Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   Vitaly Kuznetsov <vkuznets@redhat.com>
To:     Wanpeng Li <kernellwp@gmail.com>
Cc:     Paolo Bonzini <pbonzini@redhat.com>,
        Sean Christopherson <sean.j.christopherson@intel.com>,
        Wanpeng Li <wanpengli@tencent.com>,
        Jim Mattson <jmattson@google.com>,
        Joerg Roedel <joro@8bytes.org>,
        LKML <linux-kernel@vger.kernel.org>, kvm <kvm@vger.kernel.org>
Subject: Re: [PATCH v2 2/2] KVM: Pre-allocate 1 cpumask variable per cpu for both pv tlb and pv ipis
In-Reply-To: <CANRm+Cz_gskKwa0SU0PUhtacj3Ovm_MmBASDJHOECsnYz=jxkg@mail.gmail.com>
References: <CANRm+CxGOeGQ0vV9ueBgjUDvkzH29EQWLe4GQGDvOhm3idM6NQ@mail.gmail.com> <871rqtbcve.fsf@vitty.brq.redhat.com> <CANRm+Cz_gskKwa0SU0PUhtacj3Ovm_MmBASDJHOECsnYz=jxkg@mail.gmail.com>
Date:   Mon, 17 Feb 2020 15:16:46 +0100
Message-ID: <87y2t19u41.fsf@vitty.brq.redhat.com>
MIME-Version: 1.0
Content-Type: text/plain
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Wanpeng Li <kernellwp@gmail.com> writes:

> On Mon, 17 Feb 2020 at 20:46, Vitaly Kuznetsov <vkuznets@redhat.com> wrote:
>>
>> Wanpeng Li <kernellwp@gmail.com> writes:
>>
>> > From: Wanpeng Li <wanpengli@tencent.com>
>> >
>> > Nick Desaulniers Reported:
>> >
>> >   When building with:
>> >   $ make CC=clang arch/x86/ CFLAGS=-Wframe-larger-than=1000
>> >   The following warning is observed:
>> >   arch/x86/kernel/kvm.c:494:13: warning: stack frame size of 1064 bytes in
>> >   function 'kvm_send_ipi_mask_allbutself' [-Wframe-larger-than=]
>> >   static void kvm_send_ipi_mask_allbutself(const struct cpumask *mask, int
>> >   vector)
>> >               ^
>> >   Debugging with:
>> >   https://github.com/ClangBuiltLinux/frame-larger-than
>> >   via:
>> >   $ python3 frame_larger_than.py arch/x86/kernel/kvm.o \
>> >     kvm_send_ipi_mask_allbutself
>> >   points to the stack allocated `struct cpumask newmask` in
>> >   `kvm_send_ipi_mask_allbutself`. The size of a `struct cpumask` is
>> >   potentially large, as it's CONFIG_NR_CPUS divided by BITS_PER_LONG for
>> >   the target architecture. CONFIG_NR_CPUS for X86_64 can be as high as
>> >   8192, making a single instance of a `struct cpumask` 1024 B.
>> >
>> > This patch fixes it by pre-allocate 1 cpumask variable per cpu and use it for
>> > both pv tlb and pv ipis..
>> >
>> > Reported-by: Nick Desaulniers <ndesaulniers@google.com>
>> > Acked-by: Nick Desaulniers <ndesaulniers@google.com>
>> > Cc: Peter Zijlstra <peterz@infradead.org>
>> > Cc: Nick Desaulniers <ndesaulniers@google.com>
>> > Signed-off-by: Wanpeng Li <wanpengli@tencent.com>
>> > ---
>> > v1 -> v2:
>> >  * remove '!alloc' check
>> >  * use new pv check helpers
>> >
>> >  arch/x86/kernel/kvm.c | 33 +++++++++++++++++++++------------
>> >  1 file changed, 21 insertions(+), 12 deletions(-)
>> >
>> > diff --git a/arch/x86/kernel/kvm.c b/arch/x86/kernel/kvm.c
>> > index 76ea8c4..377b224 100644
>> > --- a/arch/x86/kernel/kvm.c
>> > +++ b/arch/x86/kernel/kvm.c
>> > @@ -432,6 +432,8 @@ static bool pv_tlb_flush_supported(void)
>> >          kvm_para_has_feature(KVM_FEATURE_STEAL_TIME));
>> >  }
>> >
>> > +static DEFINE_PER_CPU(cpumask_var_t, __pv_cpu_mask);
>> > +
>> >  #ifdef CONFIG_SMP
>> >
>> >  static bool pv_ipi_supported(void)
>> > @@ -510,12 +512,12 @@ static void kvm_send_ipi_mask(const struct
>> > cpumask *mask, int vector)
>> >  static void kvm_send_ipi_mask_allbutself(const struct cpumask *mask,
>> > int vector)
>> >  {
>> >      unsigned int this_cpu = smp_processor_id();
>> > -    struct cpumask new_mask;
>> > +    struct cpumask *new_mask = this_cpu_cpumask_var_ptr(__pv_cpu_mask);
>> >      const struct cpumask *local_mask;
>> >
>> > -    cpumask_copy(&new_mask, mask);
>> > -    cpumask_clear_cpu(this_cpu, &new_mask);
>> > -    local_mask = &new_mask;
>> > +    cpumask_copy(new_mask, mask);
>> > +    cpumask_clear_cpu(this_cpu, new_mask);
>> > +    local_mask = new_mask;
>> >      __send_ipi_mask(local_mask, vector);
>> >  }
>> >
>> > @@ -595,7 +597,6 @@ static void __init kvm_apf_trap_init(void)
>> >      update_intr_gate(X86_TRAP_PF, async_page_fault);
>> >  }
>> >
>> > -static DEFINE_PER_CPU(cpumask_var_t, __pv_tlb_mask);
>> >
>> >  static void kvm_flush_tlb_others(const struct cpumask *cpumask,
>> >              const struct flush_tlb_info *info)
>> > @@ -603,7 +604,7 @@ static void kvm_flush_tlb_others(const struct
>> > cpumask *cpumask,
>> >      u8 state;
>> >      int cpu;
>> >      struct kvm_steal_time *src;
>> > -    struct cpumask *flushmask = this_cpu_cpumask_var_ptr(__pv_tlb_mask);
>> > +    struct cpumask *flushmask = this_cpu_cpumask_var_ptr(__pv_cpu_mask);
>> >
>> >      cpumask_copy(flushmask, cpumask);
>> >      /*
>> > @@ -642,6 +643,7 @@ static void __init kvm_guest_init(void)
>> >      if (pv_tlb_flush_supported()) {
>> >          pv_ops.mmu.flush_tlb_others = kvm_flush_tlb_others;
>> >          pv_ops.mmu.tlb_remove_table = tlb_remove_table;
>> > +        pr_info("KVM setup pv remote TLB flush\n");
>>
>> Nit: to be consistent with __send_ipi_mask() the message should be
>> somthing like
>>
>> "KVM: switch to using PV TLB flush"
>
> There is a lot of native ops we replace by pv ops in kvm.c, I use "KVM
> setup xxx" there, like pv ipis, pv tlb flush, pv sched yield, should
> we keep consistent as before?
>

Oh, I see, it's either one or another :-) Personally, I prefer when
subsystem is delimited with ':' so if we were to change them all I'd
prefer the "KVM: switch to using PV TLB flush" (__send_ipi_mask()
style). But this looks more like a separate patch idea, we can discuss
our personal preferences there :-)

-- 
Vitaly