Michael S. Tsirkin ab9cf4996b KVM guest: guest side for eoi avoidance ...

The idea is simple: there's a bit, per APIC, in guest memory,
that tells the guest that it does not need EOI.
Guest tests it using a single est and clear operation - this is
necessary so that host can detect interrupt nesting - and if set, it can
skip the EOI MSR.

I run a simple microbenchmark to show exit reduction
(note: for testing, need to apply follow-up patch
'kvm: host side for eoi optimization' + a qemu patch
 I posted separately, on host):

Before:

Performance counter stats for 'sleep 1s':

            47,357 kvm:kvm_entry                                                [99.98%]
                 0 kvm:kvm_hypercall                                            [99.98%]
                 0 kvm:kvm_hv_hypercall                                         [99.98%]
             5,001 kvm:kvm_pio                                                  [99.98%]
                 0 kvm:kvm_cpuid                                                [99.98%]
            22,124 kvm:kvm_apic                                                 [99.98%]
            49,849 kvm:kvm_exit                                                 [99.98%]
            21,115 kvm:kvm_inj_virq                                             [99.98%]
                 0 kvm:kvm_inj_exception                                        [99.98%]
                 0 kvm:kvm_page_fault                                           [99.98%]
            22,937 kvm:kvm_msr                                                  [99.98%]
                 0 kvm:kvm_cr                                                   [99.98%]
                 0 kvm:kvm_pic_set_irq                                          [99.98%]
                 0 kvm:kvm_apic_ipi                                             [99.98%]
            22,207 kvm:kvm_apic_accept_irq                                      [99.98%]
            22,421 kvm:kvm_eoi                                                  [99.98%]
                 0 kvm:kvm_pv_eoi                                               [99.99%]
                 0 kvm:kvm_nested_vmrun                                         [99.99%]
                 0 kvm:kvm_nested_intercepts                                    [99.99%]
                 0 kvm:kvm_nested_vmexit                                        [99.99%]
                 0 kvm:kvm_nested_vmexit_inject                                    [99.99%]
                 0 kvm:kvm_nested_intr_vmexit                                    [99.99%]
                 0 kvm:kvm_invlpga                                              [99.99%]
                 0 kvm:kvm_skinit                                               [99.99%]
                57 kvm:kvm_emulate_insn                                         [99.99%]
                 0 kvm:vcpu_match_mmio                                          [99.99%]
                 0 kvm:kvm_userspace_exit                                       [99.99%]
                 2 kvm:kvm_set_irq                                              [99.99%]
                 2 kvm:kvm_ioapic_set_irq                                       [99.99%]
            23,609 kvm:kvm_msi_set_irq                                          [99.99%]
                 1 kvm:kvm_ack_irq                                              [99.99%]
               131 kvm:kvm_mmio                                                 [99.99%]
               226 kvm:kvm_fpu                                                  [100.00%]
                 0 kvm:kvm_age_page                                             [100.00%]
                 0 kvm:kvm_try_async_get_page                                    [100.00%]
                 0 kvm:kvm_async_pf_doublefault                                    [100.00%]
                 0 kvm:kvm_async_pf_not_present                                    [100.00%]
                 0 kvm:kvm_async_pf_ready                                       [100.00%]
                 0 kvm:kvm_async_pf_completed

       1.002100578 seconds time elapsed

After:

 Performance counter stats for 'sleep 1s':

            28,354 kvm:kvm_entry                                                [99.98%]
                 0 kvm:kvm_hypercall                                            [99.98%]
                 0 kvm:kvm_hv_hypercall                                         [99.98%]
             1,347 kvm:kvm_pio                                                  [99.98%]
                 0 kvm:kvm_cpuid                                                [99.98%]
             1,931 kvm:kvm_apic                                                 [99.98%]
            29,595 kvm:kvm_exit                                                 [99.98%]
            24,884 kvm:kvm_inj_virq                                             [99.98%]
                 0 kvm:kvm_inj_exception                                        [99.98%]
                 0 kvm:kvm_page_fault                                           [99.98%]
             1,986 kvm:kvm_msr                                                  [99.98%]
                 0 kvm:kvm_cr                                                   [99.98%]
                 0 kvm:kvm_pic_set_irq                                          [99.98%]
                 0 kvm:kvm_apic_ipi                                             [99.99%]
            25,953 kvm:kvm_apic_accept_irq                                      [99.99%]
            26,132 kvm:kvm_eoi                                                  [99.99%]
            26,593 kvm:kvm_pv_eoi                                               [99.99%]
                 0 kvm:kvm_nested_vmrun                                         [99.99%]
                 0 kvm:kvm_nested_intercepts                                    [99.99%]
                 0 kvm:kvm_nested_vmexit                                        [99.99%]
                 0 kvm:kvm_nested_vmexit_inject                                    [99.99%]
                 0 kvm:kvm_nested_intr_vmexit                                    [99.99%]
                 0 kvm:kvm_invlpga                                              [99.99%]
                 0 kvm:kvm_skinit                                               [99.99%]
               284 kvm:kvm_emulate_insn                                         [99.99%]
                68 kvm:vcpu_match_mmio                                          [99.99%]
                68 kvm:kvm_userspace_exit                                       [99.99%]
                 2 kvm:kvm_set_irq                                              [99.99%]
                 2 kvm:kvm_ioapic_set_irq                                       [99.99%]
            28,288 kvm:kvm_msi_set_irq                                          [99.99%]
                 1 kvm:kvm_ack_irq                                              [99.99%]
               131 kvm:kvm_mmio                                                 [100.00%]
               588 kvm:kvm_fpu                                                  [100.00%]
                 0 kvm:kvm_age_page                                             [100.00%]
                 0 kvm:kvm_try_async_get_page                                    [100.00%]
                 0 kvm:kvm_async_pf_doublefault                                    [100.00%]
                 0 kvm:kvm_async_pf_not_present                                    [100.00%]
                 0 kvm:kvm_async_pf_ready                                       [100.00%]
                 0 kvm:kvm_async_pf_completed

       1.002039622 seconds time elapsed

We see that # of exits is almost halved.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>

2012-06-25 12:38:06 +03:00

..

acpi

Merge branch 'x86-trampoline-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2012-05-29 20:14:53 -07:00

apic

Merge branch 'delete-mca' of git://git.kernel.org/pub/scm/linux/kernel/git/paulg/linux

2012-05-23 17:12:06 -07:00

cpu

Merge remote-tracking branch 'rostedt/tip/perf/urgent-2' into x86-urgent-for-linus

2012-06-01 15:55:31 -07:00

.gitignore

…

alternative.c

x86: Call stop_machine_text_poke() on all CPUs

2011-11-14 13:05:15 +01:00

amd_gart_64.c

X86 & IA64: adapt for dma_map_ops changes

2012-03-28 16:36:31 +02:00

amd_nb.c

Merge branch 'linux-next' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes/pci

2012-01-11 18:50:26 -08:00

apb_timer.c

Merge branch 'timers-clocksource-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip

2011-07-23 10:34:47 -07:00

aperture_64.c

memblock, x86: Replace memblock_x86_reserve/free_range() with generic ones

2011-07-14 11:47:53 -07:00

apm_32.c

x86: Use kick_all_cpus_sync()

2012-05-08 12:35:06 +02:00

asm-offsets_32.c

x86: Generate system call tables and unistd_*.h from tables

2011-11-17 13:35:37 -08:00

asm-offsets_64.c

x32: If configured, add x32 system calls to system call tables

2012-02-20 12:52:06 -08:00

asm-offsets.c

x86, efi: EFI boot stub support

2011-12-12 14:26:10 -08:00

audit_64.c

…

bootflag.c

…

check.c

x86: kernel/check.c simple_strtoul cleanup

2012-05-15 15:36:41 -07:00

cpuid.c

Disintegrate asm/system.h for X86

2012-03-28 18:11:12 +01:00

crash_dump_32.c

x86: remove the second argument of k[un]map_atomic()

2012-03-20 21:48:15 +08:00

crash_dump_64.c

crash_dump: export is_kdump_kernel to modules, consolidate elfcorehdr_addr, setup_elfcorehdr and saved_max_pfn

2011-03-23 19:47:19 -07:00

crash.c

x86, nmi: Wire up NMI handlers to new routines

2011-10-10 06:56:57 +02:00

devicetree.c

irq_domain/x86: Convert x86 (embedded) to use common irq_domain

2012-02-23 14:37:47 -07:00

doublefault_32.c

…

dumpstack_32.c

x86: Avoid double stack traces with show_regs()

2012-05-09 11:44:42 +02:00

dumpstack_64.c

x86: Avoid double stack traces with show_regs()

2012-05-09 11:44:42 +02:00

dumpstack.c

Merge branches 'x86-asm-for-linus', 'x86-cleanups-for-linus', 'x86-cpu-for-linus', 'x86-debug-for-linus' and 'x86-microcode-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2012-05-23 10:09:50 -07:00

e820.c

x86: print e820 physical addresses consistently with other parts of kernel

2012-05-29 16:22:20 -07:00

early_printk.c

Merge branch 'x86-platform-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2012-01-11 19:13:40 -08:00

early-quirks.c

x86, quirk: Fix SB600 revision check

2011-03-16 14:03:32 +01:00

entry_32.S

x86: get rid of calling do_notify_resume() when returning to kernel mode

2012-06-01 13:01:51 -04:00

entry_64.S

ftrace/x86: Do not change stacks in DEBUG when calling lockdep

2012-05-31 23:12:22 -04:00

ftrace.c

ftrace: Use breakpoint method to update ftrace caller

2012-05-31 23:12:19 -04:00

head32.c

x86, realmode: Move ACPI wakeup to unified realmode code

2012-05-08 11:46:05 -07:00

head64.c

x86, realmode: Move ACPI wakeup to unified realmode code

2012-05-08 11:46:05 -07:00

head_32.S

Merge branch 'x86-trampoline-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2012-05-29 20:14:53 -07:00

head_64.S

Merge branch 'x86-trampoline-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2012-05-29 20:14:53 -07:00

head.c

memblock, x86: Replace memblock_x86_reserve/free_range() with generic ones

2011-07-14 11:47:53 -07:00

hpet.c

x86: hpet: Fix copy-and-paste mistake in earlier change

2012-05-25 15:32:29 +02:00

hw_breakpoint.c

x86: Use this_cpu_ops to optimize code

2010-12-30 12:20:28 +01:00

i386_ksyms_32.c

…

i387.c

x86: replace percpu_xxx funcs with this_cpu_xxx

2012-05-14 14:15:31 -07:00

i8237.c

x86: Use syscore_ops instead of sysdev classes and sysdevs

2011-03-23 22:15:54 +01:00

i8253.c

x86: Use common i8253 clockevent

2011-07-01 10:37:14 +02:00

i8259.c

Disintegrate asm/system.h for X86

2012-03-28 18:11:12 +01:00

io_delay.c

…

ioport.c

x86: Use bitmap library functions

2011-02-17 14:59:22 +01:00

irq_32.c

x86: Use common threadinfo allocator

2012-05-08 14:08:44 +02:00

irq_64.c

x86: Add stack top margin for stack overflow checking

2011-12-07 09:27:11 +01:00

irq_work.c

irq_work: Add generic hardirq context callbacks

2010-10-18 19:58:50 +02:00

irq.c

x86: Preserve lazy irq disable semantics in fixup_irqs()

2012-03-29 15:28:47 +02:00

irqinit.c

Merge branch 'x86-cleanups-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2012-03-29 18:21:35 -07:00

jump_label.c

jump_label, x86: Fix section mismatch

2011-12-06 20:41:02 +01:00

kdebugfs.c

simple_open: automatically convert to simple_open()

2012-04-05 15:25:50 -07:00

kgdb.c

KGDB/KDB regression fixes

2012-04-04 17:26:08 -07:00

kprobes-common.h

x86/kprobes: Split out optprobe related code to kprobes-opt.c

2012-03-06 09:49:49 +01:00

kprobes-opt.c

x86/kprobes: Split out optprobe related code to kprobes-opt.c

2012-03-06 09:49:49 +01:00

kprobes.c

x86: Avoid double stack traces with show_regs()

2012-05-09 11:44:42 +02:00

kvm.c

KVM guest: guest side for eoi avoidance

2012-06-25 12:38:06 +03:00

kvmclock.c

kvmclock: remove unneeded EXPORT macro

2012-04-08 12:49:54 +03:00

ldt.c

Disintegrate asm/system.h for X86

2012-03-28 18:11:12 +01:00

machine_kexec_32.c

Disintegrate asm/system.h for X86

2012-03-28 18:11:12 +01:00

machine_kexec_64.c

x86, cleanups: Use clear_page/copy_page rather than memset/memcpy

2010-09-22 15:36:49 -07:00

Makefile

Merge branch 'x86-trampoline-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2012-05-29 20:14:53 -07:00

microcode_amd.c

x86, microcode: Ensure that module is only loaded on supported AMD CPUs

2012-04-13 11:51:05 +02:00

microcode_core.c

x86, microcode: microcode_core.c simple_strtoul cleanup

2012-05-07 11:36:49 -07:00

microcode_intel.c

x86/microcode: Ensure that module is only loaded on supported Intel CPUs

2012-05-07 14:37:14 +02:00

mmconf-fam10h_64.c

x86-64: Fix and clean up AMD Fam10 MMCONF enabling

2010-11-18 13:41:35 +01:00

module.c

Disintegrate asm/system.h for X86

2012-03-28 18:11:12 +01:00

mpparse.c

Merge branch 'x86-trampoline-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2012-05-29 20:14:53 -07:00

msr.c

Disintegrate asm/system.h for X86

2012-03-28 18:11:12 +01:00

nmi_selftest.c

Merge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2012-05-23 10:21:19 -07:00

nmi.c

x86: Reset the debug_stack update counter

2012-05-31 23:12:20 -04:00

paravirt_patch_32.c

…

paravirt_patch_64.c

…

paravirt-spinlocks.c

…

paravirt.c

x86: replace percpu_xxx funcs with this_cpu_xxx

2012-05-14 14:15:31 -07:00

pci-calgary_64.c

Merge branches 'x86-asm-for-linus', 'x86-cleanups-for-linus', 'x86-cpu-for-linus', 'x86-debug-for-linus' and 'x86-microcode-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2012-05-23 10:09:50 -07:00

pci-dma.c

X86: integrate CMA with DMA-mapping subsystem

2012-05-21 15:09:38 +02:00

pci-iommu_table.c

arch/x86/kernel/pci-iommu_table.c: Convert sprintf_symbol to %pS

2011-05-10 10:21:35 +02:00

pci-nommu.c

X86: integrate CMA with DMA-mapping subsystem

2012-05-21 15:09:38 +02:00

pci-swiotlb.c

X86 & IA64: adapt for dma_map_ops changes

2012-03-28 16:36:31 +02:00

pcspeaker.c

…

probe_roms.c

x86: Include probe_roms.h in probe_roms.c

2012-03-08 10:57:35 +01:00

process_32.c

Merge branch 'x86-fpu-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2012-05-23 10:59:07 -07:00

process_64.c

Merge branch 'x86-fpu-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2012-05-23 10:59:07 -07:00

process.c

Merge branch 'x86-fpu-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2012-05-23 10:59:07 -07:00

ptrace.c

x86, x32, ptrace: Remove PTRACE_ARCH_PRCTL for x32

2012-06-01 13:54:21 -07:00

pvclock.c

x86/pvclock: Zero last_value on resume

2010-11-28 09:33:20 +01:00

quirks.c

x86, amd: Fix up numa_node information for AMD CPU family 15h model 0-0fh northbridge functions

2011-12-05 18:13:11 +01:00

reboot_fixups_32.c

x86: Ce4100: Add reboot_fixup() for CE4100

2010-11-12 00:45:41 +01:00

reboot.c

Merge branch 'x86-trampoline-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2012-05-29 20:14:53 -07:00

relocate_kernel_32.S

kexec, x86: Fix incorrect jump back address if not preserving context

2011-07-21 11:19:28 +02:00

relocate_kernel_64.S

kexec, x86: Fix incorrect jump back address if not preserving context

2011-07-21 11:19:28 +02:00

resource.c

x86: avoid high BIOS area when allocating address space

2010-12-17 10:01:30 -08:00

rtc.c

x86/rtc, mrst: Don't register a platform RTC device for for Intel MID platforms

2011-12-05 17:09:21 +01:00

setup_percpu.c

percpu, x86: don't use PMD_SIZE as embedded atom_size on 32bit

2012-05-08 09:42:18 -07:00

setup.c

Merge branch 'x86-trampoline-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2012-05-29 20:14:53 -07:00

signal.c

x86: get rid of calling do_notify_resume() when returning to kernel mode

2012-06-01 13:01:51 -04:00

smp.c

x86/reboot: Update nonmi_ipi parameter

2012-05-14 11:49:38 +02:00

smpboot.c

Merge branch 'x86-trampoline-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2012-05-29 20:14:53 -07:00

stacktrace.c

x86: Swap save_stack_trace_regs parameters

2011-06-14 22:48:51 -04:00

step.c

x86-64: Add user_64bit_mode paravirt op

2011-08-04 16:13:49 -07:00

sys_i386_32.c

i386: Make kernel_execve() suitable for stack unwinding

2010-09-03 08:16:02 +02:00

sys_x86_64.c

Merge branch 'x86-x32-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2012-03-29 18:12:23 -07:00

syscall_32.c

x86, syscall: Re-fix typo in comment

2011-11-18 16:25:07 -08:00

syscall_64.c

x32: If configured, add x32 system calls to system call tables

2012-02-20 12:52:06 -08:00

tboot.c

x86, realmode: fixes compilation issue in tboot.c

2012-05-08 15:04:27 -07:00

tce_64.c

Disintegrate asm/system.h for X86

2012-03-28 18:11:12 +01:00

test_nx.c

x86: Eliminate various 'set but not used' warnings

2011-05-21 19:10:33 +02:00

test_rodata.c

x86, extable: Remove open-coded exception table entries in arch/x86/kernel/test_rodata.c

2012-04-20 13:51:38 -07:00

time.c

MCA: delete all remaining traces of microchannel bus support.

2012-05-17 19:06:13 -04:00

tls.c

Merge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2012-03-29 14:28:26 -07:00

tls.h

…

topology.c

x86: Fix files explicitly requiring export.h for EXPORT_SYMBOL/THIS_MODULE

2011-10-31 19:30:35 -04:00

traps.c

ftrace: Synchronize variable setting with breakpoints

2012-05-31 23:12:17 -04:00

tsc_sync.c

x86/tsc: Reduce the TSC sync check time for core-siblings

2012-02-22 11:49:40 +01:00

tsc.c

Merge branch 'timers-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2012-03-29 14:16:48 -07:00

uprobes.c

uprobes/core: Handle breakpoint and singlestep exceptions

2012-03-14 07:41:36 +01:00

verify_cpu.S

x86: Fix common misspellings

2011-03-18 10:39:30 +01:00

vm86_32.c

Merge branch 'x86-cleanups-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2012-03-29 18:21:35 -07:00

vmlinux.lds.S

x86, realmode: Move ACPI wakeup to unified realmode code

2012-05-08 11:46:05 -07:00

vsmp_64.c

x86/vsmp: Fix number of CPUs when vsmp is disabled

2012-05-14 14:42:33 +02:00

vsyscall_64.c

x86: vsyscall: Use NULL instead 0 for a pointer argument

2012-04-06 11:49:59 +02:00

vsyscall_emu_64.S

x86-64: Rework vsyscall emulation and add vsyscall= parameter

2011-08-10 19:26:46 -05:00

vsyscall_trace.h

x86-64: Add vsyscall:emulate_vsyscall trace event

2011-08-04 16:13:53 -07:00

x86_init.c

x86/apic: Replace io_apic_ops with x86_io_apic_ops.

2012-05-01 14:50:09 -04:00

x8664_ksyms_64.c

x86-64, mem: Convert memmove() to assembly file and fix return value bug

2011-01-25 16:58:39 -08:00

xsave.c

x86, xsave: remove thread_has_fpu() bug check in __sanitize_i387_state()

2012-05-16 15:17:17 -07:00