From mboxrd@z Thu Jan 1 00:00:00 1970 Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.remote-ip=95.215.58.9 ARC-Seal: i=1; d=subspace.kernel.org; s=arc20261015; a=rsa-sha256; cv=none; t=1794214180; b=ci5cfDbh3XonCUK7Mxlv+z1h+2f/gNRsyynUinHnIXMRhm+lpIwP6I93vwWOltnOLOqN cQLDezlVdZjjEXVXRglqkWUZGoPavSC+FCsnzE1Sgia6HUD8IdRBZayC1bWvPM+S81nQ0 298YSfOOv+R9VqEGm7kJJPn0hiiZv3AiVVg0POS0U0i+9IiEQbmPOjNt8JaKuiU1aJnfc oP/E4yqvaagxHrTpAQe4tofO/4jOSKpIjxI7Z4sIID9cNm3DBtrwE98f30ALEjaY/Mm5G NI2A1oKEWbsi98sr61RtFKjlu+JIsJw8QFNIALdP+6anBHchXEY8HxBE8thwwFpQM3Q== ARC-Message-Signature: i=1; d=subspace.kernel.org; s=arc20261015; a=rsa-sha256; c=relaxed/relaxed; t=1794214180; h=DMARC-Filter:X-Envelope-To:DKIM-Signature:X-Envelope-To:Received: X-Mizu-Trace-ID:X-Migadu-Flow:From:To:Cc:Subject:Date:Message-ID: X-Mailer:In-Reply-To:References:MIME-Version: Content-Transfer-Encoding; bh=uWThlNDxx8k9coc2dZrQvIwrgPoJzUnjIpsnGUgbQxc=; b=Wslmex17BVHEBrbADSVBrN/ZpghQGctR2Iy8HJ6zr/GSF1lkcn39Xkb3jaen+H0SAXp9 bdhHj/m9TpCkfe1hQJvlb9zbohT6Hx1pEy74t8PlfbOgIOOLIQYueDS9hjzS+/DbLNOao QKZJnwbY0GN8/+Gxi1ZeK/KdL+3/iofAO98bicZYzb2TiSUJqJXWsjukU8GfD6Sb1pH8I 0gWdUrKvkAA1K6c9jH2SKT1+4hfkj6QgQG+2T8iBJxkijf6co0Z7mR8l+bzDJaNleBoKY UPTWoI3PFeD5wbzWty2pXiISRi67tR5McTeMGkIeLzKX91NicHcmey/4N1HV8LLv8TQ== ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass header.from=linux.dev; spf=pass smtp.mailfrom=linux.dev; dkim=pass header.d=linux.dev header.i=@linux.dev header.a=rsa-sha256 header.s=key1 header.b=cJxh6rnL; arc=none smtp.remote-ip=95.215.58.9 Received: from mta1.migadu.com (out-9.mta1.migadu.com [95.215.58.9]) (using TLSv1.3 with cipher TLS_AES_128_GCM_SHA256 (128/128 bits) key-exchange X25519 server-signature ECDSA (prime256v1) server-digest SHA256) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id C51F2608363D for ; Mon, 9 Nov 2026 08:49:39 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.4.2 smtp.subspace.kernel.org C51F2608363D Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.dev Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=linux.dev Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key, unprotected) header.d=linux.dev header.i=@linux.dev header.a=rsa-sha256 header.s=key1 header.b=cJxh6rnL X-Envelope-To: kvm@vger.kernel.org DKIM-Signature: a=rsa-sha256; bh=tieCJB5xPHFID0PWIkTmfQAHSOPwOh5HA6tVxBq05s0=; c=simple/simple; d=linux.dev; h=from:to:subject:date:message-id:mime-version:content-type; s=key1; t=1794214178; v=1; x=1794818978; b=cJxh6rnLXbxCuZPqeeBmne21czp+pycWMjO0JRGvpH6KqJB+PFnvqTX8giIRVm/+DKCn3PR0 STh/LXHDuWfcyuZgmU4tR0QKv68PXP1N748qHX3nsQC3d6J+uEU3BA1yxbpyV9VDHCnl+69UENT sWp1j76O+mDnUvM24J3XKB7U= X-Envelope-To: kvm@vger.kernel.org Received: by smtp.migadu.com with ESMTPS id 7e8d9e8db2fd6fbd; Mon, 09 Nov 2026 08:49:37 +0000 X-Mizu-Trace-ID: 7e8d9e8db2fd6fbd X-Migadu-Flow: FLOW_OUT From: Fuad Tabba To: Marc Zyngier , Oliver Upton Cc: Will Deacon , Catalin Marinas , Mark Rutland , Paolo Bonzini , Sean Christopherson , Shuah Khan , Joey Gouly , Steffen Eiden , Suzuki K Poulose , Zenghui Yu , Vincent Donnefort , Quentin Perret , Fuad Tabba , kvmarm@lists.linux.dev, linux-arm-kernel@lists.infradead.org, kvm@vger.kernel.org, linux-kselftest@vger.kernel.org, linux-kernel@vger.kernel.org Subject: [PATCH v2 14/19] KVM: arm64: selftests: Add pkvm_irq interrupt test Date: Mon, 9 Nov 2026 08:48:49 +0000 Message-ID: <20261109084854.840576-15-fuad.tabba@linux.dev> X-Mailer: git-send-email 2.47.3 In-Reply-To: <20261109084854.840576-1-fuad.tabba@linux.dev> References: <20261109084854.840576-1-fuad.tabba@linux.dev> X-Mailing-List: kvm@vger.kernel.org List-Id: List-Post: List-Help: List-Subscribe: List-Unsubscribe: List-Owner: Precedence: list MIME-Version: 1.0 Content-Transfer-Encoding: 8bit The host emulates a protected guest's vGIC, so EL2 copies the list registers to and from the host each time the host runs the vCPU. EL2 also keeps its own copy of the guest's virtual timer state, which the host can't change. Check that the virtual timer's PPI arrives exactly once, with CVAL and CTL kept across exits, that SGIs pass between two vCPUs, and that the timer wakes a vCPU in WFI. Signed-off-by: Fuad Tabba --- tools/testing/selftests/kvm/Makefile.kvm | 1 + tools/testing/selftests/kvm/arm64/pkvm_irq.c | 268 +++++++++++++++++++ 2 files changed, 269 insertions(+) create mode 100644 tools/testing/selftests/kvm/arm64/pkvm_irq.c diff --git a/tools/testing/selftests/kvm/Makefile.kvm b/tools/testing/selftests/kvm/Makefile.kvm index 1e40e793aa6f3..13998964facd5 100644 --- a/tools/testing/selftests/kvm/Makefile.kvm +++ b/tools/testing/selftests/kvm/Makefile.kvm @@ -192,6 +192,7 @@ TEST_GEN_PROGS_arm64 += arm64/mmio_sign_ext TEST_GEN_PROGS_arm64 += arm64/page_fault_test TEST_GEN_PROGS_arm64 += arm64/pkvm_host_abi TEST_GEN_PROGS_arm64 += arm64/pkvm_id_regs +TEST_GEN_PROGS_arm64 += arm64/pkvm_irq TEST_GEN_PROGS_arm64 += arm64/pkvm_mem_share TEST_GEN_PROGS_arm64 += arm64/pkvm_psci TEST_GEN_PROGS_arm64 += arm64/pkvm_smccc diff --git a/tools/testing/selftests/kvm/arm64/pkvm_irq.c b/tools/testing/selftests/kvm/arm64/pkvm_irq.c new file mode 100644 index 0000000000000..f92fa6f0de8de --- /dev/null +++ b/tools/testing/selftests/kvm/arm64/pkvm_irq.c @@ -0,0 +1,268 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * pkvm_irq - the virtual timer's PPI and SGIs in a protected VM. + * + * Copyright (C) 2026 Google LLC + * Author: Fuad Tabba + */ + +#include + +#include +#include +#include + +#include "arch_timer.h" +#include "gic.h" +#include "gic_v3.h" +#include "kvm_util.h" +#include "pkvm.h" +#include "processor.h" +#include "test_util.h" + +#define TIMER_MS 10 +#define WAIT_MS 30000 +#define NR_EXITS 16 +#define SGI_ROUNDS 16 + +#define SGI_PING 1 +#define SGI_PONG 2 + +/* Set in the guest's copy by the host before the first run. */ +static u32 vtimer_irq; +static u64 boot_entry; +static u64 vcpu_mpidr[2]; +static u64 vcpu_ctx[2]; + +static u32 irqs[2]; +static u32 last_intid[2]; +static u32 ready; + +#define wait_until(cond) \ +do { \ + u64 __end = timer_get_cntct(VIRTUAL) + msec_to_cycles(WAIT_MS); \ + \ + while (!(cond)) \ + __GUEST_ASSERT(timer_get_cntct(VIRTUAL) < __end, \ + "timed out waiting for %s", #cond); \ +} while (0) + +static void guest_irq_handler(struct ex_regs *regs) +{ + u32 intid = gic_get_and_ack_irq(); + u32 cpu = guest_get_vcpuid(); + + if (intid == IAR_SPURIOUS) + return; + + if (intid == vtimer_irq) { + __GUEST_ASSERT(timer_get_ctl(VIRTUAL) & CTL_ISTATUS, "CNTV_CTL 0x%x", + timer_get_ctl(VIRTUAL)); + GUEST_ASSERT(timer_get_cntct(VIRTUAL) >= timer_get_cval(VIRTUAL)); + /* Mask the timer before the EOI, or its level-sensitive PPI is taken again. */ + timer_set_ctl(VIRTUAL, CTL_IMASK); + } + + WRITE_ONCE(last_intid[cpu], intid); + gic_set_eoi(intid); + WRITE_ONCE(irqs[cpu], irqs[cpu] + 1); +} + +static void guest_irq_init(unsigned int nr_cpus) +{ + local_irq_disable(); + pkvm_guest_init(); + gic_init(GIC_V3, nr_cpus); +} + +static u64 start_vtimer(void) +{ + u64 cval = timer_get_cntct(VIRTUAL) + msec_to_cycles(TIMER_MS); + + gic_irq_enable(vtimer_irq); + timer_set_cval(VIRTUAL, cval); + timer_set_ctl(VIRTUAL, CTL_ENABLE); + return cval; +} + +static void guest_vtimer(void) +{ + u64 cval; + int i; + + guest_irq_init(1); + cval = start_vtimer(); + + /* Exits before the timer fires, then with its interrupt pending but masked. */ + for (i = 0; i < NR_EXITS; i++) + GUEST_SYNC(i); + wait_until(timer_get_ctl(VIRTUAL) & CTL_ISTATUS); + for (i = 0; i < NR_EXITS; i++) + GUEST_SYNC(i); + + GUEST_ASSERT_EQ(timer_get_cval(VIRTUAL), cval); + GUEST_ASSERT_EQ(timer_get_ctl(VIRTUAL), CTL_ENABLE | CTL_ISTATUS); + GUEST_ASSERT_EQ(READ_ONCE(irqs[0]), 0); + + local_irq_enable(); + wait_until(READ_ONCE(irqs[0])); + + /* The EOI must reach the host, or it presents the interrupt again. */ + for (i = 0; i < NR_EXITS; i++) + GUEST_SYNC(i); + local_irq_disable(); + + GUEST_ASSERT_EQ(READ_ONCE(irqs[0]), 1); + GUEST_ASSERT_EQ(READ_ONCE(last_intid[0]), vtimer_irq); + GUEST_DONE(); +} + +static void send_sgi(int id, u32 intid) +{ + u64 mpidr = vcpu_mpidr[id]; + + /* The target reads this vCPU's counters once it has taken the SGI. */ + dsb(ishst); + write_sysreg_s(FIELD_PREP(ICC_SGI1R_SGI_ID_MASK, intid) | + FIELD_PREP(ICC_SGI1R_AFFINITY_1_MASK, MPIDR_AFFINITY_LEVEL(mpidr, 1)) | + FIELD_PREP(ICC_SGI1R_TARGET_LIST_MASK, BIT(MPIDR_AFFINITY_LEVEL(mpidr, 0))), + SYS_ICC_SGI1R_EL1); + isb(); +} + +static void guest_sgi_primary(void) +{ + u32 k; + + guest_irq_init(2); + gic_irq_enable(SGI_PONG); + + GUEST_ASSERT_EQ(pkvm_hvc(PSCI_0_2_FN64_CPU_ON, vcpu_mpidr[1], boot_entry, vcpu_ctx[1]), + PSCI_RET_SUCCESS); + wait_until(READ_ONCE(ready)); + + local_irq_enable(); + for (k = 1; k <= SGI_ROUNDS; k++) { + send_sgi(1, SGI_PING); + + /* vCPU 1 sends PONG only once it has taken PING, so neither is in flight. */ + wait_until(READ_ONCE(irqs[0]) >= k); + GUEST_ASSERT_EQ(READ_ONCE(irqs[0]), k); + GUEST_ASSERT_EQ(READ_ONCE(irqs[1]), k); + GUEST_ASSERT_EQ(READ_ONCE(last_intid[0]), SGI_PONG); + GUEST_ASSERT_EQ(READ_ONCE(last_intid[1]), SGI_PING); + } + local_irq_disable(); + + GUEST_DONE(); +} + +/* The primary's guest_irq_init() set up ucall and the distributor. */ +static void guest_sgi_secondary(void) +{ + u32 k; + + gic_init(GIC_V3, 2); + gic_irq_enable(SGI_PING); + local_irq_enable(); + WRITE_ONCE(ready, 1); + + for (k = 1; k <= SGI_ROUNDS; k++) { + wait_until(READ_ONCE(irqs[1]) >= k); + send_sgi(0, SGI_PONG); + } + local_irq_disable(); + + GUEST_DONE(); +} + +static void guest_wfi(void) +{ + guest_irq_init(1); + start_vtimer(); + + /* WFI can complete before the timer fires, so wait in it until the handler has run. */ + while (!READ_ONCE(irqs[0])) { + wfi(); + local_irq_enable(); + isb(); /* handle IRQ */ + local_irq_disable(); + } + + GUEST_ASSERT_EQ(READ_ONCE(irqs[0]), 1); + GUEST_ASSERT_EQ(READ_ONCE(last_intid[0]), vtimer_irq); + GUEST_DONE(); +} + +static void irq_vm_setup(struct kvm_vm *vm, struct kvm_vcpu *vcpu) +{ + vm_install_exception_handler(vm, VECTOR_IRQ_CURRENT, guest_irq_handler); + write_guest_global(vm, vtimer_irq, vcpu_get_vtimer_irq(vcpu)); +} + +static void test_one_vcpu(void *guest_code) +{ + struct kvm_vcpu *vcpu; + struct kvm_vm *vm; + + vm = pkvm_vm_create_with_one_vcpu(&vcpu, guest_code); + irq_vm_setup(vm, vcpu); + + pkvm_run_loop(vcpu, NULL); + + kvm_vm_free(vm); +} + +static void wire_mpidr(struct kvm_vm *vm, struct kvm_vcpu *vcpu) +{ + u64 mpidr = vcpu_get_reg(vcpu, KVM_ARM64_SYS_REG(SYS_MPIDR_EL1)); + + write_guest_global(vm, vcpu_mpidr[vcpu->id], mpidr & MPIDR_HWID_BITMASK); +} + +static void *vcpu_thread(void *vcpu) +{ + pkvm_run_loop(vcpu, NULL); + return NULL; +} + +static void test_sgi(void) +{ + struct kvm_vcpu *vcpus[2]; + pthread_t thread; + struct kvm_vm *vm; + gpa_t ctx; + + vm = pkvm_vm_create(2, 0, vcpus); + irq_vm_setup(vm, vcpus[0]); + + pkvm_vcpu_setup_boot(vcpus[0], guest_sgi_primary); + wire_mpidr(vm, vcpus[0]); + wire_mpidr(vm, vcpus[1]); + write_guest_global(vm, boot_entry, pkvm_idmap_gpa(vm)); + ctx = pkvm_vcpu_setup_boot_secondary(vcpus[1], guest_sgi_secondary); + write_guest_global(vm, vcpu_ctx[1], ctx); + + pthread_create(&thread, NULL, vcpu_thread, vcpus[1]); + pkvm_run_loop(vcpus[0], NULL); + pthread_join(thread, NULL); + + kvm_vm_free(vm); +} + +int main(void) +{ + TEST_REQUIRE(pkvm_supported()); + + ksft_print_header(); + ksft_set_plan(3); + + test_one_vcpu(guest_vtimer); + ksft_test_result_pass("virtual timer PPI across exits\n"); + test_sgi(); + ksft_test_result_pass("SGIs between two vCPUs: %d rounds\n", SGI_ROUNDS); + test_one_vcpu(guest_wfi); + ksft_test_result_pass("the virtual timer wakes a vCPU waiting in WFI\n"); + + ksft_finished(); +} -- 2.47.3