On Tue, May 20, 2025 at 01:23:05PM -0500, Eric Schikschneit wrote: > This will allow arch(s) that use device tree blobs to pad the end of the > device tree so they can be modified by board files at run time. This will > help prevent errors such as FDT_ERR_NOSPACE from occuring. > > Signed-off-by: Eric Schikschneit <eric.schikschn...@novatechautomation.com>
I was testing merging this and I found a few problems, sorry. > --- > arch/Kconfig | 7 +++++++ > arch/arc/dts/Makefile | 2 +- > arch/m68k/dts/Makefile | 2 +- > arch/microblaze/dts/Makefile | 2 +- > arch/nios2/dts/Makefile | 2 +- > arch/riscv/dts/Makefile | 2 +- > arch/sandbox/dts/Makefile | 2 +- > arch/x86/dts/Makefile | 2 +- > doc/board/broadcom/bcm7xxx.rst | 6 ++---- > doc/develop/makefiles.rst | 5 ++++- > scripts/Makefile.dts | 4 ++++ > 11 files changed, 24 insertions(+), 12 deletions(-) > > diff --git a/arch/Kconfig b/arch/Kconfig > index ea33d07c086..a683af3d591 100644 > --- a/arch/Kconfig > +++ b/arch/Kconfig > @@ -63,6 +63,13 @@ config SYS_CACHELINE_SIZE > default 64 if RISCV > default 32 if MIPS > > +config SYS_DTC_PAD_BYTES > + int "Size in bytes to pad device tree blob" > + default 4096 if ARC || ARM || M68K || MICROBLAZE || NIOS2 \ > + || RISCV || SANDBOX || X86 > + default 32768 if X86 && EFI_APP > + default 0 I believe this is the logic I suggested and so the problem is on me. For ARM && ARM64, this is right. But for 32bit ARM, we should be 0 and let individual boards set this as needed (none today but your board does need it). This leads to size growth and then failure on a few platforms is how I noticed this. The second problem is that the Makefiles in dts/upstream are confusingly not 100% upstream things and so also need to have -p 0x1000 taken out of them as they'll get the value this way instead. -- Tom
signature.asc
Description: PGP signature
