[U-Boot] [PATCH v3 5/5] common: Enable fdt_support in SPL build

Wed Sep 23 17:38:50 CEST 2015

Hi,

On 22 September 2015 at 21:10, Chin Liang See <clsee at altera.com> wrote:
>
> Hi,
>
> > From: marex at denx.de
> > Sent: Friday, September 11, 2015 8:45 AM
> > To: Simon Glass
> > Cc: Chin Liang See; ZY - u-boot; Dinh Nguyen; Dinh Nguyen; ZY - sr; Vikas Manocha; Jagannadh Teki; ZY - pavel
> > Subject: Re: [PATCH v3 5/5] common: Enable fdt_support in SPL build
> >
> > On Friday, September 11, 2015 at 02:14:17 AM, Simon Glass wrote:
> > > Hi,
> > >
> > > On 10 September 2015 at 04:49, Marek Vasut <marex at denx.de> wrote:
> > > > On Thursday, September 10, 2015 at 07:02:16 AM, Chin Liang See wrote:
> > > >> Enable more device tree function support such as
> > > >> fdt_getprop_u32_default_node in SPL.
> > > >>
> > > >> Signed-off-by: Chin Liang See <clsee at altera.com>
> > > >> Cc: Dinh Nguyen <dinguyen at opensource.altera.com>
> > > >> Cc: Dinh Nguyen <dinh.linux at gmail.com>
> > > >> Cc: Marek Vasut <marex at denx.de>
> > > >> Cc: Stefan Roese <sr at denx.de>
> > > >> Cc: Vikas Manocha <vikas.manocha at st.com>
> > > >> Cc: Jagannadh Teki <jteki at openedev.com>
> > > >> Cc: Pavel Machek <pavel at denx.de>
> > > >
> > > > +CC Simon, I'm worried this might blow the size of all SPLs , which is
> > > > not the thing we'd like to see.
> > > >
> > > >> ---
> > > >>
> > > >>  common/Makefile |    4 ++--
> > > >>  1 files changed, 2 insertions(+), 2 deletions(-)
> > > >>
> > > >> diff --git a/common/Makefile b/common/Makefile
> > > >> index 556fb07..9e059fe 100644
> > > >> --- a/common/Makefile
> > > >> +++ b/common/Makefile
> > > >> @@ -97,7 +97,7 @@ obj-$(CONFIG_CMD_EXT4) += cmd_ext4.o
> > > >>
> > > >>  obj-$(CONFIG_CMD_EXT2) += cmd_ext2.o
> > > >>  obj-$(CONFIG_CMD_FAT) += cmd_fat.o
> > > >>  obj-$(CONFIG_CMD_FDC) += cmd_fdc.o
> > > >>
> > > >> -obj-$(CONFIG_OF_LIBFDT) += cmd_fdt.o fdt_support.o
> > > >> +obj-$(CONFIG_OF_LIBFDT) += cmd_fdt.o
> > > >>
> > > >>  obj-$(CONFIG_CMD_FITUPD) += cmd_fitupd.o
> > > >>  obj-$(CONFIG_CMD_FLASH) += cmd_flash.o
> > > >>  ifdef CONFIG_FPGA
> > > >>
> > > >> @@ -265,7 +265,7 @@ obj-y += malloc_simple.o
> > > >>
> > > >>  endif
> > > >>  obj-y += image.o
> > > >>  obj-$(CONFIG_ANDROID_BOOT_IMAGE) += image-android.o
> > > >>
> > > >> -obj-$(CONFIG_OF_LIBFDT) += image-fdt.o
> > > >> +obj-$(CONFIG_OF_LIBFDT) += image-fdt.o  fdt_support.o
> > > >>
> > > >>  obj-$(CONFIG_FIT) += image-fit.o
> > > >>  obj-$(CONFIG_FIT_SIGNATURE) += image-sig.o
> > > >>  obj-$(CONFIG_IO_TRACE) += iotrace.o
> > > >>
> > > >> --
> > > >> 1.7.7.4
> > >
> > > It might need a bit of work - here's a buildman run:
> > >
> > > 02: common: Enable fdt_support in SPL build
> > >    powerpc:  +   P1022DS_36BIT_NAND T1024RDB_SPIFLASH
> > > P1010RDB-PA_SPIFLASH P1021RDB-PC_36BIT_NAND T2080QDS_SPIFLASH
> > > P1025RDB_SDCARD P2020RDB-PC_36BIT_SPIFLASH T1042RDB_PI_SPIFLASH
> > > P1010RDB-PA_36BIT_NAND P1010RDB-PB_SDCARD P2020RDB-PC_SPIFLASH
> > > P1020MBG-PC_36BIT_SDCARD P1024RDB_SPIFLASH T1042D4RDB_NAND
> > > T4240RDB_SDCARD P1010RDB-PB_36BIT_NAND P1022DS_36BIT_SDCARD
> > > T1024QDS_NAND P1021RDB-PC_SPIFLASH T1024RDB_NAND T1023RDB_SDCARD
> > > P1010RDB-PB_36BIT_SDCARD P1022DS_NAND P1020RDB-PC_36BIT_NAND
> > > P1020RDB-PC_SDCARD T1042RDB_PI_NAND P1020RDB-PD_SPIFLASH T2080RDB_NAND
> > > P1022DS_SDCARD P1020RDB-PC_36BIT_SPIFLASH T1042RDB_PI_SDCARD
> > > T2081QDS_NAND T4160QDS_SDCARD P1010RDB-PB_SPIFLASH T2080QDS_SDCARD
> > > P1020MBG-PC_SDCARD P1020RDB-PC_SPIFLASH T4240QDS_SDCARD
> > > P2020RDB-PC_36BIT_SDCARD T1040RDB_SDCARD P1021RDB-PC_SDCARD
> > > B4860QDS_NAND T2080RDB_SDCARD P1020RDB-PC_36BIT_SDCARD
> > > T1040D4RDB_SDCARD T1040D4RDB_SPIFLASH P1020RDB-PC_NAND T2080QDS_NAND
> > > T1023RDB_NAND P1020RDB-PD_NAND T1040D4RDB_NAND P1010RDB-PA_NAND
> > > T1040RDB_SPIFLASH P2020RDB-PC_SDCARD T2081QDS_SDCARD
> > > P1010RDB-PA_36BIT_SDCARD P1024RDB_NAND P1024RDB_SDCARD T4240QDS_NAND
> > > C29XPCIE_NAND B4420QDS_NAND P1021RDB-PC_36BIT_SPIFLASH
> > > P1025RDB_SPIFLASH T2081QDS_SPIFLASH T1040RDB_NAND P1010RDB-PA_SDCARD
> > > P1010RDB-PB_NAND T1024QDS_SDCARD T1024QDS_SPIFLASH P1021RDB-PC_NAND
> > > T1042D4RDB_SPIFLASH P1020UTM-PC_36BIT_SDCARD T4160QDS_NAND
> > > T2080RDB_SPIFLASH P1022DS_SPIFLASH T1042D4RDB_SDCARD
> > > P1021RDB-PC_36BIT_SDCARD P1020RDB-PD_SDCARD P1022DS_36BIT_SPIFLASH
> > > T1024RDB_SDCARD P1020UTM-PC_SDCARD P2020RDB-PC_36BIT_NAND
> > > P2020RDB-PC_NAND T1023RDB_SPIFLASH P1010RDB-PB_36BIT_SPIFLASH
> > > P1025RDB_NAND P1010RDB-PA_36BIT_SPIFLASH
> > >   blackfin: (for 34/35 boards)  all +16.0  rodata +16.0
> > >        x86: (for 9/9 boards)  bss -14.2  rodata +14.2
> > >    aarch64: (for 13/13 boards)  all +17.2  rodata +17.2
> > > spl/u-boot-spl:all +151.4  spl/u-boot-spl:rodata +151.4
> > >      avr32: (for 4/4 boards)  all +16.0  rodata +16.0
> > >    sandbox: (for 1/1 boards)  all +16.0  rodata +16.0
> > >       m68k: (for 48/48 boards)  all +11.7  text +11.7
> > >    powerpc: (for 412/416 boards)  all +3.6  data +1.6  rodata +2.0
> > > spl/u-boot-spl:all +5.8  spl/u-boot-spl:rodata +5.0
> > > spl/u-boot-spl:text +0.8
> > >      sparc: (for 5/5 boards)  all +32.0  data +16.0  text +16.0
> > >         sh: (for 21/22 boards)  all +16.0  rodata +16.0
> > >      nios2: (for 1/1 boards)  all +16.0  data +16.0
> > >       mips: (for 24/24 boards)  all +20.0  bss +4.0  rodata +16.0
> > >        arm: (for 531/531 boards)  all +17.7  bss +0.3  rodata +17.4
> > > spl/u-boot-spl:all +69.0  spl/u-boot-spl:rodata +69.0
> > >      nds32: (for 1/1 boards)  all +16.0  rodata +16.0
> >
> > Just what I was worried about.
> >
> > > So there are some code size changes (which may not be a big deal at 15
> > > -60 bytes) but it breaks lots of powerpc boards.
> > >
> > > Can we instead add CONFIG_SPL_OF_LIBFDT?
> >
> > What would be the difference between having CONFIG_OF_LIBFDT enabled in
> > SPL (which is the case now) and having CONFIG_SPL_OF_LIBFDT enabled ?
> > The naming seems a bit difficult to comprehend to me.
>
> Sorry for late reply as fully tight up with something.
>
> To minimize the impact, we can create fdtdec_get_uint32 in
> lib/fdtdec_common.c.
> Then we can drop this patch.
> Any thought?

Sounds good. Or you could use fdtdec_get_int().

Regards,
Simon