On Tue, Mar 23, 2010 at 8:48 AM, Sascha Hauer <s.hauer@xxxxxxxxxxxxxx> wrote:
[...]
> I wonder that this patch helps you. Does your NAND work properly after
> starting? I ask because in imxnd_probe() NFC_INT_MSK gets set again.
I agree that it's weird. But it does help. Apparantly, on i.MX21, NFC_INT_MSK
must not be set for the first (non-reset) command, otherwise NFC_INT in
NFC_CONFIG2 will never get set.
After booting, other NAND accesses do work, for instance the NAND
device ID readout and the scanning for bad blocks:
barebox 2010.03.0-00094-gc6f4ea0-dirty (Mar 22 2010 - 10:02:13)
Board: Freescale i.MX21 ADS
cfi_probe: cfi_flash base: 0xc8000000 size: 0x02000000
NAND device: Manufacturer ID: 0xec, Chip ID: 0x36 (Samsung NAND 64MiB
1,8V 8-bit)
Scanning device for bad blocks
Bad eraseblock 2133 at 0x02154000
Malloc space: 0xc0c00000 -> 0xc1000000 (size 4 MB)
Stack space : 0xc0bf8000 -> 0xc0c00000 (size 32 kB)
Open /dev/env0 No such file or directory
no valid environment found on /dev/env0. Using default environment
running /env/bin/init...
barebox:/
> I tested on i.MX27 that clearing this bit doesn't harm, so I think we do
> not need the ifdef but can clear it on every SoC, like this:
>
> diff --git a/drivers/nand/nand_imx.c b/drivers/nand/nand_imx.c
> index afd5637..5c910d8 100644
> --- a/drivers/nand/nand_imx.c
> +++ b/drivers/nand/nand_imx.c
> @@ -1085,11 +1085,11 @@ void __nand_boot_init imx_nand_load_image(void *dest, int size)
> writew(0x4, host.regs + NFC_WRPROT);
>
> tmp = readw(host.regs + NFC_CONFIG1);
> - tmp |= NFC_ECC_EN | NFC_INT_MSK;
> + tmp |= NFC_ECC_EN;
> if (nfc_is_v21())
> /* currently no support for 218 byte OOB with stronger ECC */
> tmp |= NFC_ECC_MODE;
> - tmp &= ~NFC_SP_EN;
> + tmp &= ~(NFC_SP_EN | NFC_INT_MSK);
> writew(tmp, host.regs + NFC_CONFIG1);
>
> if (nfc_is_v21()) {
> --
Tested this patch on an MX21ADS, and this is sufficient.
_______________________________________________
barebox mailing list
barebox@xxxxxxxxxxxxxxxxxxx
http://lists.infradead.org/mailman/listinfo/barebox
