Received-SPF: pass (google.com: domain of linux-kernel+bounces-153782-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:40f1:3f00::1 as permitted sender) client-ip=2604:1380:40f1:3f00::1;
From: Douglas Anderson <dianders@chromium.org>
To: Tiffany Lin <tiffany.lin@mediatek.com>,
	Andrew-CT Chen <andrew-ct.chen@mediatek.com>,
	Yunfei Dong <yunfei.dong@mediatek.com>,
	Mauro Carvalho Chehab <mchehab@kernel.org>,
	Matthias Brugger <matthias.bgg@gmail.com>,
	AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Cc: Wei-Shun Chang <weishunc@google.com>,
	Douglas Anderson <dianders@chromium.org>,
	Hans Verkuil <hverkuil-cisco@xs4all.nl>,
	Nicolas Dufresne <nicolas.dufresne@collabora.com>,
	=?UTF-8?q?N=C3=ADcolas=20F=2E=20R=2E=20A=2E=20Prado?= <nfraprado@collabora.com>,
	Rob Herring <robh@kernel.org>,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org,
	linux-media@vger.kernel.org,
	linux-mediatek@lists.infradead.org
Subject: [PATCH] media: mediatek: vcodec: Alloc DMA memory with DMA_ATTR_ALLOC_SINGLE_PAGES
Date: Mon, 22 Apr 2024 10:03:59 -0700
Message-ID: <20240422100354.1.I58b4456c014a4d678455a4ec09b908b1c71c3017@changeid>
Precedence: bulk
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit

As talked about in commit 14d3ae2efeed ("ARM: 8507/1: dma-mapping: Use
DMA_ATTR_ALLOC_SINGLE_PAGES hint to optimize alloc"), it doesn't
really make sense to try to allocate contiguous chunks of memory for
video encoding/decoding. Let's switch the Mediatek vcodec driver to
pass DMA_ATTR_ALLOC_SINGLE_PAGES and take some of the stress off the
memory subsystem.

Signed-off-by: Douglas Anderson <dianders@chromium.org>
---
NOTE: I haven't personally done massive amounts of testing with this
change, but I originally added the DMA_ATTR_ALLOC_SINGLE_PAGES flag
specifically for the video encoding / decoding cases and I know it
helped avoid memory problems in the past on other systems. Colleagues
of mine have told me that with this change memory problems are harder
to reproduce, so it seems like we should consider doing it.

 .../media/platform/mediatek/vcodec/common/mtk_vcodec_util.c    | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_util.c b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_util.c
index 9ce34a3b5ee6..3fb1d48c3e15 100644
--- a/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_util.c
+++ b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_util.c
@@ -64,7 +64,8 @@ int mtk_vcodec_mem_alloc(void *priv, struct mtk_vcodec_mem *mem)
 		id = dec_ctx->id;
 	}
 
-	mem->va = dma_alloc_coherent(&plat_dev->dev, size, &mem->dma_addr, GFP_KERNEL);
+	mem->va = dma_alloc_attrs(&plat_dev->dev, size, &mem->dma_addr,
+				  GFP_KERNEL, DMA_ATTR_ALLOC_SINGLE_PAGES);
 	if (!mem->va) {
 		mtk_v4l2_err(plat_dev, "%s dma_alloc size=%ld failed!",
 			     dev_name(&plat_dev->dev), size);
-- 
2.44.0.769.g3c40516874-goog