[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <87y1icdaoq.fsf@mail.lhotse>
Date: Tue, 15 Aug 2023 17:30:45 +1000
From: Michael Ellerman <mpe@...erman.id.au>
To: Linus Walleij <linus.walleij@...aro.org>,
Nicholas Piggin <npiggin@...il.com>,
Christophe Leroy <christophe.leroy@...roup.eu>
Cc: linuxppc-dev@...ts.ozlabs.org, linux-kernel@...r.kernel.org,
kvm@...r.kernel.org, Linus Walleij <linus.walleij@...aro.org>
Subject: Re: [PATCH] powerpc: Make virt_to_pfn() a static inline
Linus Walleij <linus.walleij@...aro.org> writes:
> Making virt_to_pfn() a static inline taking a strongly typed
> (const void *) makes the contract of a passing a pointer of that
> type to the function explicit and exposes any misuse of the
> macro virt_to_pfn() acting polymorphic and accepting many types
> such as (void *), (unitptr_t) or (unsigned long) as arguments
> without warnings.
>
> Move the virt_to_pfn() and related functions below the
> declaration of __pa() so it compiles.
>
> For symmetry do the same with pfn_to_kaddr().
>
> As the file is included right into the linker file, we need
> to surround the functions with ifndef __ASSEMBLY__ so we
> don't cause compilation errors.
>
> The conversion moreover exposes the fact that pmd_page_vaddr()
> was returning an unsigned long rather than a const void * as
> could be expected, so all the sites defining pmd_page_vaddr()
> had to be augmented as well.
...
> diff --git a/arch/powerpc/include/asm/pgtable.h b/arch/powerpc/include/asm/pgtable.h
> index 6a88bfdaa69b..a9515d3d7831 100644
> --- a/arch/powerpc/include/asm/pgtable.h
> +++ b/arch/powerpc/include/asm/pgtable.h
> @@ -60,9 +60,9 @@ static inline pgprot_t pte_pgprot(pte_t pte)
> }
>
> #ifndef pmd_page_vaddr
> -static inline unsigned long pmd_page_vaddr(pmd_t pmd)
> +static inline const void *pmd_page_vaddr(pmd_t pmd)
> {
> - return ((unsigned long)__va(pmd_val(pmd) & ~PMD_MASKED_BITS));
> + return (const void *)((unsigned long)__va(pmd_val(pmd) & ~PMD_MASKED_BITS));
This can also just be:
return __va(pmd_val(pmd) & ~PMD_MASKED_BITS);
I've squashed that in.
cheers
Powered by blists - more mailing lists