linux-kernel - Re: [PATCH 4/8] iommu/arm-smmu-v3: Add support for Substream IDs

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]

Message-ID: <104a20b7-ebb1-1569-3f6b-94438b9dbf76@arm.com>
Date:   Thu, 4 Jul 2019 10:33:52 +0100
From:   Jean-Philippe Brucker <jean-philippe.brucker@....com>
To:     Will Deacon <will@...nel.org>
Cc:     mark.rutland@....com, devicetree@...r.kernel.org,
        jacob.jun.pan@...ux.intel.com, joro@...tes.org,
        will.deacon@....com, linux-kernel@...r.kernel.org,
        eric.auger@...hat.com, iommu@...ts.linux-foundation.org,
        robh+dt@...nel.org, robin.murphy@....com,
        linux-arm-kernel@...ts.infradead.org
Subject: Re: [PATCH 4/8] iommu/arm-smmu-v3: Add support for Substream IDs

On 26/06/2019 19:00, Will Deacon wrote:
> On Mon, Jun 10, 2019 at 07:47:10PM +0100, Jean-Philippe Brucker wrote:
>> At the moment, the SMMUv3 driver implements only one stage-1 or stage-2
>> page directory per device. However SMMUv3 allows more than one address
>> space for some devices, by providing multiple stage-1 page directories. In
>> addition to the Stream ID (SID), that identifies a device, we can now have
>> Substream IDs (SSID) identifying an address space. In PCIe, SID is called
>> Requester ID (RID) and SSID is called Process Address-Space ID (PASID).
>>
>> Prepare the driver for SSID support, by adding context descriptor tables
>> in STEs (previously a single static context descriptor). A complete
>> stage-1 walk is now performed like this by the SMMU:
>>
>>       Stream tables          Ctx. tables          Page tables
>>         +--------+   ,------->+-------+   ,------->+-------+
>>         :        :   |        :       :   |        :       :
>>         +--------+   |        +-------+   |        +-------+
>>    SID->|  STE   |---'  SSID->|  CD   |---'  IOVA->|  PTE  |--> IPA
>>         +--------+            +-------+            +-------+
>>         :        :            :       :            :       :
>>         +--------+            +-------+            +-------+
>>
>> Implement a single level of context descriptor table for now, but as with
>> stream and page tables, an SSID can be split to index multiple levels of
>> tables.
>>
>> In all stream table entries, we set S1DSS=SSID0 mode, making translations
>> without an SSID use context descriptor 0. Although it would be possible by
>> setting S1DSS=BYPASS, we don't currently support SSID when user selects
>> iommu.passthrough.
> 
> I don't understand your comment here: iommu.passthrough works just as it did
> before, right, since we set bypass in the STE config field so S1DSS is not
> relevant?

Yes the comment is wrong, or at least unclear. It isn't well defined how
SSID is supposed to work with iommu.passthrough, but I guess keeping the
same behavior as non-PASID DMA is what we want (any PASID-tagged DMA
also bypasses the SMMU.)

In the comment I was referring to another possibility, supporting SVA
and auxiliary domains even when iommu.passthrough is set. That would
require allocating context tables and setting S1DSS=BYPASS. But I don't
think it's a feature anyone needs at the moment.

> I also notice that SSID0 causes transactions with SSID==0 to
> abort. Is a PASID of 0 reserved, so this doesn't matter?

Yes, PASID 0 is reserved, we start allocation at 1

> 
>> @@ -1062,33 +1143,90 @@ static u64 arm_smmu_cpu_tcr_to_cd(u64 tcr)
>>  	return val;
>>  }
>>  
>> -static void arm_smmu_write_ctx_desc(struct arm_smmu_device *smmu,
>> -				    struct arm_smmu_s1_cfg *cfg)
>> +static int arm_smmu_write_ctx_desc(struct arm_smmu_domain *smmu_domain,
>> +				   int ssid, struct arm_smmu_ctx_desc *cd)
>>  {
>>  	u64 val;
>> +	bool cd_live;
>> +	struct arm_smmu_device *smmu = smmu_domain->smmu;
>> +	__le64 *cdptr = arm_smmu_get_cd_ptr(&smmu_domain->s1_cfg, ssid);
>>  
>>  	/*
>> -	 * We don't need to issue any invalidation here, as we'll invalidate
>> -	 * the STE when installing the new entry anyway.
>> +	 * This function handles the following cases:
>> +	 *
>> +	 * (1) Install primary CD, for normal DMA traffic (SSID = 0).
>> +	 * (2) Install a secondary CD, for SID+SSID traffic.
>> +	 * (3) Update ASID of a CD. Atomically write the first 64 bits of the
>> +	 *     CD, then invalidate the old entry and mappings.
>> +	 * (4) Remove a secondary CD.
>>  	 */
>> -	val = arm_smmu_cpu_tcr_to_cd(cfg->cd.tcr) |
>> +
>> +	if (!cdptr)
>> +		return -ENOMEM;
>> +
>> +	val = le64_to_cpu(cdptr[0]);
>> +	cd_live = !!(val & CTXDESC_CD_0_V);
>> +
>> +	if (!cd) { /* (4) */
>> +		cdptr[0] = 0;
> 
> Should we be using WRITE_ONCE here? (although I notice we don't seem to
> bother for STEs either...)

Sure, that's safer

Thanks,
Jean