VMD MSI Remapping Bypass
Subsystems (5)
v2 cover VMD MSI Remapping Bypass
The Intel Volume Management Device acts similar to a PCI-to-PCI bridge in that it changes downstream devices' requester-ids to its own. As VMD supports PCIe devices, it has its own MSI/X table and transmits child device MSI/X by remapping child device MSI/X and handling like a demultiplexer.
Some newer VMD devices (Icelake Server) have an option to bypass the VMD MSI/X remapping table. This allows for better performance scaling as the child device MSI/X won't be limited by VMD's MSI/X count and IRQ handler.
V1->V2:
Updated for 5.12-next
Moved IRQ allocation and remapping enable/disable to a more logical location
V1 patches 1-4 were already merged
V1, 5/6: https://patchwork.kernel.org/project/linux-pci/patch/[email protected]/
V1, 6/6: https://patchwork.kernel.org/project/linux-pci/patch/[email protected]/
Jon Derrick (2):
iommu/vt-d: Use Real PCI DMA device for IRTE
PCI: vmd: Disable MSI/X remapping when possible
drivers/iommu/intel/irq_remapping.c | 3 +- drivers/pci/controller/vmd.c | 60 +++++++++++++++++++++++------ 2 files changed, 50 insertions(+), 13 deletions(-)
Signature · 6 lines
2.27.0 _______________________________________________ iommu mailing list [email protected] https://lists.linuxfoundation.org/mailman/listinfo/iommu
PATCH v2 1/2 iommu/vt-d: Use Real PCI DMA device for IRTE
VMD retransmits child device MSI/X with the VMD endpoint's requester-id. In order to support direct interrupt remapping of VMD child devices, ensure that the IRTE is programmed with the VMD endpoint's requester-id using pci_real_dma_dev().
--- drivers/iommu/intel/irq_remapping.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/iommu/intel/irq_remapping.c +2 -1
@@ -1276,7 +1276,8 @@ static void intel_irq_remapping_prepare_irte(struct intel_ir_data *data,break ;case X86_IRQ_ALLOC_TYPE_PCI_MSI :case X86_IRQ_ALLOC_TYPE_PCI_MSIX : -set_msi_sid (irte,msi_desc_to_pci_dev (info->desc)); +set_msi_sid (irte, +pci_real_dma_dev (msi_desc_to_pci_dev (info->desc)));break ;default :BUG_ON (1 );
Signature · 6 lines
2.27.0 _______________________________________________ iommu mailing list [email protected] https://lists.linuxfoundation.org/mailman/listinfo/iommu
Re: [PATCH v2 1/2] iommu/vt-d: Use Real PCI DMA device for IRTE
Quoted 16 lines · Jon Derrick
VMD retransmits child device MSI/X with the VMD endpoint's requester-id. In order to support direct interrupt remapping of VMD child devices, ensure that the IRTE is programmed with the VMD endpoint's requester-id using
pci_real_dma_dev().Signed-off-by Jon Derrick <[email protected]>--- drivers/iommu/intel/irq_remapping.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)drivers/iommu/intel/irq_remapping.c +0
-0
drivers/iommu/intel/irq_remapping.c static void intel_irq_remapping_prepare_irte(struct intel_ir_data *data, · line 1283
Quoted 9 lines
drivers/iommu/intel/irq_remapping.c +2
-1@@ -1276,7 +1276,8 @@ static void intel_irq_remapping_prepare_irte(struct intel_ir_data *data,break ;case X86_IRQ_ALLOC_TYPE_PCI_MSI :case X86_IRQ_ALLOC_TYPE_PCI_MSIX : -set_msi_sid (irte,msi_desc_to_pci_dev (info->desc)); +set_msi_sid (irte, +pci_real_dma_dev (msi_desc_to_pci_dev (info->desc)));break ;default :BUG_ON (1 );
Best regards,
baolu
_______________________________________________
iommu mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/iommu
PATCH v2 2/2 PCI: vmd: Disable MSI/X remapping when possible
In reply to Jon DerrickVMD will retransmit child device MSI/X using its own MSI/X table and requester-id. This limits the number of MSI/X available to the whole child device domain to the number of VMD MSI/X interrupts.
Some VMD devices have a mode where this remapping can be disabled, allowing child device interrupts to bypass processing with the VMD MSI/X domain interrupt handler and going straight the child device interrupt handler, allowing for better performance and scaling. The requester-id still gets changed to the VMD endpoint's requester-id, and the interrupt remapping handlers have been updated to properly set IRTE for child device interrupts to the VMD endpoint's context.
Some VMD platforms have existing production BIOS which rely on MSI/X remapping and won't explicitly program the MSI/X remapping bit. This re-enables MSI/X remapping on unload.
--- drivers/pci/controller/vmd.c | 60 ++++++++++++++++++++++++++++-------- 1 file changed, 48 insertions(+), 12 deletions(-)
drivers/pci/controller/vmd.c +48 -12
@@ -59,6 +59,13 @@ enum vmd_features {* be used for MSI remapping */ VMD_FEAT_OFFSET_FIRST_VECTOR = (1 <<3 ), + +/* +* Device can bypass remapping MSI/X transactions into its MSI/X table, +* avoding the requirement of a VMD MSI domain for child device +* interrupt handling +*/ + VMD_FEAT_BYPASS_MSI_REMAP = (1 << 4), };/*
@@ -306,6 +313,15 @@ static struct msi_domain_info vmd_msi_domain_info = { .chip = &vmd_msi_controller, }; +static void vmd_enable_msi_remapping (struct vmd_dev *vmd,bool enable) +{ +u16 reg; + +pci_read_config_word (vmd->dev,PCI_REG_VMCONFIG , ®); + reg = enable ? (reg & ~0x2 ) : (reg |0x2 ); +pci_write_config_word (vmd->dev,PCI_REG_VMCONFIG , reg); +} +static int vmd_create_irq_domain (struct vmd_dev *vmd) {struct fwnode_handle *fn;
@@ -325,6 +341,13 @@ static int vmd_create_irq_domain(struct vmd_dev *vmd)static void vmd_remove_irq_domain (struct vmd_dev *vmd) { +/* +* Some production BIOS won't enable remapping between soft reboots. +* Ensure remapping is restored before unloading the driver. +*/ +if (!vmd->msix_count) +vmd_enable_msi_remapping (vmd, true); +if (vmd->irq_domain) {struct fwnode_handle *fn = vmd->irq_domain->fwnode;
@@ -679,15 +702,31 @@ static int vmd_enable_domain(struct vmd_dev *vmd, unsigned long features) sd->node =pcibus_to_node (vmd->dev->bus); - ret =vmd_create_irq_domain (vmd); -if (ret) -return ret; -/* -* Override the irq domain bus token so the domain can be distinguished -* from a regular PCI/MSI domain. +* Currently MSI remapping must be enabled in guest passthrough mode +* due to some missing interrupt remapping plumbing. This is probably +* acceptable because the guest is usually CPU-limited and MSI +* remapping doesn't become a performance bottleneck. */ -irq_domain_update_bus_token (vmd->irq_domain,DOMAIN_BUS_VMD_MSI ); +if (!(features &VMD_FEAT_BYPASS_MSI_REMAP ) || offset[0 ] || offset[1 ]) { + ret =vmd_alloc_irqs (vmd); +if (ret) +return ret; + +vmd_enable_msi_remapping (vmd, true); + + ret =vmd_create_irq_domain (vmd); +if (ret) +return ret; + +/* +* Override the irq domain bus token so the domain can be +* distinguished from a regular PCI/MSI domain. +*/ +irq_domain_update_bus_token (vmd->irq_domain,DOMAIN_BUS_VMD_MSI ); + }else { +vmd_enable_msi_remapping (vmd, false); + }pci_add_resource (&resources, &vmd->resources[0 ]);pci_add_resource_offset (&resources, &vmd->resources[1 ], offset[0 ]);
@@ -753,10 +792,6 @@ static int vmd_probe(struct pci_dev *dev, const struct pci_device_id *id)if (features &VMD_FEAT_OFFSET_FIRST_VECTOR ) vmd->first_vec =1 ; - err =vmd_alloc_irqs (vmd); -if (err) -return err; -spin_lock_init (&vmd->cfg_lock);pci_set_drvdata (dev, vmd); err =vmd_enable_domain (vmd, features);
@@ -825,7 +860,8 @@ static const struct pci_device_id vmd_ids[] = { .driver_data = VMD_FEAT_HAS_MEMBAR_SHADOW_VSCAP,}, {PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_VMD_28C0), .driver_data =VMD_FEAT_HAS_MEMBAR_SHADOW | - VMD_FEAT_HAS_BUS_RESTRICTIONS,}, +VMD_FEAT_HAS_BUS_RESTRICTIONS | + VMD_FEAT_BYPASS_MSI_REMAP,}, {PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x467f), .driver_data = VMD_FEAT_HAS_MEMBAR_SHADOW_VSCAP | VMD_FEAT_HAS_BUS_RESTRICTIONS |
Signature · 6 lines
2.27.0 _______________________________________________ iommu mailing list [email protected] https://lists.linuxfoundation.org/mailman/listinfo/iommu
Re: [PATCH v2 2/2] PCI: vmd: Disable MSI/X remapping when possible
Quoted 12 lines · Jon Derrick
VMD will retransmit child device MSI/X using its own MSI/X table and requester-id. This limits the number of MSI/X available to the whole child device domain to the number of VMD MSI/X interrupts.
Some VMD devices have a mode where this remapping can be disabled, allowing child device interrupts to bypass processing with the VMD MSI/X domain interrupt handler and going straight the child device interrupt handler, allowing for better performance and scaling. The requester-id still gets changed to the VMD endpoint's requester-id, and the interrupt remapping handlers have been updated to properly set IRTE for child device interrupts to the VMD endpoint's context.
Quoted 3 lines
Some VMD platforms have existing production BIOS which rely on MSI/X remapping and won't explicitly program the MSI/X remapping bit. This re-enables MSI/X remapping on unload.
Trivial comments below. Would you mind using "MSI-X" instead of "MSI/X" so it matches the usage in the PCIe specs? Several mentions above (including subject) and below.
Quoted 10 lines
Signed-off-by Jon Derrick <[email protected]>--- drivers/pci/controller/vmd.c | 60 ++++++++++++++++++++++++++++-------- 1 file changed, 48 insertions(+), 12 deletions(-)drivers/pci/controller/vmd.c +0
-0
drivers/pci/controller/vmd.c enum vmd_features { · line 65
Quoted 7 lines
drivers/pci/controller/vmd.c +4
-0@@ -59,6 +59,13 @@ enum vmd_features {* be used for MSI remapping */ VMD_FEAT_OFFSET_FIRST_VECTOR = (1 << 3), + +/* +* Device can bypass remapping MSI/X transactions into its MSI/X table, +* avoding the requirement of a VMD MSI domain for child device
s/avoding/avoiding/
drivers/pci/controller/vmd.c enum vmd_features { · line 66
Quoted 1 line
drivers/pci/controller/vmd.c +1
-0+* interrupt handling
Maybe a period at the end of the sentence.
Quoted 6 lines
drivers/pci/controller/vmd.c +2
-0+*/ + VMD_FEAT_BYPASS_MSI_REMAP = (1 << 4), };/*
drivers/pci/controller/vmd.c static struct msi_domain_info vmd_msi_domain_info = { · line 321
Quoted 9 lines
drivers/pci/controller/vmd.c +6
-0@@ -306,6 +313,15 @@ static struct msi_domain_info vmd_msi_domain_info = { .chip = &vmd_msi_controller, }; +static void vmd_enable_msi_remapping (struct vmd_dev *vmd,bool enable) +{ +u16 reg; + +pci_read_config_word (vmd->dev,PCI_REG_VMCONFIG , ®); + reg = enable ? (reg & ~0x2 ) : (reg |0x2 );
Would be nice to have a #define for 0x2.
Quoted 83 lines
drivers/pci/controller/vmd.c +35
-12+pci_write_config_word (vmd->dev,PCI_REG_VMCONFIG , reg); +} +static int vmd_create_irq_domain (struct vmd_dev *vmd) {struct fwnode_handle *fn;@@ -325,6 +341,13 @@ static int vmd_create_irq_domain(struct vmd_dev *vmd)static void vmd_remove_irq_domain (struct vmd_dev *vmd) { +/* +* Some production BIOS won't enable remapping between soft reboots. +* Ensure remapping is restored before unloading the driver. +*/ +if (!vmd->msix_count) +vmd_enable_msi_remapping (vmd, true); +if (vmd->irq_domain) {struct fwnode_handle *fn = vmd->irq_domain->fwnode;@@ -679,15 +702,31 @@ static int vmd_enable_domain(struct vmd_dev *vmd, unsigned long features) sd->node =pcibus_to_node (vmd->dev->bus); - ret =vmd_create_irq_domain (vmd); -if (ret) -return ret; -/* -* Override the irq domain bus token so the domain can be distinguished -* from a regular PCI/MSI domain. +* Currently MSI remapping must be enabled in guest passthrough mode +* due to some missing interrupt remapping plumbing. This is probably +* acceptable because the guest is usually CPU-limited and MSI +* remapping doesn't become a performance bottleneck. */ -irq_domain_update_bus_token (vmd->irq_domain,DOMAIN_BUS_VMD_MSI ); +if (!(features &VMD_FEAT_BYPASS_MSI_REMAP ) || offset[0 ] || offset[1 ]) { + ret =vmd_alloc_irqs (vmd); +if (ret) +return ret; + +vmd_enable_msi_remapping (vmd, true); + + ret =vmd_create_irq_domain (vmd); +if (ret) +return ret; + +/* +* Override the irq domain bus token so the domain can be +* distinguished from a regular PCI/MSI domain. +*/ +irq_domain_update_bus_token (vmd->irq_domain,DOMAIN_BUS_VMD_MSI ); + }else { +vmd_enable_msi_remapping (vmd, false); + }pci_add_resource (&resources, &vmd->resources[0 ]);pci_add_resource_offset (&resources, &vmd->resources[1 ], offset[0 ]);@@ -753,10 +792,6 @@ static int vmd_probe(struct pci_dev *dev, const struct pci_device_id *id)if (features &VMD_FEAT_OFFSET_FIRST_VECTOR ) vmd->first_vec =1 ; - err =vmd_alloc_irqs (vmd); -if (err) -return err; -spin_lock_init (&vmd->cfg_lock);pci_set_drvdata (dev, vmd); err =vmd_enable_domain (vmd, features);@@ -825,7 +860,8 @@ static const struct pci_device_id vmd_ids[] = { .driver_data = VMD_FEAT_HAS_MEMBAR_SHADOW_VSCAP,}, {PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_VMD_28C0), .driver_data =VMD_FEAT_HAS_MEMBAR_SHADOW | - VMD_FEAT_HAS_BUS_RESTRICTIONS,}, +VMD_FEAT_HAS_BUS_RESTRICTIONS | + VMD_FEAT_BYPASS_MSI_REMAP,}, {PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x467f), .driver_data = VMD_FEAT_HAS_MEMBAR_SHADOW_VSCAP | VMD_FEAT_HAS_BUS_RESTRICTIONS |Signature · 1 line
2.27.0
_______________________________________________
iommu mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/iommu
Re: [PATCH v2 2/2] PCI: vmd: Disable MSI/X remapping when possible
Quoted 13 lines · Bjorn Helgaas
Quoted 12 lines · Jon Derrick
VMD will retransmit child device MSI/X using its own MSI/X table and requester-id. This limits the number of MSI/X available to the whole child device domain to the number of VMD MSI/X interrupts.
Some VMD devices have a mode where this remapping can be disabled, allowing child device interrupts to bypass processing with the VMD MSI/X domain interrupt handler and going straight the child device interrupt handler, allowing for better performance and scaling. The requester-id still gets changed to the VMD endpoint's requester-id, and the interrupt remapping handlers have been updated to properly set IRTE for child device interrupts to the VMD endpoint's context.
Quoted 7 lines
Quoted 3 lines
Some VMD platforms have existing production BIOS which rely on MSI/X remapping and won't explicitly program the MSI/X remapping bit. This re-enables MSI/X remapping on unload.
Trivial comments below. Would you mind using "MSI-X" instead of "MSI/X" so it matches the usage in the PCIe specs? Several mentions above (including subject) and below.
Thanks Bjorn, will do
Quoted 118 lines
Quoted 11 lines
Signed-off-by Jon Derrick <[email protected]>--- drivers/pci/controller/vmd.c | 60 ++++++++++++++++++++++++++++-------- 1 file changed, 48 insertions(+), 12 deletions(-)drivers/pci/controller/vmd.c +0
-0drivers/pci/controller/vmd.c enum vmd_features { · line 65
Quoted 7 lines
drivers/pci/controller/vmd.c +4
-0@@ -59,6 +59,13 @@ enum vmd_features {* be used for MSI remapping */ VMD_FEAT_OFFSET_FIRST_VECTOR = (1 << 3), + +/* +* Device can bypass remapping MSI/X transactions into its MSI/X table, +* avoding the requirement of a VMD MSI domain for child device s/avoding/avoiding/
drivers/pci/controller/vmd.c enum vmd_features { · line 66
Quoted 1 line
drivers/pci/controller/vmd.c +1
-0+* interrupt handling Maybe a period at the end of the sentence.
Quoted 6 lines
drivers/pci/controller/vmd.c +2
-0+*/ + VMD_FEAT_BYPASS_MSI_REMAP = (1 << 4), };/* drivers/pci/controller/vmd.c static struct msi_domain_info vmd_msi_domain_info = { · line 321
Quoted 9 lines
drivers/pci/controller/vmd.c +6
-0@@ -306,6 +313,15 @@ static struct msi_domain_info vmd_msi_domain_info = { .chip = &vmd_msi_controller, }; +static void vmd_enable_msi_remapping (struct vmd_dev *vmd,bool enable) +{ +u16 reg; + +pci_read_config_word (vmd->dev,PCI_REG_VMCONFIG , ®); + reg = enable ? (reg & ~0x2 ) : (reg |0x2 );Would be nice to have a #define for
0x2.Quoted 75 lines
drivers/pci/controller/vmd.c +33
-12+pci_write_config_word (vmd->dev,PCI_REG_VMCONFIG , reg); +} +static int vmd_create_irq_domain (struct vmd_dev *vmd) {struct fwnode_handle *fn;@@ -325,6 +341,13 @@ static int vmd_create_irq_domain(struct vmd_dev *vmd)static void vmd_remove_irq_domain (struct vmd_dev *vmd) { +/* +* Some production BIOS won't enable remapping between soft reboots. +* Ensure remapping is restored before unloading the driver. +*/ +if (!vmd->msix_count) +vmd_enable_msi_remapping (vmd, true); +if (vmd->irq_domain) {struct fwnode_handle *fn = vmd->irq_domain->fwnode;@@ -679,15 +702,31 @@ static int vmd_enable_domain(struct vmd_dev *vmd, unsigned long features) sd->node =pcibus_to_node (vmd->dev->bus); - ret =vmd_create_irq_domain (vmd); -if (ret) -return ret; -/* -* Override the irq domain bus token so the domain can be distinguished -* from a regular PCI/MSI domain. +* Currently MSI remapping must be enabled in guest passthrough mode +* due to some missing interrupt remapping plumbing. This is probably +* acceptable because the guest is usually CPU-limited and MSI +* remapping doesn't become a performance bottleneck. */ -irq_domain_update_bus_token (vmd->irq_domain,DOMAIN_BUS_VMD_MSI ); +if (!(features &VMD_FEAT_BYPASS_MSI_REMAP ) || offset[0 ] || offset[1 ]) { + ret =vmd_alloc_irqs (vmd); +if (ret) +return ret; + +vmd_enable_msi_remapping (vmd, true); + + ret =vmd_create_irq_domain (vmd); +if (ret) +return ret; + +/* +* Override the irq domain bus token so the domain can be +* distinguished from a regular PCI/MSI domain. +*/ +irq_domain_update_bus_token (vmd->irq_domain,DOMAIN_BUS_VMD_MSI ); + }else { +vmd_enable_msi_remapping (vmd, false); + }pci_add_resource (&resources, &vmd->resources[0 ]);pci_add_resource_offset (&resources, &vmd->resources[1 ], offset[0 ]);@@ -753,10 +792,6 @@ static int vmd_probe(struct pci_dev *dev, const struct pci_device_id *id)if (features &VMD_FEAT_OFFSET_FIRST_VECTOR ) vmd->first_vec =1 ; - err =vmd_alloc_irqs (vmd); -if (err) -return err; -spin_lock_init (&vmd->cfg_lock);pci_set_drvdata (dev, vmd); err =vmd_enable_domain (vmd, features);@@ -825,7 +860,8 @@ static const struct pci_device_id vmd_ids[] = { .driver_data = VMD_FEAT_HAS_MEMBAR_SHADOW_VSCAP,}, {PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_VMD_28C0), .driver_data = VMD_FEAT_HAS_MEMBAR_SHADOW | - VMD_FEAT_HAS_BUS_RESTRICTIONS,},
Quoted 8 lines
Quoted 8 lines
drivers/pci/controller/vmd.c +2
-0+VMD_FEAT_HAS_BUS_RESTRICTIONS | + VMD_FEAT_BYPASS_MSI_REMAP,}, {PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x467f), .driver_data = VMD_FEAT_HAS_MEMBAR_SHADOW_VSCAP | VMD_FEAT_HAS_BUS_RESTRICTIONS |Signature · 1 line
2.27.0
_______________________________________________
iommu mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/iommu
Quoted 7 lines · Jon Derrick
Jon Derrick (2):
iommu/vt-d: Use Real PCI DMA device for IRTE
PCI: vmd: Disable MSI/X remapping when possibledrivers/iommu/intel/irq_remapping.c | 3 +- drivers/pci/controller/vmd.c | 60 +++++++++++++++++++++++------ 2 files changed, 50 insertions(+), 13 deletions(-)
This probably goes via Bjorn's tree, so
_______________________________________________
iommu mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/iommu