From 38e20b07efd541a959de367dc90a17f92ce2e8a6 Mon Sep 17 00:00:00 2001 From: Sheng Yang Date: Fri, 14 May 2010 12:40:51 +0100 Subject: x86/xen: event channels delivery on HVM. Set the callback to receive evtchns from Xen, using the callback vector delivery mechanism. The traditional way for receiving event channel notifications from Xen is via the interrupts from the platform PCI device. The callback vector is a newer alternative that allow us to receive notifications on any vcpu and doesn't need any PCI support: we allocate a vector exclusively to receive events, in the vector handler we don't need to interact with the vlapic, therefore we avoid a VMEXIT. Signed-off-by: Stefano Stabellini Signed-off-by: Sheng Yang Signed-off-by: Jeremy Fitzhardinge --- include/xen/interface/features.h | 3 +++ 1 file changed, 3 insertions(+) (limited to 'include/xen/interface/features.h') diff --git a/include/xen/interface/features.h b/include/xen/interface/features.h index f51b6413b054..8ab08b91bf6f 100644 --- a/include/xen/interface/features.h +++ b/include/xen/interface/features.h @@ -41,6 +41,9 @@ /* x86: Does this Xen host support the MMU_PT_UPDATE_PRESERVE_AD hypercall? */ #define XENFEAT_mmu_pt_update_preserve_ad 5 +/* x86: Does this Xen host support the HVM callback vector type? */ +#define XENFEAT_hvm_callback_vector 8 + #define XENFEAT_NR_SUBMAPS 1 #endif /* __XEN_PUBLIC_FEATURES_H__ */ -- cgit v1.2.3 From 409771d258e9dd71c30f3c9520fd2b796ffc40f0 Mon Sep 17 00:00:00 2001 From: Stefano Stabellini Date: Fri, 14 May 2010 12:48:19 +0100 Subject: x86: Use xen_vcpuop_clockevent, xen_clocksource and xen wallclock. Use xen_vcpuop_clockevent instead of hpet and APIC timers as main clockevent device on all vcpus, use the xen wallclock time as wallclock instead of rtc and use xen_clocksource as clocksource. The pv clock algorithm needs to work correctly for the xen_clocksource and xen wallclock to be usable, only modern Xen versions offer a reliable pv clock in HVM guests (XENFEAT_hvm_safe_pvclock). Using the hpet as clocksource means a VMEXIT every time we read/write to the hpet mmio addresses, pvclock give us a better rating without VMEXITs. Same goes for the xen wallclock and xen_vcpuop_clockevent Signed-off-by: Stefano Stabellini Signed-off-by: Don Dutile Signed-off-by: Jeremy Fitzhardinge --- arch/x86/xen/enlighten.c | 14 ++-------- arch/x86/xen/suspend.c | 6 +++++ arch/x86/xen/time.c | 58 ++++++++++++++++++++++++++++++++++++---- arch/x86/xen/xen-ops.h | 7 ++--- include/xen/interface/features.h | 3 +++ 5 files changed, 66 insertions(+), 22 deletions(-) (limited to 'include/xen/interface/features.h') diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index 127c95c8d15e..a90172963884 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -935,10 +935,6 @@ static const struct pv_init_ops xen_init_ops __initdata = { .patch = xen_patch, }; -static const struct pv_time_ops xen_time_ops __initdata = { - .sched_clock = xen_sched_clock, -}; - static const struct pv_cpu_ops xen_cpu_ops __initdata = { .cpuid = xen_cpuid, @@ -1076,7 +1072,6 @@ asmlinkage void __init xen_start_kernel(void) /* Install Xen paravirt ops */ pv_info = xen_info; pv_init_ops = xen_init_ops; - pv_time_ops = xen_time_ops; pv_cpu_ops = xen_cpu_ops; pv_apic_ops = xen_apic_ops; @@ -1084,13 +1079,7 @@ asmlinkage void __init xen_start_kernel(void) x86_init.oem.arch_setup = xen_arch_setup; x86_init.oem.banner = xen_banner; - x86_init.timers.timer_init = xen_time_init; - x86_init.timers.setup_percpu_clockev = x86_init_noop; - x86_cpuinit.setup_percpu_clockev = x86_init_noop; - - x86_platform.calibrate_tsc = xen_tsc_khz; - x86_platform.get_wallclock = xen_get_wallclock; - x86_platform.set_wallclock = xen_set_wallclock; + xen_init_time_ops(); /* * Set up some pagetable state before starting to set any ptes. @@ -1327,6 +1316,7 @@ static void __init xen_hvm_guest_init(void) register_cpu_notifier(&xen_hvm_cpu_notifier); have_vcpu_info_placement = 0; x86_init.irqs.intr_init = xen_init_IRQ; + xen_hvm_init_time_ops(); } static bool __init xen_hvm_platform(void) diff --git a/arch/x86/xen/suspend.c b/arch/x86/xen/suspend.c index d07479c340f5..1d789d56877c 100644 --- a/arch/x86/xen/suspend.c +++ b/arch/x86/xen/suspend.c @@ -28,8 +28,14 @@ void xen_pre_suspend(void) void xen_hvm_post_suspend(int suspend_cancelled) { + int cpu; xen_hvm_init_shared_info(); xen_callback_vector(); + if (xen_feature(XENFEAT_hvm_safe_pvclock)) { + for_each_online_cpu(cpu) { + xen_setup_runstate_info(cpu); + } + } } void xen_post_suspend(int suspend_cancelled) diff --git a/arch/x86/xen/time.c b/arch/x86/xen/time.c index b3c6c59ed302..4780e55886a5 100644 --- a/arch/x86/xen/time.c +++ b/arch/x86/xen/time.c @@ -20,6 +20,7 @@ #include #include +#include #include #include @@ -160,7 +161,7 @@ static void do_stolen_accounting(void) * nanoseconds, which is nanoseconds the VCPU spent in RUNNING+BLOCKED * states. */ -unsigned long long xen_sched_clock(void) +static unsigned long long xen_sched_clock(void) { struct vcpu_runstate_info state; cycle_t now; @@ -195,7 +196,7 @@ unsigned long long xen_sched_clock(void) /* Get the TSC speed from Xen */ -unsigned long xen_tsc_khz(void) +static unsigned long xen_tsc_khz(void) { struct pvclock_vcpu_time_info *info = &HYPERVISOR_shared_info->vcpu_info[0].time; @@ -230,7 +231,7 @@ static void xen_read_wallclock(struct timespec *ts) put_cpu_var(xen_vcpu); } -unsigned long xen_get_wallclock(void) +static unsigned long xen_get_wallclock(void) { struct timespec ts; @@ -238,7 +239,7 @@ unsigned long xen_get_wallclock(void) return ts.tv_sec; } -int xen_set_wallclock(unsigned long now) +static int xen_set_wallclock(unsigned long now) { /* do nothing for domU */ return -1; @@ -473,7 +474,11 @@ void xen_timer_resume(void) } } -__init void xen_time_init(void) +static const struct pv_time_ops xen_time_ops __initdata = { + .sched_clock = xen_sched_clock, +}; + +static __init void xen_time_init(void) { int cpu = smp_processor_id(); struct timespec tp; @@ -497,3 +502,46 @@ __init void xen_time_init(void) xen_setup_timer(cpu); xen_setup_cpu_clockevents(); } + +__init void xen_init_time_ops(void) +{ + pv_time_ops = xen_time_ops; + + x86_init.timers.timer_init = xen_time_init; + x86_init.timers.setup_percpu_clockev = x86_init_noop; + x86_cpuinit.setup_percpu_clockev = x86_init_noop; + + x86_platform.calibrate_tsc = xen_tsc_khz; + x86_platform.get_wallclock = xen_get_wallclock; + x86_platform.set_wallclock = xen_set_wallclock; +} + +static void xen_hvm_setup_cpu_clockevents(void) +{ + int cpu = smp_processor_id(); + xen_setup_runstate_info(cpu); + xen_setup_timer(cpu); + xen_setup_cpu_clockevents(); +} + +__init void xen_hvm_init_time_ops(void) +{ + /* vector callback is needed otherwise we cannot receive interrupts + * on cpu > 0 */ + if (!xen_have_vector_callback && num_present_cpus() > 1) + return; + if (!xen_feature(XENFEAT_hvm_safe_pvclock)) { + printk(KERN_INFO "Xen doesn't support pvclock on HVM," + "disable pv timer\n"); + return; + } + + pv_time_ops = xen_time_ops; + x86_init.timers.setup_percpu_clockev = xen_time_init; + x86_cpuinit.setup_percpu_clockev = xen_hvm_setup_cpu_clockevents; + + x86_platform.calibrate_tsc = xen_tsc_khz; + x86_platform.get_wallclock = xen_get_wallclock; + x86_platform.set_wallclock = xen_set_wallclock; +} + diff --git a/arch/x86/xen/xen-ops.h b/arch/x86/xen/xen-ops.h index 01c9dd386522..089d18923d2b 100644 --- a/arch/x86/xen/xen-ops.h +++ b/arch/x86/xen/xen-ops.h @@ -49,11 +49,8 @@ void xen_setup_runstate_info(int cpu); void xen_teardown_timer(int cpu); cycle_t xen_clocksource_read(void); void xen_setup_cpu_clockevents(void); -unsigned long xen_tsc_khz(void); -void __init xen_time_init(void); -unsigned long xen_get_wallclock(void); -int xen_set_wallclock(unsigned long time); -unsigned long long xen_sched_clock(void); +void __init xen_init_time_ops(void); +void __init xen_hvm_init_time_ops(void); irqreturn_t xen_debug_interrupt(int irq, void *dev_id); diff --git a/include/xen/interface/features.h b/include/xen/interface/features.h index 8ab08b91bf6f..70d2563ab166 100644 --- a/include/xen/interface/features.h +++ b/include/xen/interface/features.h @@ -44,6 +44,9 @@ /* x86: Does this Xen host support the HVM callback vector type? */ #define XENFEAT_hvm_callback_vector 8 +/* x86: pvclock algorithm is safe to use on HVM */ +#define XENFEAT_hvm_safe_pvclock 9 + #define XENFEAT_NR_SUBMAPS 1 #endif /* __XEN_PUBLIC_FEATURES_H__ */ -- cgit v1.2.3