linux-kernel - [PATCH v5 08/11] iommu/mediatek: Add a gap for the iova regions

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 7 Mar 2023 10:35:04 +0800
From:   Yong Wu <yong.wu@...iatek.com>
To:     Joerg Roedel <joro@...tes.org>,
        Matthias Brugger <matthias.bgg@...il.com>,
        Mauro Carvalho Chehab <mchehab@...nel.org>,
        Rob Herring <robh+dt@...nel.org>
CC:     Will Deacon <will@...nel.org>, Robin Murphy <robin.murphy@....com>,
        Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
        Hans Verkuil <hverkuil@...all.nl>, <nfraprado@...labora.com>,
        <linux-media@...r.kernel.org>, <devicetree@...r.kernel.org>,
        <linux-arm-kernel@...ts.infradead.org>,
        <linux-mediatek@...ts.infradead.org>,
        <linux-kernel@...r.kernel.org>, <iommu@...ts.linux.dev>,
        "AngeloGioacchino Del Regno" 
        <angelogioacchino.delregno@...labora.com>,
        <mingyuan.ma@...iatek.com>, <yf.wang@...iatek.com>,
        <jianjiao.zeng@...iatek.com>,
        Yunfei Dong <yunfei.dong@...iatek.com>,
        kyrie wu <kyrie.wu@...iatek.corp-partner.google.com>,
        <chengci.xu@...iatek.com>, <youlin.pei@...iatek.com>,
        <anan.sun@...iatek.com>, Yong Wu <yong.wu@...iatek.com>
Subject: [PATCH v5 08/11] iommu/mediatek: Add a gap for the iova regions

As the removed property in the vcodec dt-binding, the property is:
dma-ranges = <0x1 0x0 0x0 0x40000000 0x0 0xfff00000>;

The length is 0xfff0_0000 rather than 0x1_0000_0000, this means it
requires 1M as a gap. This is because the end address for some vcodec
HW is (address + size). If the size is 4G, the end address may be
0x2_0000_0000, and the width for vcodec register only is 32, then the
HW may get the ZERO address.

Currently the consumer's dma-ranges property doesn't work, IOMMU
has to consider this case. Add a bigger gap(8M) for all the regions
to avoid it.

Signed-off-by: Yong Wu <yong.wu@...iatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@...labora.com>
---
 drivers/iommu/mtk_iommu.c | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

diff --git a/drivers/iommu/mtk_iommu.c b/drivers/iommu/mtk_iommu.c
index fbfc5e4e56a8..e0264d5f1c9a 100644
--- a/drivers/iommu/mtk_iommu.c
+++ b/drivers/iommu/mtk_iommu.c
@@ -332,8 +332,10 @@ static LIST_HEAD(m4ulist);	/* List all the M4U HWs */
 
 #define for_each_m4u(data, head)  list_for_each_entry(data, head, list)
 
+#define MTK_IOMMU_IOVA_SZ_4G		(SZ_4G - SZ_8M) /* 8M as gap */
+
 static const struct mtk_iommu_iova_region single_domain[] = {
-	{.iova_base = 0,		.size = SZ_4G},
+	{.iova_base = 0,		.size = MTK_IOMMU_IOVA_SZ_4G},
 };
 
 #define MT8192_MULTI_REGION_NR_MAX	6
@@ -342,11 +344,11 @@ static const struct mtk_iommu_iova_region single_domain[] = {
 				 MT8192_MULTI_REGION_NR_MAX : 1)
 
 static const struct mtk_iommu_iova_region mt8192_multi_dom[MT8192_MULTI_REGION_NR] = {
-	{ .iova_base = 0x0,		.size = SZ_4G},		/* 0 ~ 4G */
+	{ .iova_base = 0x0,		.size = MTK_IOMMU_IOVA_SZ_4G},	/* 0 ~ 4G,  */
 	#if IS_ENABLED(CONFIG_ARCH_DMA_ADDR_T_64BIT)
-	{ .iova_base = SZ_4G,		.size = SZ_4G},		/* 4G ~ 8G */
-	{ .iova_base = SZ_4G * 2,	.size = SZ_4G},		/* 8G ~ 12G */
-	{ .iova_base = SZ_4G * 3,	.size = SZ_4G},		/* 12G ~ 16G */
+	{ .iova_base = SZ_4G,		.size = MTK_IOMMU_IOVA_SZ_4G},	/* 4G ~ 8G */
+	{ .iova_base = SZ_4G * 2,	.size = MTK_IOMMU_IOVA_SZ_4G},	/* 8G ~ 12G */
+	{ .iova_base = SZ_4G * 3,	.size = MTK_IOMMU_IOVA_SZ_4G},	/* 12G ~ 16G */
 
 	{ .iova_base = 0x240000000ULL,	.size = 0x4000000},	/* CCU0 */
 	{ .iova_base = 0x244000000ULL,	.size = 0x4000000},	/* CCU1 */
-- 
2.25.1