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

From: Ramuthevar, Vadivel MuruganX
Date: Mon Feb 17 2020 - 04:18:17 EST


Hi Mark,

   Thank you for the review comments,  response in inline.

On 14/2/2020 9:09 PM, Mark Brown wrote:
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.
Yes, you're correct, it doesn't check unconditionally, will update the
INT flag in the INT_STATUS register after successful completion of read/write operation.
but in this case it is dedicated to qspi-interrupt,not shared with any other HW/SW interrupts.
+ 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?
Agreed!, will fix it.
+
+ 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.
Agreed, will fix it.
+ }
+ /* Invalid address return zero. */
Missing blank line.
Noted.
+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?
Good catch, will add the check in the next patch, the below check to be added.
if (cqspi->is_decoded_cs) {
          reg |= CQSPI_REG_CONFIG_DECODE_MASK;
} else {
          reg &= ~CQSPI_REG_CONFIG_DECODE_MASK;
+ 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.
Noted, will double check is there any impact if there is no data provided.
also will add the error check if return an error.
+ 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()
Yes, I'm sure that it's safe to use devm_request_irq().
- what happens if
the interrupt fires in the process of removing the device?
   This is not external interrupt which is coming from the device, its inbuilt to QSPI controller which has 2 Registers 1) INT_STATUS_REG 2) INT_MASK_REG.
   It fires an interrupt and updating INT flag bit in the interrupt status register once reached read/write completion state then irq_handler is called,
   there we are clearing the INT_FLAG bit by masking of RD/WR flag in INT_MASK register.

Regards
Vadivel