On 17 July 2016 at 14:28, Daniel Gimpelevich <daniel@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx> wrote: > On Mon, 2016-06-20 at 11:27 +0200, Jonas Gorski wrote: >> Instead of rewriting the arguments to match the UHI spec, store the >> address of a appended or UHI supplied dtb in fw_supplied_dtb. >> >> That way the original bootloader arugments are kept intact while still >> making the use of an appended dtb invisible for mach code. >> >> Mach code can still find out if it is an appended dtb by comparing >> fw_arg1 with fw_supplied_dtb. >> >> Signed-off-by: Jonas Gorski <jogo@xxxxxxxxxxx> >> --- >> v1 -> v2: >> * no changes >> >> arch/mips/ath79/setup.c | 4 ++-- >> arch/mips/bmips/setup.c | 4 ++-- >> arch/mips/include/asm/bootinfo.h | 4 ++++ >> arch/mips/kernel/head.S | 21 ++++++++++++++------- >> arch/mips/kernel/setup.c | 4 ++++ >> arch/mips/lantiq/prom.c | 4 ++-- >> arch/mips/pic32/pic32mzda/init.c | 4 ++-- >> 7 files changed, 30 insertions(+), 15 deletions(-) > [snip] >> - else if (fw_arg0 == -2) /* UHI interface */ >> - dtb = (void *)fw_arg1; >> + else if (fw_passed_dtb) /* UHI interface */ >> + dtb = (void *)fw_passed_dtb; > > I just now realized that this is also incorrect, on each platform. The > check for fw_passed_dtb should be in addition (prior) to checking for > UHI via fw_arg0 == -2, not instead of it. No it isn't, the code in head.S already does this, so there is no need to check fw_arg0 again, unless you need to know if this was an appended dtb or a UHI passed one. The whole point of using fw_passed_dtb is that you *don't* need to check individually for UHI and appended dtb. Jonas
