Skip to content

cloud-kernel
cloud-kernel

openanolis / cloud-kernel
接近 2 年 前同步成功

通知 170
Star 36
Fork 7
cloud-kernel
cloud-kernel
切换分支/标签
  1. 08 4月, 2015 3 次提交
    • P
      KVM: x86: optimize delivery of TSC deadline timer interrupt · 9c8fd1ba
      由 Paolo Bonzini 提交于 
      The newly-added tracepoint shows the following results on
the tscdeadline_latency test:

        qemu-kvm-8387  [002]  6425.558974: kvm_vcpu_wakeup:      poll time 10407 ns
        qemu-kvm-8387  [002]  6425.558984: kvm_vcpu_wakeup:      poll time 0 ns
        qemu-kvm-8387  [002]  6425.561242: kvm_vcpu_wakeup:      poll time 10477 ns
        qemu-kvm-8387  [002]  6425.561251: kvm_vcpu_wakeup:      poll time 0 ns

and so on.  This is because we need to go through kvm_vcpu_block again
after the timer IRQ is injected.  Avoid it by polling once before
entering kvm_vcpu_block.

On my machine (Xeon E5 Sandy Bridge) this removes about 500 cycles (7%)
from the latency of the TSC deadline timer.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>
      9c8fd1ba
    • P
      KVM: x86: extract blocking logic from __vcpu_run · 362c698f
      由 Paolo Bonzini 提交于 
      Rename the old __vcpu_run to vcpu_run, and extract part of it to a new
function vcpu_block.

The next patch will add a new condition in vcpu_block, avoid extra
indentation.
Reviewed-by: NDavid Matlack <dmatlack@google.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>
      362c698f
    • W
      kvm: x86: fix x86 eflags fixed bit · 35fd68a3
      由 Wanpeng Li 提交于 
      Guest can't be booted w/ ept=0, there is a message dumped as below:

If you're running a guest on an Intel machine without unrestricted mode
support, the failure can be most likely due to the guest entering an invalid
state for Intel VT. For example, the guest maybe running in big real mode
which is not supported on less recent Intel processors.

EAX=00000011 EBX=f000d2f6 ECX=00006cac EDX=000f8956
ESI=bffbdf62 EDI=00000000 EBP=00006c68 ESP=00006c68
EIP=0000d187 EFL=00000004 [-----P-] CPL=0 II=0 A20=1 SMM=0 HLT=0
ES =e000 000e0000 ffffffff 00809300 DPL=0 DS16 [-WA]
CS =f000 000f0000 ffffffff 00809b00 DPL=0 CS16 [-RA]
SS =0000 00000000 ffffffff 00809300 DPL=0 DS16 [-WA]
DS =0000 00000000 ffffffff 00809300 DPL=0 DS16 [-WA]
FS =0000 00000000 ffffffff 00809300 DPL=0 DS16 [-WA]
GS =0000 00000000 ffffffff 00809300 DPL=0 DS16 [-WA]
LDT=0000 00000000 0000ffff 00008200 DPL=0 LDT
TR =0000 00000000 0000ffff 00008b00 DPL=0 TSS32-busy
GDT=     000f6a80 00000037
IDT=     000f6abe 00000000
CR0=00000011 CR2=00000000 CR3=00000000 CR4=00000000
DR0=0000000000000000 DR1=0000000000000000 DR2=0000000000000000 DR3=0000000000000000
DR6=00000000ffff0ff0 DR7=0000000000000400
EFER=0000000000000000
Code=01 1e b8 6a 2e 0f 01 16 74 6a 0f 20 c0 66 83 c8 01 0f 22 c0 <66> ea 8f d1 0f 00 08 00 b8 10 00 00 00 8e d8 8e c0 8e d0 8e e0 8e e8 89 c8 ff e2 89 c1 b8X

X86 eflags bit 1 is fixed set, which means that 1 << 1 is set instead of 1,
this patch fix it.
Signed-off-by: NWanpeng Li <wanpeng.li@linux.intel.com>
Message-Id: <1428473294-6633-1-git-send-email-wanpeng.li@linux.intel.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>
      35fd68a3
  3. 01 4月, 2015 1 次提交
  5. 30 3月, 2015 7 次提交
  7. 27 3月, 2015 4 次提交
  9. 24 3月, 2015 1 次提交
  11. 19 3月, 2015 1 次提交
  13. 18 3月, 2015 2 次提交
  15. 14 3月, 2015 2 次提交
  17. 13 3月, 2015 1 次提交
  19. 11 3月, 2015 2 次提交
  21. 10 3月, 2015 3 次提交
  23. 03 3月, 2015 1 次提交
    • R
      KVM: SVM: fix interrupt injection (apic->isr_count always 0) · f563db4b
      由 Radim Krčmář 提交于 
      In commit b4eef9b3, we started to use hwapic_isr_update() != NULL
instead of kvm_apic_vid_enabled(vcpu->kvm).  This didn't work because
SVM had it defined and "apicv" path in apic_{set,clear}_isr() does not
change apic->isr_count, because it should always be 1.  The initial
value of apic->isr_count was based on kvm_apic_vid_enabled(vcpu->kvm),
which is always 0 for SVM, so KVM could have injected interrupts when it
shouldn't.

Fix it by implicitly setting SVM's hwapic_isr_update to NULL and make the
initial isr_count depend on hwapic_isr_update() for good measure.

Fixes: b4eef9b3 ("kvm: x86: vmx: NULL out hwapic_isr_update() in case of !enable_apicv")
Reported-and-tested-by: NBorislav Petkov <bp@suse.de>
Signed-off-by: NRadim Krčmář <rkrcmar@redhat.com>
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
      f563db4b
  25. 24 2月, 2015 2 次提交
  27. 10 2月, 2015 1 次提交
  29. 09 2月, 2015 1 次提交
  31. 06 2月, 2015 1 次提交
    • P
      kvm: add halt_poll_ns module parameter · f7819512
      由 Paolo Bonzini 提交于 
      This patch introduces a new module parameter for the KVM module; when it
is present, KVM attempts a bit of polling on every HLT before scheduling
itself out via kvm_vcpu_block.

This parameter helps a lot for latency-bound workloads---in particular
I tested it with O_DSYNC writes with a battery-backed disk in the host.
In this case, writes are fast (because the data doesn't have to go all
the way to the platters) but they cannot be merged by either the host or
the guest.  KVM's performance here is usually around 30% of bare metal,
or 50% if you use cache=directsync or cache=writethrough (these
parameters avoid that the guest sends pointless flush requests, and
at the same time they are not slow because of the battery-backed cache).
The bad performance happens because on every halt the host CPU decides
to halt itself too.  When the interrupt comes, the vCPU thread is then
migrated to a new physical CPU, and in general the latency is horrible
because the vCPU thread has to be scheduled back in.

With this patch performance reaches 60-65% of bare metal and, more
important, 99% of what you get if you use idle=poll in the guest.  This
means that the tunable gets rid of this particular bottleneck, and more
work can be done to improve performance in the kernel or QEMU.

Of course there is some price to pay; every time an otherwise idle vCPUs
is interrupted by an interrupt, it will poll unnecessarily and thus
impose a little load on the host.  The above results were obtained with
a mostly random value of the parameter (500000), and the load was around
1.5-2.5% CPU usage on one of the host's core for each idle guest vCPU.

The patch also adds a new stat, /sys/kernel/debug/kvm/halt_successful_poll,
that can be used to tune the parameter.  It counts how many HLT
instructions received an interrupt during the polling period; each
successful poll avoids that Linux schedules the VCPU thread out and back
in, and may also avoid a likely trip to C1 and back for the physical CPU.

While the VM is idle, a Linux 4 VCPU VM halts around 10 times per second.
Of these halts, almost all are failed polls.  During the benchmark,
instead, basically all halts end within the polling period, except a more
or less constant stream of 50 per second coming from vCPUs that are not
running the benchmark.  The wasted time is thus very low.  Things may
be slightly different for Windows VMs, which have a ~10 ms timer tick.

The effect is also visible on Marcelo's recently-introduced latency
test for the TSC deadline timer.  Though of course a non-RT kernel has
awful latency bounds, the latency of the timer is around 8000-10000 clock
cycles compared to 20000-120000 without setting halt_poll_ns.  For the TSC
deadline timer, thus, the effect is both a smaller average latency and
a smaller variance.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>
      f7819512
  33. 04 2月, 2015 7 次提交