Re: [PATCH v2 2/2] spi: cadence-quadpsi: Add support for the Cadence QSPI controller

From: Ramuthevar, Vadivel MuruganX
Date: Mon Nov 11 2019 - 02:32:11 EST


Hi Vignesh,

ÂÂ Thanks for review comments, sorry for the late reply since I was in one week vacation.

On 7/11/2019 1:44 PM, Vignesh Raghavendra wrote:

On 30/10/19 1:41 PM, Ramuthevar,Vadivel MuruganX wrote:
From: Ramuthevar Vadivel Murugan <vadivel.muruganx.ramuthevar@xxxxxxxxxxxxxxx>

Add support for the Cadence QSPI controller. This controller is
present in the Intel Lightning Mountain(LGM) SoCs, Altera and TI SoCs.
This driver has been tested on the Intel LGM SoCs.

This driver does not support generic SPI and also the implementation
only supports spi-mem interface to replace the existing driver in
mtd/spi-nor/cadence-quadspi.c, the existing driver only support SPI-NOR
flash memory.

Signed-off-by: Ramuthevar Vadivel Murugan <vadivel.muruganx.ramuthevar@xxxxxxxxxxxxxxx>
---
drivers/spi/Kconfig | 10 +
drivers/spi/Makefile | 1 +
drivers/spi/spi-cadence-quadspi.c | 1290 +++++++++++++++++++++++++++++++++++++
drivers/spi/spi-cadence-quadspi.h | 272 ++++++++
4 files changed, 1573 insertions(+)
create mode 100644 drivers/spi/spi-cadence-quadspi.c
create mode 100644 drivers/spi/spi-cadence-quadspi.h

diff --git a/drivers/spi/Kconfig b/drivers/spi/Kconfig
index 6f7fdcbb9151..ef4d961feb54 100644
--- a/drivers/spi/Kconfig
+++ b/drivers/spi/Kconfig
@@ -192,6 +192,16 @@ config SPI_CADENCE
This selects the Cadence SPI controller master driver
used by Xilinx Zynq and ZynqMP.
+config SPI_CADENCE_QUADSPI
+ tristate "Cadence Quad SPI controller"
+ depends on OF && (ARM || ARM64 || COMPILE_TEST || X86)
+ depends on MTD || MTD_SPI_NOR || MTD_SPI_NAND
+ help
+ Cadence QSPI is a specialized controller for connecting an SPI
+ Flash over 1/2/4-bit wide bus. This enables support for the Quad SPI
Supports Octal mode as well. So its 1/2/4/8-bit wide
Agreed!, will add it.
+ controller in master mode. This driver does not support generic SPI.
+ The implementation only supports spi-mem interface.
+
config SPI_CLPS711X
tristate "CLPS711X host SPI controller"
depends on ARCH_CLPS711X || COMPILE_TEST
diff --git a/drivers/spi/Makefile b/drivers/spi/Makefile
index bb49c9e6d0a0..84a59d75a13a 100644
--- a/drivers/spi/Makefile
+++ b/drivers/spi/Makefile
@@ -29,6 +29,7 @@ obj-$(CONFIG_SPI_BCM_QSPI) += spi-iproc-qspi.o spi-brcmstb-qspi.o spi-bcm-qspi.
obj-$(CONFIG_SPI_BITBANG) += spi-bitbang.o
obj-$(CONFIG_SPI_BUTTERFLY) += spi-butterfly.o
obj-$(CONFIG_SPI_CADENCE) += spi-cadence.o
+obj-$(CONFIG_SPI_CADENCE_QUADSPI) += spi-cadence-quadspi.o
obj-$(CONFIG_SPI_CLPS711X) += spi-clps711x.o
obj-$(CONFIG_SPI_COLDFIRE_QSPI) += spi-coldfire-qspi.o
obj-$(CONFIG_SPI_DAVINCI) += spi-davinci.o
diff --git a/drivers/spi/spi-cadence-quadspi.c b/drivers/spi/spi-cadence-quadspi.c
new file mode 100644
index 000000000000..bca391bfb58f
--- /dev/null
+++ b/drivers/spi/spi-cadence-quadspi.c
@@ -0,0 +1,1290 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Driver for Cadence QSPI Controller
+ *
+ * Copyright Altera Corporation (C) 2012-2014. All rights reserved.
+ * Copyright Intel Corporation (C) 2019. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+#include <linux/clk.h>
+#include <linux/completion.h>
+#include <linux/delay.h>
+#include <linux/dma-mapping.h>
+#include <linux/dmaengine.h>
+#include <linux/errno.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/jiffies.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/of_device.h>
+#include <linux/of.h>
+#include <linux/of_platform.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/platform_device.h>
+#include <linux/spi/spi.h>
+#include <linux/spi/spi-mem.h>
+#include <linux/unaligned/be_byteshift.h>
+
+#include "spi-cadence-quadspi.h"
+
+/* Quirks */
+#define CQSPI_NEEDS_WR_DELAY BIT(0)
+#define CQSPI_DISABLE_DAC_MODE BIT(1)
+
+#define CADENCE_QSPI_NAME "cadence-qspi"
+
+struct cqspi_driver_platdata {
+ u32 hwcaps_mask;
+ u8 quirks;
+};
+
+unsigned int cadence_qspi_init_timeout(const unsigned long timeout_in_ms)
+{
+ return jiffies + msecs_to_jiffies(timeout_in_ms);
+}
+
+unsigned int cadence_qspi_check_timeout(const unsigned long timeout)
+{
+ return time_before(jiffies, timeout);
+}
+
+static irqreturn_t cadence_qspi_irq_handler(int this_irq, void *dev)
+{
+ struct struct_cqspi *cadence_qspi = dev;
+ u32 irq_status;
+
+ /* Read interrupt status */
+ irq_status = readl(cadence_qspi->iobase + CQSPI_REG_IRQSTATUS);
+ if (!irq_status)
+ return IRQ_HANDLED;
+
+ cadence_qspi->irq_status = irq_status;
+
+ /* Clear interrupt */
+ writel(irq_status, cadence_qspi->iobase + CQSPI_REG_IRQSTATUS);
+ if (irq_status)
+ complete(&cadence_qspi->transfer_complete);
+
+ return IRQ_HANDLED;
+}
+
+static u32 cadence_qspi_cmd2addr(const unsigned char *addr_buf, u32 addr_width)
+{
+ unsigned int addr = 0;
+ int i;
+
+ /* Invalid address return zero. */
+ if (addr_width > 4)
+ return 0;
+
+ for (i = 0; i < addr_width; i++) {
+ addr = addr << 8;
+ addr |= addr_buf[i];
+ }
+
+ return addr;
+}
+
+void enable_qspi_direct_access(void *reg_base, bool enable)
+{
Please be consistent with function names. Start all functions with
cadence_qspi_. If need you could have a shorter function prefix like cqspi_

Also, this and few other functions below can be static and they are not
used outside of this module.
sure, will fix it next patch-set.
+ u32 reg;
+
+ reg = readl(reg_base + CQSPI_REG_CONFIG);
+ if (enable)
+ reg |= CQSPI_REG_CONFIG_DIRECT_MASK;
+ else
+ reg &= ~CQSPI_REG_CONFIG_DIRECT_MASK;
+
+ writel(reg, reg_base + CQSPI_REG_CONFIG);
+}
+
+void cadence_qspi_controller_enable(void *reg_base, bool enable)
+{
+ unsigned int reg;
+
+ reg = readl(reg_base + CQSPI_REG_CONFIG);
+ if (enable)
+ reg |= CQSPI_REG_CONFIG_ENABLE_MASK;
+ else
+ reg &= ~CQSPI_REG_CONFIG_ENABLE_MASK;
+
+ writel(reg, reg_base + CQSPI_REG_CONFIG);
+}
+
+static unsigned int cadence_qspi_wait_idle(void *reg_base)
+{
+ u32 timeout, count = 0;
+
+ timeout = cadence_qspi_init_timeout(CQSPI_READ_TIMEOUT_MS);
+ while (cadence_qspi_check_timeout(timeout)) {
+ if (CQSPI_REG_IS_IDLE(reg_base)) {
+ count++;
+ if (count >= CQSPI_POLL_IDLE_RETRY)
+ return 1;
+ } else {
+ count = 0;
+ }
+ }
+
+ return 0;
+}
+
Could you please keep polling routines from original driver
(drivers/mtd/spi-nor/cadence-quadspi.c). If you really want to improve
these helpers, please do those changes in a separate patch later on in
the series. This will greatly help in reviewing the code.

Also, I see that you have dropped many inline comments in the code.
These comments are quite important and I suggest not to drop them unless
they are no longer applicable

surely we will keep the polling routines and also as per your suggestion later will add the performance related patches.

To reduce the code size, removed the comments, will keep it.

+static void cadence_qspi_readdata_capture(void *reg_base, u32 bypass, u32 delay)
+{
+ unsigned int reg;
+
+ cadence_qspi_controller_enable(reg_base, 0);
+
+ reg = readl(reg_base + CQSPI_REG_READCAPTURE);
+ if (bypass)
+ reg |= (1 << CQSPI_REG_READCAPTURE_BYPASS_LSB);
+ else
+ reg &= ~(1 << CQSPI_REG_READCAPTURE_BYPASS_LSB);
+
+ reg &= ~(CQSPI_REG_READCAPTURE_DELAY_MASK
+ << CQSPI_REG_READCAPTURE_DELAY_LSB);
+ reg |= ((delay & CQSPI_REG_READCAPTURE_DELAY_MASK)
+ << CQSPI_REG_READCAPTURE_DELAY_LSB);
+ writel(reg, reg_base + CQSPI_REG_READCAPTURE);
+
+ cadence_qspi_controller_enable(reg_base, 1);
+}
+
+static void cadence_qspi_config_baudrate_div(void *reg_base, u32 ref_clk_hz,
+ u32 sclk_hz)
+{
+ unsigned int reg, div;
+
+ /* Recalculate the baudrate divisor based on QSPI specification. */
+ div = DIV_ROUND_UP(ref_clk_hz, 2 * sclk_hz) - 1;
+
+ reg = readl(reg_base + CQSPI_REG_CONFIG);
+ reg &= ~(CQSPI_REG_CONFIG_BAUD_MASK << CQSPI_REG_CONFIG_BAUD_LSB);
+ div = (div & CQSPI_REG_CONFIG_BAUD_MASK) << CQSPI_REG_CONFIG_BAUD_LSB;
+ writel(reg, reg_base + CQSPI_REG_CONFIG);
+}
+
+static void cadence_qspi_cs(void *reg_base, u32 chip_select, u32 decoder_enable)
+{
+ unsigned int reg;
+
+ cadence_qspi_controller_enable(reg_base, 0);
+
+ reg = readl(reg_base + CQSPI_REG_CONFIG);
+ /* docoder */
s/docoder/decoder
correct, the typo error
+ if (decoder_enable) {
+ reg |= CQSPI_REG_CONFIG_DECODE_MASK;
+ } else {
+ reg &= ~CQSPI_REG_CONFIG_DECODE_MASK;
+ chip_select = 0xF & ~(1 << chip_select);
+ }
+ reg &= ~(CQSPI_REG_CONFIG_CHIPSELECT_MASK
+ << CQSPI_REG_CONFIG_CHIPSELECT_LSB);
+ reg |= (chip_select & CQSPI_REG_CONFIG_CHIPSELECT_MASK)
+ << CQSPI_REG_CONFIG_CHIPSELECT_LSB;
+ writel(reg, reg_base + CQSPI_REG_CONFIG);
+ cadence_qspi_controller_enable(reg_base, 1);
+}
+
+static int cadence_qspi_exec_flash_cmd(void *reg_base, unsigned int reg)
+{
+ unsigned int timeout;
+
+ writel(reg, reg_base + CQSPI_REG_CMDCTRL);
+ reg |= CQSPI_REG_CMDCTRL_EXECUTE_MASK;
+ writel(reg, reg_base + CQSPI_REG_CMDCTRL);
+
+ timeout = cadence_qspi_init_timeout(CQSPI_TIMEOUT_MS);
+ while (cadence_qspi_check_timeout(timeout)) {
+ reg = readl(reg_base + CQSPI_REG_CMDCTRL) &
+ CQSPI_REG_CMDCTRL_INPROGRESS_MASK;
+ if (!reg)
+ break;
+ }
+
Please use cqspi_wait_for_bit() that uses readl_relaxed_poll_timeout()
here. There are couple corner cases that above code does not take care of.
sure, fix it
+ if (!cadence_qspi_wait_idle(reg_base) || reg != 0)
+ return -EIO;
+
+ return 0;
+}
+
+static int cqspi_command_read(void *reg_base, u32 txlen, const u_char *txbuf,
+ u32 addrlen, const u8 *addrbuf, u32 rxlen,
+ u8 *rxbuf, u32 flash_type)
+{
+ u32 addr_value, read_len, reg;
+ int ret;
+
+ if (!rxlen || rxlen > CQSPI_STIG_DATA_LEN_MAX || !rxbuf)
+ return -EINVAL;
+
+ reg = txbuf[0] << CQSPI_REG_CMDCTRL_OPCODE_LSB;
+ reg |= BIT(CQSPI_REG_CMDCTRL_RD_EN_LSB);
+
+ if (addrlen) {
+ reg |= BIT(CQSPI_REG_CMDCTRL_ADDR_EN_LSB);
+ reg |= ((addrlen - 1) & CQSPI_REG_CMDCTRL_ADD_BYTES_MASK)
+ << CQSPI_REG_CMDCTRL_ADD_BYTES_LSB;
+ addr_value = cadence_qspi_cmd2addr(&addrbuf[0], addrlen);
+ writel(addr_value, reg_base + CQSPI_REG_CMDADDRESS);
+ }
+ /* 0 means 1 byte. */
+ reg |= (((rxlen - 1) & CQSPI_REG_CMDCTRL_RD_BYTES_MASK)
+ << CQSPI_REG_CMDCTRL_RD_BYTES_LSB);
+ ret = cadence_qspi_exec_flash_cmd(reg_base, reg);
+ if (ret != 0)
+ return ret;
+
+ reg = readl(reg_base + CQSPI_REG_CMDREADDATALOWER);
+ read_len = (rxlen > 4) ? 4 : rxlen;
+ memcpy(rxbuf, &reg, read_len);
+ rxbuf += read_len;
+
+ if (rxlen > 4) {
+ reg = readl(reg_base + CQSPI_REG_CMDREADDATAUPPER);
+ read_len = rxlen - read_len;
+ memcpy(rxbuf, &reg, read_len);
+ }
+
+ return 0;
+}
+
+static int cqspi_command_write(void *reg_base, u32 txlen, const u8 *txbuf,
+ u32 addrlen, const u8 *addrbuf, u32 datalen,
+ const u8 *databuf, u32 flash_type)
+{
+ unsigned int addr_value, reg, data = 0;
+
+ if (!txlen || !txbuf)
+ return -EINVAL;
+
+ reg = txbuf[0] << CQSPI_REG_CMDCTRL_OPCODE_LSB;
+ if (datalen != 0) {
+ reg |= BIT(CQSPI_REG_CMDCTRL_WR_EN_LSB);
+ reg |= ((datalen - 1) & CQSPI_REG_CMDCTRL_WR_BYTES_MASK)
+ << CQSPI_REG_CMDCTRL_WR_BYTES_LSB;
+ memcpy(&data, databuf, datalen);
+ writel(data, reg_base + CQSPI_REG_CMDWRITEDATALOWER);
+ }
+ if (addrlen) {
+ reg |= BIT(CQSPI_REG_CMDCTRL_ADDR_EN_LSB);
+ reg |= ((addrlen - 1) & CQSPI_REG_CMDCTRL_ADD_BYTES_MASK)
+ << CQSPI_REG_CMDCTRL_ADD_BYTES_LSB;
+ addr_value = cadence_qspi_cmd2addr(&addrbuf[0], addrlen);
+ writel(addr_value, reg_base + CQSPI_REG_CMDADDRESS);
+ }
+
+ return cadence_qspi_exec_flash_cmd(reg_base, reg);
+}
+
+static int cqspi_indirect_read_setup(void *reg_base, u32 txlen, const u8 *txbuf,
+ u32 addrlen, const u8 *addrbuf,
+ u32 dummy_bytes, u32 flash_type,
+ u32 quad_mode)
+{
+ unsigned int addr_value, dummy_clk, reg;
+
+ if (addrlen) {
+ addr_value = cadence_qspi_cmd2addr(&addrbuf[0], addrlen);
+ writel(addr_value, reg_base + CQSPI_REG_INDIRECTRDSTARTADDR);
+ }
+
+ reg = txbuf[0] << CQSPI_REG_RD_INSTR_OPCODE_LSB;
+ reg |= (quad_mode & CQSPI_REG_RD_INSTR_TYPE_DATA_MASK) <<
+ CQSPI_REG_RD_INSTR_TYPE_DATA_LSB;
+
+ if (dummy_bytes) {
+ if (dummy_bytes > CQSPI_DUMMY_BYTES_MAX)
+ dummy_bytes = CQSPI_DUMMY_BYTES_MAX;
+ reg |= BIT(CQSPI_REG_RD_INSTR_MODE_EN_LSB);
+ writel(0xFF, reg_base + CQSPI_REG_MODE_BIT);
+ dummy_clk = dummy_bytes * CQSPI_DUMMY_CLKS_PER_BYTE;
+ dummy_clk -= CQSPI_DUMMY_CLKS_PER_BYTE;
+ if (dummy_clk)
+ reg |= (dummy_clk & CQSPI_REG_RD_INSTR_DUMMY_MASK)
+ << CQSPI_REG_RD_INSTR_DUMMY_LSB;
+ } else if (flash_type == QSPI_FLASH_TYPE_NAND) {
+ reg |= ((1 * CQSPI_DUMMY_CLKS_PER_BYTE)
+ << CQSPI_REG_RD_INSTR_DUMMY_LSB);
+ }
Why do we need flash type specific handling here? Does not spi-nand
driver provide appropriate dummy_clk values? Why is the driver always
adding 8 dummy clks always?
On our board, this needs dummy clocks to work spi-nand, will differentiate and add in the init part instead of

cqspi_indirect_read_setup.

+ writel(reg, reg_base + CQSPI_REG_RD_INSTR);
+ /* Set device size */
+ reg = readl(reg_base + CQSPI_REG_SIZE);
+ reg &= ~CQSPI_REG_SIZE_ADDRESS_MASK;
+ reg |= (addrlen - 1);
+ writel(reg, reg_base + CQSPI_REG_SIZE);
+ /* disable auto-polling */
+ reg = readl(reg_base + CQSPI_REG_WR_COMPLETION_CTRL);
+ reg |= BIT(CQSPI_REG_WR_COMPLETION_DIS_POLLING_FLD_POS);
+ writel(reg, reg_base + CQSPI_REG_WR_COMPLETION_CTRL);
+
+ return 0;
+}
+
+static void cqspi_request_mmap_dma(struct struct_cqspi *cqspi)
+{
+ dma_cap_mask_t mask;
+
+ dma_cap_zero(mask);
+ dma_cap_set(DMA_MEMCPY, mask);
+
+ cqspi->rx_chan = dma_request_chan_by_mask(&mask);
+ if (IS_ERR(cqspi->rx_chan)) {
+ dev_err(&cqspi->pdev->dev, "No Rx DMA available\n");
+ cqspi->rx_chan = NULL;
+ }
+ init_completion(&cqspi->rx_dma_complete);
+}
+
+static void cqspi_rx_dma_callback(void *param)
+{
+ struct struct_cqspi *cqspi = param;
+
+ complete(&cqspi->rx_dma_complete);
+}
+
+static int cqspi_direct_read_execute(struct struct_cqspi *cqspi, u8 *buf,
+ const u8 *addrbuf, size_t rxlen)
+{
+ struct platform_device *pdev = cqspi->pdev;
+ struct device *dev = &pdev->dev;
+ enum dma_ctrl_flags flags = DMA_CTRL_ACK | DMA_PREP_INTERRUPT;
+ dma_addr_t dma_src = (dma_addr_t)cqspi->mmap_phys_base + addrbuf[0];
+ void __iomem *ahb_base = cqspi->qspi_ahb_virt;
+ int ret = 0;
+ struct dma_async_tx_descriptor *tx;
+ dma_cookie_t cookie;
+ dma_addr_t dma_dst;
+
+ if (!cqspi->rx_chan || !virt_addr_valid(buf)) {
+ memcpy_fromio(buf, ahb_base + addrbuf[0], rxlen);
+ return 0;
+ }
+
+ dma_dst = dma_map_single(dev, buf, rxlen, DMA_FROM_DEVICE);
+ if (dma_mapping_error(dev, dma_dst)) {
+ dev_err(dev, "dma mapping failed\n");
+ return -ENOMEM;
+ }
+ tx = dmaengine_prep_dma_memcpy(cqspi->rx_chan, dma_dst, dma_src,
+ rxlen, flags);
+ if (!tx) {
+ dev_err(dev, "device_prep_dma_memcpy error\n");
+ ret = -EIO;
+ goto err_unmap;
+ }
+
+ tx->callback = cqspi_rx_dma_callback;
+ tx->callback_param = cqspi;
+ cookie = tx->tx_submit(tx);
+ reinit_completion(&cqspi->rx_dma_complete);
+
+ ret = dma_submit_error(cookie);
+ if (ret) {
+ dev_err(dev, "dma_submit_error %d\n", cookie);
+ ret = -EIO;
+ goto err_unmap;
+ }
+
+ dma_async_issue_pending(cqspi->rx_chan);
+ if (!wait_for_completion_timeout(&cqspi->rx_dma_complete,
+ msecs_to_jiffies(rxlen))) {
+ dmaengine_terminate_sync(cqspi->rx_chan);
+ dev_err(dev, "DMA wait_for_completion_timeout\n");
+ ret = -ETIMEDOUT;
+ goto err_unmap;
+ }
+
+err_unmap:
+ dma_unmap_single(dev, dma_dst, rxlen, DMA_FROM_DEVICE);
+
+ return ret;
+}
+
+static int cqspi_indirect_read_execute(struct struct_cqspi *cqspi, u32 rxlen,
+ u8 *rxbuf, u32 flash_type)
+{
+ struct platform_device *pdev = cqspi->pdev;
+ struct cqspi_platform_data *pdata = pdev->dev.platform_data;
+ void *reg_base = cqspi->iobase;
+ void *ahb_base = cqspi->qspi_ahb_virt;
+ u8 *rxbuf_end = rxbuf + rxlen;
+ u32 mod_bytes = rxlen % 4;
+ u32 bytes_to_read = 0;
+ int remaining = (int)rxlen;
+ u32 timeout, reg = 0;
+ int ret = 0;
+
+ writel(0, reg_base + CQSPI_REG_INDIRECTRDWATERMARK);
+ writel(0xa, reg_base + CQSPI_INDIRECT_TRIGGER_ADDR_RANGE_REG);
+ writel(remaining, reg_base + CQSPI_REG_INDIRECTRDBYTES);
+
+ mb();/* flush previous writes */
+
+ writel(pdata->fifo_depth - CQSPI_REG_SRAM_RESV_WORDS,
+ reg_base + CQSPI_REG_SRAMPARTITION);
+ /* Clear all interrupts. */
+ writel(CQSPI_IRQ_STATUS_MASK, reg_base + CQSPI_REG_IRQSTATUS);
+ writel(CQSPI_IRQ_MASK_RD, reg_base + CQSPI_REG_IRQMASK);
+
+ reinit_completion(&cqspi->transfer_complete);
+ writel(CQSPI_REG_INDIRECTRD_START_MASK,
+ reg_base + CQSPI_REG_INDIRECTRD);
+
+ while (remaining > 0) {
+ bytes_to_read = CQSPI_GET_RD_SRAM_LEVEL(reg_base);
+
+ while (bytes_to_read != 0) {
+ unsigned int word_remain = round_down(remaining, 4);
+
+ bytes_to_read *= CQSPI_FIFO_WIDTH;
+ bytes_to_read = bytes_to_read > remaining ?
+ remaining : bytes_to_read;
+ bytes_to_read = round_down(bytes_to_read, 4);
+ if (bytes_to_read) {
+ ioread32_rep(ahb_base, rxbuf,
+ (bytes_to_read / 4));
+ } else if (!word_remain && mod_bytes) {
+ unsigned int temp = ioread32(ahb_base);
+
+ bytes_to_read = mod_bytes;
+ memcpy(rxbuf, &temp, min((unsigned int)
+ (rxbuf_end - rxbuf), bytes_to_read));
+ }
+
+ rxbuf += bytes_to_read;
+ remaining -= bytes_to_read;
+ bytes_to_read = CQSPI_GET_RD_SRAM_LEVEL(reg_base);
+ }
+
+ if (remaining < 0)
+ reinit_completion(&cqspi->transfer_complete);
+ }
+
+ /* Check indirect done status */
+ timeout = cadence_qspi_init_timeout(CQSPI_READ_TIMEOUT_MS);
+ while (cadence_qspi_check_timeout(timeout)) {
+ reg = readl(reg_base + CQSPI_REG_INDIRECTRD);
+ if (reg & CQSPI_REG_INDIRECTRD_DONE_MASK)
+ break;
+ }
+ if (!(reg & CQSPI_REG_INDIRECTRD_DONE_MASK)) {
+ dev_err(&pdev->dev, "QSPI : Indirect read error %x\n", reg);
+ ret = -ETIMEDOUT;
+ goto failrd;
+ }
+ /* Disable interrupt */
+ writel(0, reg_base + CQSPI_REG_IRQMASK);
+ /* Clear indirect completion status */
+ writel(CQSPI_REG_INDIRECTRD_DONE_MASK, reg_base + CQSPI_REG_INDIRECTRD);
+
+ return 0;
+failrd:
+ /* Disable interrupt */
+ writel(0, reg_base + CQSPI_REG_IRQMASK);
+ /* Cancel the indirect read */
+ writel(CQSPI_REG_INDIRECTWR_CANCEL_MASK,
+ reg_base + CQSPI_REG_INDIRECTRD);
+ return ret;
+}
+
+static int cqspi_indirect_write_setup(void *reg_base, u32 txlen,
+ const u8 *txbuf, u32 addrlen,
+ const u8 *addrbuf, u32 flash_type,
+ u32 quad_mode)
+{
+ unsigned int reg;
+
+ if (!txbuf)
+ return -EINVAL;
+
+ reg = readl(reg_base + CQSPI_REG_CONFIG);
+ reg &= ~(CQSPI_REG_CONFIG_DIRECT_MASK);
+ reg &= ~(CQSPI_REG_CONFIG_DMA_MASK);
+ writel(reg, reg_base + CQSPI_REG_CONFIG);
+
+ /* Set opcode. */
+ reg = txbuf[0] << CQSPI_REG_WR_INSTR_OPCODE_LSB;
+ reg |= BIT(CQSPI_REG_WR_CONFIG_WEL_DIS_FLD_POS);
+ if (txbuf[0] == MACRONIX_4PP_CMD && flash_type == QSPI_FLASH_TYPE_NOR &&
+ quad_mode == QUAD_QIO) {
+ /* Configure the quad */
+ reg |= (quad_mode & CQSPI_REG_WR_INSTR_TYPE_ADDR_MASK) <<
+ CQSPI_REG_WR_INSTR_TYPE_ADDR_LSB;
+ }
Nope, no custom handling or interpreting of opcodes in driver.
Just set addrwidth to quad in CQSPI_REG_WR_INSTR when op->addr.buswidth
is 4.
Agreed!, add it in the next patch.

+ /* Configure the quad */
+ reg |= (quad_mode & CQSPI_REG_WR_INSTR_TYPE_DATA_MASK) <<
+ CQSPI_REG_WR_INSTR_TYPE_DATA_LSB;
+ writel(reg, reg_base + CQSPI_REG_WR_INSTR);
+ /* Setup write address. */
+ reg = cadence_qspi_cmd2addr(&addrbuf[0], addrlen);
+ writel(reg, reg_base + CQSPI_REG_INDIRECTWRSTARTADDR);
+ reg = readl(reg_base + CQSPI_REG_SIZE);
+ reg &= ~CQSPI_REG_SIZE_ADDRESS_MASK;
+ reg |= ((addrlen - 1) & CQSPI_REG_SIZE_ADDRESS_MASK);
+ writel(reg, reg_base + CQSPI_REG_SIZE);
+
+ /* disable auto-polling */
+ reg = readl(reg_base + CQSPI_REG_WR_COMPLETION_CTRL);
+ reg |= BIT(CQSPI_REG_WR_COMPLETION_DIS_POLLING_FLD_POS);
+ writel(reg, reg_base + CQSPI_REG_WR_COMPLETION_CTRL);
+
+ return 0;
+}
+
+static int cqspi_indirect_write_execute(struct struct_cqspi *cqspi, u32 txlen,
+ const u8 *txbuf, u32 flash_type)
+{
+ struct platform_device *pdev = cqspi->pdev;
+ struct cqspi_platform_data *pdata = pdev->dev.platform_data;
+ struct cqspi_flash_pdata *f_pdata =
+ &pdata->f_pdata[cqspi->current_cs];
+ void *reg_base = cqspi->iobase;
+ void *ahb_base = cqspi->qspi_ahb_virt;
+ u32 *irq_status = &cqspi->irq_status;
+ u32 page_size = f_pdata->page_size;
+ u32 write_bytes, timeout, reg = 0;
+ int remaining = (int)txlen;
+ int ret;
+
+ writel(0xa, reg_base + CQSPI_INDIRECT_TRIGGER_ADDR_RANGE_REG);
+ writel(0x0, reg_base + CQSPI_REG_INDIRECTWRWATERMARK);
+ reg = readl(reg_base + CQSPI_REG_SIZE);
+ reg &= ~(CQSPI_REG_SIZE_PAGE_MASK << CQSPI_REG_SIZE_PAGE_LSB);
+ reg &= ~(CQSPI_REG_SIZE_BLOCK_MASK << CQSPI_REG_SIZE_BLOCK_LSB);
+ reg |= (f_pdata->page_size << CQSPI_REG_SIZE_PAGE_LSB);
+ reg |= (f_pdata->block_size << CQSPI_REG_SIZE_BLOCK_LSB);
+ writel(reg, reg_base + CQSPI_REG_SIZE);
+
+ writel(remaining, reg_base + CQSPI_REG_INDIRECTWRBYTES);
+ writel(CQSPI_REG_SRAM_PARTITION_WR, reg_base + CQSPI_REG_SRAMPARTITION);
+ /* Clear all interrupts. */
+ writel(CQSPI_IRQ_STATUS_MASK, reg_base + CQSPI_REG_IRQSTATUS);
+ writel(CQSPI_IRQ_MASK_WR, reg_base + CQSPI_REG_IRQMASK);
+ reinit_completion(&cqspi->transfer_complete);
+ writel(CQSPI_REG_INDIRECTWR_START_MASK,
+ reg_base + CQSPI_REG_INDIRECTWR);
+
+ if (cqspi->wr_delay)
+ ndelay(cqspi->wr_delay);
+
+ while (remaining > 0) {
+ size_t write_words, mod_bytes;
+
+ write_bytes = remaining > page_size ? page_size : remaining;
+ write_words = write_bytes / 4;
+ mod_bytes = write_bytes % 4;
+
+ if (write_words) {
+ iowrite32_rep(ahb_base, txbuf, write_words);
+ txbuf += (write_words * 4);
+ }
+ if (mod_bytes) {
+ unsigned int temp = 0xFFFFFFFF;
+
+ memcpy(&temp, txbuf, mod_bytes);
+ iowrite32(temp, ahb_base);
+ txbuf += mod_bytes;
+ }
+ ret = wait_event_interruptible_timeout(cqspi->waitqueue,
+ *irq_status &
+ CQSPI_IRQ_MASK_WR,
+ CQSPI_TIMEOUT_MS);
This was wait_for_completion_timeout() previously. Have you tested
interrupting the sleep and verified that driver handles this case correctly?
Yes I have tested, align with

wait_for_completion_timeout

+ if (!(ret) & (*irq_status & CQSPI_IRQ_STATUS_ERR)) {
+ ret = -ETIMEDOUT;
+ goto failwr;
+ } else {
+ ret = 0;
+ }
+ remaining -= write_bytes;
+
+ if (remaining < 0)
+ reinit_completion(&cqspi->transfer_complete);
+ }
+
+ /* Check indirect done status */
+ timeout = cadence_qspi_init_timeout(CQSPI_TIMEOUT_MS);
+ while (cadence_qspi_check_timeout(timeout)) {
+ reg = readl(reg_base + CQSPI_REG_INDIRECTWR);
+ if (reg & CQSPI_REG_INDIRECTWR_DONE_MASK)
+ break;
+ }
+ if (!(reg & CQSPI_REG_INDIRECTWR_DONE_MASK)) {
+ dev_err(&pdev->dev, "QSPI: Indirect write error %x", reg);
+ ret = -ETIMEDOUT;
+ goto failwr;
+ }
+
+ return 0;
+
+failwr:
+ /* Disable interrupt. */
+ writel(0, reg_base + CQSPI_REG_IRQMASK);
+ /* Clear indirect completion status */
+ writel(CQSPI_REG_INDIRECTWR_DONE_MASK, reg_base + CQSPI_REG_INDIRECTWR);
+
+ /* Cancel the indirect write */
+ if (ret)
+ writel(CQSPI_REG_INDIRECTWR_CANCEL_MASK,
+ reg_base + CQSPI_REG_INDIRECTWR);
+
+ return ret;
+}
+
+unsigned int cadence_qspi_is_controller_ready(void *reg_base)
+{
+ return cadence_qspi_wait_idle(reg_base);
+}
+
+void cadence_qspi_controller_init(struct struct_cqspi *cqspi)
+{
+ struct platform_device *pdev = cqspi->pdev;
+ struct cqspi_platform_data *pdata = pdev->dev.platform_data;
+
+ cadence_qspi_controller_enable(cqspi->iobase, 0);
+
+ /* Configure the remap address register, no remap */
+ writel(0, cqspi->iobase + CQSPI_REG_REMAP);
+ /* Disable all interrupts. */
+ writel(0, cqspi->iobase + CQSPI_REG_IRQMASK);
+
+ /* DAC is disabled for Intel LGM SOC */
+ if (!cqspi->dac_mode) {
+ enable_qspi_direct_access(cqspi->iobase, 0);
+ } else {
+ enable_qspi_direct_access(cqspi->iobase, 1);
+ /* TODO: for TI platform to be enabled */
+ cqspi_request_mmap_dma(cqspi);
+ }
+
+ /* Load indirect trigger address. */
+ writel(pdata->trigger_address,
+ cqspi->iobase + CQSPI_REG_INDIRECTTRIGGER);
+
+ cadence_qspi_controller_enable(cqspi->iobase, 1);
+}
+
+unsigned int calculate_ticks_for_ns(u32 ref_clk_hz, u32 ns_val)
+{
+ unsigned int ticks;
+
+ ticks = ref_clk_hz / 1000; /* kHz */
+ ticks = DIV_ROUND_UP(ticks * ns_val, 1000000);
+
+ return ticks;
+}
+
+void cadence_qspi_delay(struct struct_cqspi *cqspi, u32 ref_clk, u32 sclk_hz)
+{
+ struct platform_device *pdev = cqspi->pdev;
+ struct cqspi_platform_data *pdata = pdev->dev.platform_data;
+ struct cqspi_flash_pdata *f_pdata = &pdata->f_pdata[cqspi->current_cs];
+ void __iomem *iobase = cqspi->iobase;
+ const unsigned int ref_clk_hz = pdata->master_ref_clk_hz;
+ unsigned int tshsl, tchsh, tslch, tsd2d;
+ unsigned int reg;
+ unsigned int tsclk;
+
+ cadence_qspi_controller_enable(cqspi->iobase, 0);
+ /* calculate the number of ref ticks for one sclk tick */
+ tsclk = DIV_ROUND_UP(ref_clk_hz, sclk_hz);
+
+ /* this particular value must be at least one sclk */
+ if (tshsl < tsclk)
+ tshsl = tsclk;
+
+ tchsh = calculate_ticks_for_ns(ref_clk_hz, f_pdata->tchsh_ns);
+ tslch = calculate_ticks_for_ns(ref_clk_hz, f_pdata->tslch_ns);
+ tsd2d = calculate_ticks_for_ns(ref_clk_hz, f_pdata->tsd2d_ns);
+
+ reg = ((tshsl & CQSPI_REG_DELAY_TSHSL_MASK)
+ << CQSPI_REG_DELAY_TSHSL_LSB);
+ reg |= ((tchsh & CQSPI_REG_DELAY_TCHSH_MASK)
+ << CQSPI_REG_DELAY_TCHSH_LSB);
+ reg |= ((tslch & CQSPI_REG_DELAY_TSLCH_MASK)
+ << CQSPI_REG_DELAY_TSLCH_LSB);
+ reg |= ((tsd2d & CQSPI_REG_DELAY_TSD2D_MASK)
+ << CQSPI_REG_DELAY_TSD2D_LSB);
+ writel(reg, iobase + CQSPI_REG_DELAY);
+ cadence_qspi_controller_enable(cqspi->iobase, 1);
+}
+
+void cadence_qspi_switch_chipselect(struct struct_cqspi *cqspi, u32 cs)
+{
+ struct platform_device *pdev = cqspi->pdev;
+ struct cqspi_platform_data *pdata = pdev->dev.platform_data;
+ struct cqspi_flash_pdata *f_pdata = &pdata->f_pdata[cs];
+ void __iomem *iobase = cqspi->iobase;
+ unsigned int reg;
+
+ cadence_qspi_controller_enable(cqspi->iobase, 0);
+ /* Configure page size and block size. */
+ reg = readl(iobase + CQSPI_REG_SIZE);
+ /* clear the previous value */
+ reg &= ~(CQSPI_REG_SIZE_PAGE_MASK << CQSPI_REG_SIZE_PAGE_LSB);
+ reg &= ~(CQSPI_REG_SIZE_BLOCK_MASK << CQSPI_REG_SIZE_BLOCK_LSB);
+ reg |= (f_pdata->page_size << CQSPI_REG_SIZE_PAGE_LSB);
+ reg |= (f_pdata->block_size << CQSPI_REG_SIZE_BLOCK_LSB);
+ writel(reg, iobase + CQSPI_REG_SIZE);
+ /* configure the chip select */
+ cadence_qspi_cs(iobase, cs, pdata->ext_decoder);
+ cadence_qspi_controller_enable(cqspi->iobase, 1);
+}
+
+static int cadence_qspi_mem_process(struct struct_cqspi *cqspi,
+ struct spi_mem *mem,
+ const struct spi_mem_op *op)
+{
+ struct platform_device *pdev = cqspi->pdev;
+ struct cqspi_platform_data *pdata = pdev->dev.platform_data;
+ unsigned int tmpbufsize, n_trans = 0, totalxferlen = 0;
+ void __iomem *iobase = cqspi->iobase;
+ struct spi_mem_op_cadence ops[4] = { };
+ struct spi_mem_op_cadence *cmd_ops = NULL;
+ struct spi_mem_op_cadence *data_ops = NULL;
+ struct spi_mem_op_cadence *dummy_ops = NULL;
+ struct spi_mem_op_cadence *addr_ops = NULL;
+ struct cqspi_flash_pdata *f_pdata;
+ int rdid_length, mode, ret, i;
+ u8 *tmpbuf;
+ u32 sclk;
+
+ if (cqspi->current_cs != mem->spi->chip_select) {
+ cqspi->current_cs = mem->spi->chip_select;
+ cadence_qspi_switch_chipselect(cqspi, mem->spi->chip_select);
+ }
+
+ f_pdata = &pdata->f_pdata[cqspi->current_cs];
+ if (f_pdata->flash_type == QSPI_FLASH_TYPE_NOR)
+ rdid_length = 6;
+ else
+ rdid_length = 3;
+
What? NACK... why should driver worry about rdid_length?
sorry forgot remove it.

+ tmpbufsize = sizeof(op->cmd.opcode) + op->addr.nbytes +
+ op->dummy.nbytes;
+
+ tmpbuf = kzalloc(tmpbufsize, GFP_KERNEL | GFP_DMA);
+ if (!tmpbuf)
+ return -ENOMEM;
+
+ tmpbuf[0] = op->cmd.opcode;
+ ops[n_trans].tx_buf = tmpbuf;
+ ops[n_trans].len = sizeof(op->cmd.opcode);
+ ops[n_trans].tx_nbits = op->cmd.buswidth;
+
+ n_trans++;
+ totalxferlen++;
+
+ if (op->addr.nbytes) {
+ int i;
+
+ for (i = 0; i < op->addr.nbytes; i++)
+ tmpbuf[i + 1] = op->addr.val >>
+ (8 * (op->addr.nbytes - i - 1));
+
+ ops[n_trans].tx_buf = tmpbuf + 1;
+ ops[n_trans].len = op->addr.nbytes;
+ ops[n_trans].tx_nbits = op->addr.buswidth;
+
+ n_trans++;
+ totalxferlen += op->addr.nbytes;
+ }
+ if (op->dummy.nbytes) {
+ memset(tmpbuf + op->addr.nbytes + 1, 0xff, op->dummy.nbytes);
+ ops[n_trans].tx_buf = tmpbuf + op->addr.nbytes + 1;
+ ops[n_trans].len = op->dummy.nbytes;
+ ops[n_trans].tx_nbits = op->dummy.buswidth;
+
+ n_trans++;
+ totalxferlen += op->dummy.nbytes;
+ }
+ if (op->data.nbytes) {
+ if (op->data.dir == SPI_MEM_DATA_IN) {
+ ops[n_trans].rx_buf = op->data.buf.in;
+ ops[n_trans].rx_nbits = op->data.buswidth;
+ } else {
+ ops[n_trans].tx_buf = op->data.buf.out;
+ ops[n_trans].tx_nbits = op->data.buswidth;
+ }
+
+ ops[n_trans].len = op->data.nbytes;
+ n_trans++;
+ totalxferlen += op->data.nbytes;
+ }
+
+ for (i = 0; i < n_trans; i++)
+ dev_dbg(&pdev->dev, "ops[%d] %d\n", i, ops[i].len);
+
+ switch (n_trans) {
+ case 1:
+ cmd_ops = &ops[0];
+ break;
+ case 2:
+ cmd_ops = &ops[0];
+ if (f_pdata->flash_type == QSPI_FLASH_TYPE_NAND) {
+ if (data_ops->rx_buf)
+ data_ops = &ops[1];
+ else
+ addr_ops = &ops[1];
+ } else {
+ data_ops = &ops[1];
+ }

Please get ride of all flash type specific handling... There should be
no need for it. IF there are gaps in then lets discuss and fix it in
spi-nand and spi-nor frameworks and not in the driver.

spi-mem user should just take spi_mem_op template passed by the core and
execute it. No assumptions wrt flash type or opcode should be done.
sure, will take care flash type specific handling.
+ break;
+ case 3:
+ cmd_ops = &ops[0];
+ addr_ops = &ops[1];
+ data_ops = &ops[2];
+ break;
+ case 4:
+ cmd_ops = &ops[0];
+ addr_ops = &ops[1];
+ dummy_ops = &ops[2];
+ data_ops = &ops[3];
+ break;
+ default:
+ dev_err(&pdev->dev, "Unsupported n_trans %u\n", n_trans);
+ return -EINVAL;
+ }
+
+ if (!data_ops)
+ mode = STIG_WRITE_MODE;
+ else if (data_ops->len > rdid_length)
+ mode = data_ops->tx_buf ? IDC_WRITE_MODE : IDC_READ_MODE;
+ else
+ mode = data_ops->tx_buf ? STIG_WRITE_MODE : STIG_READ_MODE;
+
Decision to use STIG was INDAC mode can be done using:

if (op->data.dir == SPI_MEM_DATA_IN && op->data.buf.in) {
if (!op->addr.nbytes)
mode = CQSPI_STIG_READ;
else
mode = IDC_READ_MODE;
} else {
if (!op->addr.nbytes || !op->data.buf.out)
mode = CQSPI_STIG_WRITE;
else
mode = IDC_WRITE_MODE;
}

No need for rdid_length etc.
Thanks for the suggestion, i will update in the next patch-set.
+ sclk = mem->spi->max_speed_hz;
+ cadence_qspi_controller_enable(iobase, 0);
+ cadence_qspi_config_baudrate_div(iobase, pdata->master_ref_clk_hz,
+ sclk);
+ cadence_qspi_delay(cqspi, pdata->master_ref_clk_hz, sclk);
+ cadence_qspi_readdata_capture(iobase, 1, f_pdata->read_delay);
+ cadence_qspi_controller_enable(iobase, 1);
+
+ /* execute transfer */
+ switch (mode) {
+ case STIG_WRITE_MODE:
+ ret = cqspi_command_write(iobase, cmd_ops->len, cmd_ops->tx_buf,
+ addr_ops ? addr_ops->len : 0,
+ addr_ops ? addr_ops->tx_buf : NULL,
+ data_ops ? data_ops->len : 0,
+ data_ops ? data_ops->tx_buf : NULL,
+ f_pdata->flash_type);
+ break;
+ case STIG_READ_MODE:
+ ret = cqspi_command_read(iobase, cmd_ops->len, cmd_ops->tx_buf,
+ addr_ops ? addr_ops->len : 0,
+ addr_ops ? addr_ops->tx_buf : NULL,
+ data_ops ? data_ops->len : 0,
+ data_ops ? data_ops->rx_buf : NULL,
+ f_pdata->flash_type);
+ break;
+ case IDC_WRITE_MODE:
+ ret = cqspi_indirect_write_setup(iobase, cmd_ops->len,
+ cmd_ops->tx_buf,
+ addr_ops ? addr_ops->len : 0,
+ addr_ops ? addr_ops->tx_buf : 0,
+ f_pdata->flash_type,
+ f_pdata->quad);
+ ret = cqspi_indirect_write_execute(cqspi, data_ops->len,
+ data_ops->tx_buf,
+ f_pdata->flash_type);
+ break;
+ case IDC_READ_MODE:
+ ret = cqspi_indirect_read_setup(iobase, cmd_ops->len,
+ cmd_ops->tx_buf,
+ addr_ops ? addr_ops->len : 0,
+ addr_ops ? addr_ops->tx_buf : 0,
+ dummy_ops ? dummy_ops->len : 0,
+ f_pdata->flash_type,
+ f_pdata->quad);
+ ret = cqspi_indirect_read_execute(cqspi, data_ops->len,
+ data_ops->rx_buf,
+ f_pdata->flash_type);
+ /* TODO: for TI DMA + DAC to be enabled */
This is a big gap and there is lot of code sharing that can be done b/w
INDAC and DAC mode.

Looks like driver isn't quite close to being ready. I suggest you also
take a look at patches for spi-mem conversion of cadence-quadspi driver
in U-Boot as well:

https://patchwork.ozlabs.org/cover/1176362/

driver works perfectly on our platform, since we do not have DMA , just left the code.

I will add those part of code in the next series of patch.

+ break;
+ default:
+ dev_err(&pdev->dev, "Unsupported mode %u\n", mode);
+ return -EINVAL;
+ }
+
+ return ret;
+}
+
+int cadence_qspi_exec_mem_op(struct spi_mem *mem, const struct spi_mem_op *op)
+{
+ struct struct_cqspi *cqspi = spi_master_get_devdata(mem->spi->master);
+ int ret;
+
+ mutex_lock(&cqspi->lock);
+ ret = cadence_qspi_mem_process(cqspi, mem, op);
+ if (ret)
+ dev_err(&mem->spi->dev, "Error: qspi transfer failed!!!.\n");
+ mutex_unlock(&cqspi->lock);
+
+ return ret;
+}
+
+static const struct spi_controller_mem_ops cadence_qspi_mem_ops = {
+ .exec_op = cadence_qspi_exec_mem_op,
+};
+
+static int cadence_qspi_setup(struct spi_device *spi)
+{
+ if (spi->chip_select > spi->master->num_chipselect) {
+ dev_err(&spi->dev, "%d CS is out of range\n", spi->chip_select);
+ return -EINVAL;
+ }
+ return 0;
+}
+
+static int cadence_qspi_of_get_pdata(struct platform_device *pdev)
+{
+ struct device_node *np = pdev->dev.of_node;
+ struct device_node *nc;
+ struct cqspi_platform_data *pdata = pdev->dev.platform_data;
+ struct cqspi_flash_pdata *f_pdata;
+ unsigned int cs;
+ unsigned int prop;
+
+ pdata->is_decoded_cs = of_property_read_bool(np, "cdns,is-decoded-cs");
+
+ if (of_property_read_u32(np, "cdns,trigger-address", &prop)) {
+ dev_err(&pdev->dev, "couldn't determine trigger-address\n");
+ return -ENXIO;
+ }
+ pdata->trigger_address = prop;
+
+ pdata->rclk_en = of_property_read_bool(np, "cdns,rclk-en");
+
+ if (of_property_read_u32(np, "cdns,qspi-phyaddr", &prop)) {
+ dev_err(&pdev->dev, "couldn't determine cdns,qspi-phyaddr\n");
+ return -ENXIO;
+ }
+ pdata->qspi_ahb_phy = prop;
+
Please dont add any new DT properties, there is no need for them. See
how this was handled in existing driver and reuse it
Okay

+ if (of_property_read_u32(np, "cdns,qspi-physize", &prop)) {
+ dev_info(&pdev->dev, "missing cdns,qspi-physize to 128 kB\n");
+ prop = 128;
+ }
+ pdata->qspi_ahb_size = prop * 1024;
+
+ if (of_property_read_u32(np, "cdns,qspi-phymask", &prop))
+ prop = CQSPI_INDIRECTTRIGGER_ADDR_MASK;
+ pdata->qspi_ahb_mask = prop;
+
+ if (of_property_read_u32(np, "bus-num", &prop)) {
+ dev_err(&pdev->dev, "couldn't determine bus-num\n");
+ return -ENXIO;
+ }
+ pdata->bus_num = prop;
+
+ if (of_property_read_u32(np, "num-chipselect", &prop)) {
+ dev_err(&pdev->dev, "couldn't determine num-chipselect\n");
+ return -ENXIO;
+ }
+ pdata->num_chipselect = prop;
+
+ if (of_property_read_u32(np, "cdns,fifo-depth", &prop)) {
+ dev_err(&pdev->dev, "couldn't determine fifo-depth\n");
+ return -ENXIO;
+ }
+ pdata->fifo_depth = prop;
+
+ if (of_property_read_u32(np, "ext-decoder", &prop)) {
+ dev_err(&pdev->dev, "couldn't determine ext-decoder\n");
+ return -ENXIO;
+ }
+ pdata->ext_decoder = prop;
+


+ pdata->enable_dma = of_property_read_bool(np, "dmas");
+ dev_info(&pdev->dev, "DMA %senabled\n", pdata->enable_dma ? "" : "NOT");
+
Slave DMA was never supported and is not needed.
remove it.
+ /* Get flash devices platform data */
+ for_each_child_of_node(np, nc) {
+ if (!of_device_is_available(nc))
+ continue;
+
+ if (of_property_read_u32(nc, "reg", &cs)) {
+ dev_err(&pdev->dev, "couldn't determine reg\n");
+ return -ENXIO;
+ }
+ f_pdata = &pdata->f_pdata[cs];
+
+ /* use compatible string to determine flash type */
+ if (of_device_is_compatible(nc, "spinand"))
+ f_pdata->flash_type = QSPI_FLASH_TYPE_NAND;
+ else
+ f_pdata->flash_type = QSPI_FLASH_TYPE_NOR;
+
+ if (of_property_read_u32(nc, "spi-rx-bus-width", &prop)) {
+ dev_err(&pdev->dev, "couldn't determine quad\n");
+ return -ENXIO;
+ }
+ f_pdata->quad = (prop == 4) ? QUAD_QIO : QUAD_SIO;
+
+ if (of_property_read_u32(nc, "page-size", &prop)) {
+ dev_err(&pdev->dev, "couldn't determine page-size\n");
+ return -ENXIO;
+ }
+ f_pdata->page_size = prop;
+
+ if (of_property_read_u32(nc, "block-size", &prop)) {
+ dev_err(&pdev->dev, "couldn't determine block-size\n");
+ return -ENXIO;
+ }
+ f_pdata->block_size = prop;
+
+ if (of_property_read_u32(nc, "cdns,read-delay", &prop)) {
+ dev_err(&pdev->dev, "couldn't determine read-delay\n");
+ return -ENXIO;
+ }
+ f_pdata->read_delay = prop;
+
+ if (of_property_read_u32(nc, "cdns,tshsl-ns", &prop)) {
+ dev_err(&pdev->dev, "couldn't determine tshsl-ns\n");
+ return -ENXIO;
+ }
+ f_pdata->tshsl_ns = prop;
+
+ if (of_property_read_u32(nc, "cdns,tsd2d-ns", &prop)) {
+ dev_err(&pdev->dev, "couldn't determine tsd2d-ns\n");
+ return -ENXIO;
+ }
+ f_pdata->tsd2d_ns = prop;
+
+ if (of_property_read_u32(nc, "cdns,tchsh-ns", &prop)) {
+ dev_err(&pdev->dev, "couldn't determine tchsh-ns\n");
+ return -ENXIO;
+ }
+ f_pdata->tchsh_ns = prop;
+
+ if (of_property_read_u32(nc, "cdns,tslch-ns", &prop)) {
+ dev_err(&pdev->dev, "couldn't determine tslch-ns\n");
+ return -ENXIO;
+ }
+ f_pdata->tslch_ns = prop;
+ }
+ return 0;
+}
+
+static int cadence_qspi_probe(struct platform_device *pdev)
+{
+ struct cqspi_platform_data *pdata;
+ struct device *dev = &pdev->dev;
+ struct struct_cqspi *cqspi;
+ struct spi_master *master;
+ struct reset_control *rstc, *rstc_ocp;
+ const struct cqspi_driver_platdata *ddata;
+ struct resource *res = NULL;
+ int ret;
+
+ master = spi_alloc_master(&pdev->dev, sizeof(*cqspi));
+ if (!master) {
+ dev_err(&pdev->dev, "spi_alloc_master failed\n");
+ return -ENOMEM;
+ }
+ master->mode_bits = SPI_CS_HIGH | SPI_CPOL | SPI_CPHA | SPI_TX_QUAD |
+ SPI_RX_QUAD | SPI_TX_DUAL | SPI_RX_DUAL;
Octal?
Add it next patch .
+ master->setup = cadence_qspi_setup;
+ master->mem_ops = &cadence_qspi_mem_ops;
+ master->dev.of_node = pdev->dev.of_node;
+ cqspi = spi_master_get_devdata(master);
+ cqspi->pdev = pdev;
+
+ pdata = kmalloc(sizeof(*pdata), GFP_KERNEL);
+ if (!pdata) {
+ ret = -ENOMEM;
+ goto err_pdata;
+ }
+ pdev->dev.platform_data = pdata;
+
+ cqspi->clk = devm_clk_get(&pdev->dev, "qspi");
+ if (IS_ERR(cqspi->clk)) {
+ dev_err(&pdev->dev, "cannot get qspi clk\n");
+ return PTR_ERR(cqspi->clk);
+ }
+ pdata->master_ref_clk_hz = clk_get_rate(cqspi->clk);
+
+ ret = clk_prepare_enable(cqspi->clk);
+ if (ret < 0) {
+ dev_err(&pdev->dev, "failed to enable qspi clock: %d\n", ret);
+ return ret;
+ }
+
+ ret = cadence_qspi_of_get_pdata(pdev);
+ if (ret) {
+ dev_err(&pdev->dev, "Get platform data failed.\n");
+ return -ENODEV;
+ }
+
+ cqspi->res = res;
+ /* Obtain and remap controller address. */
+ res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+ cqspi->iobase = devm_ioremap_resource(&pdev->dev, res);
+ if (IS_ERR(cqspi->iobase)) {
+ dev_err(dev, "Cannot remap controller address.\n");
+ return PTR_ERR(cqspi->iobase);
+ }
+
+ /* Obtain and remap AHB address. */
+ res = platform_get_resource(pdev, IORESOURCE_MEM, 1);
+ cqspi->qspi_ahb_virt = devm_ioremap_resource(dev, res);
+ if (IS_ERR(cqspi->qspi_ahb_virt)) {
+ dev_err(dev, "Cannot remap AHB address.\n");
+ return PTR_ERR(cqspi->qspi_ahb_virt);
+ }
+ cqspi->mmap_phys_base = (dma_addr_t)res->start;
+ cqspi->ahb_size = resource_size(res);
+
+ /* Obtain QSPI reset control */
+ rstc = devm_reset_control_get_optional_exclusive(dev, "qspi");
+ if (IS_ERR(rstc)) {
+ dev_err(dev, "Cannot get QSPI reset.\n");
+ return PTR_ERR(rstc);
+ }
+
+ rstc_ocp = devm_reset_control_get_optional_exclusive(dev, "qspi-ocp");
+ if (IS_ERR(rstc_ocp)) {
+ dev_err(dev, "Cannot get QSPI OCP reset.\n");
+ return PTR_ERR(rstc_ocp);
+ }
+
+ reset_control_assert(rstc);
+ reset_control_deassert(rstc);
+
+ reset_control_assert(rstc_ocp);
+ reset_control_deassert(rstc_ocp);
+
+ ddata = of_device_get_match_data(dev);
+ if (ddata && (ddata->quirks & CQSPI_NEEDS_WR_DELAY))
+ cqspi->wr_delay = 5 * DIV_ROUND_UP(NSEC_PER_SEC,
+ pdata->master_ref_clk_hz);
+
+ init_completion(&cqspi->transfer_complete);
+
+ cqspi->irq = platform_get_irq(pdev, 0);
+ if (cqspi->irq < 0) {
+ dev_err(dev, "platform_get_irq failed.\n");
+ ret = -ENXIO;
+ goto err_irq;
+ }
+ ret = devm_request_irq(dev, cqspi->irq, cadence_qspi_irq_handler, 0,
+ pdev->name, cqspi);
+ if (ret) {
+ dev_err(dev, "request_irq failed.\n");
+ goto err_irq;
+ }
+
+ master->bus_num = pdata->bus_num;
+ master->num_chipselect = pdata->num_chipselect;
+ mutex_init(&cqspi->lock);
+ platform_set_drvdata(pdev, master);
+ cadence_qspi_controller_init(cqspi);
+ cqspi->current_cs = 0;
+
+ ret = devm_spi_register_master(dev, master);
+ if (ret) {
+ dev_err(&pdev->dev, "devm_spi_register_master failed.\n");
+ goto err_of;
+ }
+
+ if (ddata && (ddata->quirks & CQSPI_DISABLE_DAC_MODE))
+ cqspi->dac_mode = false;
+
+ return 0;
+
+err_pdata:
+ kfree(pdata);
+err_irq:
+ free_irq(cqspi->irq, cqspi);
+err_of:
+ spi_master_put(master);
+ dev_err(&pdev->dev, "Cadence QSPI controller probe failed\n");
+ return ret;
+}
+
+static int cadence_qspi_remove(struct platform_device *pdev)
+{
+ struct spi_master *master = platform_get_drvdata(pdev);
+ struct struct_cqspi *cadence_qspi = spi_master_get_devdata(master);
+
+ cadence_qspi_controller_enable(cadence_qspi->iobase, 0);
+ platform_set_drvdata(pdev, NULL);
+ free_irq(cadence_qspi->irq, cadence_qspi);
+ iounmap(cadence_qspi->iobase);
+ iounmap(cadence_qspi->qspi_ahb_virt);
+ release_mem_region(cadence_qspi->res->start,
+ resource_size(cadence_qspi->res));
+ kfree(pdev->dev.platform_data);
+ spi_unregister_master(master);
+ spi_master_put(master);
+ return 0;
+}
+
+static const struct cqspi_driver_platdata k2g_qspi = {
+ .quirks = CQSPI_NEEDS_WR_DELAY,
+};
+
+static const struct cqspi_driver_platdata am654_ospi = {
+ .quirks = CQSPI_NEEDS_WR_DELAY,
+};
+
+static const struct cqspi_driver_platdata intel_lgm_qspi = {
+ .quirks = CQSPI_DISABLE_DAC_MODE,
+};
+
+#ifdef CONFIG_OF
+static const struct of_device_id cadence_qspi_of_match[] = {
+ {
+ .compatible = "cadence,qspi",
+ },
+ {
+ .compatible = "ti,k2g-qspi",
+ .data = &k2g_qspi,
+ },
+ {
+ .compatible = "ti,am654-ospi",
+ .data = &am654_ospi,
+ },
+ {
+ .compatible = "intel,lgm-qspi",
+ .data = &intel_lgm_qspi,
+ },
+ { /* end of table */}
+};
+MODULE_DEVICE_TABLE(of, cadence_qspi_of_match);
+#else
+#define cadence_qspi_of_match NULL
+#endif /* CONFIG_OF */
+
+static struct platform_driver cadence_qspi_platform_driver = {
+ .probe = cadence_qspi_probe,
+ .remove = cadence_qspi_remove,
+ .driver = {
+ .name = CADENCE_QSPI_NAME,
+ .of_match_table = cadence_qspi_of_match,
+ },
+};
+
+module_platform_driver(cadence_qspi_platform_driver);
+
+MODULE_DESCRIPTION("Cadence QSPI Controller Driver");
+MODULE_LICENSE("GPL v2");
+MODULE_ALIAS("platform:" CADENCE_QSPI_NAME);
+MODULE_AUTHOR("Ley Foon Tan <lftan@xxxxxxxxxx>");
+MODULE_AUTHOR("Graham Moore <grmoore@xxxxxxxxxxxxxxxxxxxxx>");
+MODULE_AUTHOR("Vadivel Murugan R <vadivel.muruganx.ramuthevar@xxxxxxxxx>");
diff --git a/drivers/spi/spi-cadence-quadspi.h b/drivers/spi/spi-cadence-quadspi.h
new file mode 100644
index 000000000000..c78ee9423950
--- /dev/null
+++ b/drivers/spi/spi-cadence-quadspi.h
@@ -0,0 +1,272 @@
+/* SPDX-License-Identifier: BSD-2-Clause */
NACK, should be GPLv2 and no need for boilerplate text below when SPDX
Identifier is present.
sure, keep GPLv2
+/*
+ * Driver for Cadence QSPI Controller
+ *
+ * Copyright Altera Corporation (C) 2012-2014. All rights reserved.
+ * Copyright Intel Corporation (C) 2019. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+#ifndef __CADENCE_QSPI__H__
+#define __CADENCE_QSPI__H__
+#include <linux/reset.h>
+#include <linux/spi/spi-mem.h>
+
+#define CQSPI_MAX_CHIP_SELECT (16)
+#define STIG_READ_MODE 1
+#define STIG_WRITE_MODE 2
+#define STIG_WRITE_READ_MODE 3
+#define IDC_READ_MODE 4
+#define IDC_WRITE_MODE 5
+
+#define QSPI_FLASH_TYPE_NOR 0
+#define QSPI_FLASH_TYPE_NAND 1
+

+#define QUAD_SIO 0
+#define QUAD_DIO 1
+#define QUAD_QIO 2
+
+#define QUAD_LSB 4
+
Have driver specific prefixes please.
agreed!

+#define MACRONIX_4PP_CMD 0x38
+#define MACRONIX_WRSR_CMD 0x01
+
These needs to be dropped
sure , drop it.
---
With Best Regards
Vadivel