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:   Fri, 14 Jan 2022 17:06:03 -0800
In-Reply-To: <20220115010622.3185921-1-hridya@google.com>
Message-Id: <20220115010622.3185921-6-hridya@google.com>
Mime-Version: 1.0
References: <20220115010622.3185921-1-hridya@google.com>
Subject: [RFC 5/6] dmabuf: system_heap: implement dma-buf op for GPU cgroup
 charge transfer
From:   Hridya Valsaraju <hridya@google.com>
To:     Maarten Lankhorst <maarten.lankhorst@linux.intel.com>,
        Maxime Ripard <mripard@kernel.org>,
        Thomas Zimmermann <tzimmermann@suse.de>,
        David Airlie <airlied@linux.ie>,
        Daniel Vetter <daniel@ffwll.ch>,
        Jonathan Corbet <corbet@lwn.net>,
        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 <christian@brauner.io>,
        Hridya Valsaraju <hridya@google.com>,
        Suren Baghdasaryan <surenb@google.com>,
        Sumit Semwal <sumit.semwal@linaro.org>,
        Benjamin Gaignard <benjamin.gaignard@linaro.org>,
        Liam Mark <lmark@codeaurora.org>,
        Laura Abbott <labbott@redhat.com>,
        Brian Starkey <Brian.Starkey@arm.com>,
        John Stultz <john.stultz@linaro.org>,
        "=?UTF-8?q?Christian=20K=C3=B6nig?=" <christian.koenig@amd.com>,
        Tejun Heo <tj@kernel.org>, Zefan Li <lizefan.x@bytedance.com>,
        Johannes Weiner <hannes@cmpxchg.org>,
        Dave Airlie <airlied@redhat.com>,
        Kenneth Graunke <kenneth@whitecape.org>,
        Rodrigo Vivi <rodrigo.vivi@intel.com>,
        Matthew Brost <matthew.brost@intel.com>,
        Matthew Auld <matthew.auld@intel.com>,
        Li Li <dualli@google.com>, Marco Ballesio <balejs@google.com>,
        Miguel Ojeda <ojeda@kernel.org>,
        Hang Lu <hangl@codeaurora.org>,
        Wedson Almeida Filho <wedsonaf@google.com>,
        Masahiro Yamada <masahiroy@kernel.org>,
        Andrew Morton <akpm@linux-foundation.org>,
        Nathan Chancellor <nathan@kernel.org>,
        Kees Cook <keescook@chromium.org>,
        Nick Desaulniers <ndesaulniers@google.com>,
        Chris Down <chris@chrisdown.name>,
        Vipin Sharma <vipinsh@google.com>,
        Daniel Borkmann <daniel@iogearbox.net>,
        Vlastimil Babka <vbabka@suse.cz>,
        Arnd Bergmann <arnd@arndb.de>, dri-devel@lists.freedesktop.org,
        linux-doc@vger.kernel.org, linux-kernel@vger.kernel.org,
        linux-media@vger.kernel.org, linaro-mm-sig@lists.linaro.org,
        cgroups@vger.kernel.org
Cc:     Kenny.Ho@amd.com, daniels@collabora.com, kaleshsingh@google.com,
        tjmercier@google.com
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

The DMA-BUF op can be invoked when a process that allocated a buffer
relinquishes its ownership and passes it over to another process.

Signed-off-by: Hridya Valsaraju <hridya@google.com>
---
 drivers/dma-buf/heaps/system_heap.c | 43 +++++++++++++++++++++++++++++
 1 file changed, 43 insertions(+)

diff --git a/drivers/dma-buf/heaps/system_heap.c b/drivers/dma-buf/heaps/system_heap.c
index adfdc8c576f2..70f5b98f1157 100644
--- a/drivers/dma-buf/heaps/system_heap.c
+++ b/drivers/dma-buf/heaps/system_heap.c
@@ -307,6 +307,48 @@ static void system_heap_dma_buf_release(struct dma_buf *dmabuf)
 	kfree(buffer);
 }
 
+#ifdef CONFIG_CGROUP_GPU
+static int system_heap_dma_buf_charge(struct dma_buf *dmabuf, struct gpucg *gpucg)
+{
+	struct gpucg *current_gpucg;
+	struct gpucg_device *gpucg_dev;
+	struct system_heap_buffer *buffer = dmabuf->priv;
+	size_t len = buffer->len;
+	int ret = 0;
+
+	/*
+	 * Check that the process requesting the transfer is the same as the one
+	 * to whom the buffer is currently charged to.
+	 */
+	current_gpucg = gpucg_get(current);
+	if (current_gpucg != buffer->gpucg)
+		ret = -EPERM;
+
+	gpucg_put(current_gpucg);
+	if (ret)
+		return ret;
+
+	gpucg_dev = dma_heap_get_gpucg_dev(buffer->heap);
+
+	ret = gpucg_try_charge(gpucg, gpucg_dev, len);
+	if (ret)
+		return ret;
+
+	/* uncharge the buffer from the cgroup its currently charged to. */
+	gpucg_uncharge(buffer->gpucg, gpucg_dev, buffer->len);
+	gpucg_put(buffer->gpucg);
+
+	buffer->gpucg = gpucg;
+
+	return 0;
+}
+#else
+static int system_heap_dma_buf_charge(struct dma_buf *dmabuf, struct gpucg *gpucg)
+{
+	return 0;
+}
+#endif
+
 static const struct dma_buf_ops system_heap_buf_ops = {
 	.attach = system_heap_attach,
 	.detach = system_heap_detach,
@@ -318,6 +360,7 @@ static const struct dma_buf_ops system_heap_buf_ops = {
 	.vmap = system_heap_vmap,
 	.vunmap = system_heap_vunmap,
 	.release = system_heap_dma_buf_release,
+	.charge_to_cgroup = system_heap_dma_buf_charge,
 };
 
 static struct page *alloc_largest_available(unsigned long size,
-- 
2.34.1.703.g22d0c6ccf7-goog