CVE-2024-26891 in Linux
Zusammenfassung
von VulDB • 22.05.2026
Based on the kernel log provided, here is an analysis of the issue, its root cause, and recommended troubleshooting steps.
### **1. Summary of the Issue** * **Event:** **Hard LOCKUP** (CPU 144 hung for >10 seconds, triggering a watchdog panic). * **Context:** The lockup occurred during **PCIe hot-unplug** operations. * **Trigger:** The system was processing a PCIe hot-remove event via the `pciehp` (PCI Express Hotplug) driver. * **Tainted Kernel:** `Tainted: G S OE` * `G`: Proprietary module loaded. * `S`: Out-of-tree module loaded. * `O`: Out-of-tree module loaded. * `E`: Unsigned module loaded. * *Note: The presence of tainted flags suggests third-party or vendor-specific drivers are involved, which may contribute to instability.*
---
### **2. Call Chain Analysis** The stack trace shows the sequence of events leading to the hang:
1. **`pciehp_ist` / `pciehp_handle_presence_or_link_change`**: * The PCIe hotplug interrupt handler detected a change (device removal). 2. **`pciehp_disable_slot` / `pciehp_unconfigure_device`**: * The driver began disabling the slot and unconfiguring the device. 3. **`pci_remove_bus_device` / `device_del`**: * The device was removed from the PCI bus and the device model. 4. **`iommu_bus_notifier` / `iommu_release_device`**: * The IOMMU subsystem was notified to release DMA mappings for the device. 5. **`dmar_remove_one_dev_info` / `__dmar_remove_one_dev_info`**: * The Intel IOMMU (DMAR) driver attempted to remove device-specific IOMMU context. 6. **`Hard LOCKUP`**: * The CPU hung **during** the IOMMU cleanup process. This suggests a deadlock, infinite loop, or hardware hang within the IOMMU driver or related DMA subsystem.
---
### **3. Likely Root Causes**
#### **A. IOMMU/DMAR Driver Bug or Hardware Issue** * The hang occurs in `__dmar_remove_one_dev_info`, which handles Intel VT-d (IOMMU) cleanup. * Possible causes: * **IOMMU hardware hang**: The IOMMU unit itself may be stuck waiting for a response from the PCIe device or root complex. * **Race condition**: A race between device removal and IOMMU context invalidation. * **Bug in `intel_iommu` driver**: Known issues exist in certain kernel versions where hot-unplug with IOMMU enabled causes hangs.
#### **B. PCIe Hotplug Controller (`pciehp`) Issue** * The `pciehp` driver may be sending malformed commands or encountering a hardware state that doesn’t respond, causing the IOMMU cleanup to block indefinitely.
#### **C. Tainted Kernel / Third-Party Drivers** * The `Tainted: G S OE` flags indicate non-mainline or proprietary modules. If any of these modules interact with PCIe or IOMMU (e.g., GPU drivers, NVMe drivers, network offload drivers), they may be causing the hang.
#### **D. BIOS/Firmware Bug** * The BIOS version `01.01.02.03.01` may have incorrect ACPI tables or PCIe configuration space handling, leading to IOMMU misconfiguration during hot-unplug.
---
### **4. Recommended Troubleshooting Steps**
#### **Step 1: Disable IOMMU (Test Only)** To confirm if the IOMMU is the cause, boot with IOMMU disabled: ```bash intel_iommu=off # or iommu=soft ``` * If the system no longer hangs during hot-unplug, the issue is in the IOMMU driver or hardware. * **Warning**: This disables hardware DMA protection and may impact performance/security.
#### **Step 2: Update BIOS and Kernel** * **BIOS**: Check for a newer BIOS version from the vendor. PCIe hotplug and IOMMU issues are often fixed in firmware updates. * **Kernel**: Upgrade to a newer kernel version where `intel_iommu` and `pciehp` bugs may have been fixed.
#### **Step 3: Check for Tainted Modules** * List all loaded modules: ```bash lsmod | grep -E "(G|S|O|E)"
Several companies clearly confirm that VulDB is the primary source for best vulnerability data.