Skip to contentSkip to conversation

iommu/vt-d: Force snooping improvement

PATCH In review Lu Baolu 2 participants 7 messages May 8, 2022 – May 10, 2022

Subsystems (2)
lore
Share & tools
lore

Lu Baolu

author maintainer
#0 lore raw

v4 cover iommu/vt-d: Force snooping improvement

Hi folks,

Previously, the IOMMU capability of enforcing cache coherency was queried through iommu_capable(IOMMU_CAP_CACHE_COHERENCY). This is a global capability, hence the IOMMU driver reports support for this capability only when all IOMMUs in the system has this support.

Commit 6043257b1de06 ("iommu: Introduce the domain op enforce_cache_coherency()") converts this into a per-domain test-and-set option, and the previous iommu_capable(IOMMU_CAP_CACHE_COHERENCY) is deprecated.

This is a follow-up series which improves the Intel IOMMU driver to support the per-domain scheme better.

Best regards,
baolu

Change log:
v4:
- Flush caches after changing PGSNP bit in the right way.

v3:
- https://lore.kernel.org/linux-iommu/[email protected]/
- Hold the device_domain_lock when check and set force snooping.
- Refind the commit messages.

v2:
- https://lore.kernel.org/linux-iommu/[email protected]/
- Check whether force_snooping has already been set in intel_iommu_enforce_cache_coherency().
- Set PGSNP pasid bit field during domain attaching if forcing_snooping
is set.
- Remove redundant list_empty() checks.
- Add dmar_domain->set_pte_snp and set it if force snooping is enforced
on a domain with 2nd-level translation.

v1:
- https://lore.kernel.org/linux-iommu/[email protected]
- Initial post.

Lu Baolu (4):
iommu/vt-d: Block force-snoop domain attaching if no SC support iommu/vt-d: Check domain force_snooping against attached devices iommu/vt-d: Remove domain_update_iommu_snooping() iommu/vt-d: Remove hard coding PGSNP bit in PASID entries

 include/linux/intel-iommu.h |  1 +
 drivers/iommu/intel/pasid.h |  2 +
 drivers/iommu/intel/iommu.c | 90 ++++++++++++++++++++++---------------
 drivers/iommu/intel/pasid.c | 45 +++++++++++++++++--
 4 files changed, 99 insertions(+), 39 deletions(-)
Signature · 1 line
2.25.1
author maintainer
#1 lore raw

PATCH v4 1/4 iommu/vt-d: Block force-snoop domain attaching if no SC support

In the attach_dev callback of the default domain ops, if the domain has been set force_snooping, but the iommu hardware of the device does not support SC(Snoop Control) capability, the callback should block it and return a corresponding error code.

---
 drivers/iommu/intel/iommu.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/iommu/intel/iommu.c +3 -0
index 2990f80c5e08..b4802f4055a0 100644
@@ -4367,6 +4367,9 @@ static int prepare_domain_attach_device(struct iommu_domain *domain,
 	if (!iommu)
 		return -ENODEV;
 
+	if (dmar_domain->force_snooping && !ecap_sc_support(iommu->ecap))
+		return -EOPNOTSUPP;
+
 	/* check if this iommu agaw is sufficient for max mapped address */
 	addr_width = agaw_to_width(iommu->agaw);
 	if (addr_width > cap_mgaw(iommu->cap))
Signature · 6 lines
2.25.1

_______________________________________________
iommu mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/iommu
author maintainer
#2 lore raw

PATCH v4 2/4 iommu/vt-d: Check domain force_snooping against attached devices

In reply to Lu Baolu

As domain->force_snooping only impacts the devices attached with the domain, there's no need to check against all IOMMU units. On the other hand, force_snooping could be set on a domain no matter whether it has been attached or not, and once set it is an immutable flag. If no device attached, the operation always succeeds. Then this empty domain can be only attached to a device of which the IOMMU supports snoop control.

---
 include/linux/intel-iommu.h |  1 +
 drivers/iommu/intel/pasid.h |  2 ++
 drivers/iommu/intel/iommu.c | 53 ++++++++++++++++++++++++++++++++++---
 drivers/iommu/intel/pasid.c | 42 +++++++++++++++++++++++++++++
 4 files changed, 95 insertions(+), 3 deletions(-)
include/linux/intel-iommu.h +1 -0
index 72e5d7900e71..4f29139bbfc3 100644
@@ -540,6 +540,7 @@ struct dmar_domain {
 	u8 has_iotlb_device: 1;
 	u8 iommu_coherency: 1;		/* indicate coherency of iommu access */
 	u8 force_snooping : 1;		/* Create IOPTEs with snoop control */
+	u8 set_pte_snp:1;
 
 	struct list_head devices;	/* all devices' list */
 	struct iova_domain iovad;	/* iova's that belong to this domain */
drivers/iommu/intel/pasid.h +2 -0
index ab4408c824a5..583ea67fc783 100644
@@ -123,4 +123,6 @@ void intel_pasid_tear_down_entry(struct intel_iommu *iommu,
 				 bool fault_ignore);
 int vcmd_alloc_pasid(struct intel_iommu *iommu, u32 *pasid);
 void vcmd_free_pasid(struct intel_iommu *iommu, u32 pasid);
+void intel_pasid_setup_page_snoop_control(struct intel_iommu *iommu,
+					  struct device *dev, u32 pasid);
 #endif /* __INTEL_PASID_H */
drivers/iommu/intel/iommu.c +50 -3

77 lines not on this page · raw message

drivers/iommu/intel/pasid.c +42 -0

45 lines not on this page · raw message

Signature · 6 lines
2.25.1

_______________________________________________
iommu mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/iommu
author maintainer
#5 lore raw

PATCH v4 3/4 iommu/vt-d: Remove domain_update_iommu_snooping()

In reply to Lu Baolu

The IOMMU force snooping capability is not required to be consistent among all the IOMMUs anymore. Remove force snooping capability check in the IOMMU hot-add path and domain_update_iommu_snooping() becomes a dead code now.

---
 drivers/iommu/intel/iommu.c | 34 +---------------------------------
 1 file changed, 1 insertion(+), 33 deletions(-)
drivers/iommu/intel/iommu.c +1 -33
index 048ebfbd5fcb..444d51a18c93 100644
@@ -533,33 +533,6 @@ static void domain_update_iommu_coherency(struct dmar_domain *domain)
 	rcu_read_unlock();
 }
 
-static bool domain_update_iommu_snooping(struct intel_iommu *skip)
-{
-	struct dmar_drhd_unit *drhd;
-	struct intel_iommu *iommu;
-	bool ret = true;
-
-	rcu_read_lock();
-	for_each_active_iommu(iommu, drhd) {
-		if (iommu != skip) {
-			/*
-			 * If the hardware is operating in the scalable mode,
-			 * the snooping control is always supported since we
-			 * always set PASID-table-entry.PGSNP bit if the domain
-			 * is managed outside (UNMANAGED).
-			 */
-			if (!sm_supported(iommu) &&
-			    !ecap_sc_support(iommu->ecap)) {
-				ret = false;
-				break;
-			}
-		}
-	}
-	rcu_read_unlock();
-
-	return ret;
-}
-
 static int domain_update_iommu_superpage(struct dmar_domain *domain,
 					 struct intel_iommu *skip)
 {
@@ -3606,12 +3579,7 @@ static int intel_iommu_add(struct dmar_drhd_unit *dmaru)
 			iommu->name);
 		return -ENXIO;
 	}
-	if (!ecap_sc_support(iommu->ecap) &&
-	    domain_update_iommu_snooping(iommu)) {
-		pr_warn("%s: Doesn't support snooping.\n",
-			iommu->name);
-		return -ENXIO;
-	}
+
 	sp = domain_update_iommu_superpage(NULL, iommu) - 1;
 	if (sp >= 0 && !(cap_super_page_val(iommu->cap) & (1 << sp))) {
 		pr_warn("%s: Doesn't support large page.\n",
Signature · 6 lines
2.25.1

_______________________________________________
iommu mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/iommu
author maintainer
#6 lore raw

PATCH v4 4/4 iommu/vt-d: Remove hard coding PGSNP bit in PASID entries

In reply to Lu Baolu

As enforce_cache_coherency has been introduced into the iommu_domain_ops, the kernel component which owns the iommu domain is able to opt-in its requirement for force snooping support. The iommu driver has no need to hard code the page snoop control bit in the PASID table entries anymore.

---
 drivers/iommu/intel/pasid.c | 3 ---
 1 file changed, 3 deletions(-)
drivers/iommu/intel/pasid.c +0 -3
index d19dd66a670c..cb4c1d0cf25c 100644
@@ -710,9 +710,6 @@ int intel_pasid_setup_second_level(struct intel_iommu *iommu,
 	pasid_set_fault_enable(pte);
 	pasid_set_page_snoop(pte, !!ecap_smpwc(iommu->ecap));
 
-	if (domain->domain.type == IOMMU_DOMAIN_UNMANAGED)
-		pasid_set_pgsnp(pte);
-
 	/*
 	 * Since it is a second level only translation setup, we should
 	 * set SRE bit as well (addresses are expected to be GPAs).
Signature · 6 lines
2.25.1

_______________________________________________
iommu mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/iommu
Reviewed-by on v4 2/4
#3 lore raw

RE: [PATCH v4 2/4] iommu/vt-d: Check domain force_snooping against attached devices

In reply to Lu Baolu
Quoted 11 lines

From: Lu Baolu <[email protected]>
Sent: Sunday, May 8, 2022 8:35 PM

As domain->force_snooping only impacts the devices attached with the domain, there's no need to check against all IOMMU units. On the other hand, force_snooping could be set on a domain no matter whether it has been attached or not, and once set it is an immutable flag. If no device attached, the operation always succeeds. Then this empty domain can be only attached to a device of which the IOMMU supports snoop control.

Quoted 1 line
Quoted 176 lines
---
 include/linux/intel-iommu.h |  1 +
 drivers/iommu/intel/pasid.h |  2 ++
 drivers/iommu/intel/iommu.c | 53
++++++++++++++++++++++++++++++++++---
 drivers/iommu/intel/pasid.c | 42 +++++++++++++++++++++++++++++
 4 files changed, 95 insertions(+), 3 deletions(-)
include/linux/intel-iommu.h +0 -0
index 72e5d7900e71..4f29139bbfc3 100644
@@ -540,6 +540,7 @@ struct dmar_domain {
 	u8 has_iotlb_device: 1;
 	u8 iommu_coherency: 1;		/* indicate coherency of

iommu access */
u8 force_snooping : 1; /* Create IOPTEs with snoop control
*/

include/linux/intel-iommu.h +1 -0
+	u8 set_pte_snp:1;
 
 	struct list_head devices;	/* all devices' list */
 	struct iova_domain iovad;	/* iova's that belong to this domain

*/

drivers/iommu/intel/pasid.h +2 -0
index ab4408c824a5..583ea67fc783 100644
@@ -123,4 +123,6 @@ void intel_pasid_tear_down_entry(struct
 intel_iommu *iommu,
 				 bool fault_ignore);
 int vcmd_alloc_pasid(struct intel_iommu *iommu, u32 *pasid);
 void vcmd_free_pasid(struct intel_iommu *iommu, u32 pasid);
+void intel_pasid_setup_page_snoop_control(struct intel_iommu *iommu,
+					  struct device *dev, u32 pasid);

#endif /* __INTEL_PASID_H */

drivers/iommu/intel/iommu.c +1 -1
index b4802f4055a0..048ebfbd5fcb 100644
@@ -2459,7 +2459,7 @@ static int domain_setup_first_level(struct
 intel_iommu *iommu,
 	if (level == 5)
 		flags |= PASID_FLAG_FL5LP;
 
-	if (domain->domain.type == IOMMU_DOMAIN_UNMANAGED)
+	if (domain->force_snooping)
 		flags |= PASID_FLAG_PAGE_SNOOP;
 
drivers/iommu/intel/iommu.c +1 -1
 	return intel_pasid_setup_first_level(iommu, dev, (pgd_t *)pgd, pasid,
@@ -4444,7 +4444,7 @@ static int intel_iommu_map(struct iommu_domain
 *domain,
 		prot |= DMA_PTE_READ;
 	if (iommu_prot & IOMMU_WRITE)
 		prot |= DMA_PTE_WRITE;
-	if (dmar_domain->force_snooping)
+	if (dmar_domain->set_pte_snp)
 		prot |= DMA_PTE_SNP;
 
drivers/iommu/intel/iommu.c +48 -1
 	max_addr = iova + size;
@@ -4567,13 +4567,60 @@ static phys_addr_t
 intel_iommu_iova_to_phys(struct iommu_domain *domain,
 	return phys;
 }
 
+static bool domain_support_force_snooping(struct dmar_domain *domain)
+{
+	struct device_domain_info *info;
+	bool support = true;
+
+	assert_spin_locked(&device_domain_lock);
+	list_for_each_entry(info, &domain->devices, link) {
+		if (!ecap_sc_support(info->iommu->ecap)) {
+			support = false;
+			break;
+		}
+	}
+
+	return support;
+}
+
+static void domain_set_force_snooping(struct dmar_domain *domain)
+{
+	struct device_domain_info *info;
+
+	assert_spin_locked(&device_domain_lock);
+
+	/*
+	 * Second level page table supports per-PTE snoop control. The
+	 * iommu_map() interface will handle this by setting SNP bit.
+	 */
+	if (!domain_use_first_level(domain)) {
+		domain->set_pte_snp = true;
+		return;
+	}
+
+	list_for_each_entry(info, &domain->devices, link)
+		intel_pasid_setup_page_snoop_control(info->iommu, info-
 >dev,
+						     PASID_RID2PASID);
+}
+
 static bool intel_iommu_enforce_cache_coherency(struct iommu_domain
 *domain)
 {
 	struct dmar_domain *dmar_domain = to_dmar_domain(domain);
+	unsigned long flags;
 
-	if (!domain_update_iommu_snooping(NULL))
+	if (dmar_domain->force_snooping)
+		return true;
+
+	spin_lock_irqsave(&device_domain_lock, flags);
+	if (!domain_support_force_snooping(dmar_domain)) {
+		spin_unlock_irqrestore(&device_domain_lock, flags);
 		return false;
+	}
+
+	domain_set_force_snooping(dmar_domain);
 	dmar_domain->force_snooping = true;
+	spin_unlock_irqrestore(&device_domain_lock, flags);
+
drivers/iommu/intel/iommu.c +0 -0
 	return true;
 }
drivers/iommu/intel/pasid.c +42 -1
index f8d215d85695..d19dd66a670c 100644
@@ -762,3 +762,45 @@ int intel_pasid_setup_pass_through(struct
 intel_iommu *iommu,
 
 	return 0;
 }
+
+/*
+ * Set the page snoop control for a pasid entry which has been set up.
+ */
+void intel_pasid_setup_page_snoop_control(struct intel_iommu *iommu,
+					  struct device *dev, u32 pasid)
+{
+	struct pasid_entry *pte;
+	u16 did;
+
+	spin_lock(&iommu->lock);
+	pte = intel_pasid_get_entry(dev, pasid);
+	if (WARN_ON(!pte || !pasid_pte_is_present(pte))) {
+		spin_unlock(&iommu->lock);
+		return;
+	}
+
+	pasid_set_pgsnp(pte);
+	did = pasid_get_domain_id(pte);
+	spin_unlock(&iommu->lock);
+
+	if (!ecap_coherent(iommu->ecap))
+		clflush_cache_range(pte, sizeof(*pte));
+
+	/*
+	 * VT-d spec 3.4 table23 states guides for cache invalidation:
+	 *
+	 * - PASID-selective-within-Domain PASID-cache invalidation
+	 * - PASID-selective PASID-based IOTLB invalidation
+	 * - If (pasid is RID_PASID)
+	 *    - Global Device-TLB invalidation to affected functions
+	 *   Else
+	 *    - PASID-based Device-TLB invalidation (with S=1 and
+	 *      Addr[63:12]=0x7FFFFFFF_FFFFF) to affected functions
+	 */
+	pasid_cache_invalidation_with_pasid(iommu, did, pasid);
+	qi_flush_piotlb(iommu, did, pasid, 0, -1, 0);
+
+	/* Device IOTLB doesn't need to be flushed in caching mode. */
+	if (!cap_caching_mode(iommu->cap))
+		devtlb_invalidation_with_pasid(iommu, dev, pasid);
+}
--
Quoted 1 line

2.25.1

_______________________________________________
iommu mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/iommu

author maintainer
#4 lore raw

Re: [PATCH v4 2/4] iommu/vt-d: Check domain force_snooping against attached devices

Quoted 11 lines · Kevin
Quoted 11 lines

From: Lu Baolu <[email protected]>
Sent: Sunday, May 8, 2022 8:35 PM

As domain->force_snooping only impacts the devices attached with the domain, there's no need to check against all IOMMU units. On the other hand, force_snooping could be set on a domain no matter whether it has been attached or not, and once set it is an immutable flag. If no device attached, the operation always succeeds. Then this empty domain can be only attached to a device of which the IOMMU supports snoop control.

Quoted 3 lines
Quoted 1 line

Thank you, Kevin. I will queue this series for v5.19.

Best regards,
baolu

Quoted 170 lines
Quoted 170 lines
---
  include/linux/intel-iommu.h |  1 +
  drivers/iommu/intel/pasid.h |  2 ++
  drivers/iommu/intel/iommu.c | 53
++++++++++++++++++++++++++++++++++---
  drivers/iommu/intel/pasid.c | 42 +++++++++++++++++++++++++++++
  4 files changed, 95 insertions(+), 3 deletions(-)
include/linux/intel-iommu.h +0 -0
index 72e5d7900e71..4f29139bbfc3 100644
@@ -540,6 +540,7 @@ struct dmar_domain {
  	u8 has_iotlb_device: 1;
  	u8 iommu_coherency: 1;		/* indicate coherency of

iommu access */
u8 force_snooping : 1; /* Create IOPTEs with snoop control
*/

include/linux/intel-iommu.h +1 -0
+	u8 set_pte_snp:1;
 
  	struct list_head devices;	/* all devices' list */
  	struct iova_domain iovad;	/* iova's that belong to this domain

*/

drivers/iommu/intel/pasid.h +2 -0
index ab4408c824a5..583ea67fc783 100644
@@ -123,4 +123,6 @@ void intel_pasid_tear_down_entry(struct
 intel_iommu *iommu,
  				 bool fault_ignore);
  int vcmd_alloc_pasid(struct intel_iommu *iommu, u32 *pasid);
  void vcmd_free_pasid(struct intel_iommu *iommu, u32 pasid);
+void intel_pasid_setup_page_snoop_control(struct intel_iommu *iommu,
+					  struct device *dev, u32 pasid);
  #endif /* __INTEL_PASID_H */
drivers/iommu/intel/iommu.c +1 -1
index b4802f4055a0..048ebfbd5fcb 100644
@@ -2459,7 +2459,7 @@ static int domain_setup_first_level(struct
 intel_iommu *iommu,
  	if (level == 5)
  		flags |= PASID_FLAG_FL5LP;
 
-	if (domain->domain.type == IOMMU_DOMAIN_UNMANAGED)
+	if (domain->force_snooping)
  		flags |= PASID_FLAG_PAGE_SNOOP;
 

return intel_pasid_setup_first_level(iommu, dev, (pgd_t *)pgd, pasid,

drivers/iommu/intel/iommu.c +1 -1
@@ -4444,7 +4444,7 @@ static int intel_iommu_map(struct iommu_domain
 *domain,
  		prot |= DMA_PTE_READ;
  	if (iommu_prot & IOMMU_WRITE)
  		prot |= DMA_PTE_WRITE;
-	if (dmar_domain->force_snooping)
+	if (dmar_domain->set_pte_snp)
  		prot |= DMA_PTE_SNP;
 
  	max_addr = iova + size;
drivers/iommu/intel/iommu.c +48 -1
@@ -4567,13 +4567,60 @@ static phys_addr_t
 intel_iommu_iova_to_phys(struct iommu_domain *domain,
  	return phys;
  }
 
+static bool domain_support_force_snooping(struct dmar_domain *domain)
+{
+	struct device_domain_info *info;
+	bool support = true;
+
+	assert_spin_locked(&device_domain_lock);
+	list_for_each_entry(info, &domain->devices, link) {
+		if (!ecap_sc_support(info->iommu->ecap)) {
+			support = false;
+			break;
+		}
+	}
+
+	return support;
+}
+
+static void domain_set_force_snooping(struct dmar_domain *domain)
+{
+	struct device_domain_info *info;
+
+	assert_spin_locked(&device_domain_lock);
+
+	/*
+	 * Second level page table supports per-PTE snoop control. The
+	 * iommu_map() interface will handle this by setting SNP bit.
+	 */
+	if (!domain_use_first_level(domain)) {
+		domain->set_pte_snp = true;
+		return;
+	}
+
+	list_for_each_entry(info, &domain->devices, link)
+		intel_pasid_setup_page_snoop_control(info->iommu, info-
 > dev,
+						     PASID_RID2PASID);
+}
+
  static bool intel_iommu_enforce_cache_coherency(struct iommu_domain
 *domain)
  {
  	struct dmar_domain *dmar_domain = to_dmar_domain(domain);
+	unsigned long flags;
 
-	if (!domain_update_iommu_snooping(NULL))
+	if (dmar_domain->force_snooping)
+		return true;
+
+	spin_lock_irqsave(&device_domain_lock, flags);
+	if (!domain_support_force_snooping(dmar_domain)) {
+		spin_unlock_irqrestore(&device_domain_lock, flags);
  		return false;
+	}
+
+	domain_set_force_snooping(dmar_domain);
  	dmar_domain->force_snooping = true;
+	spin_unlock_irqrestore(&device_domain_lock, flags);
+
  	return true;
  }
drivers/iommu/intel/pasid.c +36 -0
index f8d215d85695..d19dd66a670c 100644
@@ -762,3 +762,45 @@ int intel_pasid_setup_pass_through(struct
 intel_iommu *iommu,
 
  	return 0;
  }
+
+/*
+ * Set the page snoop control for a pasid entry which has been set up.
+ */
+void intel_pasid_setup_page_snoop_control(struct intel_iommu *iommu,
+					  struct device *dev, u32 pasid)
+{
+	struct pasid_entry *pte;
+	u16 did;
+
+	spin_lock(&iommu->lock);
+	pte = intel_pasid_get_entry(dev, pasid);
+	if (WARN_ON(!pte || !pasid_pte_is_present(pte))) {
+		spin_unlock(&iommu->lock);
+		return;
+	}
+
+	pasid_set_pgsnp(pte);
+	did = pasid_get_domain_id(pte);
+	spin_unlock(&iommu->lock);
+
+	if (!ecap_coherent(iommu->ecap))
+		clflush_cache_range(pte, sizeof(*pte));
+
+	/*
+	 * VT-d spec 3.4 table23 states guides for cache invalidation:
+	 *
+	 * - PASID-selective-within-Domain PASID-cache invalidation
+	 * - PASID-selective PASID-based IOTLB invalidation
+	 * - If (pasid is RID_PASID)
+	 *    - Global Device-TLB invalidation to affected functions
+	 *   Else
+	 *    - PASID-based Device-TLB invalidation (with S=1 and
+	 *      Addr[63:12]=0x7FFFFFFF_FFFFF) to affected functions
+	 */
+	pasid_cache_invalidation_with_pasid(iommu, did, pasid);
Quoted 9 lines
Quoted 8 lines
drivers/iommu/intel/pasid.c +6 -1
+	qi_flush_piotlb(iommu, did, pasid, 0, -1, 0);
+
+	/* Device IOTLB doesn't need to be flushed in caching mode. */
+	if (!cap_caching_mode(iommu->cap))
+		devtlb_invalidation_with_pasid(iommu, dev, pasid);
+}
--

2.25.1

_______________________________________________
iommu mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/iommu