When a virtual machine uses the hv timer during suspend, the kvm timer does
not advance. After a long period, if the VM is woken up, there will be a
large gap between target_expiration and now. Since each timer expiration
only advances target_expiration by one period, the timer expiration
function will be repeatedly executed.
Without the previous patch merged, the advanced target_expiration is less
than now, which causes tscdeadline to be set to a negative value. This
results in HV timer setup failure and a fallback to the SW timer. After
switching to the SW timer, apic_timer_fn is repeatedly executed within a
single clock interrupt handler, leading to a hardlockup:
NMI watchdog: Watchdog detected hard LOCKUP on cpu 45
...
RIP: 0010:advance_periodic_target_expiration+0x4d/0x80 [kvm]
...
RSP: 0018:ff4f88f5d98d8ef0 EFLAGS: 00000046
RAX: fff0103f91be678e RBX: fff0103f91be678e RCX: 00843a7d9e127bcc
RDX: 0000000000000002 RSI: 0052ca4003697505 RDI: ff440d5bfbdbd500
RBP: ff440d5956f99200 R08: ff2ff2a42deb6a84 R09: 000000000002a6c0
R10: 0122d794016332b3 R11: 0000000000000000 R12: ff440db1af39cfc0
R13: ff440db1af39cfc0 R14: ffffffffc0d4a560 R15: ff440db1af39d0f8
FS: 00007f04a6ffd700(0000) GS:ff440db1af380000(0000) knlGS:000000e38a3b8000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 000000d5651feff8 CR3: 000000684e038002 CR4: 0000000000773ee0
PKRU: 55555554
Call Trace:
apic_timer_fn+0x31/0x50 [kvm]
__hrtimer_run_queues+0x100/0x280
hrtimer_interrupt+0x100/0x210
? ttwu_do_wakeup+0x19/0x160
smp_apic_timer_interrupt+0x6a/0x130
apic_timer_interrupt+0xf/0x20
After the previous patch is merged, the HV timer can no longer fall back to
the SW timer. Additionally, while target_expiration is catching up to the
current time, the VMX-preemption timer is set to 0 before each VM entry.
According to Intel SDM 27.7.4 “VMX-Preemption Timer”: if the VMX-preemption
timer has already expired at VM entry, a VM exit will occur before any
instruction is executed. As a result, the guest cannot execute any
instructions during this period, and therefore has no opportunity to reach
vcpu_block() to switch to the SW timer. Thus, a hardlockup will not occur.
However, it is still necessary to eliminate unnecessary multiple catch-ups.
Therefore, if the advanced target_expiration is still less than now, we
catch up to now in the current handling.
Signed-off-by: fuqiang wang
---
arch/x86/kvm/lapic.c | 22 +++++++++++++++-------
1 file changed, 15 insertions(+), 7 deletions(-)
diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c
index fa07a303767c..ba30de871929 100644
--- a/arch/x86/kvm/lapic.c
+++ b/arch/x86/kvm/lapic.c
@@ -2140,17 +2140,25 @@ static void advance_periodic_target_expiration(struct kvm_lapic *apic)
apic->lapic_timer.target_expiration =
ktime_add_ns(apic->lapic_timer.target_expiration,
apic->lapic_timer.period);
- delta = ktime_sub(apic->lapic_timer.target_expiration, now);
/*
- * Don't adjust the tscdeadline if the next period has already expired,
- * e.g. due to software overhead resulting in delays larger than the
- * period. Blindly adding a negative delta could cause the deadline to
- * become excessively large due to the deadline being an unsigned value.
+ * When the vm is suspend, the hv timer also stops advancing. After it
+ * is resumed, this may result in a large delta. If the
+ * target_expiration only advances by one period each time, it will
+ * cause KVM to frequently handle timer expirations.
*/
+ if (apic->lapic_timer.period > 0 &&
+ ktime_before(apic->lapic_timer.target_expiration, now))
+ apic->lapic_timer.target_expiration = now;
+
+ delta = ktime_sub(apic->lapic_timer.target_expiration, now);
apic->lapic_timer.tscdeadline = kvm_read_l1_tsc(apic->vcpu, tscl);
- if (delta > 0)
- apic->lapic_timer.tscdeadline += nsec_to_cycles(apic->vcpu, delta);
+ /*
+ * Note: delta must not be negative. Otherwise, blindly adding a
+ * negative delta could cause the deadline to become excessively large
+ * due to the deadline being an unsigned value.
+ */
+ apic->lapic_timer.tscdeadline += nsec_to_cycles(apic->vcpu, delta);
}
static void start_sw_period(struct kvm_lapic *apic)
--
2.47.0