From da142f3d373a6ddaca0119615a8db2175ddc4121 Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Fri, 5 Dec 2025 15:26:55 -0800 Subject: [PATCH 1/2] KVM: Remove subtle "struct kvm_stats_desc" pseudo-overlay Remove KVM's internal pseudo-overlay of kvm_stats_desc, which subtly aliases the flexible name[] in the uAPI definition with a fixed-size array of the same name. The unusual embedded structure results in compiler warnings due to -Wflex-array-member-not-at-end, and also necessitates an extra level of dereferencing in KVM. To avoid the "overlay", define the uAPI structure to have a fixed-size name when building for the kernel. Opportunistically clean up the indentation for the stats macros, and replace spaces with tabs. No functional change intended. Reported-by: Gustavo A. R. Silva Closes: https://lore.kernel.org/all/aPfNKRpLfhmhYqfP@kspp Acked-by: Marc Zyngier Acked-by: Christian Borntraeger [..] Acked-by: Anup Patel Reviewed-by: Bibo Mao Acked-by: Gustavo A. R. Silva Link: https://patch.msgid.link/20251205232655.445294-1-seanjc@google.com Signed-off-by: Sean Christopherson --- arch/arm64/kvm/guest.c | 4 +- arch/loongarch/kvm/vcpu.c | 2 +- arch/loongarch/kvm/vm.c | 2 +- arch/mips/kvm/mips.c | 4 +- arch/powerpc/kvm/book3s.c | 4 +- arch/powerpc/kvm/booke.c | 4 +- arch/riscv/kvm/vcpu.c | 2 +- arch/riscv/kvm/vm.c | 2 +- arch/s390/kvm/kvm-s390.c | 4 +- arch/x86/kvm/x86.c | 4 +- include/linux/kvm_host.h | 83 +++++++++++++++++---------------------- include/uapi/linux/kvm.h | 8 ++++ virt/kvm/binary_stats.c | 2 +- virt/kvm/kvm_main.c | 20 +++++----- 14 files changed, 70 insertions(+), 75 deletions(-) diff --git a/arch/arm64/kvm/guest.c b/arch/arm64/kvm/guest.c index 1c87699fd886..332c453b87cf 100644 --- a/arch/arm64/kvm/guest.c +++ b/arch/arm64/kvm/guest.c @@ -29,7 +29,7 @@ #include "trace.h" -const struct _kvm_stats_desc kvm_vm_stats_desc[] = { +const struct kvm_stats_desc kvm_vm_stats_desc[] = { KVM_GENERIC_VM_STATS() }; @@ -42,7 +42,7 @@ const struct kvm_stats_header kvm_vm_stats_header = { sizeof(kvm_vm_stats_desc), }; -const struct _kvm_stats_desc kvm_vcpu_stats_desc[] = { +const struct kvm_stats_desc kvm_vcpu_stats_desc[] = { KVM_GENERIC_VCPU_STATS(), STATS_DESC_COUNTER(VCPU, hvc_exit_stat), STATS_DESC_COUNTER(VCPU, wfe_exit_stat), diff --git a/arch/loongarch/kvm/vcpu.c b/arch/loongarch/kvm/vcpu.c index 656b954c1134..b3547bd41d21 100644 --- a/arch/loongarch/kvm/vcpu.c +++ b/arch/loongarch/kvm/vcpu.c @@ -14,7 +14,7 @@ #define CREATE_TRACE_POINTS #include "trace.h" -const struct _kvm_stats_desc kvm_vcpu_stats_desc[] = { +const struct kvm_stats_desc kvm_vcpu_stats_desc[] = { KVM_GENERIC_VCPU_STATS(), STATS_DESC_COUNTER(VCPU, int_exits), STATS_DESC_COUNTER(VCPU, idle_exits), diff --git a/arch/loongarch/kvm/vm.c b/arch/loongarch/kvm/vm.c index 194ccbcdc3b3..7deff56e0e1a 100644 --- a/arch/loongarch/kvm/vm.c +++ b/arch/loongarch/kvm/vm.c @@ -10,7 +10,7 @@ #include #include -const struct _kvm_stats_desc kvm_vm_stats_desc[] = { +const struct kvm_stats_desc kvm_vm_stats_desc[] = { KVM_GENERIC_VM_STATS(), STATS_DESC_ICOUNTER(VM, pages), STATS_DESC_ICOUNTER(VM, hugepages), diff --git a/arch/mips/kvm/mips.c b/arch/mips/kvm/mips.c index b0fb92fda4d4..23e69baad453 100644 --- a/arch/mips/kvm/mips.c +++ b/arch/mips/kvm/mips.c @@ -38,7 +38,7 @@ #define VECTORSPACING 0x100 /* for EI/VI mode */ #endif -const struct _kvm_stats_desc kvm_vm_stats_desc[] = { +const struct kvm_stats_desc kvm_vm_stats_desc[] = { KVM_GENERIC_VM_STATS() }; @@ -51,7 +51,7 @@ const struct kvm_stats_header kvm_vm_stats_header = { sizeof(kvm_vm_stats_desc), }; -const struct _kvm_stats_desc kvm_vcpu_stats_desc[] = { +const struct kvm_stats_desc kvm_vcpu_stats_desc[] = { KVM_GENERIC_VCPU_STATS(), STATS_DESC_COUNTER(VCPU, wait_exits), STATS_DESC_COUNTER(VCPU, cache_exits), diff --git a/arch/powerpc/kvm/book3s.c b/arch/powerpc/kvm/book3s.c index d79c5d1098c0..2efbe05caed7 100644 --- a/arch/powerpc/kvm/book3s.c +++ b/arch/powerpc/kvm/book3s.c @@ -38,7 +38,7 @@ /* #define EXIT_DEBUG */ -const struct _kvm_stats_desc kvm_vm_stats_desc[] = { +const struct kvm_stats_desc kvm_vm_stats_desc[] = { KVM_GENERIC_VM_STATS(), STATS_DESC_ICOUNTER(VM, num_2M_pages), STATS_DESC_ICOUNTER(VM, num_1G_pages) @@ -53,7 +53,7 @@ const struct kvm_stats_header kvm_vm_stats_header = { sizeof(kvm_vm_stats_desc), }; -const struct _kvm_stats_desc kvm_vcpu_stats_desc[] = { +const struct kvm_stats_desc kvm_vcpu_stats_desc[] = { KVM_GENERIC_VCPU_STATS(), STATS_DESC_COUNTER(VCPU, sum_exits), STATS_DESC_COUNTER(VCPU, mmio_exits), diff --git a/arch/powerpc/kvm/booke.c b/arch/powerpc/kvm/booke.c index 3401b96be475..f3ddb24ece74 100644 --- a/arch/powerpc/kvm/booke.c +++ b/arch/powerpc/kvm/booke.c @@ -36,7 +36,7 @@ unsigned long kvmppc_booke_handlers; -const struct _kvm_stats_desc kvm_vm_stats_desc[] = { +const struct kvm_stats_desc kvm_vm_stats_desc[] = { KVM_GENERIC_VM_STATS(), STATS_DESC_ICOUNTER(VM, num_2M_pages), STATS_DESC_ICOUNTER(VM, num_1G_pages) @@ -51,7 +51,7 @@ const struct kvm_stats_header kvm_vm_stats_header = { sizeof(kvm_vm_stats_desc), }; -const struct _kvm_stats_desc kvm_vcpu_stats_desc[] = { +const struct kvm_stats_desc kvm_vcpu_stats_desc[] = { KVM_GENERIC_VCPU_STATS(), STATS_DESC_COUNTER(VCPU, sum_exits), STATS_DESC_COUNTER(VCPU, mmio_exits), diff --git a/arch/riscv/kvm/vcpu.c b/arch/riscv/kvm/vcpu.c index a55a95da54d0..fdd99ac1e714 100644 --- a/arch/riscv/kvm/vcpu.c +++ b/arch/riscv/kvm/vcpu.c @@ -24,7 +24,7 @@ #define CREATE_TRACE_POINTS #include "trace.h" -const struct _kvm_stats_desc kvm_vcpu_stats_desc[] = { +const struct kvm_stats_desc kvm_vcpu_stats_desc[] = { KVM_GENERIC_VCPU_STATS(), STATS_DESC_COUNTER(VCPU, ecall_exit_stat), STATS_DESC_COUNTER(VCPU, wfi_exit_stat), diff --git a/arch/riscv/kvm/vm.c b/arch/riscv/kvm/vm.c index 66d91ae6e9b2..715a06ae8c13 100644 --- a/arch/riscv/kvm/vm.c +++ b/arch/riscv/kvm/vm.c @@ -13,7 +13,7 @@ #include #include -const struct _kvm_stats_desc kvm_vm_stats_desc[] = { +const struct kvm_stats_desc kvm_vm_stats_desc[] = { KVM_GENERIC_VM_STATS() }; static_assert(ARRAY_SIZE(kvm_vm_stats_desc) == diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c index 56a50524b3ee..495141bf0398 100644 --- a/arch/s390/kvm/kvm-s390.c +++ b/arch/s390/kvm/kvm-s390.c @@ -64,7 +64,7 @@ #define VCPU_IRQS_MAX_BUF (sizeof(struct kvm_s390_irq) * \ (KVM_MAX_VCPUS + LOCAL_IRQS)) -const struct _kvm_stats_desc kvm_vm_stats_desc[] = { +const struct kvm_stats_desc kvm_vm_stats_desc[] = { KVM_GENERIC_VM_STATS(), STATS_DESC_COUNTER(VM, inject_io), STATS_DESC_COUNTER(VM, inject_float_mchk), @@ -90,7 +90,7 @@ const struct kvm_stats_header kvm_vm_stats_header = { sizeof(kvm_vm_stats_desc), }; -const struct _kvm_stats_desc kvm_vcpu_stats_desc[] = { +const struct kvm_stats_desc kvm_vcpu_stats_desc[] = { KVM_GENERIC_VCPU_STATS(), STATS_DESC_COUNTER(VCPU, exit_userspace), STATS_DESC_COUNTER(VCPU, exit_null), diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index ff8812f3a129..69e0a033e4ac 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -237,7 +237,7 @@ EXPORT_SYMBOL_FOR_KVM_INTERNAL(enable_ipiv); bool __read_mostly enable_device_posted_irqs = true; EXPORT_SYMBOL_FOR_KVM_INTERNAL(enable_device_posted_irqs); -const struct _kvm_stats_desc kvm_vm_stats_desc[] = { +const struct kvm_stats_desc kvm_vm_stats_desc[] = { KVM_GENERIC_VM_STATS(), STATS_DESC_COUNTER(VM, mmu_shadow_zapped), STATS_DESC_COUNTER(VM, mmu_pte_write), @@ -263,7 +263,7 @@ const struct kvm_stats_header kvm_vm_stats_header = { sizeof(kvm_vm_stats_desc), }; -const struct _kvm_stats_desc kvm_vcpu_stats_desc[] = { +const struct kvm_stats_desc kvm_vcpu_stats_desc[] = { KVM_GENERIC_VCPU_STATS(), STATS_DESC_COUNTER(VCPU, pf_taken), STATS_DESC_COUNTER(VCPU, pf_fixed), diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h index d93f75b05ae2..7428d9949382 100644 --- a/include/linux/kvm_host.h +++ b/include/linux/kvm_host.h @@ -1927,56 +1927,43 @@ enum kvm_stat_kind { struct kvm_stat_data { struct kvm *kvm; - const struct _kvm_stats_desc *desc; + const struct kvm_stats_desc *desc; enum kvm_stat_kind kind; }; -struct _kvm_stats_desc { - struct kvm_stats_desc desc; - char name[KVM_STATS_NAME_SIZE]; -}; - -#define STATS_DESC_COMMON(type, unit, base, exp, sz, bsz) \ - .flags = type | unit | base | \ - BUILD_BUG_ON_ZERO(type & ~KVM_STATS_TYPE_MASK) | \ - BUILD_BUG_ON_ZERO(unit & ~KVM_STATS_UNIT_MASK) | \ - BUILD_BUG_ON_ZERO(base & ~KVM_STATS_BASE_MASK), \ - .exponent = exp, \ - .size = sz, \ +#define STATS_DESC_COMMON(type, unit, base, exp, sz, bsz) \ + .flags = type | unit | base | \ + BUILD_BUG_ON_ZERO(type & ~KVM_STATS_TYPE_MASK) | \ + BUILD_BUG_ON_ZERO(unit & ~KVM_STATS_UNIT_MASK) | \ + BUILD_BUG_ON_ZERO(base & ~KVM_STATS_BASE_MASK), \ + .exponent = exp, \ + .size = sz, \ .bucket_size = bsz -#define VM_GENERIC_STATS_DESC(stat, type, unit, base, exp, sz, bsz) \ - { \ - { \ - STATS_DESC_COMMON(type, unit, base, exp, sz, bsz), \ - .offset = offsetof(struct kvm_vm_stat, generic.stat) \ - }, \ - .name = #stat, \ - } -#define VCPU_GENERIC_STATS_DESC(stat, type, unit, base, exp, sz, bsz) \ - { \ - { \ - STATS_DESC_COMMON(type, unit, base, exp, sz, bsz), \ - .offset = offsetof(struct kvm_vcpu_stat, generic.stat) \ - }, \ - .name = #stat, \ - } -#define VM_STATS_DESC(stat, type, unit, base, exp, sz, bsz) \ - { \ - { \ - STATS_DESC_COMMON(type, unit, base, exp, sz, bsz), \ - .offset = offsetof(struct kvm_vm_stat, stat) \ - }, \ - .name = #stat, \ - } -#define VCPU_STATS_DESC(stat, type, unit, base, exp, sz, bsz) \ - { \ - { \ - STATS_DESC_COMMON(type, unit, base, exp, sz, bsz), \ - .offset = offsetof(struct kvm_vcpu_stat, stat) \ - }, \ - .name = #stat, \ - } +#define VM_GENERIC_STATS_DESC(stat, type, unit, base, exp, sz, bsz) \ +{ \ + STATS_DESC_COMMON(type, unit, base, exp, sz, bsz), \ + .offset = offsetof(struct kvm_vm_stat, generic.stat), \ + .name = #stat, \ +} +#define VCPU_GENERIC_STATS_DESC(stat, type, unit, base, exp, sz, bsz) \ +{ \ + STATS_DESC_COMMON(type, unit, base, exp, sz, bsz), \ + .offset = offsetof(struct kvm_vcpu_stat, generic.stat), \ + .name = #stat, \ +} +#define VM_STATS_DESC(stat, type, unit, base, exp, sz, bsz) \ +{ \ + STATS_DESC_COMMON(type, unit, base, exp, sz, bsz), \ + .offset = offsetof(struct kvm_vm_stat, stat), \ + .name = #stat, \ +} +#define VCPU_STATS_DESC(stat, type, unit, base, exp, sz, bsz) \ +{ \ + STATS_DESC_COMMON(type, unit, base, exp, sz, bsz), \ + .offset = offsetof(struct kvm_vcpu_stat, stat), \ + .name = #stat, \ +} /* SCOPE: VM, VM_GENERIC, VCPU, VCPU_GENERIC */ #define STATS_DESC(SCOPE, stat, type, unit, base, exp, sz, bsz) \ SCOPE##_STATS_DESC(stat, type, unit, base, exp, sz, bsz) @@ -2053,7 +2040,7 @@ struct _kvm_stats_desc { STATS_DESC_IBOOLEAN(VCPU_GENERIC, blocking) ssize_t kvm_stats_read(char *id, const struct kvm_stats_header *header, - const struct _kvm_stats_desc *desc, + const struct kvm_stats_desc *desc, void *stats, size_t size_stats, char __user *user_buffer, size_t size, loff_t *offset); @@ -2098,9 +2085,9 @@ static inline void kvm_stats_log_hist_update(u64 *data, size_t size, u64 value) extern const struct kvm_stats_header kvm_vm_stats_header; -extern const struct _kvm_stats_desc kvm_vm_stats_desc[]; +extern const struct kvm_stats_desc kvm_vm_stats_desc[]; extern const struct kvm_stats_header kvm_vcpu_stats_header; -extern const struct _kvm_stats_desc kvm_vcpu_stats_desc[]; +extern const struct kvm_stats_desc kvm_vcpu_stats_desc[]; #ifdef CONFIG_KVM_GENERIC_MMU_NOTIFIER static inline int mmu_invalidate_retry(struct kvm *kvm, unsigned long mmu_seq) diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h index dddb781b0507..76bd54848b11 100644 --- a/include/uapi/linux/kvm.h +++ b/include/uapi/linux/kvm.h @@ -14,6 +14,10 @@ #include #include +#ifdef __KERNEL__ +#include +#endif + #define KVM_API_VERSION 12 /* @@ -1579,7 +1583,11 @@ struct kvm_stats_desc { __u16 size; __u32 offset; __u32 bucket_size; +#ifdef __KERNEL__ + char name[KVM_STATS_NAME_SIZE]; +#else char name[]; +#endif }; #define KVM_GET_STATS_FD _IO(KVMIO, 0xce) diff --git a/virt/kvm/binary_stats.c b/virt/kvm/binary_stats.c index eefca6c69f51..76ce697c773b 100644 --- a/virt/kvm/binary_stats.c +++ b/virt/kvm/binary_stats.c @@ -50,7 +50,7 @@ * Return: the number of bytes that has been successfully read */ ssize_t kvm_stats_read(char *id, const struct kvm_stats_header *header, - const struct _kvm_stats_desc *desc, + const struct kvm_stats_desc *desc, void *stats, size_t size_stats, char __user *user_buffer, size_t size, loff_t *offset) { diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index 5b5b69c97665..cf65fd82d36d 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c @@ -983,9 +983,9 @@ static void kvm_free_memslots(struct kvm *kvm, struct kvm_memslots *slots) kvm_free_memslot(kvm, memslot); } -static umode_t kvm_stats_debugfs_mode(const struct _kvm_stats_desc *pdesc) +static umode_t kvm_stats_debugfs_mode(const struct kvm_stats_desc *desc) { - switch (pdesc->desc.flags & KVM_STATS_TYPE_MASK) { + switch (desc->flags & KVM_STATS_TYPE_MASK) { case KVM_STATS_TYPE_INSTANT: return 0444; case KVM_STATS_TYPE_CUMULATIVE: @@ -1020,7 +1020,7 @@ static int kvm_create_vm_debugfs(struct kvm *kvm, const char *fdname) struct dentry *dent; char dir_name[ITOA_MAX_LEN * 2]; struct kvm_stat_data *stat_data; - const struct _kvm_stats_desc *pdesc; + const struct kvm_stats_desc *pdesc; int i, ret = -ENOMEM; int kvm_debugfs_num_entries = kvm_vm_stats_header.num_desc + kvm_vcpu_stats_header.num_desc; @@ -6186,11 +6186,11 @@ static int kvm_stat_data_get(void *data, u64 *val) switch (stat_data->kind) { case KVM_STAT_VM: r = kvm_get_stat_per_vm(stat_data->kvm, - stat_data->desc->desc.offset, val); + stat_data->desc->offset, val); break; case KVM_STAT_VCPU: r = kvm_get_stat_per_vcpu(stat_data->kvm, - stat_data->desc->desc.offset, val); + stat_data->desc->offset, val); break; } @@ -6208,11 +6208,11 @@ static int kvm_stat_data_clear(void *data, u64 val) switch (stat_data->kind) { case KVM_STAT_VM: r = kvm_clear_stat_per_vm(stat_data->kvm, - stat_data->desc->desc.offset); + stat_data->desc->offset); break; case KVM_STAT_VCPU: r = kvm_clear_stat_per_vcpu(stat_data->kvm, - stat_data->desc->desc.offset); + stat_data->desc->offset); break; } @@ -6360,7 +6360,7 @@ static void kvm_uevent_notify_change(unsigned int type, struct kvm *kvm) static void kvm_init_debug(void) { const struct file_operations *fops; - const struct _kvm_stats_desc *pdesc; + const struct kvm_stats_desc *pdesc; int i; kvm_debugfs_dir = debugfs_create_dir("kvm", NULL); @@ -6373,7 +6373,7 @@ static void kvm_init_debug(void) fops = &vm_stat_readonly_fops; debugfs_create_file(pdesc->name, kvm_stats_debugfs_mode(pdesc), kvm_debugfs_dir, - (void *)(long)pdesc->desc.offset, fops); + (void *)(long)pdesc->offset, fops); } for (i = 0; i < kvm_vcpu_stats_header.num_desc; ++i) { @@ -6384,7 +6384,7 @@ static void kvm_init_debug(void) fops = &vcpu_stat_readonly_fops; debugfs_create_file(pdesc->name, kvm_stats_debugfs_mode(pdesc), kvm_debugfs_dir, - (void *)(long)pdesc->desc.offset, fops); + (void *)(long)pdesc->offset, fops); } } From f8211e95dfda702ba81ea2e3e7a8c6c967f385fa Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Mon, 2 Mar 2026 09:02:39 -0800 Subject: [PATCH 2/2] Documentation: KVM: Formalizing taking vcpu->mutex *outside* of kvm->slots_lock Explicitly document the ordering of vcpu->mutex being taken *outside* of kvm->slots_lock. While somewhat unintuitive since vCPUs conceptually have narrower scope than VMs, the scope of the owning object (vCPU versus VM) doesn't automatically carry over to the lock. In this case, vcpu->mutex has far broader scope than kvm->slots_lock. As Paolo put it, it's a "don't worry about multiple ioctls at the same time" mutex that's intended to be taken at the outer edges of KVM. More importantly, arm64 and x86 have gained flows that take kvm->slots_lock inside of vcpu->mutex. x86's kvm_inhibit_apic_access_page() is particularly nasty, as slots_lock is taken quite deep within KVM_RUN, i.e. simply swapping the ordering isn't an option. Commit to the vcpu->mutex => kvm->slots_lock ordering, as vcpu->mutex really is intended to be a "top-level" lock, whereas kvm->slots_lock is "just" a helper lock. Opportunistically document that vcpu->mutex is also taken outside of slots_arch_lock, e.g. when allocating shadow roots on x86 (which is the entire reason slots_arch_lock exists, as shadow roots must be allocated while holding kvm->srcu) kvm_mmu_new_pgd() | -> kvm_mmu_reload() | -> kvm_mmu_load() | -> mmu_alloc_shadow_roots() | -> mmu_first_shadow_root_alloc() but also when manipulating memslots in vCPU context, e.g. when inhibiting the APIC-access page via the aforementioned kvm_inhibit_apic_access_page() kvm_inhibit_apic_access_page() | -> __x86_set_memory_region() | -> kvm_set_internal_memslot() | -> kvm_set_memory_region() | -> kvm_set_memslot() Cc: Oliver Upton Cc: Marc Zyngier Link: https://patch.msgid.link/20260302170239.596810-1-seanjc@google.com Signed-off-by: Sean Christopherson --- Documentation/virt/kvm/locking.rst | 2 ++ 1 file changed, 2 insertions(+) diff --git a/Documentation/virt/kvm/locking.rst b/Documentation/virt/kvm/locking.rst index ae8bce7fecbe..662231e958a0 100644 --- a/Documentation/virt/kvm/locking.rst +++ b/Documentation/virt/kvm/locking.rst @@ -17,6 +17,8 @@ The acquisition orders for mutexes are as follows: - kvm->lock is taken outside kvm->slots_lock and kvm->irq_lock +- vcpu->mutex is taken outside kvm->slots_lock and kvm->slots_arch_lock + - kvm->slots_lock is taken outside kvm->irq_lock, though acquiring them together is quite rare.