[PATCH v3 14/19] nvmem: microchip-otpc: Expose UID registers as 2nd nvmem device
Alexander Dahl
ada at thorsis.com
Fri Feb 20 02:38:23 PST 2026
Hello Claudiu,
thanks for your feedback on this, see my remarks inline below.
Am Sat, Jan 31, 2026 at 06:11:56PM +0200 schrieb Claudiu Beznea:
>
>
> On 1/20/26 17:44, Alexander Dahl wrote:
> > For SAM9X60 the Product UID x Register containing the Unique Product ID
> > is part of the OTPC registers.
>
> SAMA7G5 as well.
>
> > We have everything at hand here to just
> > create a trivial nvmem device for those.
>
> Please massage a bit the patch description to match https://www.kernel.org/doc/html/v6.13/process/submitting-patches.html#describe-your-changes
>
> >
> > Signed-off-by: Alexander Dahl <ada at thorsis.com>
> > ---
> >
> > Notes:
> > v3:
> > - no changes
> > v2:
> > - Use dev_err_probe() for error reporting (thanks Claudiu)
> > - Move required register definition over here from removed patch
> >
> > drivers/nvmem/microchip-otpc.c | 38 +++++++++++++++++++++++++++++++++-
> > 1 file changed, 37 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/nvmem/microchip-otpc.c b/drivers/nvmem/microchip-otpc.c
> > index 34c0d7611f3e4..b35ed13b004bb 100644
> > --- a/drivers/nvmem/microchip-otpc.c
> > +++ b/drivers/nvmem/microchip-otpc.c
> > @@ -25,10 +25,14 @@
> > #define MCHP_OTPC_HR (0x20)
> > #define MCHP_OTPC_HR_SIZE GENMASK(15, 8)
> > #define MCHP_OTPC_DR (0x24)
> > +#define MCHP_OTPC_UID0R (0x60)
> > #define MCHP_OTPC_NAME "mchp-otpc"
> > #define MCHP_OTPC_SIZE (11 * 1024)
> > +#define MCHP_OTPC_UID_NAME "mchp-uid"
> > +#define MCHP_OTPC_UID_SIZE 16
>
> I think you can drop these as they are used (individually) in a single place.
If you don't mind I would rather keep it here together with already
present similar definitions, instead of sprinkling new literals over
this c file.
>
> > +
> > /**
> > * struct mchp_otpc - OTPC private data structure
> > * @base: base address
> > @@ -230,6 +234,16 @@ static int mchp_otpc_init_packets_list(struct mchp_otpc *otpc, u32 *size)
> > return 0;
> > }
> > +static int mchp_otpc_uid_read(void *priv, unsigned int offset,
> > + void *val, size_t bytes)
> > +{
> > + struct mchp_otpc *otpc = priv;
> > +
> > + memcpy_fromio(val, otpc->base + MCHP_OTPC_UID0R + offset, bytes);
>
> Include <asm-generic/logic_io.h> for this one?
Yes, include what you use, I'd add `#include <linux/io.h>` then, seems
the usual way in drivers/nvmem.
>
> > +
> > + return 0;
> > +}
> > +
> > static struct nvmem_config mchp_nvmem_config = {
> > .name = MCHP_OTPC_NAME,
> > .type = NVMEM_TYPE_OTP,
> > @@ -239,6 +253,15 @@ static struct nvmem_config mchp_nvmem_config = {
> > .reg_read = mchp_otpc_read,
> > };
> > +static struct nvmem_config mchp_otpc_uid_nvmem_config = {
> > + .name = MCHP_OTPC_UID_NAME,
>
> Should the type be filled here?
In general I would say yes, but looking at `enum nvmem_type` none of
them seems to match. IIRC I more or less copied the approach from
drivers/soc/atmel/sfr.c here.
Greets
Alex
>
>
> > + .read_only = true,
> > + .word_size = 4,
> > + .stride = 4,
> > + .size = MCHP_OTPC_UID_SIZE,
> > + .reg_read = mchp_otpc_uid_read,
> > +};
> > +
> > static int mchp_otpc_probe(struct platform_device *pdev)
> > {
> > struct nvmem_device *nvmem;
> > @@ -269,8 +292,21 @@ static int mchp_otpc_probe(struct platform_device *pdev)
> > mchp_nvmem_config.size = size;
> > mchp_nvmem_config.priv = otpc;
> > nvmem = devm_nvmem_register(&pdev->dev, &mchp_nvmem_config);
> > + if (IS_ERR(nvmem)) {
> > + return dev_err_probe(&pdev->dev, PTR_ERR(nvmem),
> > + "Error registering OTP as nvmem device\n");
> > + }
> > - return PTR_ERR_OR_ZERO(nvmem);
> > + mchp_otpc_uid_nvmem_config.dev = otpc->dev;
> > + mchp_otpc_uid_nvmem_config.priv = otpc;
> > +
> > + nvmem = devm_nvmem_register(&pdev->dev, &mchp_otpc_uid_nvmem_config);
> > + if (IS_ERR(nvmem)) {
> > + return dev_err_probe(&pdev->dev, PTR_ERR(nvmem),
> > + "Error registering UIDxR as nvmem device\n");
> > + }
> > +
> > + return 0;
> > }
> > static const struct of_device_id __maybe_unused mchp_otpc_ids[] = {
>
More information about the linux-arm-kernel
mailing list