Re: [PATCH] acpi: Fix ARM32 platforms compile issue introduced by fw_table changes

[Sam Edwards]

On 11/21/23 07:58, Dave Jiang wrote:
> Linus reported that:
> After commit a103f46633fd the kernel stopped compiling for
> several ARM32 platforms that I am building with a bare metal
> compiler. Bare metal compilers (arm-none-eabi-) don't
> define __linux__.
>
> This is because the header <acpi/platform/acenv.h> is now
> in the include path for <linux/irq.h>:
>
>    CC      arch/arm/kernel/irq.o
>    CC      kernel/sysctl.o
>    CC      crypto/api.o
> In file included from ../include/acpi/acpi.h:22,
>                   from ../include/linux/fw_table.h:29,
>                   from ../include/linux/acpi.h:18,
>                   from ../include/linux/irqchip.h:14,
>                   from ../arch/arm/kernel/irq.c:25:
> ../include/acpi/platform/acenv.h:218:2: error: #error Unknown target environment
>    218 | #error Unknown target environment
>        |  ^~~~~
>
> The issue is caused by the introducing of splitting out the ACPI code to
> support the new generic fw_table code.
>
> Rafael suggested moving the fw_table.h include in linux/acpi.h to below
> the asm/acpi.h. The move also helped with eliminating the inclusion of
> acpi/acpi.h in fw_table.h. The unfortunate circular inclusion of
> linux/acpi.h is needed for fw_table.h due fw_table code needing the
> defined acpi structs in order to build.
>
> Fixes: a103f46633fd ("acpi: Move common tables helper functions to common lib")
> Reported-by: Linus Walleij <linus.walleij@xxxxxxxxxx>
> Suggested-by: Rafael J. Wysocki <rafael.j.wysocki@xxxxxxxxx>
> Signed-off-by: Dave Jiang <dave.jiang@xxxxxxxxx>
> ---
>   include/linux/acpi.h     |   23 ++++++++++++-----------
>   include/linux/fw_table.h |    1 -
>   2 files changed, 12 insertions(+), 12 deletions(-)
>
> diff --git a/include/linux/acpi.h b/include/linux/acpi.h
> index 54189e0e5f41..2789beb26138 100644
> --- a/include/linux/acpi.h
> +++ b/include/linux/acpi.h
> @@ -15,7 +15,6 @@
>   #include <linux/mod_devicetable.h>
>   #include <linux/property.h>
>   #include <linux/uuid.h>
> -#include <linux/fw_table.h>
>   
>   struct irq_domain;
>   struct irq_domain_ops;
> @@ -25,16 +24,6 @@ struct irq_domain_ops;
>   #endif
>   #include <acpi/acpi.h>
>   
> -#ifdef CONFIG_ACPI_TABLE_LIB
> -#define EXPORT_SYMBOL_ACPI_LIB(x) EXPORT_SYMBOL_NS_GPL(x, ACPI)
> -#define __init_or_acpilib
> -#define __initdata_or_acpilib
> -#else
> -#define EXPORT_SYMBOL_ACPI_LIB(x)
> -#define __init_or_acpilib __init
> -#define __initdata_or_acpilib __initdata
> -#endif
> -
>   #ifdef	CONFIG_ACPI
>   
>   #include <linux/list.h>
> @@ -48,6 +37,18 @@ struct irq_domain_ops;
>   #include <acpi/acpi_io.h>
>   #include <asm/acpi.h>
>   
> +#include <linux/fw_table.h>
> +
> +#ifdef CONFIG_ACPI_TABLE_LIB
> +#define EXPORT_SYMBOL_ACPI_LIB(x) EXPORT_SYMBOL_NS_GPL(x, ACPI)
> +#define __init_or_acpilib
> +#define __initdata_or_acpilib
> +#else
> +#define EXPORT_SYMBOL_ACPI_LIB(x)
> +#define __init_or_acpilib __init
> +#define __initdata_or_acpilib __initdata
> +#endif
> +
>   static inline acpi_handle acpi_device_handle(struct acpi_device *adev)
>   {
>   	return adev ? adev->handle : NULL;
> diff --git a/include/linux/fw_table.h b/include/linux/fw_table.h
> index ff8fa58d5818..a722300c215b 100644
> --- a/include/linux/fw_table.h
> +++ b/include/linux/fw_table.h
> @@ -26,7 +26,6 @@ struct acpi_subtable_proc {
>   };
>   
>   #include <linux/acpi.h>
> -#include <acpi/acpi.h>

Hi Dave,

Seems to me that the #include <linux/acpi.h> should go too, to break the 
circular including cycle. If it remains, I fear that there could be 
subtle problems in the future depending on which header is included 
first in a compilation unit. It sounds now like the only correct way to 
get fw_table.h included is transitively via linux/acpi.h (of note: 
lib/fw_table.c will have to be updated; it's the only file that 
currently breaks this rule) so that removal will just help enforce this. 
Plus, includes in the middle of non-preprocessor declarations are a 
(sometimes necessary, definitely not here) code smell, in my view.

If this include must remain for some reason, perhaps a comment should be 
added to call attention to the circular situation and provide justification?

Cheers,
Sam

>   
>   union acpi_subtable_headers {
>   	struct acpi_subtable_header common;