On Thu, Oct 22, 2026 at 04:09:06AM -0700, E Shattow wrote:
> add default mtdparts and mtdids consistent with Linux v7.3 Cadence QSPI
> driver .get_name() handler cqspi_get_name() and jh7110-common.dtsi
> 
> Signed-off-by: E Shattow <[email protected]>
> ---
> 
> This allows us to update the documentation to be slightly less likely
> that the user would type the wrong offset by using mtd part names:
> 
> -    loady && sf update $loadaddr 0 $filesize
> -        StarFive # loady && sf update $loadaddr 0 $filesize
> +    loady && sf update $loadaddr spl
> +        StarFive # loady && sf update $loadaddr spl
> 
> However that documentation update is planned for after the mtdparts
> addition lands in release. It would be preferable to derive mtdparts
> layout at runtime from device tree; if there is some way to do that
> sensibly please offer some advice how to do so in the review. Thanks!
> 
> Refer to discussion of mtd-id for SpacemiT K3 [1] and Linux v7.3 Cadence QSPI
> driver [2].
> 
> 1: https://lore.kernel.org/u-boot/[email protected]/
> 
> 2: 
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/spi/spi-cadence-quadspi.c?h=v7.3#n1724
> 
> static const char *cqspi_get_name(struct spi_mem *mem)
> {
>         struct cqspi_st *cqspi = 
> spi_controller_get_devdata(mem->spi->controller);
>         struct device *dev = &cqspi->pdev->dev;
> 
>         return devm_kasprintf(dev, GFP_KERNEL, "%s.%d", dev_name(dev),
>                               spi_get_chipselect(mem->spi, 0));
> }
> 
> In the typical situation for JH-7110 SoC then dev_name(dev) is "13010000.spi"
> and spi_get_chipselect(mem->spi, 0) returns a value of zero.
> 
> Changes in v2:
> - use Linux v7.3 cqspi_get_name() typical result for mtdparts layout
>   name; this is done for compatibility in the non-default configuration
>   of matching cmdline mtdparts mtd-id when Linux MTD_OF_PARTS is disabled
> ---
>  configs/starfive_visionfive2_defconfig | 4 ++++
>  1 file changed, 4 insertions(+)

Reviewed-by: Leo Yu-Chi Liang <[email protected]>

Reply via email to