linux-kernel - [PATCH] iommu/qcom: restore IOMMU state if needed

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-next>] [day] [month] [year] [list]

Message-Id: <20231011-msm8953-iommu-restore-v1-1-48a0c93809a2@z3ntu.xyz>
Date:   Wed, 11 Oct 2023 19:57:26 +0200
From:   Luca Weiss <luca@...tu.xyz>
To:     ~postmarketos/upstreaming@...ts.sr.ht, phone-devel@...r.kernel.org,
        Rob Clark <robdclark@...il.com>, Will Deacon <will@...nel.org>,
        Robin Murphy <robin.murphy@....com>,
        Joerg Roedel <joro@...tes.org>
Cc:     iommu@...ts.linux.dev, linux-arm-msm@...r.kernel.org,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
        Vladimir Lypak <vladimir.lypak@...il.com>,
        Luca Weiss <luca@...tu.xyz>
Subject: [PATCH] iommu/qcom: restore IOMMU state if needed

From: Vladimir Lypak <vladimir.lypak@...il.com>

If the IOMMU has a power domain then some state will be lost in
qcom_iommu_suspend and TZ will reset device if we don't call
qcom_scm_restore_sec_cfg before accessing it again.

Signed-off-by: Vladimir Lypak <vladimir.lypak@...il.com>
[luca@...tu.xyz: reword commit message a bit]
Signed-off-by: Luca Weiss <luca@...tu.xyz>
---
This patch is required for MSM8953 GPU IOMMU.

See also downstream sources:
https://git.codelinaro.org/clo/la/kernel/msm-4.9/-/commit/f8464885dafc5b780b85de29d92a08c692d3a4d0
https://git.codelinaro.org/clo/la/kernel/msm-4.9/-/blob/LA.UM.10.6.2.r1-02500-89xx.0/drivers/iommu/arm-smmu.c#L4221-4225

Since the compatibles provided by this driver (qcom,msm-iommu-v*) is
only used on msm8916, msm8939 and msm8953, and both 8916 and 8939 don't
have a power domain on the IOMMU, I also don't expect anything to break
there.
---
 drivers/iommu/arm/arm-smmu/qcom_iommu.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/drivers/iommu/arm/arm-smmu/qcom_iommu.c b/drivers/iommu/arm/arm-smmu/qcom_iommu.c
index 97b2122032b2..67abeb02cf53 100644
--- a/drivers/iommu/arm/arm-smmu/qcom_iommu.c
+++ b/drivers/iommu/arm/arm-smmu/qcom_iommu.c
@@ -900,8 +900,16 @@ static void qcom_iommu_device_remove(struct platform_device *pdev)
 static int __maybe_unused qcom_iommu_resume(struct device *dev)
 {
 	struct qcom_iommu_dev *qcom_iommu = dev_get_drvdata(dev);
+	int ret;
+
+	ret = clk_bulk_prepare_enable(CLK_NUM, qcom_iommu->clks);
+	if (ret < 0)
+		return ret;
+
+	if (dev->pm_domain)
+		return qcom_scm_restore_sec_cfg(qcom_iommu->sec_id, 0);
 
-	return clk_bulk_prepare_enable(CLK_NUM, qcom_iommu->clks);
+	return ret;
 }
 
 static int __maybe_unused qcom_iommu_suspend(struct device *dev)

---
base-commit: 2933a1156742d8c47550493a77af8e2d81cf3c84
change-id: 20231011-msm8953-iommu-restore-fabc2e31fee7

Best regards,
-- 
Luca Weiss <luca@...tu.xyz>