I want to measure shadow page table performance vs EPT. I know in kvm code path, EPT and shadow page table share some code path. There is a switch to check if EPT is enabled.
So, I turn off EPT, I think this is a way to use shadow page table in KVM VM. I exec some commands on host:
# cat /sys/module/kvm_intel/parameters/ept
Y //check if EPT is enabled, Yes it is
# rmmod kvm_intel
# modprobe kvm_intel ept=0,vpid=0 //Re-modprobe kvm_intel, but turn off EPT and VPID
# cat /sys/module/kvm_intel/parameters/ept
N // EPT is disabled
# cat /sys/module/kvm_intel/parameters/vpid
Y //VPID is still enabled
After these commads, I expect the EPT is disabled, However, I create a VM with 4 vcpu, but in VM, I use htop in VM only find i vcpu. I don't know why.
So, I turn on ept again, I can see 4 vcpu in vm using htop. But, on another server, I turn off ept, in VM, I still can find 4 vcpu.
In host, I test qemu thread num:
pstree -p | grep qemu
|-qemu-kvm(20594)-+-{qemu-kvm}(20612)
| |-{qemu-kvm}(20613)
| |-{qemu-kvm}(20614)
| |-{qemu-kvm}(20615)
| |-{qemu-kvm}(20618)
| `-{qemu-kvm}(23429)
There are still muti threads.
In KVM VM, I use lscpu to check:
# lscpu
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Byte Order: Little Endian
CPU(s): 4
On-line CPU(s) list: 0
Off-line CPU(s) list: 1-3
Thread(s) per core: 1
Core(s) per socket: 1
座: 1
NUMA 节点: 1
厂商 ID: GenuineIntel
CPU 系列: 6
型号: 62
型号名称: Intel(R) Xeon(R) CPU E5-2640 v2 @ 2.00GHz
步进: 4
CPU MHz: 1999.999
BogoMIPS: 3999.99
超管理器厂商: KVM
虚拟化类型: 完全
L1d 缓存: 32K
L1i 缓存: 32K
L2 缓存: 4096K
NUMA 节点0 CPU: 0
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon rep_good nopl eagerfpu pni pclmulqdq ssse3 cx16 pcid sse4_1 sse4_2 x2apic popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm fsgsbase tsc_adjust smep erms xsaveopt
VPID is still enabled cause your command format was incorrect