[PATCH v10 1/6] mtd: spi-nor: add Octal DTR support for Macronix flash

[Alvin Zhou]

Hi Tudor,

Tudor Ambarus <tudor.ambarus at linaro.org> 於 2024年10月4日 週五 下午5:58寫道：
>
>
>
> On 10/4/24 10:05 AM, Alvin Zhou wrote:
> >>> +     /* Read flash ID to make sure the switch was successful. */
> >>> +     ret = spi_nor_read_id(nor, 4, 4, buf, SNOR_PROTO_8_8_8_DTR);
> >> can we use nor->addr_nbytes for the second argument? Please test and
> >> confirm. No need to resend for this, just confirm and I can amend when
> >> applying.
> > The following is the process of spi_nor_scan()
> > int spi_nor_scan(...)
> > {
> > ......
> > ret = spi_nor_init_params(nor);
> > ......
> > ret = spi_nor_setup(nor, hwcaps);
> > ......
> > }
> > First, within the spi_nor_parse_sfdp() function inside
> > spi_nor_init_params(): nor->params->addr_nbytes is set based on the
> > SFDP, while nor->addr_nbytes is not available. Therefore, the second
> > argument cannot use nor->addr_nbytes but can use
> > nor->params->addr_nbytes. Additionally, For Macronix Octal NOR Flash
>
>
> nor->addr_nbytes is set in spi_nor_setup().
> spi_nor_set_octal_dtr() is called after spi_nor_setup(), thus you can
> use nor->addr_nbytes.

I apologize for the misunderstanding. Thanks for your clarification. So
using nor->addr_nbytes as the second argument is not a problem. I
have verified, thanks!
>
> > in Octal DDR mode, both the address and dummy cycles are fixed at 4
> > in READID, so setting the second and third argument to 4 is also valid.
>
> but we don't want magic numbers or states that are not tracked, so use
> the parameters set
>
> > Moreover, nor->addr_nbytes is set within the spi_nor_setup() function.
>
> yep, use it then.

Thanks,
Alvin