Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Sender: Minchan Kim <minchan.kim@gmail.com>
Date:   Wed, 11 May 2022 20:44:43 -0700
From:   Minchan Kim <minchan@kernel.org>
To:     John Hubbard <jhubbard@nvidia.com>
Cc:     "Paul E. McKenney" <paulmck@kernel.org>,
        Andrew Morton <akpm@linux-foundation.org>,
        linux-mm <linux-mm@kvack.org>,
        LKML <linux-kernel@vger.kernel.org>,
        John Dias <joaodias@google.com>,
        David Hildenbrand <david@redhat.com>
Subject: Re: [PATCH v4] mm: fix is_pinnable_page against on cma page
Message-ID: <YnyCqwY4DxcZ/NjM@google.com>
References: <20220511234534.GG1790663@paulmck-ThinkPad-P17-Gen-1>
 <ea64b36c-ecc1-74db-dd2e-909e7e507ef8@nvidia.com>
 <0d90390c-3624-4f93-f8bd-fb29e92237d3@nvidia.com>
 <20220512002207.GJ1790663@paulmck-ThinkPad-P17-Gen-1>
 <YnxUTxnCJ6EsmjEi@google.com>
 <f7ff4186-e629-a5da-3f4d-ec629b0c8dd9@nvidia.com>
 <20220512004949.GK1790663@paulmck-ThinkPad-P17-Gen-1>
 <YnxczoehQJ+x6m9Y@google.com>
 <0accce46-fac6-cdfb-db7f-d08396bf9d35@nvidia.com>
 <aafc0954-80df-dceb-03f2-2caea29cbbda@nvidia.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <aafc0954-80df-dceb-03f2-2caea29cbbda@nvidia.com>
Precedence: bulk

On Wed, May 11, 2022 at 07:18:56PM -0700, John Hubbard wrote:
> On 5/11/22 18:08, John Hubbard wrote:
> > On 5/11/22 18:03, Minchan Kim wrote:
> > > > 
> > > > Or there might be some code path that really hates a READ_ONCE() in
> > > > that place.
> > > 
> > > My worry about chaning __get_pfnblock_flags_mask is it's called
> > > multiple hot places in mm codes so I didn't want to add overhead
> > > to them.
> > 
> > ...unless it really does generate the same code as is already there,
> > right? Let me check that real quick.
> > 
> 
> It does change the generated code slightly. I don't know if this will
> affect performance here or not. But just for completeness, here you go:
> 
> free_one_page() originally has this (just showing the changed parts):
> 
>     mov    0x8(%rdx,%rax,8),%rbx
>     and    $0x3f,%ecx
>     shr    %cl,%rbx
>     and    $0x7,
> 
> 
> And after applying this diff:
> 
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index 0e42038382c1..df1f8e9a294f 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -482,7 +482,7 @@ unsigned long __get_pfnblock_flags_mask(const struct
> page *page,
>         word_bitidx = bitidx / BITS_PER_LONG;
>         bitidx &= (BITS_PER_LONG-1);
> 
> -       word = bitmap[word_bitidx];
> +       word = READ_ONCE(bitmap[word_bitidx]);
>         return (word >> bitidx) & mask;
>  }
> 
> 
> ...it now does an extra memory dereference:
> 
>     lea    0x8(%rdx,%rax,8),%rax
>     and    $0x3f,%ecx
>     mov    (%rax),%rbx
>     shr    %cl,%rbx
>     and    $0x7,%ebx
> 

Thanks for checking, John.

I don't want to have the READ_ONCE in __get_pfnblock_flags_mask 
atm even though it's an extra memory dereference for specific
architecutre and specific compiler unless other callsites *do*
need it.

We choose the choice(i.e., having __READ_ONCE in is_pinanble_
page) for *potential* cases(e.g., aggressive LTO for future)
and if it's an extra memory dereference atm, it would be multiple
instructions *potentailly* as having more change or different
compiler and/or arches now or later. It's out of scope in
this patch.