| Message ID | 1477877822-4437-3-git-send-email-wanpeng.li@hotmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Mon, Oct 31, 2016 at 09:37:02AM +0800, Wanpeng Li wrote: > From: Wanpeng Li <wanpeng.li@hotmail.com> > > | RCU used illegally from idle CPU! > | rcu_scheduler_active = 1, debug_locks = 0 > | RCU used illegally from extended quiescent state! > | no locks held by swapper/1/0. > | > | [<ffffffff9d492b95>] do_trace_write_msr+0x135/0x140 > | [<ffffffff9d06f860>] native_write_msr+0x20/0x30 > | [<ffffffff9d065fad>] native_apic_msr_eoi_write+0x1d/0x30 > | [<ffffffff9d05bd1d>] smp_reschedule_interrupt+0x1d/0x30 > | [<ffffffff9d8daec6>] reschedule_interrupt+0x96/0xa0 Please remove the text between [] and the offsets after "+..." - those are useless in a commit message. > Reschedule interrupt may be called in cpu idle state. This causes lockdep s/cpu/CPU/ > check warning above. > > As Peterz pointed out: > > | So now we're making a very frequent interrupt slower because of debug > | code. > | > | The thing is, many many smp_reschedule_interrupt() invocations don't > | actually execute anything much at all and are only send to tickle the s/send/sent/ > | return to user path (which does the actual preemption). > | > | Having to do the whole irq_enter/irq_exit dance just for this unlikely > | debug case totally blows. > > This patch converts x2apic write eoi msr to notrace to avoid the debug > codes splash and reverts irq_enter/irq_exit dance to avoid to make a very > frequent interrupt slower because of debug code. > > Suggested-by: Peter Zijlstra <peterz@infradead.org> > Suggested-by: Paolo Bonzini <pbonzini@redhat.com> > Acked-by: Paolo Bonzini <pbonzini@redhat.com> > Cc: Ingo Molnar <mingo@kernel.org> > Cc: Mike Galbraith <efault@gmx.de> > Cc: Peter Zijlstra <peterz@infradead.org> > Cc: Thomas Gleixner <tglx@linutronix.de> > Cc: Paolo Bonzini <pbonzini@redhat.com> > Cc: Borislav Petkov <bp@alien8.de> > Signed-off-by: Wanpeng Li <wanpeng.li@hotmail.com> > ---
2016-11-03 5:10 GMT+08:00 Borislav Petkov <bp@alien8.de>: > On Mon, Oct 31, 2016 at 09:37:02AM +0800, Wanpeng Li wrote: >> From: Wanpeng Li <wanpeng.li@hotmail.com> >> >> | RCU used illegally from idle CPU! >> | rcu_scheduler_active = 1, debug_locks = 0 >> | RCU used illegally from extended quiescent state! >> | no locks held by swapper/1/0. >> | >> | [<ffffffff9d492b95>] do_trace_write_msr+0x135/0x140 >> | [<ffffffff9d06f860>] native_write_msr+0x20/0x30 >> | [<ffffffff9d065fad>] native_apic_msr_eoi_write+0x1d/0x30 >> | [<ffffffff9d05bd1d>] smp_reschedule_interrupt+0x1d/0x30 >> | [<ffffffff9d8daec6>] reschedule_interrupt+0x96/0xa0 > > Please remove the text between [] and the offsets after "+..." - those > are useless in a commit message. > >> Reschedule interrupt may be called in cpu idle state. This causes lockdep > > s/cpu/CPU/ > >> check warning above. >> >> As Peterz pointed out: >> >> | So now we're making a very frequent interrupt slower because of debug >> | code. >> | >> | The thing is, many many smp_reschedule_interrupt() invocations don't >> | actually execute anything much at all and are only send to tickle the > > s/send/sent/ I will send a new version to clean it up. Regards, Wanpeng Li -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/arch/x86/include/asm/apic.h b/arch/x86/include/asm/apic.h index f5aaf6c..a5a0bcf 100644 --- a/arch/x86/include/asm/apic.h +++ b/arch/x86/include/asm/apic.h @@ -196,7 +196,7 @@ static inline void native_apic_msr_write(u32 reg, u32 v) static inline void native_apic_msr_eoi_write(u32 reg, u32 v) { - wrmsr(APIC_BASE_MSR + (APIC_EOI >> 4), APIC_EOI_ACK, 0); + wrmsr_notrace(APIC_BASE_MSR + (APIC_EOI >> 4), APIC_EOI_ACK, 0); } static inline u32 native_apic_msr_read(u32 reg) @@ -332,6 +332,7 @@ struct apic { * on write for EOI. */ void (*eoi_write)(u32 reg, u32 v); + void (*native_eoi_write)(u32 reg, u32 v); u64 (*icr_read)(void); void (*icr_write)(u32 low, u32 high); void (*wait_icr_idle)(void); diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index 88c657b..2686894 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -2263,6 +2263,7 @@ void __init apic_set_eoi_write(void (*eoi_write)(u32 reg, u32 v)) for (drv = __apicdrivers; drv < __apicdrivers_end; drv++) { /* Should happen once for each apic */ WARN_ON((*drv)->eoi_write == eoi_write); + (*drv)->native_eoi_write = (*drv)->eoi_write; (*drv)->eoi_write = eoi_write; } } diff --git a/arch/x86/kernel/kvm.c b/arch/x86/kernel/kvm.c index edbbfc8..d230513 100644 --- a/arch/x86/kernel/kvm.c +++ b/arch/x86/kernel/kvm.c @@ -308,7 +308,7 @@ static void kvm_register_steal_time(void) static DEFINE_PER_CPU(unsigned long, kvm_apic_eoi) = KVM_PV_EOI_DISABLED; -static void kvm_guest_apic_eoi_write(u32 reg, u32 val) +static notrace void kvm_guest_apic_eoi_write(u32 reg, u32 val) { /** * This relies on __test_and_clear_bit to modify the memory @@ -319,7 +319,7 @@ static void kvm_guest_apic_eoi_write(u32 reg, u32 val) */ if (__test_and_clear_bit(KVM_PV_EOI_BIT, this_cpu_ptr(&kvm_apic_eoi))) return; - apic_write(APIC_EOI, APIC_EOI_ACK); + apic->native_eoi_write(APIC_EOI, APIC_EOI_ACK); } static void kvm_guest_cpu_init(void) diff --git a/arch/x86/kernel/smp.c b/arch/x86/kernel/smp.c index c00cb64..68f8cc2 100644 --- a/arch/x86/kernel/smp.c +++ b/arch/x86/kernel/smp.c @@ -261,10 +261,8 @@ static inline void __smp_reschedule_interrupt(void) __visible void smp_reschedule_interrupt(struct pt_regs *regs) { - irq_enter(); ack_APIC_irq(); __smp_reschedule_interrupt(); - irq_exit(); /* * KVM uses this interrupt to force a cpu out of guest mode */