Re: [PATCH 0/4] Remove problematic include in <asm/set_memory.h>

[David Hildenbrand]

On 10.12.24 19:46, Kevin Brodsky wrote:
> The need for this series arose from a completely unrelated series [1].
> Long story short, that series causes <linux/mm.h> to include
> <linux/set_memory.h>, which doesn't feel too unreasonable.
>
> That works fine on arm64 and probably most other architectures, but not
> on x86 [2]: <asm/set_memory.h> itself includes <linux/mm.h>, creating a
> circular dependency.
>
> set_memory.h doesn't really need <linux/mm.h>, so removing that include
> seems like the right thing to do. That turned out to be a little more
> involved than expected, hence this series:
>
> * Patch 1-2 ensure that code doesn't rely on <asm/set_memory.h>
>    including <linux/mm.h>. The errors that these patches fix are included
>    at the end of this email.
>
> * Patch 3 removes an unused function whose declaration relied on
>    <linux/mm.h> being included.
>   
> * Patch 4 actually remove the include.
>
> I've build-tested this series with x86_64_defconfig and allyesconfig.
>
> - Kevin
>
> [1] https://lore.kernel.org/linux-hardening/20241206101110.1646108-1-kevin.brodsky@xxxxxxx/
> [2] https://lore.kernel.org/oe-kbuild-all/202412062006.C23V9ESs-lkp@xxxxxxxxx/
>
> Cc: bp@xxxxxxxxx
> Cc: dan.j.williams@xxxxxxxxx
> Cc: dave.hansen@xxxxxxxxxxxxxxx
> Cc: david@xxxxxxxxxx
> Cc: jane.chu@xxxxxxxxxx
> Cc: osalvador@xxxxxxx
> Cc: tglx@xxxxxxxxxxxxx
> ---

LGTM

Acked-by: David Hildenbrand <david@xxxxxxxxxx>

--
Cheers,

David / dhildenb