Hi Mason, Mason Yang <masonccyang@xxxxxxxxxxx> wrote on Mon, 15 Apr 2019 17:23:52 +0800: > Add a driver for Macronix MX25F0A NAND controller. > > Signed-off-by: Mason Yang <masonccyang@xxxxxxxxxxx> > --- > drivers/mtd/nand/raw/Kconfig | 6 + > drivers/mtd/nand/raw/Makefile | 1 + > drivers/mtd/nand/raw/mxic_nand.c | 294 +++++++++++++++++++++++++++++++++++++++ > 3 files changed, 301 insertions(+) > create mode 100644 drivers/mtd/nand/raw/mxic_nand.c > > diff --git a/drivers/mtd/nand/raw/Kconfig b/drivers/mtd/nand/raw/Kconfig > index e604625..e0329cc 100644 > --- a/drivers/mtd/nand/raw/Kconfig > +++ b/drivers/mtd/nand/raw/Kconfig > @@ -522,6 +522,12 @@ config MTD_NAND_QCOM > Enables support for NAND flash chips on SoCs containing the EBI2 NAND > controller. This controller is found on IPQ806x SoC. > > +config MTD_NAND_MXIC > + tristate "Macronix MX25F0A NAND controller" > + depends on HAS_IOMEM > + help > + This selects the Macronix MX25F0A NAND controller driver. > + > config MTD_NAND_MTK > tristate "Support for NAND controller on MTK SoCs" > depends on ARCH_MEDIATEK || COMPILE_TEST > diff --git a/drivers/mtd/nand/raw/Makefile b/drivers/mtd/nand/raw/Makefile > index 5a5a72f..c8a6790 100644 > --- a/drivers/mtd/nand/raw/Makefile > +++ b/drivers/mtd/nand/raw/Makefile > @@ -54,6 +54,7 @@ obj-$(CONFIG_MTD_NAND_SUNXI) += sunxi_nand.o > obj-$(CONFIG_MTD_NAND_HISI504) += hisi504_nand.o > obj-$(CONFIG_MTD_NAND_BRCMNAND) += brcmnand/ > obj-$(CONFIG_MTD_NAND_QCOM) += qcom_nandc.o > +obj-$(CONFIG_MTD_NAND_MXIC) += mxic_nand.o > obj-$(CONFIG_MTD_NAND_MTK) += mtk_ecc.o mtk_nand.o > obj-$(CONFIG_MTD_NAND_TEGRA) += tegra_nand.o > obj-$(CONFIG_MTD_NAND_STM32_FMC2) += stm32_fmc2_nand.o > diff --git a/drivers/mtd/nand/raw/mxic_nand.c b/drivers/mtd/nand/raw/mxic_nand.c > new file mode 100644 > index 0000000..689fae2 > --- /dev/null > +++ b/drivers/mtd/nand/raw/mxic_nand.c > @@ -0,0 +1,294 @@ > +// SPDX-License-Identifier: GPL-2.0 > +// > +// Copyright (C) 2019 Macronix International Co., Ltd. > +// > +// Authors: > +// Mason Yang <masonccyang@xxxxxxxxxxx> > +// zhengxunli <zhengxunli@xxxxxxxxxxx> This is not a valid name. Also if he appears here I suppose he should be credited in the module_authors() macro too. > +// As a personal taste, I prefer when the header uses /* */ and only the SPDX tag uses //. > + > +#include <linux/mfd/mxic-mx25f0a.h> > +#include <linux/mtd/rawnand.h> > +#include <linux/mtd/nand_ecc.h> > + > +#include "internals.h" > + > +struct mxic_nand_ctlr { > + struct mx25f0a_mfd *mfd; > + struct nand_controller base; > + struct nand_chip nand; Even if this controller only supports one CS (and then, one chip), please have a clear separation between the NAND controller and the NAND chip by having one structure for the NAND chips and one structure for the NAND controller. > +}; > + > +static void mxic_host_init(struct mxic_nand_ctlr *mxic) Please choose a constant prefix for all functions, right now there is: mxic_ mxic_nand_ mx25f0a_nand_ I think mxic_nand_ or mx25f0a_nand_ is wise. > +{ > + writel(DATA_STROB_EDO_EN, mxic->mfd->regs + DATA_STROB); > + writel(INT_STS_ALL, mxic->mfd->regs + INT_STS_EN); > + writel(0x0, mxic->mfd->regs + ONFI_DIN_CNT(0)); > + writel(HC_CFG_NIO(8) | HC_CFG_SLV_ACT(0) | HC_CFG_IDLE_SIO_LVL(1) | > + HC_CFG_TYPE(1, HC_CFG_TYPE_RAW_NAND) | HC_CFG_MAN_CS_EN, > + mxic->mfd->regs + HC_CFG); > + writel(0x0, mxic->mfd->regs + HC_EN); > +} > + > +static int mxic_nand_wait_ready(struct nand_chip *chip) > +{ > + struct mxic_nand_ctlr *mxic = nand_get_controller_data(chip); > + u32 sts; > + > + return readl_poll_timeout(mxic->mfd->regs + INT_STS, sts, > + sts & INT_RDY_PIN, 0, USEC_PER_SEC); > +} > + > +static void mxic_nand_select_chip(struct nand_chip *chip, int chipnr) _select_target() is preferred now > +{ > + struct mxic_nand_ctlr *mxic = nand_get_controller_data(chip); > + > + switch (chipnr) { > + case 0: > + case 1: > + writel(HC_EN_BIT, mxic->mfd->regs + HC_EN); > + writel(HC_CFG_MAN_CS_ASSERT | readl(mxic->mfd->regs + HC_CFG), > + mxic->mfd->regs + HC_CFG); In both case I would prefer a: reg = readl(...); reg &= ~xxx; reg |= yyy; writel(reg, ...); Much easier to read. > + break; > + > + case -1: > + writel(~HC_CFG_MAN_CS_ASSERT & readl(mxic->mfd->regs + HC_CFG), > + mxic->mfd->regs + HC_CFG); > + writel(0, mxic->mfd->regs + HC_EN); > + break; > + > + default: Error? > + break; > + } > +} > + > +static int mxic_nand_data_xfer(struct mxic_nand_ctlr *mxic, const void *txbuf, > + void *rxbuf, unsigned int len) > +{ There is not so much code shared, why not separating this function for rx and tx cases? > + unsigned int pos = 0; > + > + while (pos < len) { > + unsigned int nbytes = len - pos; > + u32 data = 0xffffffff; > + u32 sts; > + int ret; > + > + if (nbytes > 4) > + nbytes = 4; > + > + if (txbuf) > + memcpy(&data, txbuf + pos, nbytes); > + > + ret = readl_poll_timeout(mxic->mfd->regs + INT_STS, sts, > + sts & INT_TX_EMPTY, 0, USEC_PER_SEC); Using USEC_PER_SEC for a delay is weird > + if (ret) > + return ret; > + > + writel(data, mxic->mfd->regs + TXD(nbytes % 4)); > + > + if (rxbuf) { > + ret = readl_poll_timeout(mxic->mfd->regs + INT_STS, sts, > + sts & INT_TX_EMPTY, 0, > + USEC_PER_SEC); > + if (ret) > + return ret; > + > + ret = readl_poll_timeout(mxic->mfd->regs + INT_STS, sts, > + sts & INT_RX_NOT_EMPTY, 0, > + USEC_PER_SEC); > + if (ret) > + return ret; > + > + data = readl(mxic->mfd->regs + RXD); > + data >>= (8 * (4 - nbytes)); What is this? Are you trying to handle some endianness issue? > + memcpy(rxbuf + pos, &data, nbytes); > + WARN_ON(readl(mxic->mfd->regs + INT_STS) & > + INT_RX_NOT_EMPTY); > + } else { > + readl(mxic->mfd->regs + RXD); > + } > + WARN_ON(readl(mxic->mfd->regs + INT_STS) & INT_RX_NOT_EMPTY); > + > + pos += nbytes; > + } > + > + return 0; > +} > + > +static int mxic_nand_exec_op(struct nand_chip *chip, > + const struct nand_operation *op, bool check_only) > +{ > + struct mxic_nand_ctlr *mxic = nand_get_controller_data(chip); > + const struct nand_op_instr *instr = NULL; > + int i, len = 0, ret = 0; > + unsigned int op_id; > + unsigned char cmdcnt = 0, addr_cnt = 0, cmd_addr[8] = {0}; > + > + for (op_id = 0; op_id < op->ninstrs; op_id++) { > + instr = &op->instrs[op_id]; > + > + switch (instr->type) { > + case NAND_OP_CMD_INSTR: > + cmd_addr[len++] = instr->ctx.cmd.opcode; > + cmdcnt++; > + break; > + > + case NAND_OP_ADDR_INSTR: > + for (i = 0; i < instr->ctx.addr.naddrs; i++) > + cmd_addr[len++] = instr->ctx.addr.addrs[i]; > + addr_cnt = i; > + break; > + > + case NAND_OP_DATA_IN_INSTR: > + break; > + > + case NAND_OP_DATA_OUT_INSTR: > + writel(instr->ctx.data.len, > + mxic->mfd->regs + ONFI_DIN_CNT(0)); > + break; > + > + case NAND_OP_WAITRDY_INSTR: > + break; > + } > + } > + > + if (op_id == 5 && instr->type == NAND_OP_WAITRDY_INSTR) { > + /* > + * In case cmd-addr-data-cmd-wait in a sequence, > + * separate the 2'nd command, i.e,. nand_prog_page_op() > + */ I think this is the kind of limitation that could be described very easily with a nand_op_parser structure and used by nand_op_parser_exec_op() (see marvell_nand.c). > + writel(OP_CMD_BUSW(OP_BUSW_8) | OP_ADDR_BUSW(OP_BUSW_8) | > + OP_DATA_BUSW(OP_BUSW_8) | OP_DUMMY_CYC(0x3F) | > + OP_ADDR_BYTES(addr_cnt) | > + OP_CMD_BYTES(1), mxic->mfd->regs + SS_CTRL(0)); > + writel(0, mxic->mfd->regs + HC_EN); > + writel(HC_EN_BIT, mxic->mfd->regs + HC_EN); > + > + mxic_nand_data_xfer(mxic, cmd_addr, NULL, len - 1); > + > + mxic_nand_data_xfer(mxic, instr->ctx.data.buf.out, NULL, > + instr->ctx.data.len); > + > + writel(0, mxic->mfd->regs + HC_EN); > + writel(HC_EN_BIT, mxic->mfd->regs + HC_EN); > + mxic_nand_data_xfer(mxic, &cmd_addr[--len], NULL, 1); > + ret = mxic_nand_wait_ready(chip); > + if (ret) > + goto err_out; > + return ret; > + } > + > + if (len) { > + writel(OP_CMD_BUSW(OP_BUSW_8) | OP_ADDR_BUSW(OP_BUSW_8) | > + OP_DATA_BUSW(OP_BUSW_8) | OP_DUMMY_CYC(0x3F) | > + OP_ADDR_BYTES(addr_cnt) | > + OP_CMD_BYTES(cmdcnt > 0 ? cmdcnt : 0), > + mxic->mfd->regs + SS_CTRL(0)); > + writel(0, mxic->mfd->regs + HC_EN); > + writel(HC_EN_BIT, mxic->mfd->regs + HC_EN); > + > + mxic_nand_data_xfer(mxic, cmd_addr, NULL, len); > + } > + > + for (op_id = 0; op_id < op->ninstrs; op_id++) { > + instr = &op->instrs[op_id]; > + > + switch (instr->type) { > + case NAND_OP_CMD_INSTR: > + case NAND_OP_ADDR_INSTR: > + break; > + > + case NAND_OP_DATA_IN_INSTR: > + writel(0x0, mxic->mfd->regs + ONFI_DIN_CNT(0)); > + writel(readl(mxic->mfd->regs + SS_CTRL(0)) | OP_READ, > + mxic->mfd->regs + SS_CTRL(0)); > + mxic_nand_data_xfer(mxic, NULL, instr->ctx.data.buf.in, > + instr->ctx.data.len); > + break; > + > + case NAND_OP_DATA_OUT_INSTR: > + mxic_nand_data_xfer(mxic, instr->ctx.data.buf.out, NULL, > + instr->ctx.data.len); > + break; > + > + case NAND_OP_WAITRDY_INSTR: > + ret = mxic_nand_wait_ready(chip); > + if (ret) > + goto err_out; > + break; > + } > + } > + > +err_out: > + return ret; Ditto, please return directly if there is nothing in the error path. > +} > + > +static const struct nand_controller_ops mxic_nand_controller_ops = { > + .exec_op = mxic_nand_exec_op, > +}; > + > +static int mx25f0a_nand_probe(struct platform_device *pdev) > +{ > + struct mtd_info *mtd; > + struct mx25f0a_mfd *mfd = dev_get_drvdata(pdev->dev.parent); > + struct mxic_nand_ctlr *mxic; > + struct nand_chip *nand_chip; > + int err; > + > + mxic = devm_kzalloc(&pdev->dev, sizeof(struct mxic_nand_ctlr), > + GFP_KERNEL); mxic for a NAND controller structure is probably not a name meaningful enough. > + if (!mxic) > + return -ENOMEM; > + > + nand_chip = &mxic->nand; > + mtd = nand_to_mtd(nand_chip); > + mtd->dev.parent = pdev->dev.parent; > + nand_chip->ecc.priv = NULL; > + nand_set_flash_node(nand_chip, pdev->dev.parent->of_node); > + nand_chip->priv = mxic; > + > + mxic->mfd = mfd; > + > + nand_chip->legacy.select_chip = mxic_nand_select_chip; Please don't implement legacy interfaces. You can check in marvell_nand.c how this is handled now: b25251414f6e mtd: rawnand: marvell: Stop implementing ->select_chip() > + mxic->base.ops = &mxic_nand_controller_ops; > + nand_controller_init(&mxic->base); > + nand_chip->controller = &mxic->base; > + > + mxic_host_init(mxic); > + > + err = nand_scan(nand_chip, 1); > + if (err) > + goto fail; You can return directly as there is nothing to clean in the error path. > + > + err = mtd_device_register(mtd, NULL, 0); > + if (err) > + goto fail; Ditto > + > + platform_set_drvdata(pdev, mxic); > + > + return 0; > +fail: > + return err; > +} > + > +static int mx25f0a_nand_remove(struct platform_device *pdev) > +{ > + struct mxic_nand_ctlr *mxic = platform_get_drvdata(pdev); > + > + nand_release(&mxic->nand); > + > + return 0; > +} > + > +static struct platform_driver mx25f0a_nand_driver = { > + .probe = mx25f0a_nand_probe, Please don't align '=' on tabs. > + .remove = mx25f0a_nand_remove, > + .driver = { > + .name = "mxic-nand-ctlr", > + }, > +}; > +module_platform_driver(mx25f0a_nand_driver); mx25f0a_nand_controller_driver would be better > + > +MODULE_AUTHOR("Mason Yang <masonccyang@xxxxxxxxxxx>"); > +MODULE_DESCRIPTION("MX25F0A RAW NAND controller driver"); > +MODULE_LICENSE("GPL v2"); Thanks, Miquèl ______________________________________________________ Linux MTD discussion mailing list http://lists.infradead.org/mailman/listinfo/linux-mtd/