Skip to contentSkip to conversation

iommu/vt-d: Dump DMAR translation structure

PATCH Superseded Kyung Min Park 2 participants 3 messages Jul 22, 2021 – Jul 30, 2021

Subsystems (2)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 iommu/vt-d: Dump DMAR translation structure

When the dmar translation fault happens, the kernel prints a single line fault reason with corresponding hexadecimal code defined in the Intel VT-d specification.

Currently, when user wants to debug the translation fault in detail, debugfs is used for dumping the dmar_translation_struct, which is not available when the kernel failed to boot.

Dump the DMAR translation structure, pagewalk the IO page table and print the page table entry when the fault happens.

---
ChangeLog:
- Change from v1 to v2:
1. fix compilation issue with different IOMMU config option.

---
 drivers/iommu/intel/dmar.c  |   8 ++-
 drivers/iommu/intel/iommu.c | 107 ++++++++++++++++++++++++++++++++++++
 include/linux/dmar.h        |   4 ++
 3 files changed, 116 insertions(+), 3 deletions(-)
drivers/iommu/intel/dmar.c +5 -3
index d66f79acd14d..c74d97c5ec21 100644
@@ -1948,19 +1948,21 @@ static int dmar_fault_do_one(struct intel_iommu *iommu, int type,
 		       source_id >> 8, PCI_SLOT(source_id & 0xFF),
 		       PCI_FUNC(source_id & 0xFF), addr >> 48,
 		       fault_reason, reason);
-	else if (pasid == INVALID_IOASID)
+	else if (pasid == INVALID_IOASID) {
 		pr_err("[%s NO_PASID] Request device [0x%02x:0x%02x.%d] fault addr 0x%llx [fault reason 0x%02x] %s\n",
 		       type ? "DMA Read" : "DMA Write",
 		       source_id >> 8, PCI_SLOT(source_id & 0xFF),
 		       PCI_FUNC(source_id & 0xFF), addr,
 		       fault_reason, reason);
-	else
+		dmar_fault_dump_ptes(iommu, source_id, addr, pasid);
+	} else {
 		pr_err("[%s PASID 0x%x] Request device [0x%02x:0x%02x.%d] fault addr 0x%llx [fault reason 0x%02x] %s\n",
 		       type ? "DMA Read" : "DMA Write", pasid,
 		       source_id >> 8, PCI_SLOT(source_id & 0xFF),
 		       PCI_FUNC(source_id & 0xFF), addr,
 		       fault_reason, reason);
-
+		dmar_fault_dump_ptes(iommu, source_id, addr, pasid);
+	}
 	return 0;
 }
 
drivers/iommu/intel/iommu.c +107 -0

119 lines not on this page · raw message

include/linux/dmar.h +4 -0
index e04436a7ff27..cdcb1b6d85b9 100644
@@ -142,9 +142,13 @@ extern int dmar_parse_one_satc(struct acpi_dmar_header *hdr, void *arg);
 extern int dmar_release_one_atsr(struct acpi_dmar_header *hdr, void *arg);
 extern int dmar_iommu_hotplug(struct dmar_drhd_unit *dmaru, bool insert);
 extern int dmar_iommu_notify_scope_dev(struct dmar_pci_notify_info *info);
+void dmar_fault_dump_ptes(struct intel_iommu *iommu, u16 source_id,
+			  unsigned long long addr, u32 pasid);
 #else /* !CONFIG_INTEL_IOMMU: */
 static inline int intel_iommu_init(void) { return -ENODEV; }
 static inline void intel_iommu_shutdown(void) { }
+static inline void dmar_fault_dump_ptes(struct intel_iommu *iommu, u16 source_id,
+					unsigned long long addr, u32 pasid) {}
 
 #define	dmar_parse_one_rmrr		dmar_res_noop
 #define	dmar_parse_one_atsr		dmar_res_noop
Signature · 1 line
2.17.1
maintainer
#1 lore raw
Quoted 4 lines · Kyung Min Park

When the dmar translation fault happens, the kernel prints a single line fault reason with corresponding hexadecimal code defined in the Intel VT-d specification.

Quoted 8 lines

Currently, when user wants to debug the translation fault in detail, debugfs is used for dumping the dmar_translation_struct, which is not available when the kernel failed to boot.

Dump the DMAR translation structure, pagewalk the IO page table and print the page table entry when the fault happens.

This is too much noise in dmesg for a single dmar fault. Please hide that additional debug information behind a debug kernel option for the VT-d driver.

Thanks,

Joerg

#2 lore raw

Hi Joerg,

Quoted 17 lines · Joerg Roedel
Quoted 16 lines · Kyung Min Park

When the dmar translation fault happens, the kernel prints a single line
fault reason with corresponding hexadecimal code defined in the Intel VT-d
specification.

Currently, when user wants to debug the translation fault in detail,
debugfs is used for dumping the dmar_translation_struct, which is not
available when the kernel failed to boot.

Dump the DMAR translation structure, pagewalk the IO page table and print
the page table entry when the fault happens.

Quoted 6 lines
Quoted 1 line

This is too much noise in dmesg for a single dmar fault. Please hide that additional debug information behind a debug kernel option for the
VT-d driver.

Sure, let me change in the next version.