Received-SPF: pass (google.com: domain of linux-kernel+bounces-189272-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Date: Sat, 25 May 2024 06:12:54 +0200
From: Oscar Salvador <osalvador@suse.de>
To: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: Andrew Morton <akpm@linux-foundation.org>,
	Jason Gunthorpe <jgg@nvidia.com>, Peter Xu <peterx@redhat.com>,
	Michael Ellerman <mpe@ellerman.id.au>,
	Nicholas Piggin <npiggin@gmail.com>, linux-kernel@vger.kernel.org,
	linux-mm@kvack.org, linuxppc-dev@lists.ozlabs.org
Subject: Re: [RFC PATCH v2 10/20] powerpc/mm: Fix __find_linux_pte() on 32
 bits with PMD leaf entries
Message-ID: <ZlFlRr26AvS6n7p8@localhost.localdomain>
References: <cover.1715971869.git.christophe.leroy@csgroup.eu>
 <41bf05806501c0091a7d52e118b187416d24a76f.1715971869.git.christophe.leroy@csgroup.eu>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <41bf05806501c0091a7d52e118b187416d24a76f.1715971869.git.christophe.leroy@csgroup.eu>

On Fri, May 17, 2024 at 09:00:04PM +0200, Christophe Leroy wrote:
> Building on 32 bits with pmd_leaf() not returning always false leads
> to the following error:

I am curious though.
pmd_leaf is only defined in include/linux/pgtable.h for 32bits, and is hardcoded
to false.
I do not see where we change it in previous patches, so is this artificial?

> 
>   CC      arch/powerpc/mm/pgtable.o
> arch/powerpc/mm/pgtable.c: In function '__find_linux_pte':
> arch/powerpc/mm/pgtable.c:506:1: error: function may return address of local variable [-Werror=return-local-addr]
>   506 | }
>       | ^
> arch/powerpc/mm/pgtable.c:394:15: note: declared here
>   394 |         pud_t pud, *pudp;
>       |               ^~~
> arch/powerpc/mm/pgtable.c:394:15: note: declared here
> 
> This is due to pmd_offset() being a no-op in that case.

This is because 32bits powerpc include pgtable-nopmd.h?

> So rework it for powerpc/32 so that pXd_offset() are used on real
> pointers and not on on-stack copies.
> 
> Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
> ---
>  arch/powerpc/mm/pgtable.c | 14 ++++++++++++--
>  1 file changed, 12 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/powerpc/mm/pgtable.c b/arch/powerpc/mm/pgtable.c
> index 59f0d7706d2f..51ee508eeb5b 100644
> --- a/arch/powerpc/mm/pgtable.c
> +++ b/arch/powerpc/mm/pgtable.c
> @@ -390,8 +390,12 @@ pte_t *__find_linux_pte(pgd_t *pgdir, unsigned long ea,
>  			bool *is_thp, unsigned *hpage_shift)
>  {
>  	pgd_t *pgdp;
> -	p4d_t p4d, *p4dp;
> -	pud_t pud, *pudp;
> +	p4d_t *p4dp;
> +	pud_t *pudp;
> +#ifdef CONFIG_PPC64
> +	p4d_t p4d;
> +	pud_t pud;
> +#endif
>  	pmd_t pmd, *pmdp;
>  	pte_t *ret_pte;
>  	hugepd_t *hpdp = NULL;
> @@ -412,6 +416,7 @@ pte_t *__find_linux_pte(pgd_t *pgdir, unsigned long ea,
>  	 */
>  	pgdp = pgdir + pgd_index(ea);
>  	p4dp = p4d_offset(pgdp, ea);
> +#ifdef CONFIG_PPC64
>  	p4d  = READ_ONCE(*p4dp);
>  	pdshift = P4D_SHIFT;
>  
> @@ -452,6 +457,11 @@ pte_t *__find_linux_pte(pgd_t *pgdir, unsigned long ea,
>  
>  	pdshift = PMD_SHIFT;
>  	pmdp = pmd_offset(&pud, ea);
> +#else
> +	p4dp = p4d_offset(pgdp, ea);
> +	pudp = pud_offset(p4dp, ea);
> +	pmdp = pmd_offset(pudp, ea);

I would drop a comment on top explaining that these are no-op for 32bits,
otherwise it might not be obvious to people as why this distiction between 64 and
32bits.

Other than that looks good to me

 

-- 
Oscar Salvador
SUSE Labs