Hi Marcelo, On 10/25/2024 9:54 PM, Marcelo Tosatti wrote: > On Mon, Oct 14, 2024 at 08:17:19PM +0530, Nikunj A. Dadhania wrote: >> On 10/12/2024 1:25 PM, Isaku Yamahata wrote: >>> Problem >>> ------- >>> The current x86 KVM implementation conflicts with protected TSC because the >>> VMM can't change the TSC offset/multiplier. Disable or ignore the KVM >>> logic to change/adjust the TSC offset/multiplier somehow. >>> >>> Because KVM emulates the TSC timer or the TSC deadline timer with the TSC >>> offset/multiplier, the TSC timer interrupts are injected to the guest at the >>> wrong time if the KVM TSC offset is different from what the TDX module >>> determined. >>> >>> Originally the issue was found by cyclic test of rt-test [1] as the latency in >>> TDX case is worse than VMX value + TDX SEAMCALL overhead. It turned out that >>> the KVM TSC offset is different from what the TDX module determines. >> >> Can you provide what is the exact command line to reproduce this problem ? > > Nikunj, > > Run cyclictest, on an isolated CPU, in a VM. For the maximum latency > metric, rather than 50us, one gets 500us at times. I tried out the cyclictest after referring to the documentation[1]. Here are the results of the run on an isolated CPU in the Secure TSC-enabled SEV-SNP VM (CPUs 16-31 are isolated in the VM): $ sudo taskset -c 16-31 ./cyclictest --mlockall --priority=80 --interval=200 --distance=0 -q -D 5m T: 0 ( 1226) P:80 I:200 C:1500000 Min: 6 Act: 10 Avg: 16 Max: 150 $ VM detail: 32 vCPUs VM, guest kernel v6.12-rc3 compiled with CONFIG_PREEMPT_RT=y > > FYI: > > SEV-SNP processors (at least the one below) do not seem affected by this problem. Thanks for testing on SEV-SNP. Regards, Nikunj 1. https://wiki.linuxfoundation.org/realtime/documentation/howto/tools/cyclictest/start
