On Thursday, July 19, 2018 8:22:00 AM CEST Boris Brezillon wrote:

On Thu, 19 Jul 2018 01:57:04 +0200
Janusz Krzysztofik [off-list ref] wrote:

quoted

Initialize NWP GPIO pin low to protect the device from hazard during
probe.  Release write protection as soon as the device is under
control.

Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com>
---
 drivers/mtd/nand/raw/ams-delta.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

diff --git a/drivers/mtd/nand/raw/ams-delta.c b/drivers/mtd/nand/raw/ams-delta.c
index 48233d638d2a..6ac38e9cfa1a 100644
--- a/drivers/mtd/nand/raw/ams-delta.c
+++ b/drivers/mtd/nand/raw/ams-delta.c

@@ -219,8 +219,8 @@ static int ams_delta_init(struct platform_device *pdev)
 
 	platform_set_drvdata(pdev, priv);
 
-	/* Set chip enabled, but  */
-	priv->gpiod_nwp = devm_gpiod_get(&pdev->dev, "nwp", GPIOD_OUT_HIGH);
+	/* Set chip enabled but write protected */
+	priv->gpiod_nwp = devm_gpiod_get(&pdev->dev, "nwp", GPIOD_OUT_LOW);
 	if (IS_ERR(priv->gpiod_nwp)) {
 		err = PTR_ERR(priv->gpiod_nwp);
 		dev_err(&pdev->dev, "NWP GPIO request failed (%d)\n", err);

@@ -267,6 +267,9 @@ static int ams_delta_init(struct platform_device *pdev)
 	if (err)
 		goto out_mtd;
 
+	/* As soon as the device is found, release write protection */
+	gpiod_set_value(priv->gpiod_nwp, 1);

Please don't do that until we have a generic way to assert/deassert WP
from the core. I agree that we shouldn't write things in the
nand_scan() path, but I'd be more comfortable if the core had control
on this pin just in case we ever need to disable write-protection
during chip detection/initialization.

OK, I can drop this patch for now if that's what you suggest.
 
Thanks,
Janusz