Re: [PATCH v2] mm: use clear_user_(high)page() for arch with special user folio handling

[Mathieu Desnoyers <mathieu.desnoyers@efficios.com>]

On 2024-12-06 12:42, Zi Yan wrote:
> For architectures setting cpu_dcache_is_aliasing() to true, which require
> flushing cache, and arc, which changes folio->flags after clearing a user
> folio, __GFP_ZERO using only clear_page() is not enough to zero user
> folios and clear_user_(high)page() must be used. Otherwise, user data
> will be corrupted.
> 
> Fix it by always clearing user folios with clear_user_(high)page() when
> cpu_dcache_is_aliasing() is true or architecture is arc. Rename
> alloc_zeroed() to alloc_need_zeroing() and invert the logic to clarify its
> intend.
> 
> Fixes: 5708d96da20b ("mm: avoid zeroing user movable page twice with init_on_alloc=1")
> Reported-by: Geert Uytterhoeven <geert+renesas@glider.be>
> Closes: https://lore.kernel.org/linux-mm/CAMuHMdV1hRp_NtR5YnJo=HsfgKQeH91J537Gh4gKk3PFZhSkbA@mail.gmail.com/
> Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
> Signed-off-by: Zi Yan <ziy@nvidia.com>
> ---
>   include/linux/highmem.h |  8 +++++++-
>   include/linux/mm.h      | 17 +++++++++++++++++
>   mm/huge_memory.c        |  9 +++++----
>   mm/internal.h           |  6 ------
>   mm/memory.c             | 10 +++++-----
>   5 files changed, 34 insertions(+), 16 deletions(-)
> 
> diff --git a/include/linux/highmem.h b/include/linux/highmem.h
> index 6e452bd8e7e3..d9beb8371daa 100644
> --- a/include/linux/highmem.h
> +++ b/include/linux/highmem.h
> @@ -224,7 +224,13 @@ static inline
>   struct folio *vma_alloc_zeroed_movable_folio(struct vm_area_struct *vma,
>   				   unsigned long vaddr)
>   {
> -	return vma_alloc_folio(GFP_HIGHUSER_MOVABLE | __GFP_ZERO, 0, vma, vaddr);
> +	struct folio *folio;
> +
> +	folio = vma_alloc_folio(GFP_HIGHUSER_MOVABLE, 0, vma, vaddr);
> +	if (folio && alloc_need_zeroing())
> +		clear_user_highpage(&folio->page, vaddr);
> +
> +	return folio;
>   }
>   #endif
>   
> diff --git a/include/linux/mm.h b/include/linux/mm.h
> index c39c4945946c..ca8df5871213 100644
> --- a/include/linux/mm.h
> +++ b/include/linux/mm.h
> @@ -31,6 +31,7 @@
>   #include <linux/kasan.h>
>   #include <linux/memremap.h>
>   #include <linux/slab.h>
> +#include <linux/cacheinfo.h>
>   
>   struct mempolicy;
>   struct anon_vma;
> @@ -4175,6 +4176,22 @@ static inline int do_mseal(unsigned long start, size_t len_in, unsigned long fla
>   }
>   #endif
>   
> +/*
> + * alloc_need_zeroing checks if a user folio from page allocator needs to be
> + * zeroed or not.
> + */
> +static inline bool alloc_need_zeroing(void)
> +{
> +	/*
> +	 * for user folios, arch with cache aliasing requires cache flush and
> +	 * arc changes folio->flags, so always return false to make caller use
> +	 * clear_user_page()/clear_user_highpage()
> +	 */
> +	return (cpu_dcache_is_aliasing() || IS_ENABLED(CONFIG_ARC)) ||

Nack.

Can we please not go back to re-introducing arch-specific
conditionals in generic mm code after the cleanup I did when
introducing cpu_dcache_is_aliasing() in commit 8690bbcf3b70 ?

Based on commit eacd0e950dc2, AFAIU what you appear to need here
is to introduce a:

cpu_icache_is_aliasing() -> note the "i" for instruction cache

It would typically be directly set to

#define cpu_icache_is_aliasing() cpu_dcache_is_aliasing()

except on architecture like ARC when the icache vs dcache
is aliasing, but not dcache vs dcache.

So for ARC it would be defined as:

#define cpu_dcache_is_aliasing() false
#define cpu_icache_is_aliasing() true

And the Kconfig ARCH_HAS_CPU_CACHE_ALIASING=y would be set for ARC
again.

I'm not entirely sure if we want to go for the wording "is_aliasing"
or "is_incoherent" when talking about icache vs dcache, so I'm open
to ideas here.

Thanks,

Mathieu

> +	       !static_branch_maybe(CONFIG_INIT_ON_ALLOC_DEFAULT_ON,
> +				   &init_on_alloc);
> +}
> +
>   int arch_get_shadow_stack_status(struct task_struct *t, unsigned long __user *status);
>   int arch_set_shadow_stack_status(struct task_struct *t, unsigned long status);
>   int arch_lock_shadow_stack_status(struct task_struct *t, unsigned long status);
> diff --git a/mm/huge_memory.c b/mm/huge_memory.c
> index ee335d96fc39..107130a5413a 100644
> --- a/mm/huge_memory.c
> +++ b/mm/huge_memory.c
> @@ -1176,11 +1176,12 @@ static struct folio *vma_alloc_anon_folio_pmd(struct vm_area_struct *vma,
>   	folio_throttle_swaprate(folio, gfp);
>   
>          /*
> -	* When a folio is not zeroed during allocation (__GFP_ZERO not used),
> -	* folio_zero_user() is used to make sure that the page corresponding
> -	* to the faulting address will be hot in the cache after zeroing.
> +	* When a folio is not zeroed during allocation (__GFP_ZERO not used)
> +	* or user folios require special handling, folio_zero_user() is used to
> +	* make sure that the page corresponding to the faulting address will be
> +	* hot in the cache after zeroing.
>   	*/
> -	if (!alloc_zeroed())
> +	if (alloc_need_zeroing())
>   		folio_zero_user(folio, addr);
>   	/*
>   	 * The memory barrier inside __folio_mark_uptodate makes sure that
> diff --git a/mm/internal.h b/mm/internal.h
> index cb8d8e8e3ffa..3bd08bafad04 100644
> --- a/mm/internal.h
> +++ b/mm/internal.h
> @@ -1285,12 +1285,6 @@ void touch_pud(struct vm_area_struct *vma, unsigned long addr,
>   void touch_pmd(struct vm_area_struct *vma, unsigned long addr,
>   	       pmd_t *pmd, bool write);
>   
> -static inline bool alloc_zeroed(void)
> -{
> -	return static_branch_maybe(CONFIG_INIT_ON_ALLOC_DEFAULT_ON,
> -			&init_on_alloc);
> -}
> -
>   /*
>    * Parses a string with mem suffixes into its order. Useful to parse kernel
>    * parameters.
> diff --git a/mm/memory.c b/mm/memory.c
> index 75c2dfd04f72..cf1611791856 100644
> --- a/mm/memory.c
> +++ b/mm/memory.c
> @@ -4733,12 +4733,12 @@ static struct folio *alloc_anon_folio(struct vm_fault *vmf)
>   			folio_throttle_swaprate(folio, gfp);
>   			/*
>   			 * When a folio is not zeroed during allocation
> -			 * (__GFP_ZERO not used), folio_zero_user() is used
> -			 * to make sure that the page corresponding to the
> -			 * faulting address will be hot in the cache after
> -			 * zeroing.
> +			 * (__GFP_ZERO not used) or user folios require special
> +			 * handling, folio_zero_user() is used to make sure
> +			 * that the page corresponding to the faulting address
> +			 * will be hot in the cache after zeroing.
>   			 */
> -			if (!alloc_zeroed())
> +			if (alloc_need_zeroing())
>   				folio_zero_user(folio, vmf->address);
>   			return folio;
>   		}

-- 
Mathieu Desnoyers
EfficiOS Inc.
https://www.efficios.com