Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Wed, 16 Sep 2020 14:48:04 -0300
From:   Jason Gunthorpe <jgg@ziepe.ca>
To:     John Hubbard <jhubbard@nvidia.com>
Cc:     Peter Xu <peterx@redhat.com>,
        Linus Torvalds <torvalds@linux-foundation.org>,
        Leon Romanovsky <leonro@nvidia.com>,
        Linux-MM <linux-mm@kvack.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        "Maya B . Gokhale" <gokhale2@llnl.gov>,
        Yang Shi <yang.shi@linux.alibaba.com>,
        Marty Mcfadden <mcfadden8@llnl.gov>,
        Kirill Shutemov <kirill@shutemov.name>,
        Oleg Nesterov <oleg@redhat.com>, Jann Horn <jannh@google.com>,
        Jan Kara <jack@suse.cz>, Kirill Tkhai <ktkhai@virtuozzo.com>,
        Andrea Arcangeli <aarcange@redhat.com>,
        Christoph Hellwig <hch@lst.de>,
        Andrew Morton <akpm@linux-foundation.org>
Subject: Re: [PATCH 1/4] mm: Trial do_wp_page() simplification
Message-ID: <20200916174804.GC8409@ziepe.ca>
References: <CAHk-=wgdn5sJ0UEVZRQvj6r5kqOkU24jA_V6cPkqb9tqoAKBJg@mail.gmail.com>
 <20200914232851.GH1221970@ziepe.ca>
 <20200915145040.GA2949@xz-x1>
 <20200915160553.GJ1221970@ziepe.ca>
 <20200915182933.GM1221970@ziepe.ca>
 <20200915191346.GD2949@xz-x1>
 <20200915193838.GN1221970@ziepe.ca>
 <20200915213330.GE2949@xz-x1>
 <20200915232238.GO1221970@ziepe.ca>
 <e6c352f8-7ee9-0702-10a4-122d2c4422fc@nvidia.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <e6c352f8-7ee9-0702-10a4-122d2c4422fc@nvidia.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, Sep 15, 2020 at 06:50:46PM -0700, John Hubbard wrote:
> > 
> > It seems very strange that a physical page exclusively owned by a
> > process can become copied if pin_user_pages() is active and the
> > process did fork() at some point.
> > 
> > Could the new pin_user_pages() logic help here? eg the
> > GUP_PIN_COUNTING_BIAS stuff?
> > 
> > Could the COW code consider a refcount of GUP_PIN_COUNTING_BIAS + 1 as
> > being owned by the current mm and not needing COW? The DMA pin would
> > be 'invisible' for COW purposes?
> 
> 
> Please do be careful to use the API, rather than the implementation. The
> FOLL_PIN refcounting system results in being able to get a "maybe
> DMA-pinned", or a "definitely not DMA-pinned", via this API call:

So, what I'm thinking is something like (untested):

diff --git a/mm/memory.c b/mm/memory.c
index 469af373ae76e1..77f63183667e52 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -2889,6 +2889,26 @@ static vm_fault_t wp_page_shared(struct vm_fault *vmf)
 	return ret;
 }
 
+static bool cow_needed(struct vm_fault *vmf)
+{
+	int total_map_swapcount;
+
+	if (!reuse_swap_page(vmf->page, &total_map_swapcount)) {
+		unlock_page(vmf->page);
+		return true;
+	}
+
+	if (total_map_swapcount == 1) {
+		/*
+		 * The page is all ours. Move it to our anon_vma so the rmap
+		 * code will not search our parent or siblings.  Protected
+		 * against the rmap code by the page lock.
+		 */
+		page_move_anon_rmap(vmf->page, vmf->vma);
+	}
+	return false;
+}
+
 /*
  * This routine handles present pages, when users try to write
  * to a shared page. It is done by copying the page to a new address
@@ -2947,8 +2967,21 @@ static vm_fault_t do_wp_page(struct vm_fault *vmf)
 		if (!trylock_page(page))
 			goto copy;
 		if (PageKsm(page) || page_mapcount(page) != 1 || page_count(page) != 1) {
+			bool cow = true;
+
+			/*
+			 * If the page is DMA pinned we can't rely on the
+			 * above to know if there are other CPU references as
+			 * page_count() will be elevated by the
+			 * pin. Needlessly copying the page will cause the DMA
+			 * pin to break, try harder to avoid that.
+			 */
+			if (page_maybe_dma_pinned(page))
+				cow = cow_needed(vmf);
+
 			unlock_page(page);
-			goto copy;
+			if (cow)
+				goto copy;
 		}
 		/*
 		 * Ok, we've got the only map reference, and the only

What do you think Peter? Is this remotely close?

Seems like it preserves the fast path in most cases, the page_count &
page_maybe_dma_pinned could be further optimized down to one atomic in
non huge page cases.

Jason