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;
MIME-Version: 1.0
References: <20220419211658.11403-1-apais@linux.microsoft.com>
 <20220419211658.11403-2-apais@linux.microsoft.com> <CACRpkdZ2DFZRPHS1x0=M3_8zYvU-jpCG5Tm3863dXv51EhY+BA@mail.gmail.com>
 <CAK8P3a0j_rziihsgHnG5bHMxmPbOkAhT6_+CCE4iFZy7HzQrLw@mail.gmail.com>
 <YpCGePbo9B/Z7slV@matsya> <CAK8P3a2wD7=hgvqyS14X5p-eP+7Ajk4dFJOXgbOo8Z0r5UNYmg@mail.gmail.com>
 <YpW8J40hKwc7jwQh@matsya> <0A9EDEDC-9E6C-47F8-89C0-48DCDD3F9DE3@linux.microsoft.com>
 <CAK8P3a3vf_huJ5ysBvFDV=11E-=OxFGiDQ9ND04YKug8g6jV_A@mail.gmail.com> <708F627F-0F7D-477F-BDF7-274424501DA0@linux.microsoft.com>
In-Reply-To: <708F627F-0F7D-477F-BDF7-274424501DA0@linux.microsoft.com>
From:   Arnd Bergmann <arnd@arndb.de>
Date:   Tue, 31 May 2022 21:27:35 +0200
Message-ID: <CAK8P3a314SxhTNy6psYr_MzarOPmkjv1_KKGdeVUPpBjVkC5nA@mail.gmail.com>
Subject: Re: [RFC 1/1] drivers/dma/*: replace tasklets with workqueue
To:     Allen Pais <apais@linux.microsoft.com>
Cc:     Arnd Bergmann <arnd@arndb.de>, Vinod Koul <vkoul@kernel.org>,
        Linus Walleij <linus.walleij@linaro.org>,
        Vincent Guittot <vincent.guittot@linaro.org>,
        olivier.dautricourt@orolia.com, Stefan Roese <sr@denx.de>,
        Kees Cook <keescook@chromium.org>,
        linux-hardening@vger.kernel.org,
        Ludovic Desroches <ludovic.desroches@microchip.com>,
        Tudor Ambarus <tudor.ambarus@microchip.com>,
        Florian Fainelli <f.fainelli@gmail.com>,
        Ray Jui <rjui@broadcom.com>,
        Scott Branden <sbranden@broadcom.com>,
        bcm-kernel-feedback-list <bcm-kernel-feedback-list@broadcom.com>,
        Nicolas Saenz Julienne <nsaenz@kernel.org>,
        Paul Cercueil <paul@crapouillou.net>,
        Eugeniy.Paltsev@synopsys.com,
        Gustavo Pimentel <gustavo.pimentel@synopsys.com>,
        Viresh Kumar <vireshk@kernel.org>,
        Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
        Leo Li <leoyang.li@nxp.com>, zw@zh-kernel.org,
        Zhou Wang <wangzhou1@hisilicon.com>,
        Shawn Guo <shawnguo@kernel.org>,
        Sascha Hauer <s.hauer@pengutronix.de>,
        Sean Wang <sean.wang@mediatek.com>,
        Matthias Brugger <matthias.bgg@gmail.com>,
        =?UTF-8?Q?Andreas_F=C3=A4rber?= <afaerber@suse.de>,
        Manivannan Sadhasivam <mani@kernel.org>,
        Logan Gunthorpe <logang@deltatee.com>,
        Sanjay R Mehta <sanju.mehta@amd.com>,
        Daniel Mack <daniel@zonque.org>,
        Haojian Zhuang <haojian.zhuang@gmail.com>,
        Robert Jarzmik <robert.jarzmik@free.fr>,
        Andy Gross <agross@kernel.org>,
        Bjorn Andersson <bjorn.andersson@linaro.org>,
        Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>,
        green.wan@sifive.com, Orson Zhai <orsonzhai@gmail.com>,
        Baolin Wang <baolin.wang7@gmail.com>,
        Lyra Zhang <zhang.lyra@gmail.com>,
        Patrice CHOTARD <patrice.chotard@foss.st.com>,
        Chen-Yu Tsai <wens@csie.org>,
        =?UTF-8?Q?Jernej_=C5=A0krabec?= <jernej.skrabec@gmail.com>,
        Samuel Holland <samuel@sholland.org>,
        dmaengine@vger.kernel.org,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Tue, May 31, 2022 at 8:19 PM Allen Pais <apais@linux.microsoft.com> wrote:
> >>> That is a good idea, lot of drivers are waiting for completion which can
> >>> be signalled from hardirq, this would also reduce the hops we have and
> >>> help improve latency a bit. On the downside, some controllers provide
> >>> error information, which would need to be dealt with.
> >>
> >>
> >>   I am not an expert in dma subsystem, but by using completion from
> >> Hardirq context be a concern? Especially with latency.
> >
> > I don't see how: to the task waiting for the completion, there should
> > be no difference, and for the irq handler sending it, it just avoids
> > a few cycles going into softirq context.
>
>   Thanks for clarification.
>
>   If I have understood it correctly, your suggestion is to move the current
> Callback mechanism out to dmaengine as a generic helper function
> And introduce completion in dma_async_tx_descriptor to handle what
> Tasklets currently do.

Right: around half the callbacks are a trivial version that does nothing
other than calling complete(), so these will benefit from skipping the
softirq, while the others still need the callback.

       Arnd