kernel/linux.git, branch v7.2.10

Linux 7.2.10

2026-10-11T13:27:50+00:00

Link: https://lore.kernel.org/r/20261009140946.008206689@linuxfoundation.org Tested-by: Brett A C Sheffield Tested-by: Ronald Warsow Tested-by: Peter Schneider Tested-by: Ron Economos Tested-by: Takeshi Ogasawara Tested-by: Salvatore Bonaccorso Tested-by: Slade Watkins Tested-by: Barry K. Nathan Signed-off-by: Greg Kroah-Hartman

ipv4: free inet_opt and ireq_opt after an RCU grace period

2026-10-11T13:27:50+00:00

commit e3f33b0a1d89d738e2b913dab519226eaf76e15b upstream. tcp_v4_syn_recv_sock() transfers ownership of ireq->ireq_opt to the child socket (newinet->inet_opt) without copying it. Another cpu can concurrently process a retransmitted SYN for the same request socket, and send a SYNACK from tcp_check_req(). tcp_v4_send_synack() and inet_csk_route_req() read ireq->ireq_opt under rcu_read_lock() only, and ip_build_and_send_pkt() and ip_options_build() then read opt->optlen twice. Note that the SYNACK timer itself is not an issue: it holds a reference on its request socket, and inet_csk_reqsk_queue_drop() calls timer_delete_sync() before the child can be freed. Since commit 079096f103fa ("tcp/dccp: install syn_recv requests into ehash table"), request sockets are processed without holding the listener lock, so nothing prevents the child socket from being freed while the SYNACK is still being built. TCP child sockets do not have SOCK_RCU_FREE, and inet_sock_destruct() frees inet_opt with a plain kfree(), leading to a use-after-free in ip_options_build(). A similar issue exists with request socket migration (net.ipv4.tcp_migrate_req=1, or a BPF_SK_REUSEPORT_SELECT_OR_MIGRATE program). reqsk_timer_handler() clones the request socket with inet_reqsk_clone(), so that the old request socket and its clone share the same ireq_opt, then reqsk_migrate_reset() clears the pointer in the old one. Another cpu holding a reference on the old request socket can still be using these options (sending a SYNACK, or creating a child in tcp_v4_syn_recv_sock()) when the clone is freed, and tcp_v4_reqsk_destructor() also uses a plain kfree(). Readers already use RCU, and other paths replacing inet_opt (do_ip_setsockopt(), cipso_v4_sock_setattr()...) already use kfree_rcu(). Use kfree_rcu() in inet_sock_destruct() and tcp_v4_reqsk_destructor() as well. IPv6 is not affected by the first issue, because tcp_v6_syn_recv_sock() duplicates the options. tcp_v6_reqsk_destructor() has the same migration issue with ipv6_opt, which is only set by CALIPSO. This will be addressed in a separate patch. Fixes: 079096f103fa ("tcp/dccp: install syn_recv requests into ehash table") Fixes: c905dee62232 ("tcp: Migrate TCP_NEW_SYN_RECV requests at retransmitting SYN+ACKs.") Reported-by: Xinyang Ge Signed-off-by: Eric Dumazet Reviewed-by: Kuniyuki Iwashima Link: https://patch.msgid.link/20261001221253.2964024-1-edumazet@kernel.org Signed-off-by: Jakub Kicinski Signed-off-by: Greg Kroah-Hartman

iio: buffer-dmaengine: fix sg entry iteration when building dma_vecs

2026-10-11T13:27:50+00:00

commit 0d0ffbcc92e30a8d656ad76a6e278fa3400fed85 upstream. iio_dmaengine_buffer_submit_block() counts scatterlist entries with sg_nents_for_len(), which walks the CPU-side lengths (sg->length), but then consumes the DMA-side fields (sg_dma_address()/sg_dma_len()). After dma_map_sgtable() the two views may differ: an IOMMU can coalesce the mapping so that only the first sgt->nents entries carry valid DMA addresses, with nents < orig_nents. On x86 with an IOMMU enabled, a DMABUF block backed by two 1 MiB system-heap chunks maps to a single 2 MiB IOVA range. The CPU-side count is 2, so the loop reads one entry past the mapped set and emits a garbage vec ({addr = ~0, len = 0}). The DMA engine driver rejects the vec array (prep returns NULL), the fence is signalled with -ENOMEM, which a userspace poller cannot observe, and the block is left in ACTIVE state so every further enqueue of it fails with -EBUSY. The visible symptom is a stream of zero-filled blocks followed by a wedged buffer. Platforms without an IOMMU never hit this because nents == orig_nents. Size the vec array with sgt->nents, i.e. the DMA-mapped view, and stop the fill loop once bytes_used is covered - which is allowed to be smaller than the block size - passing the number of vecs actually filled to dmaengine_prep_peripheral_dma_vec(). One vec per mapped entry is enough since coalescing can only ever reduce the number of entries. A single mapped entry longer than the device's maximum segment size would need more than one, but the DMA API already assumes no single segment exceeds it [1], and splitting a vec down to the hardware descriptor size is the DMA engine driver's job - which both current .device_prep_peripheral_dma_vec() implementations do. [1]: commit ab2cbeb0ed30 ("iommu/dma: Handle SG length overflow better") Assisted-by: Claude:claude-fable-5 Fixes: 7a86d469983a ("iio: buffer-dmaengine: Support new DMABUF based userspace API") Signed-off-by: Michael Hennerich Signed-off-by: Nuno Sá Cc: stable@vger.kernel.org Signed-off-by: Jonathan Cameron [ sashal: Adjusted context for the missing IIO DMAengine cyclic-transfer implementation. ] Signed-off-by: Sasha Levin Signed-off-by: Greg Kroah-Hartman

xen/netfront: drop RX packets with a short Ethernet header

2026-10-11T13:27:50+00:00

commit 089e58805c452e52179482b1025a8e309a57f801 upstream. handle_incoming_queue() pulls pull_to bytes into the head before calling eth_type_trans(). pull_to is the length of the first RX slot, capped at RX_COPY_THRESHOLD, and that length comes from the backend. Nothing checks it against ETH_HLEN. If the first slot is shorter than ETH_HLEN and more slots follow, the head ends up shorter than an Ethernet header while skb->len is longer, and eth_type_trans() BUG()s in __skb_pull(). If the whole packet is shorter than ETH_HLEN, eth_type_trans() reads the header past the end of the data instead. Pull at least ETH_HLEN, and drop the packet if that fails, which also drops packets too short to hold an Ethernet header. This also checks the return value of the pull, which was ignored. Fixes: 0d160211965b ("xen: add virtual network device driver") Cc: stable@vger.kernel.org Signed-off-by: Josef Bacik Link: https://patch.msgid.link/20261007-b4-xen-netfront-short-head-v1-1-12d7113a7e4e@toxicpanda.com Signed-off-by: Jakub Kicinski Signed-off-by: Greg Kroah-Hartman

net: usb: qmi_wwan: add Quectel EG120K-EA

2026-10-11T13:27:50+00:00

commit 0f2fd31f63c65c409bd336af3a42d478a9207c1f upstream. Add support for the Quectel EG120K-EA LTE Cat.12 module (USB ID 2c7c:030b). Its QMI interface (interface 4) uses class/subclass/protocol ff/ff/ff like the other recent Quectel modules, so match it the same way. The product ID is shared with the EM060K, which the option driver already knows and uses to claim the serial interfaces. Without a qmi_wwan entry the data interface is left unbound. Tested on a GL.iNet GL-X2000, where the module is soldered down and enumerates at SuperSpeed: cdc-wdm0 and wwan0 appear and ModemManager brings up a data session. T: Bus=02 Lev=01 Prnt=01 Port=00 Cnt=01 Dev#= 2 Spd=5000 MxCh= 0 D: Ver= 3.10 Cls=00(>ifc ) Sub=00 Prot=00 MxPS= 9 #Cfgs= 1 P: Vendor=2c7c ProdID=030b Rev= 5.04 S: Manufacturer=Quectel S: Product=EG120K-EA S: SerialNumber=45546267 C:* #Ifs= 6 Cfg#= 1 Atr=a0 MxPwr=896mA I:* If#= 0 Alt= 0 #EPs= 2 Cls=ff(vend.) Sub=ff Prot=30 Driver=option E: Ad=01(O) Atr=02(Bulk) MxPS=1024 Ivl=0ms E: Ad=81(I) Atr=02(Bulk) MxPS=1024 Ivl=0ms I:* If#= 1 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=40 Driver=option E: Ad=83(I) Atr=03(Int.) MxPS= 10 Ivl=32ms E: Ad=82(I) Atr=02(Bulk) MxPS=1024 Ivl=0ms E: Ad=02(O) Atr=02(Bulk) MxPS=1024 Ivl=0ms I:* If#= 2 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=40 Driver=option E: Ad=85(I) Atr=03(Int.) MxPS= 10 Ivl=32ms E: Ad=84(I) Atr=02(Bulk) MxPS=1024 Ivl=0ms E: Ad=03(O) Atr=02(Bulk) MxPS=1024 Ivl=0ms I:* If#= 3 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=40 Driver=option E: Ad=87(I) Atr=03(Int.) MxPS= 10 Ivl=32ms E: Ad=86(I) Atr=02(Bulk) MxPS=1024 Ivl=0ms E: Ad=04(O) Atr=02(Bulk) MxPS=1024 Ivl=0ms I:* If#= 4 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=ff Driver=qmi_wwan E: Ad=88(I) Atr=03(Int.) MxPS= 8 Ivl=32ms E: Ad=8e(I) Atr=02(Bulk) MxPS=1024 Ivl=0ms E: Ad=0f(O) Atr=02(Bulk) MxPS=1024 Ivl=0ms I:* If#=12 Alt= 0 #EPs= 1 Cls=ff(vend.) Sub=ff Prot=70 Driver=(none) E: Ad=89(I) Atr=02(Bulk) MxPS=1024 Ivl=0ms Signed-off-by: Gilberto Conde Link: https://patch.msgid.link/20260921093426.2870266-1-gilbertorconde@gmail.com Signed-off-by: Jakub Kicinski Signed-off-by: Greg Kroah-Hartman

smb: client: only require read lease for size-extending preallocate

2026-10-11T13:27:50+00:00

[ Upstream commit b5d59e7270d078ab28a9c4b4d11dc3eb24671988 ] smb3_simple_falloc() refuses any fallocate that clears FALLOC_FL_KEEP_SIZE with -EOPNOTSUPP whenever the inode is not read caching: /* if file not oplocked can't be sure whether asking to extend size */ if (!CIFS_CACHE_READ(cifsi)) if (!keep_size) { ... return rc; } As with smb3_zero_range(), the read lease is only needed to trust the cached size when deciding whether the request extends the file. When it is not held, the size can instead be fetched from the server, which is authoritative, rather than refusing the request outright: after flushing, query the server's end of file and take the larger of it and the cached size for the interior-vs-extend decision. The larger of the two is used because the server's end of file reflects another client's growth while the cached size reflects this client's own writes that may not have reached the server yet; using the server size alone would wrongly treat an interior request as extending when a range flush left an extending write unwritten. smb3_simple_fallocate_range(), which performs the actual interior emulation, decided whether the range already lies past EOF from its own i_size_read(inode) rather than the old_eof computed above. In the leaseless case, that is exactly the stale, undershooting size this patch works around: a genuinely interior range can read as past EOF by that stale count, which skips the FSCTL_QUERY_ALLOCATED_RANGES check entirely and overwrites already-allocated server data with zeroes instead of only filling the holes. Pass old_eof into smb3_simple_fallocate_range() and use it for that comparison instead of re-deriving a second, inconsistent one. Rejecting interior requests is observed as generic/363 randomly failing against Windows Server with do_preallocate: fallocate: Operation not supported fsx issues an interior, non-KEEP_SIZE preallocate while the inode is transiently not read caching: the server had just downgraded the file's lease from RWH to RH after breaking the write caching, and the ensuing handle reopen/revalidation left CIFS_CACHE_READ momentarily clear. The range sat within the server's end of file, so no extend was needed, yet it was refused and fsx aborted. This keeps the emulation correct even when a genuine lease break from another client leaves the inode without read caching -- the case the -EOPNOTSUPP guard turned into a hard failure. The extra flush and round trip only happen when both keep_size is false and no read lease is held; every other case is unchanged. Fixes: 9ccf3216238c ("Add support for original fallocate") Reviewed-by: David Howells Reviewed-by: Namjae Jeon Signed-off-by: Paulo Alcantara Cc: Christian Brauner Cc: Matthew Wilcox Cc: Ronnie Sahlberg Cc: Shyam Prasad N Cc: Tom Talpey Cc: Bharath SM Cc: stable@vger.kernel.org [ sashal: Adapted cifs_resize_file_locked() calls to the two-argument helper. ] Signed-off-by: Sasha Levin Signed-off-by: Greg Kroah-Hartman

arm64: errata: Add Cortex-A725 erratum 3821522 workaround

2026-10-11T13:27:50+00:00

[ Upstream commit 45f7304679875eb000d67e194090ee7abd05c89c ] Cortex-A725 erratum 3821522 affects the CNT_CYCLES event, which can incur a significant increment error when a CPU enters and subsequently exits WFE or WFI, and may no longer track the system counter frequency. The AMEVCNTR01_EL0 counter is used as the AMU constant counter for frequency invariance and CPPC FFH feedback counters. Wire the affected Cortex-A725 range into the shared broken AMU constant-counter capability so the affected counter is treated as unavailable by returning zero in the AMU counter paths. This prevents the broken counter from being used as a reference source. The erratum can also affect PMUv3 users of the CNT_CYCLES event, but this workaround intentionally does not change PMU event handling. Hiding or rejecting the PMU event from the erratum code would change the perf-visible PMU event interface, including raw event selection, and would need a separate PMU-specific approach rather than being folded into the AMU reference-counter workaround. Cc: stable@vger.kernel.org Signed-off-by: Beata Michalska Reviewed-by: Vladimir Murzin Signed-off-by: Will Deacon Signed-off-by: Sasha Levin Signed-off-by: Greg Kroah-Hartman

arm64: errata: Factor out broken AMU const counter cap

2026-10-11T13:27:49+00:00

[ Upstream commit 573371caf7aea8582212da81dcead2c3f48fac4b ] Move the workaround from the erratum 2457168-specific cpucap to a generic broken AMU constant-counter one. This keeps the existing Cortex-A510 handling unchanged while allowing other errata with similar AMU constant counter issue to share the capability bit and call sites. Signed-off-by: Beata Michalska Reviewed-by: Vladimir Murzin Signed-off-by: Will Deacon Omit the comment update in amu_read_core_const_ctrs(), since paired CPPC FFH counter reads are not implemented here. Keep the capability rename in the existing counter-read and initialization paths, together with the shared MIDR range list needed by the additional erratum. Stable-dep-of: 45f730467987 ("arm64: errata: Add Cortex-A725 erratum 3821522 workaround") Signed-off-by: Sasha Levin Signed-off-by: Greg Kroah-Hartman

perf: Require kernel access for text poke events

2026-10-11T13:27:49+00:00

[ Upstream commit 357e8a77a501d96c9517f01f4a211eed5e9c9184 ] Perf events with exclude_kernel=1 can be opened without kernel perf access. However, exclude_kernel does not suppress text-poke sideband records. Every PERF_RECORD_TEXT_POKE is marked PERF_RECORD_MISC_KERNEL and contains a raw kernel instruction address. An unprivileged task can therefore open and mmap a task-local software event with text_poke=1. Both opening a count-only tracepoint event and configuring UDP GRO for ESP-in-UDP cause updates to inline static calls; the observer receives the relocated addresses of the modified instructions. For a known kernel image, any such address reveals the runtime kernel text base despite KASLR. Call perf_allow_kernel() whenever attr.text_poke is set, regardless of exclude_kernel. Events that neither monitor kernel execution nor request text-poke records retain their existing permissions. Fixes: e17d43b93e54 ("perf: Add perf text poke event") Assisted-by: LLM Signed-off-by: Zhengchuan Liang Signed-off-by: Peter Zijlstra (Intel) Cc: stable@vger.kernel.org Link: https://patch.msgid.link/99131354c41e23188f778b92f90363775b482395.1790573390.git.zcliangcn@gmail.com Signed-off-by: Sasha Levin Signed-off-by: Greg Kroah-Hartman

perf/core: Check kernel access when kernel callchains are requested

2026-10-11T13:27:49+00:00

[ Upstream commit a4573a3838ae4fc73b70019cfa1dac9aaea7cc2f ] perf_event_open() currently gates perf_allow_kernel() only on !attr.exclude_kernel. However, users can still request kernel callchain collection with attr.exclude_callchain_kernel == 0 even when attr.exclude_kernel == 1. That still requires kernel profiling privilege, but the existing check does not enforce it. Update the permission check to call perf_allow_kernel() when either kernel sampling is requested or kernel callchains are requested. This keeps permission checks aligned with requested data and prevents unprivileged use of kernel callchain capture. Signed-off-by: Dapeng Mi Signed-off-by: Peter Zijlstra (Intel) Link: https://patch.msgid.link/20260616044654.3468742-9-dapeng1.mi@linux.intel.com Stable-dep-of: 357e8a77a501 ("perf: Require kernel access for text poke events") Signed-off-by: Sasha Levin Signed-off-by: Greg Kroah-Hartman