From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from out-178.mta0.migadu.com (out-178.mta0.migadu.com [91.218.175.178]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 079181DE896 for ; Fri, 8 Nov 2024 22:25:15 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=91.218.175.178 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1731104718; cv=none; b=WKTPezR25ofadxr+VG3bIjlIWDPsSJi7Qg7Kl5dboBeFH+DmmGMXxH/H6lH0jb5jBMu2ov+h47afB5zD1ovwnF8ZEBAzTGLmrf7kJPGzC/F8rP9PMgvGlnV+WRNI+lVLF60lvlI1X5UHToleuM8IbBuH5O73lkbIMGDvgph0tMY= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1731104718; c=relaxed/simple; bh=rQzz07HE1N1UC3ykLn3hpT7QK0x3bNfI61Z9FbBc8t4=; h=From:To:Cc:Subject:Date:Message-Id:In-Reply-To:References: MIME-Version; b=fFs1bvHuM6DRZhlKsKmbQCtX1aQY35zWI0tSYPb3oDw9FUb4TZYSgADpXL0k4EWk/fLI7631ql/r01tpqQfIepYv9iSazgse0I0krO06VQmJpypskLqDckHr5ExxGzZMSMOT57PWkwNOlnIP5YgLmwdPT9f8vBuL+ZtcI5XS7qk= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.dev; spf=pass smtp.mailfrom=linux.dev; dkim=pass (1024-bit key) header.d=linux.dev header.i=@linux.dev header.b=xd17eUZE; arc=none smtp.client-ip=91.218.175.178 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) header.d=linux.dev header.i=@linux.dev header.b="xd17eUZE" X-Report-Abuse: Please report any abuse attempt to abuse@migadu.com and include these headers. DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linux.dev; s=key1; t=1731104714; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=OyhH4L6dg4wQz1xEhrEdJmNj7ibheAB3WYjLn2619GM=; b=xd17eUZEa52yDbkLIy0Xii2B2nEEL1iP/Bu0nU+jcpQeW8XOaLi+DU4hcjg6EKB6547ySw J2LFpTMpbs3TmVqttzrM0wTXaxo+qfx1H4RSyT/2xdKViqcuVZ86e6stv8gUT4qQEy6J3y YC85cNu//4ped2LN+SxQIXGfknNbSPw= From: Oliver Upton To: kvmarm@lists.linux.dev Cc: Marc Zyngier , Joey Gouly , Suzuki K Poulose , Zenghui Yu , Mingwei Zhang , Colton Lewis , Alexandru Elisei , Oliver Upton Subject: [PATCH 15/15] KVM: arm64: Manage software step state at load/put Date: Fri, 8 Nov 2024 14:24:19 -0800 Message-Id: <20241108222418.1677420-16-oliver.upton@linux.dev> In-Reply-To: <20241108222418.1677420-1-oliver.upton@linux.dev> References: <20241108222418.1677420-1-oliver.upton@linux.dev> Precedence: bulk X-Mailing-List: kvmarm@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Migadu-Flow: FLOW_OUT KVM takes over the guest's software step state machine if the VMM is debugging the guest, but it does the save/restore fiddling for every guest entry. Note that the only constraint on host usage of software step is that the guest's configuration remains visible to userspace via the ONE_REG ioctls. So, we can cut down on the amount of fiddling by doing this at load/put instead. Signed-off-by: Oliver Upton --- arch/arm64/include/asm/kvm_host.h | 24 ++--- arch/arm64/kvm/arm.c | 4 +- arch/arm64/kvm/debug.c | 149 ++++++++---------------------- arch/arm64/kvm/guest.c | 2 +- arch/arm64/kvm/handle_exit.c | 2 +- 5 files changed, 47 insertions(+), 134 deletions(-) diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h index 036c6de5819e..680f997be5f9 100644 --- a/arch/arm64/include/asm/kvm_host.h +++ b/arch/arm64/include/asm/kvm_host.h @@ -764,17 +764,6 @@ struct kvm_vcpu_arch { struct arch_timer_cpu timer_cpu; struct kvm_pmu pmu; - /* - * Guest registers we preserve during guest debugging. - * - * These shadow registers are updated by the kvm_handle_sys_reg - * trap handler if the guest accesses or updates them while we - * are using guest debug. - */ - struct { - bool pstate_ss; - } guest_debug_preserved; - /* vcpu power state */ struct kvm_mp_state mp_state; spinlock_t mp_state_lock; @@ -924,12 +913,14 @@ struct kvm_vcpu_arch { #define IN_WFIT __vcpu_single_flag(sflags, BIT(1)) /* vcpu system registers loaded on physical CPU */ #define SYSREGS_ON_CPU __vcpu_single_flag(sflags, BIT(2)) -/* Software step state is Active-pending */ -#define DBG_SS_ACTIVE_PENDING __vcpu_single_flag(sflags, BIT(3)) +/* Software step state is Active-pending for external debug */ +#define HOST_SS_ACTIVE_PENDING __vcpu_single_flag(sflags, BIT(3)) +/* Software step state is Active pending for guest debug */ +#define GUEST_SS_ACTIVE_PENDING __vcpu_single_flag(sflags, BIT(4)) /* PMUSERENR for the guest EL0 is on physical CPU */ -#define PMUSERENR_ON_CPU __vcpu_single_flag(sflags, BIT(4)) +#define PMUSERENR_ON_CPU __vcpu_single_flag(sflags, BIT(5)) /* WFI instruction trapped */ -#define IN_WFI __vcpu_single_flag(sflags, BIT(5)) +#define IN_WFI __vcpu_single_flag(sflags, BIT(6)) /* Pointer to the vcpu's SVE FFR for sve_{save,load}_state() */ @@ -1341,9 +1332,8 @@ static inline bool kvm_system_needs_idmapped_vectors(void) static inline void kvm_arch_sync_events(struct kvm *kvm) {} void kvm_init_host_debug_data(void); -void kvm_arm_setup_debug(struct kvm_vcpu *vcpu); -void kvm_arm_clear_debug(struct kvm_vcpu *vcpu); void kvm_vcpu_load_debug(struct kvm_vcpu *vcpu); +void kvm_vcpu_put_debug(struct kvm_vcpu *vcpu); void kvm_handle_debug_access(struct kvm_vcpu *vcpu); void kvm_debug_handle_oslar(struct kvm_vcpu *vcpu, u64 val); diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c index ec581aeb41f9..563cd0b626b9 100644 --- a/arch/arm64/kvm/arm.c +++ b/arch/arm64/kvm/arm.c @@ -622,6 +622,7 @@ void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu) void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu) { + kvm_vcpu_put_debug(vcpu); kvm_arch_vcpu_put_fp(vcpu); if (has_vhe()) kvm_vcpu_put_vhe(vcpu); @@ -1181,7 +1182,6 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu) continue; } - kvm_arm_setup_debug(vcpu); kvm_arch_vcpu_ctxflush_fp(vcpu); /************************************************************** @@ -1198,8 +1198,6 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu) * Back from guest *************************************************************/ - kvm_arm_clear_debug(vcpu); - /* * We must sync the PMU state before the vgic state so * that the vgic can properly sample the updated state of the diff --git a/arch/arm64/kvm/debug.c b/arch/arm64/kvm/debug.c index a23d214a0fb0..106382810d4b 100644 --- a/arch/arm64/kvm/debug.c +++ b/arch/arm64/kvm/debug.c @@ -16,34 +16,6 @@ #include "trace.h" -/* - * save/restore_guest_debug_regs - * - * For some debug operations we need to tweak some guest registers. As - * a result we need to save the state of those registers before we - * make those modifications. - * - * Guest access to MDSCR_EL1 is trapped by the hypervisor and handled - * after we have restored the preserved value to the main context. - * - * When single-step is enabled by userspace, we tweak PSTATE.SS on every - * guest entry. Preserve PSTATE.SS so we can restore the original value - * for the vcpu after the single-step is disabled. - */ -static void save_guest_debug_regs(struct kvm_vcpu *vcpu) -{ - vcpu->arch.guest_debug_preserved.pstate_ss = - (*vcpu_cpsr(vcpu) & DBG_SPSR_SS); -} - -static void restore_guest_debug_regs(struct kvm_vcpu *vcpu) -{ - if (vcpu->arch.guest_debug_preserved.pstate_ss) - *vcpu_cpsr(vcpu) |= DBG_SPSR_SS; - else - *vcpu_cpsr(vcpu) &= ~DBG_SPSR_SS; -} - /** * kvm_arm_setup_mdcr_el2 - configure vcpu mdcr_el2 value * @@ -84,89 +56,6 @@ static void kvm_arm_setup_mdcr_el2(struct kvm_vcpu *vcpu) vcpu->arch.mdcr_el2 |= MDCR_EL2_TDA; } -/** - * kvm_arm_setup_debug - set up debug related stuff - * - * @vcpu: the vcpu pointer - * - * This is called before each entry into the hypervisor to setup any - * debug related registers. - * - * Additionally, KVM only traps guest accesses to the debug registers if - * the guest is not actively using them. Since the guest must not interfere - * with the hardware state when debugging the guest, we must ensure that - * trapping is enabled whenever we are debugging the guest using the - * debug registers. - */ - -void kvm_arm_setup_debug(struct kvm_vcpu *vcpu) -{ - unsigned long mdscr; - - trace_kvm_arm_setup_debug(vcpu, vcpu->guest_debug); - - /* Check if we need to use the debug registers. */ - if (vcpu->guest_debug || kvm_vcpu_os_lock_enabled(vcpu)) { - /* Save guest debug state */ - save_guest_debug_regs(vcpu); - - /* - * Single Step (ARM ARM D2.12.3 The software step state - * machine) - * - * If we are doing Single Step we need to manipulate - * the guest's MDSCR_EL1.SS and PSTATE.SS. Once the - * step has occurred the hypervisor will trap the - * debug exception and we return to userspace. - * - * If the guest attempts to single step its userspace - * we would have to deal with a trapped exception - * while in the guest kernel. Because this would be - * hard to unwind we suppress the guest's ability to - * do so by masking MDSCR_EL.SS. - * - * This confuses guest debuggers which use - * single-step behind the scenes but everything - * returns to normal once the host is no longer - * debugging the system. - */ - if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP) { - /* - * If the software step state at the last guest exit - * was Active-pending, we don't set DBG_SPSR_SS so - * that the state is maintained (to not run another - * single-step until the pending Software Step - * exception is taken). - */ - if (!vcpu_get_flag(vcpu, DBG_SS_ACTIVE_PENDING)) - *vcpu_cpsr(vcpu) |= DBG_SPSR_SS; - else - *vcpu_cpsr(vcpu) &= ~DBG_SPSR_SS; - } - } -} - -void kvm_arm_clear_debug(struct kvm_vcpu *vcpu) -{ - trace_kvm_arm_clear_debug(vcpu->guest_debug); - - /* - * Restore the guest's debug registers if we were using them. - */ - if (vcpu->guest_debug || kvm_vcpu_os_lock_enabled(vcpu)) { - if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP) { - if (!(*vcpu_cpsr(vcpu) & DBG_SPSR_SS)) - /* - * Mark the vcpu as ACTIVE_PENDING - * until Software Step exception is taken. - */ - vcpu_set_flag(vcpu, DBG_SS_ACTIVE_PENDING); - } - - restore_guest_debug_regs(vcpu); - } -} - void kvm_init_host_debug_data(void) { u64 dfr0 = read_sysreg(id_aa64dfr0_el1); @@ -197,8 +86,24 @@ void kvm_vcpu_load_debug(struct kvm_vcpu *vcpu) if (vcpu->guest_debug || kvm_vcpu_os_lock_enabled(vcpu)) { mdscr = MDSCR_EL1_TDCC; - if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP) + /* + * Steal the guest's single-step state machine if userspace wants + * single-step the guest. + */ + if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP) { mdscr |= MDSCR_EL1_SS; + + if (*vcpu_cpsr(vcpu) & DBG_SPSR_SS) + vcpu_clear_flag(vcpu, GUEST_SS_ACTIVE_PENDING); + else + vcpu_set_flag(vcpu, GUEST_SS_ACTIVE_PENDING); + + if (!vcpu_get_flag(vcpu, HOST_SS_ACTIVE_PENDING)) + *vcpu_cpsr(vcpu) |= DBG_SPSR_SS; + else + *vcpu_cpsr(vcpu) &= ~DBG_SPSR_SS; + } + if (vcpu->guest_debug & KVM_GUESTDBG_USE_HW) mdscr |= MDSCR_EL1_MDE; @@ -220,6 +125,26 @@ void kvm_vcpu_load_debug(struct kvm_vcpu *vcpu) kvm_arm_setup_mdcr_el2(vcpu); } +void kvm_vcpu_put_debug(struct kvm_vcpu *vcpu) +{ + if (likely(!(vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP))) + return; + + /* + * Save the host's single-step state machine and restore the guest's + * before potentially returning to userspace. + */ + if (!(*vcpu_cpsr(vcpu) & DBG_SPSR_SS)) + vcpu_set_flag(vcpu, HOST_SS_ACTIVE_PENDING); + else + vcpu_clear_flag(vcpu, HOST_SS_ACTIVE_PENDING); + + if (vcpu_get_flag(vcpu, GUEST_SS_ACTIVE_PENDING)) + *vcpu_cpsr(vcpu) &= ~DBG_SPSR_SS; + else + *vcpu_cpsr(vcpu) |= DBG_SPSR_SS; +} + void kvm_handle_debug_access(struct kvm_vcpu *vcpu) { if (kvm_host_owns_debug_regs(vcpu)) diff --git a/arch/arm64/kvm/guest.c b/arch/arm64/kvm/guest.c index b72eec27adc7..bea56c23bf98 100644 --- a/arch/arm64/kvm/guest.c +++ b/arch/arm64/kvm/guest.c @@ -926,7 +926,7 @@ int kvm_arch_vcpu_ioctl_set_guest_debug(struct kvm_vcpu *vcpu, if (!(dbg->control & KVM_GUESTDBG_ENABLE)) { vcpu->guest_debug = 0; - vcpu_clear_flag(vcpu, DBG_SS_ACTIVE_PENDING); + vcpu_clear_flag(vcpu, HOST_SS_ACTIVE_PENDING); return 0; } diff --git a/arch/arm64/kvm/handle_exit.c b/arch/arm64/kvm/handle_exit.c index d7c2990e7c9e..1e302f0c8903 100644 --- a/arch/arm64/kvm/handle_exit.c +++ b/arch/arm64/kvm/handle_exit.c @@ -193,7 +193,7 @@ static int kvm_handle_guest_debug(struct kvm_vcpu *vcpu) run->debug.arch.far = vcpu->arch.fault.far_el2; break; case ESR_ELx_EC_SOFTSTP_LOW: - vcpu_clear_flag(vcpu, DBG_SS_ACTIVE_PENDING); + *vcpu_cpsr(vcpu) |= DBG_SPSR_SS; break; } -- 2.39.5