Vulnerabilidades do Linux Linux

🔔 Criar alerta RSS Verificador de versão

CVEs publicadas que afetam Linux Linux: 10465 registro(s), com versões afetadas e correções.

CVE-2026-89733 Alta · 7.8 11/09/2026

usb: gadget: uvc: fix dangling pointers in uvc_function_bind() and uvc_function_unbind()

In the Linux kernel, the following vulnerability has been resolved: usb: gadget: uvc: fix dangling pointers in uvc_function_bind() and uvc_function_unbind() In uvc_function_bind() error path, we use usb_ep_free_request which uses uvc->control_req but does not set it to NULL afterwards. Thus, uvc->control_req is a dangling pointer causing a UAF. Also we do not set the uvc->control_buf pointer to NULL after freeing it, which is another dangling pointer. Fix it by setting uvc->control_req to NULL after we run usb_ep_free_request() and uvc->control_buf to NULL after kfree. Do the same for uvc_function_unbind().

CVE-2026-89732 Desconhecida 11/09/2026

usb: gadget: f_fs: Prevent deadlock during ep0 read loop

In the Linux kernel, the following vulnerability has been resolved: usb: gadget: f_fs: Prevent deadlock during ep0 read loop Currently, ffs_ep0_read() holds ffs->mutex when it prepares to go to sleep waiting for an event. When no setup events are pending, it calls wait_event_interruptible_exclusive_locked_irq() with the mutex still held. The wait macro deliberately drops the waitqueue spinlock before sleeping but does not drop the mutex. If a userspace daemon is polling ep0 via read() and the gadget is asynchronously torn down via configfs (e.g., echo "" > UDC), a deadlock can occur: 1. The configfs teardown calls functionfs_unbind(), which queues a FUNCTIONFS_UNBIND event. 2. The daemon wakes up, consumes the event, and drops the mutex. 3. However, if the daemon loops and immediately issues another read() before exiting, it reacquires ffs->mutex and again goes into an interruptible sleep. 4. Meanwhile, functionfs_unbind() continues execution and attempts to acquire ffs->mutex to tear down ep0req. 5. The kernel deadlocks because the configfs thread is stuck in an uninterruptible sleep waiting for the mutex, while the userspace daemon is in an interruptible sleep holding the mutex forever because no more events will arrive. To fix this, we drop both the waitqueue spinlock and ffs->mutex before going to sleep, and use wait_event_interruptible_exclusive() instead. Upon waking up, we jump back to the `retry` label to safely reacquire the mutex and re-evaluate the state machine. By not sleeping with ffs->mutex held, we natively decouple gadget teardowns (which require the mutex) from userspace polling.

CVE-2026-89731 Alta · 7.1 11/09/2026

cxl/ras: Fix cxl_rch_get_aer_info() out-of-bounds AER register read

In the Linux kernel, the following vulnerability has been resolved: cxl/ras: Fix cxl_rch_get_aer_info() out-of-bounds AER register read cxl_rch_get_aer_info() copies the RCH Downstream Port AER capability from the RCRB MMIO block using a readl() loop bounded by sizeof(struct aer_capability_regs). This struct is a software layout and its embedded struct pcie_tlp_log is larger than the on-wire AER capability. As a result the loop reads past the mapped AER register block. The over-read also populates the software-only tail fields including header_log.header_len. An out-of-range header_len passed to pcie_print_tlp_log() can then loop past the header log buffer and cause a second out-of-bounds read. The read was correct when introduced, but struct pcie_tlp_log has since grown (Header Log and TLP Prefix Log sizes, header_len and flit fields), so sizeof(struct aer_capability_regs) no longer matches the physical AER capability. Bound the read to the physical AER registers, header through the 16 byte Header Log. Zero the destination first so the software-only fields are deterministic.

CVE-2026-89730 Desconhecida 11/09/2026

fpga: altera-cvp: Avoid out-of-bounds read in trailing byte write

In the Linux kernel, the following vulnerability has been resolved: fpga: altera-cvp: Avoid out-of-bounds read in trailing byte write The trailing byte path in altera_cvp_send_block() dereferences a u32 pointer even when only 1-3 bytes remain in the input buffer. If the buffer ends at a page or scatterlist boundary, this can read past the valid image data and fault. Copy the remaining bytes into a zero-initialized u32 before writing the final word so only valid bytes are read from the input buffer.

CVE-2026-89729 Alta · 8.8 11/09/2026

HID: sensor-hub: Fix out-of-bounds write in sensor_hub_get_feature

In the Linux kernel, the following vulnerability has been resolved: HID: sensor-hub: Fix out-of-bounds write in sensor_hub_get_feature sensor_hub_get_feature() clamps its return value to the caller's buffer size, but the copy loop still copies field->report_size / 8 bytes for each report value. A malicious HID descriptor can advertise a large feature field size while an IIO caller supplies a small stack buffer, such as a single s32, causing an out-of-bounds write. HID core stores parsed report values in __s32 slots and clamps extracted values to 32 bits. Reject feature fields that require more than one slot per value, guard the total byte count calculation, and clamp each per-value copy to the remaining caller buffer.

CVE-2026-89728 Desconhecida 11/09/2026

i3c: renesas: Fix out-of-bounds access for newdevs mask

In the Linux kernel, the following vulnerability has been resolved: i3c: renesas: Fix out-of-bounds access for newdevs mask When software initiates DAA (Dynamic Address Assignment), the controller reports the result via the NRSPQP (Normal Response Queue Port Register). The data length field of the response descriptor, which is accessible through the NRSPQP register, indicates the number of devices remaining after DAA. Consequently, when the bus is empty, this field contains the maximum number of devices supported by the controller (8 for the Renesas I3C controller). Adjust the condition that computes the newly discovered devices bitmask to prevent an out-of-bounds when the I3C bus is empty.

CVE-2026-89727 Desconhecida 11/09/2026

KVM: arm64: GICv2: Don't WARN on out-of-range GICV_DIR INTID

In the Linux kernel, the following vulnerability has been resolved: KVM: arm64: GICv2: Don't WARN on out-of-range GICV_DIR INTID vgic_v2_deactivate() passes the INTID a guest wrote to GICV_DIR straight to vgic_get_vcpu_irq(), and treats a failed lookup as a "can't happen" condition with WARN_ON_ONCE(). The guest can make it happen at will, though: for any INTID outside of the implemented SGI, PPI and SPI ranges the lookup returns NULL, since GICv2 has no LPIs. A guest running with EOImode==1 writing such an INTID to GICV_DIR triggers the WARN, and panics hosts running with panic_on_warn. Drop the WARN and ignore failed lookups.

CVE-2026-89726 Desconhecida 11/09/2026

lib/ucs2_string.c: fix out-of-bounds read in ucs2_strnlen()

In the Linux kernel, the following vulnerability has been resolved: lib/ucs2_string.c: fix out-of-bounds read in ucs2_strnlen() Patch series "lib/ucs2_string.c: fix out-of-bounds read in ucs2_strnlen()", v2. This series fixes an off-by-one out-of-bounds read in ucs2_strnlen(). The first patch is the real fix, the second patch comes as a bonus and fixes the code indentation. This patch (of 2): ucs2_strnlen() checks the current character before checking whether the caller-provided maximum length has been reached. If the input is not NUL-terminated within that bound, the loop can read one ucs2_char_t past the limit. Test the length before dereferencing to prevent an off-by-one out-of-bounds read.

CVE-2026-89725 Alta · 8.8 11/09/2026

media: cec: stm32: prevent out-of-bounds write on RX overflow

In the Linux kernel, the following vulnerability has been resolved: media: cec: stm32: prevent out-of-bounds write on RX overflow stm32_rx_done() appends each received CEC byte to rx_msg.msg[] using rx_msg.len as the write index, incrementing it on every RXBR (receive-byte-ready) interrupt without checking it against the buffer size: cec->rx_msg.msg[cec->rx_msg.len++] = val & 0xFF; rx_msg.msg[] is a fixed CEC_MAX_MSG_SIZE (16) byte array in struct cec_msg, and rx_msg.len is only reset on RXACKE/RXOVR or after a completed message (RXEND). The number of bytes received before RXEND is decided by the remote CEC device (it sets EOM), not by the driver. A peer that keeps sending bytes without ending the message drives RXBR repeatedly, pushing rx_msg.len past 16 and writing peer-controlled bytes out of bounds into the surrounding memory. This is reachable in normal operation once the driver has probed and receiving is enabled, from the IRQ thread, without any local privilege. The length check in the CEC core runs on the consumer side, after the byte has been stored, so it does not prevent the overflow. Bound the index in the driver before the store, as the other platform CEC drivers already do (e.g. tegra_cec), dropping the excess bytes of an overlong frame. Found by static analysis tool CodeQL.

CVE-2026-89724 Alta · 7.8 11/09/2026

media: vicodec: fix out-of-bounds write in FWHT encoder

In the Linux kernel, the following vulnerability has been resolved: media: vicodec: fix out-of-bounds write in FWHT encoder vidioc_s_fmt_vid_out() sizes the encoder CAPTURE buffer from the compressed descriptor pixfmt_fwht, whose sizeimage_mult is 3: coded_w * coded_h * 3 + sizeof(struct fwht_cframe_hdr). fwht_encode_frame() encodes one plane per component, and an incompressible plane takes the FWHT_FRAME_UNENCODED path in encode_plane(), copying the plane verbatim. For a 4-component pixel format all four planes are full resolution (width_div == height_div == 1), so a frame that forces every plane through the unencoded fallback writes sizeof(struct fwht_cframe_hdr) + 4 * coded_w * coded_h bytes, overrunning the plane by coded_w * coded_h, which can result in corruption of adjacent kernel heap memory. Bump pixfmt_fwht.sizeimage_mult from 3 to 4, matching the largest components_num among the supported raw formats, so the capture buffer is always large enough for the unencoded fallback.

CVE-2026-89723 Alta · 7.8 11/09/2026

nilfs2: fix slab-out-of-bounds in nilfs_direct_propagate after truncation

In the Linux kernel, the following vulnerability has been resolved: nilfs2: fix slab-out-of-bounds in nilfs_direct_propagate after truncation Shuangpeng Bai reported that KASAN detected a slab-out-of-bounds error in nilfs_direct_propagate() during testing. Analysis revealed that after truncating a file, a node block immediately below the B-tree root was not deleted. Instead, it remained in the B-tree node cache in a dirty state. The log writer subsequently detected this block and incorrectly invoked nilfs_direct_propagate() on it, which is designed to handle only data blocks in direct mapping. B-tree nodes in the cache are managed by virtual block numbers, and their logical keys typically exceed the range expected by direct mapping. Consequently, processing such a node as a direct mapping entry triggers a slab-out-of-bounds access. The root cause is that when a B-tree mapping collapses into a direct mapping during truncation, an intermediate node block pointed to by the root node is left behind as garbage instead of being explicitly deleted. This resolves the issue by adding a nilfs_btree_discard() operation to delete the remaining intermediate node block during the conversion. A 'deform' flag is added to the bop_delete interface to explicitly signal that the deletion is part of a mapping transformation. This allows the B-tree mapping implementation to perform the necessary cleanup and discarding of the residual node structure that would be otherwise be left orphaned after the transition.

CVE-2026-89722 Desconhecida 11/09/2026

PCI/sysfs: Fix out-of-bounds read in pci_write_legacy_io()

In the Linux kernel, the following vulnerability has been resolved: PCI/sysfs: Fix out-of-bounds read in pci_write_legacy_io() pci_write_legacy_io() loads 4 bytes from the kernfs write buffer regardless of how many bytes userspace wrote: if (count != 1 && count != 2 && count != 4) return -EINVAL; return pci_legacy_write(bus, off, *(u32 *)buf, count); kernfs_fop_write_iter() allocates the buffer with kmalloc(len + 1), so a 1-byte write to the legacy_io sysfs file allocates 2 bytes and the unconditional u32 load reads up to 2 bytes past the end of the allocation, which KASAN reports as a slab-out-of-bounds read. Similarly, a 2-byte write overreads by 1 byte. Thus, read only the number of bytes requested using get_unaligned_le16() and get_unaligned_le32() for the 2 and 4 byte cases, interpreting the buffer as little-endian to match the byte ordering of PCI I/O port space. The PowerPC implementation previously compensated for the generic code's native-endian 32-bit load by shifting the value into place for the 1 and 2 byte cases. The shifts were only correct on big-endian kernels. On little-endian PowerPC (POWER8 and later), they extracted the wrong bytes, so a 1-byte write wrote an out-of-bounds byte instead of the requested value. On big-endian, the native load also caused out_le16() and out_le32() to reverse the user's bytes on the wire for 2 and 4 byte writes. The little-endian helpers resolve both issues, so the shifts are removed. No changes are needed for the Alpha platform. The legacy_io file is root-only and exists only on Alpha and PowerPC, the two architectures that define HAVE_PCI_LEGACY.

CVE-2026-89721 Desconhecida 11/09/2026

phy: rockchip-samsung-dcphy: fix out-of-range max_register

In the Linux kernel, the following vulnerability has been resolved: phy: rockchip-samsung-dcphy: fix out-of-range max_register The PHY register block is 64KB, so with a register stride of 4 the last accessible register sits at offset 0xfffc. max_register names 0x10000, one register past the end of the mapping: dumping the registers through the regmap debugfs interface reads beyond the ioremapped region and oopses on the unmapped page. The oops fires with the regmap lock held, so later PHY operations deadlock.

CVE-2026-89720 Alta · 7.7 11/09/2026

ubifs: fix out-of-bounds read in signature length check

In the Linux kernel, the following vulnerability has been resolved: ubifs: fix out-of-bounds read in signature length check ubifs_sb_verify_signature() bounds the on-disk ubifs_sig_node->len field before handing the signature payload to verify_pkcs7_signature(), but the check has the wrong sign: if (le32_to_cpu(signode->len) > snod->len + sizeof(struct ubifs_sig_node)) The signature bytes start sizeof(struct ubifs_sig_node) (UBIFS_SIG_NODE_SZ, 64 bytes) into the node, so the payload is at most snod->len - sizeof(struct ubifs_sig_node) bytes long. Adding the header size instead of subtracting it accepts a declared length up to 2 * UBIFS_SIG_NODE_SZ larger than the node actually holds -- past the end of c->sbuf, which is vmalloc(c->leb_size). verify_pkcs7_signature() -> pkcs7_parse_message() -> asn1_ber_decoder() is then handed that inflated length and reads beyond the allocation while walking the DER headers. The node length comes straight from the mounted image, so a crafted signed UBIFS image reaches this via ubifs_read_superblock() before the signature is cryptographically checked. snod->len is guaranteed to be >= UBIFS_SIG_NODE_SZ by the node scanner (c->ranges[UBIFS_SIG_NODE].min_len == UBIFS_SIG_NODE_SZ), so the corrected subtraction cannot underflow. Legitimately signed images are unaffected: a correct superblock never declares a signature longer than the node it is embedded in.

CVE-2026-89719 Desconhecida 11/09/2026

zram: fix out-of-bounds access in read_block_state()

In the Linux kernel, the following vulnerability has been resolved: zram: fix out-of-bounds access in read_block_state() read_block_state() calculates nr_pages before taking dev_lock. If the device is reset and reinitialized with a smaller disksize before lock acquisition, nr_pages still describes the old table. The subsequent loop can then call slot_lock() past the end of the newly allocated table. Read disksize after acquiring dev_lock and checking that the device is initialized. The read lock then keeps the table and its bound stable for the duration of the scan.

CVE-2026-89717 Desconhecida 11/09/2026

zram: set default primary compressor in zram_destroy_comps()

In the Linux kernel, the following vulnerability has been resolved: zram: set default primary compressor in zram_destroy_comps() Patch series "zram: fix zram issues reported by sashiko". Sashiko drove by and reported [1] a couple of zram issues: a possible BUG_ON() in zlib code due to missing winbits range validation and one possible NULL-ptr dereference in zcomp. Both are low risk yet still worth fixing. This patch (of 2): zram_destroy_comps() resets all compressors and leaves them set to NULL, including the primary one, which is invalid device state, as now comp_algorithm_show()->strcmp() can be called on a NULL compressor. Set default primary compressor in zram_destroy_comps().

CVE-2026-89716 Desconhecida 11/09/2026

zram: validate deflate params

In the Linux kernel, the following vulnerability has been resolved: zram: validate deflate params We must validate user-supplied deflate winbits before we pass it to zlib_deflate_workspacesize(), which triggers BUG_ON() if winbits value is outside of valid ranges.

CVE-2026-89718 Desconhecida 11/09/2026

zram: fix out-of-bounds access in writeback_store()

In the Linux kernel, the following vulnerability has been resolved: zram: fix out-of-bounds access in writeback_store() Patch series "zram: fix stale scan bounds after reinitialization". Both writeback_store() and read_block_state() derive their table scan bounds from zram->disksize before acquiring dev_lock. If the device is reset and reinitialized with a smaller disksize between that read and lock acquisition, the bound can describe the old table while the scan operates on the new one. This can lead to out-of-bounds slot accesses. Move both bound calculations under dev_lock so each bound remains consistent with the table throughout its scan. Keep the fixes separate because the affected interfaces originate from different commits and can be backported independently. This patch (of 2): writeback_store() calculates the table scan bounds before taking dev_lock. A reset followed by reconfiguration with a smaller disksize can therefore replace zram->table while writeback_store() is waiting for the lock. Once it acquires the lock, it sees an initialized device but scans the new table using the old upper bound, resulting in an out-of-bounds access. Calculate the number of pages while holding dev_lock so the scan bound matches the table protected by the lock.

CVE-2026-89715 Desconhecida 11/09/2026

NFS/localio: fix ref leak on nfs_uuid_add_file failure

In the Linux kernel, the following vulnerability has been resolved: NFS/localio: fix ref leak on nfs_uuid_add_file failure When nfs_uuid_add_file() races with nfs_uuid_put() tearing down uuid->net, it returns -ENXIO without publishing nfl->nfs_uuid via rcu_assign_pointer(). nfs_open_local_fh() then enters its error branch and only releases the slot's file ref and its paired net ref plus its own entry-time net ref, while the close path is a no-op: nfs_close_local_fh() nfs_uuid = rcu_dereference(nfl->nfs_uuid); if (!nfs_uuid) { rcu_read_unlock(); return; } /* always */ nfsd_open_local_fh() returns localio holding a caller-owned +1 nfsd_file reference (from nfsd_file_get() after nfsd_file_acquire_local()) and an entry-time nfsd_net reference (from its first nfsd_net_try_get()) embedded as nf->nf_net. Both are leaked on the failure path, pinning one nfsd_file (and the underlying struct file, dentry, inode) and one nfsd_net_ref per occurrence, which blocks nfsd_net and netns teardown. Fix by releasing the caller-owned file ref and its net ref through the existing helper, using a stack-local RCU pointer so the helper can xchg it out, then returning -ENXIO so callers do not dereference a localio whose slot has been cleared: struct nfsd_file __rcu *tmp = RCU_INITIALIZER(localio); nfs_to_nfsd_file_put_local(pnf); nfs_to_nfsd_file_put_local(&tmp); localio = ERR_PTR(-ENXIO); The trailing nfs_to_nfsd_net_put(net) continues to release the outer net ref, so all three nfsd_net_try_get() increments are balanced on the error branch.

CVE-2026-89714 Desconhecida 11/09/2026

NFS: fix delegation_hash_table leak when nfs4_server_common_setup() fails

In the Linux kernel, the following vulnerability has been resolved: NFS: fix delegation_hash_table leak when nfs4_server_common_setup() fails nfs4_server_common_setup() allocates server->delegation_hash_table first, but server->destroy - the only path that frees the table via nfs4_destroy_server() - is not assigned until the very end of the function. If any intermediate step fails (the is_ds_only_client() check, nfs4_init_session(), nfs4_get_rootfh(), or nfs_probe_server()), the function returns with server->destroy still NULL, so the caller's nfs_free_server() skips the destroy callback and the hash table is leaked (4 KiB per attempt with the default delegation watermark). This is trivially reachable from userspace: every failed NFSv4 mount leaks one allocation. A client that persistently retries a mount that cannot succeed leaks kernel memory without bound. Observed in production where a Longhorn backup poller retried mount.nfs4 against an NFSv3-only server roughly 10 times per second, leaking ~3.4 GiB of unreclaimable slab (kmalloc-rnd-13-4k) per day; the node accumulated 12 GiB of leaked slab before the source was identified via the kmem:kmalloc tracepoint (call_site=nfs4_delegation_hash_alloc). Reproducer: # server exports NFSv3 only (or export path absent for v4) while :; do mount -t nfs4 <server>:/missing /mnt; done # watch SUnreclaim in /proc/meminfo grow 4 KiB per iteration Free the table on the error paths between the allocation and the assignment of server->destroy.

CVE-2026-89713 Crítica · 9.1 11/09/2026

NFSD: check truncate permission under inode lock

In the Linux kernel, the following vulnerability has been resolved: NFSD: check truncate permission under inode lock nfsd_setattr() checks whether a size update needs NFSD_MAY_TRUNC before it takes inode_lock(). The comparison uses the file size sampled by that unlocked read, but the actual ATTR_SIZE update is applied later under inode_lock() by notify_change(). This leaves a TOCTOU window for append-only files. If a client sends a SETATTR that does not shrink the file at the time of the unlocked sample, a concurrent append can extend the file before nfsd_setattr() takes inode_lock(). notify_change() then applies a real truncation without the NFSD_MAY_TRUNC check that rejects IS_APPEND(inode). The VFS truncate syscall paths perform their own append-only checks before calling notify_change(), so NFSD must make this decision against the locked size it is about to change. Split the write-count acquisition from the truncation permission check. Keep get_write_access() before the locked setattr work, then recheck whether the requested size is below i_size_read(inode) after inode_lock() has been acquired and before notify_change(ATTR_SIZE). This also avoids the plain unlocked inode->i_size load.

CVE-2026-89712 Crítica · 9.8 11/09/2026

NFSD: restart ssc_expire_umount walk after dropping nfsd_ssc_lock

In the Linux kernel, the following vulnerability has been resolved: NFSD: restart ssc_expire_umount walk after dropping nfsd_ssc_lock nfsd4_ssc_expire_umount() walks nn->nfsd_ssc_mount_list with list_for_each_entry_safe(ni, tmp, ...). For each expired entry it sets nsui_busy = true, drops nfsd_ssc_lock to run mntput() on the source vfsmount, then reacquires the lock to list_del + kfree the entry and continue iterating via the macro's saved tmp pointer. The nsui_busy flag protects the current ni from concurrent nfsd4_ssc_setup_dul() finders during the lock-drop window, but it does not pin tmp. Another nfsd RPC thread that fails its source- server mount and reaches nfsd4_ssc_cancel_dul() will, during that same window, take nfsd_ssc_lock, list_del + kfree its own ssc_umount item, and release the lock. If that item is the saved tmp of the expire walk, the next iteration dereferences a freed nfsd4_ssc_umount_item. Restart the walk from the head after the mntput() unlock window so no saved next pointer survives the lock-drop. The list is bounded by the number of active inter-server source mounts (typically small) and the expire delayed-work runs periodically rather than per-IO, so the restart is cheap.

CVE-2026-89711 Alta · 8.2 11/09/2026

NFSD: remove flawed WARN_ON_ONCE from nfsd_mode_check

In the Linux kernel, the following vulnerability has been resolved: NFSD: remove flawed WARN_ON_ONCE from nfsd_mode_check The header for commit e75b23f9e323 ("nfsd: check d_can_lookup in fh_verify of directories") details the assumption that justified adding the WARN_ON_ONCE to nfsd_mode_check(), that assumption is invalid (in the case of NFS reexport). When NFSD exports an NFS filesystem it is very possible for nfsd_mode_check() to encounter a @dentry that doesn't have i_op->lookup (see nfs_fhget()'s NFS_ATTR_FATTR_MOUNTPOINT and NFS_ATTR_FATTR_V4_REFERRAL handling, and d_flags_for_inode()). So remove nfsd_mode_check()'s WARN_ON_ONCE(). The nfserr_notdir return on that branch must stay. It guards the subsequent lookup_one_unlocked() -> __lookup_slow() path, which calls inode->i_op->lookup() with no NULL check, so returning nfserr_notdir is what keeps a client LOOKUP into such a @dentry from dereferencing a NULL method pointer.

CVE-2026-89710 Desconhecida 11/09/2026

NFSv4.1: fix layout segment leak on the pnfs_layout_process() forget path

In the Linux kernel, the following vulnerability has been resolved: NFSv4.1: fix layout segment leak on the pnfs_layout_process() forget path When the server returns a new layout stateid while a valid one is still held, pnfs_layout_process() calls pnfs_mark_matching_lsegs_return() on the on-stack free_me list and jumps to out_forget. Segments whose reference count drops to zero are unlinked from lo->plh_segs and moved to free_me by mark_lseg_invalid(); for an idle cached segment the layout header holds the only reference, so this happens on the first decrement. out_forget never drains free_me -- only the success path calls pnfs_free_lseg_list(). Commit 814b84971388 ("pNFS/NFSv4: Fix a layout segment leak in pnfs_layout_process()") added the drain; commit 08bd8dbe8882 ("pNFS/NFSv4: Try to return invalid layout in pnfs_layout_process()") removed it while switching the destination to lo->plh_return_segs, which is drained elsewhere. Commit fb700ef02676 ("NFSv4.1: Simplify layout return in pnfs_layout_process()") switched the destination back to free_me without restoring the drain. Restore the pnfs_free_lseg_list() call.

CVE-2026-89708 Crítica · 9.8 11/09/2026

nfsd: RCU-protect cl_cb_session to fix use-after-free on session teardown

In the Linux kernel, the following vulnerability has been resolved: nfsd: RCU-protect cl_cb_session to fix use-after-free on session teardown After a DESTROY_SESSION the per-session teardown path can free a session while rpciod still holds an inflight callback rpc_task that dereferences clp->cl_cb_session. nfsd4_probe_callback_sync() flushes cl_callback_wq, but once nfsd4_run_cb_work() has called rpc_call_async() the rpc_task lives on rpciod; flushing the workqueue does not wait for it. rpc_shutdown_client() does drain rpciod tasks, but uses a 1-second wait_event_timeout — tasks stuck in rpc_delay() (e.g. 2-second NFS4ERR_DELAY retries) can outlive the drain. destroy path rpciod ------------ ------ unhash_session(ses) nfsd4_probe_callback_sync(clp) flush_workqueue(cl_callback_wq) /* returns; rpc_task still live */ nfsd4_put_session_locked(ses) free_session(ses) -> kfree(ses) nfsd4_cb_sequence_done() reads cb_clp->cl_cb_session /* freed slab */ A second window exists in nfsd4_process_cb_update(). When __nfsd4_find_backchannel() returns NULL because unhash_session() has already removed the destroyed session from cl_sessions, setup_callback_client() takes the v4.1 early return so clp->cl_cb_session = ses never fires and the field retains a pointer to the about-to-be-freed session. Fix both by converting cl_cb_session to an RCU-protected pointer: - Move the cl_cb_session = ses assignment in setup_callback_client() to after rpc_create() succeeds, so it is only published when a working backchannel exists. Clear cl_cb_session on the error return in nfsd4_process_cb_update(). Both stores use rcu_assign_pointer(). - Annotate cl_cb_session with __rcu. All rpciod-side readers use rcu_read_lock()/rcu_dereference() and check for NULL, bailing to the appropriate error or requeue path: encode_cb_sequence4args(), decode_cb_sequence4resok(), nfsd41_cb_get_slot(), nfsd41_cb_release_slot(), nfsd4_cb_prepare(), and nfsd4_cb_sequence_done(). - Switch __free_session() from kfree() to kfree_rcu() so the session slab is not reclaimed until after an RCU grace period, guaranteeing that rpciod readers inside rcu_read_lock() never dereference freed memory. - Pass the session pointer to the nfsd_cb_seq_status and nfsd_cb_free_slot tracepoints instead of having them re-read cl_cb_session. - nfsd4_cb_prepare() calls rpc_exit() when the session is NULL, routing through the done/release path to requeue the callback.