Re: (no subject)

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Conditionally compiling the OF 'isa' bus translator based on `CONFIG_ISA` breaks device tree address translation for non-ISA architectures (e.g., ARM64) that use 'isa' nodes for LPC or legacy I/O bridges.
--

commit 3f535ae9ca6c863a12c1d6dbcfcb4358a6f6ed4b
Author: Daniel Palmer <daniel@thingy.jp>

of/address: Drop ISA parts when !CONFIG_ISA

This patch conditionally compiles the ISA parts of the Open Firmware address
translator based on CONFIG_ISA to reduce code size.

> diff --git a/drivers/of/address.c b/drivers/of/address.c
> index cf4aab11e9b1f..9b2f0d462fb59 100644
> --- a/drivers/of/address.c
> +++ b/drivers/of/address.c
[ ... ]
> @@ -349,6 +351,7 @@ static const struct of_bus of_busses[] = {
>  		.get_flags = of_bus_pci_get_flags,
>  	},
>  #endif /* CONFIG_PCI */
> +#ifdef CONFIG_ISA
>  	/* ISA */
>  	{
>  		.name = "isa",
> @@ -360,6 +363,7 @@ static const struct of_bus of_busses[] = {
>  		.flag_cells = 1,
>  		.get_flags = of_bus_isa_get_flags,
>  	},
> +#endif /* CONFIG_ISA */

[Severity: High]
Does this break device tree address translation for non-ISA architectures (like
ARM64) that use 'isa' nodes for LPC or legacy I/O bridges?

Platforms like Hisilicon hip06/hip07 don't have CONFIG_ISA defined but still
use 'isa' nodes. If this bus entry is compiled out, they will fall back to the
'default' OF bus translator.

The 'default' bus translator's get_flags method (of_bus_default_get_flags)
unconditionally returns IORESOURCE_MEM. This causes __of_address_to_resource()
to misclassify the I/O port and call of_translate_address() instead of
of_translate_ioport().

Because Hisilicon LPC registers an indirect I/O host, of_translate_address()
then fails with OF_BAD_ADDR (-EINVAL), causing probe failures for devices like
UART and IPMI.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260626171747.202269-1-daniel@thingy.jp?part=1