intel-iommu: Prevent devices with RMRRs from being placed into SI Domain

PATCH In review Tom Mingarelli 3 participants 3 messages Nov 20, 2012 IOMMU

Share & tools

Tom Mingarelli

author

2012-11-20 19:43 UTC #0 lore raw

PATCH v4 1/1 intel-iommu: Prevent devices with RMRRs from being placed into SI Domain

This patch is to prevent non-USB devices that have RMRRs associated with them from being placed into the SI Domain during init. This fixes the issue where the RMRR info for devices being placed in and out of the SI Domain gets lost.

Signed-off-by Thomas Mingarelli <[email protected]>

Tested-by Shuah Khan <[email protected]>

---
PATCH v1: https://lkml.org/lkml/2012/6/15/204
PATCH v2: https://lkml.org/lkml/2012/9/18/354
PATCH v3: https://lkml.org/lkml/2012/10/16/375

 drivers/iommu/intel-iommu.c |   31 +++++++++++++++++++++++++++++++
 1 files changed, 31 insertions(+), 0 deletions(-)

drivers/iommu/intel-iommu.c +31 -0

index d4a4cd4..8c064df 100644
@@ -2320,8 +2320,39 @@ static int domain_add_dev_info(struct dmar_domain *domain,
 	return 0;
 }
 
+static bool device_has_rmrr(struct pci_dev *dev)
+{
+	struct dmar_rmrr_unit *rmrr;
+	int i;
+
+	for_each_rmrr_units(rmrr) {
+		for (i = 0; i < rmrr->devices_cnt; i++) {
+			/*
+			 * Return TRUE if this RMRR contains the device that
+			 * is passed in.
+			 */
+			if (rmrr->devices[i] == dev)
+				return true;
+		}
+	}
+	return false;
+}
+
 static int iommu_should_identity_map(struct pci_dev *pdev, int startup)
 {
+
+	/*
+	 * We want to prevent any device associated with an RMRR from
+	 * getting placed into the SI Domain. This is done because
+	 * problems exist when devices are moved in and out of domains
+	 * and their respective RMRR info is lost. We exempt USB devices
+	 * from this process due to their usage of RMRRs that are known
+	 * to not be needed after BIOS hand-off to OS.
+	 */
+	if (device_has_rmrr(pdev) &&
+	    (pdev->class >> 8) != PCI_CLASS_SERIAL_USB)
+		return 0;
+
 	if ((iommu_identity_mapping & IDENTMAP_AZALIA) && IS_AZALIA(pdev))
 		return 1;

Signature · 1 line

1.7.1

Alex Williamson

Reviewed-by on v4 1/1

2012-11-20 20:19 UTC #1 lore raw

Quoted 6 lines · Tom Mingarelli

This patch is to prevent non-USB devices that have RMRRs associated with them from being placed into the SI Domain during init. This fixes the issue where the RMRR info for devices being placed in and out of the SI Domain gets lost.
Signed-off-by Thomas Mingarelli <[email protected]>
Tested-by Shuah Khan <[email protected]>

Quoted 7 lines

---
PATCH v1: https://lkml.org/lkml/2012/6/15/204
PATCH v2: https://lkml.org/lkml/2012/9/18/354
PATCH v3: https://lkml.org/lkml/2012/10/16/375
 drivers/iommu/intel-iommu.c |   31 +++++++++++++++++++++++++++++++
 1 files changed, 31 insertions(+), 0 deletions(-)

Looks ok to me.

Reviewed-by Alex Williamson <[email protected]>

Quoted 45 lines

drivers/iommu/intel-iommu.c +31 -0

index d4a4cd4..8c064df 100644
@@ -2320,8 +2320,39 @@ static int domain_add_dev_info(struct dmar_domain *domain,
 	return 0;
 }
 
+static bool device_has_rmrr(struct pci_dev *dev)
+{
+	struct dmar_rmrr_unit *rmrr;
+	int i;
+
+	for_each_rmrr_units(rmrr) {
+		for (i = 0; i < rmrr->devices_cnt; i++) {
+			/*
+			 * Return TRUE if this RMRR contains the device that
+			 * is passed in.
+			 */
+			if (rmrr->devices[i] == dev)
+				return true;
+		}
+	}
+	return false;
+}
+
 static int iommu_should_identity_map(struct pci_dev *pdev, int startup)
 {
+
+	/*
+	 * We want to prevent any device associated with an RMRR from
+	 * getting placed into the SI Domain. This is done because
+	 * problems exist when devices are moved in and out of domains
+	 * and their respective RMRR info is lost. We exempt USB devices
+	 * from this process due to their usage of RMRRs that are known
+	 * to not be needed after BIOS hand-off to OS.
+	 */
+	if (device_has_rmrr(pdev) &&
+	    (pdev->class >> 8) != PCI_CLASS_SERIAL_USB)
+		return 0;
+
 	if ((iommu_identity_mapping & IDENTMAP_AZALIA) && IS_AZALIA(pdev))
 		return 1;

Don Dutile

Reviewed-by Donald Dutile on v4 1/1

2012-11-20 20:37 UTC #2 lore raw

In reply to Tom Mingarelli

Quoted 6 lines · Tom Mingarelli

This patch is to prevent non-USB devices that have RMRRs associated with them from being placed into the SI Domain during init. This fixes the issue where the RMRR info for devices being placed in and out of the SI Domain gets lost.
Signed-off-by Thomas Mingarelli<[email protected]>
Tested-by Shuah Khan<[email protected]>

Quoted 8 lines

---
PATCH v1: https://lkml.org/lkml/2012/6/15/204
PATCH v2: https://lkml.org/lkml/2012/9/18/354
PATCH v3: https://lkml.org/lkml/2012/10/16/375
  drivers/iommu/intel-iommu.c |   31 +++++++++++++++++++++++++++++++
  1 files changed, 31 insertions(+), 0 deletions(-)

Thanks for the efforts in getting this completed.
Looks good to me.

Reviewed-by Donald Dutile <[email protected]>

Quoted 44 lines

drivers/iommu/intel-iommu.c +31 -0

index d4a4cd4..8c064df 100644
@@ -2320,8 +2320,39 @@ static int domain_add_dev_info(struct dmar_domain *domain,
  	return 0;
  }
 
+static bool device_has_rmrr(struct pci_dev *dev)
+{
+	struct dmar_rmrr_unit *rmrr;
+	int i;
+
+	for_each_rmrr_units(rmrr) {
+		for (i = 0; i<  rmrr->devices_cnt; i++) {
+			/*
+			 * Return TRUE if this RMRR contains the device that
+			 * is passed in.
+			 */
+			if (rmrr->devices[i] == dev)
+				return true;
+		}
+	}
+	return false;
+}
+
  static int iommu_should_identity_map(struct pci_dev *pdev, int startup)
  {
+
+	/*
+	 * We want to prevent any device associated with an RMRR from
+	 * getting placed into the SI Domain. This is done because
+	 * problems exist when devices are moved in and out of domains
+	 * and their respective RMRR info is lost. We exempt USB devices
+	 * from this process due to their usage of RMRRs that are known
+	 * to not be needed after BIOS hand-off to OS.
+	 */
+	if (device_has_rmrr(pdev)&&
+	    (pdev->class>>  8) != PCI_CLASS_SERIAL_USB)
+		return 0;
+
  	if ((iommu_identity_mapping&  IDENTMAP_AZALIA)&&  IS_AZALIA(pdev))
  		return 1;