Date:   Thu, 21 Mar 2019 11:00:45 +0000
From:   Naga Sureshkumar Relli <nagasure@...inx.com>
To:     Vignesh Raghavendra <vigneshr@...com>,
        "broonie@...nel.org" <broonie@...nel.org>,
        "bbrezillon@...nel.org" <bbrezillon@...nel.org>
CC:     "linux-spi@...r.kernel.org" <linux-spi@...r.kernel.org>,
        "dwmw2@...radead.org" <dwmw2@...radead.org>,
        "marek.vasut@...il.com" <marek.vasut@...il.com>,
        "richard@....at" <richard@....at>,
        "linux-mtd@...ts.infradead.org" <linux-mtd@...ts.infradead.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        Michal Simek <michals@...inx.com>,
        "nagasuresh12@...il.com" <nagasuresh12@...il.com>
Subject: RE: [RFC PATCH 2/2] spi: spi-mem: Add support for Zynq QSPI
 controller

Hi Vignesh,

Thanks for the review.

> -----Original Message-----
> From: linux-spi-owner@...r.kernel.org <linux-spi-owner@...r.kernel.org> On Behalf Of
> Vignesh Raghavendra
> Sent: Wednesday, March 20, 2019 10:42 PM
> To: Naga Sureshkumar Relli <nagasure@...inx.com>; broonie@...nel.org;
> bbrezillon@...nel.org
> Cc: linux-spi@...r.kernel.org; dwmw2@...radead.org; marek.vasut@...il.com;
> richard@....at; linux-mtd@...ts.infradead.org; linux-kernel@...r.kernel.org; Michal Simek
> <michals@...inx.com>; nagasuresh12@...il.com
> Subject: Re: [RFC PATCH 2/2] spi: spi-mem: Add support for Zynq QSPI controller
> 
> Hi,
> 
> Few more comments that I missed last time around.
> 
> On 28/02/19 12:32 PM, Naga Sureshkumar Relli wrote:
> > Add support for QSPI controller driver used by Xilinx Zynq SOC.
> >
> > Signed-off-by: Naga Sureshkumar Relli
> > <naga.sureshkumar.relli@...inx.com>
> > ---
> >  drivers/spi/Kconfig         |   8 +
> >  drivers/spi/Makefile        |   1 +
> >  drivers/spi/spi-zynq-qspi.c | 780
> > ++++++++++++++++++++++++++++++++++++++++++++
> >  3 files changed, 789 insertions(+)
> >  create mode 100644 drivers/spi/spi-zynq-qspi.c
> >
> > diff --git a/drivers/spi/Kconfig b/drivers/spi/Kconfig index
> > 9f89cb1..f12c880 100644
> > --- a/drivers/spi/Kconfig
> > +++ b/drivers/spi/Kconfig
> > @@ -816,6 +816,14 @@ config SPI_XTENSA_XTFPGA
> >  	  16 bit words in SPI mode 0, automatically asserting CS on transfer
> >  	  start and deasserting on end.
> >
> > +config SPI_ZYNQ_QSPI
> > +	tristate "Xilinx Zynq QSPI controller"
> > +	depends on ARCH_ZYNQ
> 
> Could you enable the driver for compile testing?
> From quick look this should enough:
Ok. I will update it in next version.
> 
> depends on ARCH_ZYNQ || COMPILE_TEST
> 
> > +	help
> > +	  This enables support for the Zynq Quad SPI controller
> > +	  in master mode.
> > +	  This controller only supports SPI memory interface.
> > +
> >  config SPI_ZYNQMP_GQSPI
> >  	tristate "Xilinx ZynqMP GQSPI controller"
> >  	depends on SPI_MASTER && HAS_DMA
> 
> [...]
> 
> > +/**
> > + * zynq_qspi_exec_mem_op() - Initiates the QSPI transfer
> > + * @mem: the SPI memory
> > + * @op: the memory operation to execute
> > + *
> > + * Executes a memory operation.
> > + *
> > + * This function first selects the chip and starts the memory operation.
> > + *
> > + * Return: 0 in case of success, a negative error code otherwise.
> > + */
> > +static int zynq_qspi_exec_mem_op(struct spi_mem *mem,
> > +				 const struct spi_mem_op *op)
> > +{
> > +	struct zynq_qspi *xqspi = spi_controller_get_devdata(mem->spi->master);
> > +	int err = 0, i;
> > +	u8 *tmpbuf;
> > +
> > +	dev_dbg(xqspi->dev, "cmd:%#x mode:%d.%d.%d.%d\n",
> > +		op->cmd.opcode, op->cmd.buswidth, op->addr.buswidth,
> > +		op->dummy.buswidth, op->data.buswidth);
> > +
> > +	zynq_qspi_chipselect(mem->spi, true);
> > +	zynq_qspi_config_op(xqspi, mem->spi);
> > +
> > +	if (op->cmd.opcode) {
> > +		reinit_completion(&xqspi->data_completion);
> > +		xqspi->txbuf = (u8 *)&op->cmd.opcode;
> > +		xqspi->rxbuf = NULL;
> > +		xqspi->tx_bytes = sizeof(op->cmd.opcode);
> > +		xqspi->rx_bytes = sizeof(op->cmd.opcode);
> > +		zynq_qspi_write_op(xqspi, ZYNQ_QSPI_FIFO_DEPTH, true);
> > +		zynq_qspi_write(xqspi, ZYNQ_QSPI_IEN_OFFSET,
> > +				ZYNQ_QSPI_IXR_RXTX_MASK);
> > +		if (!wait_for_completion_interruptible_timeout(&xqspi->data_completion,
> > +							       msecs_to_jiffies(1000))) {
> > +			err = -ETIMEDOUT;
> > +		}
> > +	}
> > +	zynq_qspi_config_op(xqspi, mem->spi);
> 
> Why is it required to call zynq_qspi_config_op() b/w every stage of single flash operation? Is
> it not enough to do once at the beginning?
Yes, we can do it once. 
I will update it.

> 
> > +	if (op->addr.nbytes) {
> > +		for (i = 0; i < op->addr.nbytes; i++) {
> > +			xqspi->txbuf[i] = op->addr.val >>
> > +					(8 * (op->addr.nbytes - i - 1));
> > +		}
> > +
> > +		reinit_completion(&xqspi->data_completion);
> > +		xqspi->rxbuf = NULL;
> > +		xqspi->tx_bytes = op->addr.nbytes;
> > +		xqspi->rx_bytes = op->addr.nbytes;
> > +		zynq_qspi_write_op(xqspi, ZYNQ_QSPI_FIFO_DEPTH, true);
> > +		zynq_qspi_write(xqspi, ZYNQ_QSPI_IEN_OFFSET,
> > +				ZYNQ_QSPI_IXR_RXTX_MASK);
> > +		if (!wait_for_completion_interruptible_timeout(&xqspi->data_completion,
> > +							       msecs_to_jiffies(1000))) {
> > +			err = -ETIMEDOUT;
> > +		}
> > +	}
> > +	zynq_qspi_config_op(xqspi, mem->spi);
> > +	if (op->dummy.nbytes) {
> > +		tmpbuf = kzalloc(op->dummy.nbytes, GFP_KERNEL);
> > +		memset(tmpbuf, 0xff, op->dummy.nbytes);
> > +		reinit_completion(&xqspi->data_completion);
> > +		xqspi->txbuf = tmpbuf;
> > +		xqspi->rxbuf = NULL;
> > +		xqspi->tx_bytes = op->dummy.nbytes;
> > +		xqspi->rx_bytes = op->dummy.nbytes;
> > +		zynq_qspi_write_op(xqspi, ZYNQ_QSPI_FIFO_DEPTH, true);
> > +		zynq_qspi_write(xqspi, ZYNQ_QSPI_IEN_OFFSET,
> > +				ZYNQ_QSPI_IXR_RXTX_MASK);
> > +		if (!wait_for_completion_interruptible_timeout(&xqspi->data_completion,
> > +							       msecs_to_jiffies(1000))) {
> > +			err = -ETIMEDOUT;
> > +		}
> > +		kfree(tmpbuf);
> > +	}
> > +	zynq_qspi_config_op(xqspi, mem->spi);
> > +	if (op->data.nbytes) {
> > +		reinit_completion(&xqspi->data_completion);
> > +		if (op->data.dir == SPI_MEM_DATA_OUT) {
> > +			xqspi->txbuf = (u8 *)op->data.buf.out;
> > +			xqspi->tx_bytes = op->data.nbytes;
> > +			xqspi->rxbuf = NULL;
> > +			xqspi->rx_bytes = op->data.nbytes;
> > +		} else {
> > +			xqspi->txbuf = NULL;
> > +			xqspi->rxbuf = (u8 *)op->data.buf.in;
> > +			xqspi->rx_bytes = op->data.nbytes;
> > +			xqspi->tx_bytes = op->data.nbytes;
> > +		}
> > +
> > +		zynq_qspi_write_op(xqspi, ZYNQ_QSPI_FIFO_DEPTH, true);
> > +		zynq_qspi_write(xqspi, ZYNQ_QSPI_IEN_OFFSET,
> > +				ZYNQ_QSPI_IXR_RXTX_MASK);
> > +		if (!wait_for_completion_interruptible_timeout(&xqspi->data_completion,
> > +							       msecs_to_jiffies(1000))) {
> > +			err = -ETIMEDOUT;
> > +		}
> > +	}
> > +	zynq_qspi_chipselect(mem->spi, false);
> > +
> > +	return err;
> > +}
> > +
> > +static const struct spi_controller_mem_ops zynq_qspi_mem_ops = {
> > +	.supports_op = zynq_qspi_supports_op,
> > +	.exec_op = zynq_qspi_exec_mem_op,
> > +};
> > +
> > +/**
> > + * zynq_qspi_probe - Probe method for the QSPI driver
> > + * @pdev:	Pointer to the platform_device structure
> > + *
> > + * This function initializes the driver data structures and the hardware.
> > + *
> > + * Return:	0 on success and error value on failure
> > + */
> > +static int zynq_qspi_probe(struct platform_device *pdev) {
> > +	int ret = 0;
> > +	struct spi_controller *ctlr;
> > +	struct device *dev = &pdev->dev;
> > +	struct device_node *np = dev->of_node;
> > +	struct zynq_qspi *xqspi;
> > +	struct resource *res;
> > +	u32 num_cs;
> > +
> > +	ctlr = spi_alloc_master(&pdev->dev, sizeof(*xqspi));
> > +	if (!ctlr)
> > +		return -ENOMEM;
> > +
> > +	xqspi = spi_controller_get_devdata(ctlr);
> > +	xqspi->dev = dev;
> > +	platform_set_drvdata(pdev, xqspi);
> > +	ctlr->mem_ops = &zynq_qspi_mem_ops;
> 
> 
> ctlr->mem_ops is initialized twice. See below..
Ok. I will correct it.
> 
> > +	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> > +	xqspi->regs = devm_ioremap_resource(&pdev->dev, res);
> > +	if (IS_ERR(xqspi->regs)) {
> > +		ret = PTR_ERR(xqspi->regs);
> > +		goto remove_master;
> > +	}
> > +
> > +	xqspi->pclk = devm_clk_get(&pdev->dev, "pclk");
> > +	if (IS_ERR(xqspi->pclk)) {
> > +		dev_err(&pdev->dev, "pclk clock not found.\n");
> > +		ret = PTR_ERR(xqspi->pclk);
> > +		goto remove_master;
> > +	}
> > +
> > +	init_completion(&xqspi->data_completion);
> > +
> > +	xqspi->refclk = devm_clk_get(&pdev->dev, "ref_clk");
> > +	if (IS_ERR(xqspi->refclk)) {
> > +		dev_err(&pdev->dev, "ref_clk clock not found.\n");
> > +		ret = PTR_ERR(xqspi->refclk);
> > +		goto remove_master;
> > +	}
> > +
> > +	ret = clk_prepare_enable(xqspi->pclk);
> > +	if (ret) {
> > +		dev_err(&pdev->dev, "Unable to enable APB clock.\n");
> > +		goto remove_master;
> > +	}
> > +
> > +	ret = clk_prepare_enable(xqspi->refclk);
> > +	if (ret) {
> > +		dev_err(&pdev->dev, "Unable to enable device clock.\n");
> > +		goto clk_dis_pclk;
> > +	}
> > +
> > +	/* QSPI controller initializations */
> > +	zynq_qspi_init_hw(xqspi);
> > +
> > +	xqspi->irq = platform_get_irq(pdev, 0);
> > +	if (xqspi->irq <= 0) {
> > +		ret = -ENXIO;
> > +		dev_err(&pdev->dev, "irq resource not found\n");
> > +		goto remove_master;
> > +	}
> > +	ret = devm_request_irq(&pdev->dev, xqspi->irq, zynq_qspi_irq,
> > +			       0, pdev->name, xqspi);
> > +	if (ret != 0) {
> > +		ret = -ENXIO;
> > +		dev_err(&pdev->dev, "request_irq failed\n");
> > +		goto remove_master;
> > +	}
> > +
> > +	ret = of_property_read_u32(pdev->dev.of_node, "num-cs",
> > +				   &num_cs);
> > +	if (ret < 0)
> > +		ctlr->num_chipselect = ZYNQ_QSPI_DEFAULT_NUM_CS;
> > +	else
> > +		ctlr->num_chipselect = num_cs;
> > +	ctlr->mode_bits =  SPI_RX_DUAL | SPI_RX_QUAD |
> > +			    SPI_TX_DUAL | SPI_TX_QUAD;
> > +	ctlr->mem_ops = &zynq_qspi_mem_ops;
> 
> And here again...
Ok. will remove it from here.

Thanks,
Naga Sureshkumar Relli
> 
> > +	ctlr->setup = zynq_qspi_setup_op;
> > +	ctlr->max_speed_hz = clk_get_rate(xqspi->refclk) / 2;
> > +	ctlr->dev.of_node = np;
> > +	ret = spi_register_controller(ctlr);
> > +	if (ret) {
> > +		dev_err(&pdev->dev, "spi_register_master failed\n");
> > +		goto clk_dis_all;
> > +	}
> > +
> > +	return ret;
> > +
> > +clk_dis_all:
> > +	clk_disable_unprepare(xqspi->refclk);
> > +clk_dis_pclk:
> > +	clk_disable_unprepare(xqspi->pclk);
> > +remove_master:
> > +	spi_controller_put(ctlr);
> > +
> > +	return ret;
> > +}
> > +
> > +/**
> > + * zynq_qspi_remove - Remove method for the QSPI driver
> > + * @pdev:	Pointer to the platform_device structure
> > + *
> > + * This function is called if a device is physically removed from the
> > +system or
> > + * if the driver module is being unloaded. It frees all resources
> > +allocated to
> > + * the device.
> > + *
> > + * Return:	0 on success and error value on failure
> > + */
> > +static int zynq_qspi_remove(struct platform_device *pdev) {
> > +	struct zynq_qspi *xqspi = platform_get_drvdata(pdev);
> > +
> > +	zynq_qspi_write(xqspi, ZYNQ_QSPI_ENABLE_OFFSET, 0);
> > +
> > +	clk_disable_unprepare(xqspi->refclk);
> > +	clk_disable_unprepare(xqspi->pclk);
> > +
> > +	return 0;
> > +}
> > +
> > +static const struct of_device_id zynq_qspi_of_match[] = {
> > +	{ .compatible = "xlnx,zynq-qspi-1.0", },
> > +	{ /* end of table */ }
> > +};
> > +
> > +MODULE_DEVICE_TABLE(of, zynq_qspi_of_match);
> > +
> > +/*
> > + * zynq_qspi_driver - This structure defines the QSPI platform driver
> > +*/ static struct platform_driver zynq_qspi_driver = {
> > +	.probe = zynq_qspi_probe,
> > +	.remove = zynq_qspi_remove,
> > +	.driver = {
> > +		.name = "zynq-qspi",
> > +		.of_match_table = zynq_qspi_of_match,
> > +	},
> > +};
> > +
> > +module_platform_driver(zynq_qspi_driver);
> > +
> > +MODULE_AUTHOR("Xilinx, Inc.");
> > +MODULE_DESCRIPTION("Xilinx Zynq QSPI driver"); MODULE_LICENSE("GPL");
> >
> 
> --
> Regards
> Vignesh

Powered by blists - more mailing lists