Hi Ahmad, On Mon, Jun 28, 2021 at 08:40:32AM +0200, Ahmad Fatoum wrote: > The Kconfig option already warns that the current behavior of > machine_id_set_hashable() overriding previous calls can lead to the > machine-id changing over updates. We don't yet have this problem in > practice, because the only two upstream users are for bsec and ocotp, > which are efuse drivers for different SoCs. On the other hand, users > may want to get the unique ID from an EEPROM and with deep probe > support, the initcall ordering will be independent of the actual probe > order. > > Work around this issue by introducing a way for each board to explicitly > reference a nvmem cell that should be hashed to produce the machine-id. > > If no such device tree property is supplied, the last call to > machine_id_set_hashable() will be used as before. > > Cc: Bastian Krause <bst@xxxxxxxxxxxxxx> > Cc: Uwe Kleine-König <ukl@xxxxxxxxxxxxxx> > Signed-off-by: Ahmad Fatoum <a.fatoum@xxxxxxxxxxxxxx> > --- > common/Kconfig | 13 ++++++---- > common/machine_id.c | 41 ++++++++++++++++++++++++++++--- > drivers/nvmem/core.c | 44 ++++++++++++++++++++++++---------- > drivers/of/base.c | 11 +++++++++ > include/linux/nvmem-consumer.h | 5 ++++ > include/of.h | 6 +++++ This lacks a change in Documentation/devicetree/bindings/ :) Regards, 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 | _______________________________________________ barebox mailing list barebox@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/barebox