[PATCH] cmd: fdt: use U-Boot's FDT by default

[Simon Glass]

Hi Caleb,

On Fri, 6 Sept 2024 at 03:31, Caleb Connolly <caleb.connolly at linaro.org> wrote:
>
>
>
> On 03/09/2024 20:14, Tom Rini wrote:
> > On Sat, Aug 31, 2024 at 05:46:19PM +0100, Caleb Connolly wrote:
> >
> >> When using the FDT command to inspect an arbitrary FDT in memory, it
> >> will always be necessary to explicitly set the FDT address. However it
> >> is also quite likely that the command is being used to inspect U-Boot's
> >> own FDT. Simplify that common workflow of running "fdt addr -c" to get
> >> the control address and set it by just making $fdtcontroladdr the
> >> default FDT if there isn't one.
> >>
> >> Signed-off-by: Caleb Connolly <caleb.connolly at linaro.org>
> >> ---
> >>   cmd/fdt.c | 9 +++++++++
> >>   1 file changed, 9 insertions(+)
> >>
> >> diff --git a/cmd/fdt.c b/cmd/fdt.c
> >> index d16b141ce32d..8909706e2483 100644
> >> --- a/cmd/fdt.c
> >> +++ b/cmd/fdt.c
> >> @@ -276,8 +276,17 @@ static int do_fdt(struct cmd_tbl *cmdtp, int flag, int argc, char *const argv[])
> >>
> >>              return CMD_RET_SUCCESS;
> >>      }
> >>
> >> +    /* Try using U-Boot's FDT by default */
> >> +    if (!working_fdt) {
> >> +            struct fdt_header *addr;
> >> +
> >> +            addr = (void *)env_get_hex("fdtcontroladdr", 0);
> >> +            if (addr && fdt_check_header(&addr))
> >> +                    set_working_fdt_addr((phys_addr_t)addr);
> >> +    }
> >> +
> >>      if (!working_fdt) {
> >>              puts("No FDT memory address configured. Please configure\n"
> >>                   "the FDT address via \"fdt addr <address>\" command.\n"
> >>                   "Aborting!\n");
> >
> > Setting aside the behavior change (which I am thinking about), this
> > makes the next check of !working_fdt dead code which should be removed.
>
> I wasn't sure if we could safely assume that fdtcontroladdr always
> points to a valid FDT, if that's true then yes this can be dropped.

$ ./tools/qconfig.py -f ~OF_CONTROL
11 matches
integratorap_cm720t integratorap_cm920t integratorap_cm926ejs
integratorap_cm946es integratorcp_cm1136 integratorcp_cm920t
integratorcp_cm926ejs integratorcp_cm946es mx6memcal work_92105 xtfpga

So yes, there are boards where it would not be set.

Regards,
Simon