public inbox for kvm@vger.kernel.org
 help / color / mirror / Atom feed
From: Auger Eric <eric.auger@redhat.com>
To: Kunkun Jiang <jiangkunkun@huawei.com>,
	eric.auger.pro@gmail.com, iommu@lists.linux-foundation.org,
	linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
	kvmarm@lists.cs.columbia.edu, will@kernel.org, maz@kernel.org,
	robin.murphy@arm.com, joro@8bytes.org,
	alex.williamson@redhat.com, tn@semihalf.com,
	zhukeqian1@huawei.com
Cc: jacob.jun.pan@linux.intel.com, yi.l.liu@intel.com,
	wangxingang5@huawei.com, jean-philippe@linaro.org,
	zhangfei.gao@linaro.org, zhangfei.gao@gmail.com,
	vivek.gautam@arm.com, shameerali.kolothum.thodi@huawei.com,
	yuzenghui@huawei.com, nicoleotsuka@gmail.com,
	lushenming@huawei.com, vsethi@nvidia.com,
	wanghaibin.wang@huawei.com
Subject: Re: [PATCH v14 06/13] iommu/smmuv3: Allow stage 1 invalidation with unmanaged ASIDs
Date: Thu, 8 Apr 2021 14:30:13 +0200	[thread overview]
Message-ID: <0246aec2-162d-0584-3ca4-b9c304ef3c8a@redhat.com> (raw)
In-Reply-To: <901720e6-6ca5-eb9a-1f24-0ca479bcfecc@huawei.com>

Hi Kunkun,

On 4/1/21 2:37 PM, Kunkun Jiang wrote:
> Hi Eric,
> 
> On 2021/2/24 4:56, Eric Auger wrote:
>> With nested stage support, soon we will need to invalidate
>> S1 contexts and ranges tagged with an unmanaged asid, this
>> latter being managed by the guest. So let's introduce 2 helpers
>> that allow to invalidate with externally managed ASIDs
>>
>> Signed-off-by: Eric Auger <eric.auger@redhat.com>
>>
>> ---
>>
>> v13 -> v14
>> - Actually send the NH_ASID command (reported by Xingang Wang)
>> ---
>>   drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c | 38 ++++++++++++++++-----
>>   1 file changed, 29 insertions(+), 9 deletions(-)
>>
>> diff --git a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c
>> b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c
>> index 5579ec4fccc8..4c19a1114de4 100644
>> --- a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c
>> +++ b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c
>> @@ -1843,9 +1843,9 @@ int arm_smmu_atc_inv_domain(struct
>> arm_smmu_domain *smmu_domain, int ssid,
>>   }
>>     /* IO_PGTABLE API */
>> -static void arm_smmu_tlb_inv_context(void *cookie)
>> +static void __arm_smmu_tlb_inv_context(struct arm_smmu_domain
>> *smmu_domain,
>> +                       int ext_asid)
>>   {
>> -    struct arm_smmu_domain *smmu_domain = cookie;
>>       struct arm_smmu_device *smmu = smmu_domain->smmu;
>>       struct arm_smmu_cmdq_ent cmd;
>>   @@ -1856,7 +1856,13 @@ static void arm_smmu_tlb_inv_context(void
>> *cookie)
>>        * insertion to guarantee those are observed before the TLBI. Do be
>>        * careful, 007.
>>        */
>> -    if (smmu_domain->stage == ARM_SMMU_DOMAIN_S1) {
>> +    if (ext_asid >= 0) { /* guest stage 1 invalidation */
>> +        cmd.opcode    = CMDQ_OP_TLBI_NH_ASID;
>> +        cmd.tlbi.asid    = ext_asid;
>> +        cmd.tlbi.vmid    = smmu_domain->s2_cfg.vmid;
>> +        arm_smmu_cmdq_issue_cmd(smmu, &cmd);
>> +        arm_smmu_cmdq_issue_sync(smmu);
>> +    } else if (smmu_domain->stage == ARM_SMMU_DOMAIN_S1) {
>>           arm_smmu_tlb_inv_asid(smmu, smmu_domain->s1_cfg.cd.asid);
>>       } else {
>>           cmd.opcode    = CMDQ_OP_TLBI_S12_VMALL;
>> @@ -1867,6 +1873,13 @@ static void arm_smmu_tlb_inv_context(void *cookie)
>>       arm_smmu_atc_inv_domain(smmu_domain, 0, 0, 0);
>>   }
>>   +static void arm_smmu_tlb_inv_context(void *cookie)
>> +{
>> +    struct arm_smmu_domain *smmu_domain = cookie;
>> +
>> +    __arm_smmu_tlb_inv_context(smmu_domain, -1);
>> +}
>> +
>>   static void __arm_smmu_tlb_inv_range(struct arm_smmu_cmdq_ent *cmd,
>>                        unsigned long iova, size_t size,
>>                        size_t granule,
>> @@ -1926,9 +1939,10 @@ static void __arm_smmu_tlb_inv_range(struct
>> arm_smmu_cmdq_ent *cmd,
>>       arm_smmu_cmdq_batch_submit(smmu, &cmds);
>>   }
>>   
> Here is the part of code in __arm_smmu_tlb_inv_range():
>>         if (smmu->features & ARM_SMMU_FEAT_RANGE_INV) {
>>                 /* Get the leaf page size */
>>                 tg = __ffs(smmu_domain->domain.pgsize_bitmap);
>>
>>                 /* Convert page size of 12,14,16 (log2) to 1,2,3 */
>>                 cmd->tlbi.tg = (tg - 10) / 2;
>>
>>                 /* Determine what level the granule is at */
>>                 cmd->tlbi.ttl = 4 - ((ilog2(granule) - 3) / (tg - 3));
>>
>>                 num_pages = size >> tg;
>>         }
> When pSMMU supports RIL, we get the leaf page size by __ffs(smmu_domain->
> domain.pgsize_bitmap). In nested mode, it is determined by host
> PAGE_SIZE. If
> the host kernel and guest kernel has different translation granule (e.g.
> host 16K,
> guest 4K), __arm_smmu_tlb_inv_range() will issue an incorrect tlbi command.
> 
> Do you have any idea about this issue?

I think this is the same issue as the one reported by Chenxiang

https://lore.kernel.org/lkml/15938ed5-2095-e903-a290-333c299015a2@hisilicon.com/

In case RIL is not supported by the host, next version will use the
smallest pSMMU supported page size, as done in __arm_smmu_tlb_inv_range

Thanks

Eric

> 
> Best Regards,
> Kunkun Jiang
>> -static void arm_smmu_tlb_inv_range_domain(unsigned long iova, size_t
>> size,
>> -                      size_t granule, bool leaf,
>> -                      struct arm_smmu_domain *smmu_domain)
>> +static void
>> +arm_smmu_tlb_inv_range_domain(unsigned long iova, size_t size,
>> +                  size_t granule, bool leaf, int ext_asid,
>> +                  struct arm_smmu_domain *smmu_domain)
>>   {
>>       struct arm_smmu_cmdq_ent cmd = {
>>           .tlbi = {
>> @@ -1936,7 +1950,12 @@ static void
>> arm_smmu_tlb_inv_range_domain(unsigned long iova, size_t size,
>>           },
>>       };
>>   -    if (smmu_domain->stage == ARM_SMMU_DOMAIN_S1) {
>> +    if (ext_asid >= 0) {  /* guest stage 1 invalidation */
>> +        cmd.opcode    = smmu_domain->smmu->features &
>> ARM_SMMU_FEAT_E2H ?
>> +                  CMDQ_OP_TLBI_EL2_VA : CMDQ_OP_TLBI_NH_VA;
>> +        cmd.tlbi.asid    = ext_asid;
>> +        cmd.tlbi.vmid    = smmu_domain->s2_cfg.vmid;
>> +    } else if (smmu_domain->stage == ARM_SMMU_DOMAIN_S1) {
>>           cmd.opcode    = smmu_domain->smmu->features &
>> ARM_SMMU_FEAT_E2H ?
>>                     CMDQ_OP_TLBI_EL2_VA : CMDQ_OP_TLBI_NH_VA;
>>           cmd.tlbi.asid    = smmu_domain->s1_cfg.cd.asid;
>> @@ -1944,6 +1963,7 @@ static void
>> arm_smmu_tlb_inv_range_domain(unsigned long iova, size_t size,
>>           cmd.opcode    = CMDQ_OP_TLBI_S2_IPA;
>>           cmd.tlbi.vmid    = smmu_domain->s2_cfg.vmid;
>>       }
>> +
>>       __arm_smmu_tlb_inv_range(&cmd, iova, size, granule, smmu_domain);
>>         /*
>> @@ -1982,7 +2002,7 @@ static void arm_smmu_tlb_inv_page_nosync(struct
>> iommu_iotlb_gather *gather,
>>   static void arm_smmu_tlb_inv_walk(unsigned long iova, size_t size,
>>                     size_t granule, void *cookie)
>>   {
>> -    arm_smmu_tlb_inv_range_domain(iova, size, granule, false, cookie);
>> +    arm_smmu_tlb_inv_range_domain(iova, size, granule, false, -1,
>> cookie);
>>   }
>>     static const struct iommu_flush_ops arm_smmu_flush_ops = {
>> @@ -2523,7 +2543,7 @@ static void arm_smmu_iotlb_sync(struct
>> iommu_domain *domain,
>>         arm_smmu_tlb_inv_range_domain(gather->start,
>>                         gather->end - gather->start + 1,
>> -                      gather->pgsize, true, smmu_domain);
>> +                      gather->pgsize, true, -1, smmu_domain);
>>   }
>>     static phys_addr_t
> 
> 


  reply	other threads:[~2021-04-08 12:30 UTC|newest]

Thread overview: 46+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-02-23 20:56 [PATCH v14 00/13] SMMUv3 Nested Stage Setup (IOMMU part) Eric Auger
2021-02-23 20:56 ` [PATCH v14 01/13] iommu: Introduce attach/detach_pasid_table API Eric Auger
2021-02-23 20:56 ` [PATCH v14 02/13] iommu: Introduce bind/unbind_guest_msi Eric Auger
2021-02-23 20:56 ` [PATCH v14 03/13] iommu/smmuv3: Allow s1 and s2 configs to coexist Eric Auger
2021-02-23 20:56 ` [PATCH v14 04/13] iommu/smmuv3: Get prepared for nested stage support Eric Auger
2021-02-23 20:56 ` [PATCH v14 05/13] iommu/smmuv3: Implement attach/detach_pasid_table Eric Auger
2021-03-02  8:35   ` Keqian Zhu
2021-03-19 13:15     ` Auger Eric
2021-03-22  6:23       ` Keqian Zhu
2021-02-23 20:56 ` [PATCH v14 06/13] iommu/smmuv3: Allow stage 1 invalidation with unmanaged ASIDs Eric Auger
2021-03-30  9:17   ` Zenghui Yu
2021-04-01  9:38     ` Auger Eric
2021-04-01 12:37   ` Kunkun Jiang
2021-04-08 12:30     ` Auger Eric [this message]
2021-04-09  4:48       ` Kunkun Jiang
2021-04-09  8:31         ` Auger Eric
2021-04-09  9:43           ` Kunkun Jiang
2021-02-23 20:56 ` [PATCH v14 07/13] iommu/smmuv3: Implement cache_invalidate Eric Auger
     [not found]   ` <c10c6405-5efe-5a41-2b3a-f3af85a528ba@hisilicon.com>
2021-03-19 17:36     ` Auger Eric
2021-03-22  6:40       ` [Linuxarm] " chenxiang (M)
2021-03-22  9:05         ` Auger Eric
2021-03-23  1:28           ` chenxiang (M)
2021-04-01  6:11   ` Zenghui Yu
2021-04-01 12:06     ` Auger Eric
2021-02-23 20:56 ` [PATCH v14 08/13] dma-iommu: Implement NESTED_MSI cookie Eric Auger
2021-04-07  7:39   ` Zenghui Yu
2021-04-10 13:34     ` Auger Eric
2021-02-23 20:56 ` [PATCH v14 09/13] iommu/smmuv3: Nested mode single MSI doorbell per domain enforcement Eric Auger
2021-02-23 20:56 ` [PATCH v14 10/13] iommu/smmuv3: Enforce incompatibility between nested mode and HW MSI regions Eric Auger
2021-02-23 20:56 ` [PATCH v14 11/13] iommu/smmuv3: Implement bind/unbind_guest_msi Eric Auger
2021-02-23 20:56 ` [PATCH v14 12/13] iommu/smmuv3: report additional recoverable faults Eric Auger
2021-02-23 20:56 ` [PATCH v14 13/13] iommu/smmuv3: Accept configs with more than one context descriptor Eric Auger
2021-03-30  9:23   ` Zenghui Yu
2021-04-01 11:48     ` Auger Eric
2021-04-01 12:38       ` Shameerali Kolothum Thodi
2021-04-01 13:20         ` Auger Eric
2021-02-25 16:06 ` [PATCH v14 00/13] SMMUv3 Nested Stage Setup (IOMMU part) Auger Eric
2021-04-22 15:04   ` Sumit Gupta
2021-04-23 13:00     ` Jean-Philippe Brucker
2021-04-23 17:16       ` Sumit Gupta
2021-04-23 17:58         ` Krishna Reddy
2021-04-23 18:19           ` Sumit Gupta
2021-04-24  9:06           ` Marc Zyngier
2021-04-24 11:29             ` Sumit Gupta
2021-03-18  0:16 ` Krishna Reddy
2021-03-19 13:17   ` Auger Eric

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=0246aec2-162d-0584-3ca4-b9c304ef3c8a@redhat.com \
    --to=eric.auger@redhat.com \
    --cc=alex.williamson@redhat.com \
    --cc=eric.auger.pro@gmail.com \
    --cc=iommu@lists.linux-foundation.org \
    --cc=jacob.jun.pan@linux.intel.com \
    --cc=jean-philippe@linaro.org \
    --cc=jiangkunkun@huawei.com \
    --cc=joro@8bytes.org \
    --cc=kvm@vger.kernel.org \
    --cc=kvmarm@lists.cs.columbia.edu \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lushenming@huawei.com \
    --cc=maz@kernel.org \
    --cc=nicoleotsuka@gmail.com \
    --cc=robin.murphy@arm.com \
    --cc=shameerali.kolothum.thodi@huawei.com \
    --cc=tn@semihalf.com \
    --cc=vivek.gautam@arm.com \
    --cc=vsethi@nvidia.com \
    --cc=wanghaibin.wang@huawei.com \
    --cc=wangxingang5@huawei.com \
    --cc=will@kernel.org \
    --cc=yi.l.liu@intel.com \
    --cc=yuzenghui@huawei.com \
    --cc=zhangfei.gao@gmail.com \
    --cc=zhangfei.gao@linaro.org \
    --cc=zhukeqian1@huawei.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox