Following up on IRC conversation, for the record: > Le 1 mai 2021 à 11:27, Thibaut <ha...@slashdirt.org> a écrit :
[…] > Yes, but only the first 4K is used. What’s needed is to check whether or not > the rb_hardconfig driver will successfully process a 4K-only block from an 8K > partition. I *think* it should, but haven’t looked back at the code or tested > that yet. After checking, the rb_hardconfig driver will handle this case with no problem. > If that’s the case then adapting routerbootpart is relatively trivial. There are two ways to deal with the hap-ac2 situation: Either fix it globally, by making the hard_config partition auto-extend up to the next defined partition via the following patch: --- a/target/linux/generic/files/drivers/mtd/parsers/routerbootpart.c +++ b/target/linux/generic/files/drivers/mtd/parsers/routerbootpart.c @@ -99,7 +99,7 @@ static struct routerboot_dynpart rb_dynparts[] = { .magic = RB_MAGIC_HARD, // stored in CPU-endianness on flash .size_fixup = NULL, .offset = 0x0, - .size = RB_BLOCK_SIZE, + .size = 0x0, .found = false, }, { .name = "soft_config », This would of course need to be confirmed working on all currently supported devices (I can’t think of a situation where this would break, but one never knows). Alternatively, overriding the size= parameter in the hap-ac2 DTS should also work. It’s likely this change will be needed for other ipq-based mikrotik devices. HTH, Thibaut _______________________________________________ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel