[PATCH v2] soc: renesas: make ARCH_R9A07G043 depend on required options

[PATCH v2] soc: renesas: make ARCH_R9A07G043 depend on required options

[Conor Dooley]

From: Conor Dooley <conor.dooley@microchip.com>

Randy reported a randconfig build issue against linux-next:
WARNING: unmet direct dependencies detected for ERRATA_ANDES
  Depends on [n]: RISCV_ALTERNATIVE [=n] && RISCV_SBI [=y]
  Selected by [y]:
  - ARCH_R9A07G043 [=y] && SOC_RENESAS [=y] && RISCV [=y] && NONPORTABLE [=y] && RISCV_SBI [=y]

../arch/riscv/errata/andes/errata.c:59:54: warning: 'struct alt_entry' declared inside parameter list will not be visible outside of this definition or declaration
   59 | void __init_or_module andes_errata_patch_func(struct alt_entry *begin, struct alt_entry *end,

On RISC-V, alternatives are not usable in XIP kernels, which this
randconfig happened to select. Rather than add a check for whether
alternatives are available before selecting the ERRATA_ANDES config
option, rework the R9A07G043 Kconfig entry to depend on the
configuration options required to support its non-standard cache
coherency implementation.

Without these options enabled, the SoC is effectively non-functional to
begin with, so there's an extra benefit in preventing the creation of
non-functional kernels.

The "if RISCV_DMA_NONCOHERENT" can be dropped, as ERRATA_ANDES_CMO will
select it.

Reported-by: Randy Dunlap <rdunlap@infradead.org>
Closes: https://lore.kernel.org/all/09a6b0f0-76a1-45e3-ab52-329c47393d1d@infradead.org/
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
---

I dropped Randy's t-b etc since this patch is quite different.

v2: drop the extra condition on the select of ERRATA_ANDES, move instead
to depending on required options.

CC: Geert Uytterhoeven <geert+renesas@glider.be>
CC: Magnus Damm <magnus.damm@gmail.com>
CC: Paul Walmsley <paul.walmsley@sifive.com>
CC: Palmer Dabbelt <palmer@dabbelt.com>
CC: Albert Ou <aou@eecs.berkeley.edu>
CC: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
CC: linux-renesas-soc@vger.kernel.org
CC: linux-kernel@vger.kernel.org
CC: linux-riscv@lists.infradead.org
---
 drivers/soc/renesas/Kconfig | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/soc/renesas/Kconfig b/drivers/soc/renesas/Kconfig
index 7b74de732718..adedf02897c6 100644
--- a/drivers/soc/renesas/Kconfig
+++ b/drivers/soc/renesas/Kconfig
@@ -340,10 +340,12 @@ if RISCV
 config ARCH_R9A07G043
 	bool "RISC-V Platform support for RZ/Five"
 	depends on NONPORTABLE
+	depends on RISCV_ALTERNATIVE
+	depends on RISCV_SBI
 	select ARCH_RZG2L
-	select AX45MP_L2_CACHE if RISCV_DMA_NONCOHERENT
+	select AX45MP_L2_CACHE
 	select DMA_GLOBAL_POOL
-	select ERRATA_ANDES if RISCV_SBI
+	select ERRATA_ANDES
 	select ERRATA_ANDES_CMO if ERRATA_ANDES
 	help
 	  This enables support for the Renesas RZ/Five SoC.
-- 
2.39.2

Re: [PATCH v2] soc: renesas: make ARCH_R9A07G043 depend on required options

[Geert Uytterhoeven]

Hi Conor,

On Thu, Oct 12, 2023 at 1:11 PM Conor Dooley <conor@kernel.org> wrote:
> From: Conor Dooley <conor.dooley@microchip.com>
>
> Randy reported a randconfig build issue against linux-next:
> WARNING: unmet direct dependencies detected for ERRATA_ANDES
>   Depends on [n]: RISCV_ALTERNATIVE [=n] && RISCV_SBI [=y]
>   Selected by [y]:
>   - ARCH_R9A07G043 [=y] && SOC_RENESAS [=y] && RISCV [=y] && NONPORTABLE [=y] && RISCV_SBI [=y]
>
> ../arch/riscv/errata/andes/errata.c:59:54: warning: 'struct alt_entry' declared inside parameter list will not be visible outside of this definition or declaration
>    59 | void __init_or_module andes_errata_patch_func(struct alt_entry *begin, struct alt_entry *end,
>
> On RISC-V, alternatives are not usable in XIP kernels, which this
> randconfig happened to select. Rather than add a check for whether
> alternatives are available before selecting the ERRATA_ANDES config
> option, rework the R9A07G043 Kconfig entry to depend on the
> configuration options required to support its non-standard cache
> coherency implementation.
>
> Without these options enabled, the SoC is effectively non-functional to
> begin with, so there's an extra benefit in preventing the creation of
> non-functional kernels.
>
> The "if RISCV_DMA_NONCOHERENT" can be dropped, as ERRATA_ANDES_CMO will
> select it.
>
> Reported-by: Randy Dunlap <rdunlap@infradead.org>
> Closes: https://lore.kernel.org/all/09a6b0f0-76a1-45e3-ab52-329c47393d1d@infradead.org/
> Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
> ---
>
> I dropped Randy's t-b etc since this patch is quite different.
>
> v2: drop the extra condition on the select of ERRATA_ANDES, move instead
> to depending on required options.

Thanks for the update!

Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
i.e. will queue in renesas-fixes for v6.6.

> --- a/drivers/soc/renesas/Kconfig
> +++ b/drivers/soc/renesas/Kconfig
> @@ -340,10 +340,12 @@ if RISCV
>  config ARCH_R9A07G043
>         bool "RISC-V Platform support for RZ/Five"
>         depends on NONPORTABLE
> +       depends on RISCV_ALTERNATIVE
> +       depends on RISCV_SBI
>         select ARCH_RZG2L
> -       select AX45MP_L2_CACHE if RISCV_DMA_NONCOHERENT
> +       select AX45MP_L2_CACHE
>         select DMA_GLOBAL_POOL
> -       select ERRATA_ANDES if RISCV_SBI
> +       select ERRATA_ANDES
>         select ERRATA_ANDES_CMO if ERRATA_ANDES

As ERRATA_ANDES is now selected unconditionally, the test
for it can be removed.  I can do that while applying.

>         help
>           This enables support for the Renesas RZ/Five SoC.

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds