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.1627468308.git.robin.murphy@arm.com> <CAD=FV=U5a=aj5NPOGywVoEGQLuza9T6RhBhw9NT2AT2qxaKzYw@mail.gmail.com>
In-Reply-To: <CAD=FV=U5a=aj5NPOGywVoEGQLuza9T6RhBhw9NT2AT2qxaKzYw@mail.gmail.com>
From:   Doug Anderson <dianders@chromium.org>
Date:   Thu, 29 Jul 2021 17:06:59 -0700
Message-ID: <CAD=FV=Xek6_Qp7U=PrvGgVTQ9BqDwN34QVKAhgrMVCh=PjvBqg@mail.gmail.com>
Subject: Re: [PATCH v2 00/24] iommu: Refactor DMA domain strictness
To:     Robin Murphy <robin.murphy@arm.com>
Cc:     Joerg Roedel <joro@8bytes.org>, Will Deacon <will@kernel.org>,
        "list@263.net:IOMMU DRIVERS <iommu@lists.linux-foundation.org>, Joerg
        Roedel <joro@8bytes.org>," <iommu@lists.linux-foundation.org>,
        Linux ARM <linux-arm-kernel@lists.infradead.org>,
        LKML <linux-kernel@vger.kernel.org>,
        suravee.suthikulpanit@amd.com, Lu Baolu <baolu.lu@linux.intel.com>,
        John Garry <john.garry@huawei.com>,
        Marek Szyprowski <m.szyprowski@samsung.com>,
        Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>,
        Geert Uytterhoeven <geert+renesas@glider.be>,
        Yong Wu <yong.wu@mediatek.com>,
        Heiko Stuebner <heiko@sntech.de>,
        Chunyan Zhang <chunyan.zhang@unisoc.com>,
        Maxime Ripard <mripard@kernel.org>,
        Jean-Philippe Brucker <jean-philippe@linaro.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

Hi,

On Thu, Jul 29, 2021 at 3:33 PM Doug Anderson <dianders@chromium.org> wrote:
>
> I was definitely getting some inconsistencies in my tests where the
> eMMC speeds were getting into a bad state, but I don't believe it's
> related to your patch series.

I think this was just me being an idiot. I forgot that I'd been
running with KASAN, so that explains why my speeds were so much slower
than usual and probably also explains how it could get in a bad state
(I guess it also explains why sugov was eating up 30% of my CPU time
since that went away too!). No mystery here aside from why it took me
this long to realize it.

I'm now getting ~213 MB/s without forcing it to lazy and ~261 MB/s
with forcing it to lazy through sysfs (and without any other cpufreq
hacks).

-Doug