Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <cover.1591885760.git.afzal.mohd.ma@gmail.com> <9e1de19f35e2d5e1d115c9ec3b7c3284b4a4e077.1591885760.git.afzal.mohd.ma@gmail.com>
 <CAK8P3a1XUJHC0kG_Qwh4D4AoxTgCL5ggHd=45yNSmzaYWLUWXw@mail.gmail.com> <20200612135538.GA13399@afzalpc>
In-Reply-To: <20200612135538.GA13399@afzalpc>
From:   Arnd Bergmann <arnd@arndb.de>
Date:   Fri, 12 Jun 2020 22:07:28 +0200
Message-ID: <CAK8P3a25ffh_2Y1xKDbkL2xU9nLpGbEq7j6xHdODEwUtavgdwA@mail.gmail.com>
Subject: Re: [RFC 1/3] lib: copy_{from,to}_user using gup & kmap_atomic()
To:     afzal mohammed <afzal.mohd.ma@gmail.com>
Cc:     Russell King - ARM Linux admin <linux@armlinux.org.uk>,
        Linus Walleij <linus.walleij@linaro.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        Linux-MM <linux-mm@kvack.org>,
        Linux ARM <linux-arm-kernel@lists.infradead.org>,
        Nicolas Pitre <nico@fluxnic.net>,
        Catalin Marinas <catalin.marinas@arm.com>,
        Will Deacon <will@kernel.org>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Fri, Jun 12, 2020 at 3:55 PM afzal mohammed <afzal.mohd.ma@gmail.com> wrote:
> On Fri, Jun 12, 2020 at 02:02:13PM +0200, Arnd Bergmann wrote:
> > On Fri, Jun 12, 2020 at 12:18 PM afzal mohammed <afzal.mohd.ma@gmail.com> wrote:
>
> > > Roughly a one-third drop in performance. Disabling highmem improves
> > > performance only slightly.
>
> > There are probably some things that can be done to optimize it,
> > but I guess most of the overhead is from the page table operations
> > and cannot be avoided.
>
> Ingo's series did a follow_page() first, then as a fallback did it
> invoke get_user_pages(), i will try that way as well.

Right, that could help, in particular for the small copies. I think a lot
of usercopy calls are only for a few bytes, though this is of course
highly workload dependent and you might only care about the large
ones.

> Yes, i too feel get_user_pages_fast() path is the most time consuming,
> will instrument & check.
>
> > What was the exact 'dd' command you used, in particular the block size?
> > Note that by default, 'dd' will request 512 bytes at a time, so you usually
> > only access a single page. It would be interesting to see the overhead with
> > other typical or extreme block sizes, e.g. '1', '64', '4K', '64K' or '1M'.
>
> It was the default(512), more test results follows (in MB/s),
>
>                 512     1K      4K      16K     32K     64K     1M
>
> w/o series      30      46      89      95      90      85      65
>
> w/ series       22      36      72      79      78      75      61
>
> perf drop       26%     21%     19%     16%     13%     12%    6%
>
> Hmm, results ain't that bad :)

There is also still hope of optimizing small aligned copies like

set_ttbr0(user_ttbr);
ldm();
set_ttbr0(kernel_ttbr);
stm();

which could do e.g. 32 bytes at a time, but with more overhead
if you have to loop around it.

        Arnd