Hi Jon,

On Mon, Jun 18, 2012 at 21:31:46, Hunter, Jon wrote:

quoted

@@ -95,10 +89,6 @@ static int omap2_onenand_set_async_mode(int cs, void __iomem *onenand_base)
 	if (err)
 		return err;
 
-	/* Ensure sync read and sync write are disabled */
-	reg = readw(onenand_base + ONENAND_REG_SYS_CFG1);
-	reg &= ~ONENAND_SYS_CFG1_SYNC_READ & ~ONENAND_SYS_CFG1_SYNC_WRITE;
-	writew(reg, onenand_base + ONENAND_REG_SYS_CFG1);

Sorry if I was not clear, but I was still thinking that we keep the
above instance of setting async mode.

:

quoted

 static int gpmc_onenand_setup(void __iomem *onenand_base, int *freq_ptr)
 {
 	struct device *dev = &gpmc_onenand_device.dev;
+	u32 reg;
+
+	/* Ensure sync read and sync write are disabled */
+	reg = readw(onenand_base + ONENAND_REG_SYS_CFG1);
+	reg &= ~ONENAND_SYS_CFG1_SYNC_READ & ~ONENAND_SYS_CFG1_SYNC_WRITE;
+	writew(reg, onenand_base + ONENAND_REG_SYS_CFG1);

... and here we should just call omap2_onenand_set_async_mode(), ...

If omap2_onenand_set_async_mode is called from gpmc_onenand_setup & removed
from gpmc_onenand_init, when using new gpmc driver interface, i.e. using
gpmc_onenand_update [1], we lost opportunity to provide gpmc driver with
configuration details for async mode causing a big warning about each
timings & configurations that was left by bootloader (which is meant only
for cases where Kernel can't configure GPMC). Of course, we can put
omap2_onenand_set_async_mode in gpmc_onenand_update too, but then
unnecessarily, omap2_onenand_sey_async_mode would be called twice.

Please note that gpmc_onenand_setup is a callback by onenand driver.

Regards
Afzal

quoted

 
+	omap2_onenand_set_async_mode(gpmc_onenand_data->cs);
+

...  then remove the above call.

[1] http://www.mail-archive.com/linux-omap at vger.kernel.org/msg69919.html