Thread (16 messages) 16 messages, 5 authors, 2022-11-25

Re: [PATCH v4 2/2] arm64: dts: mt8192: Add vcodec lat and core nodes

From: Nícolas F. R. A. Prado <hidden>
Date: 2022-11-21 14:25:06
Also in: linux-devicetree, linux-media, linux-mediatek, lkml

On Mon, Nov 21, 2022 at 12:30:37PM +0000, Allen-KH Cheng (程冠勳) wrote:
Hi Nícolas,

On Fri, 2022-11-18 at 09:10 -0500, Nícolas F. R. A. Prado wrote:
quoted
On Fri, Sep 30, 2022 at 07:22:37PM +0800, Allen-KH Cheng wrote:
quoted
Add vcodec lat and core nodes for mt8192 SoC.

Signed-off-by: Allen-KH Cheng <redacted>
Tested-by: Chen-Yu Tsai <wenst@chromium.org>
Reviewed-by: AngeloGioacchino Del Regno <
angelogioacchino.delregno@collabora.com>
---
 arch/arm64/boot/dts/mediatek/mt8192.dtsi | 60
++++++++++++++++++++++++
 1 file changed, 60 insertions(+)
diff --git a/arch/arm64/boot/dts/mediatek/mt8192.dtsi
b/arch/arm64/boot/dts/mediatek/mt8192.dtsi
index 6b20376191a7..92a20f87468b 100644
--- a/arch/arm64/boot/dts/mediatek/mt8192.dtsi
+++ b/arch/arm64/boot/dts/mediatek/mt8192.dtsi
@@ -1449,6 +1449,66 @@
 			power-domains = <&spm
MT8192_POWER_DOMAIN_ISP2>;
 		};
 
+		vcodec_dec: video-codec@16000000 {
+			compatible = "mediatek,mt8192-vcodec-dec";
+			reg = <0 0x16000000 0 0x1000>;
+			mediatek,scp = <&scp>;
+			iommus = <&iommu0 M4U_PORT_L4_VDEC_MC_EXT>;
+			dma-ranges = <0x1 0x0 0x0 0x40000000 0x0
0xfff00000>;
Hi,

since commit 951d48855d86 ("of: Make of_dma_get_range() work on bus
nodes") [1]
was merged this no longer works as is. Running the fluster codec
tests results
in IOMMU faults:

	[  386.233976] mtk-iommu 1401d000.m4u: fault type=0x280
iova=0x1fcdc0000 pa=0x0 master=0x500041c(larb=4 port=7) layer=0 read
	[  386.250666] mtk_vdec_worker(),241:
[MTK_V4L2][ERROR]  <===[138], src_buf[0] sz=0x298 pts=0
vdec_if_decode() ret=1 res_chg=0===>

The issue is that the DMA configuration supplied by dma-ranges is now
looked for
in the parent node, so the vcodec_dec node no longer gets the
configuration it
expected.

That said, given that the node already uses the IOMMU for the address
translations (iommus property), there shouldn't even be a dma-ranges
property.
Indeed simply removing the dma-ranges property from this node fixes
the issue
and gets the decoder working again.

Thanks,
Nícolas

[1] 
https://urldefense.com/v3/__https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=f1ad5338a4d57fe1fe6475003acb8c70bf9d1bdf__;!!CTRNKA9wMg0ARbw!xdhhnvXMY5-BjI2BXPHQI-Hw8zgtZ1lvFyFFv7KtNuCDxW17VC7RqAaW9B_uXsQucT1sLk_DUl-c99ijF9dF8QXbJQ$
 
Noted!

The paret node should be

vcodec_dec: video-codec@16000000 {
	compatible = "mediatek,mt8192-vcodec-dec";
	reg = <0 0x16000000 0 0x1000>;
	mediatek,scp = <&scp>;
	iommus = <&iommu0 M4U_PORT_L4_VDEC_MC_EXT>;
	#address-cells = <2>;
	#size-cells = <2>;
	ranges = <0 0 0 0x16000000 0 0x26000>;
	
	video-codec-lat@10000 {
...

}

Am I right?
Yes, that's right.

Thanks,
Nícolas

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help