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;
From:   "Fabio M. De Francesco" <fmdefrancesco@gmail.com>
To:     Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        =?UTF-8?q?Arve=20Hj=C3=B8nnev=C3=A5g?= <arve@android.com>,
        Todd Kjos <tkjos@android.com>,
        Martijn Coenen <maco@android.com>,
        Joel Fernandes <joel@joelfernandes.org>,
        Christian Brauner <brauner@kernel.org>,
        Hridya Valsaraju <hridya@google.com>,
        Suren Baghdasaryan <surenb@google.com>,
        Ira Weiny <ira.weiny@intel.com>, linux-kernel@vger.kernel.org,
        Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Cc:     "Fabio M. De Francesco" <fmdefrancesco@gmail.com>,
        Todd Kjos <tkjos@google.com>
Subject: [PATCH v2 3/3] binder: Use memcpy_{to,from}_page() in binder_alloc_do_buffer_copy()
Date:   Mon, 25 Apr 2022 19:57:54 +0200
Message-Id: <20220425175754.8180-4-fmdefrancesco@gmail.com>
In-Reply-To: <20220425175754.8180-1-fmdefrancesco@gmail.com>
References: <20220425175754.8180-1-fmdefrancesco@gmail.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

The use of kmap_atomic() is being deprecated in favor of kmap_local_page()
where it is feasible. Each call of kmap_atomic() in the kernel creates
a non-preemptible section and disable pagefaults. This could be a source
of unwanted latency, so kmap_local_page() should be preferred.

With kmap_local_page(), the mapping is per thread, CPU local and not
globally visible. Furthermore, the mapping can be acquired from any context
(including interrupts). binder_alloc_do_buffer_copy() is a function where
the use of kmap_local_page() in place of kmap_atomic() is correctly suited.

Use kmap_local_page() / kunmap_local() in place of kmap_atomic() /
kunmap_atomic() but, instead of open coding the mappings and call memcpy()
to and from the virtual addresses of the mapped pages, prefer the use of
the memcpy_{to,from}_page() wrappers (as suggested by Christophe
Jaillet).

Cc: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Acked-by: Todd Kjos <tkjos@google.com>
Signed-off-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
---

v1->v2: Add Todd Kjos's tag in the commit message (thanks!); re-write
the subject which referred to another function (my fault!); use the
memcpy_{to,from}_page() (as suggested by Christophe - thanks!); extend
and rework the commit message to say something about the use of the
above-mentioned wrappers.

 drivers/android/binder_alloc.c | 13 ++-----------
 1 file changed, 2 insertions(+), 11 deletions(-)

diff --git a/drivers/android/binder_alloc.c b/drivers/android/binder_alloc.c
index 0875c463c002..5649a0371a1f 100644
--- a/drivers/android/binder_alloc.c
+++ b/drivers/android/binder_alloc.c
@@ -1244,23 +1244,14 @@ static int binder_alloc_do_buffer_copy(struct binder_alloc *alloc,
 		unsigned long size;
 		struct page *page;
 		pgoff_t pgoff;
-		void *tmpptr;
-		void *base_ptr;
 
 		page = binder_alloc_get_page(alloc, buffer,
 					     buffer_offset, &pgoff);
 		size = min_t(size_t, bytes, PAGE_SIZE - pgoff);
-		base_ptr = kmap_atomic(page);
-		tmpptr = base_ptr + pgoff;
 		if (to_buffer)
-			memcpy(tmpptr, ptr, size);
+			memcpy_to_page(page, pgoff, ptr, size);
 		else
-			memcpy(ptr, tmpptr, size);
-		/*
-		 * kunmap_atomic() takes care of flushing the cache
-		 * if this device has VIVT cache arch
-		 */
-		kunmap_atomic(base_ptr);
+			memcpy_from_page(ptr, page, pgoff, size);
 		bytes -= size;
 		pgoff = 0;
 		ptr = ptr + size;
-- 
2.34.1