Skip to contentSkip to conversation

iommu/vt-d: Fix lockdep splat due to klist iteration in atomic context

PATCH In review Lu Baolu 1 participant 2 messages Aug 19, 2022 – Aug 23, 2022

Subsystems (2)
lore
Share & tools
lore

Lu Baolu

author maintainer
#0 lore raw

PATCH 1/1 iommu/vt-d: Fix lockdep splat due to klist iteration in atomic context

With CONFIG_INTEL_IOMMU_DEBUGFS enabled, below lockdep splat are seen when an I/O fault occurs on a machine with an Intel IOMMU in it.

DMAR: DRHD: handling fault status reg 3
DMAR: [DMA Write NO_PASID] Request device [00:1a.0] fault addr 0x0

       [fault reason 0x05] PTE Write access is not set
 DMAR: Dump dmar0 table entries for IOVA 0x0
 DMAR: root entry: 0x0000000127f42001
 DMAR: context entry: hi 0x0000000000001502, low 0x000000012d8ab001
 ================================
 WARNING: inconsistent lock state
 5.20.0-0.rc0.20220812git7ebfc85e2cd7.10.fc38.x86_64 #1 Not tainted
 --------------------------------
 inconsistent {HARDIRQ-ON-W} -> {IN-HARDIRQ-W} usage.
 rngd/1006 [HC1[1]:SC0[0]:HE0:SE1] takes:
 ff177021416f2d78 (&k->k_lock){?.+.}-{2:2}, at: klist_next+0x1b/0x160
 {HARDIRQ-ON-W} state was registered at:
   lock_acquire+0xce/0x2d0
   _raw_spin_lock+0x33/0x80
   klist_add_tail+0x46/0x80
   bus_add_device+0xee/0x150
   device_add+0x39d/0x9a0
   add_memory_block+0x108/0x1d0
   memory_dev_init+0xe1/0x117
   driver_init+0x43/0x4d
   kernel_init_freeable+0x1c2/0x2cc
   kernel_init+0x16/0x140
   ret_from_fork+0x1f/0x30
 irq event stamp: 7812
 hardirqs last  enabled at (7811): [<ffffffff85000e86>] asm_sysvec_apic_timer_interrupt+0x16/0x20
 hardirqs last disabled at (7812): [<ffffffff84f16894>] irqentry_enter+0x54/0x60
 softirqs last  enabled at (7794): [<ffffffff840ff669>] __irq_exit_rcu+0xf9/0x170
 softirqs last disabled at (7787): [<ffffffff840ff669>] __irq_exit_rcu+0xf9/0x170

The klist iterator functions using spin_*lock_irq*() but the klist insertion functions using spin_*lock(), combined with the Intel DMAR IOMMU driver iterating over klists from atomic (hardirq) context, where pci_get_domain_bus_and_slot() calls into bus_find_device() which iterates over klists.

As currently there's no plan to fix the klist to make it safe to use in atomic context, this fixes the lockdep splat by avoid calling pci_get_domain_bus_and_slot() in the hardirq context.

---
 drivers/iommu/intel/iommu.c | 47 +++++++++++++++----------------------
 1 file changed, 19 insertions(+), 28 deletions(-)
drivers/iommu/intel/iommu.c +19 -28
index f12046f4b9aa..795529feac01 100644
@@ -794,32 +794,11 @@ static void free_context_table(struct intel_iommu *iommu)
 }
 
 #ifdef CONFIG_DMAR_DEBUG
-static void pgtable_walk(struct intel_iommu *iommu, unsigned long pfn, u8 bus, u8 devfn)
+static void pgtable_walk(struct intel_iommu *iommu, unsigned long pfn,
+			 u8 bus, u8 devfn, struct dma_pte *parent, int level)
 {
-	struct device_domain_info *info;
-	struct dma_pte *parent, *pte;
-	struct dmar_domain *domain;
-	struct pci_dev *pdev;
-	int offset, level;
-
-	pdev = pci_get_domain_bus_and_slot(iommu->segment, bus, devfn);
-	if (!pdev)
-		return;
-
-	info = dev_iommu_priv_get(&pdev->dev);
-	if (!info || !info->domain) {
-		pr_info("device [%02x:%02x.%d] not probed\n",
-			bus, PCI_SLOT(devfn), PCI_FUNC(devfn));
-		return;
-	}
-
-	domain = info->domain;
-	level = agaw_to_level(domain->agaw);
-	parent = domain->pgd;
-	if (!parent) {
-		pr_info("no page table setup\n");
-		return;
-	}
+	struct dma_pte *pte;
+	int offset;
 
 	while (1) {
 		offset = pfn_level_offset(pfn, level);
@@ -846,9 +825,10 @@ void dmar_fault_dump_ptes(struct intel_iommu *iommu, u16 source_id,
 	struct pasid_entry *entries, *pte;
 	struct context_entry *ctx_entry;
 	struct root_entry *rt_entry;
+	int i, dir_index, index, level;
 	u8 devfn = source_id & 0xff;
 	u8 bus = source_id >> 8;
-	int i, dir_index, index;
+	struct dma_pte *pgtable;
 
 	pr_info("Dump %s table entries for IOVA 0x%llx\n", iommu->name, addr);
 
@@ -876,8 +856,11 @@ void dmar_fault_dump_ptes(struct intel_iommu *iommu, u16 source_id,
 		ctx_entry->hi, ctx_entry->lo);
 
 	/* legacy mode does not require PASID entries */
-	if (!sm_supported(iommu))
+	if (!sm_supported(iommu)) {
+		level = agaw_to_level(ctx_entry->hi & 7);
+		pgtable = phys_to_virt(ctx_entry->lo & VTD_PAGE_MASK);
 		goto pgtable_walk;
+	}
 
 	/* get the pointer to pasid directory entry */
 	dir = phys_to_virt(ctx_entry->lo & VTD_PAGE_MASK);
@@ -904,8 +887,16 @@ void dmar_fault_dump_ptes(struct intel_iommu *iommu, u16 source_id,
 	for (i = 0; i < ARRAY_SIZE(pte->val); i++)
 		pr_info("pasid table entry[%d]: 0x%016llx\n", i, pte->val[i]);
 
+	if (pasid_pte_get_pgtt(pte) == PASID_ENTRY_PGTT_FL_ONLY) {
+		level = pte->val[2] & BIT_ULL(2) ? 5 : 4;
+		pgtable = phys_to_virt(pte->val[2] & VTD_PAGE_MASK);
+	} else {
+		level = agaw_to_level((pte->val[0] >> 2) & 0x7);
+		pgtable = phys_to_virt(pte->val[0] & VTD_PAGE_MASK);
+	}
+
 pgtable_walk:
-	pgtable_walk(iommu, addr >> VTD_PAGE_SHIFT, bus, devfn);
+	pgtable_walk(iommu, addr >> VTD_PAGE_SHIFT, bus, devfn, pgtable, level);
 }
 #endif
 
Signature · 1 line
2.25.1
author maintainer
#1 lore raw
Quoted 44 lines · Lu Baolu

With CONFIG_INTEL_IOMMU_DEBUGFS enabled, below lockdep splat are seen when an I/O fault occurs on a machine with an Intel IOMMU in it.

DMAR: DRHD: handling fault status reg 3
DMAR: [DMA Write NO_PASID] Request device [00:1a.0] fault addr 0x0

        [fault reason 0x05] PTE Write access is not set
  DMAR: Dump dmar0 table entries for IOVA 0x0
  DMAR: root entry: 0x0000000127f42001
  DMAR: context entry: hi 0x0000000000001502, low 0x000000012d8ab001
  ================================
  WARNING: inconsistent lock state
  5.20.0-0.rc0.20220812git7ebfc85e2cd7.10.fc38.x86_64 #1 Not tainted
  --------------------------------
  inconsistent {HARDIRQ-ON-W} -> {IN-HARDIRQ-W} usage.
  rngd/1006 [HC1[1]:SC0[0]:HE0:SE1] takes:
  ff177021416f2d78 (&k->k_lock){?.+.}-{2:2}, at: klist_next+0x1b/0x160
  {HARDIRQ-ON-W} state was registered at:
    lock_acquire+0xce/0x2d0
    _raw_spin_lock+0x33/0x80
    klist_add_tail+0x46/0x80
    bus_add_device+0xee/0x150
    device_add+0x39d/0x9a0
    add_memory_block+0x108/0x1d0
    memory_dev_init+0xe1/0x117
    driver_init+0x43/0x4d
    kernel_init_freeable+0x1c2/0x2cc
    kernel_init+0x16/0x140
    ret_from_fork+0x1f/0x30
  irq event stamp: 7812
  hardirqs last  enabled at (7811): [<ffffffff85000e86>] asm_sysvec_apic_timer_interrupt+0x16/0x20
  hardirqs last disabled at (7812): [<ffffffff84f16894>] irqentry_enter+0x54/0x60
  softirqs last  enabled at (7794): [<ffffffff840ff669>] __irq_exit_rcu+0xf9/0x170
  softirqs last disabled at (7787): [<ffffffff840ff669>] __irq_exit_rcu+0xf9/0x170

The klist iterator functions using spin_*lock_irq*() but the klist insertion functions using spin_*lock(), combined with the Intel DMAR IOMMU driver iterating over klists from atomic (hardirq) context, where pci_get_domain_bus_and_slot() calls into bus_find_device() which iterates over klists.

As currently there's no plan to fix the klist to make it safe to use in atomic context, this fixes the lockdep splat by avoid calling pci_get_domain_bus_and_slot() in the hardirq context.

Quoted 5 lines

This patch has been queued:

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

Best regards,
baolu