Re: [PATCH v9 2/2] spi: cadence-quadpsi: Add support for the Cadence QSPI controller

[Mark Brown <broonie@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 2976 bytes --]

On Fri, Feb 14, 2020 at 07:46:18PM +0800, Ramuthevar,Vadivel MuruganX wrote:

> +static irqreturn_t cqspi_irq_handler(int this_irq, void *dev)
> +{
> +	struct cqspi_st *cqspi = dev;
> +	unsigned int irq_status;
> +
> +	/* Read interrupt status */
> +	irq_status = readl(cqspi->iobase + CQSPI_REG_IRQSTATUS);
> +
> +	/* Clear interrupt */
> +	writel(irq_status, cqspi->iobase + CQSPI_REG_IRQSTATUS);
> +
> +	irq_status &= CQSPI_IRQ_MASK_RD | CQSPI_IRQ_MASK_WR;
> +
> +	if (irq_status)
> +		complete(&cqspi->transfer_complete);
> +
> +	return IRQ_HANDLED;
> +}

This will unconditionally handle the interrupt regardless of if the
hardware was actually flagging an interrupt which will break shared
interrupts and the fault handling code in genirq.

> +	tmpbufsize = op->addr.nbytes + op->dummy.nbytes;
> +	tmpbuf = kzalloc(tmpbufsize, GFP_KERNEL | GFP_DMA);
> +	if (!tmpbuf)
> +		return -ENOMEM;

I'm not clear where tmpbuf gets freed or passed out of this function?

> +
> +	if (op->addr.nbytes) {
> +		for (i = 0; i < op->addr.nbytes; i++)
> +			tmpbuf[i] = op->addr.val >> (8 * (op->addr.nbytes - i - 1));
> +
> +		addr_buf = tmpbuf;

We assign tmpbuf to addr_buf here but addr_buf just gets read from so
it's not via that AFAICT.

> +	}
> +	/* Invalid address return zero. */

Missing blank line.

> +static void cqspi_chipselect(struct cqspi_flash_pdata *f_pdata)
> +{
> +	struct cqspi_st *cqspi = f_pdata->cqspi;
> +	void __iomem *reg_base = cqspi->iobase;
> +	unsigned int chip_select = f_pdata->cs;
> +	unsigned int reg;
> +
> +	reg = readl(reg_base + CQSPI_REG_CONFIG);
> +	reg &= ~CQSPI_REG_CONFIG_DECODE_MASK;
> +
> +	/* Convert CS if without decoder.
> +	 * CS0 to 4b'1110
> +	 * CS1 to 4b'1101
> +	 * CS2 to 4b'1011
> +	 * CS3 to 4b'0111
> +	 */
> +	chip_select = 0xF & ~(1 << chip_select);

This says "if without decoder" but there's no conditionals here, what if
we do have a decoder?

> +	cqspi->master_ref_clk_hz = clk_get_rate(cqspi->clk);
> +	ddata  = of_device_get_match_data(dev);
> +	if (ddata) {
> +		if (ddata->quirks & CQSPI_NEEDS_WR_DELAY)
> +			cqspi->wr_delay = 5 * DIV_ROUND_UP(NSEC_PER_SEC,
> +						cqspi->master_ref_clk_hz);
> +		if (ddata->hwcaps_mask & CQSPI_SUPPORTS_OCTAL)
> +			master->mode_bits |= SPI_RX_OCTAL;
> +		if (!(ddata->quirks & CQSPI_DISABLE_DAC_MODE))
> +			cqspi->use_dac_mode = true;
> +		if (ddata->quirks & CQSPI_NEEDS_ADDR_SWAP) {
> +			master->bus_num = 0;
> +			master->num_chipselect = 2;
> +		}
> +	}

Given that the driver appears to unconditionally dereference match data
in other places I'd expect this to return an error if there's none,
otherwise we'll oops in those other code paths later on.

> +	ret = devm_request_irq(dev, irq, cqspi_irq_handler, 0,
> +			       pdev->name, cqspi);
> +	if (ret) {
> +		dev_err(dev, "Cannot request IRQ.\n");
> +		goto probe_reset_failed;
> +	}

Are you sure that it's safe to use devm_request_irq() - what happens if
the interrupt fires in the process of removing the device?

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]