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;
MIME-Version: 1.0
References: <20190423065706.15430-1-namit@vmware.com> <CALCETrUohp2rMr9EcoAYm2M=Ktu4Bg2sPfUWMy5JqM-GXsDOBQ@mail.gmail.com>
 <44A66873-0834-4794-81A5-3B2131314AA4@vmware.com>
In-Reply-To: <44A66873-0834-4794-81A5-3B2131314AA4@vmware.com>
From:   Andy Lutomirski <luto@kernel.org>
Date:   Tue, 23 Apr 2019 10:23:47 -0700
Message-ID: <CALCETrXLjL4vMwQP1tweWvk_+dBvZFdE85ckt8TZSVw2XyW_yQ@mail.gmail.com>
Subject: Re: [PATCH] x86/mm/tlb: Remove flush_tlb_info from the stack
To:     Nadav Amit <namit@vmware.com>
Cc:     Andy Lutomirski <luto@kernel.org>,
        Peter Zijlstra <peterz@infradead.org>,
        Borislav Petkov <bp@alien8.de>, Ingo Molnar <mingo@redhat.com>,
        Thomas Gleixner <tglx@linutronix.de>, X86 ML <x86@kernel.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Dave Hansen <dave.hansen@intel.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, Apr 23, 2019 at 9:56 AM Nadav Amit <namit@vmware.com> wrote:
>
> > On Apr 23, 2019, at 9:50 AM, Andy Lutomirski <luto@kernel.org> wrote:
> >
> > On Tue, Apr 23, 2019 at 12:12 AM Nadav Amit <namit@vmware.com> wrote:
>https://git.kernel.org/pub/scm/linux/kernel/git/luto/linux.git/log/?h=3Dx8=
6/fixes >> Remove flush_tlb_info variables from the stack. This allows to a=
lign
> >> flush_tlb_info to cache-line and avoid potentially unnecessary cache
> >> line movements. It also allows to have a fixed virtual-to-physical
> >> translation of the variables, which reduces TLB misses.
> >>
> >> Use per-CPU struct for flush_tlb_mm_range() and
> >> flush_tlb_kernel_range(). Add debug assertions to ensure there are
> >> no nested TLB flushes that might overwrite the per-CPU data. For
> >> arch_tlbbatch_flush(), use a const struct.
> >>
> >> Results when running a microbenchmarks that performs 10^6 MADV_DONTEED
> >> operations and touching a page, in which 3 additional threads run a
> >> busy-wait loop (5 runs):
> >
> > Can you add a memset(,,,. 0, sizeof(struct flush_tlb_info)) everywhere
> > you grab it?  Or, even better, perhaps do something like:
> >
> > static inline struct flush_tlb_info *get_flush_tlb_info(void)
> > {
> >  /* check reentrancy, make sure that we use smp_processor_id() or
> > otherwise assert that we're bound to a single CPU. */
> >  struct flush_tlb_info *ptr =3D this_cpu_ptr(...);
> >  memset(ptr, 0, sizeof(*ptr));
> >  return ptr;
> > }
> >
> > static inline void put_flush_tlb_info(void)
> > {
> > /* finish checking reentrancy. */
> > }
>
> I=E2=80=99ll check if the compiler is smart enough to avoid redundant ass=
ignments,
> and if it is not, I=E2=80=99ll just give all the struct arguments to
> get_flush_tlb_info() instead of memset() if you don=E2=80=99t mind.

Sounds good.

>
> I also want to give a try for parallelizing the remote and local
> invocations, which really annoys me every time I look at the code.

Yes please!