OSDN Git Service

KVM: X86: Fix softlockup when get the current kvmclock
authorWanpeng Li <kernellwp@gmail.com>
Mon, 20 Nov 2017 22:55:05 +0000 (14:55 -0800)
committerPaolo Bonzini <pbonzini@redhat.com>
Mon, 27 Nov 2017 16:32:53 +0000 (17:32 +0100)
commite70b57a6ce4e8b92a56a615ae79bdb2bd66035e7
tree32e4f258fb9dca5655b7b0fe974672e9024f89fe
parent12806ba937382fdfdbad62a399aa2dce65c10fcd
KVM: X86: Fix softlockup when get the current kvmclock

 watchdog: BUG: soft lockup - CPU#6 stuck for 22s! [qemu-system-x86:10185]
 CPU: 6 PID: 10185 Comm: qemu-system-x86 Tainted: G           OE   4.14.0-rc4+ #4
 RIP: 0010:kvm_get_time_scale+0x4e/0xa0 [kvm]
 Call Trace:
  get_time_ref_counter+0x5a/0x80 [kvm]
  kvm_hv_process_stimers+0x120/0x5f0 [kvm]
  kvm_arch_vcpu_ioctl_run+0x4b4/0x1690 [kvm]
  kvm_vcpu_ioctl+0x33a/0x620 [kvm]
  do_vfs_ioctl+0xa1/0x5d0
  SyS_ioctl+0x79/0x90
  entry_SYSCALL_64_fastpath+0x1e/0xa9

This can be reproduced when running kvm-unit-tests/hyperv_stimer.flat and
cpu-hotplug stress simultaneously. __this_cpu_read(cpu_tsc_khz) returns 0
(set in kvmclock_cpu_down_prep()) when the pCPU is unhotplug which results
in kvm_get_time_scale() gets into an infinite loop.

This patch fixes it by treating the unhotplug pCPU as not using master clock.

Reviewed-by: Radim Krčmář <rkrcmar@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Radim Krčmář <rkrcmar@redhat.com>
Signed-off-by: Wanpeng Li <wanpeng.li@hotmail.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
arch/x86/kvm/x86.c