From 7af192c954017499ec163bc9dbaaee2e593d7ef2 Mon Sep 17 00:00:00 2001 From: Gerd Hoffmann Date: Tue, 3 Jun 2008 16:17:29 +0200 Subject: x86: Add structs and functions for paravirt clocksource This patch adds structs for the paravirt clocksource ABI used by both xen and kvm (pvclock-abi.h). It also adds some helper functions to read system time and wall clock time from a paravirtual clocksource (pvclock.[ch]). They are based on the xen code. They are enabled using CONFIG_PARAVIRT_CLOCK. Subsequent patches of this series will put the code in use. Signed-off-by: Gerd Hoffmann Acked-by: Jeremy Fitzhardinge Signed-off-by: Avi Kivity --- arch/x86/Kconfig | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'arch/x86/Kconfig') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 52e18e6d2ba0..f94bca6ff47f 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -410,6 +410,10 @@ config PARAVIRT over full virtualization. However, when run without a hypervisor the kernel is theoretically slower and slightly larger. +config PARAVIRT_CLOCK + bool + default n + endif config MEMTEST_BOOTPARAM -- cgit v1.2.3-55-g7522 From f6e16d5ad463d15f285666f588cfe49495c692d9 Mon Sep 17 00:00:00 2001 From: Gerd Hoffmann Date: Tue, 3 Jun 2008 16:17:32 +0200 Subject: x86: KVM guest: Use the paravirt clocksource structs and functions This patch updates the kvm host code to use the pvclock structs and functions, thereby making it compatible with Xen. The patch also fixes an initialization bug: on SMP systems the per-cpu has two different locations early at boot and after CPU bringup. kvmclock must take that in account when registering the physical address within the host. Signed-off-by: Gerd Hoffmann Signed-off-by: Avi Kivity --- arch/x86/Kconfig | 1 + arch/x86/kernel/kvmclock.c | 89 +++++++++++++++++----------------------------- 2 files changed, 34 insertions(+), 56 deletions(-) (limited to 'arch/x86/Kconfig') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index f94bca6ff47f..e0edaaa6920a 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -383,6 +383,7 @@ config VMI config KVM_CLOCK bool "KVM paravirtualized clock" select PARAVIRT + select PARAVIRT_CLOCK depends on !(X86_VISWS || X86_VOYAGER) help Turning on this option will allow you to run a paravirtualized clock diff --git a/arch/x86/kernel/kvmclock.c b/arch/x86/kernel/kvmclock.c index 08a30986d472..87edf1ceb1df 100644 --- a/arch/x86/kernel/kvmclock.c +++ b/arch/x86/kernel/kvmclock.c @@ -18,6 +18,7 @@ #include #include +#include #include #include #include @@ -36,18 +37,9 @@ static int parse_no_kvmclock(char *arg) early_param("no-kvmclock", parse_no_kvmclock); /* The hypervisor will put information about time periodically here */ -static DEFINE_PER_CPU_SHARED_ALIGNED(struct kvm_vcpu_time_info, hv_clock); -#define get_clock(cpu, field) per_cpu(hv_clock, cpu).field +static DEFINE_PER_CPU_SHARED_ALIGNED(struct pvclock_vcpu_time_info, hv_clock); +static struct pvclock_wall_clock wall_clock; -static inline u64 kvm_get_delta(u64 last_tsc) -{ - int cpu = smp_processor_id(); - u64 delta = native_read_tsc() - last_tsc; - return (delta * get_clock(cpu, tsc_to_system_mul)) >> KVM_SCALE; -} - -static struct kvm_wall_clock wall_clock; -static cycle_t kvm_clock_read(void); /* * The wallclock is the time of day when we booted. Since then, some time may * have elapsed since the hypervisor wrote the data. So we try to account for @@ -55,64 +47,37 @@ static cycle_t kvm_clock_read(void); */ static unsigned long kvm_get_wallclock(void) { - u32 wc_sec, wc_nsec; - u64 delta; + struct pvclock_vcpu_time_info *vcpu_time; struct timespec ts; - int version, nsec; int low, high; low = (int)__pa(&wall_clock); high = ((u64)__pa(&wall_clock) >> 32); + native_write_msr(MSR_KVM_WALL_CLOCK, low, high); - delta = kvm_clock_read(); + vcpu_time = &get_cpu_var(hv_clock); + pvclock_read_wallclock(&wall_clock, vcpu_time, &ts); + put_cpu_var(hv_clock); - native_write_msr(MSR_KVM_WALL_CLOCK, low, high); - do { - version = wall_clock.wc_version; - rmb(); - wc_sec = wall_clock.wc_sec; - wc_nsec = wall_clock.wc_nsec; - rmb(); - } while ((wall_clock.wc_version != version) || (version & 1)); - - delta = kvm_clock_read() - delta; - delta += wc_nsec; - nsec = do_div(delta, NSEC_PER_SEC); - set_normalized_timespec(&ts, wc_sec + delta, nsec); - /* - * Of all mechanisms of time adjustment I've tested, this one - * was the champion! - */ - return ts.tv_sec + 1; + return ts.tv_sec; } static int kvm_set_wallclock(unsigned long now) { - return 0; + return -1; } -/* - * This is our read_clock function. The host puts an tsc timestamp each time - * it updates a new time. Without the tsc adjustment, we can have a situation - * in which a vcpu starts to run earlier (smaller system_time), but probes - * time later (compared to another vcpu), leading to backwards time - */ static cycle_t kvm_clock_read(void) { - u64 last_tsc, now; - int cpu; + struct pvclock_vcpu_time_info *src; + cycle_t ret; - preempt_disable(); - cpu = smp_processor_id(); - - last_tsc = get_clock(cpu, tsc_timestamp); - now = get_clock(cpu, system_time); - - now += kvm_get_delta(last_tsc); - preempt_enable(); - - return now; + src = &get_cpu_var(hv_clock); + ret = pvclock_clocksource_read(src); + put_cpu_var(hv_clock); + return ret; } + static struct clocksource kvm_clock = { .name = "kvm-clock", .read = kvm_clock_read, @@ -123,13 +88,14 @@ static struct clocksource kvm_clock = { .flags = CLOCK_SOURCE_IS_CONTINUOUS, }; -static int kvm_register_clock(void) +static int kvm_register_clock(char *txt) { int cpu = smp_processor_id(); int low, high; low = (int)__pa(&per_cpu(hv_clock, cpu)) | 1; high = ((u64)__pa(&per_cpu(hv_clock, cpu)) >> 32); - + printk(KERN_INFO "kvm-clock: cpu %d, msr %x:%x, %s\n", + cpu, high, low, txt); return native_write_msr_safe(MSR_KVM_SYSTEM_TIME, low, high); } @@ -140,12 +106,20 @@ static void kvm_setup_secondary_clock(void) * Now that the first cpu already had this clocksource initialized, * we shouldn't fail. */ - WARN_ON(kvm_register_clock()); + WARN_ON(kvm_register_clock("secondary cpu clock")); /* ok, done with our trickery, call native */ setup_secondary_APIC_clock(); } #endif +#ifdef CONFIG_SMP +void __init kvm_smp_prepare_boot_cpu(void) +{ + WARN_ON(kvm_register_clock("primary cpu clock")); + native_smp_prepare_boot_cpu(); +} +#endif + /* * After the clock is registered, the host will keep writing to the * registered memory location. If the guest happens to shutdown, this memory @@ -174,13 +148,16 @@ void __init kvmclock_init(void) return; if (kvmclock && kvm_para_has_feature(KVM_FEATURE_CLOCKSOURCE)) { - if (kvm_register_clock()) + if (kvm_register_clock("boot clock")) return; pv_time_ops.get_wallclock = kvm_get_wallclock; pv_time_ops.set_wallclock = kvm_set_wallclock; pv_time_ops.sched_clock = kvm_clock_read; #ifdef CONFIG_X86_LOCAL_APIC pv_apic_ops.setup_secondary_clock = kvm_setup_secondary_clock; +#endif +#ifdef CONFIG_SMP + smp_ops.smp_prepare_boot_cpu = kvm_smp_prepare_boot_cpu; #endif machine_ops.shutdown = kvm_shutdown; #ifdef CONFIG_KEXEC -- cgit v1.2.3-55-g7522 From efac41894df57d32b483ac622d03541b5b2692c0 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 1 Jul 2008 08:56:32 +0200 Subject: x86: fix NODES_SHIFT Kconfig range commit 4323838215184f5a2f081e0d17b8d60731b03164 x86: change size of node ids from u8 to s16 set the range for NODES_SHIFT to 1..15. The possible range is 1..9 Fixes Bugzilla #10726 Reported-by: Dave Jones Signed-off-by: Thomas Gleixner --- arch/x86/Kconfig | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86/Kconfig') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 52e18e6d2ba0..8a07f417f5f4 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -961,8 +961,8 @@ config NUMA_EMU number of nodes. This is only useful for debugging. config NODES_SHIFT - int "Max num nodes shift(1-15)" - range 1 15 if X86_64 + int "Max num nodes shift(1-9)" + range 1 9 if X86_64 default "6" if X86_64 default "4" if X86_NUMAQ default "3" -- cgit v1.2.3-55-g7522