kernel/git/stable/linux.git/drivers/vfio, branch linux-7.2.y

vfio/pci: Avoid mapping BARs for devices with non-mappable BARs

2026-09-14T11:40:21+00:00

[ Upstream commit 6edd69813b2d8c30c24a86052065a5dbf390a3cd ] vfio_pci_core_map_bars() calls pci_iomap() to set up BAR resources, but not all devices support having their BARs mapped by the CPU. The non_mappable_bars flag indicates that a PCI device's BARs cannot be accessed by the CPU. The ISM device on s390 is one such device. The BAR size for an ISM device is 256 TiB, and attempting to map the BAR will lead to warnings: vmalloc_node_range for size 281474976714752 failed: Address range restricted to 0x2110bab00000 - 0x21903ab00000 Use pdev->non_mappable_bars to skip pci_iomap() for such devices. This flag is set by the PCI core at enumeration time and already serves the same purpose in vfio_pci_probe_mmaps(). Fixes: 05f2a68b407a ("vfio/pci: Set up BAR resources and maps in vfio_pci_core_enable()") Reported-by: Christian Borntraeger Signed-off-by: Farhan Ali Reviewed-by: Niklas Schnelle Reviewed-by: Matthew Rosato Link: https://lore.kernel.org/r/20260729181116.1373-1-alifm@linux.ibm.com Signed-off-by: Alex Williamson Signed-off-by: Sasha Levin

vfio/pci: clear vdev->msi_perm after freeing it on init failure

2026-09-14T11:39:37+00:00

[ Upstream commit dc77acfeb979dded39b247b60fef0399536bfa77 ] vfio_msi_cap_len() lazily allocates the per-device MSI permission table: vdev->msi_perm = kmalloc_obj(struct perm_bits, GFP_KERNEL_ACCOUNT); if (!vdev->msi_perm) return -ENOMEM; ret = init_pci_cap_msi_perm(vdev->msi_perm, len, flags); if (ret) { kfree(vdev->msi_perm); return ret; /* vdev->msi_perm left dangling */ } When init_pci_cap_msi_perm() -> alloc_perm_bits() fails with -ENOMEM, the error path frees vdev->msi_perm but leaves the freed pointer stored in it. vdev->msi_perm is not re-zeroed later because struct vfio_pci_core_device is per-device and persists across open/close cycles, and the vfio_config_init() error path returns without calling vfio_config_free(). So the dangling pointer outlives the failed open. That leads to two use-after-frees on the same device: 1. Reuse. The next vfio_config_init() sees the stale pointer at "if (vdev->msi_perm) return len;" and reuses the freed object. MSI config accesses in vfio_pci_config_rw_single() then dereference and call the freed perm->readfn / perm->writefn function pointers. 2. Double free. A later vfio_config_free() runs free_perm_bits() and kfree() on the already-freed object. Fix it by NULLing vdev->msi_perm after the kfree(), matching the NULL-after-free discipline already used in free_perm_bits() and vfio_config_free(). BUG: KASAN: slab-use-after-free in vfio_pci_config_rw_single (drivers/vfio/pci/vfio_pci_config.c:1961) Read of size 8 at addr ffff88800fcc88d0 by task exploit/143 Call Trace: ... kasan_report (mm/kasan/report.c:595) vfio_pci_config_rw_single (drivers/vfio/pci/vfio_pci_config.c:1961) vfio_pci_config_rw (drivers/vfio/pci/vfio_pci_config.c:1986) vfio_pci_rw (drivers/vfio/pci/vfio_pci_core.c:1599) vfs_read (fs/read_write.c:572) __x64_sys_pread64 (fs/read_write.c:764) do_syscall_64 (arch/x86/entry/syscall_64.c:94) ... Followed on device close by a double free of the same object: Oops: general protection fault, probably for non-canonical address 0x1f63e0e8000008: 0000 [#1] SMP KASAN NOPTI RIP: 0010:kfree (mm/slub.c:6711) Call Trace: vfio_config_free (drivers/vfio/pci/vfio_pci_config.c:1861) vfio_pci_core_disable (drivers/vfio/pci/vfio_pci_core.c:685) vfio_pci_core_close_device (drivers/vfio/pci/vfio_pci_core.c:777) vfio_df_close (drivers/vfio/vfio_main.c:602) vfio_device_fops_release (drivers/vfio/vfio_main.c:648) __fput (fs/file_table.c:512) __x64_sys_close (fs/open.c:1496) do_syscall_64 (arch/x86/entry/syscall_64.c:94) ... Kernel panic - not syncing: Fatal exception Fixes: 30ea32ab1951 ("vfio/pci: Fix potential memory leak in vfio_msi_cap_len") Reported-by: Weiming Shi Assisted-by: Claude:claude-opus-4-8 Signed-off-by: Xiang Mei Link: https://lore.kernel.org/r/20260705014010.1297885-1-xmei5@asu.edu Signed-off-by: Alex Williamson Signed-off-by: Sasha Levin

vfio/pci: Expose latched module parameter policy in debugfs

2026-06-29T20:26:31+00:00

The nointxmask and disable_idle_d3 module parameters remain writable, but vfio-pci now latches their values into each device at init. Once a device is registered, changing the module parameter only affects future devices, leaving no direct way to confirm the effective policy for an existing device. Add a pci debugfs directory under the VFIO device debugfs root and report the per-device nointxmask and disable_idle_d3 values. These are read-only debugfs views and use the same Y/N bool output convention as the module parameters. Read-only vfio-pci parameters, such as disable_vga, are not exposed here because they cannot drift from the latched device value, therefore the existing module parameter exposure via sysfs is sufficient. Note that while only vfio-pci currently provides these options, the implementation is in vfio-pci-core and therefore properly reflects the device policy in the core, regardless of driver. Assisted-by: OpenAI Codex:gpt-5 Cc: Guixin Liu Signed-off-by: Alex Williamson Reviewed-by: Kevin Tian Link: https://lore.kernel.org/r/20260615191241.688297-7-alex.williamson@nvidia.com Signed-off-by: Alex Williamson

vfio: Remove device debugfs before releasing devres

2026-06-29T20:26:28+00:00

VFIO device debugfs files created with debugfs_create_devm_seqfile() store a devres allocated debugfs_devm_entry as inode private data. vfio_unregister_group_dev() currently calls vfio_device_del() before vfio_device_debugfs_exit(), but device_del() releases devres. This can leave debugfs entries visible with stale inode private data while unregister waits for userspace references to drain. Remove the per-device debugfs tree before vfio_device_del(). The debugfs view is diagnostic only, so losing it at the start of unregister is preferable to preserving entries whose backing storage may already have been released. Complete the teardown by clearing the per-device debugfs root after removal. This matches the global debugfs root cleanup and prevents future users from mistaking a removed dentry for a live debugfs tree during the remainder of unregister. Fixes: 2202844e4468 ("vfio/migration: Add debugfs to live migration driver") Reported-by: Sashiko AI Review Link: https://lore.kernel.org/r/20260615192725.6A2221F000E9@smtp.kernel.org Cc: stable@vger.kernel.org Cc: Longfang Liu Assisted-by: OpenAI Codex:gpt-5 Signed-off-by: Alex Williamson Reviewed-by: Kevin Tian Link: https://lore.kernel.org/r/20260615204717.735302-1-alex.williamson@nvidia.com Signed-off-by: Alex Williamson

vfio/pci: Latch all module parameters per device

2026-06-29T19:46:50+00:00

The vfio-pci module parameters of disable_idle_d3, nointxmask, and disable_vga latch vfio-pci policy into vfio-pci-core globals each time the vfio-pci module is initialized. The disable_idle_d3 parameter has already migrated to a per-device flag in order to provide consistency for refcounted PM operations for the lifetime of the device registration. Pull the remaining vfio-pci module-parameter policy out of vfio-pci-core into per-device flags set at device initialization. This also restores the mutable aspect of the disable_idle_d3 and nointxmask module parameters for vfio-pci, with the caveat that the parameters are latched into the device at probe. A notable change for variant drivers is that their devices are no longer affected by vfio-pci module parameters and those drivers may need to adopt similar module parameters if any devices have a hidden dependency on vfio-pci setting non-default policy. Assisted-by: Claude:claude-opus-4-8 Acked-by: Chengwen Feng Signed-off-by: Alex Williamson Reviewed-by: Kevin Tian Link: https://lore.kernel.org/r/20260615191241.688297-6-alex.williamson@nvidia.com Signed-off-by: Alex Williamson

vfio/mlx5: Fix racy bitfields and tighten struct layout

2026-06-29T19:46:50+00:00

Bitfield operations are not atomic, they use a read-modify-write pattern, therefore we should be careful not to pack bitfields that can be concurrently updated into the same storage unit. This split takes a binary approach: flags that are only modified pre/post open/close remain bitfields, flags modified from user action, including actions that reach across to another device (ex. reset) use dedicated storage units. Note mlx5_vhca_page_tracker.status is relocated to fill the alignment hole this split exposes. Bitfield justifications: migrate_cap: written only in mlx5vf_cmd_set_migratable() at probe chunk_mode: written only in mlx5vf_cmd_set_migratable() at probe mig_state_cap: written only in mlx5vf_cmd_set_migratable() at probe Dedicated storage units: mdev_detach: written in the VF attach/detach event notifier mlx5fv_vf_event() at runtime log_active: written in mlx5vf_start_page_tracker()/ mlx5vf_stop_page_tracker() during runtime dirty tracking deferred_reset: written in mlx5vf_state_mutex_unlock()/ mlx5vf_pci_aer_reset_done() during runtime reset handling is_err: set by tracker error handling and dirty-log polling at runtime object_changed: set by tracker event handling and cleared by dirty-log polling at runtime Fixes: 61a2f1460fd0 ("vfio/mlx5: Manage the VF attach/detach callback from the PF") Fixes: 79c3cf279926 ("vfio/mlx5: Init QP based resources for dirty tracking") Fixes: f886473071d6 ("vfio/mlx5: Add support for tracker object change event") Cc: Yishai Hadas Cc: stable@vger.kernel.org Assisted-by: Claude:claude-opus-4-8 Signed-off-by: Alex Williamson Reviewed-by: Kevin Tian Link: https://lore.kernel.org/r/20260615191241.688297-5-alex.williamson@nvidia.com Signed-off-by: Alex Williamson

vfio/pci: Release the VGA arbiter client on register_device() failure

2026-06-29T19:46:49+00:00

The re-order in the Fixes commit below displaced vfio_pci_vga_init() as the last failure point of what is now vfio_pci_core_register_device() without introducing an unwind for the VGA arbiter registration. In current kernels this is mostly benign because vfio_pci_set_decode() only uses pci_dev state, but the original failure path could leave a callback with a freed vdev cookie. The stale registration also becomes unsafe again once the callback follows drvdata to the vfio device. Add the required VGA unwind callout. Fixes: 4aeec3984ddc ("vfio/pci: Re-order vfio_pci_probe()") Cc: stable@vger.kernel.org Assisted-by: Claude:claude-opus-4-8 Signed-off-by: Alex Williamson Reviewed-by: Kevin Tian Link: https://lore.kernel.org/r/20260615191241.688297-3-alex.williamson@nvidia.com Signed-off-by: Alex Williamson

vfio/pci: Latch disable_idle_d3 per device

2026-06-29T19:46:49+00:00

When disable_idle_d3 was introduced in vfio-pci, it directly manipulated the device power state with pci_set_power_state(). There were no refcounts to maintain or balanced operations, we could unconditionally bring the device to D0 and conditionally move it to D3hot. Therefore the module parameter was made writable. Later, in commit c61302aa48f7 ("vfio/pci: Move module parameters to vfio_pci.c"), as part of the vfio-pci-core split, the writable aspect of the module parameter was nullified. The parameter value could still be changed through sysfs, but the vfio-pci driver latched the values into vfio-pci-core globals at module init. Loading the vfio-pci module, or unloading and reloading, with non-default or different values could change the globals relative to existing devices bound to vfio-pci variant drivers. Runtime PM was introduced in commit 7ab5e10eda02 ("vfio/pci: Move the unused device into low power state with runtime PM"), which marks the point where power states became refcounted. PM get and put operations need to be balanced, but the same module operations noted above can change the global variables relative to those devices already bound to vfio-pci variant drivers. This introduces a window where PM operations can now become unbalanced. To resolve this with a narrow footprint for stable backports, the disable_idle_d3 flag is latched into the vfio_pci_core_device at the time of initialization, such that the device always operates with a consistent value. NB. vfio_pci_dev_set_try_reset() now unconditionally raises the runtime PM usage count around bus reset to account for disable_idle_d3 becoming a per-device rather than global flag. When this flag is set, the additional get/put pair is harmless and allows continued use of the shared vfio_pci_dev_set_pm_runtime_get() helper. Fixes: 7ab5e10eda02 ("vfio/pci: Move the unused device into low power state with runtime PM") Cc: stable@vger.kernel.org Assisted-by: Claude:claude-opus-4-8 Signed-off-by: Alex Williamson Reviewed-by: Kevin Tian Link: https://lore.kernel.org/r/20260615191241.688297-2-alex.williamson@nvidia.com Signed-off-by: Alex Williamson

vfio/qat: fix f_pos race in qat_vf_resume_write()

2026-06-10T20:33:05+00:00

qat_vf_resume_write() checks filp->f_pos before taking migf->lock, but copies into the migration-state buffer after taking the lock and re-reading the shared file position. Two concurrent writers could therefore pass the bounds check with the old offset, then have the second writer copy after the first advanced f_pos, writing past the end of the migration-state buffer. Take migf->lock before doing the boundary checks. Fixes: bb208810b1ab ("vfio/qat: Add vfio_pci driver for Intel QAT SR-IOV VF devices") Reviewed-by: Ahsan Atta Signed-off-by: Giovanni Cabiddu Link: https://lore.kernel.org/r/20260608151317.136613-1-giovanni.cabiddu@intel.com Signed-off-by: Alex Williamson

vfio: prevent infinite loop in vfio_mig_get_next_state() on blocked arc

2026-06-05T16:45:41+00:00

vfio_mig_get_next_state() walks vfio_from_fsm_table[] one step at a time, looping to skip optional states the device does not support until *next_fsm is supported. A blocked transition is encoded as VFIO_DEVICE_STATE_ERROR, which the trailing return reports as -EINVAL. The skip loop does not account for the ERROR sentinel. state_flags_table[ERROR] is ~0U and vfio_from_fsm_table[ERROR][*] is ERROR, so once *next_fsm becomes ERROR the loop condition stays true and *next_fsm never changes. The blocked arcs STOP_COPY -> PRE_COPY and STOP_COPY -> PRE_COPY_P2P map to ERROR yet pass the support check on a precopy-capable device, causing the loop to spin forever while holding the driver state mutex. This can result in a soft lockup, and a panic with softlockup_panic set. Terminate the skip loop on the ERROR sentinel so a blocked transition falls through to the existing return and reports -EINVAL. Fixes: 4db52602a607 ("vfio: Extend the device migration protocol with PRE_COPY") Reported-by: Yuhao Jiang Cc: stable@vger.kernel.org Signed-off-by: Junrui Luo Reviewed-by: Kevin Tian Reviewed-by: Jason Gunthorpe Link: https://lore.kernel.org/r/SYBPR01MB7881290BBDE79B61AE6A017FAF122@SYBPR01MB7881.ausprd01.prod.outlook.com Signed-off-by: Alex Williamson