On Mon, 10 Dec 2018 10:35:35 +0000
Schrempf Frieder [off-list ref] wrote:

quoted

quoted

quoted

+
+static int nxp_fspi_exec_op(struct spi_mem *mem, const struct
+spi_mem_op *op) {
+	struct nxp_fspi *f = spi_controller_get_devdata(mem->spi->master);
+	int err = 0;
+
+	mutex_lock(&f->lock);
+
+	/* Wait for controller being ready. */
+	err = fspi_readl_poll_tout(f, f->iobase + FSPI_STS0,
+				   FSPI_STS0_ARB_IDLE, 1, POLL_TOUT, true);
+	WARN_ON(err);
+
+	nxp_fspi_select_mem(f, mem->spi);
+
+	nxp_fspi_prepare_lut(f, op);
+	/*
+	 * If we have large chunks of data, we read them through the AHB bus
+	 * by accessing the mapped memory. In all other cases we use
+	 * IP commands to access the flash.
+	 */
+	if (op->data.nbytes > (f->devtype_data->rxfifo - 4) &&
+	    op->data.dir == SPI_MEM_DATA_IN) {
+		nxp_fspi_read_ahb(f, op);
+	} else {
+		if (op->data.nbytes && op->data.dir == SPI_MEM_DATA_OUT)
+			nxp_fspi_fill_txfifo(f, op);
+
+		err = nxp_fspi_do_op(f, op);
+
+		/* Invalidate the data in the AHB buffer. */
+		if (op->data.nbytes && op->data.dir == SPI_MEM_DATA_OUT)
+			nxp_fspi_invalid(f);  

E.g. in case of an erase operation or a NAND load page operation, the
invalidation is not triggered, but flash/buffer contents have changed.
So I'm not sure if this is enough...  

Ok, would change this and have invalidate for all operations.  

Maybe you can find out the correct way through testing with NOR and NAND.

Or just invalidate the buffer every time you're doing a read through the
AHB. This should always work.

I also think we should quickly move to a model where AHB accesses are
reserved for dirmap, and regular spi-mem op are limited to non-ahb
reads.

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel