From nobody Tue Dec 16 21:28:44 2025 Received: from OSPPR02CU001.outbound.protection.outlook.com (mail-norwayeastazon11013069.outbound.protection.outlook.com [40.107.159.69]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 3CE972D6E70; Wed, 3 Dec 2025 09:00:48 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=fail smtp.client-ip=40.107.159.69 ARC-Seal: i=2; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1764752450; cv=fail; b=b8SZ0WgsSKNnx8nJWfdTobrNPBX3XkonokMVwG2tzS3Gkci3l5bZp/a2c5bmwxgbdRPG49c8Ab84pydYkz+3SYUbaUjzuqzoAtHJsD5k95UEb+odkB5uji7boGTz/ZY4YwuQJIl4ob1PfewvFciIzDX+06bMzLoUiYRHfPY0D8g= ARC-Message-Signature: i=2; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1764752450; c=relaxed/simple; bh=EborCir4qam+eEPdlgjnryTrttyY6rs698YXrBjWPhI=; h=From:To:Cc:Subject:Date:Message-Id:In-Reply-To:References: Content-Type:MIME-Version; b=Q2gdmv+iSkGLe69DLFyCCOoJJDr2ykyVoEviqZ6NjeWbdqS/R1DKx+eWfbzPRUtpaYgapzY4IKVpC1XWdfIYHPAEk6VEDe//FxbExQwgyVJ3XC0sGj8F6IBrCbozQAciAmQNIJYXYIcOW2Qu+6EBQOht8f9U8ntibnid0NQGC+s= ARC-Authentication-Results: i=2; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=nxp.com; spf=pass smtp.mailfrom=nxp.com; dkim=pass (2048-bit key) header.d=nxp.com header.i=@nxp.com header.b=d3IwLcUa; arc=fail smtp.client-ip=40.107.159.69 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=nxp.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=nxp.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=nxp.com header.i=@nxp.com header.b="d3IwLcUa" ARC-Seal: i=1; a=rsa-sha256; s=arcselector10001; d=microsoft.com; cv=none; b=KsRHCKVwDs5RWJTskc01hPZNOm5rtQ43BT/gCyUJo8hxmzuTesU54DDGufV/3FZ/waErSZeGhyvEsU0uXNZT2Ag3ZLi+ILerwlZBrlynSHw4nCLPOZSuvO/8Kv824pZbUAaY3w9kk8hVzRl4NnmJ/vJmKX/B5pmXI24iuKvJFFTTZISppBvjfMQNu03oR8IumnBUCJOwmXAhs9054T+Y8gEGDEQOCAClBHhmlXgt5wMeMfwEqrGF8w4UntrvsHndedn4Z7kQIq04k8XnP3vYWD5yuDIx3OZ7q4UNORMJGaQK+x5OHTQfFHrEbDOBRJuy+ly1NL6Ptdr0kuoovNDTLg== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=microsoft.com; s=arcselector10001; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-AntiSpam-MessageData-ChunkCount:X-MS-Exchange-AntiSpam-MessageData-0:X-MS-Exchange-AntiSpam-MessageData-1; bh=QVkeB7iAyOpsyZzok3gHzZf6pNc/u9xhlnAI6NFRhJU=; b=copRFGO8ADtFKsmqxktusDIv8rHcs0z4GB+eK6dyKu8Vrlxgk89c6D5Cf+tFjsnDRj4qd2KY/36mIx6Q6Uv+7mirMDTWbjKg45lGDZExvex4vZZ1r+goDJ4sY94l2VgKOnB534gFUzzVnFirsoHQoMuTe4hSVrmygpjWWvztDENgOC7KXJhQ1hMdg/nEs3xykCE4e+De2J9F2rb45Ty5Y+p6BEfM8c+WftX5uq5jCsVZxBuF7+vLIEX8qWoUlDHVDeFYN9WvQy8kIhnwiOsSCAnCBI2qvMx3pDvf+1ZSKzEfOjclxpYKbOZoa+WSBMEasvhcQC5LU1F5MdDZbwI3Vg== ARC-Authentication-Results: i=1; mx.microsoft.com 1; spf=pass smtp.mailfrom=nxp.com; dmarc=pass action=none header.from=nxp.com; dkim=pass header.d=nxp.com; arc=none DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=nxp.com; s=selector1; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-SenderADCheck; bh=QVkeB7iAyOpsyZzok3gHzZf6pNc/u9xhlnAI6NFRhJU=; b=d3IwLcUa9LgKwMUvgPbu5A8js1QYwAUCspp2OmUunYHNLarfunpHXwoRIDDHektNEma/UQPBsg5g/9o6QCXWIc0n9t8GTqozwT82Vzdi8ZUN8FRIdQ3PkoE2EshmZ7WN1sIjJ1QZjYR8l6YK/H/smqb1me+/pf35eh7++Dkw1CHFB4aVFvJyaKDJW1F8pdwCIzqodh2k/zFESEr+6/3x37+j+lSR2eztMuiQL0kgmt/cKi4tTZxBU3kD/e5iMvhhbvHtH1F6OSnw/Bhn+s6nEhilAqSifpOmjYHZEZnUZq6SpiumZgvckZgkIwlj3xAv0imk76IGZIYWAYg8Uw/g4Q== Authentication-Results: dkim=none (message not signed) header.d=none;dmarc=none action=none header.from=nxp.com; Received: from VI2PR04MB11147.eurprd04.prod.outlook.com (2603:10a6:800:293::14) by AM0PR04MB6868.eurprd04.prod.outlook.com (2603:10a6:208:18c::22) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.9366.17; Wed, 3 Dec 2025 09:00:45 +0000 Received: from VI2PR04MB11147.eurprd04.prod.outlook.com ([fe80::75ad:fac7:cfe7:b687]) by VI2PR04MB11147.eurprd04.prod.outlook.com ([fe80::75ad:fac7:cfe7:b687%6]) with mapi id 15.20.9366.012; Wed, 3 Dec 2025 09:00:45 +0000 From: Carlos Song To: frank.li@nxp.com, mkl@pengutronix.de, broonie@kernel.org, shawnguo@kernel.org, s.hauer@pengutronix.de, kernel@pengutronix.de, festevam@gmail.com, kees@kernel.org, gustavoars@kernel.org Cc: linux-spi@vger.kernel.org, imx@lists.linux.dev, linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, linux-hardening@vger.kernel.org, Carlos Song , Frank Li Subject: [PATCH v3 5/6] spi: imx: support dynamic burst length for ECSPI DMA mode Date: Wed, 3 Dec 2025 16:59:48 +0800 Message-Id: <20251203085949.2922166-6-carlos.song@nxp.com> X-Mailer: git-send-email 2.34.1 In-Reply-To: <20251203085949.2922166-1-carlos.song@nxp.com> References: <20251203085949.2922166-1-carlos.song@nxp.com> Content-Transfer-Encoding: quoted-printable X-ClientProxiedBy: AM0P190CA0017.EURP190.PROD.OUTLOOK.COM (2603:10a6:208:190::27) To VI2PR04MB11147.eurprd04.prod.outlook.com (2603:10a6:800:293::14) Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-MS-PublicTrafficType: Email X-MS-TrafficTypeDiagnostic: VI2PR04MB11147:EE_|AM0PR04MB6868:EE_ X-MS-Office365-Filtering-Correlation-Id: 987927b4-afa8-49f8-1f14-08de324a7168 X-MS-Exchange-SenderADCheck: 1 X-MS-Exchange-AntiSpam-Relay: 0 X-Microsoft-Antispam: BCL:0;ARA:13230040|376014|7416014|52116014|19092799006|366016|1800799024|38350700014; X-Microsoft-Antispam-Message-Info: =?us-ascii?Q?RPGfSagLjFq4jhu/FrbTlCkyklXej9aXLZMQeEmkQ3hF1wJLs4NMC+cdJU4+?= =?us-ascii?Q?gq2OQFkPcEef+jjQ4N6+a5/H7vxVX3/Lqgyzt8BNoMXQvNRQ29nDDUy6QEiL?= =?us-ascii?Q?1/CagiBzG1tFCVP6omTUr+iwwon5zgmdtCSioCajHHgWsy8EsYVi3xy8YOhO?= =?us-ascii?Q?8wxmC2D/LJShXUk8MusQ2QfRa9nu1DCei8tWTWWl6NF4nQFPTgMCP/V2fuks?= =?us-ascii?Q?67/B84rPCa7RAmxElv/hkZRI49T01zk5V2FPSpxRkaW2Izxucabl7vSNo3iY?= =?us-ascii?Q?ifkDeCqDXUpiCAyNAnQ1qdsl1Wq/xrF3kJKSXMPMH5/cHBy2OOKGWp/bUADW?= =?us-ascii?Q?r2g2Mnz67vGAPrMnTYLlL7mNyjAaYRIx8VHkWB+eW3AoOYBsU/4uInWi9n44?= =?us-ascii?Q?RCS8V/BPst5rRgD0VwvsB4X0oDMGyvL13XCPJpd0HnMevKXxdMm+7ZI1Wjav?= =?us-ascii?Q?y/CO2x+xsnsVh7sxeX3c/HuxzpBlemHIDVIFgUU7kMD9HApA43EBCBX1EwPs?= =?us-ascii?Q?7IuLFUE7W2sOiOW8YTFyqJlEGaMEBlnNTITRvI6eNVxa1BkNylGem9XgCGLg?= =?us-ascii?Q?Jwjk12bNuwG2Gnzm2H6YrLs3ZgD4zj6RO0+JaHLOxaZ80aZaYTkY3XSOHLfw?= =?us-ascii?Q?CizBiIbIFoITtlzjayEeohlxM4RhhFFykT3yCdshRkY4MLzkl5zC/Te3xGeY?= =?us-ascii?Q?aYk03D4WkBcWsGRp+0A2vvEwXtUq9NRkTTIIZLCNmHVT1aq1ZejsJQ6h9wID?= =?us-ascii?Q?b6VVTNST99IBJqELR+3h20UeUDXzPxu18US0SnlhVZl7QEztHkbAZHsr/PDD?= =?us-ascii?Q?/nj/cCwZhDxn0i86b5HRSx7EJFs7TBxsZigZs6cr2S4zsxGOefYhtYH65hxy?= =?us-ascii?Q?cMq1ZRXzXdhPw4nWC6iKYGzEXINtvu29bhmKqYFRXPzk6EnKYWqoWg1U4Q1N?= =?us-ascii?Q?ojE+znME82IGBB809DG8GTR1Kukc1jBLBaJBtIPRXuooe/swr0+/yPaiRq1o?= =?us-ascii?Q?I/vddw929A+MhA43zxBgXcO79r3FxolHoZAMM2hO1TilbTVGbBLiOmC/+DLS?= =?us-ascii?Q?jhRva2G+iPYIoWjzyAIXNrRMgRsUVdhOtR/H/yTXlVXAk05TRh46J6okXa1J?= =?us-ascii?Q?Irdu85osKc5kRJmK7EWSDv1l3THS/7XcNvli47sLNdLypZCBz/7HhRztGurx?= =?us-ascii?Q?uuhR5tYfAZyPYOPExkNz+cLh9rnCow0Qh+jkExFVFLM486D0ePQVHd4IdH2V?= =?us-ascii?Q?go0e/a8j1HgYxSOXuj3+bDKL6W8Ko2SCPEnpbODXwYMi7c5yVx/NYuG1Kop0?= =?us-ascii?Q?RkbyUZSsAg+JKcCTAOgeb8ocqv5I7xNHZVZkDUVYf0yjbH4DbtWPF6+2IGNk?= =?us-ascii?Q?VnKg3ly/fHWX5w4m8KT+VlOaPP+/osgIc8q819DIx39VtNIt3aOZlq+Gucd9?= =?us-ascii?Q?NO4rsjntXKsOsWu0FUg2fI0ed19qLvlU4qJpmvNnVnP3j3Ws4yNVF4e0lXOo?= =?us-ascii?Q?TYsra57pj5sgmgFesP9A03Pzv3cAG0Piiv89?= X-Forefront-Antispam-Report: CIP:255.255.255.255;CTRY:;LANG:en;SCL:1;SRV:;IPV:NLI;SFV:NSPM;H:VI2PR04MB11147.eurprd04.prod.outlook.com;PTR:;CAT:NONE;SFS:(13230040)(376014)(7416014)(52116014)(19092799006)(366016)(1800799024)(38350700014);DIR:OUT;SFP:1101; X-MS-Exchange-AntiSpam-MessageData-ChunkCount: 1 X-MS-Exchange-AntiSpam-MessageData-0: =?us-ascii?Q?UlbJaMN5bRXALsrKXrpPjYYzhULqHaGIUfpa2HZet10vQVVf0P56XeXkoFPE?= =?us-ascii?Q?ecImTA2405ooMSBkLW5guVKN82lQcnzGHbws5W/LpOGuZSSkcnFD4VgKQBAa?= =?us-ascii?Q?W+kN8Xe9ofuCbtgLyCWRq27gxHgCFhnNvHOnS16C4fv1w7X1rbtyC0O3PH1O?= =?us-ascii?Q?/mSjG4CabHr6kRJB6BLyoUsiDE6YZ2oEGBJ82/nOGPCCC9u3c4QEhhQ1gpB+?= =?us-ascii?Q?oN6uvX4wmCCx66RMesXzRvmweZb5m55cEWlskz7vFsO1XHic+JeqRv2CgTZl?= =?us-ascii?Q?F8ENRnEmlsx2v1T6eOqRXPGQPSfRntN7EW/qXozydmt2uMQoKtCmXB/rzrAR?= =?us-ascii?Q?/M5Yd0+jX4YY4RDftm8hVCNDcuWApMnBqlGLeS2C3+lcObEgYy2FyEnCGZpC?= =?us-ascii?Q?Nx+XpqxjanJh2Oi8vCRJz8ByxqvDXixSFkoFF9R6oCfPmqPqAzuk7mafzSSN?= =?us-ascii?Q?49vV3oQZ6ZhGVbhYWilCRIDcnCmYBQdSCDRGbn3123uCh4iC6beOagrkjFiZ?= =?us-ascii?Q?td3JNvPkCty21jZrxkoNORz8YqobuOwZqr+ySS1bkjWGCdGWvKvKzBZQySDg?= =?us-ascii?Q?jlKvILDAcsQFYCTjwhgS4M1IdxRfMlq2cTHhRyByO0ujlKyUVrmXXzATCRo4?= =?us-ascii?Q?4RIFt7uN9b4oxbXEM3IK8teL/4cSC5lRxX/9k1Z0H5aE5tQtUulDKwDbJZcI?= =?us-ascii?Q?YN3dU8mNssf/inn4hOi9srXvmAxvFYl+UEp+CvToxvCKUu5ek6HOKxjO2OrB?= =?us-ascii?Q?IwgWKb7simg2+67t+G3XJJqAfsiLK1QeI9A1Sccy43XnkFYD2bIqAcAQYabo?= =?us-ascii?Q?g0UU6+OSt70DPvL8bbkayHtnJNCDWidDQvMtcWfuaYd6ZS6fxiTw0tqOhfX+?= =?us-ascii?Q?07dNhTm5JMqHJ4xn9Do9AAlXQHTJv/0pi3FQTsenX9kcj2OYJk2X2atrdTny?= =?us-ascii?Q?crfkytVPndcRR+uFIvMUScZlhaudAnJhnVpOlyJpsDqgN64f5Zve2ybckC1O?= =?us-ascii?Q?G4X/bVuDhz8eKNvjVsPhnryvCGOoeNO+NCtZ2LQNammkQOJrls2H5lqTkkfj?= =?us-ascii?Q?js6d/2Hni/SI6kJEw56ZIPIfNmoEl1DR6aRPtkg58aqBhczTWx+33/YQvsEe?= =?us-ascii?Q?MlmrR0T2BkkEdNaLA3A+qzHJ3OBV6RPNtH+LKZxAOt7oLZ0LLP3PdE+6AOZF?= =?us-ascii?Q?pXsfoQ4em5KJjVt1CWoA3VxYiazsrbRTEJrWs98vd579CBbqfYhpXR6yHrCc?= =?us-ascii?Q?RQcTM3RVlBfncH/qXP5CNKcxx7DJxV0nZWY8OkfPrHkKz3eaFHA0jPvo7xOA?= =?us-ascii?Q?DiRV18N7x0+Lv5qtjq1IABuEXqF6XWUojJwJhyoR/fxKBCyPRK/yRndrL/oS?= =?us-ascii?Q?tmK3nWeP2PoXtgV9p7OBkjh8vYGHDxp/fbPCvXDBBfIpopAGXio3h1LMPzwA?= =?us-ascii?Q?oYckrGVtgkbRJIwZodt8zjUs0dKr47BBO+k1hFrMl0ADcOilse1wJeLh4IPg?= =?us-ascii?Q?SCFQmcRbmOI2H0C2pa2MtGD9O7NpJKt+SfbJKJPr0Yb4bQ4T7kzWfu4WO1b8?= =?us-ascii?Q?za+R7S0DNZcmZp/cjc93uLO+kItgO5IJ0S9OAm1V?= X-OriginatorOrg: nxp.com X-MS-Exchange-CrossTenant-Network-Message-Id: 987927b4-afa8-49f8-1f14-08de324a7168 X-MS-Exchange-CrossTenant-AuthSource: VI2PR04MB11147.eurprd04.prod.outlook.com X-MS-Exchange-CrossTenant-AuthAs: Internal X-MS-Exchange-CrossTenant-OriginalArrivalTime: 03 Dec 2025 09:00:45.3592 (UTC) X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted X-MS-Exchange-CrossTenant-Id: 686ea1d3-bc2b-4c6f-a92c-d99c5c301635 X-MS-Exchange-CrossTenant-MailboxType: HOSTED X-MS-Exchange-CrossTenant-UserPrincipalName: 4ye57I6jhS73wQL7/RGIFFrSHJN6Xqs0iFpfg1OKeXBF0k52Hnb2EAUCek+Cr38HEyQa+SufF4ApFXo+SiUmdA== X-MS-Exchange-Transport-CrossTenantHeadersStamped: AM0PR04MB6868 Content-Type: text/plain; charset="utf-8" ECSPI transfers only one word per frame in DMA mode, causing SCLK stalls between words due to BURST_LENGTH updates, which significantly impacts performance. To improve throughput, configure BURST_LENGTH as large as possible (up to 512 bytes per frame) instead of word length. This avoids delays between words. When transfer length is not 4-byte aligned, use bounce buffers to align data for DMA. TX uses aligned words for TXFIFO, while RX trims DMA buffer data after transfer completion. Introduce a new dma_package structure to store: 1. BURST_LENGTH values for each DMA request 2. Variables for DMA submission 3. DMA transmission length and actual data length Handle three cases: - len <=3D 512 bytes: one package, BURST_LENGTH =3D len * 8 - 1 - len > 512 and aligned: one package, BURST_LENGTH =3D max (512 bytes) - len > 512 and unaligned: two packages, second for tail data Performance test (spidev_test @10MHz, 4KB): Before: tx/rx ~6651.9 kbps After: tx/rx ~9922.2 kbps (~50% improvement) For compatibility with slow SPI devices, add configurable word delay in DMA mode. When word delay is set, dynamic burst is disabled and BURST_LENGTH equals word length. Signed-off-by: Carlos Song Reviewed-by: Frank Li --- drivers/spi/spi-imx.c | 413 ++++++++++++++++++++++++++++++++++++++---- 1 file changed, 377 insertions(+), 36 deletions(-) diff --git a/drivers/spi/spi-imx.c b/drivers/spi/spi-imx.c index 42f64d9535c9..04a492e7dbeb 100644 --- a/drivers/spi/spi-imx.c +++ b/drivers/spi/spi-imx.c @@ -60,6 +60,7 @@ MODULE_PARM_DESC(polling_limit_us, #define MX51_ECSPI_CTRL_MAX_BURST 512 /* The maximum bytes that IMX53_ECSPI can transfer in target mode.*/ #define MX53_MAX_TRANSFER_BYTES 512 +#define BYTES_PER_32BITS_WORD 4 =20 enum spi_imx_devtype { IMX1_CSPI, @@ -95,6 +96,16 @@ struct spi_imx_devtype_data { enum spi_imx_devtype devtype; }; =20 +struct dma_data_package { + u32 cmd_word; + void *dma_rx_buf; + void *dma_tx_buf; + dma_addr_t dma_tx_addr; + dma_addr_t dma_rx_addr; + int dma_len; + int data_len; +}; + struct spi_imx_data { struct spi_controller *controller; struct device *dev; @@ -130,6 +141,9 @@ struct spi_imx_data { u32 wml; struct completion dma_rx_completion; struct completion dma_tx_completion; + size_t dma_package_num; + struct dma_data_package *dma_data; + int rx_offset; =20 const struct spi_imx_devtype_data *devtype_data; }; @@ -189,6 +203,9 @@ MXC_SPI_BUF_TX(u16) MXC_SPI_BUF_RX(u32) MXC_SPI_BUF_TX(u32) =20 +/* Align to cache line to avoid swiotlo bounce */ +#define DMA_CACHE_ALIGNED_LEN(x) ALIGN((x), dma_get_cache_alignment()) + /* First entry is reserved, second entry is valid only if SDHC_SPIEN is set * (which is currently not the case in this driver) */ @@ -253,6 +270,14 @@ static bool spi_imx_can_dma(struct spi_controller *con= troller, struct spi_device if (transfer->len < spi_imx->devtype_data->fifo_size) return false; =20 + /* DMA only can transmit data in bytes */ + if (spi_imx->bits_per_word !=3D 8 && spi_imx->bits_per_word !=3D 16 && + spi_imx->bits_per_word !=3D 32) + return false; + + if (transfer->len >=3D MAX_SDMA_BD_BYTES) + return false; + spi_imx->dynamic_burst =3D 0; =20 return true; @@ -1398,8 +1423,6 @@ static int spi_imx_sdma_init(struct device *dev, stru= ct spi_imx_data *spi_imx, =20 init_completion(&spi_imx->dma_rx_completion); init_completion(&spi_imx->dma_tx_completion); - controller->can_dma =3D spi_imx_can_dma; - controller->max_dma_len =3D MAX_SDMA_BD_BYTES; spi_imx->controller->flags =3D SPI_CONTROLLER_MUST_RX | SPI_CONTROLLER_MUST_TX; =20 @@ -1437,10 +1460,259 @@ static int spi_imx_calculate_timeout(struct spi_im= x_data *spi_imx, int size) return secs_to_jiffies(2 * timeout); } =20 +static void spi_imx_dma_unmap(struct spi_imx_data *spi_imx, + struct dma_data_package *dma_data) +{ + struct device *tx_dev =3D spi_imx->controller->dma_tx->device->dev; + struct device *rx_dev =3D spi_imx->controller->dma_rx->device->dev; + + dma_unmap_single(tx_dev, dma_data->dma_tx_addr, + DMA_CACHE_ALIGNED_LEN(dma_data->dma_len), + DMA_TO_DEVICE); + dma_unmap_single(rx_dev, dma_data->dma_rx_addr, + DMA_CACHE_ALIGNED_LEN(dma_data->dma_len), + DMA_FROM_DEVICE); +} + +static void spi_imx_dma_rx_data_handle(struct spi_imx_data *spi_imx, + struct dma_data_package *dma_data, void *rx_buf, + bool word_delay) +{ + void *copy_ptr; + int unaligned; + + /* + * On little-endian CPUs, adjust byte order: + * - Swap bytes when bpw =3D 8 + * - Swap half-words when bpw =3D 16 + * This ensures correct data ordering for DMA transfers. + */ +#ifdef __LITTLE_ENDIAN + if (!word_delay) { + unsigned int bytes_per_word =3D spi_imx_bytes_per_word(spi_imx->bits_per= _word); + u32 *temp =3D dma_data->dma_rx_buf; + + for (int i =3D 0; i < DIV_ROUND_UP(dma_data->dma_len, sizeof(*temp)); i+= +) { + if (bytes_per_word =3D=3D 1) + swab32s(temp + i); + else if (bytes_per_word =3D=3D 2) + swahw32s(temp + i); + } + } +#endif + + /* + * When dynamic burst enabled, DMA RX always receives 32-bit words from R= XFIFO with + * buswidth =3D 4, but when data_len is not 4-bytes alignment, the RM sho= ws when + * burst length =3D 32*n + m bits, a SPI burst contains the m LSB in firs= t word and all + * 32 bits in other n words. So if garbage bytes in the first word, trim = first word then + * copy the actual data to rx_buf. + */ + if (dma_data->data_len % BYTES_PER_32BITS_WORD && !word_delay) { + unaligned =3D dma_data->data_len % BYTES_PER_32BITS_WORD; + copy_ptr =3D (u8 *)dma_data->dma_rx_buf + BYTES_PER_32BITS_WORD - unalig= ned; + } else { + copy_ptr =3D dma_data->dma_rx_buf; + } + + memcpy(rx_buf, copy_ptr, dma_data->data_len); +} + +static int spi_imx_dma_map(struct spi_imx_data *spi_imx, + struct dma_data_package *dma_data) +{ + struct spi_controller *controller =3D spi_imx->controller; + struct device *tx_dev =3D controller->dma_tx->device->dev; + struct device *rx_dev =3D controller->dma_rx->device->dev; + int ret; + + dma_data->dma_tx_addr =3D dma_map_single(tx_dev, dma_data->dma_tx_buf, + DMA_CACHE_ALIGNED_LEN(dma_data->dma_len), + DMA_TO_DEVICE); + ret =3D dma_mapping_error(tx_dev, dma_data->dma_tx_addr); + if (ret < 0) { + dev_err(spi_imx->dev, "DMA TX map failed %d\n", ret); + return ret; + } + + dma_data->dma_rx_addr =3D dma_map_single(rx_dev, dma_data->dma_rx_buf, + DMA_CACHE_ALIGNED_LEN(dma_data->dma_len), + DMA_FROM_DEVICE); + ret =3D dma_mapping_error(rx_dev, dma_data->dma_rx_addr); + if (ret < 0) { + dev_err(spi_imx->dev, "DMA RX map failed %d\n", ret); + dma_unmap_single(tx_dev, dma_data->dma_tx_addr, + DMA_CACHE_ALIGNED_LEN(dma_data->dma_len), + DMA_TO_DEVICE); + return ret; + } + + return 0; +} + +static int spi_imx_dma_tx_data_handle(struct spi_imx_data *spi_imx, + struct dma_data_package *dma_data, + const void *tx_buf, + bool word_delay) +{ + void *copy_ptr; + int unaligned; + + if (word_delay) { + dma_data->dma_len =3D dma_data->data_len; + } else { + /* + * As per the reference manual, when burst length =3D 32*n + m bits, ECS= PI + * sends m LSB bits in the first word, followed by n full 32-bit words. + * Since actual data may not be 4-byte aligned, allocate DMA TX/RX buffe= rs + * to ensure alignment. For TX, DMA pushes 4-byte aligned words to TXFIF= O, + * while ECSPI uses BURST_LENGTH settings to maintain correct bit count. + * For RX, DMA always receives 32-bit words from RXFIFO, when data len is + * not 4-byte aligned, trim the first word to drop garbage bytes, then g= roup + * all transfer DMA bounse buffer and copy all valid data to rx_buf. + */ + dma_data->dma_len =3D ALIGN(dma_data->data_len, BYTES_PER_32BITS_WORD); + } + + dma_data->dma_tx_buf =3D kzalloc(dma_data->dma_len, GFP_KERNEL); + if (!dma_data->dma_tx_buf) + return -ENOMEM; + + dma_data->dma_rx_buf =3D kzalloc(dma_data->dma_len, GFP_KERNEL); + if (!dma_data->dma_rx_buf) { + kfree(dma_data->dma_tx_buf); + return -ENOMEM; + } + + if (dma_data->data_len % BYTES_PER_32BITS_WORD && !word_delay) { + unaligned =3D dma_data->data_len % BYTES_PER_32BITS_WORD; + copy_ptr =3D (u8 *)dma_data->dma_tx_buf + BYTES_PER_32BITS_WORD - unalig= ned; + } else { + copy_ptr =3D dma_data->dma_tx_buf; + } + + memcpy(copy_ptr, tx_buf, dma_data->data_len); + + /* + * When word_delay is enabled, DMA transfers an entire word in one minor = loop. + * In this case, no data requires additional handling. + */ + if (word_delay) + return 0; + +#ifdef __LITTLE_ENDIAN + /* + * On little-endian CPUs, adjust byte order: + * - Swap bytes when bpw =3D 8 + * - Swap half-words when bpw =3D 16 + * This ensures correct data ordering for DMA transfers. + */ + unsigned int bytes_per_word =3D spi_imx_bytes_per_word(spi_imx->bits_per_= word); + u32 *temp =3D dma_data->dma_tx_buf; + + for (int i =3D 0; i < DIV_ROUND_UP(dma_data->dma_len, sizeof(*temp)); i++= ) { + if (bytes_per_word =3D=3D 1) + swab32s(temp + i); + else if (bytes_per_word =3D=3D 2) + swahw32s(temp + i); + } +#endif + + return 0; +} + +static int spi_imx_dma_data_prepare(struct spi_imx_data *spi_imx, + struct spi_transfer *transfer, + bool word_delay) +{ + u32 pre_bl, tail_bl; + u32 ctrl; + int ret; + + /* + * ECSPI supports a maximum burst of 512 bytes. When xfer->len exceeds 512 + * and is not a multiple of 512, a tail transfer is required. BURST_LEGTH + * is used for SPI HW to maintain correct bit count. BURST_LENGTH should + * update with data length. After DMA request submit, SPI can not update = the + * BURST_LENGTH, in this case, we must split two package, update the regi= ster + * then setup second DMA transfer. + */ + ctrl =3D readl(spi_imx->base + MX51_ECSPI_CTRL); + if (word_delay) { + /* + * When SPI IMX need to support word delay, according to "Sample Period = Control + * Register" shows, The Sample Period Control Register (ECSPI_PERIODREG) + * provides software a way to insert delays (wait states) between consec= utive + * SPI transfers. As a result, ECSPI can only transfer one word per fram= e, and + * the delay occurs between frames. + */ + spi_imx->dma_package_num =3D 1; + pre_bl =3D spi_imx->bits_per_word - 1; + } else if (transfer->len <=3D MX51_ECSPI_CTRL_MAX_BURST) { + spi_imx->dma_package_num =3D 1; + pre_bl =3D transfer->len * BITS_PER_BYTE - 1; + } else if (!(transfer->len % MX51_ECSPI_CTRL_MAX_BURST)) { + spi_imx->dma_package_num =3D 1; + pre_bl =3D MX51_ECSPI_CTRL_MAX_BURST * BITS_PER_BYTE - 1; + } else { + spi_imx->dma_package_num =3D 2; + pre_bl =3D MX51_ECSPI_CTRL_MAX_BURST * BITS_PER_BYTE - 1; + tail_bl =3D (transfer->len % MX51_ECSPI_CTRL_MAX_BURST) * BITS_PER_BYTE = - 1; + } + + spi_imx->dma_data =3D kmalloc_array(spi_imx->dma_package_num, + sizeof(struct dma_data_package), + GFP_KERNEL | __GFP_ZERO); + if (!spi_imx->dma_data) { + dev_err(spi_imx->dev, "Failed to allocate DMA package buffer!\n"); + return -ENOMEM; + } + + if (spi_imx->dma_package_num =3D=3D 1) { + ctrl &=3D ~MX51_ECSPI_CTRL_BL_MASK; + ctrl |=3D pre_bl << MX51_ECSPI_CTRL_BL_OFFSET; + spi_imx->dma_data[0].cmd_word =3D ctrl; + spi_imx->dma_data[0].data_len =3D transfer->len; + ret =3D spi_imx_dma_tx_data_handle(spi_imx, &spi_imx->dma_data[0], trans= fer->tx_buf, + word_delay); + if (ret) { + kfree(spi_imx->dma_data); + return ret; + } + } else { + ctrl &=3D ~MX51_ECSPI_CTRL_BL_MASK; + ctrl |=3D pre_bl << MX51_ECSPI_CTRL_BL_OFFSET; + spi_imx->dma_data[0].cmd_word =3D ctrl; + spi_imx->dma_data[0].data_len =3D round_down(transfer->len, + MX51_ECSPI_CTRL_MAX_BURST); + ret =3D spi_imx_dma_tx_data_handle(spi_imx, &spi_imx->dma_data[0], trans= fer->tx_buf, + false); + if (ret) { + kfree(spi_imx->dma_data); + return ret; + } + + ctrl &=3D ~MX51_ECSPI_CTRL_BL_MASK; + ctrl |=3D tail_bl << MX51_ECSPI_CTRL_BL_OFFSET; + spi_imx->dma_data[1].cmd_word =3D ctrl; + spi_imx->dma_data[1].data_len =3D transfer->len % MX51_ECSPI_CTRL_MAX_BU= RST; + ret =3D spi_imx_dma_tx_data_handle(spi_imx, &spi_imx->dma_data[1], + transfer->tx_buf + spi_imx->dma_data[0].data_len, + false); + if (ret) { + kfree(spi_imx->dma_data[0].dma_tx_buf); + kfree(spi_imx->dma_data[0].dma_rx_buf); + kfree(spi_imx->dma_data); + } + } + + return 0; +} + static int spi_imx_dma_submit(struct spi_imx_data *spi_imx, + struct dma_data_package *dma_data, struct spi_transfer *transfer) { - struct sg_table *tx =3D &transfer->tx_sg, *rx =3D &transfer->rx_sg; struct spi_controller *controller =3D spi_imx->controller; struct dma_async_tx_descriptor *desc_tx, *desc_rx; unsigned long transfer_timeout; @@ -1451,9 +1723,9 @@ static int spi_imx_dma_submit(struct spi_imx_data *sp= i_imx, * The TX DMA setup starts the transfer, so make sure RX is configured * before TX. */ - desc_rx =3D dmaengine_prep_slave_sg(controller->dma_rx, - rx->sgl, rx->nents, DMA_DEV_TO_MEM, - DMA_PREP_INTERRUPT | DMA_CTRL_ACK); + desc_rx =3D dmaengine_prep_slave_single(controller->dma_rx, dma_data->dma= _rx_addr, + dma_data->dma_len, DMA_DEV_TO_MEM, + DMA_PREP_INTERRUPT | DMA_CTRL_ACK); if (!desc_rx) { transfer->error |=3D SPI_TRANS_FAIL_NO_START; return -EINVAL; @@ -1471,9 +1743,9 @@ static int spi_imx_dma_submit(struct spi_imx_data *sp= i_imx, reinit_completion(&spi_imx->dma_rx_completion); dma_async_issue_pending(controller->dma_rx); =20 - desc_tx =3D dmaengine_prep_slave_sg(controller->dma_tx, - tx->sgl, tx->nents, DMA_MEM_TO_DEV, - DMA_PREP_INTERRUPT | DMA_CTRL_ACK); + desc_tx =3D dmaengine_prep_slave_single(controller->dma_tx, dma_data->dma= _tx_addr, + dma_data->dma_len, DMA_MEM_TO_DEV, + DMA_PREP_INTERRUPT | DMA_CTRL_ACK); if (!desc_tx) goto dmaengine_terminate_rx; =20 @@ -1521,16 +1793,16 @@ static int spi_imx_dma_submit(struct spi_imx_data *= spi_imx, } =20 static void spi_imx_dma_max_wml_find(struct spi_imx_data *spi_imx, - struct spi_transfer *transfer) + struct dma_data_package *dma_data, + bool word_delay) { - struct sg_table *rx =3D &transfer->rx_sg; - struct scatterlist *last_sg =3D sg_last(rx->sgl, rx->nents); - unsigned int bytes_per_word, i; + unsigned int bytes_per_word =3D word_delay ? + spi_imx_bytes_per_word(spi_imx->bits_per_word) : + BYTES_PER_32BITS_WORD; + unsigned int i; =20 - /* Get the right burst length from the last sg to ensure no tail data */ - bytes_per_word =3D spi_imx_bytes_per_word(transfer->bits_per_word); for (i =3D spi_imx->devtype_data->fifo_size / 2; i > 0; i--) { - if (!(sg_dma_len(last_sg) % (i * bytes_per_word))) + if (!dma_data->dma_len % (i * bytes_per_word)) break; } /* Use 1 as wml in case no available burst length got */ @@ -1540,25 +1812,29 @@ static void spi_imx_dma_max_wml_find(struct spi_imx= _data *spi_imx, spi_imx->wml =3D i; } =20 -static int spi_imx_dma_configure(struct spi_controller *controller) +static int spi_imx_dma_configure(struct spi_controller *controller, bool w= ord_delay) { int ret; enum dma_slave_buswidth buswidth; struct dma_slave_config rx =3D {}, tx =3D {}; struct spi_imx_data *spi_imx =3D spi_controller_get_devdata(controller); =20 - switch (spi_imx_bytes_per_word(spi_imx->bits_per_word)) { - case 4: + if (word_delay) { + switch (spi_imx_bytes_per_word(spi_imx->bits_per_word)) { + case 4: + buswidth =3D DMA_SLAVE_BUSWIDTH_4_BYTES; + break; + case 2: + buswidth =3D DMA_SLAVE_BUSWIDTH_2_BYTES; + break; + case 1: + buswidth =3D DMA_SLAVE_BUSWIDTH_1_BYTE; + break; + default: + return -EINVAL; + } + } else { buswidth =3D DMA_SLAVE_BUSWIDTH_4_BYTES; - break; - case 2: - buswidth =3D DMA_SLAVE_BUSWIDTH_2_BYTES; - break; - case 1: - buswidth =3D DMA_SLAVE_BUSWIDTH_1_BYTE; - break; - default: - return -EINVAL; } =20 tx.direction =3D DMA_MEM_TO_DEV; @@ -1584,15 +1860,17 @@ static int spi_imx_dma_configure(struct spi_control= ler *controller) return 0; } =20 -static int spi_imx_dma_transfer(struct spi_imx_data *spi_imx, - struct spi_transfer *transfer) +static int spi_imx_dma_package_transfer(struct spi_imx_data *spi_imx, + struct dma_data_package *dma_data, + struct spi_transfer *transfer, + bool word_delay) { struct spi_controller *controller =3D spi_imx->controller; int ret; =20 - spi_imx_dma_max_wml_find(spi_imx, transfer); + spi_imx_dma_max_wml_find(spi_imx, dma_data, word_delay); =20 - ret =3D spi_imx_dma_configure(controller); + ret =3D spi_imx_dma_configure(controller, word_delay); if (ret) goto dma_failure_no_start; =20 @@ -1603,10 +1881,17 @@ static int spi_imx_dma_transfer(struct spi_imx_data= *spi_imx, } spi_imx->devtype_data->setup_wml(spi_imx); =20 - ret =3D spi_imx_dma_submit(spi_imx, transfer); + ret =3D spi_imx_dma_submit(spi_imx, dma_data, transfer); if (ret) return ret; =20 + /* Trim the DMA RX buffer and copy the actual data to rx_buf */ + dma_sync_single_for_cpu(controller->dma_rx->device->dev, dma_data->dma_rx= _addr, + dma_data->dma_len, DMA_FROM_DEVICE); + spi_imx_dma_rx_data_handle(spi_imx, dma_data, transfer->rx_buf + spi_imx-= >rx_offset, + word_delay); + spi_imx->rx_offset +=3D dma_data->data_len; + return 0; /* fallback to pio */ dma_failure_no_start: @@ -1614,6 +1899,57 @@ static int spi_imx_dma_transfer(struct spi_imx_data = *spi_imx, return ret; } =20 +static int spi_imx_dma_transfer(struct spi_imx_data *spi_imx, + struct spi_transfer *transfer) +{ + bool word_delay =3D transfer->word_delay.value !=3D 0; + int ret; + int i; + + ret =3D spi_imx_dma_data_prepare(spi_imx, transfer, word_delay); + if (ret < 0) { + transfer->error |=3D SPI_TRANS_FAIL_NO_START; + dev_err(spi_imx->dev, "DMA data prepare fail\n"); + goto fallback_pio; + } + + spi_imx->rx_offset =3D 0; + + /* Each dma_package performs a separate DMA transfer once */ + for (i =3D 0; i < spi_imx->dma_package_num; i++) { + ret =3D spi_imx_dma_map(spi_imx, &spi_imx->dma_data[i]); + if (ret < 0) { + if (i =3D=3D 0) + transfer->error |=3D SPI_TRANS_FAIL_NO_START; + dev_err(spi_imx->dev, "DMA map fail\n"); + break; + } + + /* Update the CTRL register BL field */ + writel(spi_imx->dma_data[i].cmd_word, spi_imx->base + MX51_ECSPI_CTRL); + + ret =3D spi_imx_dma_package_transfer(spi_imx, &spi_imx->dma_data[i], + transfer, word_delay); + + /* Whether the dma transmission is successful or not, dma unmap is neces= sary */ + spi_imx_dma_unmap(spi_imx, &spi_imx->dma_data[i]); + + if (ret < 0) { + dev_dbg(spi_imx->dev, "DMA %d transfer not really finish\n", i); + break; + } + } + + for (int j =3D 0; j < spi_imx->dma_package_num; j++) { + kfree(spi_imx->dma_data[j].dma_tx_buf); + kfree(spi_imx->dma_data[j].dma_rx_buf); + } + kfree(spi_imx->dma_data); + +fallback_pio: + return ret; +} + static int spi_imx_pio_transfer(struct spi_device *spi, struct spi_transfer *transfer) { @@ -1780,9 +2116,14 @@ static int spi_imx_transfer_one(struct spi_controlle= r *controller, * transfer, the SPI transfer has already been mapped, so we * have to do the DMA transfer here. */ - if (spi_imx->usedma) - return spi_imx_dma_transfer(spi_imx, transfer); - + if (spi_imx->usedma) { + ret =3D spi_imx_dma_transfer(spi_imx, transfer); + if (transfer->error & SPI_TRANS_FAIL_NO_START) { + spi_imx->usedma =3D false; + return spi_imx_pio_transfer(spi, transfer); + } + return ret; + } /* run in polling mode for short transfers */ if (transfer->len =3D=3D 1 || (polling_limit_us && spi_imx_transfer_estimate_time_us(transfer) < polling_limit_us)) --=20 2.34.1