| Message ID | 1527752569-18020-4-git-send-email-thunder.leizhen@huawei.com (mailing list archive) |
|---|---|
| State | Not Applicable, archived |
| Delegated to: | Andy Gross |
| Headers | show |
On 31/05/18 08:42, Zhen Lei wrote: > In common, a IOMMU unmap operation follow the below steps: > 1. remove the mapping in page table of the specified iova range > 2. execute tlbi command to invalid the mapping which is cached in TLB > 3. wait for the above tlbi operation to be finished > 4. free the IOVA resource > 5. free the physical memory resource > > This maybe a problem when unmap is very frequently, the combination of tlbi > and wait operation will consume a lot of time. A feasible method is put off > tlbi and iova-free operation, when accumulating to a certain number or > reaching a specified time, execute only one tlbi_all command to clean up > TLB, then free the backup IOVAs. Mark as non-strict mode. > > But it must be noted that, although the mapping has already been removed in > the page table, it maybe still exist in TLB. And the freed physical memory > may also be reused for others. So a attacker can persistent access to memory > based on the just freed IOVA, to obtain sensible data or corrupt memory. So > the VFIO should always choose the strict mode. > > This patch just add a new parameter for the unmap operation, to help the > upper functions capable choose which mode to be applied. This seems like it might be better handled by a flag in io_pgtable_cfg->quirks. This interface change on its own looks rather invasive, and teh fact that it ends up only being used to pass through a constant property of the domain (which is already known by the point io_pgtable_alloc() is called) implies that it is indeed the wrong level of abstraction. > No functional changes. > > Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com> > --- > drivers/iommu/arm-smmu-v3.c | 2 +- > drivers/iommu/arm-smmu.c | 2 +- > drivers/iommu/io-pgtable-arm-v7s.c | 6 +++--- > drivers/iommu/io-pgtable-arm.c | 6 +++--- > drivers/iommu/io-pgtable.h | 2 +- > drivers/iommu/ipmmu-vmsa.c | 2 +- > drivers/iommu/msm_iommu.c | 2 +- > drivers/iommu/mtk_iommu.c | 2 +- > drivers/iommu/qcom_iommu.c | 2 +- > include/linux/iommu.h | 2 ++ Plus things specific to io-pgtable shouldn't really be spilling into the core API header either. Robin. > 10 files changed, 15 insertions(+), 13 deletions(-) > > diff --git a/drivers/iommu/arm-smmu-v3.c b/drivers/iommu/arm-smmu-v3.c > index 4402187..59b3387 100644 > --- a/drivers/iommu/arm-smmu-v3.c > +++ b/drivers/iommu/arm-smmu-v3.c > @@ -1767,7 +1767,7 @@ static int arm_smmu_map(struct iommu_domain *domain, unsigned long iova, > if (!ops) > return 0; > > - return ops->unmap(ops, iova, size); > + return ops->unmap(ops, iova, size, IOMMU_STRICT); > } > > static void arm_smmu_flush_iotlb_all(struct iommu_domain *domain) > diff --git a/drivers/iommu/arm-smmu.c b/drivers/iommu/arm-smmu.c > index 69e7c60..253e807 100644 > --- a/drivers/iommu/arm-smmu.c > +++ b/drivers/iommu/arm-smmu.c > @@ -1249,7 +1249,7 @@ static size_t arm_smmu_unmap(struct iommu_domain *domain, unsigned long iova, > if (!ops) > return 0; > > - return ops->unmap(ops, iova, size); > + return ops->unmap(ops, iova, size, IOMMU_STRICT); > } > > static void arm_smmu_iotlb_sync(struct iommu_domain *domain) > diff --git a/drivers/iommu/io-pgtable-arm-v7s.c b/drivers/iommu/io-pgtable-arm-v7s.c > index 10e4a3d..799eced 100644 > --- a/drivers/iommu/io-pgtable-arm-v7s.c > +++ b/drivers/iommu/io-pgtable-arm-v7s.c > @@ -658,7 +658,7 @@ static size_t __arm_v7s_unmap(struct arm_v7s_io_pgtable *data, > } > > static size_t arm_v7s_unmap(struct io_pgtable_ops *ops, unsigned long iova, > - size_t size) > + size_t size, int strict) > { > struct arm_v7s_io_pgtable *data = io_pgtable_ops_to_data(ops); > > @@ -883,7 +883,7 @@ static int __init arm_v7s_do_selftests(void) > size = 1UL << __ffs(cfg.pgsize_bitmap); > while (i < loopnr) { > iova_start = i * SZ_16M; > - if (ops->unmap(ops, iova_start + size, size) != size) > + if (ops->unmap(ops, iova_start + size, size, IOMMU_STRICT) != size) > return __FAIL(ops); > > /* Remap of partial unmap */ > @@ -902,7 +902,7 @@ static int __init arm_v7s_do_selftests(void) > while (i != BITS_PER_LONG) { > size = 1UL << i; > > - if (ops->unmap(ops, iova, size) != size) > + if (ops->unmap(ops, iova, size, IOMMU_STRICT) != size) > return __FAIL(ops); > > if (ops->iova_to_phys(ops, iova + 42)) > diff --git a/drivers/iommu/io-pgtable-arm.c b/drivers/iommu/io-pgtable-arm.c > index 39c2a05..e0f52db 100644 > --- a/drivers/iommu/io-pgtable-arm.c > +++ b/drivers/iommu/io-pgtable-arm.c > @@ -624,7 +624,7 @@ static size_t __arm_lpae_unmap(struct arm_lpae_io_pgtable *data, > } > > static size_t arm_lpae_unmap(struct io_pgtable_ops *ops, unsigned long iova, > - size_t size) > + size_t size, int strict) > { > struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops); > arm_lpae_iopte *ptep = data->pgd; > @@ -1108,7 +1108,7 @@ static int __init arm_lpae_run_tests(struct io_pgtable_cfg *cfg) > > /* Partial unmap */ > size = 1UL << __ffs(cfg->pgsize_bitmap); > - if (ops->unmap(ops, SZ_1G + size, size) != size) > + if (ops->unmap(ops, SZ_1G + size, size, IOMMU_STRICT) != size) > return __FAIL(ops, i); > > /* Remap of partial unmap */ > @@ -1124,7 +1124,7 @@ static int __init arm_lpae_run_tests(struct io_pgtable_cfg *cfg) > while (j != BITS_PER_LONG) { > size = 1UL << j; > > - if (ops->unmap(ops, iova, size) != size) > + if (ops->unmap(ops, iova, size, IOMMU_STRICT) != size) > return __FAIL(ops, i); > > if (ops->iova_to_phys(ops, iova + 42)) > diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h > index 2df7909..2908806 100644 > --- a/drivers/iommu/io-pgtable.h > +++ b/drivers/iommu/io-pgtable.h > @@ -120,7 +120,7 @@ struct io_pgtable_ops { > int (*map)(struct io_pgtable_ops *ops, unsigned long iova, > phys_addr_t paddr, size_t size, int prot); > size_t (*unmap)(struct io_pgtable_ops *ops, unsigned long iova, > - size_t size); > + size_t size, int strict); > phys_addr_t (*iova_to_phys)(struct io_pgtable_ops *ops, > unsigned long iova); > }; > diff --git a/drivers/iommu/ipmmu-vmsa.c b/drivers/iommu/ipmmu-vmsa.c > index 40ae6e8..e6d9e11 100644 > --- a/drivers/iommu/ipmmu-vmsa.c > +++ b/drivers/iommu/ipmmu-vmsa.c > @@ -716,7 +716,7 @@ static size_t ipmmu_unmap(struct iommu_domain *io_domain, unsigned long iova, > { > struct ipmmu_vmsa_domain *domain = to_vmsa_domain(io_domain); > > - return domain->iop->unmap(domain->iop, iova, size); > + return domain->iop->unmap(domain->iop, iova, size, IOMMU_STRICT); > } > > static void ipmmu_iotlb_sync(struct iommu_domain *io_domain) > diff --git a/drivers/iommu/msm_iommu.c b/drivers/iommu/msm_iommu.c > index 0d33504..180fa3d 100644 > --- a/drivers/iommu/msm_iommu.c > +++ b/drivers/iommu/msm_iommu.c > @@ -532,7 +532,7 @@ static size_t msm_iommu_unmap(struct iommu_domain *domain, unsigned long iova, > unsigned long flags; > > spin_lock_irqsave(&priv->pgtlock, flags); > - len = priv->iop->unmap(priv->iop, iova, len); > + len = priv->iop->unmap(priv->iop, iova, len, IOMMU_STRICT); > spin_unlock_irqrestore(&priv->pgtlock, flags); > > return len; > diff --git a/drivers/iommu/mtk_iommu.c b/drivers/iommu/mtk_iommu.c > index f2832a1..54661ed 100644 > --- a/drivers/iommu/mtk_iommu.c > +++ b/drivers/iommu/mtk_iommu.c > @@ -386,7 +386,7 @@ static size_t mtk_iommu_unmap(struct iommu_domain *domain, > size_t unmapsz; > > spin_lock_irqsave(&dom->pgtlock, flags); > - unmapsz = dom->iop->unmap(dom->iop, iova, size); > + unmapsz = dom->iop->unmap(dom->iop, iova, size, IOMMU_STRICT); > spin_unlock_irqrestore(&dom->pgtlock, flags); > > return unmapsz; > diff --git a/drivers/iommu/qcom_iommu.c b/drivers/iommu/qcom_iommu.c > index 65b9c99..90abde1 100644 > --- a/drivers/iommu/qcom_iommu.c > +++ b/drivers/iommu/qcom_iommu.c > @@ -444,7 +444,7 @@ static size_t qcom_iommu_unmap(struct iommu_domain *domain, unsigned long iova, > */ > pm_runtime_get_sync(qcom_domain->iommu->dev); > spin_lock_irqsave(&qcom_domain->pgtbl_lock, flags); > - ret = ops->unmap(ops, iova, size); > + ret = ops->unmap(ops, iova, size, IOMMU_STRICT); > spin_unlock_irqrestore(&qcom_domain->pgtbl_lock, flags); > pm_runtime_put_sync(qcom_domain->iommu->dev); > > diff --git a/include/linux/iommu.h b/include/linux/iommu.h > index 19938ee..39b3150 100644 > --- a/include/linux/iommu.h > +++ b/include/linux/iommu.h > @@ -86,6 +86,8 @@ struct iommu_domain_geometry { > #define IOMMU_DOMAIN_DMA (__IOMMU_DOMAIN_PAGING | \ > __IOMMU_DOMAIN_DMA_API) > > +#define IOMMU_STRICT 1 > + > struct iommu_domain { > unsigned type; > const struct iommu_ops *ops; > -- To unsubscribe from this list: send the line "unsubscribe linux-arm-msm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 2018/5/31 21:04, Robin Murphy wrote: > On 31/05/18 08:42, Zhen Lei wrote: >> In common, a IOMMU unmap operation follow the below steps: >> 1. remove the mapping in page table of the specified iova range >> 2. execute tlbi command to invalid the mapping which is cached in TLB >> 3. wait for the above tlbi operation to be finished >> 4. free the IOVA resource >> 5. free the physical memory resource >> >> This maybe a problem when unmap is very frequently, the combination of tlbi >> and wait operation will consume a lot of time. A feasible method is put off >> tlbi and iova-free operation, when accumulating to a certain number or >> reaching a specified time, execute only one tlbi_all command to clean up >> TLB, then free the backup IOVAs. Mark as non-strict mode. >> >> But it must be noted that, although the mapping has already been removed in >> the page table, it maybe still exist in TLB. And the freed physical memory >> may also be reused for others. So a attacker can persistent access to memory >> based on the just freed IOVA, to obtain sensible data or corrupt memory. So >> the VFIO should always choose the strict mode. >> >> This patch just add a new parameter for the unmap operation, to help the >> upper functions capable choose which mode to be applied. > > This seems like it might be better handled by a flag in io_pgtable_cfg->quirks. This interface change on its own looks rather invasive, and teh fact that it ends up only being used to pass through a constant property of the domain (which is already known by the point io_pgtable_alloc() is called) implies that it is indeed the wrong level of abstraction. > Sound good. Thanks for your suggestion, I will try it in v2. >> No functional changes. >> >> Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com> >> --- >> drivers/iommu/arm-smmu-v3.c | 2 +- >> drivers/iommu/arm-smmu.c | 2 +- >> drivers/iommu/io-pgtable-arm-v7s.c | 6 +++--- >> drivers/iommu/io-pgtable-arm.c | 6 +++--- >> drivers/iommu/io-pgtable.h | 2 +- >> drivers/iommu/ipmmu-vmsa.c | 2 +- >> drivers/iommu/msm_iommu.c | 2 +- >> drivers/iommu/mtk_iommu.c | 2 +- >> drivers/iommu/qcom_iommu.c | 2 +- >> include/linux/iommu.h | 2 ++ > > Plus things specific to io-pgtable shouldn't really be spilling into the core API header either. > > Robin. > >> 10 files changed, 15 insertions(+), 13 deletions(-) >> >> diff --git a/drivers/iommu/arm-smmu-v3.c b/drivers/iommu/arm-smmu-v3.c >> index 4402187..59b3387 100644 >> --- a/drivers/iommu/arm-smmu-v3.c >> +++ b/drivers/iommu/arm-smmu-v3.c >> @@ -1767,7 +1767,7 @@ static int arm_smmu_map(struct iommu_domain *domain, unsigned long iova, >> if (!ops) >> return 0; >> - return ops->unmap(ops, iova, size); >> + return ops->unmap(ops, iova, size, IOMMU_STRICT); >> } >> static void arm_smmu_flush_iotlb_all(struct iommu_domain *domain) >> diff --git a/drivers/iommu/arm-smmu.c b/drivers/iommu/arm-smmu.c >> index 69e7c60..253e807 100644 >> --- a/drivers/iommu/arm-smmu.c >> +++ b/drivers/iommu/arm-smmu.c >> @@ -1249,7 +1249,7 @@ static size_t arm_smmu_unmap(struct iommu_domain *domain, unsigned long iova, >> if (!ops) >> return 0; >> - return ops->unmap(ops, iova, size); >> + return ops->unmap(ops, iova, size, IOMMU_STRICT); >> } >> static void arm_smmu_iotlb_sync(struct iommu_domain *domain) >> diff --git a/drivers/iommu/io-pgtable-arm-v7s.c b/drivers/iommu/io-pgtable-arm-v7s.c >> index 10e4a3d..799eced 100644 >> --- a/drivers/iommu/io-pgtable-arm-v7s.c >> +++ b/drivers/iommu/io-pgtable-arm-v7s.c >> @@ -658,7 +658,7 @@ static size_t __arm_v7s_unmap(struct arm_v7s_io_pgtable *data, >> } >> static size_t arm_v7s_unmap(struct io_pgtable_ops *ops, unsigned long iova, >> - size_t size) >> + size_t size, int strict) >> { >> struct arm_v7s_io_pgtable *data = io_pgtable_ops_to_data(ops); >> @@ -883,7 +883,7 @@ static int __init arm_v7s_do_selftests(void) >> size = 1UL << __ffs(cfg.pgsize_bitmap); >> while (i < loopnr) { >> iova_start = i * SZ_16M; >> - if (ops->unmap(ops, iova_start + size, size) != size) >> + if (ops->unmap(ops, iova_start + size, size, IOMMU_STRICT) != size) >> return __FAIL(ops); >> /* Remap of partial unmap */ >> @@ -902,7 +902,7 @@ static int __init arm_v7s_do_selftests(void) >> while (i != BITS_PER_LONG) { >> size = 1UL << i; >> - if (ops->unmap(ops, iova, size) != size) >> + if (ops->unmap(ops, iova, size, IOMMU_STRICT) != size) >> return __FAIL(ops); >> if (ops->iova_to_phys(ops, iova + 42)) >> diff --git a/drivers/iommu/io-pgtable-arm.c b/drivers/iommu/io-pgtable-arm.c >> index 39c2a05..e0f52db 100644 >> --- a/drivers/iommu/io-pgtable-arm.c >> +++ b/drivers/iommu/io-pgtable-arm.c >> @@ -624,7 +624,7 @@ static size_t __arm_lpae_unmap(struct arm_lpae_io_pgtable *data, >> } >> static size_t arm_lpae_unmap(struct io_pgtable_ops *ops, unsigned long iova, >> - size_t size) >> + size_t size, int strict) >> { >> struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops); >> arm_lpae_iopte *ptep = data->pgd; >> @@ -1108,7 +1108,7 @@ static int __init arm_lpae_run_tests(struct io_pgtable_cfg *cfg) >> /* Partial unmap */ >> size = 1UL << __ffs(cfg->pgsize_bitmap); >> - if (ops->unmap(ops, SZ_1G + size, size) != size) >> + if (ops->unmap(ops, SZ_1G + size, size, IOMMU_STRICT) != size) >> return __FAIL(ops, i); >> /* Remap of partial unmap */ >> @@ -1124,7 +1124,7 @@ static int __init arm_lpae_run_tests(struct io_pgtable_cfg *cfg) >> while (j != BITS_PER_LONG) { >> size = 1UL << j; >> - if (ops->unmap(ops, iova, size) != size) >> + if (ops->unmap(ops, iova, size, IOMMU_STRICT) != size) >> return __FAIL(ops, i); >> if (ops->iova_to_phys(ops, iova + 42)) >> diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h >> index 2df7909..2908806 100644 >> --- a/drivers/iommu/io-pgtable.h >> +++ b/drivers/iommu/io-pgtable.h >> @@ -120,7 +120,7 @@ struct io_pgtable_ops { >> int (*map)(struct io_pgtable_ops *ops, unsigned long iova, >> phys_addr_t paddr, size_t size, int prot); >> size_t (*unmap)(struct io_pgtable_ops *ops, unsigned long iova, >> - size_t size); >> + size_t size, int strict); >> phys_addr_t (*iova_to_phys)(struct io_pgtable_ops *ops, >> unsigned long iova); >> }; >> diff --git a/drivers/iommu/ipmmu-vmsa.c b/drivers/iommu/ipmmu-vmsa.c >> index 40ae6e8..e6d9e11 100644 >> --- a/drivers/iommu/ipmmu-vmsa.c >> +++ b/drivers/iommu/ipmmu-vmsa.c >> @@ -716,7 +716,7 @@ static size_t ipmmu_unmap(struct iommu_domain *io_domain, unsigned long iova, >> { >> struct ipmmu_vmsa_domain *domain = to_vmsa_domain(io_domain); >> - return domain->iop->unmap(domain->iop, iova, size); >> + return domain->iop->unmap(domain->iop, iova, size, IOMMU_STRICT); >> } >> static void ipmmu_iotlb_sync(struct iommu_domain *io_domain) >> diff --git a/drivers/iommu/msm_iommu.c b/drivers/iommu/msm_iommu.c >> index 0d33504..180fa3d 100644 >> --- a/drivers/iommu/msm_iommu.c >> +++ b/drivers/iommu/msm_iommu.c >> @@ -532,7 +532,7 @@ static size_t msm_iommu_unmap(struct iommu_domain *domain, unsigned long iova, >> unsigned long flags; >> spin_lock_irqsave(&priv->pgtlock, flags); >> - len = priv->iop->unmap(priv->iop, iova, len); >> + len = priv->iop->unmap(priv->iop, iova, len, IOMMU_STRICT); >> spin_unlock_irqrestore(&priv->pgtlock, flags); >> return len; >> diff --git a/drivers/iommu/mtk_iommu.c b/drivers/iommu/mtk_iommu.c >> index f2832a1..54661ed 100644 >> --- a/drivers/iommu/mtk_iommu.c >> +++ b/drivers/iommu/mtk_iommu.c >> @@ -386,7 +386,7 @@ static size_t mtk_iommu_unmap(struct iommu_domain *domain, >> size_t unmapsz; >> spin_lock_irqsave(&dom->pgtlock, flags); >> - unmapsz = dom->iop->unmap(dom->iop, iova, size); >> + unmapsz = dom->iop->unmap(dom->iop, iova, size, IOMMU_STRICT); >> spin_unlock_irqrestore(&dom->pgtlock, flags); >> return unmapsz; >> diff --git a/drivers/iommu/qcom_iommu.c b/drivers/iommu/qcom_iommu.c >> index 65b9c99..90abde1 100644 >> --- a/drivers/iommu/qcom_iommu.c >> +++ b/drivers/iommu/qcom_iommu.c >> @@ -444,7 +444,7 @@ static size_t qcom_iommu_unmap(struct iommu_domain *domain, unsigned long iova, >> */ >> pm_runtime_get_sync(qcom_domain->iommu->dev); >> spin_lock_irqsave(&qcom_domain->pgtbl_lock, flags); >> - ret = ops->unmap(ops, iova, size); >> + ret = ops->unmap(ops, iova, size, IOMMU_STRICT); >> spin_unlock_irqrestore(&qcom_domain->pgtbl_lock, flags); >> pm_runtime_put_sync(qcom_domain->iommu->dev); >> diff --git a/include/linux/iommu.h b/include/linux/iommu.h >> index 19938ee..39b3150 100644 >> --- a/include/linux/iommu.h >> +++ b/include/linux/iommu.h >> @@ -86,6 +86,8 @@ struct iommu_domain_geometry { >> #define IOMMU_DOMAIN_DMA (__IOMMU_DOMAIN_PAGING | \ >> __IOMMU_DOMAIN_DMA_API) >> +#define IOMMU_STRICT 1 >> + >> struct iommu_domain { >> unsigned type; >> const struct iommu_ops *ops; >> > > . >
diff --git a/drivers/iommu/arm-smmu-v3.c b/drivers/iommu/arm-smmu-v3.c index 4402187..59b3387 100644 --- a/drivers/iommu/arm-smmu-v3.c +++ b/drivers/iommu/arm-smmu-v3.c @@ -1767,7 +1767,7 @@ static int arm_smmu_map(struct iommu_domain *domain, unsigned long iova, if (!ops) return 0; - return ops->unmap(ops, iova, size); + return ops->unmap(ops, iova, size, IOMMU_STRICT); } static void arm_smmu_flush_iotlb_all(struct iommu_domain *domain) diff --git a/drivers/iommu/arm-smmu.c b/drivers/iommu/arm-smmu.c index 69e7c60..253e807 100644 --- a/drivers/iommu/arm-smmu.c +++ b/drivers/iommu/arm-smmu.c @@ -1249,7 +1249,7 @@ static size_t arm_smmu_unmap(struct iommu_domain *domain, unsigned long iova, if (!ops) return 0; - return ops->unmap(ops, iova, size); + return ops->unmap(ops, iova, size, IOMMU_STRICT); } static void arm_smmu_iotlb_sync(struct iommu_domain *domain) diff --git a/drivers/iommu/io-pgtable-arm-v7s.c b/drivers/iommu/io-pgtable-arm-v7s.c index 10e4a3d..799eced 100644 --- a/drivers/iommu/io-pgtable-arm-v7s.c +++ b/drivers/iommu/io-pgtable-arm-v7s.c @@ -658,7 +658,7 @@ static size_t __arm_v7s_unmap(struct arm_v7s_io_pgtable *data, } static size_t arm_v7s_unmap(struct io_pgtable_ops *ops, unsigned long iova, - size_t size) + size_t size, int strict) { struct arm_v7s_io_pgtable *data = io_pgtable_ops_to_data(ops); @@ -883,7 +883,7 @@ static int __init arm_v7s_do_selftests(void) size = 1UL << __ffs(cfg.pgsize_bitmap); while (i < loopnr) { iova_start = i * SZ_16M; - if (ops->unmap(ops, iova_start + size, size) != size) + if (ops->unmap(ops, iova_start + size, size, IOMMU_STRICT) != size) return __FAIL(ops); /* Remap of partial unmap */ @@ -902,7 +902,7 @@ static int __init arm_v7s_do_selftests(void) while (i != BITS_PER_LONG) { size = 1UL << i; - if (ops->unmap(ops, iova, size) != size) + if (ops->unmap(ops, iova, size, IOMMU_STRICT) != size) return __FAIL(ops); if (ops->iova_to_phys(ops, iova + 42)) diff --git a/drivers/iommu/io-pgtable-arm.c b/drivers/iommu/io-pgtable-arm.c index 39c2a05..e0f52db 100644 --- a/drivers/iommu/io-pgtable-arm.c +++ b/drivers/iommu/io-pgtable-arm.c @@ -624,7 +624,7 @@ static size_t __arm_lpae_unmap(struct arm_lpae_io_pgtable *data, } static size_t arm_lpae_unmap(struct io_pgtable_ops *ops, unsigned long iova, - size_t size) + size_t size, int strict) { struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops); arm_lpae_iopte *ptep = data->pgd; @@ -1108,7 +1108,7 @@ static int __init arm_lpae_run_tests(struct io_pgtable_cfg *cfg) /* Partial unmap */ size = 1UL << __ffs(cfg->pgsize_bitmap); - if (ops->unmap(ops, SZ_1G + size, size) != size) + if (ops->unmap(ops, SZ_1G + size, size, IOMMU_STRICT) != size) return __FAIL(ops, i); /* Remap of partial unmap */ @@ -1124,7 +1124,7 @@ static int __init arm_lpae_run_tests(struct io_pgtable_cfg *cfg) while (j != BITS_PER_LONG) { size = 1UL << j; - if (ops->unmap(ops, iova, size) != size) + if (ops->unmap(ops, iova, size, IOMMU_STRICT) != size) return __FAIL(ops, i); if (ops->iova_to_phys(ops, iova + 42)) diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h index 2df7909..2908806 100644 --- a/drivers/iommu/io-pgtable.h +++ b/drivers/iommu/io-pgtable.h @@ -120,7 +120,7 @@ struct io_pgtable_ops { int (*map)(struct io_pgtable_ops *ops, unsigned long iova, phys_addr_t paddr, size_t size, int prot); size_t (*unmap)(struct io_pgtable_ops *ops, unsigned long iova, - size_t size); + size_t size, int strict); phys_addr_t (*iova_to_phys)(struct io_pgtable_ops *ops, unsigned long iova); }; diff --git a/drivers/iommu/ipmmu-vmsa.c b/drivers/iommu/ipmmu-vmsa.c index 40ae6e8..e6d9e11 100644 --- a/drivers/iommu/ipmmu-vmsa.c +++ b/drivers/iommu/ipmmu-vmsa.c @@ -716,7 +716,7 @@ static size_t ipmmu_unmap(struct iommu_domain *io_domain, unsigned long iova, { struct ipmmu_vmsa_domain *domain = to_vmsa_domain(io_domain); - return domain->iop->unmap(domain->iop, iova, size); + return domain->iop->unmap(domain->iop, iova, size, IOMMU_STRICT); } static void ipmmu_iotlb_sync(struct iommu_domain *io_domain) diff --git a/drivers/iommu/msm_iommu.c b/drivers/iommu/msm_iommu.c index 0d33504..180fa3d 100644 --- a/drivers/iommu/msm_iommu.c +++ b/drivers/iommu/msm_iommu.c @@ -532,7 +532,7 @@ static size_t msm_iommu_unmap(struct iommu_domain *domain, unsigned long iova, unsigned long flags; spin_lock_irqsave(&priv->pgtlock, flags); - len = priv->iop->unmap(priv->iop, iova, len); + len = priv->iop->unmap(priv->iop, iova, len, IOMMU_STRICT); spin_unlock_irqrestore(&priv->pgtlock, flags); return len; diff --git a/drivers/iommu/mtk_iommu.c b/drivers/iommu/mtk_iommu.c index f2832a1..54661ed 100644 --- a/drivers/iommu/mtk_iommu.c +++ b/drivers/iommu/mtk_iommu.c @@ -386,7 +386,7 @@ static size_t mtk_iommu_unmap(struct iommu_domain *domain, size_t unmapsz; spin_lock_irqsave(&dom->pgtlock, flags); - unmapsz = dom->iop->unmap(dom->iop, iova, size); + unmapsz = dom->iop->unmap(dom->iop, iova, size, IOMMU_STRICT); spin_unlock_irqrestore(&dom->pgtlock, flags); return unmapsz; diff --git a/drivers/iommu/qcom_iommu.c b/drivers/iommu/qcom_iommu.c index 65b9c99..90abde1 100644 --- a/drivers/iommu/qcom_iommu.c +++ b/drivers/iommu/qcom_iommu.c @@ -444,7 +444,7 @@ static size_t qcom_iommu_unmap(struct iommu_domain *domain, unsigned long iova, */ pm_runtime_get_sync(qcom_domain->iommu->dev); spin_lock_irqsave(&qcom_domain->pgtbl_lock, flags); - ret = ops->unmap(ops, iova, size); + ret = ops->unmap(ops, iova, size, IOMMU_STRICT); spin_unlock_irqrestore(&qcom_domain->pgtbl_lock, flags); pm_runtime_put_sync(qcom_domain->iommu->dev); diff --git a/include/linux/iommu.h b/include/linux/iommu.h index 19938ee..39b3150 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -86,6 +86,8 @@ struct iommu_domain_geometry { #define IOMMU_DOMAIN_DMA (__IOMMU_DOMAIN_PAGING | \ __IOMMU_DOMAIN_DMA_API) +#define IOMMU_STRICT 1 + struct iommu_domain { unsigned type; const struct iommu_ops *ops;
In common, a IOMMU unmap operation follow the below steps: 1. remove the mapping in page table of the specified iova range 2. execute tlbi command to invalid the mapping which is cached in TLB 3. wait for the above tlbi operation to be finished 4. free the IOVA resource 5. free the physical memory resource This maybe a problem when unmap is very frequently, the combination of tlbi and wait operation will consume a lot of time. A feasible method is put off tlbi and iova-free operation, when accumulating to a certain number or reaching a specified time, execute only one tlbi_all command to clean up TLB, then free the backup IOVAs. Mark as non-strict mode. But it must be noted that, although the mapping has already been removed in the page table, it maybe still exist in TLB. And the freed physical memory may also be reused for others. So a attacker can persistent access to memory based on the just freed IOVA, to obtain sensible data or corrupt memory. So the VFIO should always choose the strict mode. This patch just add a new parameter for the unmap operation, to help the upper functions capable choose which mode to be applied. No functional changes. Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com> --- drivers/iommu/arm-smmu-v3.c | 2 +- drivers/iommu/arm-smmu.c | 2 +- drivers/iommu/io-pgtable-arm-v7s.c | 6 +++--- drivers/iommu/io-pgtable-arm.c | 6 +++--- drivers/iommu/io-pgtable.h | 2 +- drivers/iommu/ipmmu-vmsa.c | 2 +- drivers/iommu/msm_iommu.c | 2 +- drivers/iommu/mtk_iommu.c | 2 +- drivers/iommu/qcom_iommu.c | 2 +- include/linux/iommu.h | 2 ++ 10 files changed, 15 insertions(+), 13 deletions(-)