user/sven/linux.git/include, branch v6.17.3

mm/ksm: fix flag-dropping behavior in ksm_madvise

2025-10-15T10:04:19Z

commit f04aad36a07cc17b7a5d5b9a2d386ce6fae63e93 upstream. syzkaller discovered the following crash: (kernel BUG) [ 44.607039] ------------[ cut here ]------------ [ 44.607422] kernel BUG at mm/userfaultfd.c:2067! [ 44.608148] Oops: invalid opcode: 0000 [#1] SMP DEBUG_PAGEALLOC KASAN NOPTI [ 44.608814] CPU: 1 UID: 0 PID: 2475 Comm: reproducer Not tainted 6.16.0-rc6 #1 PREEMPT(none) [ 44.609635] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.16.3-0-ga6ed6b701f0a-prebuilt.qemu.org 04/01/2014 [ 44.610695] RIP: 0010:userfaultfd_release_all+0x3a8/0x460 [ 44.617726] Call Trace: [ 44.617926] [ 44.619284] userfaultfd_release+0xef/0x1b0 [ 44.620976] __fput+0x3f9/0xb60 [ 44.621240] fput_close_sync+0x110/0x210 [ 44.622222] __x64_sys_close+0x8f/0x120 [ 44.622530] do_syscall_64+0x5b/0x2f0 [ 44.622840] entry_SYSCALL_64_after_hwframe+0x76/0x7e [ 44.623244] RIP: 0033:0x7f365bb3f227 Kernel panics because it detects UFFD inconsistency during userfaultfd_release_all(). Specifically, a VMA which has a valid pointer to vma->vm_userfaultfd_ctx, but no UFFD flags in vma->vm_flags. The inconsistency is caused in ksm_madvise(): when user calls madvise() with MADV_UNMEARGEABLE on a VMA that is registered for UFFD in MINOR mode, it accidentally clears all flags stored in the upper 32 bits of vma->vm_flags. Assuming x86_64 kernel build, unsigned long is 64-bit and unsigned int and int are 32-bit wide. This setup causes the following mishap during the &= ~VM_MERGEABLE assignment. VM_MERGEABLE is a 32-bit constant of type unsigned int, 0x8000'0000. After ~ is applied, it becomes 0x7fff'ffff unsigned int, which is then promoted to unsigned long before the & operation. This promotion fills upper 32 bits with leading 0s, as we're doing unsigned conversion (and even for a signed conversion, this wouldn't help as the leading bit is 0). & operation thus ends up AND-ing vm_flags with 0x0000'0000'7fff'ffff instead of intended 0xffff'ffff'7fff'ffff and hence accidentally clears the upper 32-bits of its value. Fix it by changing `VM_MERGEABLE` constant to unsigned long, using the BIT() macro. Note: other VM_* flags are not affected: This only happens to the VM_MERGEABLE flag, as the other VM_* flags are all constants of type int and after ~ operation, they end up with leading 1 and are thus converted to unsigned long with leading 1s. Note 2: After commit 31defc3b01d9 ("userfaultfd: remove (VM_)BUG_ON()s"), this is no longer a kernel BUG, but a WARNING at the same place: [ 45.595973] WARNING: CPU: 1 PID: 2474 at mm/userfaultfd.c:2067 but the root-cause (flag-drop) remains the same. [akpm@linux-foundation.org: rust bindgen wasn't able to handle BIT(), from Miguel] Link: https://lore.kernel.org/oe-kbuild-all/202510030449.VfSaAjvd-lkp@intel.com/ Link: https://lkml.kernel.org/r/20251001090353.57523-2-acsjakub@amazon.de Fixes: 7677f7fd8be7 ("userfaultfd: add minor fault registration mode") Signed-off-by: Jakub Acs Signed-off-by: Miguel Ojeda Acked-by: David Hildenbrand Acked-by: SeongJae Park Tested-by: Alice Ryhl Tested-by: Miguel Ojeda Cc: Xu Xin Cc: Chengming Zhou Cc: Peter Xu Cc: Axel Rasmussen Cc: Signed-off-by: Andrew Morton Signed-off-by: Greg Kroah-Hartman

nfs/localio: avoid issuing misaligned IO using O_DIRECT

2025-10-15T10:04:13Z

[ Upstream commit 25ba2b84c38f624151a3ba36e56d41c39b9223ad ] Add nfsd_file_dio_alignment and use it to avoid issuing misaligned IO using O_DIRECT. Any misaligned DIO falls back to using buffered IO. Because misaligned DIO is now handled safely, remove the nfs modparam 'localio_O_DIRECT_semantics' that was added to require users opt-in to the requirement that all O_DIRECT be properly DIO-aligned. Also, introduce nfs_iov_iter_aligned_bvec() which is a variant of iov_iter_aligned_bvec() that also verifies the offset associated with an iov_iter is DIO-aligned. NOTE: in a parallel effort, iov_iter_aligned_bvec() is being removed along with iov_iter_is_aligned(). Lastly, add pr_info_ratelimited if underlying filesystem returns -EINVAL because it was made to try O_DIRECT for IO that is not DIO-aligned (shouldn't happen, so its best to be louder if it does). Fixes: 3feec68563d ("nfs/localio: add direct IO enablement with sync and async IO support") Signed-off-by: Mike Snitzer Reviewed-by: Jeff Layton Signed-off-by: Anna Schumaker Signed-off-by: Sasha Levin

NFSD: filecache: add STATX_DIOALIGN and STATX_DIO_READ_ALIGN support

2025-10-15T10:04:13Z

[ Upstream commit d11f6cd1bb4a416b4515702d020a7480ac667f0f ] Use STATX_DIOALIGN and STATX_DIO_READ_ALIGN to get DIO alignment attributes from the underlying filesystem and store them in the associated nfsd_file. This is done when the nfsd_file is first opened for each regular file. Signed-off-by: Mike Snitzer Reviewed-by: Jeff Layton Reviewed-by: NeilBrown Signed-off-by: Chuck Lever Acked-by: Chuck Lever Signed-off-by: Anna Schumaker Stable-dep-of: 25ba2b84c38f ("nfs/localio: avoid issuing misaligned IO using O_DIRECT") Signed-off-by: Sasha Levin

bonding: fix xfrm offload feature setup on active-backup mode

2025-10-15T10:04:13Z

[ Upstream commit 5b66169f6be4847008c0aea50885ff0632151479 ] The active-backup bonding mode supports XFRM ESP offload. However, when a bond is added using command like `ip link add bond0 type bond mode 1 miimon 100`, the `ethtool -k` command shows that the XFRM ESP offload is disabled. This occurs because, in bond_newlink(), we change bond link first and register bond device later. So the XFRM feature update in bond_option_mode_set() is not called as the bond device is not yet registered, leading to the offload feature not being set successfully. To resolve this issue, we can modify the code order in bond_newlink() to ensure that the bond device is registered first before changing the bond link parameters. This change will allow the XFRM ESP offload feature to be correctly enabled. Fixes: 007ab5345545 ("bonding: fix feature flag setting at init time") Signed-off-by: Hangbin Liu Link: https://patch.msgid.link/20250925023304.472186-1-liuhangbin@gmail.com Signed-off-by: Paolo Abeni Signed-off-by: Sasha Levin

scsi: ufs: core: Fix data race in CPU latency PM QoS request handling

2025-10-15T10:04:08Z

[ Upstream commit 79dde5f7dc7c038eec903745dc1550cd4139980e ] The cpu_latency_qos_add/remove/update_request interfaces lack internal synchronization by design, requiring the caller to ensure thread safety. The current implementation relies on the 'pm_qos_enabled' flag, which is insufficient to prevent concurrent access and cannot serve as a proper synchronization mechanism. This has led to data races and list corruption issues. A typical race condition call trace is: [Thread A] ufshcd_pm_qos_exit() --> cpu_latency_qos_remove_request() --> cpu_latency_qos_apply(); --> pm_qos_update_target() --> plist_del <--(1) delete plist node --> memset(req, 0, sizeof(*req)); --> hba->pm_qos_enabled = false; [Thread B] ufshcd_devfreq_target --> ufshcd_devfreq_scale --> ufshcd_scale_clks --> ufshcd_pm_qos_update <--(2) pm_qos_enabled is true --> cpu_latency_qos_update_request --> pm_qos_update_target --> plist_del <--(3) plist node use-after-free Introduces a dedicated mutex to serialize PM QoS operations, preventing data races and ensuring safe access to PM QoS resources, including sysfs interface reads. Fixes: 2777e73fc154 ("scsi: ufs: core: Add CPU latency QoS support for UFS driver") Signed-off-by: Zhongqiu Han Reviewed-by: Bart Van Assche Tested-by: Huan Tang Signed-off-by: Martin K. Petersen Signed-off-by: Sasha Levin

coresight: Avoid enable programming clock duplicately

2025-10-15T10:04:07Z

[ Upstream commit d091c6312561821f216ced63a7ad17c946b6d335 ] The programming clock is enabled by AMBA bus driver before a dynamic probe. As a result, a CoreSight driver may redundantly enable the same clock. To avoid this, add a check for device type and skip enabling the programming clock for AMBA devices. The returned NULL pointer will be tolerated by the drivers. Fixes: 73d779a03a76 ("coresight: etm4x: Change etm4_platform_driver driver for MMIO devices") Reviewed-by: Anshuman Khandual Reviewed-by: Yeoreum Yun Tested-by: James Clark Signed-off-by: Leo Yan Signed-off-by: Suzuki K Poulose Link: https://lore.kernel.org/r/20250731-arm_cs_fix_clock_v4-v6-6-1dfe10bb3f6f@arm.com Signed-off-by: Sasha Levin

coresight: Appropriately disable programming clocks

2025-10-15T10:04:06Z

[ Upstream commit 1abc1b212effe920f4729353880c8e03f1d76b4b ] Some CoreSight components have programming clocks (pclk) and are enabled using clk_get() and clk_prepare_enable(). However, in many cases, these clocks are not disabled when modules exit and only released by clk_put(). To fix the issue, this commit refactors programming clock by replacing clk_get() and clk_prepare_enable() with devm_clk_get_optional_enabled() for enabling APB clock. If the "apb_pclk" clock is not found, a NULL pointer is returned, and the function proceeds to attempt enabling the "apb" clock. Since ACPI platforms rely on firmware to manage clocks, returning a NULL pointer in this case leaves clock management to the firmware rather than the driver. This effectively avoids a clock imbalance issue during module removal - where the clock could be disabled twice: once during the ACPI runtime suspend and again during the devm resource release. Callers are updated to reuse the returned error value. With the change, programming clocks are managed as resources in driver model layer, allowing clock cleanup to be handled automatically. As a result, manual cleanup operations are no longer needed and are removed from the Coresight drivers. Fixes: 73d779a03a76 ("coresight: etm4x: Change etm4_platform_driver driver for MMIO devices") Reviewed-by: Yeoreum Yun Tested-by: James Clark Signed-off-by: Leo Yan Signed-off-by: Suzuki K Poulose Link: https://lore.kernel.org/r/20250731-arm_cs_fix_clock_v4-v6-4-1dfe10bb3f6f@arm.com Signed-off-by: Sasha Levin

crypto: comp - Use same definition of context alloc and free ops

2025-10-15T10:04:04Z

[ Upstream commit f75f66683ded09f7135aef2e763c245a07c8271a ] In commit 42d9f6c77479 ("crypto: acomp - Move scomp stream allocation code into acomp"), the crypto_acomp_streams struct was made to rely on having the alloc_ctx and free_ctx operations defined in the same order as the scomp_alg struct. But in that same commit, the alloc_ctx and free_ctx members of scomp_alg may be randomized by structure layout randomization, since they are contained in a pure ops structure (containing only function pointers). If the pointers within scomp_alg are randomized, but those in crypto_acomp_streams are not, then the order may no longer match. This fixes the problem by removing the union from scomp_alg so that both crypto_acomp_streams and scomp_alg will share the same definition of alloc_ctx and free_ctx, ensuring they will always have the same layout. Signed-off-by: Dan Moulding Suggested-by: Herbert Xu Fixes: 42d9f6c77479 ("crypto: acomp - Move scomp stream allocation code into acomp") Signed-off-by: Herbert Xu Signed-off-by: Sasha Levin

HID: hidraw: tighten ioctl command parsing

2025-10-15T10:04:00Z

[ Upstream commit 75d5546f60b36900051d75ee623fceccbeb6750c ] The handling for variable-length ioctl commands in hidraw_ioctl() is rather complex and the check for the data direction is incomplete. Simplify this code by factoring out the various ioctls grouped by dir and size, and using a switch() statement with the size masked out, to ensure the rest of the command is correctly matched. Fixes: 9188e79ec3fd ("HID: add phys and name ioctls to hidraw") Reported-by: Arnd Bergmann Signed-off-by: Benjamin Tissoires Signed-off-by: Jiri Kosina Signed-off-by: Sasha Levin

ipv6: snmp: do not use SNMP_MIB_SENTINEL anymore

2025-10-15T10:03:54Z

[ Upstream commit ceac1fb2290d230eb83aff3761058c559440de13 ] Use ARRAY_SIZE(), so that we know the limit at compile time. Following patch needs this preliminary change. Signed-off-by: Eric Dumazet Reviewed-by: Sabrina Dubroca Link: https://patch.msgid.link/20250905165813.1470708-3-edumazet@google.com Signed-off-by: Jakub Kicinski Stable-dep-of: 2fab94bcf313 ("ipv6: snmp: do not track per idev ICMP6_MIB_RATELIMITHOST") Signed-off-by: Sasha Levin