From patchwork Wed Dec 14 09:03:24 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: =?utf-8?q?J=C3=B6rg_Krause?= X-Patchwork-Id: 87991 Delivered-To: patch@linaro.org Received: by 10.140.20.101 with SMTP id 92csp88213qgi; Wed, 14 Dec 2016 01:05:31 -0800 (PST) X-Received: by 10.84.142.131 with SMTP id 3mr204201181plx.124.1481706331006; Wed, 14 Dec 2016 01:05:31 -0800 (PST) Return-Path: Received: from bombadil.infradead.org (bombadil.infradead.org. [2001:1868:205::9]) by mx.google.com with ESMTPS id s136si51777216pgs.329.2016.12.14.01.05.30 for (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Wed, 14 Dec 2016 01:05:30 -0800 (PST) Received-SPF: pass (google.com: best guess record for domain of linux-arm-kernel-bounces+patch=linaro.org@lists.infradead.org designates 2001:1868:205::9 as permitted sender) client-ip=2001:1868:205::9; Authentication-Results: mx.google.com; dkim=neutral (body hash did not verify) header.i=@embedded.rocks; spf=pass (google.com: best guess record for domain of linux-arm-kernel-bounces+patch=linaro.org@lists.infradead.org designates 2001:1868:205::9 as permitted sender) smtp.mailfrom=linux-arm-kernel-bounces+patch=linaro.org@lists.infradead.org Received: from localhost ([127.0.0.1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.85_2 #1 (Red Hat Linux)) id 1cH5U0-0006Sg-Au; Wed, 14 Dec 2016 09:04:16 +0000 Received: from mout02.posteo.de ([185.67.36.142]) by bombadil.infradead.org with esmtps (Exim 4.85_2 #1 (Red Hat Linux)) id 1cH5Ts-0006Fu-IJ for linux-arm-kernel@lists.infradead.org; Wed, 14 Dec 2016 09:04:11 +0000 Received: from customer (localhost [127.0.0.1]) by submission (posteo.de) with ESMTPSA id 3tdrGJ1b2rz106f; Wed, 14 Dec 2016 10:03:36 +0100 (CET) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=embedded.rocks; h=mime-version:x-mailer:content-type:content-type:references :in-reply-to:date:date:from:from:subject:subject:message-id :received:received; s=default; t=1481706214; x=1482311015; bh=wN uY/hqQYuiyx9gJzNzzvAkgYTv3auU53K+oy8XlOJY=; b=2HKhZ7zirS1CtR9/Ak 76F4WIEypOiwZ1TGgc9m82ua7fVtlSuqyJApyJtuiiM74yoXeYqtcsC/HmAesGj6 XPzNNRXjtLEq4e7TwFEuCVan/qAx9dBCNKGhPXpU/CYlyyk3p4IxbRP7ysHvVJ60 KY90fbrcRBXeiJ226XydF913DpuR/tAfJl0gWeodqdfxmSZpgEMRNaO0iV0/S56T RF46oQXUkB7k2mytwlXINTUY9tYDQABvxRAO+vax9lOxFBrSS7fKxLvgURpxDWL/ xZkWYIhTUnZleIIdGrg+MtPVqPzBKSDB3FMM0udCbeY+GavcXhExvXzBl/8jqLmL 8tKw== Received: from mail.embedded.rocks ([127.0.0.1]) by localhost (mail.embedded.rocks [127.0.0.1]) (amavisd-new, port 10025) with ESMTP id mA6P2hEuPe8X; Wed, 14 Dec 2016 10:03:34 +0100 (CET) Received: from hp (ip5b40bc83.dynamic.kabel-deutschland.de [91.64.188.131]) (Authenticated sender: joerg.krause@embedded.rocks) by mail.embedded.rocks (Postfix) with ESMTPSA; Wed, 14 Dec 2016 10:03:34 +0100 (CET) Message-ID: <1481706204.3994.4.camel@embedded.rocks> Subject: Re: mmc: core: complete/wait_for_completion performance From: =?ISO-8859-1?Q?J=F6rg?= Krause To: Stefan Wahren Date: Wed, 14 Dec 2016 10:03:24 +0100 In-Reply-To: <123257138.400786.e58aee3b-9fc2-4b39-a030-c2409c5b92fc.open-xchange@email.1und1.de> References: <1479644869.2653.3.camel@embedded.rocks> <585759233.283839.1cb53b4d-2805-48ea-aef1-dd282306d108.open-xchange@email.1und1.de> <1479652929.2841.1.camel@embedded.rocks> <187975187.249177.bccdc17e-e9c6-48c2-aeaf-3b81f1b61ec7.open-xchange@email.1und1.de> <1479669034.1975.1.camel@embedded.rocks> <1198138554.59982.63209ba1-8fb4-4a13-9ee0-f746a192f4c7.open-xchange@email.1und1.de> <1481095953.17027.0.camel@embedded.rocks> <123257138.400786.e58aee3b-9fc2-4b39-a030-c2409c5b92fc.open-xchange@email.1und1.de> X-Mailer: Evolution 3.22.2 Mime-Version: 1.0 X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20161214_010409_022324_AD7580B7 X-CRM114-Status: GOOD ( 35.59 ) X-Spam-Score: -2.5 (--) X-Spam-Report: SpamAssassin version 3.4.0 on bombadil.infradead.org summary: Content analysis details: (-2.5 points) pts rule name description ---- ---------------------- -------------------------------------------------- -0.7 RCVD_IN_DNSWL_LOW RBL: Sender listed at http://www.dnswl.org/, low trust [185.67.36.142 listed in list.dnswl.org] -1.9 BAYES_00 BODY: Bayes spam probability is 0 to 1% [score: 0.0000] 0.1 DKIM_SIGNED Message has a DKIM or DK signature, not necessarily valid 0.0 T_DKIM_INVALID DKIM-Signature header exists but is not valid X-BeenThere: linux-arm-kernel@lists.infradead.org X-Mailman-Version: 2.1.20 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Russell King , Ulf Hansson , linux-arm-kernel@lists.infradead.org Sender: "linux-arm-kernel" Errors-To: linux-arm-kernel-bounces+patch=linaro.org@lists.infradead.org Hi Stefan, On Wed, 2016-12-07 at 20:23 +0100, Stefan Wahren wrote: > Hi Jörg, > > > Jörg Krause hat am 7. Dezember 2016 > > um 08:32 > > geschrieben: > > > > > > Hit Stefan, > > > > On Sat, 2016-11-26 at 20:10 +0100, Stefan Wahren wrote: > > > Hi Jörg, > > > > > > > Jörg Krause hat am 20. November > > > > 2016 > > > > um 20:10 > > > > geschrieben: > > > > > > > > > > > > On Sun, 2016-11-20 at 16:44 +0100, Stefan Wahren wrote: > > > > > > Jörg Krause hat am 20. > > > > > > November > > > > > > 2016 > > > > > > um 15:42 > > > > > > geschrieben: > > > > > > > > > > > > > > > > > > Hi Stefan, > > > > > > > > > > > > On Sun, 2016-11-20 at 14:28 +0100, Stefan Wahren wrote: > > > > > > > Hi Jörg, > > > > > > > > > > > > > > > Jörg Krause hat am 20. > > > > > > > > November > > > > > > > > 2016 > > > > > > > > um 13:27 > > > > > > > > geschrieben: > > > > > > > > > > > > > > > > > > > > > > > > Hi, > > > > > > > > > > > > > > > > I started the discussion on this mailing list in > > > > > > > > another > > > > > > > > thread > > > > > > > > [1], > > > > > > > > but I'd like to move it to a new thread, because it > > > > > > > > might > > > > > > > > be > > > > > > > > mmc > > > > > > > > specific. > > > > > > > > > > > > > > > > The issue is that I am noticed low wifi network > > > > > > > > throughput > > > > > > > > on > > > > > > > > an > > > > > > > > i.MX28 > > > > > > > > board with the mainline kernel (v4.7.10, about 6 Mbps) > > > > > > > > compared > > > > > > > > to > > > > > > > > the > > > > > > > > vendor kernel (Freescale v2.6.35.3, about 20 Mbps). The > > > > > > > > wifi > > > > > > > > chip > > > > > > > > is > > > > > > > > attached using the SDIO interface. > > > > > > > > > > > > > > > > I started investigation where the bottleneck in the > > > > > > > > mainline > > > > > > > > kernel might come from. Therefore I checked that the > > > > > > > > configs > > > > > > > > and > > > > > > > > settings for the interfaces and drivers are the same. > > > > > > > > They > > > > > > > > are. > > > > > > > > > > > > > > so you're not using the mxs_defconfig settings anymore? > > > > > > > > > > > > No, I changed the settings. > > > > > > > > > > > > > > > > What happens to performance to if you change the following > > > > > settings > > > > > to the same > > > > > like in mxs_defconfig? > > > > > > > > > > CONFIG_PREEMPT_VOLUNTARY=y > > > > > CONFIG_DEFAULT_IOSCHED="noop" > > > > > > > > No much change at all. The time difference between complete() > > > > and > > > > wait_for_complete() decreases in best case to 110 us, but also > > > > varies > > > > to above 130 us. > > > > > > just a weird idea. Did you tried to add MMC_CAP_CMD23 into the > > > caps > > > [1]? > > > > > > [1] - http://lxr.free-electrons.com/source/drivers/mmc/host/mxs-m > > > mc.c > > > ?v=4.8#L642 > > > > I tried, but it did not improved the timing or throughput. However, > > many thanks for the input. > > > > Jörg > > did you try cyclictest [1]? Not yet. Not sure what to measure and which values to compare here. > > Beside the time for a request the amount of requests for the complete > iperf test > would we interesting. Maybe there are retries. > > I'm still interested in your PIO mode patches for mxs-mmc even > without clean up. Actually, the patch does not implement a PIO mode, but drops DMA and uses polling instead. I've attached the patch. > [1] - https://git.kernel.org/cgit/utils/rt-tests/rt-tests.git/ Best regards Jörg _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel >From 816b9a4fc08280ca971fa60825e9bb0ceb42ef0b Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?J=C3=B6rg=20Krause?= Date: Mon, 31 Oct 2016 21:27:11 +0100 Subject: [PATCH] mmc: mxs-mmc: drop DMA and use polling mode MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Jörg Krause --- drivers/mmc/host/mxs-mmc.c | 141 ++++++++++++++++++++++++++++++++++++++++++-- include/linux/spi/mxs-spi.h | 11 ++++ 2 files changed, 148 insertions(+), 4 deletions(-) diff --git a/drivers/mmc/host/mxs-mmc.c b/drivers/mmc/host/mxs-mmc.c index 44ecebd..c4f5ff0 100644 --- a/drivers/mmc/host/mxs-mmc.c +++ b/drivers/mmc/host/mxs-mmc.c @@ -71,6 +71,9 @@ struct mxs_mmc_host { spinlock_t lock; int sdio_irq_en; bool broken_cd; + + struct completion dma_done; + u32 status; }; static int mxs_mmc_get_cd(struct mmc_host *mmc) @@ -256,22 +259,83 @@ static struct dma_async_tx_descriptor *mxs_mmc_prep_dma( return desc; } +/* + * Check for MMC command errors + * Returns error code or zerro if no errors + */ +static inline int mxs_mmc_cmd_error(u32 status) +{ + int err = 0; + + if (status & BM_SSP_STATUS_TIMEOUT) + err = -ETIMEDOUT; + else if (status & BM_SSP_STATUS_RESP_TIMEOUT) + err = -ETIMEDOUT; + else if (status & BM_SSP_STATUS_RESP_CRC_ERR) + err = -EILSEQ; + else if (status & BM_SSP_STATUS_RESP_ERR) + err = -EIO; + + return err; +} + static void mxs_mmc_bc(struct mxs_mmc_host *host) { struct mxs_ssp *ssp = &host->ssp; struct mmc_command *cmd = host->cmd; struct dma_async_tx_descriptor *desc; - u32 ctrl0, cmd0, cmd1; + u32 ctrl0, ctrl1, cmd0, cmd1, c1; ctrl0 = BM_SSP_CTRL0_ENABLE | BM_SSP_CTRL0_IGNORE_CRC; + ctrl1 = + BM_SSP_CTRL1_DMA_ENABLE | + BM_SSP_CTRL1_RECV_TIMEOUT_IRQ_EN | + BM_SSP_CTRL1_DATA_CRC_IRQ_EN | + BM_SSP_CTRL1_DATA_TIMEOUT_IRQ_EN | + BM_SSP_CTRL1_RESP_TIMEOUT_IRQ_EN | + BM_SSP_CTRL1_RESP_ERR_IRQ_EN; cmd0 = BF_SSP(cmd->opcode, CMD0_CMD) | BM_SSP_CMD0_APPEND_8CYC; - cmd1 = cmd->arg; + cmd1 = BF_SSP(cmd->arg, CMD1_CMD); if (host->sdio_irq_en) { ctrl0 |= BM_SSP_CTRL0_SDIO_IRQ_CHECK; cmd0 |= BM_SSP_CMD0_CONT_CLKING_EN | BM_SSP_CMD0_SLOW_CLKING_EN; } + /* following IO operations */ + writel(ctrl0, ssp->base + HW_SSP_CTRL0); + writel(cmd0, ssp->base + HW_SSP_CMD0); + writel(cmd1, ssp->base + HW_SSP_CMD1); + + /* clear these bits */ + writel(ctrl1, ssp->base + HW_SSP_CTRL1(ssp) + STMP_OFFSET_REG_CLR); + + init_completion(&host->dma_done); + + writel(BM_SSP_CTRL0_RUN, + ssp->base + HW_SSP_CTRL0 + STMP_OFFSET_REG_SET); + + while (readl(ssp->base + HW_SSP_CTRL0) & BM_SSP_CTRL0_RUN) + continue; + while (readl(ssp->base + HW_SSP_STATUS(ssp)) & BM_SSP_STATUS_BUSY) + continue; + + host->status = readl(ssp->base + HW_SSP_STATUS(ssp)); + c1 = readl(ssp->base + HW_SSP_CTRL1(ssp)); + writel(c1 & MXS_MMC_IRQ_BITS, + ssp->base + HW_SSP_CTRL1(ssp) + STMP_OFFSET_REG_CLR); + + /* reenable these bits */ + writel(ctrl1, ssp->base + HW_SSP_CTRL1(ssp) + STMP_OFFSET_REG_SET); + /* end IO operations */ + + cmd->error = mxs_mmc_cmd_error(host->status); + + if (cmd->error) { + dev_warn(mmc_dev(host->mmc), "BC command error 0x%x\n", cmd->error); + } + +#if 0 ssp->ssp_pio_words[0] = ctrl0; ssp->ssp_pio_words[1] = cmd0; ssp->ssp_pio_words[2] = cmd1; @@ -288,6 +352,7 @@ static void mxs_mmc_bc(struct mxs_mmc_host *host) out: dev_warn(mmc_dev(host->mmc), "%s: failed to prep dma\n", __func__); +#endif } static void mxs_mmc_ac(struct mxs_mmc_host *host) @@ -296,7 +361,7 @@ static void mxs_mmc_ac(struct mxs_mmc_host *host) struct mmc_command *cmd = host->cmd; struct dma_async_tx_descriptor *desc; u32 ignore_crc, get_resp, long_resp; - u32 ctrl0, cmd0, cmd1; + u32 ctrl0, ctrl1, cmd0, cmd1, c1; ignore_crc = (mmc_resp_type(cmd) & MMC_RSP_CRC) ? 0 : BM_SSP_CTRL0_IGNORE_CRC; @@ -306,14 +371,81 @@ static void mxs_mmc_ac(struct mxs_mmc_host *host) BM_SSP_CTRL0_LONG_RESP : 0; ctrl0 = BM_SSP_CTRL0_ENABLE | ignore_crc | get_resp | long_resp; + ctrl1 = BM_SSP_CTRL1_DMA_ENABLE | + BM_SSP_CTRL1_RECV_TIMEOUT_IRQ_EN | + BM_SSP_CTRL1_DATA_CRC_IRQ_EN | + BM_SSP_CTRL1_DATA_TIMEOUT_IRQ_EN | + BM_SSP_CTRL1_RESP_TIMEOUT_IRQ_EN | + BM_SSP_CTRL1_RESP_ERR_IRQ_EN; cmd0 = BF_SSP(cmd->opcode, CMD0_CMD); - cmd1 = cmd->arg; + cmd1 = BF_SSP(cmd->arg, CMD1_CMD); if (host->sdio_irq_en) { ctrl0 |= BM_SSP_CTRL0_SDIO_IRQ_CHECK; cmd0 |= BM_SSP_CMD0_CONT_CLKING_EN | BM_SSP_CMD0_SLOW_CLKING_EN; } + /* following IO operations */ + writel(ctrl0, ssp->base + HW_SSP_CTRL0); + writel(cmd0, ssp->base + HW_SSP_CMD0); + writel(cmd1, ssp->base + HW_SSP_CMD1); + + /* clear these bits */ + writel(ctrl1, ssp->base + HW_SSP_CTRL1(ssp) + STMP_OFFSET_REG_CLR); + + init_completion(&host->dma_done); + + writel(BM_SSP_CTRL0_RUN, + ssp->base + HW_SSP_CTRL0 + STMP_OFFSET_REG_SET); + + while (readl(ssp->base + HW_SSP_CTRL0) & BM_SSP_CTRL0_RUN) + continue; + while (readl(ssp->base + HW_SSP_STATUS(ssp)) & BM_SSP_STATUS_BUSY) + continue; + + host->status = readl(ssp->base + HW_SSP_STATUS(ssp)); + c1 = readl(ssp->base + HW_SSP_CTRL1(ssp)); + writel(c1 & MXS_MMC_IRQ_BITS, + ssp->base + HW_SSP_CTRL1(ssp) + STMP_OFFSET_REG_CLR); + + /* reenable these bits */ + writel(ctrl1, ssp->base + HW_SSP_CTRL1(ssp) + STMP_OFFSET_REG_SET); + /* end IO operations */ + + switch (mmc_resp_type(cmd)) { + case MMC_RSP_NONE: + while (readl(ssp->base + HW_SSP_CTRL0) & BM_SSP_CTRL0_RUN) + continue; + break; + case MMC_RSP_R1: + case MMC_RSP_R1B: + case MMC_RSP_R3: + cmd->resp[0] = readl(ssp->base + HW_SSP_SDRESP0(ssp)); + break; + case MMC_RSP_R2: + cmd->resp[3] = + readl(ssp->base + HW_SSP_SDRESP0(ssp)); + cmd->resp[2] = + readl(ssp->base + HW_SSP_SDRESP1(ssp)); + cmd->resp[1] = + readl(ssp->base + HW_SSP_SDRESP2(ssp)); + cmd->resp[0] = + readl(ssp->base + HW_SSP_SDRESP3(ssp)); + break; + default: + dev_warn(mmc_dev(host->mmc), "Unsupported response type 0x%x\n", + mmc_resp_type(cmd)); + BUG(); + break; + } + + cmd->error = mxs_mmc_cmd_error(host->status); + + if (cmd->error) { + dev_warn(mmc_dev(host->mmc), "AC command error 0x%x\n", cmd->error); + } + +#if 0 ssp->ssp_pio_words[0] = ctrl0; ssp->ssp_pio_words[1] = cmd0; ssp->ssp_pio_words[2] = cmd1; @@ -330,6 +462,7 @@ static void mxs_mmc_ac(struct mxs_mmc_host *host) out: dev_warn(mmc_dev(host->mmc), "%s: failed to prep dma\n", __func__); +#endif } static unsigned short mxs_ns_to_ssp_ticks(unsigned clock_rate, unsigned ns) diff --git a/include/linux/spi/mxs-spi.h b/include/linux/spi/mxs-spi.h index 381d368..83cb802 100644 --- a/include/linux/spi/mxs-spi.h +++ b/include/linux/spi/mxs-spi.h @@ -53,6 +53,10 @@ #define BP_SSP_CMD0_CMD 0 #define BM_SSP_CMD0_CMD 0xff #define HW_SSP_CMD1 0x020 + +#define BP_SSP_CMD1_CMD 0 +#define BM_SSP_CMD1_CMD 0xFFFFFFFF + #define HW_SSP_XFER_SIZE 0x030 #define HW_SSP_BLOCK_SIZE 0x040 #define BP_SSP_BLOCK_SIZE_BLOCK_COUNT 4 @@ -117,6 +121,13 @@ #define BM_SSP_STATUS_SDIO_IRQ (1 << 17) #define BM_SSP_STATUS_FIFO_EMPTY (1 << 5) +#define BM_SSP_STATUS_RESP_CRC_ERR 0x00010000 +#define BM_SSP_STATUS_RESP_ERR 0x00008000 +#define BM_SSP_STATUS_RESP_TIMEOUT 0x00004000 +#define BM_SSP_STATUS_DATA_CRC_ERR 0x00002000 +#define BM_SSP_STATUS_TIMEOUT 0x00001000 +#define BM_SSP_STATUS_BUSY 0x00000001 + #define BF_SSP(value, field) (((value) << BP_SSP_##field) & BM_SSP_##field) #define SSP_PIO_NUM 3 -- 2.10.2