On 4 June 2015 at 17:05, Ezequiel Garcia [off-list ref] wrote:

Joachim,

I tested the driver and found an issue with it.

On 05/31/2015 06:06 PM, Joachim Eastwood wrote:

quoted

+
+     spifi->mtd.priv  = &spifi->nor;
+     spifi->nor.mtd   = &spifi->mtd;
+     spifi->nor.dev   = spifi->dev;
+     spifi->nor.priv  = spifi;
+     spifi->nor.read  = nxp_spifi_read;
+     spifi->nor.write = nxp_spifi_write;
+     spifi->nor.erase = nxp_spifi_erase;
+     spifi->nor.read_reg  = nxp_spifi_read_reg;
+     spifi->nor.write_reg = nxp_spifi_write_reg;
+
+     ret = of_modalias_node(np, modalias, sizeof(modalias));
+     if (ret < 0) {
+             dev_err(spifi->dev, "unable to get device modalias\n");
+             return ret;
+     }
+
+     /*
+      * The first read on a hard reset isn't reliable so do a
+      * dummy read of the id before calling spi_nor_scan().
+      * The reason for this problem is unknown.
+      *
+      * The official NXP spifilib uses more or less the same
+      * workaround that is applied here by reading the device
+      * id multiple times.
+      */
+     nxp_spifi_dummy_id_read(&spifi->nor);
+

Actually, I think there's an issue here. See how m25p80 deals
with modalias:

        [..]
        else if (!strcmp(spi->modalias, "spi-nor"))
                flash_name = NULL; /* auto-detect */
        else
                flash_name = spi->modalias;

        ret = spi_nor_scan(nor, flash_name, mode);
        if (ret)
                return ret;

I think the "auto-detect" quirk is really needed, or you are will
otherwise fail to probe with the generic "jedec,spi-nor" compatible
string.

I see.

Seem easy enough to fix. I'll send out new version after doing some
more tests here.

Note that this code was copied from fsl-quadspi so I assume it has the
same problem.


Thanks for testing and catching the issue, Ezequiel.


regards,
Joachim Eastwood