Re: [merged] mm-slub-init_on_free=1-should-wipe-freelist-ptr-for-bulk-allocations.patch removed from -mm tree

Re: [merged] mm-slub-init_on_free=1-should-wipe-freelist-ptr-for-bulk-allocations.patch removed from -mm tree

[Thibaut Sautereau]

On Tue, Oct 15, 2019 at 11:14:42AM -0700, akpm@linux-foundation.org wrote:
> 
> The patch titled
>      Subject: mm/slub.c: init_on_free=1 should wipe freelist ptr for bulk allocations
> has been removed from the -mm tree.  Its filename was
>      mm-slub-init_on_free=1-should-wipe-freelist-ptr-for-bulk-allocations.patch
> 
> This patch was dropped because it was merged into mainline or a subsystem tree
> 
> ------------------------------------------------------
> From: Alexander Potapenko <glider@google.com>
> Subject: mm/slub.c: init_on_free=1 should wipe freelist ptr for bulk allocations
> 
> slab_alloc_node() already zeroed out the freelist pointer if init_on_free
> was on.  Thibaut Sautereau noticed that the same needs to be done for
> kmem_cache_alloc_bulk(), which performs the allocations separately.
> 
> kmem_cache_alloc_bulk() is currently used in two places in the kernel, so
> this change is unlikely to have a major performance impact.
> 
> SLAB doesn't require a similar change, as auto-initialization makes the
> allocator store the freelist pointers off-slab.
> 
> Link: http://lkml.kernel.org/r/20191007091605.30530-1-glider@google.com
> Fixes: 6471384af2a6 ("mm: security: introduce init_on_alloc=1 and init_on_free=1 boot options")
> Signed-off-by: Alexander Potapenko <glider@google.com>
> Reported-by: Thibaut Sautereau <thibaut@sautereau.fr>
> Reported-by: Kees Cook <keescook@chromium.org>
> Cc: Christoph Lameter <cl@linux.com>
> Cc: Laura Abbott <labbott@redhat.com>
> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
> ---
> 
>  mm/slub.c |   22 ++++++++++++++++------
>  1 file changed, 16 insertions(+), 6 deletions(-)
> 
> --- a/mm/slub.c~mm-slub-init_on_free=1-should-wipe-freelist-ptr-for-bulk-allocations
> +++ a/mm/slub.c
> @@ -2672,6 +2672,17 @@ static void *__slab_alloc(struct kmem_ca
>  }
>  
>  /*
> + * If the object has been wiped upon free, make sure it's fully initialized by
> + * zeroing out freelist pointer.
> + */
> +static __always_inline void maybe_wipe_obj_freeptr(struct kmem_cache *s,
> +						   void *obj)
> +{
> +	if (unlikely(slab_want_init_on_free(s)) && obj)
> +		memset((void *)((char *)obj + s->offset), 0, sizeof(void *));
> +}
> +
> +/*
>   * Inlined fastpath so that allocation functions (kmalloc, kmem_cache_alloc)
>   * have the fastpath folded into their functions. So no function call
>   * overhead for requests that can be satisfied on the fastpath.
> @@ -2759,12 +2770,8 @@ redo:
>  		prefetch_freepointer(s, next_object);
>  		stat(s, ALLOC_FASTPATH);
>  	}
> -	/*
> -	 * If the object has been wiped upon free, make sure it's fully
> -	 * initialized by zeroing out freelist pointer.
> -	 */
> -	if (unlikely(slab_want_init_on_free(s)) && object)
> -		memset(object + s->offset, 0, sizeof(void *));
> +
> +	maybe_wipe_obj_freeptr(s, object);
>  
>  	if (unlikely(slab_want_init_on_alloc(gfpflags, s)) && object)
>  		memset(object, 0, s->object_size);
> @@ -3178,10 +3185,13 @@ int kmem_cache_alloc_bulk(struct kmem_ca
>  				goto error;
>  
>  			c = this_cpu_ptr(s->cpu_slab);
> +			maybe_wipe_obj_freeptr(s, p[i]);
> +
>  			continue; /* goto for-loop */
>  		}
>  		c->freelist = get_freepointer(s, object);
>  		p[i] = object;
> +		maybe_wipe_obj_freeptr(s, p[i]);
>  	}
>  	c->tid = next_tid(c->tid);
>  	local_irq_enable();
> _

Can this be backported to stable 5.3 please? It's commit 0f181f9fbea8
upstream. Thanks!

-- 
Thibaut Sautereau

Re: [merged] mm-slub-init_on_free=1-should-wipe-freelist-ptr-for-bulk-allocations.patch removed from -mm tree

[Sasha Levin]

On Tue, Nov 26, 2019 at 09:35:38PM +0100, Thibaut Sautereau wrote:
>On Tue, Oct 15, 2019 at 11:14:42AM -0700, akpm@linux-foundation.org wrote:
>>
>> The patch titled
>>      Subject: mm/slub.c: init_on_free=1 should wipe freelist ptr for bulk allocations
>> has been removed from the -mm tree.  Its filename was
>>      mm-slub-init_on_free=1-should-wipe-freelist-ptr-for-bulk-allocations.patch
>>
>> This patch was dropped because it was merged into mainline or a subsystem tree
>>
>> ------------------------------------------------------
>> From: Alexander Potapenko <glider@google.com>
>> Subject: mm/slub.c: init_on_free=1 should wipe freelist ptr for bulk allocations
>>
>> slab_alloc_node() already zeroed out the freelist pointer if init_on_free
>> was on.  Thibaut Sautereau noticed that the same needs to be done for
>> kmem_cache_alloc_bulk(), which performs the allocations separately.
>>
>> kmem_cache_alloc_bulk() is currently used in two places in the kernel, so
>> this change is unlikely to have a major performance impact.
>>
>> SLAB doesn't require a similar change, as auto-initialization makes the
>> allocator store the freelist pointers off-slab.
>>
>> Link: http://lkml.kernel.org/r/20191007091605.30530-1-glider@google.com
>> Fixes: 6471384af2a6 ("mm: security: introduce init_on_alloc=1 and init_on_free=1 boot options")
>> Signed-off-by: Alexander Potapenko <glider@google.com>
>> Reported-by: Thibaut Sautereau <thibaut@sautereau.fr>
>> Reported-by: Kees Cook <keescook@chromium.org>
>> Cc: Christoph Lameter <cl@linux.com>
>> Cc: Laura Abbott <labbott@redhat.com>
>> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
>> ---
>>
>>  mm/slub.c |   22 ++++++++++++++++------
>>  1 file changed, 16 insertions(+), 6 deletions(-)
>>
>> --- a/mm/slub.c~mm-slub-init_on_free=1-should-wipe-freelist-ptr-for-bulk-allocations
>> +++ a/mm/slub.c
>> @@ -2672,6 +2672,17 @@ static void *__slab_alloc(struct kmem_ca
>>  }
>>
>>  /*
>> + * If the object has been wiped upon free, make sure it's fully initialized by
>> + * zeroing out freelist pointer.
>> + */
>> +static __always_inline void maybe_wipe_obj_freeptr(struct kmem_cache *s,
>> +						   void *obj)
>> +{
>> +	if (unlikely(slab_want_init_on_free(s)) && obj)
>> +		memset((void *)((char *)obj + s->offset), 0, sizeof(void *));
>> +}
>> +
>> +/*
>>   * Inlined fastpath so that allocation functions (kmalloc, kmem_cache_alloc)
>>   * have the fastpath folded into their functions. So no function call
>>   * overhead for requests that can be satisfied on the fastpath.
>> @@ -2759,12 +2770,8 @@ redo:
>>  		prefetch_freepointer(s, next_object);
>>  		stat(s, ALLOC_FASTPATH);
>>  	}
>> -	/*
>> -	 * If the object has been wiped upon free, make sure it's fully
>> -	 * initialized by zeroing out freelist pointer.
>> -	 */
>> -	if (unlikely(slab_want_init_on_free(s)) && object)
>> -		memset(object + s->offset, 0, sizeof(void *));
>> +
>> +	maybe_wipe_obj_freeptr(s, object);
>>
>>  	if (unlikely(slab_want_init_on_alloc(gfpflags, s)) && object)
>>  		memset(object, 0, s->object_size);
>> @@ -3178,10 +3185,13 @@ int kmem_cache_alloc_bulk(struct kmem_ca
>>  				goto error;
>>
>>  			c = this_cpu_ptr(s->cpu_slab);
>> +			maybe_wipe_obj_freeptr(s, p[i]);
>> +
>>  			continue; /* goto for-loop */
>>  		}
>>  		c->freelist = get_freepointer(s, object);
>>  		p[i] = object;
>> +		maybe_wipe_obj_freeptr(s, p[i]);
>>  	}
>>  	c->tid = next_tid(c->tid);
>>  	local_irq_enable();
>> _
>
>Can this be backported to stable 5.3 please? It's commit 0f181f9fbea8
>upstream. Thanks!

Sure, I'll queue it up for 5.3, thanks!

-- 
Thanks,
Sasha