linux-kernel - Re: [PATCH v3 4/4] arm64: dts: marvell: add SMMU support

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:   Fri, 3 Jul 2020 11:33:20 +0200
From:   Tomasz Nowicki <tn@...ihalf.com>
To:     Robin Murphy <robin.murphy@....com>, will@...nel.org,
        joro@...tes.org, gregory.clement@...tlin.com, robh+dt@...nel.org,
        hannah@...vell.com
Cc:     linux-kernel@...r.kernel.org, iommu@...ts.linux-foundation.org,
        devicetree@...r.kernel.org, catalin.marinas@....com,
        nadavh@...vell.com, linux-arm-kernel@...ts.infradead.org,
        mw@...ihalf.com
Subject: Re: [PATCH v3 4/4] arm64: dts: marvell: add SMMU support

On 03.07.2020 11:16, Robin Murphy wrote:
> On 2020-07-02 21:16, Tomasz Nowicki wrote:
>> From: Marcin Wojtas <mw@...ihalf.com>
>>
>> Add IOMMU node for Marvell AP806 based SoCs together with platform
>> and PCI device Stream ID mapping.
>>
>> Signed-off-by: Marcin Wojtas <mw@...ihalf.com>
>> Signed-off-by: Tomasz Nowicki <tn@...ihalf.com>
>> ---
>>   arch/arm64/boot/dts/marvell/armada-8040.dtsi  | 36 +++++++++++++++++++
>>   arch/arm64/boot/dts/marvell/armada-ap80x.dtsi | 17 +++++++++
>>   2 files changed, 53 insertions(+)
>>
>> diff --git a/arch/arm64/boot/dts/marvell/armada-8040.dtsi 
>> b/arch/arm64/boot/dts/marvell/armada-8040.dtsi
>> index 7699b19224c2..25c1df709f72 100644
>> --- a/arch/arm64/boot/dts/marvell/armada-8040.dtsi
>> +++ b/arch/arm64/boot/dts/marvell/armada-8040.dtsi
>> @@ -23,3 +23,39 @@
>>   &cp0_rtc {
>>       status = "disabled";
>>   };
>> +
>> +&cp0_usb3_0 {
>> +    iommus = <&smmu 0x440>;
>> +};
>> +
>> +&cp0_usb3_1 {
>> +    iommus = <&smmu 0x441>;
>> +};
>> +
>> +&cp0_sata0 {
>> +    iommus = <&smmu 0x444>;
>> +};
>> +
>> +&cp0_sdhci0 {
>> +    iommus = <&smmu 0x445>;
>> +};
>> +
>> +&cp1_sata0 {
>> +    iommus = <&smmu 0x454>;
>> +};
>> +
>> +&cp1_usb3_0 {
>> +    iommus = <&smmu 0x450>;
>> +};
>> +
>> +&cp1_usb3_1 {
>> +    iommus = <&smmu 0x451>;
>> +};
>> +
>> +&cp0_pcie0 {
>> +    iommu-map =
>> +        <0x0   &smmu 0x480 0x20>,
>> +        <0x100 &smmu 0x4a0 0x20>,
>> +        <0x200 &smmu 0x4c0 0x20>;
>> +    iommu-map-mask = <0x031f>;
> 
> Nice! I do like a good compressed mapping :D
> 
>> +};
>> diff --git a/arch/arm64/boot/dts/marvell/armada-ap80x.dtsi 
>> b/arch/arm64/boot/dts/marvell/armada-ap80x.dtsi
>> index 7f9b9a647717..ded8b8082d79 100644
>> --- a/arch/arm64/boot/dts/marvell/armada-ap80x.dtsi
>> +++ b/arch/arm64/boot/dts/marvell/armada-ap80x.dtsi
>> @@ -56,6 +56,23 @@
>>               compatible = "simple-bus";
>>               ranges = <0x0 0x0 0xf0000000 0x1000000>;
>> +            smmu: iommu@...0000 {
>> +                compatible = "marvell,ap806-smmu-500", "arm,mmu-500";
>> +                reg = <0x100000 0x100000>;
>> +                dma-coherent;
>> +                #iommu-cells = <1>;
>> +                #global-interrupts = <1>;
>> +                interrupts = <GIC_SPI 6 IRQ_TYPE_LEVEL_HIGH>,
>> +                         <GIC_SPI 6 IRQ_TYPE_LEVEL_HIGH>,
>> +                         <GIC_SPI 6 IRQ_TYPE_LEVEL_HIGH>,
>> +                         <GIC_SPI 6 IRQ_TYPE_LEVEL_HIGH>,
>> +                         <GIC_SPI 6 IRQ_TYPE_LEVEL_HIGH>,
>> +                         <GIC_SPI 6 IRQ_TYPE_LEVEL_HIGH>,
>> +                         <GIC_SPI 6 IRQ_TYPE_LEVEL_HIGH>,
>> +                         <GIC_SPI 6 IRQ_TYPE_LEVEL_HIGH>,
>> +                         <GIC_SPI 6 IRQ_TYPE_LEVEL_HIGH>;
> 
> I'd recommend you have the node disabled by default here, then 
> explicitly enable it in armada-8040.dtsi where you add the Stream IDs. 
> Otherwise it will also end up enabled for 8020, 70x0, etc. where 
> disable_bypass will then catastrophically break everything.
> 

Good point! I will fix this.

Thanks,
Tomasz