Skip to contentSkip to conversation

PCI/VMD: Fix config addressing with bus offsets

PATCH In review Jon Derrick 1 participant 2 messages Jun 7, 2019 – Jun 10, 2019 1 message not in the archive

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 PCI/VMD: Fix config addressing with bus offsets

VMD config space addressing relies on mapping the BDF of the target into the VMD config bar. When using bus number offsets to number the VMD domain, the offset needs to be ignored in order to correctly map devices to their config space.

Fixes 2a5a9c9a20f9 ("PCI: vmd: Add offset to bus numbers if necessary")
Cc <[email protected]> # v4.18+
---
 drivers/pci/controller/vmd.c | 16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)
drivers/pci/controller/vmd.c +9 -7
index 999a5509e57e..6c80b9098dd0 100644
@@ -94,6 +94,7 @@ struct vmd_dev {
 	struct resource		resources[3];
 	struct irq_domain	*irq_domain;
 	struct pci_bus		*bus;
+	u8			busn_start;
 
 	struct dma_map_ops	dma_ops;
 	struct dma_domain	dma_domain;
@@ -440,7 +441,8 @@ static char __iomem *vmd_cfg_addr(struct vmd_dev *vmd, struct pci_bus *bus,
 				  unsigned int devfn, int reg, int len)
 {
 	char __iomem *addr = vmd->cfgbar +
-			     (bus->number << 20) + (devfn << 12) + reg;
+			     ((bus->number - vmd->busn_start) << 20) +
+			     (devfn << 12) + reg;
 
 	if ((addr - vmd->cfgbar) + len >=
 	    resource_size(&vmd->dev->resource[VMD_CFGBAR]))
@@ -563,7 +565,7 @@ static int vmd_enable_domain(struct vmd_dev *vmd, unsigned long features)
 	unsigned long flags;
 	LIST_HEAD(resources);
 	resource_size_t offset[2] = {0};
-	resource_size_t membar2_offset = 0x2000, busn_start = 0;
+	resource_size_t membar2_offset = 0x2000;
 	struct pci_bus *child;
 
 	/*
@@ -606,14 +608,14 @@ static int vmd_enable_domain(struct vmd_dev *vmd, unsigned long features)
 		pci_read_config_dword(vmd->dev, PCI_REG_VMCONFIG, &vmconfig);
 		if (BUS_RESTRICT_CAP(vmcap) &&
 		    (BUS_RESTRICT_CFG(vmconfig) == 0x1))
-			busn_start = 128;
+			vmd->busn_start = 128;
 	}
 
 	res = &vmd->dev->resource[VMD_CFGBAR];
 	vmd->resources[0] = (struct resource) {
 		.name  = "VMD CFGBAR",
-		.start = busn_start,
-		.end   = busn_start + (resource_size(res) >> 20) - 1,
+		.start = vmd->busn_start,
+		.end   = vmd->busn_start + (resource_size(res) >> 20) - 1,
 		.flags = IORESOURCE_BUS | IORESOURCE_PCI_FIXED,
 	};
 
@@ -681,8 +683,8 @@ static int vmd_enable_domain(struct vmd_dev *vmd, unsigned long features)
 	pci_add_resource_offset(&resources, &vmd->resources[1], offset[0]);
 	pci_add_resource_offset(&resources, &vmd->resources[2], offset[1]);
 
-	vmd->bus = pci_create_root_bus(&vmd->dev->dev, busn_start, &vmd_ops,
-				       sd, &resources);
+	vmd->bus = pci_create_root_bus(&vmd->dev->dev, vmd->busn_start,
+				       &vmd_ops, sd, &resources);
 	if (!vmd->bus) {
 		pci_free_resource_list(&resources);
 		irq_domain_remove(vmd->irq_domain);
Signature · 1 line
2.20.1

A message that is not in the archive (sent off-list or not yet received): [email protected]

#2 lore raw
Quoted 14 lines · Sasha Levin

Hi,

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag, fixing commit: 2a5a9c9a20f9 PCI: vmd: Add offset to bus numbers if necessary.

The bot has tested the following trees: v5.1.7, v4.19.48.

v5.1.7: Build OK!
v4.19.48: Failed to apply! Possible dependencies:
0294951030eb ("PCI/VMD: Configure MPS settings before adding devices")

Quoted 5 lines

How should we proceed with this patch?

--
Thanks,
Sasha

Thanks Sasha-bot,

I'll resend for kernels without 0294951030eb