On Mon, Apr 14, 2025 at 04:57:35PM +0200, Ahmad Fatoum wrote:
> Hello Sascha,
>
> On 4/14/25 16:49, Sascha Hauer wrote:
> > On Fri, Apr 11, 2025 at 09:40:39AM +0200, Ahmad Fatoum wrote:
> >> barebox TLV is a scheme for storing factory data on non-volatile
> >> storage. Unlike state, it's meant to be read-only and if content
> >> is limited to already specified tags, it can be extended later on,
> >> without modifying the bootloader binary.
> >>
> >> Signed-off-by: Ahmad Fatoum <a.fat...@pengutronix.de>
> >> ---
> >> +struct tlv_mapping barebox_tlv_v1_mappings[] = {
> >> + { 0x0002, tlv_format_str, "device-hardware-release" },
> >> + { 0x0003, tlv_format_dec, "factory-timestamp" },
> >
> > The numbers in the examples suggest that the format is in epoch time,
> > but shouldn't this be documented?
>
> Good point. Yes, it's seconds since the UNIX epoch.
> Will document it for v2.
>
> >> + { 0x0004, tlv_handle_serial, "device-serial-number"},
> >> + { 0x0005, tlv_format_dec, "modification" },
> >> + { 0x0006, tlv_format_str, "featureset" },
> >
> > I think things like "device-hardware-release", "modification" and
> > "featureset"
> > need (better) documentation, at least when they are supposed to be generic.
>
> In the end the revision scheme and the concrete interpretation of bits
> depends on the project, but I can add some generic documentation inline,
> e.g.:
>
> device-hardware-release: Detailed release information string for the device
>
> modification: 0: Device is unmodified. Any set bits indicate a
> board-specific modification after assembly
Ok, the schema file describes this as a boolean.
>
> featureset: Any set bits indicate enablement of board-specific features
This is interpreted as a string with the example in test/self/tlv.dts
being "base,PoE"
... Just to emphasize we should document this :)
Sascha
--
Pengutronix e.K. | |
Steuerwalder Str. 21 | http://www.pengutronix.de/ |
31137 Hildesheim, Germany | Phone: +49-5121-206917-0 |
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
