Re: [PATCH v2 1/4] mm, madvise: simplify anon_name handling

[Vlastimil Babka <vbabka@suse.cz>]

On 6/24/25 17:26, Lorenzo Stoakes wrote:
> On Tue, Jun 24, 2025 at 03:03:45PM +0200, Vlastimil Babka wrote:
>> Since the introduction in 9a10064f5625 ("mm: add a field to store names
>> for private anonymous memory") the code to set anon_name on a vma has
>> been using madvise_update_vma() to call replace_anon_vma_name(). Since
>> the former is called also by a number of other madvise behaviours that
>> do not set a new anon_name, they have been passing the existing
>> anon_name of the vma to make replace_vma_anon_name() a no-op.
>>
>> This is rather wasteful as it needs anon_vma_name_eq() to determine the
>> no-op situations, and checks for when replace_vma_anon_name() is allowed
>> (the vma is anon/shmem) duplicate the checks already done earlier in
>> madvise_vma_behavior(). It has also lead to commit 942341dcc574 ("mm:
>> fix use-after-free when anon vma name is used after vma is freed")
>> adding anon_name refcount get/put operations exactly to the cases that
>> actually do not change anon_name - just so the replace_vma_anon_name()
>> can keep safely determining it has nothing to do.
>>
>> The recent madvise cleanups made this suboptimal handling very obvious,
>> but happily also allow for an easy fix. madvise_update_vma() now has the
>> complete information whether it's been called to set a new anon_name, so
>> stop passing it the existing vma's name and doing the refcount get/put
>> in its only caller madvise_vma_behavior().
>>
>> In madvise_update_vma() itself, limit calling of replace_anon_vma_name()
>> only to cases where we are setting a new name, otherwise we know it's a
>> no-op. We can rely solely on the __MADV_SET_ANON_VMA_NAME behaviour and
>> can remove the duplicate checks for vma being anon/shmem that were done
>> already in madvise_vma_behavior().
>>
>> Additionally, by using vma_modify_flags() when not modifying the
>> anon_name, avoid explicitly passing the existing vma's anon_name and
>> storing a pointer to it in struct madv_behavior or a local variable.
>> This prevents the danger of accessing a freed anon_name after vma
>> merging, previously fixed by commit 942341dcc574.
>>
>> Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
> 
> Cheers, LGTM so:
> 
> Reviewed-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
> 
> I made sure to do some stress-ng --madvise testing with this series (+ my recent
> fix) applied :P all good.

Thanks for that!

>> ---
>>  mm/madvise.c | 37 +++++++++++++------------------------
>>  1 file changed, 13 insertions(+), 24 deletions(-)
>>
>> diff --git a/mm/madvise.c b/mm/madvise.c
>> index 4491bf080f55d6d1aeffb2ff0b8fdd28904af950..fca0e9b3e844ad766e83ac04cc0d7f4099c74005 100644
>> --- a/mm/madvise.c
>> +++ b/mm/madvise.c
>> @@ -176,25 +176,29 @@ static int replace_anon_vma_name(struct vm_area_struct *vma,
>>  }
>>  #endif /* CONFIG_ANON_VMA_NAME */
>>  /*
>> - * Update the vm_flags on region of a vma, splitting it or merging it as
>> - * necessary.  Must be called with mmap_lock held for writing;
>> - * Caller should ensure anon_name stability by raising its refcount even when
>> - * anon_name belongs to a valid vma because this function might free that vma.
>> + * Update the vm_flags and/or anon_name on region of a vma, splitting it or
>> + * merging it as necessary. Must be called with mmap_lock held for writing.
>>   */
>>  static int madvise_update_vma(vm_flags_t new_flags,
>>  		struct madvise_behavior *madv_behavior)
>>  {
>> -	int error;
>>  	struct vm_area_struct *vma = madv_behavior->vma;
>>  	struct madvise_behavior_range *range = &madv_behavior->range;
>>  	struct anon_vma_name *anon_name = madv_behavior->anon_name;
>> +	bool set_new_anon_name = madv_behavior->behavior == __MADV_SET_ANON_VMA_NAME;
>>  	VMA_ITERATOR(vmi, madv_behavior->mm, range->start);
>>
>> -	if (new_flags == vma->vm_flags && anon_vma_name_eq(anon_vma_name(vma), anon_name))
>> +	if (new_flags == vma->vm_flags && (!set_new_anon_name ||
>> +			anon_vma_name_eq(anon_vma_name(vma), anon_name)))
>>  		return 0;
>>
>> -	vma = vma_modify_flags_name(&vmi, madv_behavior->prev, vma,
>> +	if (set_new_anon_name)
>> +		vma = vma_modify_flags_name(&vmi, madv_behavior->prev, vma,
>>  			range->start, range->end, new_flags, anon_name);
> 
> I will do a follow up (doesn't really belong here I'd say as involves change to
> modify code) that makes this vma_modify_name() or vma_modify_anon_name() as this
> is the only caller and we don't care about flags here :)

Oh right, good idea!