Zhenzhong Duan <[email protected]> writes:
> There are cases where a guest tries to switch spinlocks to bare metal
> behavior (e.g. by setting "xen_nopvspin" on XEN platform and
> "hv_nopvspin" on HYPER_V).
>
> That feature is missed on KVM, add a new parameter "nopvspin" to disable
> PV spinlocks for KVM guest.
>
> The new 'nopvspin' parameter will also replace Xen and Hyper-V specific
> parameters in future patches.
>
> Define variable nopvsin as global because it will be used in future
> patches as above.
>
> Signed-off-by: Zhenzhong Duan <[email protected]>
> Cc: Jonathan Corbet <[email protected]>
> Cc: Thomas Gleixner <[email protected]>
> Cc: Ingo Molnar <[email protected]>
> Cc: Borislav Petkov <[email protected]>
> Cc: "H. Peter Anvin" <[email protected]>
> Cc: Paolo Bonzini <[email protected]>
> Cc: Radim Krcmar <[email protected]>
> Cc: Sean Christopherson <[email protected]>
> Cc: Vitaly Kuznetsov <[email protected]>
> Cc: Wanpeng Li <[email protected]>
> Cc: Jim Mattson <[email protected]>
> Cc: Joerg Roedel <[email protected]>
> Cc: Peter Zijlstra <[email protected]>
> Cc: Will Deacon <[email protected]>
> ---
> Documentation/admin-guide/kernel-parameters.txt | 5 ++++
> arch/x86/include/asm/qspinlock.h | 1 +
> arch/x86/kernel/kvm.c | 34
> ++++++++++++++++++++++---
> kernel/locking/qspinlock.c | 7 +++++
> 4 files changed, 43 insertions(+), 4 deletions(-)
>
> diff --git a/Documentation/admin-guide/kernel-parameters.txt
> b/Documentation/admin-guide/kernel-parameters.txt
> index a84a83f..bd49ed2 100644
> --- a/Documentation/admin-guide/kernel-parameters.txt
> +++ b/Documentation/admin-guide/kernel-parameters.txt
> @@ -5334,6 +5334,11 @@
> as generic guest with no PV drivers. Currently support
> XEN HVM, KVM, HYPER_V and VMWARE guest.
>
> + nopvspin [X86,KVM]
> + Disables the qspinlock slow path using PV optimizations
> + which allow the hypervisor to 'idle' the guest on lock
> + contention.
> +
> xirc2ps_cs= [NET,PCMCIA]
> Format:
>
> <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
> diff --git a/arch/x86/include/asm/qspinlock.h
> b/arch/x86/include/asm/qspinlock.h
> index 444d6fd..d86ab94 100644
> --- a/arch/x86/include/asm/qspinlock.h
> +++ b/arch/x86/include/asm/qspinlock.h
> @@ -32,6 +32,7 @@ static __always_inline u32
> queued_fetch_set_pending_acquire(struct qspinlock *lo
> extern void __pv_init_lock_hash(void);
> extern void __pv_queued_spin_lock_slowpath(struct qspinlock *lock, u32 val);
> extern void __raw_callee_save___pv_queued_spin_unlock(struct qspinlock
> *lock);
> +extern bool nopvspin;
>
> #define queued_spin_unlock queued_spin_unlock
> /**
> diff --git a/arch/x86/kernel/kvm.c b/arch/x86/kernel/kvm.c
> index 249f14a..e9c76d8 100644
> --- a/arch/x86/kernel/kvm.c
> +++ b/arch/x86/kernel/kvm.c
> @@ -825,18 +825,44 @@ __visible bool __kvm_vcpu_is_preempted(long cpu)
> */
> void __init kvm_spinlock_init(void)
> {
> - /* Does host kernel support KVM_FEATURE_PV_UNHALT? */
> - if (!kvm_para_has_feature(KVM_FEATURE_PV_UNHALT))
> + /*
> + * PV spinlocks is disabled if no host side support, then native
> + * qspinlock will be used. As native qspinlock is a fair lock, there is
> + * lock holder preemption issue using it in a guest, imaging one pCPU
> + * running 10 vCPUs of same guest contending same lock.
> + *
> + * virt_spin_lock() is introduced as an optimization for that scenario
> + * which is enabled by virt_spin_lock_key key. To use that optimization,
> + * virt_spin_lock_key isn't disabled here.
> + */
My take (if I properly understood what you say) would be:
"In case host doesn't support KVM_FEATURE_PV_UNHALT there is still an
advantage of keeping virt_spin_lock_key enabled: virt_spin_lock() is
preferred over native qspinlock when vCPU is preempted."
> + if (!kvm_para_has_feature(KVM_FEATURE_PV_UNHALT)) {
> + pr_info("PV spinlocks disabled, no host support.\n");
> return;
> + }
>
> + /*
> + * Disable PV qspinlock and use native qspinlock when dedicated pCPUs
> + * are available.
> + */
> if (kvm_para_has_hint(KVM_HINTS_REALTIME)) {
> + pr_info("PV spinlocks disabled with KVM_HINTS_REALTIME
> hints.\n");
> + static_branch_disable(&virt_spin_lock_key);
> + return;
> + }
> +
> + if (num_possible_cpus() == 1) {
> + pr_info("PV spinlocks disabled, single CPU.\n");
> static_branch_disable(&virt_spin_lock_key);
> return;
> }
>
> - /* Don't use the pvqspinlock code if there is only 1 vCPU. */
> - if (num_possible_cpus() == 1)
> + if (nopvspin) {
> + pr_info("PV spinlocks disabled, forced by \"nopvspin\"
> parameter.\n");
> + static_branch_disable(&virt_spin_lock_key);
> return;
You could've replaced this 'static_branch_disable(); return;' pattern
with a goto to the end of the function to save a few lines but this
looks good anyways.
Reviewed-by: Vitaly Kuznetsov <[email protected]>
> + }
> +
> + pr_info("PV spinlocks enabled\n");
>
> __pv_init_lock_hash();
> pv_ops.lock.queued_spin_lock_slowpath = __pv_queued_spin_lock_slowpath;
> diff --git a/kernel/locking/qspinlock.c b/kernel/locking/qspinlock.c
> index 2473f10..75193d6 100644
> --- a/kernel/locking/qspinlock.c
> +++ b/kernel/locking/qspinlock.c
> @@ -580,4 +580,11 @@ void queued_spin_lock_slowpath(struct qspinlock *lock,
> u32 val)
> #include "qspinlock_paravirt.h"
> #include "qspinlock.c"
>
> +bool nopvspin __initdata;
> +static __init int parse_nopvspin(char *arg)
> +{
> + nopvspin = true;
> + return 0;
> +}
> +early_param("nopvspin", parse_nopvspin);
> #endif
--
Vitaly
