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:   Fri, 26 May 2023 18:42:20 -0400
From:   Peter Xu <peterx@redhat.com>
To:     Hugh Dickins <hughd@google.com>
Cc:     Andrew Morton <akpm@linux-foundation.org>,
        Mike Kravetz <mike.kravetz@oracle.com>,
        Mike Rapoport <rppt@kernel.org>,
        "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>,
        Matthew Wilcox <willy@infradead.org>,
        David Hildenbrand <david@redhat.com>,
        Suren Baghdasaryan <surenb@google.com>,
        Qi Zheng <zhengqi.arch@bytedance.com>,
        Yang Shi <shy828301@gmail.com>,
        Mel Gorman <mgorman@techsingularity.net>,
        Peter Zijlstra <peterz@infradead.org>,
        Will Deacon <will@kernel.org>, Yu Zhao <yuzhao@google.com>,
        Alistair Popple <apopple@nvidia.com>,
        Ralph Campbell <rcampbell@nvidia.com>,
        Ira Weiny <ira.weiny@intel.com>,
        Steven Price <steven.price@arm.com>,
        SeongJae Park <sj@kernel.org>,
        Naoya Horiguchi <naoya.horiguchi@nec.com>,
        Christophe Leroy <christophe.leroy@csgroup.eu>,
        Zack Rusin <zackr@vmware.com>, Jason Gunthorpe <jgg@ziepe.ca>,
        Axel Rasmussen <axelrasmussen@google.com>,
        Anshuman Khandual <anshuman.khandual@arm.com>,
        Pasha Tatashin <pasha.tatashin@soleen.com>,
        Miaohe Lin <linmiaohe@huawei.com>,
        Minchan Kim <minchan@kernel.org>,
        Christoph Hellwig <hch@infradead.org>,
        Song Liu <song@kernel.org>,
        Thomas Hellstrom <thomas.hellstrom@linux.intel.com>,
        linux-kernel@vger.kernel.org, linux-mm@kvack.org
Subject: Re: [PATCH 03/31] mm/pgtable: kmap_local_page() instead of
 kmap_atomic()
Message-ID: <ZHE1zLLF1jfcw8OI@x1n>
References: <68a97fbe-5c1e-7ac6-72c-7b9c6290b370@google.com>
 <9df4aba7-fd2f-2da3-1543-fc6b4b42f5b9@google.com>
 <ZHExQo3KO5PcVR76@x1n>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <ZHExQo3KO5PcVR76@x1n>
Precedence: bulk

On Fri, May 26, 2023 at 06:22:58PM -0400, Peter Xu wrote:
> On Sun, May 21, 2023 at 09:52:31PM -0700, Hugh Dickins wrote:
> > pte_offset_map() was still using kmap_atomic(): update it to the
> > preferred kmap_local_page() before making further changes there, in case
> > we need this as a bisection point; but I doubt it can cause any trouble.
> > 
> > Signed-off-by: Hugh Dickins <hughd@google.com>
> > ---
> >  include/linux/pgtable.h | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/include/linux/pgtable.h b/include/linux/pgtable.h
> > index 8ec27fe69dc8..94235ff2706e 100644
> > --- a/include/linux/pgtable.h
> > +++ b/include/linux/pgtable.h
> > @@ -96,9 +96,9 @@ static inline pte_t *pte_offset_kernel(pmd_t *pmd, unsigned long address)
> >  
> >  #if defined(CONFIG_HIGHPTE)
> >  #define pte_offset_map(dir, address)				\
> > -	((pte_t *)kmap_atomic(pmd_page(*(dir))) +		\
> > +	((pte_t *)kmap_local_page(pmd_page(*(dir))) +		\
> >  	 pte_index((address)))
> > -#define pte_unmap(pte) kunmap_atomic((pte))
> > +#define pte_unmap(pte) kunmap_local((pte))
> >  #else
> >  #define pte_offset_map(dir, address)	pte_offset_kernel((dir), (address))
> >  #define pte_unmap(pte) ((void)(pte))	/* NOP */
> 
> (I think this could be a dumb question if this patch has been running there
>  for years downstream, but still..)
> 
> I assume one major difference of using kmap_local() is page fault will not
> be disabled, while kmap_atomic() will.
> 
> Meanwhile, pte_offset_map() is also used by pte_offset_map_lock(), which
> means before this patch CONFIG_HIGHPTE systems will disable pgfault before
> taking pgtable lock for it, while it will stop doing so after the change.
> 
> Then what happens if a page fault happens on the same pgtable lock range
> that is already taken by the thread context?  What stops the deadlock from
> happening?

Ah, stupid me.  I think such a page fault just cannot happen when holding
the pgtable lock..  I believe the same applies to !HIGHPTE..

Sorry about the noise.

-- 
Peter Xu