Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Wed, 30 Nov 2022 11:09:10 -0500
From:   Peter Xu <peterx@redhat.com>
To:     David Hildenbrand <david@redhat.com>
Cc:     linux-mm@kvack.org, linux-kernel@vger.kernel.org,
        James Houghton <jthoughton@google.com>,
        Jann Horn <jannh@google.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        Andrea Arcangeli <aarcange@redhat.com>,
        Rik van Riel <riel@surriel.com>,
        Nadav Amit <nadav.amit@gmail.com>,
        Miaohe Lin <linmiaohe@huawei.com>,
        Muchun Song <songmuchun@bytedance.com>,
        Mike Kravetz <mike.kravetz@oracle.com>
Subject: Re: [PATCH 03/10] mm/hugetlb: Document huge_pte_offset usage
Message-ID: <Y4eAJuZRG0CLP7PW@x1n>
References: <20221129193526.3588187-1-peterx@redhat.com>
 <20221129193526.3588187-4-peterx@redhat.com>
 <afce1bdf-6b5f-3393-cafa-81148277773d@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <afce1bdf-6b5f-3393-cafa-81148277773d@redhat.com>
Precedence: bulk

On Wed, Nov 30, 2022 at 11:24:34AM +0100, David Hildenbrand wrote:
> On 29.11.22 20:35, Peter Xu wrote:
> > huge_pte_offset() is potentially a pgtable walker, looking up pte_t* for a
> > hugetlb address.
> > 
> > Normally, it's always safe to walk a generic pgtable as long as we're with
> > the mmap lock held for either read or write, because that guarantees the
> > pgtable pages will always be valid during the process.
> > 
> > But it's not true for hugetlbfs, especially shared: hugetlbfs can have its
> > pgtable freed by pmd unsharing, it means that even with mmap lock held for
> > current mm, the PMD pgtable page can still go away from under us if pmd
> > unsharing is possible during the walk.
> > 
> > So we have two ways to make it safe even for a shared mapping:
> > 
> >    (1) If we're with the hugetlb vma lock held for either read/write, it's
> >        okay because pmd unshare cannot happen at all.
> > 
> >    (2) If we're with the i_mmap_rwsem lock held for either read/write, it's
> >        okay because even if pmd unshare can happen, the pgtable page cannot
> >        be freed from under us.
> > 
> > Document it.
> > 
> > Signed-off-by: Peter Xu <peterx@redhat.com>
> > ---
> >   include/linux/hugetlb.h | 32 ++++++++++++++++++++++++++++++++
> >   1 file changed, 32 insertions(+)
> > 
> > diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h
> > index 551834cd5299..81efd9b9baa2 100644
> > --- a/include/linux/hugetlb.h
> > +++ b/include/linux/hugetlb.h
> > @@ -192,6 +192,38 @@ extern struct list_head huge_boot_pages;
> >   pte_t *huge_pte_alloc(struct mm_struct *mm, struct vm_area_struct *vma,
> >   			unsigned long addr, unsigned long sz);
> > +/*
> > + * huge_pte_offset(): Walk the hugetlb pgtable until the last level PTE.
> > + * Returns the pte_t* if found, or NULL if the address is not mapped.
> > + *
> > + * Since this function will walk all the pgtable pages (including not only
> > + * high-level pgtable page, but also PUD entry that can be unshared
> > + * concurrently for VM_SHARED), the caller of this function should be
> > + * responsible of its thread safety.  One can follow this rule:
> > + *
> > + *  (1) For private mappings: pmd unsharing is not possible, so it'll
> > + *      always be safe if we're with the mmap sem for either read or write.
> > + *      This is normally always the case, IOW we don't need to do anything
> > + *      special.
> 
> Maybe worth mentioning that hugetlb_vma_lock_read() and friends already
> optimize for private mappings, to not take the VMA lock if not required.

Yes we can.  I assume this is not super urgent so I'll hold a while to see
whether there's anything else that needs amending for the documents.

Btw, even with hugetlb_vma_lock_read() checking SHARED for a private only
code path it's still better to not take the lock at all, because that still
contains a function jump which will be unnecesary.

Thanks,

-- 
Peter Xu