault(struct kvm_vcpu *vcpu, struct x86_exception *fault) { struct vmcs12 *vmcs12 = get_vmcs12(vcpu); + struct vcpu_vmx *vmx = to_vmx(vcpu); u32 exit_reason; + unsigned long exit_qualification = vcpu->arch.exit_qualification; - if (fault->error_code & PFERR_RSVD_MASK) + if (vmx->nested.pml_full) { + exit_reason = EXIT_REASON_PML_FULL; + vmx->nested.pml_full = false; + exit_qualification &= INTR_INFO_UNBLOCK_NMI; + } else if (fault->error_code & PFERR_RSVD_MASK) exit_reason = EXIT_REASON_EPT_MISCONFIG; else exit_reason = EXIT_REASON_EPT_VIOLATION; - nested_vmx_vmexit(vcpu, exit_reason, 0, vcpu->arch.exit_qualification); + + nested_vmx_vmexit(vcpu, exit_reason, 0, exit_qualification); vmcs12->guest_physical_address = fault->address; } @@ -9713,6 +9738,22 @@ static int nested_vmx_check_msr_switch_controls(struct kvm_vcpu *vcpu, return 0; } +static int nested_vmx_check_pml_controls(struct kvm_vcpu *vcpu, + struct vmcs12 *vmcs12) +{ + u64 address = vmcs12->pml_address; + int maxphyaddr = cpuid_maxphyaddr(vcpu); + + if (nested_cpu_has2(vmcs12, SECONDARY_EXEC_ENABLE_PML)) { + if (!nested_cpu_has_ept(vmcs12) || + !IS_ALIGNED(address, 4096) || + address >> maxphyaddr) + return -EINVAL; + } + + return 0; +} + static int nested_vmx_msr_check_common(struct kvm_vcpu *vcpu, struct vmx_msr_entry *e) { @@ -9886,7 +9927,7 @@ static int prepare_vmcs02(struct kvm_vcpu *vcpu, struct vmcs12 *vmcs12, bool from_vmentry, u32 *entry_failure_code) { struct vcpu_vmx *vmx = to_vmx(vcpu); - u32 exec_control; + u32 exec_control, vmcs12_exec_ctrl; vmcs_write16(GUEST_ES_SELECTOR, vmcs12->guest_es_selector); vmcs_write16(GUEST_CS_SELECTOR, vmcs12->guest_cs_selector); @@ -10017,8 +10058,11 @@ static int prepare_vmcs02(struct kvm_vcpu *vcpu, struct vmcs12 *vmcs12, SECONDARY_EXEC_VIRTUAL_INTR_DELIVERY | SECONDARY_EXEC_APIC_REGISTER_VIRT); if (nested_cpu_has(vmcs12, - CPU_BASED_ACTIVATE_SECONDARY_CONTROLS)) - exec_control |= vmcs12->secondary_vm_exec_control; + CPU_BASED_ACTIVATE_SECONDARY_CONTROLS)) { + vmcs12_exec_ctrl = vmcs12->secondary_vm_exec_control & + ~SECONDARY_EXEC_ENABLE_PML; + exec_control |= vmcs12_exec_ctrl; + } if (exec_control & SECONDARY_EXEC_VIRTUAL_INTR_DELIVERY) { vmcs_write64(EOI_EXIT_BITMAP0, @@ -10248,6 +10292,9 @@ static int check_vmentry_prereqs(struct kvm_vcpu *vcpu, struct vmcs12 *vmcs12) if (nested_vmx_check_msr_switch_controls(vcpu, vmcs12)) return VMXERR_ENTRY_INVALID_CONTROL_FIELD; + if (nested_vmx_check_pml_controls(vcpu, vmcs12)) + return VMXERR_ENTRY_INVALID_CONTROL_FIELD; + if (!vmx_control_verify(vmcs12->cpu_based_vm_exec_control, vmx->nested.nested_vmx_procbased_ctls_low, vmx->nested.nested_vmx_procbased_ctls_high) || @@ -10266,6 +10313,9 @@ static int check_vmentry_prereqs(struct kvm_vcpu *vcpu, struct vmcs12 *vmcs12) vmx->nested.nested_vmx_entry_ctls_high)) return VMXERR_ENTRY_INVALID_CONTROL_FIELD; + if (vmcs12->cr3_target_count > nested_cpu_vmx_misc_cr3_count(vcpu)) + return VMXERR_ENTRY_INVALID_CONTROL_FIELD; + if (!nested_host_cr0_valid(vcpu, vmcs12->host_cr0) || !nested_host_cr4_valid(vcpu, vmcs12->host_cr4) || !nested_cr3_valid(vcpu, vmcs12->host_cr3)) @@ -11143,6 +11193,46 @@ static void vmx_flush_log_dirty(struct kvm *kvm) kvm_flush_pml_buffers(kvm); } +static int vmx_write_pml_buffer(struct kvm_vcpu *vcpu) +{ + struct vmcs12 *vmcs12; + struct vcpu_vmx *vmx = to_vmx(vcpu); + gpa_t gpa; + struct page *page = NULL; + u64 *pml_address; + + if (is_guest_mode(vcpu)) { + WARN_ON_ONCE(vmx->nested.pml_full); + + /* + * Check if PML is enabled for the nested guest. + * Whether eptp bit 6 is set is already checked + * as part of A/D emulation. + */ + vmcs12 = get_vmcs12(vcpu); + if (!nested_cpu_has_pml(vmcs12)) + return 0; + + if (vmcs12->guest_pml_index > PML_ENTITY_NUM) { + vmx->nested.pml_full = true; + return 1; + } + + gpa = vmcs_read64(GUEST_PHYSICAL_ADDRESS) & ~0xFFFull; + + page = nested_get_page(vcpu, vmcs12->pml_address); + if (!page) + return 0; + + pml_address = kmap(page); + pml_address[vmcs12->guest_pml_index--] = gpa; + kunmap(page); + nested_release_page_clean(page); + } + + return 0; +} + static void vmx_enable_log_dirty_pt_masked(struct kvm *kvm, struct kvm_memory_slot *memslot, gfn_t offset, unsigned long mask) @@ -11502,6 +11592,7 @@ static struct kvm_x86_ops vmx_x86_ops __ro_after_init = { .slot_disable_log_dirty = vmx_slot_disable_log_dirty, .flush_log_dirty = vmx_flush_log_dirty, .enable_log_dirty_pt_masked = vmx_enable_log_dirty_pt_masked, + .write_log_dirty = vmx_write_pml_buffer, .pre_block = vmx_pre_block, .post_block = vmx_post_block, diff --git a/include/kvm/arm_vgic.h b/include/kvm/arm_vgic.h index 581a59ea7e34..97b8d3728b31 100644 --- a/include/kvm/arm_vgic.h +++ b/include/kvm/arm_vgic.h @@ -148,7 +148,6 @@ struct vgic_its { gpa_t vgic_its_base; bool enabled; - bool initialized; struct vgic_io_device iodev; struct kvm_device *dev; @@ -162,6 +161,9 @@ struct vgic_its { u32 creadr; u32 cwriter; + /* migration ABI revision in use */ + u32 abi_rev; + /* Protects the device and collection lists */ struct mutex its_lock; struct list_head device_list; @@ -283,6 +285,7 @@ extern struct static_key_false vgic_v2_cpuif_trap; int kvm_vgic_addr(struct kvm *kvm, unsigned long type, u64 *addr, bool write); void kvm_vgic_early_init(struct kvm *kvm); +int kvm_vgic_vcpu_init(struct kvm_vcpu *vcpu); int kvm_vgic_create(struct kvm *kvm, u32 type); void kvm_vgic_destroy(struct kvm *kvm); void kvm_vgic_vcpu_early_init(struct kvm_vcpu *vcpu); diff --git a/include/linux/irqchip/arm-gic-v3.h b/include/linux/irqchip/arm-gic-v3.h index 97cbca19430d..fffb91202bc9 100644 --- a/include/linux/irqchip/arm-gic-v3.h +++ b/include/linux/irqchip/arm-gic-v3.h @@ -132,6 +132,9 @@ #define GIC_BASER_SHAREABILITY(reg, type) \ (GIC_BASER_##type << reg##_SHAREABILITY_SHIFT) +/* encode a size field of width @w containing @n - 1 units */ +#define GIC_ENCODE_SZ(n, w) (((unsigned long)(n) - 1) & GENMASK_ULL(((w) - 1), 0)) + #define GICR_PROPBASER_SHAREABILITY_SHIFT (10) #define GICR_PROPBASER_INNER_CACHEABILITY_SHIFT (7) #define GICR_PROPBASER_OUTER_CACHEABILITY_SHIFT (56) @@ -156,6 +159,8 @@ #define GICR_PROPBASER_RaWaWb GIC_BASER_CACHEABILITY(GICR_PROPBASER, INNER, RaWaWb) #define GICR_PROPBASER_IDBITS_MASK (0x1f) +#define GICR_PROPBASER_ADDRESS(x) ((x) & GENMASK_ULL(51, 12)) +#define GICR_PENDBASER_ADDRESS(x) ((x) & GENMASK_ULL(51, 16)) #define GICR_PENDBASER_SHAREABILITY_SHIFT (10) #define GICR_PENDBASER_INNER_CACHEABILITY_SHIFT (7) @@ -232,12 +237,18 @@ #define GITS_CTLR_QUIESCENT (1U << 31) #define GITS_TYPER_PLPIS (1UL << 0) +#define GITS_TYPER_ITT_ENTRY_SIZE_SHIFT 4 #define GITS_TYPER_IDBITS_SHIFT 8 #define GITS_TYPER_DEVBITS_SHIFT 13 #define GITS_TYPER_DEVBITS(r) ((((r) >> GITS_TYPER_DEVBITS_SHIFT) & 0x1f) + 1) #define GITS_TYPER_PTA (1UL << 19) #define GITS_TYPER_HWCOLLCNT_SHIFT 24 +#define GITS_IIDR_REV_SHIFT 12 +#define GITS_IIDR_REV_MASK (0xf << GITS_IIDR_REV_SHIFT) +#define GITS_IIDR_REV(r) (((r) >> GITS_IIDR_REV_SHIFT) & 0xf) +#define GITS_IIDR_PRODUCTID_SHIFT 24 + #define GITS_CBASER_VALID (1ULL << 63) #define GITS_CBASER_SHAREABILITY_SHIFT (10) #define GITS_CBASER_INNER_CACHEABILITY_SHIFT (59) @@ -290,6 +301,7 @@ #define GITS_BASER_TYPE(r) (((r) >> GITS_BASER_TYPE_SHIFT) & 7) #define GITS_BASER_ENTRY_SIZE_SHIFT (48) #define GITS_BASER_ENTRY_SIZE(r) ((((r) >> GITS_BASER_ENTRY_SIZE_SHIFT) & 0x1f) + 1) +#define GITS_BASER_ENTRY_SIZE_MASK GENMASK_ULL(52, 48) #define GITS_BASER_SHAREABILITY_SHIFT (10) #define GITS_BASER_InnerShareable \ GIC_BASER_SHAREABILITY(GITS_BASER, InnerShareable) @@ -337,9 +349,11 @@ #define E_ITS_INT_UNMAPPED_INTERRUPT 0x010307 #define E_ITS_CLEAR_UNMAPPED_INTERRUPT 0x010507 #define E_ITS_MAPD_DEVICE_OOR 0x010801 +#define E_ITS_MAPD_ITTSIZE_OOR 0x010802 #define E_ITS_MAPC_PROCNUM_OOR 0x010902 #define E_ITS_MAPC_COLLECTION_OOR 0x010903 #define E_ITS_MAPTI_UNMAPPED_DEVICE 0x010a04 +#define E_ITS_MAPTI_ID_OOR 0x010a05 #define E_ITS_MAPTI_PHYSICALID_OOR 0x010a06 #define E_ITS_INV_UNMAPPED_INTERRUPT 0x010c07 #define E_ITS_INVALL_UNMAPPED_COLLECTION 0x010d09 diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h index 2b12b2683359..8c0664309815 100644 --- a/include/linux/kvm_host.h +++ b/include/linux/kvm_host.h @@ -499,6 +499,17 @@ static inline struct kvm_vcpu *kvm_get_vcpu_by_id(struct kvm *kvm, int id) return NULL; } +static inline int kvm_vcpu_get_idx(struct kvm_vcpu *vcpu) +{ + struct kvm_vcpu *tmp; + int idx; + + kvm_for_each_vcpu(idx, tmp, vcpu->kvm) + if (tmp == vcpu) + return idx; + BUG(); +} + #define kvm_for_each_memslot(memslot, slots) \ for (memslot = &slots->memslots[0]; \ memslot < slots->memslots + KVM_MEM_SLOTS_NUM && memslot->npages;\ @@ -1167,7 +1178,6 @@ int kvm_register_device_ops(struct kvm_device_ops *ops, u32 type); void kvm_unregister_device_ops(u32 type); extern struct kvm_device_ops kvm_mpic_ops; -extern struct kvm_device_ops kvm_xics_ops; extern struct kvm_device_ops kvm_arm_vgic_v2_ops; extern struct kvm_device_ops kvm_arm_vgic_v3_ops; diff --git a/arch/arm/kvm/arm.c b/virt/kvm/arm/arm.c index 8a31906bdc9b..3417e184c8e1 100644 --- a/arch/arm/kvm/arm.c +++ b/virt/kvm/arm/arm.c @@ -332,7 +332,7 @@ int kvm_arch_vcpu_init(struct kvm_vcpu *vcpu) kvm_arm_reset_debug_ptr(vcpu); - return 0; + return kvm_vgic_vcpu_init(vcpu); } void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu) diff --git a/arch/arm/kvm/mmio.c b/virt/kvm/arm/mmio.c index b6e715fd3c90..b6e715fd3c90 100644 --- a/arch/arm/kvm/mmio.c +++ b/virt/kvm/arm/mmio.c diff --git a/arch/arm/kvm/mmu.c b/virt/kvm/arm/mmu.c index 313ee646480f..313ee646480f 100644 --- a/arch/arm/kvm/mmu.c +++ b/virt/kvm/arm/mmu.c diff --git a/arch/arm/kvm/perf.c b/virt/kvm/arm/perf.c index 1a3849da0b4b..1a3849da0b4b 100644 --- a/arch/arm/kvm/perf.c +++ b/virt/kvm/arm/perf.c diff --git a/arch/arm/kvm/psci.c b/virt/kvm/arm/psci.c index a08d7a93aebb..a08d7a93aebb 100644 --- a/arch/arm/kvm/psci.c +++ b/virt/kvm/arm/psci.c diff --git a/virt/kvm/arm/trace.h b/virt/kvm/arm/trace.h index 37d8b98867d5..f7dc5ddd6847 100644 --- a/virt/kvm/arm/trace.h +++ b/virt/kvm/arm/trace.h @@ -7,26 +7,250 @@ #define TRACE_SYSTEM kvm /* - * Tracepoints for vgic + * Tracepoints for entry/exit to guest */ -TRACE_EVENT(vgic_update_irq_pending, - TP_PROTO(unsigned long vcpu_id, __u32 irq, bool level), - TP_ARGS(vcpu_id, irq, level), +TRACE_EVENT(kvm_entry, + TP_PROTO(unsigned long vcpu_pc), + TP_ARGS(vcpu_pc), TP_STRUCT__entry( - __field( unsigned long, vcpu_id ) - __field( __u32, irq ) - __field( bool, level ) + __field( unsigned long, vcpu_pc ) ), TP_fast_assign( - __entry->vcpu_id = vcpu_id; - __entry->irq = irq; + __entry->vcpu_pc = vcpu_pc; + ), + + TP_printk("PC: 0x%08lx", __entry->vcpu_pc) +); + +TRACE_EVENT(kvm_exit, + TP_PROTO(int idx, unsigned int exit_reason, unsigned long vcpu_pc), + TP_ARGS(idx, exit_reason, vcpu_pc), + + TP_STRUCT__entry( + __field( int, idx ) + __field( unsigned int, exit_reason ) + __field( unsigned long, vcpu_pc ) + ), + + TP_fast_assign( + __entry->idx = idx; + __entry->exit_reason = exit_reason; + __entry->vcpu_pc = vcpu_pc; + ), + + TP_printk("%s: HSR_EC: 0x%04x (%s), PC: 0x%08lx", + __print_symbolic(__entry->idx, kvm_arm_exception_type), + __entry->exit_reason, + __print_symbolic(__entry->exit_reason, kvm_arm_exception_class), + __entry->vcpu_pc) +); + +TRACE_EVENT(kvm_guest_fault, + TP_PROTO(unsigned long vcpu_pc, unsigned long hsr, + unsigned long hxfar, + unsigned long long ipa), + TP_ARGS(vcpu_pc, hsr, hxfar, ipa), + + TP_STRUCT__entry( + __field( unsigned long, vcpu_pc ) + __field( unsigned long, hsr ) + __field( unsigned long, hxfar ) + __field( unsigned long long, ipa ) + ), + + TP_fast_assign( + __entry->vcpu_pc = vcpu_pc; + __entry->hsr = hsr; + __entry->hxfar = hxfar; + __entry->ipa = ipa; + ), + + TP_printk("ipa %#llx, hsr %#08lx, hxfar %#08lx, pc %#08lx", + __entry->ipa, __entry->hsr, + __entry->hxfar, __entry->vcpu_pc) +); + +TRACE_EVENT(kvm_access_fault, + TP_PROTO(unsigned long ipa), + TP_ARGS(ipa), + + TP_STRUCT__entry( + __field( unsigned long, ipa ) + ), + + TP_fast_assign( + __entry->ipa = ipa; + ), + + TP_printk("IPA: %lx", __entry->ipa) +); + +TRACE_EVENT(kvm_irq_line, + TP_PROTO(unsigned int type, int vcpu_idx, int irq_num, int level), + TP_ARGS(type, vcpu_idx, irq_num, level), + + TP_STRUCT__entry( + __field( unsigned int, type ) + __field( int, vcpu_idx ) + __field( int, irq_num ) + __field( int, level ) + ), + + TP_fast_assign( + __entry->type = type; + __entry->vcpu_idx = vcpu_idx; + __entry->irq_num = irq_num; __entry->level = level; ), - TP_printk("VCPU: %ld, IRQ %d, level: %d", - __entry->vcpu_id, __entry->irq, __entry->level) + TP_printk("Inject %s interrupt (%d), vcpu->idx: %d, num: %d, level: %d", + (__entry->type == KVM_ARM_IRQ_TYPE_CPU) ? "CPU" : + (__entry->type == KVM_ARM_IRQ_TYPE_PPI) ? "VGIC PPI" : + (__entry->type == KVM_ARM_IRQ_TYPE_SPI) ? "VGIC SPI" : "UNKNOWN", + __entry->type, __entry->vcpu_idx, __entry->irq_num, __entry->level) +); + +TRACE_EVENT(kvm_mmio_emulate, + TP_PROTO(unsigned long vcpu_pc, unsigned long instr, + unsigned long cpsr), + TP_ARGS(vcpu_pc, instr, cpsr), + + TP_STRUCT__entry( + __field( unsigned long, vcpu_pc ) + __field( unsigned long, instr ) + __field( unsigned long, cpsr ) + ), + + TP_fast_assign( + __entry->vcpu_pc = vcpu_pc; + __entry->instr = instr; + __entry->cpsr = cpsr; + ), + + TP_printk("Emulate MMIO at: 0x%08lx (instr: %08lx, cpsr: %08lx)", + __entry->vcpu_pc, __entry->instr, __entry->cpsr) +); + +TRACE_EVENT(kvm_unmap_hva, + TP_PROTO(unsigned long hva), + TP_ARGS(hva), + + TP_STRUCT__entry( + __field( unsigned long, hva ) + ), + + TP_fast_assign( + __entry->hva = hva; + ), + + TP_printk("mmu notifier unmap hva: %#08lx", __entry->hva) +); + +TRACE_EVENT(kvm_unmap_hva_range, + TP_PROTO(unsigned long start, unsigned long end), + TP_ARGS(start, end), + + TP_STRUCT__entry( + __field( unsigned long, start ) + __field( unsigned long, end ) + ), + + TP_fast_assign( + __entry->start = start; + __entry->end = end; + ), + + TP_printk("mmu notifier unmap range: %#08lx -- %#08lx", + __entry->start, __entry->end) +); + +TRACE_EVENT(kvm_set_spte_hva, + TP_PROTO(unsigned long hva), + TP_ARGS(hva), + + TP_STRUCT__entry( + __field( unsigned long, hva ) + ), + + TP_fast_assign( + __entry->hva = hva; + ), + + TP_printk("mmu notifier set pte hva: %#08lx", __entry->hva) +); + +TRACE_EVENT(kvm_age_hva, + TP_PROTO(unsigned long start, unsigned long end), + TP_ARGS(start, end), + + TP_STRUCT__entry( + __field( unsigned long, start ) + __field( unsigned long, end ) + ), + + TP_fast_assign( + __entry->start = start; + __entry->end = end; + ), + + TP_printk("mmu notifier age hva: %#08lx -- %#08lx", + __entry->start, __entry->end) +); + +TRACE_EVENT(kvm_test_age_hva, + TP_PROTO(unsigned long hva), + TP_ARGS(hva), + + TP_STRUCT__entry( + __field( unsigned long, hva ) + ), + + TP_fast_assign( + __entry->hva = hva; + ), + + TP_printk("mmu notifier test age hva: %#08lx", __entry->hva) +); + +TRACE_EVENT(kvm_set_way_flush, + TP_PROTO(unsigned long vcpu_pc, bool cache), + TP_ARGS(vcpu_pc, cache), + + TP_STRUCT__entry( + __field( unsigned long, vcpu_pc ) + __field( bool, cache ) + ), + + TP_fast_assign( + __entry->vcpu_pc = vcpu_pc; + __entry->cache = cache; + ), + + TP_printk("S/W flush at 0x%016lx (cache %s)", + __entry->vcpu_pc, __entry->cache ? "on" : "off") +); + +TRACE_EVENT(kvm_toggle_cache, + TP_PROTO(unsigned long vcpu_pc, bool was, bool now), + TP_ARGS(vcpu_pc, was, now), + + TP_STRUCT__entry( + __field( unsigned long, vcpu_pc ) + __field( bool, was ) + __field( bool, now ) + ), + + TP_fast_assign( + __entry->vcpu_pc = vcpu_pc; + __entry->was = was; + __entry->now = now; + ), + + TP_printk("VM op at 0x%016lx (cache was %s, now %s)", + __entry->vcpu_pc, __entry->was ? "on" : "off", + __entry->now ? "on" : "off") ); /* diff --git a/virt/kvm/arm/vgic/trace.h b/virt/kvm/arm/vgic/trace.h new file mode 100644 index 000000000000..ed3229282888 --- /dev/null +++ b/virt/kvm/arm/vgic/trace.h @@ -0,0 +1,37 @@ +#if !defined(_TRACE_VGIC_H) || defined(TRACE_HEADER_MULTI_READ) +#define _TRACE_VGIC_H + +#include <linux/tracepoint.h> + +#undef TRACE_SYSTEM +#define TRACE_SYSTEM kvm + +TRACE_EVENT(vgic_update_irq_pending, + TP_PROTO(unsigned long vcpu_id, __u32 irq, bool level), + TP_ARGS(vcpu_id, irq, level), + + TP_STRUCT__entry( + __field( unsigned long, vcpu_id ) + __field( __u32, irq ) + __field( bool, level ) + ), + + TP_fast_assign( + __entry->vcpu_id = vcpu_id; + __entry->irq = irq; + __entry->level = level; + ), + + TP_printk("VCPU: %ld, IRQ %d, level: %d", + __entry->vcpu_id, __entry->irq, __entry->level) +); + +#endif /* _TRACE_VGIC_H */ + +#undef TRACE_INCLUDE_PATH +#define TRACE_INCLUDE_PATH ../../../virt/kvm/arm/vgic +#undef TRACE_INCLUDE_FILE +#define TRACE_INCLUDE_FILE trace + +/* This part must be outside protection */ +#include <trace/define_trace.h> diff --git a/virt/kvm/arm/vgic/vgic-init.c b/virt/kvm/arm/vgic/vgic-init.c index 25fd1b942c11..dc68e2e424ab 100644 --- a/virt/kvm/arm/vgic/vgic-init.c +++ b/virt/kvm/arm/vgic/vgic-init.c @@ -227,10 +227,27 @@ static int kvm_vgic_dist_init(struct kvm *kvm, unsigned int nr_spis) } /** - * kvm_vgic_vcpu_init() - Enable the VCPU interface - * @vcpu: the VCPU which's VGIC should be enabled + * kvm_vgic_vcpu_init() - Register VCPU-specific KVM iodevs + * @vcpu: pointer to the VCPU being created and initialized */ -static void kvm_vgic_vcpu_init(struct kvm_vcpu *vcpu) +int kvm_vgic_vcpu_init(struct kvm_vcpu *vcpu) +{ + int ret = 0; + struct vgic_dist *dist = &vcpu->kvm->arch.vgic; + + if (!irqchip_in_kernel(vcpu->kvm)) + return 0; + + /* + * If we are creating a VCPU with a GICv3 we must also register the + * KVM io device for the redistributor that belongs to this VCPU. + */ + if (dist->vgic_model == KVM_DEV_TYPE_ARM_VGIC_V3) + ret = vgic_register_redist_iodev(vcpu); + return ret; +} + +static void kvm_vgic_vcpu_enable(struct kvm_vcpu *vcpu) { if (kvm_vgic_global_state.type == VGIC_V2) vgic_v2_enable(vcpu); @@ -269,7 +286,7 @@ int vgic_init(struct kvm *kvm) dist->msis_require_devid = true; kvm_for_each_vcpu(i, vcpu, kvm) - kvm_vgic_vcpu_init(vcpu); + kvm_vgic_vcpu_enable(vcpu); ret = kvm_vgic_setup_default_irq_routing(kvm); if (ret) diff --git a/virt/kvm/arm/vgic/vgic-its.c b/virt/kvm/arm/vgic/vgic-its.c index 8d1da1af4b09..2dff288b3a66 100644 --- a/virt/kvm/arm/vgic/vgic-its.c +++ b/virt/kvm/arm/vgic/vgic-its.c @@ -23,6 +23,7 @@ #include <linux/interrupt.h> #include <linux/list.h> #include <linux/uaccess.h> +#include <linux/list_sort.h> #include <linux/irqchip/arm-gic-v3.h> @@ -33,6 +34,12 @@ #include "vgic.h" #include "vgic-mmio.h" +static int vgic_its_save_tables_v0(struct vgic_its *its); +static int vgic_its_restore_tables_v0(struct vgic_its *its); +static int vgic_its_commit_v0(struct vgic_its *its); +static int update_lpi_config(struct kvm *kvm, struct vgic_irq *irq, + struct kvm_vcpu *filter_vcpu); + /* * Creates a new (reference to a) struct vgic_irq for a given LPI. * If this LPI is already mapped on another ITS, we increase its refcount @@ -40,10 +47,12 @@ * If this is a "new" LPI, we allocate and initialize a new struct vgic_irq. * This function returns a pointer to the _unlocked_ structure. */ -static struct vgic_irq *vgic_add_lpi(struct kvm *kvm, u32 intid) +static struct vgic_irq *vgic_add_lpi(struct kvm *kvm, u32 intid, + struct kvm_vcpu *vcpu) { struct vgic_dist *dist = &kvm->arch.vgic; struct vgic_irq *irq = vgic_get_irq(kvm, NULL, intid), *oldirq; + int ret; /* In this case there is no put, since we keep the reference. */ if (irq) @@ -60,6 +69,7 @@ static struct vgic_irq *vgic_add_lpi(struct kvm *kvm, u32 intid) irq->config = VGIC_CONFIG_EDGE; kref_init(&irq->refcount); irq->intid = intid; + irq->target_vcpu = vcpu; spin_lock(&dist->lpi_list_lock); @@ -91,6 +101,19 @@ static struct vgic_irq *vgic_add_lpi(struct kvm *kvm, u32 intid) out_unlock: spin_unlock(&dist->lpi_list_lock); + /* + * We "cache" the configuration table entries in our struct vgic_irq's. + * However we only have those structs for mapped IRQs, so we read in + * the respective config data from memory here upon mapping the LPI. + */ + ret = update_lpi_config(kvm, irq, NULL); + if (ret) + return ERR_PTR(ret); + + ret = vgic_v3_lpi_sync_pending_status(kvm, irq); + if (ret) + return ERR_PTR(ret); + return irq; } @@ -99,6 +122,8 @@ struct its_device { /* the head for the list of ITTEs */ struct list_head itt_head; + u32 num_eventid_bits; + gpa_t itt_addr; u32 device_id; }; @@ -114,8 +139,8 @@ struct its_collection { #define its_is_collection_mapped(coll) ((coll) && \ ((coll)->target_addr != COLLECTION_NOT_MAPPED)) -struct its_itte { - struct list_head itte_list; +struct its_ite { + struct list_head ite_list; struct vgic_irq *irq; struct its_collection *collection; @@ -123,6 +148,50 @@ struct its_itte { u32 event_id; }; +/** + * struct vgic_its_abi - ITS abi ops and settings + * @cte_esz: collection table entry size + * @dte_esz: device table entry size + * @ite_esz: interrupt translation table entry size + * @save tables: save the ITS tables into guest RAM + * @restore_tables: restore the ITS internal structs from tables + * stored in guest RAM + * @commit: initialize the registers which expose the ABI settings, + * especially the entry sizes + */ +struct vgic_its_abi { + int cte_esz; + int dte_esz; + int ite_esz; + int (*save_tables)(struct vgic_its *its); + int (*restore_tables)(struct vgic_its *its); + int (*commit)(struct vgic_its *its); +}; + +static const struct vgic_its_abi its_table_abi_versions[] = { + [0] = {.cte_esz = 8, .dte_esz = 8, .ite_esz = 8, + .save_tables = vgic_its_save_tables_v0, + .restore_tables = vgic_its_restore_tables_v0, + .commit = vgic_its_commit_v0, + }, +}; + +#define NR_ITS_ABIS ARRAY_SIZE(its_table_abi_versions) + +inline const struct vgic_its_abi *vgic_its_get_abi(struct vgic_its *its) +{ + return &its_table_abi_versions[its->abi_rev]; +} + +int vgic_its_set_abi(struct vgic_its *its, int rev) +{ + const struct vgic_its_abi *abi; + + its->abi_rev = rev; + abi = vgic_its_get_abi(its); + return abi->commit(its); +} + /* * Find and returns a device in the device table for an ITS. * Must be called with the its_lock mutex held. @@ -143,27 +212,27 @@ static struct its_device *find_its_device(struct vgic_its *its, u32 device_id) * Device ID/Event ID pair on an ITS. * Must be called with the its_lock mutex held. */ -static struct its_itte *find_itte(struct vgic_its *its, u32 device_id, +static struct its_ite *find_ite(struct vgic_its *its, u32 device_id, u32 event_id) { struct its_device *device; - struct its_itte *itte; + struct its_ite *ite; device = find_its_device(its, device_id); if (device == NULL) return NULL; - list_for_each_entry(itte, &device->itt_head, itte_list) - if (itte->event_id == event_id) - return itte; + list_for_each_entry(ite, &device->itt_head, ite_list) + if (ite->event_id == event_id) + return ite; return NULL; } /* To be used as an iterator this macro misses the enclosing parentheses */ -#define for_each_lpi_its(dev, itte, its) \ +#define for_each_lpi_its(dev, ite, its) \ list_for_each_entry(dev, &(its)->device_list, dev_list) \ - list_for_each_entry(itte, &(dev)->itt_head, itte_list) + list_for_each_entry(ite, &(dev)->itt_head, ite_list) /* * We only implement 48 bits of PA at the moment, although the ITS @@ -171,11 +240,14 @@ static struct its_itte *find_itte(struct vgic_its *its, u32 device_id, */ #define BASER_ADDRESS(x) ((x) & GENMASK_ULL(47, 16)) #define CBASER_ADDRESS(x) ((x) & GENMASK_ULL(47, 12)) -#define PENDBASER_ADDRESS(x) ((x) & GENMASK_ULL(47, 16)) -#define PROPBASER_ADDRESS(x) ((x) & GENMASK_ULL(47, 12)) #define GIC_LPI_OFFSET 8192 +#define VITS_TYPER_IDBITS 16 +#define VITS_TYPER_DEVBITS 16 +#define VITS_DTE_MAX_DEVID_OFFSET (BIT(14) - 1) +#define VITS_ITE_MAX_EVENTID_OFFSET (BIT(16) - 1) + /* * Finds and returns a collection in the ITS collection table. * Must be called with the its_lock mutex held. @@ -204,7 +276,7 @@ static struct its_collection *find_collection(struct vgic_its *its, int coll_id) static int update_lpi_config(struct kvm *kvm, struct vgic_irq *irq, struct kvm_vcpu *filter_vcpu) { - u64 propbase = PROPBASER_ADDRESS(kvm->arch.vgic.propbaser); + u64 propbase = GICR_PROPBASER_ADDRESS(kvm->arch.vgic.propbaser); u8 prop; int ret; @@ -229,13 +301,13 @@ static int update_lpi_config(struct kvm *kvm, struct vgic_irq *irq, } /* - * Create a snapshot of the current LPI list, so that we can enumerate all - * LPIs without holding any lock. - * Returns the array length and puts the kmalloc'ed array into intid_ptr. + * Create a snapshot of the current LPIs targeting @vcpu, so that we can + * enumerate those LPIs without holding any lock. + * Returns their number and puts the kmalloc'ed array into intid_ptr. */ -static int vgic_copy_lpi_list(struct kvm *kvm, u32 **intid_ptr) +static int vgic_copy_lpi_list(struct kvm_vcpu *vcpu, u32 **intid_ptr) { - struct vgic_dist *dist = &kvm->arch.vgic; + struct vgic_dist *dist = &vcpu->kvm->arch.vgic; struct vgic_irq *irq; u32 *intids; int irq_count = dist->lpi_list_count, i = 0; @@ -254,14 +326,14 @@ static int vgic_copy_lpi_list(struct kvm *kvm, u32 **intid_ptr) spin_lock(&dist->lpi_list_lock); list_for_each_entry(irq, &dist->lpi_list_head, lpi_list) { /* We don't need to "get" the IRQ, as we hold the list lock. */ - intids[i] = irq->intid; - if (++i == irq_count) - break; + if (irq->target_vcpu != vcpu) + continue; + intids[i++] = irq->intid; } spin_unlock(&dist->lpi_list_lock); *intid_ptr = intids; - return irq_count; + return i; } /* @@ -270,18 +342,18 @@ static int vgic_copy_lpi_list(struct kvm *kvm, u32 **intid_ptr) * Needs to be called whenever either the collection for a LPIs has * changed or the collection itself got retargeted. */ -static void update_affinity_itte(struct kvm *kvm, struct its_itte *itte) +static void update_affinity_ite(struct kvm *kvm, struct its_ite *ite) { struct kvm_vcpu *vcpu; - if (!its_is_collection_mapped(itte->collection)) + if (!its_is_collection_mapped(ite->collection)) return; - vcpu = kvm_get_vcpu(kvm, itte->collection->target_addr); + vcpu = kvm_get_vcpu(kvm, ite->collection->target_addr); - spin_lock(&itte->irq->irq_lock); - itte->irq->target_vcpu = vcpu; - spin_unlock(&itte->irq->irq_lock); + spin_lock(&ite->irq->irq_lock); + ite->irq->target_vcpu = vcpu; + spin_unlock(&ite->irq->irq_lock); } /* @@ -292,13 +364,13 @@ static void update_affinity_collection(struct kvm *kvm, struct vgic_its *its, struct its_collection *coll) { struct its_device *device; - struct its_itte *itte; + struct its_ite *ite; - for_each_lpi_its(device, itte, its) { - if (!itte->collection || coll != itte->collection) + for_each_lpi_its(device, ite, its) { + if (!ite->collection || coll != ite->collection) continue; - update_affinity_itte(kvm, itte); + update_affinity_ite(kvm, ite); } } @@ -310,20 +382,20 @@ static u32 max_lpis_propbaser(u64 propbaser) } /* - * Scan the whole LPI pending table and sync the pending bit in there + * Sync the pending table pending bit of LPIs targeting @vcpu * with our own data structures. This relies on the LPI being * mapped before. */ static int its_sync_lpi_pending_table(struct kvm_vcpu *vcpu) { - gpa_t pendbase = PENDBASER_ADDRESS(vcpu->arch.vgic_cpu.pendbaser); + gpa_t pendbase = GICR_PENDBASER_ADDRESS(vcpu->arch.vgic_cpu.pendbaser); struct vgic_irq *irq; int last_byte_offset = -1; int ret = 0; u32 *intids; int nr_irqs, i; - nr_irqs = vgic_copy_lpi_list(vcpu->kvm, &intids); + nr_irqs = vgic_copy_lpi_list(vcpu, &intids); if (nr_irqs < 0) return nr_irqs; @@ -364,6 +436,7 @@ static unsigned long vgic_mmio_read_its_typer(struct kvm *kvm, struct vgic_its *its, gpa_t addr, unsigned int len) { + const struct vgic_its_abi *abi = vgic_its_get_abi(its); u64 reg = GITS_TYPER_PLPIS; /* @@ -374,8 +447,9 @@ static unsigned long vgic_mmio_read_its_typer(struct kvm *kvm, * To avoid memory waste in the guest, we keep the number of IDBits and * DevBits low - as least for the time being. */ - reg |= 0x0f << GITS_TYPER_DEVBITS_SHIFT; - reg |= 0x0f << GITS_TYPER_IDBITS_SHIFT; + reg |= GIC_ENCODE_SZ(VITS_TYPER_DEVBITS, 5) << GITS_TYPER_DEVBITS_SHIFT; + reg |= GIC_ENCODE_SZ(VITS_TYPER_IDBITS, 5) << GITS_TYPER_IDBITS_SHIFT; + reg |= GIC_ENCODE_SZ(abi->ite_esz, 4) << GITS_TYPER_ITT_ENTRY_SIZE_SHIFT; return extract_bytes(reg, addr & 7, len); } @@ -384,7 +458,23 @@ static unsigned long vgic_mmio_read_its_iidr(struct kvm *kvm, struct vgic_its *its, gpa_t addr, unsigned int len) { - return (PRODUCT_ID_KVM << 24) | (IMPLEMENTER_ARM << 0); + u32 val; + + val = (its->abi_rev << GITS_IIDR_REV_SHIFT) & GITS_IIDR_REV_MASK; + val |= (PRODUCT_ID_KVM << GITS_IIDR_PRODUCTID_SHIFT) | IMPLEMENTER_ARM; + return val; +} + +static int vgic_mmio_uaccess_write_its_iidr(struct kvm *kvm, + struct vgic_its *its, + gpa_t addr, unsigned int len, + unsigned long val) +{ + u32 rev = GITS_IIDR_REV(val); + + if (rev >= NR_ITS_ABIS) + return -EINVAL; + return vgic_its_set_abi(its, rev); } static unsigned long vgic_mmio_read_its_idregs(struct kvm *kvm, @@ -425,25 +515,25 @@ static int vgic_its_trigger_msi(struct kvm *kvm, struct vgic_its *its, u32 devid, u32 eventid) { struct kvm_vcpu *vcpu; - struct its_itte *itte; + struct its_ite *ite; if (!its->enabled) return -EBUSY; - itte = find_itte(its, devid, eventid); - if (!itte || !its_is_collection_mapped(itte->collection)) + ite = find_ite(its, devid, eventid); + if (!ite || !its_is_collection_mapped(ite->collection)) return E_ITS_INT_UNMAPPED_INTERRUPT; - vcpu = kvm_get_vcpu(kvm, itte->collection->target_addr); + vcpu = kvm_get_vcpu(kvm, ite->collection->target_addr); if (!vcpu) return E_ITS_INT_UNMAPPED_INTERRUPT; if (!vcpu->arch.vgic_cpu.lpis_enabled) return -EBUSY; - spin_lock(&itte->irq->irq_lock); - itte->irq->pending_latch = true; - vgic_queue_irq_unlock(kvm, itte->irq); + spin_lock(&ite->irq->irq_lock); + ite->irq->pending_latch = true; + vgic_queue_irq_unlock(kvm, ite->irq); return 0; } @@ -511,15 +601,15 @@ int vgic_its_inject_msi(struct kvm *kvm, struct kvm_msi *msi) } /* Requires the its_lock to be held. */ -static void its_free_itte(struct kvm *kvm, struct its_itte *itte) +static void its_free_ite(struct kvm *kvm, struct its_ite *ite) { - list_del(&itte->itte_list); + list_del(&ite->ite_list); /* This put matches the get in vgic_add_lpi. */ - if (itte->irq) - vgic_put_irq(kvm, itte->irq); + if (ite->irq) + vgic_put_irq(kvm, ite->irq); - kfree(itte); + kfree(ite); } static u64 its_cmd_mask_field(u64 *its_cmd, int word, int shift, int size) @@ -529,9 +619,11 @@ static u64 its_cmd_mask_field(u64 *its_cmd, int word, int shift, int size) #define its_cmd_get_command(cmd) its_cmd_mask_field(cmd, 0, 0, 8) #define its_cmd_get_deviceid(cmd) its_cmd_mask_field(cmd, 0, 32, 32) +#define its_cmd_get_size(cmd) (its_cmd_mask_field(cmd, 1, 0, 5) + 1) #define its_cmd_get_id(cmd) its_cmd_mask_field(cmd, 1, 0, 32) #define its_cmd_get_physical_id(cmd) its_cmd_mask_field(cmd, 1, 32, 32) #define its_cmd_get_collection(cmd) its_cmd_mask_field(cmd, 2, 0, 16) +#define its_cmd_get_ittaddr(cmd) (its_cmd_mask_field(cmd, 2, 8, 44) << 8) #define its_cmd_get_target_addr(cmd) its_cmd_mask_field(cmd, 2, 16, 32) #define its_cmd_get_validbit(cmd) its_cmd_mask_field(cmd, 2, 63, 1) @@ -544,17 +636,17 @@ static int vgic_its_cmd_handle_discard(struct kvm *kvm, struct vgic_its *its, { u32 device_id = its_cmd_get_deviceid(its_cmd); u32 event_id = its_cmd_get_id(its_cmd); - struct its_itte *itte; + struct its_ite *ite; - itte = find_itte(its, device_id, event_id); - if (itte && itte->collection) { + ite = find_ite(its, device_id, event_id); + if (ite && ite->collection) { /* * Though the spec talks about removing the pending state, we * don't bother here since we clear the ITTE anyway and the * pending state is a property of the ITTE struct. */ - its_free_itte(kvm, itte); + its_free_ite(kvm, ite); return 0; } @@ -572,26 +664,26 @@ static int vgic_its_cmd_handle_movi(struct kvm *kvm, struct vgic_its *its, u32 event_id = its_cmd_get_id(its_cmd); u32 coll_id = its_cmd_get_collection(its_cmd); struct kvm_vcpu *vcpu; - struct its_itte *itte; + struct its_ite *ite; struct its_collection *collection; - itte = find_itte(its, device_id, event_id); - if (!itte) + ite = find_ite(its, device_id, event_id); + if (!ite) return E_ITS_MOVI_UNMAPPED_INTERRUPT; - if (!its_is_collection_mapped(itte->collection)) + if (!its_is_collection_mapped(ite->collection)) return E_ITS_MOVI_UNMAPPED_COLLECTION; collection = find_collection(its, coll_id); if (!its_is_collection_mapped(collection)) return E_ITS_MOVI_UNMAPPED_COLLECTION; - itte->collection = collection; + ite->collection = collection; vcpu = kvm_get_vcpu(kvm, collection->target_addr); - spin_lock(&itte->irq->irq_lock); - itte->irq->target_vcpu = vcpu; - spin_unlock(&itte->irq->irq_lock); + spin_lock(&ite->irq->irq_lock); + ite->irq->target_vcpu = vcpu; + spin_unlock(&ite->irq->irq_lock); return 0; } @@ -600,16 +692,31 @@ static int vgic_its_cmd_handle_movi(struct kvm *kvm, struct vgic_its *its, * Check whether an ID can be stored into the corresponding guest table. * For a direct table this is pretty easy, but gets a bit nasty for * indirect tables. We check whether the resulting guest physical address - * is actually valid (covered by a memslot and guest accessbible). + * is actually valid (covered by a memslot and guest accessible). * For this we have to read the respective first level entry. */ -static bool vgic_its_check_id(struct vgic_its *its, u64 baser, int id) +static bool vgic_its_check_id(struct vgic_its *its, u64 baser, u32 id, + gpa_t *eaddr) { int l1_tbl_size = GITS_BASER_NR_PAGES(baser) * SZ_64K; + u64 indirect_ptr, type = GITS_BASER_TYPE(baser); + int esz = GITS_BASER_ENTRY_SIZE(baser); int index; - u64 indirect_ptr; gfn_t gfn; - int esz = GITS_BASER_ENTRY_SIZE(baser); + + switch (type) { + case GITS_BASER_TYPE_DEVICE: + if (id >= BIT_ULL(VITS_TYPER_DEVBITS)) + return false; + break; + case GITS_BASER_TYPE_COLLECTION: + /* as GITS_TYPER.CIL == 0, ITS supports 16-bit collection ID */ + if (id >= BIT_ULL(16)) + return false; + break; + default: + return false; + } if (!(baser & GITS_BASER_INDIRECT)) { phys_addr_t addr; @@ -620,6 +727,8 @@ static bool vgic_its_check_id(struct vgic_its *its, u64 baser, int id) addr = BASER_ADDRESS(baser) + id * esz; gfn = addr >> PAGE_SHIFT; + if (eaddr) + *eaddr = addr; return kvm_is_visible_gfn(its->dev->kvm, gfn); } @@ -652,6 +761,8 @@ static bool vgic_its_check_id(struct vgic_its *its, u64 baser, int id) indirect_ptr += index * esz; gfn = indirect_ptr >> PAGE_SHIFT; + if (eaddr) + *eaddr = indirect_ptr; return kvm_is_visible_gfn(its->dev->kvm, gfn); } @@ -661,7 +772,7 @@ static int vgic_its_alloc_collection(struct vgic_its *its, { struct its_collection *collection; - if (!vgic_its_check_id(its, its->baser_coll_table, coll_id)) + if (!vgic_its_check_id(its, its->baser_coll_table, coll_id, NULL)) return E_ITS_MAPC_COLLECTION_OOR; collection = kzalloc(sizeof(*collection), GFP_KERNEL); @@ -679,7 +790,7 @@ static void vgic_its_free_collection(struct vgic_its *its, u32 coll_id) { struct its_collection *collection; struct its_device *device; - struct its_itte *itte; + struct its_ite *ite; /* * Clearing the mapping for that collection ID removes the @@ -690,15 +801,34 @@ static void vgic_its_free_collection(struct vgic_its *its, u32 coll_id) if (!collection) return; - for_each_lpi_its(device, itte, its) - if (itte->collection && - itte->collection->collection_id == coll_id) - itte->collection = NULL; + for_each_lpi_its(device, ite, its) + if (ite->collection && + ite->collection->collection_id == coll_id) + ite->collection = NULL; list_del(&collection->coll_list); kfree(collection); } +/* Must be called with its_lock mutex held */ +static struct its_ite *vgic_its_alloc_ite(struct its_device *device, + struct its_collection *collection, + u32 lpi_id, u32 event_id) +{ + struct its_ite *ite; + + ite = kzalloc(sizeof(*ite), GFP_KERNEL); + if (!ite) + return ERR_PTR(-ENOMEM); + + ite->event_id = event_id; + ite->collection = collection; + ite->lpi = lpi_id; + + list_add_tail(&ite->ite_list, &device->itt_head); + return ite; +} + /* * The MAPTI and MAPI commands map LPIs to ITTEs. * Must be called with its_lock mutex held. @@ -709,16 +839,20 @@ static int vgic_its_cmd_handle_mapi(struct kvm *kvm, struct vgic_its *its, u32 device_id = its_cmd_get_deviceid(its_cmd); u32 event_id = its_cmd_get_id(its_cmd); u32 coll_id = its_cmd_get_collection(its_cmd); - struct its_itte *itte; + struct its_ite *ite; + struct kvm_vcpu *vcpu = NULL; struct its_device *device; struct its_collection *collection, *new_coll = NULL; - int lpi_nr; struct vgic_irq *irq; + int lpi_nr; device = find_its_device(its, device_id); if (!device) return E_ITS_MAPTI_UNMAPPED_DEVICE; + if (event_id >= BIT_ULL(device->num_eventid_bits)) + return E_ITS_MAPTI_ID_OOR; + if (its_cmd_get_command(its_cmd) == GITS_CMD_MAPTI) lpi_nr = its_cmd_get_physical_id(its_cmd); else @@ -728,7 +862,7 @@ static int vgic_its_cmd_handle_mapi(struct kvm *kvm, struct vgic_its *its, return E_ITS_MAPTI_PHYSICALID_OOR; /* If there is an existing mapping, behavior is UNPREDICTABLE. */ - if (find_itte(its, device_id, event_id)) + if (find_ite(its, device_id, event_id)) return 0; collection = find_collection(its, coll_id); @@ -739,36 +873,24 @@ static int vgic_its_cmd_handle_mapi(struct kvm *kvm, struct vgic_its *its, new_coll = collection; } - itte = kzalloc(sizeof(struct its_itte), GFP_KERNEL); - if (!itte) { + ite = vgic_its_alloc_ite(device, collection, lpi_nr, event_id); + if (IS_ERR(ite)) { if (new_coll) vgic_its_free_collection(its, coll_id); - return -ENOMEM; + return PTR_ERR(ite); } - itte->event_id = event_id; - list_add_tail(&itte->itte_list, &device->itt_head); - - itte->collection = collection; - itte->lpi = lpi_nr; + if (its_is_collection_mapped(collection)) + vcpu = kvm_get_vcpu(kvm, collection->target_addr); - irq = vgic_add_lpi(kvm, lpi_nr); + irq = vgic_add_lpi(kvm, lpi_nr, vcpu); if (IS_ERR(irq)) { if (new_coll) vgic_its_free_collection(its, coll_id); - its_free_itte(kvm, itte); + its_free_ite(kvm, ite); return PTR_ERR(irq); } - itte->irq = irq; - - update_affinity_itte(kvm, itte); - - /* - * We "cache" the configuration table entries in out struct vgic_irq's. - * However we only have those structs for mapped IRQs, so we read in - * the respective config data from memory here upon mapping the LPI. - */ - update_lpi_config(kvm, itte->irq, NULL); + ite->irq = irq; return 0; } @@ -776,20 +898,40 @@ static int vgic_its_cmd_handle_mapi(struct kvm *kvm, struct vgic_its *its, /* Requires the its_lock to be held. */ static void vgic_its_unmap_device(struct kvm *kvm, struct its_device *device) { - struct its_itte *itte, *temp; + struct its_ite *ite, *temp; /* * The spec says that unmapping a device with still valid * ITTEs associated is UNPREDICTABLE. We remove all ITTEs, * since we cannot leave the memory unreferenced. */ - list_for_each_entry_safe(itte, temp, &device->itt_head, itte_list) - its_free_itte(kvm, itte); + list_for_each_entry_safe(ite, temp, &device->itt_head, ite_list) + its_free_ite(kvm, ite); list_del(&device->dev_list); kfree(device); } +/* Must be called with its_lock mutex held */ +static struct its_device *vgic_its_alloc_device(struct vgic_its *its, + u32 device_id, gpa_t itt_addr, + u8 num_eventid_bits) +{ + struct its_device *device; + + device = kzalloc(sizeof(*device), GFP_KERNEL); + if (!device) + return ERR_PTR(-ENOMEM); + + device->device_id = device_id; + device->itt_addr = itt_addr; + device->num_eventid_bits = num_eventid_bits; + INIT_LIST_HEAD(&device->itt_head); + + list_add_tail(&device->dev_list, &its->device_list); + return device; +} + /* * MAPD maps or unmaps a device ID to Interrupt Translation Tables (ITTs). * Must be called with the its_lock mutex held. @@ -799,11 +941,16 @@ static int vgic_its_cmd_handle_mapd(struct kvm *kvm, struct vgic_its *its, { u32 device_id = its_cmd_get_deviceid(its_cmd); bool valid = its_cmd_get_validbit(its_cmd); + u8 num_eventid_bits = its_cmd_get_size(its_cmd); + gpa_t itt_addr = its_cmd_get_ittaddr(its_cmd); struct its_device *device; - if (!vgic_its_check_id(its, its->baser_device_table, device_id)) + if (!vgic_its_check_id(its, its->baser_device_table, device_id, NULL)) return E_ITS_MAPD_DEVICE_OOR; + if (valid && num_eventid_bits > VITS_TYPER_IDBITS) + return E_ITS_MAPD_ITTSIZE_OOR; + device = find_its_device(its, device_id); /* @@ -821,14 +968,10 @@ static int vgic_its_cmd_handle_mapd(struct kvm *kvm, struct vgic_its *its, if (!valid) return 0; - device = kzalloc(sizeof(struct its_device), GFP_KERNEL); - if (!device) - return -ENOMEM; - - device->device_id = device_id; - INIT_LIST_HEAD(&device->itt_head); - - list_add_tail(&device->dev_list, &its->device_list); + device = vgic_its_alloc_device(its, device_id, itt_addr, + num_eventid_bits); + if (IS_ERR(device)) + return PTR_ERR(device); return 0; } @@ -883,14 +1026,14 @@ static int vgic_its_cmd_handle_clear(struct kvm *kvm, struct vgic_its *its, { u32 device_id = its_cmd_get_deviceid(its_cmd); u32 event_id = its_cmd_get_id(its_cmd); - struct its_itte *itte; + struct its_ite *ite; - itte = find_itte(its, device_id, event_id); - if (!itte) + ite = find_ite(its, device_id, event_id); + if (!ite) return E_ITS_CLEAR_UNMAPPED_INTERRUPT; - itte->irq->pending_latch = false; + ite->irq->pending_latch = false; return 0; } @@ -904,14 +1047,14 @@ static int vgic_its_cmd_handle_inv(struct kvm *kvm, struct vgic_its *its, { u32 device_id = its_cmd_get_deviceid(its_cmd); u32 event_id = its_cmd_get_id(its_cmd); - struct its_itte *itte; + struct its_ite *ite; - itte = find_itte(its, device_id, event_id); - if (!itte) + ite = find_ite(its, device_id, event_id); + if (!ite) return E_ITS_INV_UNMAPPED_INTERRUPT; - return update_lpi_config(kvm, itte->irq, NULL); + return update_lpi_config(kvm, ite->irq, NULL); } /* @@ -938,7 +1081,7 @@ static int vgic_its_cmd_handle_invall(struct kvm *kvm, struct vgic_its *its, vcpu = kvm_get_vcpu(kvm, collection->target_addr); - irq_count = vgic_copy_lpi_list(kvm, &intids); + irq_count = vgic_copy_lpi_list(vcpu, &intids); if (irq_count < 0) return irq_count; @@ -1213,6 +1356,33 @@ static unsigned long vgic_mmio_read_its_creadr(struct kvm *kvm, return extract_bytes(its->creadr, addr & 0x7, len); } +static int vgic_mmio_uaccess_write_its_creadr(struct kvm *kvm, + struct vgic_its *its, + gpa_t addr, unsigned int len, + unsigned long val) +{ + u32 cmd_offset; + int ret = 0; + + mutex_lock(&its->cmd_lock); + + if (its->enabled) { + ret = -EBUSY; + goto out; + } + + cmd_offset = ITS_CMD_OFFSET(val); + if (cmd_offset >= ITS_CMD_BUFFER_SIZE(its->cbaser)) { + ret = -EINVAL; + goto out; + } + + its->creadr = cmd_offset; +out: + mutex_unlock(&its->cmd_lock); + return ret; +} + #define BASER_INDEX(addr) (((addr) / sizeof(u64)) & 0x7) static unsigned long vgic_mmio_read_its_baser(struct kvm *kvm, struct vgic_its *its, @@ -1241,6 +1411,7 @@ static void vgic_mmio_write_its_baser(struct kvm *kvm, gpa_t addr, unsigned int len, unsigned long val) { + const struct vgic_its_abi *abi = vgic_its_get_abi(its); u64 entry_size, device_type; u64 reg, *regptr, clearbits = 0; @@ -1251,12 +1422,12 @@ static void vgic_mmio_write_its_baser(struct kvm *kvm, switch (BASER_INDEX(addr)) { case 0: regptr = &its->baser_device_table; - entry_size = 8; + entry_size = abi->dte_esz; device_type = GITS_BASER_TYPE_DEVICE; break; case 1: regptr = &its->baser_coll_table; - entry_size = 8; + entry_size = abi->cte_esz; device_type = GITS_BASER_TYPE_COLLECTION; clearbits = GITS_BASER_INDIRECT; break; @@ -1317,6 +1488,16 @@ static void vgic_mmio_write_its_ctlr(struct kvm *kvm, struct vgic_its *its, .its_write = wr, \ } +#define REGISTER_ITS_DESC_UACCESS(off, rd, wr, uwr, length, acc)\ +{ \ + .reg_offset = off, \ + .len = length, \ + .access_flags = acc, \ + .its_read = rd, \ + .its_write = wr, \ + .uaccess_its_write = uwr, \ +} + static void its_mmio_write_wi(struct kvm *kvm, struct vgic_its *its, gpa_t addr, unsigned int len, unsigned long val) { @@ -1327,8 +1508,9 @@ static struct vgic_register_region its_registers[] = { REGISTER_ITS_DESC(GITS_CTLR, vgic_mmio_read_its_ctlr, vgic_mmio_write_its_ctlr, 4, VGIC_ACCESS_32bit), - REGISTER_ITS_DESC(GITS_IIDR, - vgic_mmio_read_its_iidr, its_mmio_write_wi, 4, + REGISTER_ITS_DESC_UACCESS(GITS_IIDR, + vgic_mmio_read_its_iidr, its_mmio_write_wi, + vgic_mmio_uaccess_write_its_iidr, 4, VGIC_ACCESS_32bit), REGISTER_ITS_DESC(GITS_TYPER, vgic_mmio_read_its_typer, its_mmio_write_wi, 8, @@ -1339,8 +1521,9 @@ static struct vgic_register_region its_registers[] = { REGISTER_ITS_DESC(GITS_CWRITER, vgic_mmio_read_its_cwriter, vgic_mmio_write_its_cwriter, 8, VGIC_ACCESS_64bit | VGIC_ACCESS_32bit), - REGISTER_ITS_DESC(GITS_CREADR, - vgic_mmio_read_its_creadr, its_mmio_write_wi, 8, + REGISTER_ITS_DESC_UACCESS(GITS_CREADR, + vgic_mmio_read_its_creadr, its_mmio_write_wi, + vgic_mmio_uaccess_write_its_creadr, 8, VGIC_ACCESS_64bit | VGIC_ACCESS_32bit), REGISTER_ITS_DESC(GITS_BASER, vgic_mmio_read_its_baser, vgic_mmio_write_its_baser, 0x40, @@ -1357,17 +1540,19 @@ void vgic_enable_lpis(struct kvm_vcpu *vcpu) its_sync_lpi_pending_table(vcpu); } -static int vgic_register_its_iodev(struct kvm *kvm, struct vgic_its *its) +static int vgic_register_its_iodev(struct kvm *kvm, struct vgic_its *its, + u64 addr) { struct vgic_io_device *iodev = &its->iodev; int ret; - if (!its->initialized) - return -EBUSY; - - if (IS_VGIC_ADDR_UNDEF(its->vgic_its_base)) - return -ENXIO; + mutex_lock(&kvm->slots_lock); + if (!IS_VGIC_ADDR_UNDEF(its->vgic_its_base)) { + ret = -EBUSY; + goto out; + } + its->vgic_its_base = addr; iodev->regions = its_registers; iodev->nr_regions = ARRAY_SIZE(its_registers); kvm_iodevice_init(&iodev->dev, &kvm_io_gic_ops); @@ -1375,9 +1560,9 @@ static int vgic_register_its_iodev(struct kvm *kvm, struct vgic_its *its) iodev->base_addr = its->vgic_its_base; iodev->iodev_type = IODEV_ITS; iodev->its = its; - mutex_lock(&kvm->slots_lock); ret = kvm_io_bus_register_dev(kvm, KVM_MMIO_BUS, iodev->base_addr, KVM_VGIC_V3_ITS_SIZE, &iodev->dev); +out: mutex_unlock(&kvm->slots_lock); return ret; @@ -1387,7 +1572,6 @@ static int vgic_register_its_iodev(struct kvm *kvm, struct vgic_its *its) (GIC_BASER_CACHEABILITY(GITS_BASER, INNER, RaWb) | \ GIC_BASER_CACHEABILITY(GITS_BASER, OUTER, SameAsInner) | \ GIC_BASER_SHAREABILITY(GITS_BASER, InnerShareable) | \ - ((8ULL - 1) << GITS_BASER_ENTRY_SIZE_SHIFT) | \ GITS_BASER_PAGE_SIZE_64K) #define INITIAL_PROPBASER_VALUE \ @@ -1415,7 +1599,6 @@ static int vgic_its_create(struct kvm_device *dev, u32 type) INIT_LIST_HEAD(&its->collection_list); dev->kvm->arch.vgic.has_its = true; - its->initialized = false; its->enabled = false; its->dev = dev; @@ -1427,16 +1610,23 @@ static int vgic_its_create(struct kvm_device *dev, u32 type) dev->private = its; - return 0; + return vgic_its_set_abi(its, NR_ITS_ABIS - 1); +} + +static void vgic_its_free_device(struct kvm *kvm, struct its_device *dev) +{ + struct its_ite *ite, *tmp; + + list_for_each_entry_safe(ite, tmp, &dev->itt_head, ite_list) + its_free_ite(kvm, ite); + list_del(&dev->dev_list); + kfree(dev); } static void vgic_its_destroy(struct kvm_device *kvm_dev) { struct kvm *kvm = kvm_dev->kvm; struct vgic_its *its = kvm_dev->private; - struct its_device *dev; - struct its_itte *itte; - struct list_head *dev_cur, *dev_temp; struct list_head *cur, *temp; /* @@ -1447,25 +1637,710 @@ static void vgic_its_destroy(struct kvm_device *kvm_dev) return; mutex_lock(&its->its_lock); - list_for_each_safe(dev_cur, dev_temp, &its->device_list) { - dev = container_of(dev_cur, struct its_device, dev_list); - list_for_each_safe(cur, temp, &dev->itt_head) { - itte = (container_of(cur, struct its_itte, itte_list)); - its_free_itte(kvm, itte); - } - list_del(dev_cur); - kfree(dev); + list_for_each_safe(cur, temp, &its->device_list) { + struct its_device *dev; + + dev = list_entry(cur, struct its_device, dev_list); + vgic_its_free_device(kvm, dev); } list_for_each_safe(cur, temp, &its->collection_list) { + struct its_collection *coll; + + coll = list_entry(cur, struct its_collection, coll_list); list_del(cur); - kfree(container_of(cur, struct its_collection, coll_list)); + kfree(coll); } mutex_unlock(&its->its_lock); kfree(its); } +int vgic_its_has_attr_regs(struct kvm_device *dev, + struct kvm_device_attr *attr) +{ + const struct vgic_register_region *region; + gpa_t offset = attr->attr; + int align; + + align = (offset < GITS_TYPER) || (offset >= GITS_PIDR4) ? 0x3 : 0x7; + + if (offset & align) + return -EINVAL; + + region = vgic_find_mmio_region(its_registers, + ARRAY_SIZE(its_registers), + offset); + if (!region) + return -ENXIO; + + return 0; +} + +int vgic_its_attr_regs_access(struct kvm_device *dev, + struct kvm_device_attr *attr, + u64 *reg, bool is_write) +{ + const struct vgic_register_region *region; + struct vgic_its *its; + gpa_t addr, offset; + unsigned int len; + int align, ret = 0; + + its = dev->private; + offset = attr->attr; + + /* + * Although the spec supports upper/lower 32-bit accesses to + * 64-bit ITS registers, the userspace ABI requires 64-bit + * accesses to all 64-bit wide registers. We therefore only + * support 32-bit accesses to GITS_CTLR, GITS_IIDR and GITS ID + * registers + */ + if ((offset < GITS_TYPER) || (offset >= GITS_PIDR4)) + align = 0x3; + else + align = 0x7; + + if (offset & align) + return -EINVAL; + + mutex_lock(&dev->kvm->lock); + + if (IS_VGIC_ADDR_UNDEF(its->vgic_its_base)) { + ret = -ENXIO; + goto out; + } + + region = vgic_find_mmio_region(its_registers, + ARRAY_SIZE(its_registers), + offset); + if (!region) { + ret = -ENXIO; + goto out; + } + + if (!lock_all_vcpus(dev->kvm)) { + ret = -EBUSY; + goto out; + } + + addr = its->vgic_its_base + offset; + + len = region->access_flags & VGIC_ACCESS_64bit ? 8 : 4; + + if (is_write) { + if (region->uaccess_its_write) + ret = region->uaccess_its_write(dev->kvm, its, addr, + len, *reg); + else + region->its_write(dev->kvm, its, addr, len, *reg); + } else { + *reg = region->its_read(dev->kvm, its, addr, len); + } + unlock_all_vcpus(dev->kvm); +out: + mutex_unlock(&dev->kvm->lock); + return ret; +} + +static u32 compute_next_devid_offset(struct list_head *h, + struct its_device *dev) +{ + struct its_device *next; + u32 next_offset; + + if (list_is_last(&dev->dev_list, h)) + return 0; + next = list_next_entry(dev, dev_list); + next_offset = next->device_id - dev->device_id; + + return min_t(u32, next_offset, VITS_DTE_MAX_DEVID_OFFSET); +} + +static u32 compute_next_eventid_offset(struct list_head *h, struct its_ite *ite) +{ + struct its_ite *next; + u32 next_offset; + + if (list_is_last(&ite->ite_list, h)) + return 0; + next = list_next_entry(ite, ite_list); + next_offset = next->event_id - ite->event_id; + + return min_t(u32, next_offset, VITS_ITE_MAX_EVENTID_OFFSET); +} + +/** + * entry_fn_t - Callback called on a table entry restore path + * @its: its handle + * @id: id of the entry + * @entry: pointer to the entry + * @opaque: pointer to an opaque data + * + * Return: < 0 on error, 0 if last element was identified, id offset to next + * element otherwise + */ +typedef int (*entry_fn_t)(struct vgic_its *its, u32 id, void *entry, + void *opaque); + +/** + * scan_its_table - Scan a contiguous table in guest RAM and applies a function + * to each entry + * + * @its: its handle + * @base: base gpa of the table + * @size: size of the table in bytes + * @esz: entry size in bytes + * @start_id: the ID of the first entry in the table + * (non zero for 2d level tables) + * @fn: function to apply on each entry + * + * Return: < 0 on error, 0 if last element was identified, 1 otherwise + * (the last element may not be found on second level tables) + */ +static int scan_its_table(struct vgic_its *its, gpa_t base, int size, int esz, + int start_id, entry_fn_t fn, void *opaque) +{ + void *entry = kzalloc(esz, GFP_KERNEL); + struct kvm *kvm = its->dev->kvm; + unsigned long len = size; + int id = start_id; + gpa_t gpa = base; + int ret; + + while (len > 0) { + int next_offset; + size_t byte_offset; + + ret = kvm_read_guest(kvm, gpa, entry, esz); + if (ret) + goto out; + + next_offset = fn(its, id, entry, opaque); + if (next_offset <= 0) { + ret = next_offset; + goto out; + } + + byte_offset = next_offset * esz; + id += next_offset; + gpa += byte_offset; + len -= byte_offset; + } + ret = 1; + +out: + kfree(entry); + return ret; +} + +/** + * vgic_its_save_ite - Save an interrupt translation entry at @gpa + */ +static int vgic_its_save_ite(struct vgic_its *its, struct its_device *dev, + struct its_ite *ite, gpa_t gpa, int ite_esz) +{ + struct kvm *kvm = its->dev->kvm; + u32 next_offset; + u64 val; + + next_offset = compute_next_eventid_offset(&dev->itt_head, ite); + val = ((u64)next_offset << KVM_ITS_ITE_NEXT_SHIFT) | + ((u64)ite->lpi << KVM_ITS_ITE_PINTID_SHIFT) | + ite->collection->collection_id; + val = cpu_to_le64(val); + return kvm_write_guest(kvm, gpa, &val, ite_esz); +} + +/** + * vgic_its_restore_ite - restore an interrupt translation entry + * @event_id: id used for indexing + * @ptr: pointer to the ITE entry + * @opaque: pointer to the its_device + */ +static int vgic_its_restore_ite(struct vgic_its *its, u32 event_id, + void *ptr, void *opaque) +{ + struct its_device *dev = (struct its_device *)opaque; + struct its_collection *collection; + struct kvm *kvm = its->dev->kvm; + struct kvm_vcpu *vcpu = NULL; + u64 val; + u64 *p = (u64 *)ptr; + struct vgic_irq *irq; + u32 coll_id, lpi_id; + struct its_ite *ite; + u32 offset; + + val = *p; + + val = le64_to_cpu(val); + + coll_id = val & KVM_ITS_ITE_ICID_MASK; + lpi_id = (val & KVM_ITS_ITE_PINTID_MASK) >> KVM_ITS_ITE_PINTID_SHIFT; + + if (!lpi_id) + return 1; /* invalid entry, no choice but to scan next entry */ + + if (lpi_id < VGIC_MIN_LPI) + return -EINVAL; + + offset = val >> KVM_ITS_ITE_NEXT_SHIFT; + if (event_id + offset >= BIT_ULL(dev->num_eventid_bits)) + return -EINVAL; + + collection = find_collection(its, coll_id); + if (!collection) + return -EINVAL; + + ite = vgic_its_alloc_ite(dev, collection, lpi_id, event_id); + if (IS_ERR(ite)) + return PTR_ERR(ite); + + if (its_is_collection_mapped(collection)) + vcpu = kvm_get_vcpu(kvm, collection->target_addr); + + irq = vgic_add_lpi(kvm, lpi_id, vcpu); + if (IS_ERR(irq)) + return PTR_ERR(irq); + ite->irq = irq; + + return offset; +} + +static int vgic_its_ite_cmp(void *priv, struct list_head *a, + struct list_head *b) +{ + struct its_ite *itea = container_of(a, struct its_ite, ite_list); + struct its_ite *iteb = container_of(b, struct its_ite, ite_list); + + if (itea->event_id < iteb->event_id) + return -1; + else + return 1; +} + +static int vgic_its_save_itt(struct vgic_its *its, struct its_device *device) +{ + const struct vgic_its_abi *abi = vgic_its_get_abi(its); + gpa_t base = device->itt_addr; + struct its_ite *ite; + int ret; + int ite_esz = abi->ite_esz; + + list_sort(NULL, &device->itt_head, vgic_its_ite_cmp); + + list_for_each_entry(ite, &device->itt_head, ite_list) { + gpa_t gpa = base + ite->event_id * ite_esz; + + ret = vgic_its_save_ite(its, device, ite, gpa, ite_esz); + if (ret) + return ret; + } + return 0; +} + +static int vgic_its_restore_itt(struct vgic_its *its, struct its_device *dev) +{ + const struct vgic_its_abi *abi = vgic_its_get_abi(its); + gpa_t base = dev->itt_addr; + int ret; + int ite_esz = abi->ite_esz; + size_t max_size = BIT_ULL(dev->num_eventid_bits) * ite_esz; + + ret = scan_its_table(its, base, max_size, ite_esz, 0, + vgic_its_restore_ite, dev); + + return ret; +} + +/** + * vgic_its_save_dte - Save a device table entry at a given GPA + * + * @its: ITS handle + * @dev: ITS device + * @ptr: GPA + */ +static int vgic_its_save_dte(struct vgic_its *its, struct its_device *dev, + gpa_t ptr, int dte_esz) +{ + struct kvm *kvm = its->dev->kvm; + u64 val, itt_addr_field; + u32 next_offset; + + itt_addr_field = dev->itt_addr >> 8; + next_offset = compute_next_devid_offset(&its->device_list, dev); + val = (1ULL << KVM_ITS_DTE_VALID_SHIFT | + ((u64)next_offset << KVM_ITS_DTE_NEXT_SHIFT) | + (itt_addr_field << KVM_ITS_DTE_ITTADDR_SHIFT) | + (dev->num_eventid_bits - 1)); + val = cpu_to_le64(val); + return kvm_write_guest(kvm, ptr, &val, dte_esz); +} + +/** + * vgic_its_restore_dte - restore a device table entry + * + * @its: its handle + * @id: device id the DTE corresponds to + * @ptr: kernel VA where the 8 byte DTE is located + * @opaque: unused + * + * Return: < 0 on error, 0 if the dte is the last one, id offset to the + * next dte otherwise + */ +static int vgic_its_restore_dte(struct vgic_its *its, u32 id, + void *ptr, void *opaque) +{ + struct its_device *dev; + gpa_t itt_addr; + u8 num_eventid_bits; + u64 entry = *(u64 *)ptr; + bool valid; + u32 offset; + int ret; + + entry = le64_to_cpu(entry); + + valid = entry >> KVM_ITS_DTE_VALID_SHIFT; + num_eventid_bits = (entry & KVM_ITS_DTE_SIZE_MASK) + 1; + itt_addr = ((entry & KVM_ITS_DTE_ITTADDR_MASK) + >> KVM_ITS_DTE_ITTADDR_SHIFT) << 8; + + if (!valid) + return 1; + + /* dte entry is valid */ + offset = (entry & KVM_ITS_DTE_NEXT_MASK) >> KVM_ITS_DTE_NEXT_SHIFT; + + dev = vgic_its_alloc_device(its, id, itt_addr, num_eventid_bits); + if (IS_ERR(dev)) + return PTR_ERR(dev); + + ret = vgic_its_restore_itt(its, dev); + if (ret) { + vgic_its_free_device(its->dev->kvm, dev); + return ret; + } + + return offset; +} + +static int vgic_its_device_cmp(void *priv, struct list_head *a, + struct list_head *b) +{ + struct its_device *deva = container_of(a, struct its_device, dev_list); + struct its_device *devb = container_of(b, struct its_device, dev_list); + + if (deva->device_id < devb->device_id) + return -1; + else + return 1; +} + +/** + * vgic_its_save_device_tables - Save the device table and all ITT + * into guest RAM + * + * L1/L2 handling is hidden by vgic_its_check_id() helper which directly + * returns the GPA of the device entry + */ +static int vgic_its_save_device_tables(struct vgic_its *its) +{ + const struct vgic_its_abi *abi = vgic_its_get_abi(its); + struct its_device *dev; + int dte_esz = abi->dte_esz; + u64 baser; + + baser = its->baser_device_table; + + list_sort(NULL, &its->device_list, vgic_its_device_cmp); + + list_for_each_entry(dev, &its->device_list, dev_list) { + int ret; + gpa_t eaddr; + + if (!vgic_its_check_id(its, baser, + dev->device_id, &eaddr)) + return -EINVAL; + + ret = vgic_its_save_itt(its, dev); + if (ret) + return ret; + + ret = vgic_its_save_dte(its, dev, eaddr, dte_esz); + if (ret) + return ret; + } + return 0; +} + +/** + * handle_l1_dte - callback used for L1 device table entries (2 stage case) + * + * @its: its handle + * @id: index of the entry in the L1 table + * @addr: kernel VA + * @opaque: unused + * + * L1 table entries are scanned by steps of 1 entry + * Return < 0 if error, 0 if last dte was found when scanning the L2 + * table, +1 otherwise (meaning next L1 entry must be scanned) + */ +static int handle_l1_dte(struct vgic_its *its, u32 id, void *addr, + void *opaque) +{ + const struct vgic_its_abi *abi = vgic_its_get_abi(its); + int l2_start_id = id * (SZ_64K / abi->dte_esz); + u64 entry = *(u64 *)addr; + int dte_esz = abi->dte_esz; + gpa_t gpa; + int ret; + + entry = le64_to_cpu(entry); + + if (!(entry & KVM_ITS_L1E_VALID_MASK)) + return 1; + + gpa = entry & KVM_ITS_L1E_ADDR_MASK; + + ret = scan_its_table(its, gpa, SZ_64K, dte_esz, + l2_start_id, vgic_its_restore_dte, NULL); + + if (ret <= 0) + return ret; + + return 1; +} + +/** + * vgic_its_restore_device_tables - Restore the device table and all ITT + * from guest RAM to internal data structs + */ +static int vgic_its_restore_device_tables(struct vgic_its *its) +{ + const struct vgic_its_abi *abi = vgic_its_get_abi(its); + u64 baser = its->baser_device_table; + int l1_esz, ret; + int l1_tbl_size = GITS_BASER_NR_PAGES(baser) * SZ_64K; + gpa_t l1_gpa; + + if (!(baser & GITS_BASER_VALID)) + return 0; + + l1_gpa = BASER_ADDRESS(baser); + + if (baser & GITS_BASER_INDIRECT) { + l1_esz = GITS_LVL1_ENTRY_SIZE; + ret = scan_its_table(its, l1_gpa, l1_tbl_size, l1_esz, 0, + handle_l1_dte, NULL); + } else { + l1_esz = abi->dte_esz; + ret = scan_its_table(its, l1_gpa, l1_tbl_size, l1_esz, 0, + vgic_its_restore_dte, NULL); + } + + if (ret > 0) + ret = -EINVAL; + + return ret; +} + +static int vgic_its_save_cte(struct vgic_its *its, + struct its_collection *collection, + gpa_t gpa, int esz) +{ + u64 val; + + val = (1ULL << KVM_ITS_CTE_VALID_SHIFT | + ((u64)collection->target_addr << KVM_ITS_CTE_RDBASE_SHIFT) | + collection->collection_id); + val = cpu_to_le64(val); + return kvm_write_guest(its->dev->kvm, gpa, &val, esz); +} + +static int vgic_its_restore_cte(struct vgic_its *its, gpa_t gpa, int esz) +{ + struct its_collection *collection; + struct kvm *kvm = its->dev->kvm; + u32 target_addr, coll_id; + u64 val; + int ret; + + BUG_ON(esz > sizeof(val)); + ret = kvm_read_guest(kvm, gpa, &val, esz); + if (ret) + return ret; + val = le64_to_cpu(val); + if (!(val & KVM_ITS_CTE_VALID_MASK)) + return 0; + + target_addr = (u32)(val >> KVM_ITS_CTE_RDBASE_SHIFT); + coll_id = val & KVM_ITS_CTE_ICID_MASK; + + if (target_addr >= atomic_read(&kvm->online_vcpus)) + return -EINVAL; + + collection = find_collection(its, coll_id); + if (collection) + return -EEXIST; + ret = vgic_its_alloc_collection(its, &collection, coll_id); + if (ret) + return ret; + collection->target_addr = target_addr; + return 1; +} + +/** + * vgic_its_save_collection_table - Save the collection table into + * guest RAM + */ +static int vgic_its_save_collection_table(struct vgic_its *its) +{ + const struct vgic_its_abi *abi = vgic_its_get_abi(its); + struct its_collection *collection; + u64 val; + gpa_t gpa; + size_t max_size, filled = 0; + int ret, cte_esz = abi->cte_esz; + + gpa = BASER_ADDRESS(its->baser_coll_table); + if (!gpa) + return 0; + + max_size = GITS_BASER_NR_PAGES(its->baser_coll_table) * SZ_64K; + + list_for_each_entry(collection, &its->collection_list, coll_list) { + ret = vgic_its_save_cte(its, collection, gpa, cte_esz); + if (ret) + return ret; + gpa += cte_esz; + filled += cte_esz; + } + + if (filled == max_size) + return 0; + + /* + * table is not fully filled, add a last dummy element + * with valid bit unset + */ + val = 0; + BUG_ON(cte_esz > sizeof(val)); + ret = kvm_write_guest(its->dev->kvm, gpa, &val, cte_esz); + return ret; +} + +/** + * vgic_its_restore_collection_table - reads the collection table + * in guest memory and restores the ITS internal state. Requires the + * BASER registers to be restored before. + */ +static int vgic_its_restore_collection_table(struct vgic_its *its) +{ + const struct vgic_its_abi *abi = vgic_its_get_abi(its); + int cte_esz = abi->cte_esz; + size_t max_size, read = 0; + gpa_t gpa; + int ret; + + if (!(its->baser_coll_table & GITS_BASER_VALID)) + return 0; + + gpa = BASER_ADDRESS(its->baser_coll_table); + + max_size = GITS_BASER_NR_PAGES(its->baser_coll_table) * SZ_64K; + + while (read < max_size) { + ret = vgic_its_restore_cte(its, gpa, cte_esz); + if (ret <= 0) + break; + gpa += cte_esz; + read += cte_esz; + } + return ret; +} + +/** + * vgic_its_save_tables_v0 - Save the ITS tables into guest ARM + * according to v0 ABI + */ +static int vgic_its_save_tables_v0(struct vgic_its *its) +{ + struct kvm *kvm = its->dev->kvm; + int ret; + + mutex_lock(&kvm->lock); + mutex_lock(&its->its_lock); + + if (!lock_all_vcpus(kvm)) { + mutex_unlock(&its->its_lock); + mutex_unlock(&kvm->lock); + return -EBUSY; + } + + ret = vgic_its_save_device_tables(its); + if (ret) + goto out; + + ret = vgic_its_save_collection_table(its); + +out: + unlock_all_vcpus(kvm); + mutex_unlock(&its->its_lock); + mutex_unlock(&kvm->lock); + return ret; +} + +/** + * vgic_its_restore_tables_v0 - Restore the ITS tables from guest RAM + * to internal data structs according to V0 ABI + * + */ +static int vgic_its_restore_tables_v0(struct vgic_its *its) +{ + struct kvm *kvm = its->dev->kvm; + int ret; + + mutex_lock(&kvm->lock); + mutex_lock(&its->its_lock); + + if (!lock_all_vcpus(kvm)) { + mutex_unlock(&its->its_lock); + mutex_unlock(&kvm->lock); + return -EBUSY; + } + + ret = vgic_its_restore_collection_table(its); + if (ret) + goto out; + + ret = vgic_its_restore_device_tables(its); +out: + unlock_all_vcpus(kvm); + mutex_unlock(&its->its_lock); + mutex_unlock(&kvm->lock); + + return ret; +} + +static int vgic_its_commit_v0(struct vgic_its *its) +{ + const struct vgic_its_abi *abi; + + abi = vgic_its_get_abi(its); + its->baser_coll_table &= ~GITS_BASER_ENTRY_SIZE_MASK; + its->baser_device_table &= ~GITS_BASER_ENTRY_SIZE_MASK; + + its->baser_coll_table |= (GIC_ENCODE_SZ(abi->cte_esz, 5) + << GITS_BASER_ENTRY_SIZE_SHIFT); + + its->baser_device_table |= (GIC_ENCODE_SZ(abi->dte_esz, 5) + << GITS_BASER_ENTRY_SIZE_SHIFT); + return 0; +} + static int vgic_its_has_attr(struct kvm_device *dev, struct kvm_device_attr *attr) { @@ -1480,8 +2355,14 @@ static int vgic_its_has_attr(struct kvm_device *dev, switch (attr->attr) { case KVM_DEV_ARM_VGIC_CTRL_INIT: return 0; + case KVM_DEV_ARM_ITS_SAVE_TABLES: + return 0; + case KVM_DEV_ARM_ITS_RESTORE_TABLES: + return 0; } break; + case KVM_DEV_ARM_VGIC_GRP_ITS_REGS: + return vgic_its_has_attr_regs(dev, attr); } return -ENXIO; } @@ -1509,18 +2390,30 @@ static int vgic_its_set_attr(struct kvm_device *dev, if (ret) return ret; - its->vgic_its_base = addr; - - return 0; + return vgic_register_its_iodev(dev->kvm, its, addr); } - case KVM_DEV_ARM_VGIC_GRP_CTRL: + case KVM_DEV_ARM_VGIC_GRP_CTRL: { + const struct vgic_its_abi *abi = vgic_its_get_abi(its); + switch (attr->attr) { case KVM_DEV_ARM_VGIC_CTRL_INIT: - its->initialized = true; - + /* Nothing to do */ return 0; + case KVM_DEV_ARM_ITS_SAVE_TABLES: + return abi->save_tables(its); + case KVM_DEV_ARM_ITS_RESTORE_TABLES: + return abi->restore_tables(its); } - break; + } + case KVM_DEV_ARM_VGIC_GRP_ITS_REGS: { + u64 __user *uaddr = (u64 __user *)(long)attr->addr; + u64 reg; + + if (get_user(reg, uaddr)) + return -EFAULT; + + return vgic_its_attr_regs_access(dev, attr, ®, true); + } } return -ENXIO; } @@ -1541,10 +2434,20 @@ static int vgic_its_get_attr(struct kvm_device *dev, if (copy_to_user(uaddr, &addr, sizeof(addr))) return -EFAULT; break; + } + case KVM_DEV_ARM_VGIC_GRP_ITS_REGS: { + u64 __user *uaddr = (u64 __user *)(long)attr->addr; + u64 reg; + int ret; + + ret = vgic_its_attr_regs_access(dev, attr, ®, false); + if (ret) + return ret; + return put_user(reg, uaddr); + } default: return -ENXIO; } - } return 0; } @@ -1563,30 +2466,3 @@ int kvm_vgic_register_its_device(void) return kvm_register_device_ops(&kvm_arm_vgic_its_ops, KVM_DEV_TYPE_ARM_VGIC_ITS); } - -/* - * Registers all ITSes with the kvm_io_bus framework. - * To follow the existing VGIC initialization sequence, this has to be - * done as late as possible, just before the first VCPU runs. - */ -int vgic_register_its_iodevs(struct kvm *kvm) -{ - struct kvm_device *dev; - int ret = 0; - - list_for_each_entry(dev, &kvm->devices, vm_node) { - if (dev->ops != &kvm_arm_vgic_its_ops) - continue; - - ret = vgic_register_its_iodev(kvm, dev->private); - if (ret) - return ret; - /* - * We don't need to care about tearing down previously - * registered ITSes, as the kvm_io_bus framework removes - * them for us if the VM gets destroyed. - */ - } - - return ret; -} diff --git a/virt/kvm/arm/vgic/vgic-kvm-device.c b/virt/kvm/arm/vgic/vgic-kvm-device.c index d181d2baee9c..10ae6f394b71 100644 --- a/virt/kvm/arm/vgic/vgic-kvm-device.c +++ b/virt/kvm/arm/vgic/vgic-kvm-device.c @@ -37,6 +37,14 @@ int vgic_check_ioaddr(struct kvm *kvm, phys_addr_t *ioaddr, return 0; } +static int vgic_check_type(struct kvm *kvm, int type_needed) +{ + if (kvm->arch.vgic.vgic_model != type_needed) + return -ENODEV; + else + return 0; +} + /** * kvm_vgic_addr - set or get vgic VM base addresses * @kvm: pointer to the vm struct @@ -57,40 +65,41 @@ int kvm_vgic_addr(struct kvm *kvm, unsigned long type, u64 *addr, bool write) { int r = 0; struct vgic_dist *vgic = &kvm->arch.vgic; - int type_needed; phys_addr_t *addr_ptr, alignment; mutex_lock(&kvm->lock); switch (type) { case KVM_VGIC_V2_ADDR_TYPE_DIST: - type_needed = KVM_DEV_TYPE_ARM_VGIC_V2; + r = vgic_check_type(kvm, KVM_DEV_TYPE_ARM_VGIC_V2); addr_ptr = &vgic->vgic_dist_base; alignment = SZ_4K; break; case KVM_VGIC_V2_ADDR_TYPE_CPU: - type_needed = KVM_DEV_TYPE_ARM_VGIC_V2; + r = vgic_check_type(kvm, KVM_DEV_TYPE_ARM_VGIC_V2); addr_ptr = &vgic->vgic_cpu_base; alignment = SZ_4K; break; case KVM_VGIC_V3_ADDR_TYPE_DIST: - type_needed = KVM_DEV_TYPE_ARM_VGIC_V3; + r = vgic_check_type(kvm, KVM_DEV_TYPE_ARM_VGIC_V3); addr_ptr = &vgic->vgic_dist_base; alignment = SZ_64K; break; case KVM_VGIC_V3_ADDR_TYPE_REDIST: - type_needed = KVM_DEV_TYPE_ARM_VGIC_V3; + r = vgic_check_type(kvm, KVM_DEV_TYPE_ARM_VGIC_V3); + if (r) + break; + if (write) { + r = vgic_v3_set_redist_base(kvm, *addr); + goto out; + } addr_ptr = &vgic->vgic_redist_base; - alignment = SZ_64K; break; default: r = -ENODEV; - goto out; } - if (vgic->vgic_model != type_needed) { - r = -ENODEV; + if (r) goto out; - } if (write) { r = vgic_check_ioaddr(kvm, addr_ptr, *addr, alignment); @@ -259,13 +268,13 @@ static void unlock_vcpus(struct kvm *kvm, int vcpu_lock_idx) } } -static void unlock_all_vcpus(struct kvm *kvm) +void unlock_all_vcpus(struct kvm *kvm) { unlock_vcpus(kvm, atomic_read(&kvm->online_vcpus) - 1); } /* Returns true if all vcpus were locked, false otherwise */ -static bool lock_all_vcpus(struct kvm *kvm) +bool lock_all_vcpus(struct kvm *kvm) { struct kvm_vcpu *tmp_vcpu; int c; @@ -580,6 +589,24 @@ static int vgic_v3_set_attr(struct kvm_device *dev, reg = tmp32; return vgic_v3_attr_regs_access(dev, attr, ®, true); } + case KVM_DEV_ARM_VGIC_GRP_CTRL: { + int ret; + + switch (attr->attr) { + case KVM_DEV_ARM_VGIC_SAVE_PENDING_TABLES: + mutex_lock(&dev->kvm->lock); + + if (!lock_all_vcpus(dev->kvm)) { + mutex_unlock(&dev->kvm->lock); + return -EBUSY; + } + ret = vgic_v3_save_pending_tables(dev->kvm); + unlock_all_vcpus(dev->kvm); + mutex_unlock(&dev->kvm->lock); + return ret; + } + break; + } } return -ENXIO; } @@ -658,6 +685,8 @@ static int vgic_v3_has_attr(struct kvm_device *dev, switch (attr->attr) { case KVM_DEV_ARM_VGIC_CTRL_INIT: return 0; + case KVM_DEV_ARM_VGIC_SAVE_PENDING_TABLES: + return 0; } } return -ENXIO; diff --git a/virt/kvm/arm/vgic/vgic-mmio-v3.c b/virt/kvm/arm/vgic/vgic-mmio-v3.c index 6afb3b484886..99da1a207c19 100644 --- a/virt/kvm/arm/vgic/vgic-mmio-v3.c +++ b/virt/kvm/arm/vgic/vgic-mmio-v3.c @@ -556,67 +556,130 @@ unsigned int vgic_v3_init_dist_iodev(struct vgic_io_device *dev) return SZ_64K; } -int vgic_register_redist_iodevs(struct kvm *kvm, gpa_t redist_base_address) +/** + * vgic_register_redist_iodev - register a single redist iodev + * @vcpu: The VCPU to which the redistributor belongs + * + * Register a KVM iodev for this VCPU's redistributor using the address + * provided. + * + * Return 0 on success, -ERRNO otherwise. + */ +int vgic_register_redist_iodev(struct kvm_vcpu *vcpu) +{ + struct kvm *kvm = vcpu->kvm; + struct vgic_dist *vgic = &kvm->arch.vgic; + struct vgic_io_device *rd_dev = &vcpu->arch.vgic_cpu.rd_iodev; + struct vgic_io_device *sgi_dev = &vcpu->arch.vgic_cpu.sgi_iodev; + gpa_t rd_base, sgi_base; + int ret; + + /* + * We may be creating VCPUs before having set the base address for the + * redistributor region, in which case we will come back to this + * function for all VCPUs when the base address is set. Just return + * without doing any work for now. + */ + if (IS_VGIC_ADDR_UNDEF(vgic->vgic_redist_base)) + return 0; + + if (!vgic_v3_check_base(kvm)) + return -EINVAL; + + rd_base = vgic->vgic_redist_base + kvm_vcpu_get_idx(vcpu) * SZ_64K * 2; + sgi_base = rd_base + SZ_64K; + + kvm_iodevice_init(&rd_dev->dev, &kvm_io_gic_ops); + rd_dev->base_addr = rd_base; + rd_dev->iodev_type = IODEV_REDIST; + rd_dev->regions = vgic_v3_rdbase_registers; + rd_dev->nr_regions = ARRAY_SIZE(vgic_v3_rdbase_registers); + rd_dev->redist_vcpu = vcpu; + + mutex_lock(&kvm->slots_lock); + ret = kvm_io_bus_register_dev(kvm, KVM_MMIO_BUS, rd_base, + SZ_64K, &rd_dev->dev); + mutex_unlock(&kvm->slots_lock); + + if (ret) + return ret; + + kvm_iodevice_init(&sgi_dev->dev, &kvm_io_gic_ops); + sgi_dev->base_addr = sgi_base; + sgi_dev->iodev_type = IODEV_REDIST; + sgi_dev->regions = vgic_v3_sgibase_registers; + sgi_dev->nr_regions = ARRAY_SIZE(vgic_v3_sgibase_registers); + sgi_dev->redist_vcpu = vcpu; + + mutex_lock(&kvm->slots_lock); + ret = kvm_io_bus_register_dev(kvm, KVM_MMIO_BUS, sgi_base, + SZ_64K, &sgi_dev->dev); + mutex_unlock(&kvm->slots_lock); + if (ret) + kvm_io_bus_unregister_dev(kvm, KVM_MMIO_BUS, + &rd_dev->dev); + + return ret; +} + +static void vgic_unregister_redist_iodev(struct kvm_vcpu *vcpu) +{ + struct vgic_io_device *rd_dev = &vcpu->arch.vgic_cpu.rd_iodev; + struct vgic_io_device *sgi_dev = &vcpu->arch.vgic_cpu.sgi_iodev; + + kvm_io_bus_unregister_dev(vcpu->kvm, KVM_MMIO_BUS, &rd_dev->dev); + kvm_io_bus_unregister_dev(vcpu->kvm, KVM_MMIO_BUS, &sgi_dev->dev); +} + +static int vgic_register_all_redist_iodevs(struct kvm *kvm) { struct kvm_vcpu *vcpu; int c, ret = 0; kvm_for_each_vcpu(c, vcpu, kvm) { - gpa_t rd_base = redist_base_address + c * SZ_64K * 2; - gpa_t sgi_base = rd_base + SZ_64K; - struct vgic_io_device *rd_dev = &vcpu->arch.vgic_cpu.rd_iodev; - struct vgic_io_device *sgi_dev = &vcpu->arch.vgic_cpu.sgi_iodev; - - kvm_iodevice_init(&rd_dev->dev, &kvm_io_gic_ops); - rd_dev->base_addr = rd_base; - rd_dev->iodev_type = IODEV_REDIST; - rd_dev->regions = vgic_v3_rdbase_registers; - rd_dev->nr_regions = ARRAY_SIZE(vgic_v3_rdbase_registers); - rd_dev->redist_vcpu = vcpu; - - mutex_lock(&kvm->slots_lock); - ret = kvm_io_bus_register_dev(kvm, KVM_MMIO_BUS, rd_base, - SZ_64K, &rd_dev->dev); - mutex_unlock(&kvm->slots_lock); - + ret = vgic_register_redist_iodev(vcpu); if (ret) break; - - kvm_iodevice_init(&sgi_dev->dev, &kvm_io_gic_ops); - sgi_dev->base_addr = sgi_base; - sgi_dev->iodev_type = IODEV_REDIST; - sgi_dev->regions = vgic_v3_sgibase_registers; - sgi_dev->nr_regions = ARRAY_SIZE(vgic_v3_sgibase_registers); - sgi_dev->redist_vcpu = vcpu; - - mutex_lock(&kvm->slots_lock); - ret = kvm_io_bus_register_dev(kvm, KVM_MMIO_BUS, sgi_base, - SZ_64K, &sgi_dev->dev); - mutex_unlock(&kvm->slots_lock); - if (ret) { - kvm_io_bus_unregister_dev(kvm, KVM_MMIO_BUS, - &rd_dev->dev); - break; - } } if (ret) { /* The current c failed, so we start with the previous one. */ for (c--; c >= 0; c--) { - struct vgic_cpu *vgic_cpu; - vcpu = kvm_get_vcpu(kvm, c); - vgic_cpu = &vcpu->arch.vgic_cpu; - kvm_io_bus_unregister_dev(kvm, KVM_MMIO_BUS, - &vgic_cpu->rd_iodev.dev); - kvm_io_bus_unregister_dev(kvm, KVM_MMIO_BUS, - &vgic_cpu->sgi_iodev.dev); + vgic_unregister_redist_iodev(vcpu); } } return ret; } +int vgic_v3_set_redist_base(struct kvm *kvm, u64 addr) +{ + struct vgic_dist *vgic = &kvm->arch.vgic; + int ret; + + /* vgic_check_ioaddr makes sure we don't do this twice */ + ret = vgic_check_ioaddr(kvm, &vgic->vgic_redist_base, addr, SZ_64K); + if (ret) + return ret; + + vgic->vgic_redist_base = addr; + if (!vgic_v3_check_base(kvm)) { + vgic->vgic_redist_base = VGIC_ADDR_UNDEF; + return -EINVAL; + } + + /* + * Register iodevs for each existing VCPU. Adding more VCPUs + * afterwards will register the iodevs when needed. + */ + ret = vgic_register_all_redist_iodevs(kvm); + if (ret) + return ret; + + return 0; +} + int vgic_v3_has_attr_regs(struct kvm_device *dev, struct kvm_device_attr *attr) { const struct vgic_register_region *region; diff --git a/virt/kvm/arm/vgic/vgic-mmio.c b/virt/kvm/arm/vgic/vgic-mmio.c index 2a5db1352722..1c17b2a2f105 100644 --- a/virt/kvm/arm/vgic/vgic-mmio.c +++ b/virt/kvm/arm/vgic/vgic-mmio.c @@ -446,13 +446,12 @@ static int match_region(const void *key, const void *elt) return 0; } -/* Find the proper register handler entry given a certain address offset. */ -static const struct vgic_register_region * -vgic_find_mmio_region(const struct vgic_register_region *region, int nr_regions, - unsigned int offset) +const struct vgic_register_region * +vgic_find_mmio_region(const struct vgic_register_region *regions, + int nr_regions, unsigned int offset) { - return bsearch((void *)(uintptr_t)offset, region, nr_regions, - sizeof(region[0]), match_region); + return bsearch((void *)(uintptr_t)offset, regions, nr_regions, + sizeof(regions[0]), match_region); } void vgic_set_vmcr(struct kvm_vcpu *vcpu, struct vgic_vmcr *vmcr) diff --git a/virt/kvm/arm/vgic/vgic-mmio.h b/virt/kvm/arm/vgic/vgic-mmio.h index 98bb566b660a..ea4171acdef3 100644 --- a/virt/kvm/arm/vgic/vgic-mmio.h +++ b/virt/kvm/arm/vgic/vgic-mmio.h @@ -36,8 +36,13 @@ struct vgic_register_region { }; unsigned long (*uaccess_read)(struct kvm_vcpu *vcpu, gpa_t addr, unsigned int len); - void (*uaccess_write)(struct kvm_vcpu *vcpu, gpa_t addr, - unsigned int len, unsigned long val); + union { + void (*uaccess_write)(struct kvm_vcpu *vcpu, gpa_t addr, + unsigned int len, unsigned long val); + int (*uaccess_its_write)(struct kvm *kvm, struct vgic_its *its, + gpa_t addr, unsigned int len, + unsigned long val); + }; }; extern struct kvm_io_device_ops kvm_io_gic_ops; @@ -192,4 +197,9 @@ u64 vgic_sanitise_shareability(u64 reg); u64 vgic_sanitise_field(u64 reg, u64 field_mask, int field_shift, u64 (*sanitise_fn)(u64)); +/* Find the proper register handler entry given a certain address offset */ +const struct vgic_register_region * +vgic_find_mmio_region(const struct vgic_register_region *regions, + int nr_regions, unsigned int offset); + #endif diff --git a/virt/kvm/arm/vgic/vgic-v3.c b/virt/kvm/arm/vgic/vgic-v3.c index df1503650300..8fa737edde6f 100644 --- a/virt/kvm/arm/vgic/vgic-v3.c +++ b/virt/kvm/arm/vgic/vgic-v3.c @@ -234,19 +234,125 @@ void vgic_v3_enable(struct kvm_vcpu *vcpu) vgic_v3->vgic_hcr = ICH_HCR_EN; } -/* check for overlapping regions and for regions crossing the end of memory */ -static bool vgic_v3_check_base(struct kvm *kvm) +int vgic_v3_lpi_sync_pending_status(struct kvm *kvm, struct vgic_irq *irq) +{ + struct kvm_vcpu *vcpu; + int byte_offset, bit_nr; + gpa_t pendbase, ptr; + bool status; + u8 val; + int ret; + +retry: + vcpu = irq->target_vcpu; + if (!vcpu) + return 0; + + pendbase = GICR_PENDBASER_ADDRESS(vcpu->arch.vgic_cpu.pendbaser); + + byte_offset = irq->intid / BITS_PER_BYTE; + bit_nr = irq->intid % BITS_PER_BYTE; + ptr = pendbase + byte_offset; + + ret = kvm_read_guest(kvm, ptr, &val, 1); + if (ret) + return ret; + + status = val & (1 << bit_nr); + + spin_lock(&irq->irq_lock); + if (irq->target_vcpu != vcpu) { + spin_unlock(&irq->irq_lock); + goto retry; + } + irq->pending_latch = status; + vgic_queue_irq_unlock(vcpu->kvm, irq); + + if (status) { + /* clear consumed data */ + val &= ~(1 << bit_nr); + ret = kvm_write_guest(kvm, ptr, &val, 1); + if (ret) + return ret; + } + return 0; +} + +/** + * vgic_its_save_pending_tables - Save the pending tables into guest RAM + * kvm lock and all vcpu lock must be held + */ +int vgic_v3_save_pending_tables(struct kvm *kvm) +{ + struct vgic_dist *dist = &kvm->arch.vgic; + int last_byte_offset = -1; + struct vgic_irq *irq; + int ret; + + list_for_each_entry(irq, &dist->lpi_list_head, lpi_list) { + int byte_offset, bit_nr; + struct kvm_vcpu *vcpu; + gpa_t pendbase, ptr; + bool stored; + u8 val; + + vcpu = irq->target_vcpu; + if (!vcpu) + continue; + + pendbase = GICR_PENDBASER_ADDRESS(vcpu->arch.vgic_cpu.pendbaser); + + byte_offset = irq->intid / BITS_PER_BYTE; + bit_nr = irq->intid % BITS_PER_BYTE; + ptr = pendbase + byte_offset; + + if (byte_offset != last_byte_offset) { + ret = kvm_read_guest(kvm, ptr, &val, 1); + if (ret) + return ret; + last_byte_offset = byte_offset; + } + + stored = val & (1U << bit_nr); + if (stored == irq->pending_latch) + continue; + + if (irq->pending_latch) + val |= 1 << bit_nr; + else + val &= ~(1 << bit_nr); + + ret = kvm_write_guest(kvm, ptr, &val, 1); + if (ret) + return ret; + } + return 0; +} + +/* + * Check for overlapping regions and for regions crossing the end of memory + * for base addresses which have already been set. + */ +bool vgic_v3_check_base(struct kvm *kvm) { struct vgic_dist *d = &kvm->arch.vgic; gpa_t redist_size = KVM_VGIC_V3_REDIST_SIZE; redist_size *= atomic_read(&kvm->online_vcpus); - if (d->vgic_dist_base + KVM_VGIC_V3_DIST_SIZE < d->vgic_dist_base) + if (!IS_VGIC_ADDR_UNDEF(d->vgic_dist_base) && + d->vgic_dist_base + KVM_VGIC_V3_DIST_SIZE < d->vgic_dist_base) return false; - if (d->vgic_redist_base + redist_size < d->vgic_redist_base) + + if (!IS_VGIC_ADDR_UNDEF(d->vgic_redist_base) && + d->vgic_redist_base + redist_size < d->vgic_redist_base) return false; + /* Both base addresses must be set to check if they overlap */ + if (IS_VGIC_ADDR_UNDEF(d->vgic_dist_base) || + IS_VGIC_ADDR_UNDEF(d->vgic_redist_base)) + return true; + if (d->vgic_dist_base + KVM_VGIC_V3_DIST_SIZE <= d->vgic_redist_base) return true; if (d->vgic_redist_base + redist_size <= d->vgic_dist_base) @@ -291,20 +397,6 @@ int vgic_v3_map_resources(struct kvm *kvm) goto out; } - ret = vgic_register_redist_iodevs(kvm, dist->vgic_redist_base); - if (ret) { - kvm_err("Unable to register VGICv3 redist MMIO regions\n"); - goto out; - } - - if (vgic_has_its(kvm)) { - ret = vgic_register_its_iodevs(kvm); - if (ret) { - kvm_err("Unable to register VGIC ITS MMIO regions\n"); - goto out; - } - } - dist->ready = true; out: diff --git a/virt/kvm/arm/vgic/vgic.c b/virt/kvm/arm/vgic/vgic.c index 4346bc7d08dc..83b24d20ff8f 100644 --- a/virt/kvm/arm/vgic/vgic.c +++ b/virt/kvm/arm/vgic/vgic.c @@ -21,7 +21,7 @@ #include "vgic.h" #define CREATE_TRACE_POINTS -#include "../trace.h" +#include "trace.h" #ifdef CONFIG_DEBUG_SPINLOCK #define DEBUG_SPINLOCK_BUG_ON(p) BUG_ON(p) diff --git a/virt/kvm/arm/vgic/vgic.h b/virt/kvm/arm/vgic/vgic.h index 799fd651b260..da83e4caa272 100644 --- a/virt/kvm/arm/vgic/vgic.h +++ b/virt/kvm/arm/vgic/vgic.h @@ -73,6 +73,29 @@ KVM_REG_ARM_VGIC_SYSREG_CRM_MASK | \ KVM_REG_ARM_VGIC_SYSREG_OP2_MASK) +/* + * As per Documentation/virtual/kvm/devices/arm-vgic-its.txt, + * below macros are defined for ITS table entry encoding. + */ +#define KVM_ITS_CTE_VALID_SHIFT 63 +#define KVM_ITS_CTE_VALID_MASK BIT_ULL(63) +#define KVM_ITS_CTE_RDBASE_SHIFT 16 +#define KVM_ITS_CTE_ICID_MASK GENMASK_ULL(15, 0) +#define KVM_ITS_ITE_NEXT_SHIFT 48 +#define KVM_ITS_ITE_PINTID_SHIFT 16 +#define KVM_ITS_ITE_PINTID_MASK GENMASK_ULL(47, 16) +#define KVM_ITS_ITE_ICID_MASK GENMASK_ULL(15, 0) +#define KVM_ITS_DTE_VALID_SHIFT 63 +#define KVM_ITS_DTE_VALID_MASK BIT_ULL(63) +#define KVM_ITS_DTE_NEXT_SHIFT 49 +#define KVM_ITS_DTE_NEXT_MASK GENMASK_ULL(62, 49) +#define KVM_ITS_DTE_ITTADDR_SHIFT 5 +#define KVM_ITS_DTE_ITTADDR_MASK GENMASK_ULL(48, 5) +#define KVM_ITS_DTE_SIZE_MASK GENMASK_ULL(4, 0) +#define KVM_ITS_L1E_VALID_MASK BIT_ULL(63) +/* we only support 64 kB translation table page size */ +#define KVM_ITS_L1E_ADDR_MASK GENMASK_ULL(51, 16) + static inline bool irq_is_pending(struct vgic_irq *irq) { if (irq->config == VGIC_CONFIG_EDGE) @@ -157,12 +180,15 @@ void vgic_v3_get_vmcr(struct kvm_vcpu *vcpu, struct vgic_vmcr *vmcr); void vgic_v3_enable(struct kvm_vcpu *vcpu); int vgic_v3_probe(const struct gic_kvm_info *info); int vgic_v3_map_resources(struct kvm *kvm); -int vgic_register_redist_iodevs(struct kvm *kvm, gpa_t dist_base_address); +int vgic_v3_lpi_sync_pending_status(struct kvm *kvm, struct vgic_irq *irq); +int vgic_v3_save_pending_tables(struct kvm *kvm); +int vgic_v3_set_redist_base(struct kvm *kvm, u64 addr); +int vgic_register_redist_iodev(struct kvm_vcpu *vcpu); +bool vgic_v3_check_base(struct kvm *kvm); void vgic_v3_load(struct kvm_vcpu *vcpu); void vgic_v3_put(struct kvm_vcpu *vcpu); -int vgic_register_its_iodevs(struct kvm *kvm); bool vgic_has_its(struct kvm *kvm); int kvm_vgic_register_its_device(void); void vgic_enable_lpis(struct kvm_vcpu *vcpu); @@ -187,4 +213,7 @@ int vgic_init(struct kvm *kvm); int vgic_debug_init(struct kvm *kvm); int vgic_debug_destroy(struct kvm *kvm); +bool lock_all_vcpus(struct kvm *kvm); +void unlock_all_vcpus(struct kvm *kvm); + #endif diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index b3d151ee2a67..f0fe9d02f6bb 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c @@ -2836,10 +2836,6 @@ static struct kvm_device_ops *kvm_device_ops_table[KVM_DEV_TYPE_MAX] = { [KVM_DEV_TYPE_FSL_MPIC_20] = &kvm_mpic_ops, [KVM_DEV_TYPE_FSL_MPIC_42] = &kvm_mpic_ops, #endif - -#ifdef CONFIG_KVM_XICS - [KVM_DEV_TYPE_XICS] = &kvm_xics_ops, -#endif }; int kvm_register_device_ops(struct kvm_device_ops *ops, u32 type) @@ -3715,7 +3711,7 @@ static const struct file_operations vm_stat_get_per_vm_fops = { .release = kvm_debugfs_release, .read = simple_attr_read, .write = simple_attr_write, - .llseek = generic_file_llseek, + .llseek = no_llseek, }; static int vcpu_stat_get_per_vm(void *data, u64 *val) @@ -3760,7 +3756,7 @@ static const struct file_operations vcpu_stat_get_per_vm_fops = { .release = kvm_debugfs_release, .read = simple_attr_read, .write = simple_attr_write, - .llseek = generic_file_llseek, + .llseek = no_llseek, }; static const struct file_operations *stat_fops_per_vm[] = { |