| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-Id: <20210426090644.2218834-7-hikalium@chromium.org>
Date: Mon, 26 Apr 2021 18:06:45 +0900
From: Hikaru Nishida <hikalium@...omium.org>
To: kvm@...r.kernel.org
Cc: suleiman@...gle.com, Hikaru Nishida <hikalium@...omium.org>,
Borislav Petkov <bp@...en8.de>,
"H. Peter Anvin" <hpa@...or.com>, Ingo Molnar <mingo@...hat.com>,
Jim Mattson <jmattson@...gle.com>,
Joerg Roedel <joro@...tes.org>,
John Stultz <john.stultz@...aro.org>,
Paolo Bonzini <pbonzini@...hat.com>,
Sean Christopherson <seanjc@...gle.com>,
Stephen Boyd <sboyd@...nel.org>,
Thomas Gleixner <tglx@...utronix.de>,
Vitaly Kuznetsov <vkuznets@...hat.com>,
Wanpeng Li <wanpengli@...cent.com>,
linux-kernel@...r.kernel.org, x86@...nel.org
Subject: [RFC PATCH 6/6] x86/kvm: Add a guest side support for virtual suspend time injection
This patch implements virtual suspend time injection support for kvm
guests. If this functionality is enabled and the host supports
KVM_FEATURE_HOST_SUSPEND_TIME,
the guest will register struct kvm_host_suspend_time through MSR to
monitor how much time we spend during the host's suspension. If the
duration is increased, the guest will advance CLOCK_BOOTTIME to match
with the host's suspension.
Signed-off-by: Hikaru Nishida <hikalium@...omium.org>
---
arch/x86/include/asm/kvm_para.h | 9 +++++++++
arch/x86/kernel/kvmclock.c | 25 +++++++++++++++++++++++++
include/linux/timekeeper_internal.h | 4 ++++
kernel/time/timekeeping.c | 27 +++++++++++++++++++++++++++
4 files changed, 65 insertions(+)
diff --git a/arch/x86/include/asm/kvm_para.h b/arch/x86/include/asm/kvm_para.h
index 338119852512..e552efa931a8 100644
--- a/arch/x86/include/asm/kvm_para.h
+++ b/arch/x86/include/asm/kvm_para.h
@@ -18,6 +18,15 @@ static inline bool kvm_check_and_clear_guest_paused(void)
}
#endif /* CONFIG_KVM_GUEST */
+#ifdef CONFIG_KVM_VIRT_SUSPEND_TIMING_GUEST
+u64 kvm_get_suspend_time(void);
+#else
+static inline u64 kvm_get_suspend_time(void)
+{
+ return 0;
+}
+#endif /* CONFIG_KVM_VIRT_SUSPEND_TIMING_GUEST */
+
#define KVM_HYPERCALL \
ALTERNATIVE("vmcall", "vmmcall", X86_FEATURE_VMMCALL)
diff --git a/arch/x86/kernel/kvmclock.c b/arch/x86/kernel/kvmclock.c
index 1fc0962c89c0..630460beb05a 100644
--- a/arch/x86/kernel/kvmclock.c
+++ b/arch/x86/kernel/kvmclock.c
@@ -49,6 +49,9 @@ early_param("no-kvmclock-vsyscall", parse_no_kvmclock_vsyscall);
static struct pvclock_vsyscall_time_info
hv_clock_boot[HVC_BOOT_ARRAY_SIZE] __bss_decrypted __aligned(PAGE_SIZE);
+#ifdef CONFIG_KVM_VIRT_SUSPEND_TIMING_GUEST
+static struct kvm_host_suspend_time suspend_time __bss_decrypted;
+#endif
static struct pvclock_wall_clock wall_clock __bss_decrypted;
static DEFINE_PER_CPU(struct pvclock_vsyscall_time_info *, hv_clock_per_cpu);
static struct pvclock_vsyscall_time_info *hvclock_mem;
@@ -164,6 +167,20 @@ static int kvm_cs_enable(struct clocksource *cs)
return 0;
}
+#ifdef CONFIG_KVM_VIRT_SUSPEND_TIMING_GUEST
+/*
+ * kvm_get_suspend_time_and_clear_request - This function
+ * checks how much suspend time is injected by the host.
+ * A returned value is a total time passed during the guest in a suspend
+ * state while this guest is running. (Not a duration of the last host
+ * suspend.)
+ */
+u64 kvm_get_suspend_time(void)
+{
+ return suspend_time.suspend_time_ns;
+}
+#endif /* CONFIG_KVM_VIRT_SUSPEND_TIMING_GUEST */
+
struct clocksource kvm_clock = {
.name = "kvm-clock",
.read = kvm_clock_get_cycles,
@@ -324,6 +341,14 @@ void __init kvmclock_init(void)
return;
}
+#ifdef CONFIG_KVM_VIRT_SUSPEND_TIMING_GUEST
+ if (kvm_para_has_feature(KVM_FEATURE_HOST_SUSPEND_TIME)) {
+ /* Register the suspend time structure */
+ wrmsrl(MSR_KVM_HOST_SUSPEND_TIME,
+ slow_virt_to_phys(&suspend_time) | KVM_MSR_ENABLED);
+ }
+#endif
+
if (cpuhp_setup_state(CPUHP_BP_PREPARE_DYN, "kvmclock:setup_percpu",
kvmclock_setup_percpu, NULL) < 0) {
return;
diff --git a/include/linux/timekeeper_internal.h b/include/linux/timekeeper_internal.h
index 84ff2844df2a..a5fd515f0a9d 100644
--- a/include/linux/timekeeper_internal.h
+++ b/include/linux/timekeeper_internal.h
@@ -124,6 +124,10 @@ struct timekeeper {
u32 ntp_err_mult;
/* Flag used to avoid updating NTP twice with same second */
u32 skip_second_overflow;
+#ifdef CONFIG_KVM_VIRT_SUSPEND_TIMING_GUEST
+ /* suspend_time_injected keeps the duration injected through kvm */
+ u64 suspend_time_injected;
+#endif
#ifdef CONFIG_DEBUG_TIMEKEEPING
long last_warning;
/*
diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c
index 342a032ad552..6b89e0d42596 100644
--- a/kernel/time/timekeeping.c
+++ b/kernel/time/timekeeping.c
@@ -2114,6 +2114,29 @@ static u64 logarithmic_accumulation(struct timekeeper *tk, u64 offset,
return offset;
}
+#ifdef CONFIG_KVM_VIRT_SUSPEND_TIMING_GUEST
+/*
+ * timekeeping_inject_suspend_time - Inject virtual suspend time
+ * if it is requested by kvm host.
+ * This function should be called under holding timekeeper_lock and
+ * only from timekeeping_advance().
+ */
+static void timekeeping_inject_virtual_suspend_time(struct timekeeper *tk)
+{
+ struct timespec64 delta;
+ u64 suspend_time;
+
+ suspend_time = kvm_get_suspend_time();
+ if (suspend_time <= tk->suspend_time_injected) {
+ /* Sufficient amount of suspend time is already injected. */
+ return;
+ }
+ delta = ns_to_timespec64(suspend_time - tk->suspend_time_injected);
+ __timekeeping_inject_sleeptime(tk, &delta);
+ tk->suspend_time_injected = suspend_time;
+}
+#endif
+
/*
* timekeeping_advance - Updates the timekeeper to the current time and
* current NTP tick length
@@ -2143,6 +2166,10 @@ static void timekeeping_advance(enum timekeeping_adv_mode mode)
/* Do some additional sanity checking */
timekeeping_check_update(tk, offset);
+#ifdef CONFIG_KVM_VIRT_SUSPEND_TIMING_GUEST
+ timekeeping_inject_virtual_suspend_time(tk);
+#endif
+
/*
* With NO_HZ we may have to accumulate many cycle_intervals
* (think "ticks") worth of time at once. To do this efficiently,
--
2.31.1.498.g6c1eba8ee3d-goog
Powered by blists - more mailing lists