Debugging dtoc?

[Simon Glass]

Hi again,

On Mon, 26 Jul 2021 at 08:06, Simon Glass <sjg at chromium.org> wrote:
>
> Hi Tom,
>
> On Sun, 25 Jul 2021 at 15:10, Tom Rini <trini at konsulko.com> wrote:
> >
> > So, I'm trying to fix the problem on am335x_evm (and some family
> > configs) with needing SPL_OF_CONTROL enabled.  This is mostly fine just
> > enabling the option, except on am335x_evm itself, which is the
> > kitchen-sink config and overflows memory.  I've gone with switching to
> > SPL_OF_PLATDATA there as am335x in general has all of the U_BOOT_DRVINFO
> > entries it needs I believe.  But, with the following patch:
> >
> > diff --git a/arch/arm/dts/am335x-evm-u-boot.dtsi b/arch/arm/dts/am335x-evm-u-boot.dtsi
> > index 4cf5f9928d58..514f682cac99 100644
> > --- a/arch/arm/dts/am335x-evm-u-boot.dtsi
> > +++ b/arch/arm/dts/am335x-evm-u-boot.dtsi
> > @@ -8,6 +8,7 @@
> >  &l4_per {
> >
> >         segment at 300000 {
> > +               u-boot,dm-pre-reloc;
> >
> >                 target-module at e000 {
> >                         u-boot,dm-pre-reloc;
> > diff --git a/configs/am335x_boneblack_vboot_defconfig b/configs/am335x_boneblack_vboot_defconfig
> > index a0baeec79edd..ffeefd1a0087 100644
> > --- a/configs/am335x_boneblack_vboot_defconfig
> > +++ b/configs/am335x_boneblack_vboot_defconfig
> > @@ -31,6 +31,7 @@ CONFIG_CMD_SPL=y
> >  # CONFIG_CMD_SETEXPR is not set
> >  CONFIG_BOOTP_DNS2=y
> >  CONFIG_OF_CONTROL=y
> > +CONFIG_SPL_OF_CONTROL=y
> >  CONFIG_ENV_OVERWRITE=y
> >  CONFIG_ENV_IS_IN_MMC=y
> >  CONFIG_SYS_REDUNDAND_ENVIRONMENT=y
> > diff --git a/configs/am335x_evm_defconfig b/configs/am335x_evm_defconfig
> > index a33efff42a74..f35b2a02f56b 100644
> > --- a/configs/am335x_evm_defconfig
> > +++ b/configs/am335x_evm_defconfig
> > @@ -37,13 +37,16 @@ CONFIG_MTDIDS_DEFAULT="nand0=nand.0"
> >  CONFIG_MTDPARTS_DEFAULT="mtdparts=nand.0:128k(NAND.SPL),128k(NAND.SPL.backup1),128k(NAND.SPL.backup2),128k(NAND.SPL.backup3),256k(NAND.u-boot-spl-os),1m(NAND.u-boot),128k(NAND.u-boot-env),128k(NAND.u-boot-env.backup1),8m(NAND.kernel),-(NAND.file-system)"
> >  # CONFIG_SPL_EFI_PARTITION is not set
> >  CONFIG_OF_CONTROL=y
> > +CONFIG_SPL_OF_CONTROL=y
> >  CONFIG_OF_LIST="am335x-evm am335x-bone am335x-boneblack am335x-evmsk am335x-bonegreen am335x-icev2 am335x-pocketbeagle"
> > +CONFIG_SPL_OF_PLATDATA=y
> >  CONFIG_ENV_OVERWRITE=y
> >  CONFIG_SYS_RELOC_GD_ENV_ADDR=y
> >  CONFIG_ENV_VARS_UBOOT_RUNTIME_CONFIG=y
> >  CONFIG_SPL_ENV_IS_NOWHERE=y
> >  CONFIG_VERSION_VARIABLE=y
> >  CONFIG_BOOTP_SEND_HOSTNAME=y
> > +# CONFIG_SPL_SIMPLE_BUS is not set
> >  CONFIG_BOOTCOUNT_LIMIT=y
> >  CONFIG_CLK=y
> >  CONFIG_CLK_CDCE9XX=y
> > diff --git a/configs/am335x_evm_spiboot_defconfig b/configs/am335x_evm_spiboot_defconfig
> > index 8f0c330674a9..4a2a56a9af9e 100644
> > --- a/configs/am335x_evm_spiboot_defconfig
> > +++ b/configs/am335x_evm_spiboot_defconfig
> > @@ -32,6 +32,7 @@ CONFIG_BOOTP_DNS2=y
> >  CONFIG_CMD_MTDPARTS=y
> >  # CONFIG_SPL_EFI_PARTITION is not set
> >  CONFIG_OF_CONTROL=y
> > +CONFIG_SPL_OF_CONTROL=y
> >  CONFIG_OF_LIST="am335x-evm am335x-bone am335x-boneblack am335x-evmsk am335x-bonegreen am335x-icev2 am335x-pocketbeagle"
> >  CONFIG_ENV_OVERWRITE=y
> >  # CONFIG_ENV_IS_IN_FAT is not set
> >
> > I get the following failure and I don't see how to debug this:
> >   DTOC    spl/dts/dt-plat.c
> > Traceback (most recent call last):
> >   File "./tools/dtoc/dtoc", line 115, in <module>
> >     args.phase, instantiate=args.instantiate)
> >   File "/home/trini/work/u-boot/u-boot/tools/dtoc/../dtoc/dtb_platdata.py", line 1223, in run_steps
> >     outfile.method(plat)
> >   File "/home/trini/work/u-boot/u-boot/tools/dtoc/../dtoc/dtb_platdata.py", line 1081, in generate_plat
> >     self.output_node_plat(node)
> >   File "/home/trini/work/u-boot/u-boot/tools/dtoc/../dtoc/dtb_platdata.py", line 1023, in output_node_plat
> >     self._output_values(node)
> >   File "/home/trini/work/u-boot/u-boot/tools/dtoc/../dtoc/dtb_platdata.py", line 812, in _output_values
> >     self._output_prop(node, node.props[pname])
> >   File "/home/trini/work/u-boot/u-boot/tools/dtoc/../dtoc/dtb_platdata.py", line 798, in _output_prop
> >     self._output_list(node, prop)
> >   File "/home/trini/work/u-boot/u-boot/tools/dtoc/../dtoc/dtb_platdata.py", line 628, in _output_list
> >     vals.append(get_value(prop.type, val))
> >   File "/home/trini/work/u-boot/u-boot/tools/dtoc/../dtoc/dtb_platdata.py", line 126, in get_value
> >     val = '%#x' % fdt_util.fdt32_to_cpu(value)
> >   File "/home/trini/work/u-boot/u-boot/tools/dtoc/../dtoc/fdt_util.py", line 28, in fdt32_to_cpu
> >     return struct.unpack('>I', val)[0]
> > TypeError: a bytes-like object is required, not 'bool'
> > scripts/Makefile.spl:352: recipe for target 'spl/dts/dt-plat.c' failed
> > make[1]: *** [spl/dts/dt-plat.c] Error 1
> > make[1]: *** Deleting file 'spl/dts/dt-plat.c'
> > Makefile:1999: recipe for target 'spl/u-boot-spl' failed
> > make: *** [spl/u-boot-spl] Error 2
>
> That seems like a bug, where perhaps it is seeing a property with no
> value so it thinks it is a bool, but then somehow tries to gets its
> int value.
>
> I added a print to  _output_list()
>
>         else:
>             print('node', node.path, prop.name)
>             for val in prop.value:
>                 vals.append(get_value(prop.type, val))
>
> and see that it is the 'ranges' property. Definitely seems like a bug
> but I'll have to dig into it I think. The Prop() class selects a bool
> type when the value is empty, but somehow that isn't happening.

This is actually a little tricky, but I'm looking at it.

Regards,
Simon