6.12.60-rc1 review

This is the start of the stable review cycle for the 6.12.60 release. There are 112 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let me know.

Responses should be made by Sat, 29 Nov 2025 14:40:08 +0000. Anything received after that time might be too late.

The whole patch series can be found in one patch at:

	https://www.kernel.org/pub/linux/kernel/v6.x/stable-review/patch-6.12.60-rc1.gz

or in the git tree and branch at:
git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git linux-6.12.y
and the diffstat can be found below.

thanks,

greg k-h

-------------
Pseudo-Shortlog of commits:

Greg Kroah-Hartman <[email protected]>

    Linux 6.12.60-rc1

Fangzhi Zuo <[email protected]>
drm/amd/display: Prevent Gating DTBCLK before It Is Properly Latched

Charlene Liu <[email protected]>
drm/amd/display: Insert dccg log for easy debug

Charlene Liu <[email protected]>
drm/amd/display: disable DPP RCG before DPP CLK enable

Charlene Liu <[email protected]>
drm/amd/display: avoid reset DTBCLK at clock init

Darrick J. Wong <[email protected]>
xfs: fix out of bounds memory read error in symlink repair

Marcelo Moreira <[email protected]>
xfs: Replace strncpy with memcpy

Eric Dumazet <[email protected]>
mptcp: fix a race in mptcp_pm_del_add_timer()

Imre Deak <[email protected]>
drm/i915/dp_mst: Disable Panel Replay

Martin Kaiser <[email protected]>
maple_tree: fix tracepoint string pointers

Jari Ruusu <[email protected]>
tty/vt: fix up incorrect backport to stable releases

Henrique Carvalho <[email protected]>
smb: client: fix incomplete backport in cfids_invalidation_worker()

Samuel Zhang <[email protected]>
drm/amdgpu: fix gpu page fault after hibernation on PF passthrough

Zhang Chujun <[email protected]>
tracing/tools: Fix incorrcet short option in usage text for --threads

Nishanth Menon <[email protected]>
net: ethernet: ti: netcp: Standardize knav_dma_open_channel to return NULL on error

René Rebe <[email protected]>
ALSA: usb-audio: fix uac2 clock source at terminal parser

Heiko Carstens <[email protected]>
s390/mm: Fix __ptep_rdp() inline assembly

Shuicheng Lin <[email protected]>
drm/xe: Prevent BIT() overflow when handling invalid prefetch region

Wentao Guan <[email protected]>
Revert "RDMA/irdma: Update Kconfig"

Marc Zyngier <[email protected]>
KVM: arm64: Make all 32bit ID registers fully writable

Takashi Iwai <[email protected]>
ALSA: usb-audio: Fix missing unlock at error path of maxpacksize check

Jakub Horký <[email protected]>
kconfig/nconf: Initialize the default locale at startup

Jakub Horký <[email protected]>
kconfig/mconf: Initialize the default locale at startup

Shahar Shitrit <[email protected]>
net: tls: Cancel RX async resync request on rcd_delta overflow

Carlos Llamas <[email protected]>
blk-crypto: use BLK_STS_INVAL for alignment errors

Shahar Shitrit <[email protected]>
net: tls: Change async resync helpers argument

Po-Hsu Lin <[email protected]>
selftests: net: use BASH for bareudp testing

Borislav Petkov (AMD) <[email protected]>
x86/microcode/AMD: Limit Entrysign signature checking to known generations

Bart Van Assche <[email protected]>
scsi: core: Fix a regression triggered by scsi_host_busy()

Steve French <[email protected]>
cifs: fix typo in enable_gcm_256 module parameter

Rafał Miłecki <[email protected]>
bcma: don't register devices disabled in OF

Michal Luczaj <[email protected]>
vsock: Ignore signal/timeout on connect() if already established

Shaurya Rane <[email protected]>
cifs: fix memory leak in smb3_fs_context_parse_param error path

Thomas Weißschuh <[email protected]>
LoongArch: Use UAPI types in ptrace UAPI header

Kuniyuki Iwashima <[email protected]>
af_unix: Read sk_peek_offset() again after sleeping in unix_stream_read_generic().

Kuniyuki Iwashima <[email protected]>

    af_unix: Cache state->msg in unix_stream_read_generic().

Pradyumn Rahar <[email protected]>
net/mlx5: Clean up only new IRQ glue on request_irq() failure

Shay Drory <[email protected]>
devlink: rate: Unset parent pointer in devl_rate_nodes_destroy

Jared Kangas <[email protected]>
pinctrl: s32cc: initialize gpio_pin_config::list after kmalloc()

Jared Kangas <[email protected]>
pinctrl: s32cc: fix uninitialized memory in s32_pinctrl_desc

Grzegorz Nitka <[email protected]>
ice: fix PTP cleanup on driver removal in error path

Emil Tantilov <[email protected]>
idpf: fix possible vport_config NULL pointer deref in remove

Pavel Zhigulin <[email protected]>
net: qlogic/qede: fix potential out-of-bounds read in qede_tpa_cont() and qede_tpa_end()

Haotian Zhang <[email protected]>
platform/x86/intel/speed_select_if: Convert PCIBIOS_* return codes to errnos

Ido Schimmel <[email protected]>
selftests: net: lib: Do not overwrite error messages

Aleksei Nikiforov <[email protected]>
s390/ctcm: Fix double-kfree

Shin'ichiro Kawasaki <[email protected]>
nvme-multipath: fix lockdep WARN due to partition scan work

Chen Pei <[email protected]>
tools: riscv: Fixed misalignment of CSR related definitions

Ilya Maximets <[email protected]>
net: openvswitch: remove never-working support for setting nsh fields

Pavel Zhigulin <[email protected]>
net: mlxsw: linecards: fix missing error check in mlxsw_linecard_devlink_info_get()

Pavel Zhigulin <[email protected]>
net: dsa: hellcreek: fix missing error handling in LED registration

Prateek Agarwal <[email protected]>
drm/tegra: Add call to put_pid()

Zilin Guan <[email protected]>
mlxsw: spectrum: Fix memory leak in mlxsw_sp_flower_stats()

Armin Wolf <[email protected]>
platform/x86: msi-wmi-platform: Fix typo in WMI GUID

Armin Wolf <[email protected]>
platform/x86: msi-wmi-platform: Only load on MSI devices

Haotian Zhang <[email protected]>
pinctrl: cirrus: Fix fwnode leak in cs42l43_pin_probe()

Jianbo Liu <[email protected]>
xfrm: Prevent locally generated packets from direct output in tunnel mode

Jianbo Liu <[email protected]>
xfrm: Determine inner GSO type from packet inner protocol

Yu-Chun Lin <[email protected]>
pinctrl: realtek: Select REGMAP_MMIO for RTD driver

Sabrina Dubroca <[email protected]>
xfrm: set err and extack on failure to create pcpu SA

Sabrina Dubroca <[email protected]>
xfrm: drop SA reference in xfrm_state_update if dir doesn't match

Ivan Lipski <[email protected]>
drm/amd/display: Clear the CUR_ENABLE register on DCN20 on DPP5

Fangzhi Zuo <[email protected]>
drm/amd/display: Fix pbn to kbps Conversion

Mario Limonciello (AMD) <[email protected]>
drm/amd/display: Move sleep into each retry for retrieve_link_cap()

Mario Limonciello (AMD) <[email protected]>
drm/amd/display: Increase DPCD read retries

Yifan Zha <[email protected]>
drm/amdgpu: Skip emit de meta data on gfx11 with rs64 enabled

Mario Limonciello <[email protected]>
drm/amd: Skip power ungate during suspend for VPE

Robert McClinton <[email protected]>
drm/radeon: delete radeon_fence_process in is_signaled, no deadlock

Ma Ke <[email protected]>
drm/tegra: dc: Fix reference leak in tegra_dc_couple()

Paolo Abeni <[email protected]>
mptcp: do not fallback when OoO is present

Paolo Abeni <[email protected]>
mptcp: decouple mptcp fastclose from tcp close

Paolo Abeni <[email protected]>
mptcp: avoid unneeded subflow-level drops

Matthieu Baerts (NGI0) <[email protected]>
selftests: mptcp: join: userspace: longer timeout

Matthieu Baerts (NGI0) <[email protected]>
selftests: mptcp: join: endpoints: longer timeout

Paolo Abeni <[email protected]>
mptcp: fix premature close in case of fallback

Paolo Abeni <[email protected]>
mptcp: fix duplicate reset on fastclose

Paolo Abeni <[email protected]>
mptcp: fix ack generation for fallback msk

Eric Dumazet <[email protected]>
mptcp: fix race condition in mptcp_schedule_work()

Huacai Chen <[email protected]>
LoongArch: Don't panic if no valid cache info for PCI

Krzysztof Kozlowski <[email protected]>
dt-bindings: pinctrl: toshiba,visconti: Fix number of items in groups

Maciej W. Rozycki <[email protected]>
MIPS: Malta: Fix !EVA SOC-it PCI MMIO

Hamza Mahfooz <[email protected]>
scsi: target: tcm_loop: Fix segfault in tcm_loop_tpg_address_show()

Bart Van Assche <[email protected]>
scsi: sg: Do not sleep in atomic context

Ewan D. Milne <[email protected]>

    nvme: nvme-fc: Ensure ->ioerr_work is cancelled in nvme_fc_delete_ctrl()

Ewan D. Milne <[email protected]>
nvme: nvme-fc: move tagset removal to nvme_fc_delete_ctrl()

Nam Cao <[email protected]>
nouveau/firmware: Add missing kfree() of nvkm_falcon_fw::boot

Vlastimil Babka <[email protected]>
mm/mempool: fix poisoning order>0 pages with HIGHMEM

Seungjin Bae <[email protected]>
Input: pegasus-notetaker - fix potential out-of-bounds access

Dan Carpenter <[email protected]>
Input: imx_sc_key - fix memory corruption on unload

Hans de Goede <[email protected]>
Input: goodix - add support for ACPI ID GDIX1003

Tzung-Bi Shih <[email protected]>
Input: cros_ec_keyb - fix an invalid memory access

Diogo Ivo <[email protected]>
Revert "drm/tegra: dsi: Clear enable register if powered by bootloader"

Oleksij Rempel <[email protected]>
net: dsa: microchip: lan937x: Fix RGMII delay tuning

Andrey Vatoropin <[email protected]>
be2net: pass wrb_params in case of OS2BMC

Yihang Li <[email protected]>
ata: libata-scsi: Add missing scsi_device_put() in ata_scsi_dev_rescan()

Henrique Carvalho <[email protected]>
smb: client: introduce close_cached_dir_locked()

Maciej W. Rozycki <[email protected]>
MIPS: mm: Prevent a TLB shutdown on initial uniquification

Niklas Cassel <[email protected]>
ata: libata-scsi: Fix system suspend for a security locked drive

Jiayuan Chen <[email protected]>
mptcp: Fix proto fallback detection with BPF

Jiayuan Chen <[email protected]>
mptcp: Disallow MPTCP subflows from sockmap

Yongpeng Yang <[email protected]>
exfat: check return value of sb_min_blocksize in exfat_read_boot_sector

Mike Yuan <[email protected]>
shmem: fix tmpfs reconfiguration (remount) when noswap is set

Yongpeng Yang <[email protected]>
isofs: check the return value of sb_min_blocksize() in isofs_fill_super

Dan Carpenter <[email protected]>
mtdchar: fix integer overflow in read/write ioctls

Niravkumar L Rabara <[email protected]>
mtd: rawnand: cadence: fix DMA device NULL pointer dereference

Quentin Schulz <[email protected]>
arm64: dts: rockchip: disable HS400 on RK3588 Tiger

Quentin Schulz <[email protected]>

    arm64: dts: rockchip: include rk3399-base instead of rk3399 in rk3399-op1

Mykola Kvach <[email protected]>
arm64: dts: rockchip: fix PCIe 3.3V regulator voltage on orangepi-5

Diederik de Haas <[email protected]>
arm64: dts: rockchip: Fix vccio4-supply on rk3566-pinetab2

Zhang Heng <[email protected]>

    HID: quirks: work around VID/PID conflict for 0x4c4a/0x4155

Mario Limonciello (AMD) <[email protected]>
HID: amd_sfh: Stop sensor before starting

Yipeng Zou <[email protected]>
timers: Fix NULL function pointer race in timer_shutdown_sync()

Sebastian Ene <[email protected]>
KVM: arm64: Check the untrusted offset in FF-A memory share

-------------

Diffstat:

 .../bindings/pinctrl/toshiba,visconti-pinctrl.yaml |  26 +++---
 Documentation/wmi/driver-development-guide.rst     |   1 +
 Makefile                                           |   4 +-
 arch/arm64/boot/dts/rockchip/rk3399-op1.dtsi       |   2 +-
 arch/arm64/boot/dts/rockchip/rk3566-pinetab2.dtsi  |   2 +-
 arch/arm64/boot/dts/rockchip/rk3588-tiger.dtsi     |   4 +-
 .../arm64/boot/dts/rockchip/rk3588s-orangepi-5.dts |   4 +-
 arch/arm64/kvm/hyp/nvhe/ffa.c                      |   9 +-
 arch/arm64/kvm/sys_regs.c                          |  63 +++++++------
 arch/loongarch/include/uapi/asm/ptrace.h           |  40 ++++----
 arch/loongarch/pci/pci.c                           |   8 +-
 arch/mips/mm/tlb-r4k.c                             | 102 +++++++++++++--------
 arch/mips/mti-malta/malta-init.c                   |  20 ++--
 arch/s390/include/asm/pgtable.h                    |  12 +--
 arch/s390/mm/pgtable.c                             |   4 +-
 arch/x86/kernel/cpu/microcode/amd.c                |  20 +++-
 block/blk-crypto.c                                 |   2 +-
 drivers/ata/libata-scsi.c                          |  11 ++-
 drivers/bcma/main.c                                |   6 ++
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c         |   3 +-
 drivers/gpu/drm/amd/amdgpu/aqua_vanjaram.c         |   3 +-
 drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c             |   4 +-
 drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c            |   4 +-
 .../amd/display/amdgpu_dm/amdgpu_dm_mst_types.c    |  59 +++++-------
 .../amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c   |  20 ++--
 .../gpu/drm/amd/display/dc/dccg/dcn35/dcn35_dccg.c |  60 ++++++++----
 .../drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c    |   8 ++
 .../drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c    |  21 +++--
 .../display/dc/link/protocols/link_dp_capability.c |  11 ++-
 drivers/gpu/drm/i915/display/intel_psr.c           |   4 +
 drivers/gpu/drm/nouveau/nvkm/falcon/fw.c           |   2 +
 drivers/gpu/drm/radeon/radeon_fence.c              |   7 --
 drivers/gpu/drm/tegra/dc.c                         |   1 +
 drivers/gpu/drm/tegra/dsi.c                        |   9 --
 drivers/gpu/drm/tegra/uapi.c                       |   7 +-
 drivers/gpu/drm/xe/xe_vm.c                         |   4 +-
 drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_init.c      |   2 +
 drivers/hid/hid-ids.h                              |   4 +-
 drivers/hid/hid-quirks.c                           |  13 ++-
 drivers/infiniband/hw/irdma/Kconfig                |   7 +-
 drivers/input/keyboard/cros_ec_keyb.c              |   6 ++
 drivers/input/keyboard/imx_sc_key.c                |   2 +-
 drivers/input/tablet/pegasus_notetaker.c           |   9 ++
 drivers/input/touchscreen/goodix.c                 |   1 +
 drivers/mtd/mtdchar.c                              |   6 +-
 drivers/mtd/nand/raw/cadence-nand-controller.c     |   3 +-
 drivers/net/dsa/hirschmann/hellcreek_ptp.c         |  14 ++-
 drivers/net/dsa/microchip/lan937x_main.c           |   1 +
 drivers/net/ethernet/emulex/benet/be_main.c        |   7 +-
 drivers/net/ethernet/intel/ice/ice_ptp.c           |  22 ++++-
 drivers/net/ethernet/intel/idpf/idpf_main.c        |   2 +
 .../ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c |   9 +-
 drivers/net/ethernet/mellanox/mlx5/core/pci_irq.c  |   6 +-
 .../net/ethernet/mellanox/mlxsw/core_linecards.c   |   2 +
 .../net/ethernet/mellanox/mlxsw/spectrum_flower.c  |   6 +-
 drivers/net/ethernet/qlogic/qede/qede_fp.c         |   5 +-
 drivers/net/ethernet/ti/netcp_core.c               |  10 +-
 drivers/nvme/host/fc.c                             |  15 +--
 drivers/nvme/host/multipath.c                      |   2 +-
 drivers/perf/riscv_pmu_sbi.c                       |   2 +-
 drivers/pinctrl/cirrus/pinctrl-cs42l43.c           |  21 ++++-
 drivers/pinctrl/nxp/pinctrl-s32cc.c                |   3 +-
 drivers/pinctrl/realtek/Kconfig                    |   1 +
 drivers/platform/x86/Kconfig                       |   1 +
 .../x86/intel/speed_select_if/isst_if_mmio.c       |   4 +-
 drivers/platform/x86/msi-wmi-platform.c            |  43 ++++++++-
 drivers/s390/net/ctcm_mpc.c                        |   1 -
 drivers/scsi/hosts.c                               |   5 +-
 drivers/scsi/sg.c                                  |  10 +-
 drivers/soc/ti/knav_dma.c                          |  14 +--
 drivers/target/loopback/tcm_loop.c                 |   3 +
 drivers/tty/vt/vt_ioctl.c                          |   4 +-
 fs/exfat/super.c                                   |   5 +-
 fs/isofs/inode.c                                   |   5 +
 fs/smb/client/cached_dir.c                         |  43 ++++++++-
 fs/smb/client/cifsfs.c                             |   2 +-
 fs/smb/client/fs_context.c                         |   4 +
 fs/xfs/scrub/symlink_repair.c                      |   4 +-
 include/linux/ata.h                                |   1 +
 include/net/tls.h                                  |  25 ++---
 include/net/xfrm.h                                 |   3 +-
 kernel/time/timer.c                                |   7 +-
 lib/maple_tree.c                                   |  30 +++---
 mm/mempool.c                                       |  32 +++++--
 mm/shmem.c                                         |  15 ++-
 net/devlink/rate.c                                 |   4 +-
 net/ipv4/esp4_offload.c                            |   6 +-
 net/ipv6/esp6_offload.c                            |   6 +-
 net/mptcp/options.c                                |  54 ++++++++++-
 net/mptcp/pm_netlink.c                             |  20 ++--
 net/mptcp/protocol.c                               |  84 +++++++++++------
 net/mptcp/protocol.h                               |   3 +-
 net/mptcp/subflow.c                                |   8 ++
 net/openvswitch/actions.c                          |  68 +-------------
 net/openvswitch/flow_netlink.c                     |  64 ++-----------
 net/openvswitch/flow_netlink.h                     |   2 -
 net/tls/tls_device.c                               |   4 +-
 net/unix/af_unix.c                                 |  36 ++++----
 net/vmw_vsock/af_vsock.c                           |  40 ++++++--
 net/xfrm/xfrm_output.c                             |   6 +-
 net/xfrm/xfrm_state.c                              |   8 +-
 net/xfrm/xfrm_user.c                               |   5 +-
 scripts/kconfig/mconf.c                            |   3 +
 scripts/kconfig/nconf.c                            |   3 +
 sound/usb/endpoint.c                               |   3 +-
 sound/usb/mixer.c                                  |   2 +-
 tools/arch/riscv/include/asm/csr.h                 |   5 +-
 tools/testing/selftests/net/bareudp.sh             |   2 +-
 .../selftests/net/forwarding/lib_sh_test.sh        |   7 ++
 tools/testing/selftests/net/lib.sh                 |   2 +-
 tools/testing/selftests/net/mptcp/mptcp_join.sh    |  18 ++--
 tools/tracing/latency/latency-collector.c          |   2 +-
 112 files changed, 914 insertions(+), 560 deletions(-)

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Sebastian Ene <[email protected]>

commit 103e17aac09cdd358133f9e00998b75d6c1f1518 upstream.

Verify the offset to prevent OOB access in the hypervisor FF-A buffer in case an untrusted large enough value [U32_MAX - sizeof(struct ffa_composite_mem_region) + 1, U32_MAX] is set from the host kernel.

---
 arch/arm64/kvm/hyp/nvhe/ffa.c |    9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

@@ -437,7 +437,7 @@ static void __do_ffa_mem_xfer(const u64
 	struct ffa_mem_region_attributes *ep_mem_access;
 	struct ffa_composite_mem_region *reg;
 	struct ffa_mem_region *buf;
-	u32 offset, nr_ranges;
+	u32 offset, nr_ranges, checked_offset;
 	int ret = 0;
 
 	if (addr_mbz || npages_mbz || fraglen > len ||

@@ -474,7 +474,12 @@ static void __do_ffa_mem_xfer(const u64
 		goto out_unlock;
 	}
 
-	if (fraglen < offset + sizeof(struct ffa_composite_mem_region)) {
+	if (check_add_overflow(offset, sizeof(struct ffa_composite_mem_region), &checked_offset)) {
+		ret = FFA_RET_INVALID_PARAMETERS;
+		goto out_unlock;
+	}
+
+	if (fraglen < checked_offset) {
 		ret = FFA_RET_INVALID_PARAMETERS;
 		goto out_unlock;
 	}

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Yipeng Zou <[email protected]>

commit 20739af07383e6eb1ec59dcd70b72ebfa9ac362c upstream.

There is a race condition between timer_shutdown_sync() and timer expiration that can lead to hitting a WARN_ON in expire_timers().

The issue occurs when timer_shutdown_sync() clears the timer function to NULL while the timer is still running on another CPU. The race scenario looks like this:

CPU0 CPU1

					<SOFTIRQ>
					lock_timer_base()
					expire_timers()
					base->running_timer = timer;
					unlock_timer_base()
					[call_timer_fn enter]
					mod_timer()
					...

timer_shutdown_sync()
lock_timer_base()
// For now, will not detach the timer but only clear its function to NULL if (base->running_timer != timer)

	ret = detach_if_pending(timer, base, true);
if (shutdown)
	timer->function = NULL;
unlock_timer_base()
					[call_timer_fn exit]
					lock_timer_base()
					base->running_timer = NULL;
					unlock_timer_base()
					...
					// Now timer is pending while its function set to NULL.
					// next timer trigger
					<SOFTIRQ>
					expire_timers()
					WARN_ON_ONCE(!fn) // hit
					...
lock_timer_base()
// Now timer will detach
if (base->running_timer != timer)
	ret = detach_if_pending(timer, base, true);
if (shutdown)
	timer->function = NULL;

unlock_timer_base()

The problem is that timer_shutdown_sync() clears the timer function regardless of whether the timer is currently running. This can leave a pending timer with a NULL function pointer, which triggers the WARN_ON_ONCE(!fn) check in expire_timers().

Fix this by only clearing the timer function when actually detaching the timer. If the timer is running, leave the function pointer intact, which is safe because the timer will be properly detached when it finishes running.

---
 kernel/time/timer.c |    7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

@@ -1505,10 +1505,11 @@ static int __try_to_del_timer_sync(struc
 
 	base = lock_timer_base(timer, &flags);
 
-	if (base->running_timer != timer)
+	if (base->running_timer != timer) {
 		ret = detach_if_pending(timer, base, true);
-	if (shutdown)
-		timer->function = NULL;
+		if (shutdown)
+			timer->function = NULL;
+	}
 
 	raw_spin_unlock_irqrestore(&base->lock, flags);
 

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Mario Limonciello (AMD) <[email protected]>

commit 4d3a13afa8b64dc49293b3eab3e7beac11072c12 upstream.

Titas reports that the accelerometer sensor on their laptop only works after a warm boot or unloading/reloading the amd-sfh kernel module.

Presumably the sensor is in a bad state on cold boot and failing to start, so explicitly stop it before starting.

---
 drivers/hid/amd-sfh-hid/sfh1_1/amd_sfh_init.c |    2 ++
 1 file changed, 2 insertions(+)

@@ -172,6 +172,8 @@ static int amd_sfh1_1_hid_client_init(st
 		if (rc)
 			goto cleanup;
 
+		mp2_ops->stop(privdata, cl_data->sensor_idx[i]);
+		amd_sfh_wait_for_response(privdata, cl_data->sensor_idx[i], DISABLE_SENSOR);
 		writel(0, privdata->mmio + amd_get_p2c_val(privdata, 0));
 		mp2_ops->start(privdata, info);
 		status = amd_sfh_wait_for_response

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Zhang Heng <[email protected]>

commit beab067dbcff642243291fd528355d64c41dc3b2 upstream.

Based on available evidence, the USB ID 4c4a:4155 used by multiple devices has been attributed to Jieli. The commit 1a8953f4f774 ("HID: Add IGNORE quirk for SMARTLINKTECHNOLOGY") affected touchscreen functionality. Added checks for manufacturer and serial number to maintain microphone compatibility, enabling both devices to function properly.

---
 drivers/hid/hid-ids.h    |    4 ++--
 drivers/hid/hid-quirks.c |   13 ++++++++++++-
 2 files changed, 14 insertions(+), 3 deletions(-)

@@ -1528,7 +1528,7 @@
 #define USB_VENDOR_ID_SIGNOTEC			0x2133
 #define USB_DEVICE_ID_SIGNOTEC_VIEWSONIC_PD1011	0x0018
 
-#define USB_VENDOR_ID_SMARTLINKTECHNOLOGY              0x4c4a
-#define USB_DEVICE_ID_SMARTLINKTECHNOLOGY_4155         0x4155
+#define USB_VENDOR_ID_JIELI_SDK_DEFAULT		0x4c4a
+#define USB_DEVICE_ID_JIELI_SDK_4155		0x4155
 
 #endif

@@ -900,7 +900,6 @@ static const struct hid_device_id hid_ig
 #endif
 	{ HID_USB_DEVICE(USB_VENDOR_ID_YEALINK, USB_DEVICE_ID_YEALINK_P1K_P4K_B2K) },
 	{ HID_USB_DEVICE(USB_VENDOR_ID_QUANTA, USB_DEVICE_ID_QUANTA_HP_5MP_CAMERA_5473) },
-	{ HID_USB_DEVICE(USB_VENDOR_ID_SMARTLINKTECHNOLOGY, USB_DEVICE_ID_SMARTLINKTECHNOLOGY_4155) },
 	{ }
 };
 

@@ -1057,6 +1056,18 @@ bool hid_ignore(struct hid_device *hdev)
 					     strlen(elan_acpi_id[i].id)))
 					return true;
 		break;
+	case USB_VENDOR_ID_JIELI_SDK_DEFAULT:
+		/*
+		 * Multiple USB devices with identical IDs (mic & touchscreen).
+		 * The touch screen requires hid core processing, but the
+		 * microphone does not. They can be distinguished by manufacturer
+		 * and serial number.
+		 */
+		if (hdev->product == USB_DEVICE_ID_JIELI_SDK_4155 &&
+		    strncmp(hdev->name, "SmartlinkTechnology", 19) == 0 &&
+		    strncmp(hdev->uniq, "20201111000001", 14) == 0)
+			return true;
+		break;
 	}
 
 	if (hdev->type == HID_TYPE_USBMOUSE &&

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Diederik de Haas <[email protected]>

commit 03c7e964a02e388ee168c804add7404eda23908c upstream.

Page 13 of the PineTab2 v2 schematic dd 20230417 shows VCCIO4's power source is VCCIO_WL. Page 19 shows that VCCIO_WL is connected to VCCA1V8_PMU, so fix the PineTab2 dtsi to reflect that.

---
 arch/arm64/boot/dts/rockchip/rk3566-pinetab2.dtsi |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -789,7 +789,7 @@
 	vccio1-supply = <&vccio_acodec>;
 	vccio2-supply = <&vcc_1v8>;
 	vccio3-supply = <&vccio_sd>;
-	vccio4-supply = <&vcc_1v8>;
+	vccio4-supply = <&vcca1v8_pmu>;
 	vccio5-supply = <&vcc_1v8>;
 	vccio6-supply = <&vcc1v8_dvp>;
 	vccio7-supply = <&vcc_3v3>;

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Mykola Kvach <[email protected]>

commit b5414520793e68d266fdd97a84989d9831156aad upstream.

The vcc3v3_pcie20 fixed regulator powers the PCIe device-side 3.3V rail for pcie2x1l2 via vpcie3v3-supply. The DTS mistakenly set its regulator-min/max-microvolt to 1800000 (1.8 V). Correct both to 3300000 (3.3 V) to match the rail name, the PCIe/M.2 power requirement, and the actual hardware wiring on Orange Pi 5.

---
 arch/arm64/boot/dts/rockchip/rk3588s-orangepi-5.dts |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

@@ -85,8 +85,8 @@
 		gpios = <&gpio0 RK_PC5 GPIO_ACTIVE_HIGH>;
 		regulator-name = "vcc3v3_pcie20";
 		regulator-boot-on;
-		regulator-min-microvolt = <1800000>;
-		regulator-max-microvolt = <1800000>;
+		regulator-min-microvolt = <3300000>;
+		regulator-max-microvolt = <3300000>;
 		startup-delay-us = <50000>;
 		vin-supply = <&vcc5v0_sys>;
 	};

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Quentin Schulz <[email protected]>

commit 08d70143e3033d267507deb98a5fd187df3e6640 upstream.

In commit 296602b8e5f7 ("arm64: dts: rockchip: Move RK3399 OPPs to dtsi files for SoC variants"), everything shared between variants of RK3399 was put into rk3399-base.dtsi and the rest in variant-specific DTSI, such as rk3399-t, rk3399-op1, rk3399, etc.
Therefore, the variant-specific DTSI should include rk3399-base.dtsi and not another variant's DTSI.

rk3399-op1 wrongly includes rk3399 (a variant) DTSI instead of rk3399-base DTSI, let's fix this oversight by including the intended DTSI.

Fortunately, this had no impact on the resulting DTB since all nodes were named the same and all node properties were overridden in rk3399-op1.dtsi. This was checked by doing a checksum of rk3399-op1 DTBs before and after this commit.

No intended change in behavior.

---
 arch/arm64/boot/dts/rockchip/rk3399-op1.dtsi | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index c4f4f1ff6117..9da6fd82e46b 100644
@@ -3,7 +3,7 @@
  * Copyright (c) 2016-2017 Fuzhou Rockchip Electronics Co., Ltd
  */
 
-#include "rk3399.dtsi"
+#include "rk3399-base.dtsi"
 
 / {
 	cluster0_opp: opp-table-0 {

2.52.0

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Quentin Schulz <[email protected]>

commit baa18d577cd445145039e731d3de0fa49ca57204 upstream.

We've had reports from the field that some RK3588 Tiger have random issues with eMMC errors.

Applying commit a28352cf2d2f ("mmc: sdhci-of-dwcmshc: Change DLL_STRBIN_TAPNUM_DEFAULT to 0x4") didn't help and seemed to have made things worse for our board.

Our HW department checked the eMMC lines and reported that they are too long and don't look great so signal integrity is probably not the best.

Note that not all Tigers with the same eMMC chip have errors, so the suspicion is that we're really on the edge in terms of signal integrity and only a handful devices are failing. Additionally, we have RK3588 Jaguars with the same eMMC chip but the layout is different and we also haven't received reports about those so far.

Lowering the max-frequency to 150MHz from 200MHz instead of simply disabling HS400 was briefly tested and seem to work as well. We've disabled HS400 downstream and haven't received reports since so we'll go with that instead of lowering the max-frequency.

---
 arch/arm64/boot/dts/rockchip/rk3588-tiger.dtsi |    4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

@@ -359,14 +359,12 @@
 	cap-mmc-highspeed;
 	mmc-ddr-1_8v;
 	mmc-hs200-1_8v;
-	mmc-hs400-1_8v;
-	mmc-hs400-enhanced-strobe;
 	mmc-pwrseq = <&emmc_pwrseq>;
 	no-sdio;
 	no-sd;
 	non-removable;
 	pinctrl-names = "default";
-	pinctrl-0 = <&emmc_bus8 &emmc_cmd &emmc_clk &emmc_data_strobe>;
+	pinctrl-0 = <&emmc_bus8 &emmc_cmd &emmc_clk>;
 	vmmc-supply = <&vcc_3v3_s3>;
 	vqmmc-supply = <&vcc_1v8_s3>;
 	status = "okay";

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Niravkumar L Rabara <[email protected]>

commit 5c56bf214af85ca042bf97f8584aab2151035840 upstream.

The DMA device pointer dma_dev was being dereferenced before ensuring that cdns_ctrl->dmac is properly initialized.

Move the assignment of dma_dev after successfully acquiring the DMA channel to ensure the pointer is valid before use.

---
 drivers/mtd/nand/raw/cadence-nand-controller.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

@@ -2871,7 +2871,7 @@ cadence_nand_irq_cleanup(int irqnum, str
 static int cadence_nand_init(struct cdns_nand_ctrl *cdns_ctrl)
 {
 	dma_cap_mask_t mask;
-	struct dma_device *dma_dev = cdns_ctrl->dmac->device;
+	struct dma_device *dma_dev;
 	int ret;
 
 	cdns_ctrl->cdma_desc = dma_alloc_coherent(cdns_ctrl->dev,

@@ -2915,6 +2915,7 @@ static int cadence_nand_init(struct cdns
 		}
 	}
 
+	dma_dev = cdns_ctrl->dmac->device;
 	cdns_ctrl->io.iova_dma = dma_map_resource(dma_dev->dev, cdns_ctrl->io.dma,
 						  cdns_ctrl->io.size,
 						  DMA_BIDIRECTIONAL, 0);

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Dan Carpenter <[email protected]>

commit e4185bed738da755b191aa3f2e16e8b48450e1b8 upstream.

The "req.start" and "req.len" variables are u64 values that come from the user at the start of the function. We mask away the high 32 bits of "req.len" so that's capped at U32_MAX but the "req.start" variable can go up to U64_MAX which means that the addition can still integer overflow.

Use check_add_overflow() to fix this bug.

---
 drivers/mtd/mtdchar.c |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

@@ -599,6 +599,7 @@ mtdchar_write_ioctl(struct mtd_info *mtd
 	uint8_t *datbuf = NULL, *oobbuf = NULL;
 	size_t datbuf_len, oobbuf_len;
 	int ret = 0;
+	u64 end;
 
 	if (copy_from_user(&req, argp, sizeof(req)))
 		return -EFAULT;

@@ -618,7 +619,7 @@ mtdchar_write_ioctl(struct mtd_info *mtd
 	req.len &= 0xffffffff;
 	req.ooblen &= 0xffffffff;
 
-	if (req.start + req.len > mtd->size)
+	if (check_add_overflow(req.start, req.len, &end) || end > mtd->size)
 		return -EINVAL;
 
 	datbuf_len = min_t(size_t, req.len, mtd->erasesize);

@@ -698,6 +699,7 @@ mtdchar_read_ioctl(struct mtd_info *mtd,
 	size_t datbuf_len, oobbuf_len;
 	size_t orig_len, orig_ooblen;
 	int ret = 0;
+	u64 end;
 
 	if (copy_from_user(&req, argp, sizeof(req)))
 		return -EFAULT;

@@ -724,7 +726,7 @@ mtdchar_read_ioctl(struct mtd_info *mtd,
 	req.len &= 0xffffffff;
 	req.ooblen &= 0xffffffff;
 
-	if (req.start + req.len > mtd->size) {
+	if (check_add_overflow(req.start, req.len, &end) || end > mtd->size) {
 		ret = -EINVAL;
 		goto out;
 	}

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Yongpeng Yang <[email protected]>

commit e106e269c5cb38315eb0a0e7e38f71e9b20c8c66 upstream.

sb_min_blocksize() may return 0. Check its return value to avoid opt->blocksize and sb->s_blocksize is 0.

---
 fs/isofs/inode.c |    5 +++++
 1 file changed, 5 insertions(+)

@@ -610,6 +610,11 @@ static int isofs_fill_super(struct super
 		goto out_freesbi;
 	}
 	opt->blocksize = sb_min_blocksize(s, opt->blocksize);
+	if (!opt->blocksize) {
+		printk(KERN_ERR
+		       "ISOFS: unable to set blocksize\n");
+		goto out_freesbi;
+	}
 
 	sbi->s_high_sierra = 0; /* default is iso9660 */
 	sbi->s_session = opt->session;

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Mike Yuan <[email protected]>

commit 3cd1548a278c7d6a9bdef1f1866e7cf66bfd3518 upstream.

In systemd we're trying to switch the internal credentials setup logic to new mount API [1], and I noticed fsconfig(FSCONFIG_CMD_RECONFIGURE) consistently fails on tmpfs with noswap option. This can be trivially reproduced with the following:

```
int fs_fd = fsopen("tmpfs", 0);
fsconfig(fs_fd, FSCONFIG_SET_FLAG, "noswap", NULL, 0);
fsconfig(fs_fd, FSCONFIG_CMD_CREATE, NULL, NULL, 0);
fsmount(fs_fd, 0, 0);
fsconfig(fs_fd, FSCONFIG_CMD_RECONFIGURE, NULL, NULL, 0);  <------ EINVAL
```

After some digging the culprit is shmem_reconfigure() rejecting !(ctx->seen & SHMEM_SEEN_NOSWAP) && sbinfo->noswap, which is bogus as ctx->seen serves as a mask for whether certain options are touched at all. On top of that, noswap option doesn't use fsparam_flag_no, hence it's not really possible to "reenable" swap to begin with. Drop the check and redundant SHMEM_SEEN_NOSWAP flag.

[1] https://github.com/systemd/systemd/pull/39637

---
 mm/shmem.c |   15 +++++++--------
 1 file changed, 7 insertions(+), 8 deletions(-)

@@ -127,8 +127,7 @@ struct shmem_options {
 #define SHMEM_SEEN_INODES 2
 #define SHMEM_SEEN_HUGE 4
 #define SHMEM_SEEN_INUMS 8
-#define SHMEM_SEEN_NOSWAP 16
-#define SHMEM_SEEN_QUOTA 32
+#define SHMEM_SEEN_QUOTA 16
 };
 
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE

@@ -4330,7 +4329,6 @@ static int shmem_parse_one(struct fs_con
 				       "Turning off swap in unprivileged tmpfs mounts unsupported");
 		}
 		ctx->noswap = true;
-		ctx->seen |= SHMEM_SEEN_NOSWAP;
 		break;
 	case Opt_quota:
 		if (fc->user_ns != &init_user_ns)

@@ -4480,14 +4478,15 @@ static int shmem_reconfigure(struct fs_c
 		err = "Current inum too high to switch to 32-bit inums";
 		goto out;
 	}
-	if ((ctx->seen & SHMEM_SEEN_NOSWAP) && ctx->noswap && !sbinfo->noswap) {
+
+	/*
+	 * "noswap" doesn't use fsparam_flag_no, i.e. there's no "swap"
+	 * counterpart for (re-)enabling swap.
+	 */
+	if (ctx->noswap && !sbinfo->noswap) {
 		err = "Cannot disable swap on remount";
 		goto out;
 	}
-	if (!(ctx->seen & SHMEM_SEEN_NOSWAP) && !ctx->noswap && sbinfo->noswap) {
-		err = "Cannot enable swap on remount if it was disabled on first mount";
-		goto out;
-	}
 
 	if (ctx->seen & SHMEM_SEEN_QUOTA &&
 	    !sb_any_quota_loaded(fc->root->d_sb)) {

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Yongpeng Yang <[email protected]>

commit f2c1f631630e01821fe4c3fdf6077bc7a8284f82 upstream.

sb_min_blocksize() may return 0. Check its return value to avoid accessing the filesystem super block when sb->s_blocksize is 0.

---
 fs/exfat/super.c |    5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

@@ -452,7 +452,10 @@ static int exfat_read_boot_sector(struct
 	struct exfat_sb_info *sbi = EXFAT_SB(sb);
 
 	/* set block size to read super block */
-	sb_min_blocksize(sb, 512);
+	if (!sb_min_blocksize(sb, 512)) {
+		exfat_err(sb, "unable to set blocksize");
+		return -EINVAL;
+	}
 
 	/* read boot sector */
 	sbi->boot_bh = sb_bread(sb, 0);

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Jiayuan Chen <[email protected]>

commit fbade4bd08ba52cbc74a71c4e86e736f059f99f7 upstream.

The sockmap feature allows bpf syscall from userspace, or based on bpf sockops, replacing the sk_prot of sockets during protocol stack processing with sockmap's custom read/write interfaces.
'''
tcp_rcv_state_process()
subflow_syn_recv_sock()

    tcp_init_transfer(BPF_SOCK_OPS_PASSIVE_ESTABLISHED_CB)
      bpf_skops_established       <== sockops
        bpf_sock_map_update(sk)   <== call bpf helper
          tcp_bpf_update_proto()  <== update sk_prot

'''
Consider two scenarios:

1. When the server has MPTCP enabled and the client also requests MPTCP, the sk passed to the BPF program is a subflow sk. Since subflows only handle partial data, replacing their sk_prot is meaningless and will cause traffic disruption.

2. When the server has MPTCP enabled but the client sends a TCP SYN without MPTCP, subflow_syn_recv_sock() performs a fallback on the subflow, replacing the subflow sk's sk_prot with the native sk_prot. '''
subflow_ulp_fallback()
subflow_drop_ctx()
mptcp_subflow_ops_undo_override()
'''
Subsequently, accept::mptcp_stream_accept::mptcp_fallback_tcp_ops() converts the subflow to plain TCP.

For the first case, we should prevent it from being combined with sockmap by setting sk_prot->psock_update_sk_prot to NULL, which will be blocked by sockmap's own flow.

For the second case, since subflow_syn_recv_sock() has already restored sk_prot to native tcp_prot/tcpv6_prot, no further action is needed.

---
 net/mptcp/subflow.c |    8 ++++++++
 1 file changed, 8 insertions(+)

@@ -2150,6 +2150,10 @@ void __init mptcp_subflow_init(void)
 	tcp_prot_override = tcp_prot;
 	tcp_prot_override.release_cb = tcp_release_cb_override;
 	tcp_prot_override.diag_destroy = tcp_abort_override;
+#ifdef CONFIG_BPF_SYSCALL
+	/* Disable sockmap processing for subflows */
+	tcp_prot_override.psock_update_sk_prot = NULL;
+#endif
 
 #if IS_ENABLED(CONFIG_MPTCP_IPV6)
 	/* In struct mptcp_subflow_request_sock, we assume the TCP request sock

@@ -2186,6 +2190,10 @@ void __init mptcp_subflow_init(void)
 	tcpv6_prot_override = tcpv6_prot;
 	tcpv6_prot_override.release_cb = tcp_release_cb_override;
 	tcpv6_prot_override.diag_destroy = tcp_abort_override;
+#ifdef CONFIG_BPF_SYSCALL
+	/* Disable sockmap processing for subflows */
+	tcpv6_prot_override.psock_update_sk_prot = NULL;
+#endif
 #endif
 
 	mptcp_diag_subflow_init(&subflow_ulp_ops);

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Jiayuan Chen <[email protected]>

commit c77b3b79a92e3345aa1ee296180d1af4e7031f8f upstream.

The sockmap feature allows bpf syscall from userspace, or based on bpf sockops, replacing the sk_prot of sockets during protocol stack processing with sockmap's custom read/write interfaces.
'''
tcp_rcv_state_process()
syn_recv_sock()/subflow_syn_recv_sock()

    tcp_init_transfer(BPF_SOCK_OPS_PASSIVE_ESTABLISHED_CB)
      bpf_skops_established       <== sockops
        bpf_sock_map_update(sk)   <== call bpf helper
          tcp_bpf_update_proto()  <== update sk_prot

'''

When the server has MPTCP enabled but the client sends a TCP SYN without MPTCP, subflow_syn_recv_sock() performs a fallback on the subflow, replacing the subflow sk's sk_prot with the native sk_prot. '''
subflow_syn_recv_sock()
subflow_ulp_fallback()
subflow_drop_ctx()
mptcp_subflow_ops_undo_override()
'''

Then, this subflow can be normally used by sockmap, which replaces the native sk_prot with sockmap's custom sk_prot. The issue occurs when the user executes accept::mptcp_stream_accept::mptcp_fallback_tcp_ops(). Here, it uses sk->sk_prot to compare with the native sk_prot, but this is incorrect when sockmap is used, as we may incorrectly set sk->sk_socket->ops.

This fix uses the more generic sk_family for the comparison instead.

Additionally, this also prevents a WARNING from occurring:

result from ./scripts/decode_stacktrace.sh:

------------[ cut here ]------------
WARNING: CPU: 0 PID: 337 at net/mptcp/protocol.c:68 mptcp_stream_accept \
(net/mptcp/protocol.c:4005)
Modules linked in:

...

PKRU: 55555554

Call Trace:
<TASK>

do_accept (net/socket.c:1989)
__sys_accept4 (net/socket.c:2028 net/socket.c:2057) __x64_sys_accept (net/socket.c:2067)
x64_sys_call (arch/x86/entry/syscall_64.c:41)
do_syscall_64 (arch/x86/entry/syscall_64.c:63 arch/x86/entry/syscall_64.c:94) entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:130)

RIP: 0033:0x7f87ac92b83d

---[ end trace 0000000000000000 ]---

---
 net/mptcp/protocol.c |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

@@ -57,11 +57,13 @@ static u64 mptcp_wnd_end(const struct mp
 
 static const struct proto_ops *mptcp_fallback_tcp_ops(const struct sock *sk)
 {
+	unsigned short family = READ_ONCE(sk->sk_family);
+
 #if IS_ENABLED(CONFIG_MPTCP_IPV6)
-	if (sk->sk_prot == &tcpv6_prot)
+	if (family == AF_INET6)
 		return &inet6_stream_ops;
 #endif
-	WARN_ON_ONCE(sk->sk_prot != &tcp_prot);
+	WARN_ON_ONCE(family != AF_INET);
 	return &inet_stream_ops;
 }
 

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Niklas Cassel <[email protected]>

commit b11890683380a36b8488229f818d5e76e8204587 upstream.

Commit cf3fc037623c ("ata: libata-scsi: Fix ata_to_sense_error() status handling") fixed ata_to_sense_error() to properly generate sense key ABORTED COMMAND (without any additional sense code), instead of the previous bogus sense key ILLEGAL REQUEST with the additional sense code UNALIGNED WRITE COMMAND, for a failed command.

However, this broke suspend for Security locked drives (drives that have Security enabled, and have not been Security unlocked by boot firmware).

The reason for this is that the SCSI disk driver, for the Synchronize Cache command only, treats any sense data with sense key ILLEGAL REQUEST as a successful command (regardless of ASC / ASCQ).

After commit cf3fc037623c ("ata: libata-scsi: Fix ata_to_sense_error() status handling") the code that treats any sense data with sense key ILLEGAL REQUEST as a successful command is no longer applicable, so the command fails, which causes the system suspend to be aborted:

sd 1:0:0:0: PM: dpm_run_callback(): scsi_bus_suspend returns -5 sd 1:0:0:0: PM: failed to suspend async: error -5
PM: Some devices failed to suspend, or early wake event detected

To make suspend work once again, for a Security locked device only, return sense data LOGICAL UNIT ACCESS NOT AUTHORIZED, the actual sense data which a real SCSI device would have returned if locked. The SCSI disk driver treats this sense data as a successful command.

---
 drivers/ata/libata-scsi.c |    7 +++++++
 include/linux/ata.h       |    1 +
 2 files changed, 8 insertions(+)

@@ -991,6 +991,13 @@ static void ata_gen_ata_sense(struct ata
 		return;
 	}
 
+	if (ata_id_is_locked(dev->id)) {
+		/* Security locked */
+		/* LOGICAL UNIT ACCESS NOT AUTHORIZED */
+		ata_scsi_set_sense(dev, cmd, DATA_PROTECT, 0x74, 0x71);
+		return;
+	}
+
 	if (!(qc->flags & ATA_QCFLAG_RTF_FILLED)) {
 		ata_dev_dbg(dev,
 			    "Missing result TF: reporting aborted command\n");

@@ -566,6 +566,7 @@ struct ata_bmdma_prd {
 #define ata_id_has_ncq(id)	((id)[ATA_ID_SATA_CAPABILITY] & (1 << 8))
 #define ata_id_queue_depth(id)	(((id)[ATA_ID_QUEUE_DEPTH] & 0x1f) + 1)
 #define ata_id_removable(id)	((id)[ATA_ID_CONFIG] & (1 << 7))
+#define ata_id_is_locked(id)	(((id)[ATA_ID_DLF] & 0x7) == 0x7)
 #define ata_id_has_atapi_AN(id)	\
 	((((id)[ATA_ID_SATA_CAPABILITY] != 0x0000) && \
 	  ((id)[ATA_ID_SATA_CAPABILITY] != 0xffff)) && \

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Maciej W. Rozycki <[email protected]>

commit 9f048fa487409e364cf866c957cf0b0d782ca5a3 upstream.

Depending on the particular CPU implementation a TLB shutdown may occur if multiple matching entries are detected upon the execution of a TLBP or the TLBWI/TLBWR instructions. Given that we don't know what entries we have been handed we need to be very careful with the initial TLB setup and avoid all these instructions.

Therefore read all the TLB entries one by one with the TLBR instruction, bypassing the content addressing logic, and truncate any large pages in place so as to avoid a case in the second step where an incoming entry for a large page at a lower address overlaps with a replacement entry chosen at another index. Then preinitialize the TLB using addresses outside our usual unique range and avoiding clashes with any entries received, before making the usual call to local_flush_tlb_all().

This fixes (at least) R4x00 cores if TLBP hits multiple matching TLB entries (SGI IP22 PROM for examples sets up all TLBs to the same virtual address).

---
 arch/mips/mm/tlb-r4k.c |  102 ++++++++++++++++++++++++++++++-------------------
 1 file changed, 64 insertions(+), 38 deletions(-)

@@ -15,6 +15,7 @@
 #include <linux/mm.h>
 #include <linux/hugetlb.h>
 #include <linux/export.h>
+#include <linux/sort.h>
 
 #include <asm/cpu.h>
 #include <asm/cpu-type.h>

@@ -508,55 +509,79 @@ static int __init set_ntlb(char *str)
 
 __setup("ntlb=", set_ntlb);
 
-/* Initialise all TLB entries with unique values */
+
+/* Comparison function for EntryHi VPN fields.  */
+static int r4k_vpn_cmp(const void *a, const void *b)
+{
+	long v = *(unsigned long *)a - *(unsigned long *)b;
+	int s = sizeof(long) > sizeof(int) ? sizeof(long) * 8 - 1: 0;
+	return s ? (v != 0) | v >> s : v;
+}
+
+/*
+ * Initialise all TLB entries with unique values that do not clash with
+ * what we have been handed over and what we'll be using ourselves.
+ */
 static void r4k_tlb_uniquify(void)
 {
-	int entry = num_wired_entries();
+	unsigned long tlb_vpns[1 << MIPS_CONF1_TLBS_SIZE];
+	int tlbsize = current_cpu_data.tlbsize;
+	int start = num_wired_entries();
+	unsigned long vpn_mask;
+	int cnt, ent, idx, i;
+
+	vpn_mask = GENMASK(cpu_vmbits - 1, 13);
+	vpn_mask |= IS_ENABLED(CONFIG_64BIT) ? 3ULL << 62 : 1 << 31;
 
 	htw_stop();
-	write_c0_entrylo0(0);
-	write_c0_entrylo1(0);
 
-	while (entry < current_cpu_data.tlbsize) {
-		unsigned long asid_mask = cpu_asid_mask(&current_cpu_data);
-		unsigned long asid = 0;
-		int idx;
+	for (i = start, cnt = 0; i < tlbsize; i++, cnt++) {
+		unsigned long vpn;
 
-		/* Skip wired MMID to make ginvt_mmid work */
-		if (cpu_has_mmid)
-			asid = MMID_KERNEL_WIRED + 1;
+		write_c0_index(i);
+		mtc0_tlbr_hazard();
+		tlb_read();
+		tlb_read_hazard();
+		vpn = read_c0_entryhi();
+		vpn &= vpn_mask & PAGE_MASK;
+		tlb_vpns[cnt] = vpn;
 
-		/* Check for match before using UNIQUE_ENTRYHI */
-		do {
-			if (cpu_has_mmid) {
-				write_c0_memorymapid(asid);
-				write_c0_entryhi(UNIQUE_ENTRYHI(entry));
-			} else {
-				write_c0_entryhi(UNIQUE_ENTRYHI(entry) | asid);
-			}
-			mtc0_tlbw_hazard();
-			tlb_probe();
-			tlb_probe_hazard();
-			idx = read_c0_index();
-			/* No match or match is on current entry */
-			if (idx < 0 || idx == entry)
-				break;
-			/*
-			 * If we hit a match, we need to try again with
-			 * a different ASID.
-			 */
-			asid++;
-		} while (asid < asid_mask);
-
-		if (idx >= 0 && idx != entry)
-			panic("Unable to uniquify TLB entry %d", idx);
-
-		write_c0_index(entry);
+		/* Prevent any large pages from overlapping regular ones.  */
+		write_c0_pagemask(read_c0_pagemask() & PM_DEFAULT_MASK);
 		mtc0_tlbw_hazard();
 		tlb_write_indexed();
-		entry++;
+		tlbw_use_hazard();
 	}
 
+	sort(tlb_vpns, cnt, sizeof(tlb_vpns[0]), r4k_vpn_cmp, NULL);
+
+	write_c0_pagemask(PM_DEFAULT_MASK);
+	write_c0_entrylo0(0);
+	write_c0_entrylo1(0);
+
+	idx = 0;
+	ent = tlbsize;
+	for (i = start; i < tlbsize; i++)
+		while (1) {
+			unsigned long entryhi, vpn;
+
+			entryhi = UNIQUE_ENTRYHI(ent);
+			vpn = entryhi & vpn_mask & PAGE_MASK;
+
+			if (idx >= cnt || vpn < tlb_vpns[idx]) {
+				write_c0_entryhi(entryhi);
+				write_c0_index(i);
+				mtc0_tlbw_hazard();
+				tlb_write_indexed();
+				ent++;
+				break;
+			} else if (vpn == tlb_vpns[idx]) {
+				ent++;
+			} else {
+				idx++;
+			}
+		}
+
 	tlbw_use_hazard();
 	htw_start();
 	flush_micro_tlb();

@@ -602,6 +627,7 @@ static void r4k_tlb_configure(void)
 
 	/* From this point on the ARC firmware is dead.	 */
 	r4k_tlb_uniquify();
+	local_flush_tlb_all();
 
 	/* Did I tell you that ARC SUCKS?  */
 }

So this caused a regression for a subset of systems. A fix is in review,
so I think this will best be dropped at this point and then cherry-picked along the fix. I've arranged for this to happen via patch prerequisites already. Please let me know if there's an issue with this approach.

Maciej

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Henrique Carvalho <[email protected]>

commit a9d1f38df7ecd0e21233447c9cc6fa1799eddaf3 upstream.

Replace close_cached_dir() calls under cfid_list_lock with a new close_cached_dir_locked() variant that uses kref_put() instead of kref_put_lock() to avoid recursive locking when dropping references.

While the existing code works if the refcount >= 2 invariant holds, this area has proven error-prone. Make deadlocks impossible and WARN on invariant violations.

---
 fs/smb/client/cached_dir.c |   41 ++++++++++++++++++++++++++++++++++++++---
 1 file changed, 38 insertions(+), 3 deletions(-)

@@ -16,6 +16,7 @@ static struct cached_fid *init_cached_di
 static void free_cached_dir(struct cached_fid *cfid);
 static void smb2_close_cached_fid(struct kref *ref);
 static void cfids_laundromat_worker(struct work_struct *work);
+static void close_cached_dir_locked(struct cached_fid *cfid);
 
 struct cached_dir_dentry {
 	struct list_head entry;

@@ -362,7 +363,7 @@ out:
 			 * lease. Release one here, and the second below.
 			 */
 			cfid->has_lease = false;
-			close_cached_dir(cfid);
+			close_cached_dir_locked(cfid);
 		}
 		spin_unlock(&cfids->cfid_list_lock);
 

@@ -448,18 +449,52 @@ void drop_cached_dir_by_name(const unsig
 	spin_lock(&cfid->cfids->cfid_list_lock);
 	if (cfid->has_lease) {
 		cfid->has_lease = false;
-		close_cached_dir(cfid);
+		close_cached_dir_locked(cfid);
 	}
 	spin_unlock(&cfid->cfids->cfid_list_lock);
 	close_cached_dir(cfid);
 }
 
-
+/**
+ * close_cached_dir - drop a reference of a cached dir
+ *
+ * The release function will be called with cfid_list_lock held to remove the
+ * cached dirs from the list before any other thread can take another @cfid
+ * ref. Must not be called with cfid_list_lock held; use
+ * close_cached_dir_locked() called instead.
+ *
+ * @cfid: cached dir
+ */
 void close_cached_dir(struct cached_fid *cfid)
 {
+	lockdep_assert_not_held(&cfid->cfids->cfid_list_lock);
 	kref_put_lock(&cfid->refcount, smb2_close_cached_fid, &cfid->cfids->cfid_list_lock);
 }
 
+/**
+ * close_cached_dir_locked - put a reference of a cached dir with
+ * cfid_list_lock held
+ *
+ * Calling close_cached_dir() with cfid_list_lock held has the potential effect
+ * of causing a deadlock if the invariant of refcount >= 2 is false.
+ *
+ * This function is used in paths that hold cfid_list_lock and expect at least
+ * two references. If that invariant is violated, WARNs and returns without
+ * dropping a reference; the final put must still go through
+ * close_cached_dir().
+ *
+ * @cfid: cached dir
+ */
+static void close_cached_dir_locked(struct cached_fid *cfid)
+{
+	lockdep_assert_held(&cfid->cfids->cfid_list_lock);
+
+	if (WARN_ON(kref_read(&cfid->refcount) < 2))
+		return;
+
+	kref_put(&cfid->refcount, smb2_close_cached_fid);
+}
+
 /*
  * Called from cifs_kill_sb when we unmount a share
  */

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Yihang Li <[email protected]>

commit b32cc17d607e8ae7af037303fe101368cb4dc44c upstream.

Call scsi_device_put() in ata_scsi_dev_rescan() if the device or its queue are not running.

---
 drivers/ata/libata-scsi.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

@@ -4807,8 +4807,10 @@ void ata_scsi_dev_rescan(struct work_str
 			spin_unlock_irqrestore(ap->lock, flags);
 			if (do_resume) {
 				ret = scsi_resume_device(sdev);
-				if (ret == -EWOULDBLOCK)
+				if (ret == -EWOULDBLOCK) {
+					scsi_device_put(sdev);
 					goto unlock_scan;
+				}
 				dev->flags &= ~ATA_DFLAG_RESUMING;
 			}
 			ret = scsi_rescan_device(sdev);

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Andrey Vatoropin <[email protected]>

commit 7d277a7a58578dd62fd546ddaef459ec24ccae36 upstream.

be_insert_vlan_in_pkt() is called with the wrb_params argument being NULL at be_send_pkt_to_bmc() call site.  This may lead to dereferencing a NULL pointer when processing a workaround for specific packet, as commit bc0c3405abbb ("be2net: fix a Tx stall bug caused by a specific ipv6 packet") states.

The correct way would be to pass the wrb_params from be_xmit().

---
 drivers/net/ethernet/emulex/benet/be_main.c |    7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

@@ -1296,7 +1296,8 @@ static void be_xmit_flush(struct be_adap
 		(adapter->bmc_filt_mask & BMC_FILT_MULTICAST)
 
 static bool be_send_pkt_to_bmc(struct be_adapter *adapter,
-			       struct sk_buff **skb)
+			       struct sk_buff **skb,
+			       struct be_wrb_params *wrb_params)
 {
 	struct ethhdr *eh = (struct ethhdr *)(*skb)->data;
 	bool os2bmc = false;

@@ -1360,7 +1361,7 @@ done:
 	 * to BMC, asic expects the vlan to be inline in the packet.
 	 */
 	if (os2bmc)
-		*skb = be_insert_vlan_in_pkt(adapter, *skb, NULL);
+		*skb = be_insert_vlan_in_pkt(adapter, *skb, wrb_params);
 
 	return os2bmc;
 }

@@ -1387,7 +1388,7 @@ static netdev_tx_t be_xmit(struct sk_buf
 	/* if os2bmc is enabled and if the pkt is destined to bmc,
 	 * enqueue the pkt a 2nd time with mgmt bit set.
 	 */
-	if (be_send_pkt_to_bmc(adapter, &skb)) {
+	if (be_send_pkt_to_bmc(adapter, &skb, &wrb_params)) {
 		BE_WRB_F_SET(wrb_params.features, OS2BMC, 1);
 		wrb_cnt = be_xmit_enqueue(adapter, txo, skb, &wrb_params);
 		if (unlikely(!wrb_cnt))

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Oleksij Rempel <[email protected]>

commit 3ceb6ac2116ecda1c5d779bb73271479e70fccb4 upstream.

Correct RGMII delay application logic in lan937x_set_tune_adj().

The function was missing data16 &= ~PORT_TUNE_ADJ before setting the new delay value. This caused the new value to be bitwise-OR'd with the existing PORT_TUNE_ADJ field instead of replacing it.

For example, when setting the RGMII 2 TX delay on port 4, the intended TUNE_ADJUST value of 0 (RGMII_2_TX_DELAY_2NS) was incorrectly OR'd with the default 0x1B (from register value 0xDA3), leaving the delay at the wrong setting.

This patch adds the missing mask to clear the field, ensuring the correct delay value is written. Physical measurements on the RGMII TX lines confirm the fix, showing the delay changing from ~1ns (before change) to ~2ns.

While testing on i.MX 8MP showed this was within the platform's timing tolerance, it did not match the intended hardware-characterized value.

---
 drivers/net/dsa/microchip/lan937x_main.c |    1 +
 1 file changed, 1 insertion(+)

@@ -339,6 +339,7 @@ static void lan937x_set_tune_adj(struct
 	ksz_pread16(dev, port, reg, &data16);
 
 	/* Update tune Adjust */
+	data16 &= ~PORT_TUNE_ADJ;
 	data16 |= FIELD_PREP(PORT_TUNE_ADJ, val);
 	ksz_pwrite16(dev, port, reg, data16);
 

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Diogo Ivo <[email protected]>

commit 660b299bed2a2a55a1f9102d029549d0235f881c upstream.

Commit b6bcbce33596 ("soc/tegra: pmc: Ensure power-domains are in a known state") was introduced so that all power domains get initialized to a known working state when booting and it does this by shutting them down (including asserting resets and disabling clocks) before registering each power domain with the genpd framework, leaving it to each driver to later on power its needed domains.

This caused the Google Pixel C to hang when booting due to a workaround in the DSI driver introduced in commit b22fd0b9639e ("drm/tegra: dsi:
Clear enable register if powered by bootloader") meant to handle the case where the bootloader enabled the DSI hardware module. The workaround relies on reading a hardware register to determine the current status and after b6bcbce33596 that now happens in a powered down state thus leading to the boot hang.

Fix this by reverting b22fd0b9639e since currently we are guaranteed that the hardware will be fully reset by the time we start enabling the DSI module.

---
 drivers/gpu/drm/tegra/dsi.c |    9 ---------
 1 file changed, 9 deletions(-)

@@ -913,15 +913,6 @@ static void tegra_dsi_encoder_enable(str
 	u32 value;
 	int err;
 
-	/* If the bootloader enabled DSI it needs to be disabled
-	 * in order for the panel initialization commands to be
-	 * properly sent.
-	 */
-	value = tegra_dsi_readl(dsi, DSI_POWER_CONTROL);
-
-	if (value & DSI_POWER_CONTROL_ENABLE)
-		tegra_dsi_disable(dsi);
-
 	err = tegra_dsi_prepare(dsi);
 	if (err < 0) {
 		dev_err(dsi->dev, "failed to prepare: %d\n", err);

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Tzung-Bi Shih <[email protected]>

commit e08969c4d65ac31297fcb4d31d4808c789152f68 upstream.

If cros_ec_keyb_register_matrix() isn't called (due to buttons_switches_only) in cros_ec_keyb_probe(), ckdev->idev remains NULL. An invalid memory access is observed in cros_ec_keyb_process() when receiving an EC_MKBP_EVENT_KEY_MATRIX event in cros_ec_keyb_work() in such case.

  Unable to handle kernel read from unreadable memory at virtual address 0000000000000028
  ...
  x3 : 0000000000000000 x2 : 0000000000000000
  x1 : 0000000000000000 x0 : 0000000000000000
  Call trace:
  input_event
  cros_ec_keyb_work
  blocking_notifier_call_chain
  ec_irq_thread

It's still unknown about why the kernel receives such malformed event, in any cases, the kernel shouldn't access ckdev->idev and friends if the driver doesn't intend to initialize them.

---
 drivers/input/keyboard/cros_ec_keyb.c |    6 ++++++
 1 file changed, 6 insertions(+)

@@ -261,6 +261,12 @@ static int cros_ec_keyb_work(struct noti
 	case EC_MKBP_EVENT_KEY_MATRIX:
 		pm_wakeup_event(ckdev->dev, 0);
 
+		if (!ckdev->idev) {
+			dev_warn_once(ckdev->dev,
+				      "Unexpected key matrix event\n");
+			return NOTIFY_OK;
+		}
+
 		if (ckdev->ec->event_size != ckdev->cols) {
 			dev_err(ckdev->dev,
 				"Discarded incomplete key matrix event.\n");

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Hans de Goede <[email protected]>

commit c6d99e488117201c63efd747ce17b80687c3f5a9 upstream.

Some newer devices use an ACPI hardware ID of GDIX1003 for their Goodix touchscreen controller, instead of GDIX1001 / GDIX1002. Add GDIX1003 to the goodix_acpi_match[] table.

---
 drivers/input/touchscreen/goodix.c |    1 +
 1 file changed, 1 insertion(+)

@@ -1519,6 +1519,7 @@ MODULE_DEVICE_TABLE(i2c, goodix_ts_id);
 static const struct acpi_device_id goodix_acpi_match[] = {
 	{ "GDIX1001", 0 },
 	{ "GDIX1002", 0 },
+	{ "GDIX1003", 0 },
 	{ "GDX9110", 0 },
 	{ }
 };

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Dan Carpenter <[email protected]>

commit d83f1512758f4ef6fc5e83219fe7eeeb6b428ea4 upstream.

This is supposed to be "priv" but we accidentally pass "&priv" which is an address in the stack and so it will lead to memory corruption when the imx_sc_key_action() function is called. Remove the &.

---
 drivers/input/keyboard/imx_sc_key.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -158,7 +158,7 @@ static int imx_sc_key_probe(struct platf
 		return error;
 	}
 
-	error = devm_add_action_or_reset(&pdev->dev, imx_sc_key_action, &priv);
+	error = devm_add_action_or_reset(&pdev->dev, imx_sc_key_action, priv);
 	if (error)
 		return error;
 

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Seungjin Bae <[email protected]>

commit 69aeb507312306f73495598a055293fa749d454e upstream.

In the pegasus_notetaker driver, the pegasus_probe() function allocates the URB transfer buffer using the wMaxPacketSize value from the endpoint descriptor. An attacker can use a malicious USB descriptor to force the allocation of a very small buffer.

Subsequently, if the device sends an interrupt packet with a specific pattern (e.g., where the first byte is 0x80 or 0x42),
the pegasus_parse_packet() function parses the packet without checking the allocated buffer size. This leads to an out-of-bounds memory access.

---
 drivers/input/tablet/pegasus_notetaker.c |    9 +++++++++
 1 file changed, 9 insertions(+)

@@ -63,6 +63,9 @@
 #define BUTTON_PRESSED			0xb5
 #define COMMAND_VERSION			0xa9
 
+/* 1 Status + 1 Color + 2 X + 2 Y = 6 bytes */
+#define NOTETAKER_PACKET_SIZE		6
+
 /* in xy data packet */
 #define BATTERY_NO_REPORT		0x40
 #define BATTERY_LOW			0x41

@@ -303,6 +306,12 @@ static int pegasus_probe(struct usb_inte
 	}
 
 	pegasus->data_len = usb_maxpacket(dev, pipe);
+	if (pegasus->data_len < NOTETAKER_PACKET_SIZE) {
+		dev_err(&intf->dev, "packet size is too small (%d)\n",
+			pegasus->data_len);
+		error = -EINVAL;
+		goto err_free_mem;
+	}
 
 	pegasus->data = usb_alloc_coherent(dev, pegasus->data_len, GFP_KERNEL,
 					   &pegasus->data_dma);

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Vlastimil Babka <[email protected]>

commit ec33b59542d96830e3c89845ff833cf7b25ef172 upstream.

The kernel test has reported:

  BUG: unable to handle page fault for address: fffba000
  #PF: supervisor write access in kernel mode
  #PF: error_code(0x0002) - not-present page
  *pde = 03171067 *pte = 00000000
  Oops: Oops: 0002 [#1]
  CPU: 0 UID: 0 PID: 1 Comm: swapper/0 Tainted: G                T   6.18.0-rc2-00031-gec7f31b2a2d3 #1 NONE  a1d066dfe789f54bc7645c7989957d2bdee593ca
  Tainted: [T]=RANDSTRUCT
  Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.3-debian-1.16.3-2 04/01/2014
  EIP: memset (arch/x86/include/asm/string_32.h:168 arch/x86/lib/memcpy_32.c:17)
  Code: a5 8b 4d f4 83 e1 03 74 02 f3 a4 83 c4 04 5e 5f 5d 2e e9 73 41 01 00 90 90 90 3e 8d 74 26 00 55 89 e5 57 56 89 c6 89 d0 89 f7 <f3> aa 89 f0 5e 5f 5d 2e e9 53 41 01 00 cc cc cc 55 89 e5 53 57 56
  EAX: 0000006b EBX: 00000015 ECX: 001fefff EDX: 0000006b
  ESI: fffb9000 EDI: fffba000 EBP: c611fbf0 ESP: c611fbe8
  DS: 007b ES: 007b FS: 0000 GS: 0000 SS: 0068 EFLAGS: 00010287
  CR0: 80050033 CR2: fffba000 CR3: 0316e000 CR4: 00040690
  Call Trace:
   poison_element (mm/mempool.c:83 mm/mempool.c:102)
   mempool_init_node (mm/mempool.c:142 mm/mempool.c:226)
   mempool_init_noprof (mm/mempool.c:250 (discriminator 1))
   ? mempool_alloc_pages (mm/mempool.c:640)
   bio_integrity_initfn (block/bio-integrity.c:483 (discriminator 8))
   ? mempool_alloc_pages (mm/mempool.c:640)
   do_one_initcall (init/main.c:1283)

Christoph found out this is due to the poisoning code not dealing properly with CONFIG_HIGHMEM because only the first page is mapped but then the whole potentially high-order page is accessed.

We could give up on HIGHMEM here, but it's straightforward to fix this with a loop that's mapping, poisoning or checking and unmapping individual pages.

---
 mm/mempool.c |   32 ++++++++++++++++++++++++++------
 1 file changed, 26 insertions(+), 6 deletions(-)

@@ -68,10 +68,20 @@ static void check_element(mempool_t *poo
 	} else if (pool->free == mempool_free_pages) {
 		/* Mempools backed by page allocator */
 		int order = (int)(long)pool->pool_data;
-		void *addr = kmap_local_page((struct page *)element);
 
-		__check_element(pool, addr, 1UL << (PAGE_SHIFT + order));
-		kunmap_local(addr);
+#ifdef CONFIG_HIGHMEM
+		for (int i = 0; i < (1 << order); i++) {
+			struct page *page = (struct page *)element;
+			void *addr = kmap_local_page(page + i);
+
+			__check_element(pool, addr, PAGE_SIZE);
+			kunmap_local(addr);
+		}
+#else
+		void *addr = page_address((struct page *)element);
+
+		__check_element(pool, addr, PAGE_SIZE << order);
+#endif
 	}
 }
 

@@ -97,10 +107,20 @@ static void poison_element(mempool_t *po
 	} else if (pool->alloc == mempool_alloc_pages) {
 		/* Mempools backed by page allocator */
 		int order = (int)(long)pool->pool_data;
-		void *addr = kmap_local_page((struct page *)element);
 
-		__poison_element(addr, 1UL << (PAGE_SHIFT + order));
-		kunmap_local(addr);
+#ifdef CONFIG_HIGHMEM
+		for (int i = 0; i < (1 << order); i++) {
+			struct page *page = (struct page *)element;
+			void *addr = kmap_local_page(page + i);
+
+			__poison_element(addr, PAGE_SIZE);
+			kunmap_local(addr);
+		}
+#else
+		void *addr = page_address((struct page *)element);
+
+		__poison_element(addr, PAGE_SIZE << order);
+#endif
 	}
 }
 #else /* CONFIG_SLUB_DEBUG_ON */

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Nam Cao <[email protected]>

commit 949f1fd2225baefbea2995afa807dba5cbdb6bd3 upstream.

nvkm_falcon_fw::boot is allocated, but no one frees it. This causes a kmemleak warning.

Make sure this data is deallocated.

---
 drivers/gpu/drm/nouveau/nvkm/falcon/fw.c |    2 ++
 1 file changed, 2 insertions(+)

@@ -159,6 +159,8 @@ nvkm_falcon_fw_dtor(struct nvkm_falcon_f
 	nvkm_memory_unref(&fw->inst);
 	nvkm_falcon_fw_dtor_sigs(fw);
 	nvkm_firmware_dtor(&fw->fw);
+	kfree(fw->boot);
+	fw->boot = NULL;
 }
 
 static const struct nvkm_firmware_func

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Ewan D. Milne <[email protected]>

commit ea3442efabd0aa3930c5bab73c3901ef38ef6ac3 upstream.

Now target is removed from nvme_fc_ctrl_free() which is the ctrl->ref release handler. And even admin queue is unquiesced there, this way is definitely wrong because the ctr->ref is grabbed when submitting command.

And Marco observed that nvme_fc_ctrl_free() can be called from request completion code path, and trigger kernel warning since request completes from softirq context.

Fix the issue by moveing target removal into nvme_fc_delete_ctrl(), which is also aligned with nvme-tcp and nvme-rdma.

Patch originally proposed by Ming Lei, then modified to move the tagset removal down to after nvme_fc_delete_association() after further testing.

---
 drivers/nvme/host/fc.c |   13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

@@ -2349,17 +2349,11 @@ nvme_fc_ctrl_free(struct kref *ref)
 		container_of(ref, struct nvme_fc_ctrl, ref);
 	unsigned long flags;
 
-	if (ctrl->ctrl.tagset)
-		nvme_remove_io_tag_set(&ctrl->ctrl);
-
 	/* remove from rport list */
 	spin_lock_irqsave(&ctrl->rport->lock, flags);
 	list_del(&ctrl->ctrl_list);
 	spin_unlock_irqrestore(&ctrl->rport->lock, flags);
 
-	nvme_unquiesce_admin_queue(&ctrl->ctrl);
-	nvme_remove_admin_tag_set(&ctrl->ctrl);
-
 	kfree(ctrl->queues);
 
 	put_device(ctrl->dev);

@@ -3255,11 +3249,18 @@ nvme_fc_delete_ctrl(struct nvme_ctrl *nc
 
 	cancel_work_sync(&ctrl->ioerr_work);
 	cancel_delayed_work_sync(&ctrl->connect_work);
+
 	/*
 	 * kill the association on the link side.  this will block
 	 * waiting for io to terminate
 	 */
 	nvme_fc_delete_association(ctrl);
+
+	if (ctrl->ctrl.tagset)
+		nvme_remove_io_tag_set(&ctrl->ctrl);
+
+	nvme_unquiesce_admin_queue(&ctrl->ctrl);
+	nvme_remove_admin_tag_set(&ctrl->ctrl);
 }
 
 static void

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Ewan D. Milne <[email protected]>

commit 0a2c5495b6d1ecb0fa18ef6631450f391a888256 upstream.

nvme_fc_delete_assocation() waits for pending I/O to complete before returning, and an error can cause ->ioerr_work to be queued after cancel_work_sync() had been called. Move the call to cancel_work_sync() to be after nvme_fc_delete_association() to ensure ->ioerr_work is not running when the nvme_fc_ctrl object is freed. Otherwise the following can occur:

[ 1135.911754] list_del corruption, ff2d24c8093f31f8->next is NULL
[ 1135.917705] ------------[ cut here ]------------
[ 1135.922336] kernel BUG at lib/list_debug.c:52!
[ 1135.926784] Oops: invalid opcode: 0000 [#1] SMP NOPTI
[ 1135.931851] CPU: 48 UID: 0 PID: 726 Comm: kworker/u449:23 Kdump: loaded Not tainted 6.12.0 #1 PREEMPT(voluntary)
[ 1135.943490] Hardware name: Dell Inc. PowerEdge R660/0HGTK9, BIOS 2.5.4 01/16/2025
[ 1135.950969] Workqueue:  0x0 (nvme-wq)
[ 1135.954673] RIP: 0010:__list_del_entry_valid_or_report.cold+0xf/0x6f
[ 1135.961041] Code: c7 c7 98 68 72 94 e8 26 45 fe ff 0f 0b 48 c7 c7 70 68 72 94 e8 18 45 fe ff 0f 0b 48 89 fe 48 c7 c7 80 69 72 94 e8 07 45 fe ff <0f> 0b 48 89 d1 48 c7 c7 a0 6a 72 94 48 89 c2 e8 f3 44 fe ff 0f 0b
[ 1135.979788] RSP: 0018:ff579b19482d3e50 EFLAGS: 00010046
[ 1135.985015] RAX: 0000000000000033 RBX: ff2d24c8093f31f0 RCX: 0000000000000000
[ 1135.992148] RDX: 0000000000000000 RSI: ff2d24d6bfa1d0c0 RDI: ff2d24d6bfa1d0c0
[ 1135.999278] RBP: ff2d24c8093f31f8 R08: 0000000000000000 R09: ffffffff951e2b08
[ 1136.006413] R10: ffffffff95122ac8 R11: 0000000000000003 R12: ff2d24c78697c100
[ 1136.013546] R13: fffffffffffffff8 R14: 0000000000000000 R15: ff2d24c78697c0c0
[ 1136.020677] FS:  0000000000000000(0000) GS:ff2d24d6bfa00000(0000) knlGS:0000000000000000
[ 1136.028765] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1136.034510] CR2: 00007fd207f90b80 CR3: 000000163ea22003 CR4: 0000000000f73ef0
[ 1136.041641] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 1136.048776] DR3: 0000000000000000 DR6: 00000000fffe07f0 DR7: 0000000000000400
[ 1136.055910] PKRU: 55555554
[ 1136.058623] Call Trace:
[ 1136.061074]  <TASK>
[ 1136.063179]  ? show_trace_log_lvl+0x1b0/0x2f0
[ 1136.067540]  ? show_trace_log_lvl+0x1b0/0x2f0
[ 1136.071898]  ? move_linked_works+0x4a/0xa0
[ 1136.075998]  ? __list_del_entry_valid_or_report.cold+0xf/0x6f
[ 1136.081744]  ? __die_body.cold+0x8/0x12
[ 1136.085584]  ? die+0x2e/0x50
[ 1136.088469]  ? do_trap+0xca/0x110
[ 1136.091789]  ? do_error_trap+0x65/0x80
[ 1136.095543]  ? __list_del_entry_valid_or_report.cold+0xf/0x6f
[ 1136.101289]  ? exc_invalid_op+0x50/0x70
[ 1136.105127]  ? __list_del_entry_valid_or_report.cold+0xf/0x6f
[ 1136.110874]  ? asm_exc_invalid_op+0x1a/0x20
[ 1136.115059]  ? __list_del_entry_valid_or_report.cold+0xf/0x6f
[ 1136.120806]  move_linked_works+0x4a/0xa0
[ 1136.124733]  worker_thread+0x216/0x3a0
[ 1136.128485]  ? __pfx_worker_thread+0x10/0x10
[ 1136.132758]  kthread+0xfa/0x240
[ 1136.135904]  ? __pfx_kthread+0x10/0x10
[ 1136.139657]  ret_from_fork+0x31/0x50
[ 1136.143236]  ? __pfx_kthread+0x10/0x10
[ 1136.146988]  ret_from_fork_asm+0x1a/0x30
[ 1136.150915]  </TASK>

---
 drivers/nvme/host/fc.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -3247,7 +3247,6 @@ nvme_fc_delete_ctrl(struct nvme_ctrl *nc
 {
 	struct nvme_fc_ctrl *ctrl = to_fc_ctrl(nctrl);
 
-	cancel_work_sync(&ctrl->ioerr_work);
 	cancel_delayed_work_sync(&ctrl->connect_work);
 
 	/*

@@ -3255,6 +3254,7 @@ nvme_fc_delete_ctrl(struct nvme_ctrl *nc
 	 * waiting for io to terminate
 	 */
 	nvme_fc_delete_association(ctrl);
+	cancel_work_sync(&ctrl->ioerr_work);
 
 	if (ctrl->ctrl.tagset)
 		nvme_remove_io_tag_set(&ctrl->ctrl);

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Bart Van Assche <[email protected]>

commit 90449f2d1e1f020835cba5417234636937dd657e upstream.

sg_finish_rem_req() calls blk_rq_unmap_user(). The latter function may sleep. Hence, call sg_finish_rem_req() with interrupts enabled instead of disabled.

---
 drivers/scsi/sg.c |   10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

@@ -2210,9 +2210,17 @@ sg_remove_sfp_usercontext(struct work_st
 	write_lock_irqsave(&sfp->rq_list_lock, iflags);
 	while (!list_empty(&sfp->rq_list)) {
 		srp = list_first_entry(&sfp->rq_list, Sg_request, entry);
-		sg_finish_rem_req(srp);
 		list_del(&srp->entry);
+		write_unlock_irqrestore(&sfp->rq_list_lock, iflags);
+
+		sg_finish_rem_req(srp);
+		/*
+		 * sg_rq_end_io() uses srp->parentfp. Hence, only clear
+		 * srp->parentfp after blk_mq_free_request() has been called.
+		 */
 		srp->parentfp = NULL;
+
+		write_lock_irqsave(&sfp->rq_list_lock, iflags);
 	}
 	write_unlock_irqrestore(&sfp->rq_list_lock, iflags);
 

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Hamza Mahfooz <[email protected]>

commit e6965188f84a7883e6a0d3448e86b0cf29b24dfc upstream.

If the allocation of tl_hba->sh fails in tcm_loop_driver_probe() and we attempt to dereference it in tcm_loop_tpg_address_show() we will get a segfault, see below for an example. So, check tl_hba->sh before dereferencing it.

Unable to allocate struct scsi_host

  BUG: kernel NULL pointer dereference, address: 0000000000000194
  #PF: supervisor read access in kernel mode
  #PF: error_code(0x0000) - not-present page
  PGD 0 P4D 0
  Oops: 0000 [#1] PREEMPT SMP NOPTI
  CPU: 1 PID: 8356 Comm: tokio-runtime-w Not tainted 6.6.104.2-4.azl3 #1
  Hardware name: Microsoft Corporation Virtual Machine/Virtual Machine, BIOS Hyper-V UEFI Release v4.1 09/28/2024
  RIP: 0010:tcm_loop_tpg_address_show+0x2e/0x50 [tcm_loop]
...
  Call Trace:
   <TASK>
   configfs_read_iter+0x12d/0x1d0 [configfs]
   vfs_read+0x1b5/0x300
   ksys_read+0x6f/0xf0

...

---
 drivers/target/loopback/tcm_loop.c |    3 +++
 1 file changed, 3 insertions(+)

@@ -893,6 +893,9 @@ static ssize_t tcm_loop_tpg_address_show
 			struct tcm_loop_tpg, tl_se_tpg);
 	struct tcm_loop_hba *tl_hba = tl_tpg->tl_hba;
 
+	if (!tl_hba->sh)
+		return -ENODEV;
+
 	return snprintf(page, PAGE_SIZE, "%d:0:%d\n",
 			tl_hba->sh->host_no, tl_tpg->tl_tpgt);
 }

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Maciej W. Rozycki <[email protected]>

commit ebd729fef31620e0bf74cbf8a4c7fda73a2a4e7e upstream.

Fix a regression that has caused accesses to the PCI MMIO window to complete unclaimed in non-EVA configurations with the SOC-it family of system controllers, preventing PCI devices from working that use MMIO.

In the non-EVA case PHYS_OFFSET is set to 0, meaning that PCI_BAR0 is set with an empty mask (and PCI_HEAD4 matches addresses starting from 0 accordingly). Consequently all addresses are matched for incoming DMA accesses from PCI. This seems to confuse the system controller's logic and outgoing bus cycles targeting the PCI MMIO window seem not to make it to the intended devices.

This happens as well when a wider mask is used with PCI_BAR0, such as 0x80000000 or 0xe0000000, that makes addresses match that overlap with the PCI MMIO window, which starts at 0x10000000 in our configuration.

Set the mask in PCI_BAR0 to 0xf0000000 for non-EVA then, covering the non-EVA maximum 256 MiB of RAM, which is what YAMON does and which used to work correctly up to the offending commit. Set PCI_P2SCMSKL to match PCI_BAR0 as required by the system controller's specification, and match PCI_P2SCMAPL to PCI_HEAD4 for identity mapping.

Verified with:

Core board type/revision =      0x0d (Core74K) / 0x01
System controller/revision =    MIPS SOC-it 101 OCP / 1.3   SDR-FW-4:1
Processor Company ID/options =  0x01 (MIPS Technologies, Inc.) / 0x1c
Processor ID/revision =         0x97 (MIPS 74Kf) / 0x4c

for non-EVA and with:

Core board type/revision =      0x0c (CoreFPGA-5) / 0x00
System controller/revision =    MIPS ROC-it2 / 0.0   FW-1:1 (CLK_unknown) GIC
Processor Company ID/options =  0x01 (MIPS Technologies, Inc.) / 0x00
Processor ID/revision =         0xa0 (MIPS interAptiv UP) / 0x20

for EVA/non-EVA, fixing:

defxx 0000:00:12.0: assign IRQ: got 10
defxx: v1.12 2021/03/10 Lawrence V. Stefani and others 0000:00:12.0: Could not read adapter factory MAC address!

vs:

defxx 0000:00:12.0: assign IRQ: got 10
defxx: v1.12 2021/03/10 Lawrence V. Stefani and others
0000:00:12.0: DEFPA at MMIO addr = 0x10142000, IRQ = 10, Hardware addr = 00-00-f8-xx-xx-xx 0000:00:12.0: registered as fddi0

for non-EVA and causing no change for EVA.

---
 arch/mips/mti-malta/malta-init.c |   20 +++++++++++++-------
 1 file changed, 13 insertions(+), 7 deletions(-)

@@ -241,16 +241,22 @@ mips_pci_controller:
 #endif
 
 		/*
-		 * Setup the Malta max (2GB) memory for PCI DMA in host bridge
-		 * in transparent addressing mode.
+		 * Set up memory mapping in host bridge for PCI DMA masters,
+		 * in transparent addressing mode.  For EVA use the Malta
+		 * maximum of 2 GiB memory in the alias space at 0x80000000
+		 * as per PHYS_OFFSET.  Otherwise use 256 MiB of memory in
+		 * the regular space, avoiding mapping the PCI MMIO window
+		 * for DMA as it seems to confuse the system controller's
+		 * logic, causing PCI MMIO to stop working.
 		 */
-		mask = PHYS_OFFSET | PCI_BASE_ADDRESS_MEM_PREFETCH;
-		MSC_WRITE(MSC01_PCI_BAR0, mask);
-		MSC_WRITE(MSC01_PCI_HEAD4, mask);
+		mask = PHYS_OFFSET ? PHYS_OFFSET : 0xf0000000;
+		MSC_WRITE(MSC01_PCI_BAR0,
+			  mask | PCI_BASE_ADDRESS_MEM_PREFETCH);
+		MSC_WRITE(MSC01_PCI_HEAD4,
+			  PHYS_OFFSET | PCI_BASE_ADDRESS_MEM_PREFETCH);
 
-		mask &= MSC01_PCI_BAR0_SIZE_MSK;
 		MSC_WRITE(MSC01_PCI_P2SCMSKL, mask);
-		MSC_WRITE(MSC01_PCI_P2SCMAPL, mask);
+		MSC_WRITE(MSC01_PCI_P2SCMAPL, PHYS_OFFSET);
 
 		/* Don't handle target retries indefinitely.  */
 		if ((data & MSC01_PCI_CFG_MAXRTRY_MSK) ==

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Krzysztof Kozlowski <[email protected]>

commit 316e361b5d2cdeb8d778983794a1c6eadcb26814 upstream.

The "groups" property can hold multiple entries (e.g. toshiba/tmpv7708-rm-mbrc.dts file), so allow that by dropping incorrect type (pinmux-node.yaml schema already defines that as string-array) and adding constraints for items. This fixes dtbs_check warnings like:

toshiba/tmpv7708-rm-mbrc.dtb: pinctrl@24190000 (toshiba,tmpv7708-pinctrl):

    pwm-pins:groups: ['pwm0_gpio16_grp', 'pwm1_gpio17_grp', 'pwm2_gpio18_grp', 'pwm3_gpio19_grp'] is too long

---
 Documentation/devicetree/bindings/pinctrl/toshiba,visconti-pinctrl.yaml |   26 +++++-----
 1 file changed, 14 insertions(+), 12 deletions(-)

@@ -50,18 +50,20 @@ patternProperties:
       groups:
         description:
           Name of the pin group to use for the functions.
-        $ref: /schemas/types.yaml#/definitions/string
-        enum: [i2c0_grp, i2c1_grp, i2c2_grp, i2c3_grp, i2c4_grp,
-               i2c5_grp, i2c6_grp, i2c7_grp, i2c8_grp,
-               spi0_grp, spi0_cs0_grp, spi0_cs1_grp, spi0_cs2_grp,
-               spi1_grp, spi2_grp, spi3_grp, spi4_grp, spi5_grp, spi6_grp,
-               uart0_grp, uart1_grp, uart2_grp, uart3_grp,
-               pwm0_gpio4_grp, pwm0_gpio8_grp, pwm0_gpio12_grp,
-               pwm0_gpio16_grp, pwm1_gpio5_grp, pwm1_gpio9_grp,
-               pwm1_gpio13_grp, pwm1_gpio17_grp, pwm2_gpio6_grp,
-               pwm2_gpio10_grp, pwm2_gpio14_grp, pwm2_gpio18_grp,
-               pwm3_gpio7_grp, pwm3_gpio11_grp, pwm3_gpio15_grp,
-               pwm3_gpio19_grp, pcmif_out_grp, pcmif_in_grp]
+        items:
+          enum: [i2c0_grp, i2c1_grp, i2c2_grp, i2c3_grp, i2c4_grp,
+                 i2c5_grp, i2c6_grp, i2c7_grp, i2c8_grp,
+                 spi0_grp, spi0_cs0_grp, spi0_cs1_grp, spi0_cs2_grp,
+                 spi1_grp, spi2_grp, spi3_grp, spi4_grp, spi5_grp, spi6_grp,
+                 uart0_grp, uart1_grp, uart2_grp, uart3_grp,
+                 pwm0_gpio4_grp, pwm0_gpio8_grp, pwm0_gpio12_grp,
+                 pwm0_gpio16_grp, pwm1_gpio5_grp, pwm1_gpio9_grp,
+                 pwm1_gpio13_grp, pwm1_gpio17_grp, pwm2_gpio6_grp,
+                 pwm2_gpio10_grp, pwm2_gpio14_grp, pwm2_gpio18_grp,
+                 pwm3_gpio7_grp, pwm3_gpio11_grp, pwm3_gpio15_grp,
+                 pwm3_gpio19_grp, pcmif_out_grp, pcmif_in_grp]
+        minItems: 1
+        maxItems: 8
 
       drive-strength:
         enum: [2, 4, 6, 8, 16, 24, 32]

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Huacai Chen <[email protected]>

commit a6b533adfc05ba15360631e019d3e18275080275 upstream.

If there is no valid cache info detected (may happen in virtual machine) for pci_dfl_cache_line_size, kernel shouldn't panic. Because in the PCI core it will be evaluated to (L1_CACHE_BYTES >> 2).

---
 arch/loongarch/pci/pci.c |    8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

@@ -51,11 +51,11 @@ static int __init pcibios_init(void)
 	 */
 	lsize = cpu_last_level_cache_line_size();
 
-	BUG_ON(!lsize);
+	if (lsize) {
+		pci_dfl_cache_line_size = lsize >> 2;
 
-	pci_dfl_cache_line_size = lsize >> 2;
-
-	pr_debug("PCI: pci_cache_line_size set to %d bytes\n", lsize);
+		pr_debug("PCI: pci_cache_line_size set to %d bytes\n", lsize);
+	}
 
 	return 0;
 }

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Eric Dumazet <[email protected]>

commit 035bca3f017ee9dea3a5a756e77a6f7138cc6eea upstream.

syzbot reported use-after-free in mptcp_schedule_work() [1]

Issue here is that mptcp_schedule_work() schedules a work, then gets a refcount on sk->sk_refcnt if the work was scheduled. This refcount will be released by mptcp_worker().

[A] if (schedule_work(...)) {

[B]     sock_hold(sk);
        return true;
    }

Problem is that mptcp_worker() can run immediately and complete before [B]

We need instead :

    sock_hold(sk);
    if (schedule_work(...))
        return true;
    sock_put(sk);

[1]
refcount_t: addition on 0; use-after-free.

 WARNING: CPU: 1 PID: 29 at lib/refcount.c:25 refcount_warn_saturate+0xfa/0x1d0 lib/refcount.c:25
Call Trace:
 <TASK>
 __refcount_add include/linux/refcount.h:-1 [inline]
  __refcount_inc include/linux/refcount.h:366 [inline]
  refcount_inc include/linux/refcount.h:383 [inline]
  sock_hold include/net/sock.h:816 [inline]
  mptcp_schedule_work+0x164/0x1a0 net/mptcp/protocol.c:943
  mptcp_tout_timer+0x21/0xa0 net/mptcp/protocol.c:2316
  call_timer_fn+0x17e/0x5f0 kernel/time/timer.c:1747
  expire_timers kernel/time/timer.c:1798 [inline]
  __run_timers kernel/time/timer.c:2372 [inline]
  __run_timer_base+0x648/0x970 kernel/time/timer.c:2384
  run_timer_base kernel/time/timer.c:2393 [inline]
  run_timer_softirq+0xb7/0x180 kernel/time/timer.c:2403
  handle_softirqs+0x22f/0x710 kernel/softirq.c:622
  __do_softirq kernel/softirq.c:656 [inline]
  run_ktimerd+0xcf/0x190 kernel/softirq.c:1138
  smpboot_thread_fn+0x542/0xa60 kernel/smpboot.c:160
  kthread+0x711/0x8a0 kernel/kthread.c:463
  ret_from_fork+0x4bc/0x870 arch/x86/kernel/process.c:158
  ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:245

---
 net/mptcp/protocol.c |   19 ++++++++++++-------
 1 file changed, 12 insertions(+), 7 deletions(-)

@@ -963,14 +963,19 @@ static void mptcp_reset_rtx_timer(struct
 
 bool mptcp_schedule_work(struct sock *sk)
 {
-	if (inet_sk_state_load(sk) != TCP_CLOSE &&
-	    schedule_work(&mptcp_sk(sk)->work)) {
-		/* each subflow already holds a reference to the sk, and the
-		 * workqueue is invoked by a subflow, so sk can't go away here.
-		 */
-		sock_hold(sk);
+	if (inet_sk_state_load(sk) == TCP_CLOSE)
+		return false;
+
+	/* Get a reference on this socket, mptcp_worker() will release it.
+	 * As mptcp_worker() might complete before us, we can not avoid
+	 * a sock_hold()/sock_put() if schedule_work() returns false.
+	 */
+	sock_hold(sk);
+
+	if (schedule_work(&mptcp_sk(sk)->work))
 		return true;
-	}
+
+	sock_put(sk);
 	return false;
 }
 

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Paolo Abeni <[email protected]>

commit 5e15395f6d9ec07395866c5511f4b4ac566c0c9b upstream.

mptcp_cleanup_rbuf() needs to know the last most recent, mptcp-level rcv_wnd sent, and such information is tracked into the msk->old_wspace field, updated at ack transmission time by mptcp_write_options().

Fallback socket do not add any mptcp options, such helper is never invoked, and msk->old_wspace value remain stale. That in turn makes ack generation at recvmsg() time quite random.

Address the issue ensuring mptcp_write_options() is invoked even for fallback sockets, and just update the needed info in such a case.

The issue went unnoticed for a long time, as mptcp currently overshots the fallback socket receive buffer autotune significantly. It is going to change in the near future.

---
 net/mptcp/options.c |   23 ++++++++++++++++++++++-
 1 file changed, 22 insertions(+), 1 deletion(-)

@@ -839,8 +839,11 @@ bool mptcp_established_options(struct so
 
 	opts->suboptions = 0;
 
+	/* Force later mptcp_write_options(), but do not use any actual
+	 * option space.
+	 */
 	if (unlikely(__mptcp_check_fallback(msk) && !mptcp_check_infinite_map(skb)))
-		return false;
+		return true;
 
 	if (unlikely(skb && TCP_SKB_CB(skb)->tcp_flags & TCPHDR_RST)) {
 		if (mptcp_established_options_fastclose(sk, &opt_size, remaining, opts) ||

@@ -1319,6 +1322,20 @@ update_wspace:
 	WRITE_ONCE(msk->old_wspace, tp->rcv_wnd);
 }
 
+static void mptcp_track_rwin(struct tcp_sock *tp)
+{
+	const struct sock *ssk = (const struct sock *)tp;
+	struct mptcp_subflow_context *subflow;
+	struct mptcp_sock *msk;
+
+	if (!ssk)
+		return;
+
+	subflow = mptcp_subflow_ctx(ssk);
+	msk = mptcp_sk(subflow->conn);
+	WRITE_ONCE(msk->old_wspace, tp->rcv_wnd);
+}
+
 __sum16 __mptcp_make_csum(u64 data_seq, u32 subflow_seq, u16 data_len, __wsum sum)
 {
 	struct csum_pseudo_header header;

@@ -1611,6 +1628,10 @@ mp_rst:
 				      opts->reset_transient,
 				      opts->reset_reason);
 		return;
+	} else if (unlikely(!opts->suboptions)) {
+		/* Fallback to TCP */
+		mptcp_track_rwin(tp);
+		return;
 	}
 
 	if (OPTION_MPTCP_PRIO & opts->suboptions) {

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Paolo Abeni <[email protected]>

commit ae155060247be8dcae3802a95bd1bdf93ab3215d upstream.

The CI reports sporadic failures of the fastclose self-tests. The root cause is a duplicate reset, not carrying the relevant MPTCP option. In the failing scenario the bad reset is received by the peer before the fastclose one, preventing the reception of the latter.

Indeed there is window of opportunity at fastclose time for the following race:

mptcp_do_fastclose

    __mptcp_close_ssk
      __tcp_close()
        tcp_set_state() [1]
        tcp_send_active_reset() [2]

After [1] the stack will send reset to in-flight data reaching the now closed port. Such reset may race with [2].

Address the issue explicitly sending a single reset on fastclose before explicitly moving the subflow to close status.

---
 net/mptcp/protocol.c |   36 +++++++++++++++++++++++-------------
 1 file changed, 23 insertions(+), 13 deletions(-)

@@ -2461,7 +2461,6 @@ bool __mptcp_retransmit_pending_data(str
 
 /* flags for __mptcp_close_ssk() */
 #define MPTCP_CF_PUSH		BIT(1)
-#define MPTCP_CF_FASTCLOSE	BIT(2)
 
 /* be sure to send a reset only if the caller asked for it, also
  * clean completely the subflow status when the subflow reaches

@@ -2472,7 +2471,7 @@ static void __mptcp_subflow_disconnect(s
 				       unsigned int flags)
 {
 	if (((1 << ssk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN)) ||
-	    (flags & MPTCP_CF_FASTCLOSE)) {
+	    subflow->send_fastclose) {
 		/* The MPTCP code never wait on the subflow sockets, TCP-level
 		 * disconnect should never fail
 		 */

@@ -2519,14 +2518,8 @@ static void __mptcp_close_ssk(struct soc
 
 	lock_sock_nested(ssk, SINGLE_DEPTH_NESTING);
 
-	if ((flags & MPTCP_CF_FASTCLOSE) && !__mptcp_check_fallback(msk)) {
-		/* be sure to force the tcp_close path
-		 * to generate the egress reset
-		 */
-		ssk->sk_lingertime = 0;
-		sock_set_flag(ssk, SOCK_LINGER);
-		subflow->send_fastclose = 1;
-	}
+	if (subflow->send_fastclose && ssk->sk_state != TCP_CLOSE)
+		tcp_set_state(ssk, TCP_CLOSE);
 
 	need_push = (flags & MPTCP_CF_PUSH) && __mptcp_retransmit_pending_data(sk);
 	if (!dispose_it) {

@@ -2829,9 +2822,26 @@ static void mptcp_do_fastclose(struct so
 	struct mptcp_sock *msk = mptcp_sk(sk);
 
 	mptcp_set_state(sk, TCP_CLOSE);
-	mptcp_for_each_subflow_safe(msk, subflow, tmp)
-		__mptcp_close_ssk(sk, mptcp_subflow_tcp_sock(subflow),
-				  subflow, MPTCP_CF_FASTCLOSE);
+
+	/* Explicitly send the fastclose reset as need */
+	if (__mptcp_check_fallback(msk))
+		return;
+
+	mptcp_for_each_subflow_safe(msk, subflow, tmp) {
+		struct sock *ssk = mptcp_subflow_tcp_sock(subflow);
+
+		lock_sock(ssk);
+
+		/* Some subflow socket states don't allow/need a reset.*/
+		if ((1 << ssk->sk_state) & (TCPF_LISTEN | TCPF_CLOSE))
+			goto unlock;
+
+		subflow->send_fastclose = 1;
+		tcp_send_active_reset(ssk, ssk->sk_allocation,
+				      SK_RST_REASON_TCP_ABORT_ON_CLOSE);
+unlock:
+		release_sock(ssk);
+	}
 }
 
 static void mptcp_worker(struct work_struct *work)

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Paolo Abeni <[email protected]>

commit 17393fa7b7086664be519e7230cb6ed7ec7d9462 upstream.

I'm observing very frequent self-tests failures in case of fallback when running on a CONFIG_PREEMPT kernel.

The root cause is that subflow_sched_work_if_closed() closes any subflow as soon as it is half-closed and has no incoming data pending.

That works well for regular subflows - MPTCP needs bi-directional connectivity to operate on a given subflow - but for fallback socket is race prone.

When TCP peer closes the connection before the MPTCP one, subflow_sched_work_if_closed() will schedule the MPTCP worker to gracefully close the subflow, and shortly after will do another schedule to inject and process a dummy incoming DATA_FIN.

On CONFIG_PREEMPT kernel, the MPTCP worker can kick-in and close the fallback subflow before subflow_sched_work_if_closed() is able to create the dummy DATA_FIN, unexpectedly interrupting the transfer.

Address the issue explicitly avoiding closing fallback subflows on when the peer is only half-closed.

Note that, when the subflow is able to create the DATA_FIN before the worker invocation, the worker will change the msk state before trying to close the subflow and will skip the latter operation as the msk will not match anymore the precondition in __mptcp_close_subflow().

---
 net/mptcp/protocol.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

@@ -2615,7 +2615,8 @@ static void __mptcp_close_subflow(struct
 
 		if (ssk_state != TCP_CLOSE &&
 		    (ssk_state != TCP_CLOSE_WAIT ||
-		     inet_sk_state_load(sk) != TCP_ESTABLISHED))
+		     inet_sk_state_load(sk) != TCP_ESTABLISHED ||
+		     __mptcp_check_fallback(msk)))
 			continue;
 
 		/* 'subflow_data_ready' will re-sched once rx queue is empty */

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Matthieu Baerts (NGI0) <[email protected]>

commit fb13c6bb810ca871964e062cf91882d1c83db509 upstream.

In rare cases, when the test environment is very slow, some endpoints tests can fail because some expected events have not been seen.

Because the tests are expecting a long on-going connection, and they are not waiting for the end of the transfer, it is fine to have a longer timeout, and even go over the default one. This connection will be killed at the end, after the verifications: increasing the timeout doesn't change anything, apart from avoiding it to end before the end of the verifications.

To play it safe, all endpoints tests not waiting for the end of the transfer are now having a longer timeout: 2 minutes.

The Fixes commit was making the connection longer, but still, the default timeout would have stopped it after 1 minute, which might not be enough in very slow environments.

---
 tools/testing/selftests/net/mptcp/mptcp_join.sh |    8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

@@ -3728,7 +3728,7 @@ endpoint_tests()
 		pm_nl_set_limits $ns1 2 2
 		pm_nl_set_limits $ns2 2 2
 		pm_nl_add_endpoint $ns1 10.0.2.1 flags signal
-		{ test_linkfail=128 speed=slow \
+		{ timeout_test=120 test_linkfail=128 speed=slow \
 			run_tests $ns1 $ns2 10.0.1.1 & } 2>/dev/null
 		local tests_pid=$!
 

@@ -3755,7 +3755,7 @@ endpoint_tests()
 		pm_nl_set_limits $ns2 0 3
 		pm_nl_add_endpoint $ns2 10.0.1.2 id 1 dev ns2eth1 flags subflow
 		pm_nl_add_endpoint $ns2 10.0.2.2 id 2 dev ns2eth2 flags subflow
-		{ test_linkfail=128 speed=5 \
+		{ timeout_test=120 test_linkfail=128 speed=5 \
 			run_tests $ns1 $ns2 10.0.1.1 & } 2>/dev/null
 		local tests_pid=$!
 

@@ -3833,7 +3833,7 @@ endpoint_tests()
 		# broadcast IP: no packet for this address will be received on ns1
 		pm_nl_add_endpoint $ns1 224.0.0.1 id 2 flags signal
 		pm_nl_add_endpoint $ns1 10.0.1.1 id 42 flags signal
-		{ test_linkfail=128 speed=5 \
+		{ timeout_test=120 test_linkfail=128 speed=5 \
 			run_tests $ns1 $ns2 10.0.1.1 & } 2>/dev/null
 		local tests_pid=$!
 

@@ -3906,7 +3906,7 @@ endpoint_tests()
 		# broadcast IP: no packet for this address will be received on ns1
 		pm_nl_add_endpoint $ns1 224.0.0.1 id 2 flags signal
 		pm_nl_add_endpoint $ns2 10.0.3.2 id 3 flags subflow
-		{ test_linkfail=128 speed=20 \
+		{ timeout_test=120 test_linkfail=128 speed=20 \
 			run_tests $ns1 $ns2 10.0.1.1 & } 2>/dev/null
 		local tests_pid=$!
 

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Matthieu Baerts (NGI0) <[email protected]>

commit 0e4ec14dc1ee4b1ec347729c225c3ca950f2bcf6 upstream.

In rare cases, when the test environment is very slow, some userspace tests can fail because some expected events have not been seen.

Because the tests are expecting a long on-going connection, and they are not waiting for the end of the transfer, it is fine to have a longer timeout, and even go over the default one. This connection will be killed at the end, after the verifications: increasing the timeout doesn't change anything, apart from avoiding it to end before the end of the verifications.

To play it safe, all userspace tests not waiting for the end of the transfer are now having a longer timeout: 2 minutes.

The Fixes commit was making the connection longer, but still, the default timeout would have stopped it after 1 minute, which might not be enough in very slow environments.

---
 tools/testing/selftests/net/mptcp/mptcp_join.sh |   10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

@@ -3591,7 +3591,7 @@ userspace_tests()
 	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
 		set_userspace_pm $ns1
 		pm_nl_set_limits $ns2 2 2
-		{ test_linkfail=128 speed=5 \
+		{ timeout_test=120 test_linkfail=128 speed=5 \
 			run_tests $ns1 $ns2 10.0.1.1 & } 2>/dev/null
 		local tests_pid=$!
 		wait_mpj $ns1

@@ -3624,7 +3624,7 @@ userspace_tests()
 	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
 		set_userspace_pm $ns2
 		pm_nl_set_limits $ns1 0 1
-		{ test_linkfail=128 speed=5 \
+		{ timeout_test=120 test_linkfail=128 speed=5 \
 			run_tests $ns1 $ns2 10.0.1.1 & } 2>/dev/null
 		local tests_pid=$!
 		wait_mpj $ns2

@@ -3652,7 +3652,7 @@ userspace_tests()
 	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
 		set_userspace_pm $ns2
 		pm_nl_set_limits $ns1 0 1
-		{ test_linkfail=128 speed=5 \
+		{ timeout_test=120 test_linkfail=128 speed=5 \
 			run_tests $ns1 $ns2 10.0.1.1 & } 2>/dev/null
 		local tests_pid=$!
 		wait_mpj $ns2

@@ -3673,7 +3673,7 @@ userspace_tests()
 	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
 		set_userspace_pm $ns2
 		pm_nl_set_limits $ns1 0 1
-		{ test_linkfail=128 speed=5 \
+		{ timeout_test=120 test_linkfail=128 speed=5 \
 			run_tests $ns1 $ns2 10.0.1.1 & } 2>/dev/null
 		local tests_pid=$!
 		wait_mpj $ns2

@@ -3697,7 +3697,7 @@ userspace_tests()
 	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
 		set_userspace_pm $ns1
 		pm_nl_set_limits $ns2 1 1
-		{ test_linkfail=128 speed=5 \
+		{ timeout_test=120 test_linkfail=128 speed=5 \
 			run_tests $ns1 $ns2 10.0.1.1 & } 2>/dev/null
 		local tests_pid=$!
 		wait_mpj $ns1

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Paolo Abeni <[email protected]>

commit 4f102d747cadd8f595f2b25882eed9bec1675fb1 upstream.

The rcv window is shared among all the subflows. Currently, MPTCP sync the TCP-level rcv window with the MPTCP one at tcp_transmit_skb() time.

The above means that incoming data may sporadically observe outdated TCP-level rcv window and being wrongly dropped by TCP.

Address the issue checking for the edge condition before queuing the data at TCP level, and eventually syncing the rcv window as needed.

Note that the issue is actually present from the very first MPTCP implementation, but backports older than the blamed commit below will range from impossible to useless.

Before:

  $ nstat -n; sleep 1; nstat -z TcpExtBeyondWindow
  TcpExtBeyondWindow              14                 0.0

After:

  $ nstat -n; sleep 1; nstat -z TcpExtBeyondWindow
  TcpExtBeyondWindow              0                  0.0

---
 net/mptcp/options.c  |   31 +++++++++++++++++++++++++++++++
 net/mptcp/protocol.h |    1 +
 2 files changed, 32 insertions(+)

@@ -1044,6 +1044,31 @@ static void __mptcp_snd_una_update(struc
 	WRITE_ONCE(msk->snd_una, new_snd_una);
 }
 
+static void rwin_update(struct mptcp_sock *msk, struct sock *ssk,
+			struct sk_buff *skb)
+{
+	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(ssk);
+	struct tcp_sock *tp = tcp_sk(ssk);
+	u64 mptcp_rcv_wnd;
+
+	/* Avoid touching extra cachelines if TCP is going to accept this
+	 * skb without filling the TCP-level window even with a possibly
+	 * outdated mptcp-level rwin.
+	 */
+	if (!skb->len || skb->len < tcp_receive_window(tp))
+		return;
+
+	mptcp_rcv_wnd = atomic64_read(&msk->rcv_wnd_sent);
+	if (!after64(mptcp_rcv_wnd, subflow->rcv_wnd_sent))
+		return;
+
+	/* Some other subflow grew the mptcp-level rwin since rcv_wup,
+	 * resync.
+	 */
+	tp->rcv_wnd += mptcp_rcv_wnd - subflow->rcv_wnd_sent;
+	subflow->rcv_wnd_sent = mptcp_rcv_wnd;
+}
+
 static void ack_update_msk(struct mptcp_sock *msk,
 			   struct sock *ssk,
 			   struct mptcp_options_received *mp_opt)

@@ -1211,6 +1236,7 @@ bool mptcp_incoming_options(struct sock
 	 */
 	if (mp_opt.use_ack)
 		ack_update_msk(msk, sk, &mp_opt);
+	rwin_update(msk, sk, skb);
 
 	/* Zero-data-length packets are dropped by the caller and not
 	 * propagated to the MPTCP layer, so the skb extension does not

@@ -1297,6 +1323,10 @@ static void mptcp_set_rwin(struct tcp_so
 
 	if (rcv_wnd_new != rcv_wnd_old) {
 raise_win:
+		/* The msk-level rcv wnd is after the tcp level one,
+		 * sync the latter.
+		 */
+		rcv_wnd_new = rcv_wnd_old;
 		win = rcv_wnd_old - ack_seq;
 		tp->rcv_wnd = min_t(u64, win, U32_MAX);
 		new_win = tp->rcv_wnd;

@@ -1320,6 +1350,7 @@ raise_win:
 
 update_wspace:
 	WRITE_ONCE(msk->old_wspace, tp->rcv_wnd);
+	subflow->rcv_wnd_sent = rcv_wnd_new;
 }
 
 static void mptcp_track_rwin(struct tcp_sock *tp)

@@ -508,6 +508,7 @@ struct mptcp_subflow_context {
 	u64	remote_key;
 	u64	idsn;
 	u64	map_seq;
+	u64	rcv_wnd_sent;
 	u32	snd_isn;
 	u32	token;
 	u32	rel_write_seq;

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Paolo Abeni <[email protected]>

commit fff0c87996672816a84c3386797a5e69751c5888 upstream.

With the current fastclose implementation, the mptcp_do_fastclose() helper is in charge of two distinct actions: send the fastclose reset and cleanup the subflows.

Formally decouple the two steps, ensuring that mptcp explicitly closes all the subflows after the mentioned helper.

This will make the upcoming fix simpler, and allows dropping the 2nd argument from mptcp_destroy_common(). The Fixes tag is then the same as in the next commit to help with the backports.

---
 net/mptcp/protocol.c |   13 +++++++++----
 net/mptcp/protocol.h |    2 +-
 2 files changed, 10 insertions(+), 5 deletions(-)

@@ -2869,7 +2869,11 @@ static void mptcp_worker(struct work_str
 		__mptcp_close_subflow(sk);
 
 	if (mptcp_close_tout_expired(sk)) {
+		struct mptcp_subflow_context *subflow, *tmp;
+
 		mptcp_do_fastclose(sk);
+		mptcp_for_each_subflow_safe(msk, subflow, tmp)
+			__mptcp_close_ssk(sk, subflow->tcp_sock, subflow, 0);
 		mptcp_close_wake_up(sk);
 	}
 

@@ -3301,7 +3305,8 @@ static int mptcp_disconnect(struct sock
 	/* msk->subflow is still intact, the following will not free the first
 	 * subflow
 	 */
-	mptcp_destroy_common(msk, MPTCP_CF_FASTCLOSE);
+	mptcp_do_fastclose(sk);
+	mptcp_destroy_common(msk);
 
 	/* The first subflow is already in TCP_CLOSE status, the following
 	 * can't overlap with a fallback anymore

@@ -3483,7 +3488,7 @@ void mptcp_rcv_space_init(struct mptcp_s
 		msk->rcvq_space.space = TCP_INIT_CWND * TCP_MSS_DEFAULT;
 }
 
-void mptcp_destroy_common(struct mptcp_sock *msk, unsigned int flags)
+void mptcp_destroy_common(struct mptcp_sock *msk)
 {
 	struct mptcp_subflow_context *subflow, *tmp;
 	struct sock *sk = (struct sock *)msk;

@@ -3492,7 +3497,7 @@ void mptcp_destroy_common(struct mptcp_s
 
 	/* join list will be eventually flushed (with rst) at sock lock release time */
 	mptcp_for_each_subflow_safe(msk, subflow, tmp)
-		__mptcp_close_ssk(sk, mptcp_subflow_tcp_sock(subflow), subflow, flags);
+		__mptcp_close_ssk(sk, mptcp_subflow_tcp_sock(subflow), subflow, 0);
 
 	/* move to sk_receive_queue, sk_stream_kill_queues will purge it */
 	mptcp_data_lock(sk);

@@ -3517,7 +3522,7 @@ static void mptcp_destroy(struct sock *s
 
 	/* allow the following to close even the initial subflow */
 	msk->free_first = 1;
-	mptcp_destroy_common(msk, 0);
+	mptcp_destroy_common(msk);
 	sk_sockets_allocated_dec(sk);
 }
 

@@ -968,7 +968,7 @@ static inline void mptcp_propagate_sndbu
 	local_bh_enable();
 }
 
-void mptcp_destroy_common(struct mptcp_sock *msk, unsigned int flags);
+void mptcp_destroy_common(struct mptcp_sock *msk);
 
 #define MPTCP_TOKEN_MAX_RETRIES	4
 

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Paolo Abeni <[email protected]>

commit 1bba3f219c5e8c29e63afa3c1fc24f875ebec119 upstream.

In case of DSS corruption, the MPTCP protocol tries to avoid the subflow reset if fallback is possible. Such corruptions happen in the receive path; to ensure fallback is possible the stack additionally needs to check for OoO data, otherwise the fallback will break the data stream.

---
 net/mptcp/protocol.c |    7 +++++++
 1 file changed, 7 insertions(+)

@@ -906,6 +906,13 @@ static bool __mptcp_finish_join(struct m
 	if (sk->sk_state != TCP_ESTABLISHED)
 		return false;
 
+	/* The caller possibly is not holding the msk socket lock, but
+	 * in the fallback case only the current subflow is touching
+	 * the OoO queue.
+	 */
+	if (!RB_EMPTY_ROOT(&msk->out_of_order_queue))
+		return false;
+
 	spin_lock_bh(&msk->fallback_lock);
 	if (!msk->allow_subflows) {
 		spin_unlock_bh(&msk->fallback_lock);

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Ma Ke <[email protected]>

commit 4c5376b4b143c4834ebd392aef2215847752b16a upstream.

driver_find_device() calls get_device() to increment the reference count once a matching device is found, but there is no put_device() to balance the reference count. To avoid reference count leakage, add put_device() to decrease the reference count.

Found by code review.

---
 drivers/gpu/drm/tegra/dc.c |    1 +
 1 file changed, 1 insertion(+)

@@ -3147,6 +3147,7 @@ static int tegra_dc_couple(struct tegra_
 		dc->client.parent = &parent->client;
 
 		dev_dbg(dc->dev, "coupled to %s\n", dev_name(companion));
+		put_device(companion);
 	}
 
 	return 0;

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Robert McClinton <[email protected]>

commit 9eb00b5f5697bd56baa3222c7a1426fa15bacfb5 upstream.

Delete the attempt to progress the queue when checking if fence is signaled. This avoids deadlock.

dma-fence_ops::signaled can be called with the fence lock in unknown state. For radeon, the fence lock is also the wait queue lock. This can cause a self deadlock when signaled() tries to make forward progress on the wait queue. But advancing the queue is unneeded because incorrectly returning false from signaled() is perfectly acceptable.

---
 drivers/gpu/drm/radeon/radeon_fence.c |    7 -------
 1 file changed, 7 deletions(-)

@@ -360,13 +360,6 @@ static bool radeon_fence_is_signaled(str
 	if (atomic64_read(&rdev->fence_drv[ring].last_seq) >= seq)
 		return true;
 
-	if (down_read_trylock(&rdev->exclusive_lock)) {
-		radeon_fence_process(rdev, ring);
-		up_read(&rdev->exclusive_lock);
-
-		if (atomic64_read(&rdev->fence_drv[ring].last_seq) >= seq)
-			return true;
-	}
 	return false;
 }
 

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Mario Limonciello <[email protected]>

commit 31ab31433c9bd2f255c48dc6cb9a99845c58b1e4 upstream.

During the suspend sequence VPE is already going to be power gated as part of vpe_suspend(). It's unnecessary to call during calls to amdgpu_device_set_pg_state().

It actually can expose a race condition with the firmware if s0i3 sequence starts as well. Drop these calls.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

@@ -3090,10 +3090,11 @@ int amdgpu_device_set_pg_state(struct am
 		    (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GFX ||
 		     adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_SDMA))
 			continue;
-		/* skip CG for VCE/UVD, it's handled specially */
+		/* skip CG for VCE/UVD/VPE, it's handled specially */
 		if (adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_UVD &&
 		    adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_VCE &&
 		    adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_VCN &&
+		    adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_VPE &&
 		    adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_JPEG &&
 		    adev->ip_blocks[i].version->funcs->set_powergating_state) {
 			/* enable powergating to save power */

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Yifan Zha <[email protected]>

commit 80d8a9ad1587b64c545d515ab6cb7ecb9908e1b3 upstream.

[Why]
Accoreding to CP updated to RS64 on gfx11,
WRITE_DATA with PREEMPTION_META_MEMORY(dst_sel=8) is illegal for CP FW. That packet is used for MCBP on F32 based system.
So it would lead to incorrect GRBM write and FW is not handling that extra case correctly.

[How]
With gfx11 rs64 enabled, skip emit de meta data.

---
 drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

@@ -5642,9 +5642,9 @@ static void gfx_v11_0_ring_emit_ib_gfx(s
 		if (flags & AMDGPU_IB_PREEMPTED)
 			control |= INDIRECT_BUFFER_PRE_RESUME(1);
 
-		if (vmid)
+		if (vmid && !ring->adev->gfx.rs64_enable)
 			gfx_v11_0_ring_emit_de_meta(ring,
-				    (!amdgpu_sriov_vf(ring->adev) && flags & AMDGPU_IB_PREEMPTED) ? true : false);
+				!amdgpu_sriov_vf(ring->adev) && (flags & AMDGPU_IB_PREEMPTED));
 	}
 
 	if (ring->is_mes_queue)

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Mario Limonciello (AMD) <[email protected]>

commit 8612badc331bcab2068baefa69e1458085ed89e3 upstream.

[Why]
Empirical measurement of some monitors that fail to read EDID while booting shows that the number of retries with a 30ms delay between tries is as high as 16.

[How]
Increase number of retries to 20.

---
 drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -1587,7 +1587,7 @@ static bool retrieve_link_cap(struct dc_
 	union edp_configuration_cap edp_config_cap;
 	union dp_downstream_port_present ds_port = { 0 };
 	enum dc_status status = DC_ERROR_UNEXPECTED;
-	uint32_t read_dpcd_retry_cnt = 3;
+	uint32_t read_dpcd_retry_cnt = 20;
 	int i;
 	struct dp_sink_hw_fw_revision dp_hw_fw_revision;
 	const uint32_t post_oui_delay = 30; // 30ms

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Mario Limonciello (AMD) <[email protected]>

commit 71ad9054c1f241be63f9d11df8cbd0aa0352fe16 upstream.

[Why]
When a monitor is booting it's possible that it isn't ready to retrieve link caps and this can lead to an EDID read failure:

```
[drm:retrieve_link_cap [amdgpu]] *ERROR* retrieve_link_cap: Read receiver caps dpcd data failed.
amdgpu 0000:c5:00.0: [drm] *ERROR* No EDID read.
```

[How]
Rather than msleep once and try a few times, msleep each time. Should be no changes for existing working monitors, but should correct reading caps on a monitor that is slow to boot.

---
 drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c |    9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

@@ -1633,12 +1633,13 @@ static bool retrieve_link_cap(struct dc_
 	status = dpcd_get_tunneling_device_data(link);
 
 	dpcd_set_source_specific_data(link);
-	/* Sink may need to configure internals based on vendor, so allow some
-	 * time before proceeding with possibly vendor specific transactions
-	 */
-	msleep(post_oui_delay);
 
 	for (i = 0; i < read_dpcd_retry_cnt; i++) {
+		/*
+		 * Sink may need to configure internals based on vendor, so allow some
+		 * time before proceeding with possibly vendor specific transactions
+		 */
+		msleep(post_oui_delay);
 		status = core_link_read_dpcd(
 				link,
 				DP_DPCD_REV,

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Fangzhi Zuo <[email protected]>

commit 1788ef30725da53face7e311cdf62ad65fababcd upstream.

[Why]
Existing routine has two conversion sequence, pbn_to_kbps and kbps_to_pbn with margin.
Non of those has without-margin calculation.

kbps_to_pbn with margin conversion includes fec overhead which has already been included in pbn_div calculation with 0.994 factor considered. It is a double counted fec overhead factor that causes potential bw loss.

[How]
Add without-margin calculation.
Fix fec overhead double counted issue.

---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c |   59 ++++--------
 1 file changed, 23 insertions(+), 36 deletions(-)

@@ -846,26 +846,28 @@ struct dsc_mst_fairness_params {
 };
 
 #if defined(CONFIG_DRM_AMD_DC_FP)
-static uint16_t get_fec_overhead_multiplier(struct dc_link *dc_link)
+static uint64_t kbps_to_pbn(int kbps, bool is_peak_pbn)
 {
-	u8 link_coding_cap;
-	uint16_t fec_overhead_multiplier_x1000 = PBN_FEC_OVERHEAD_MULTIPLIER_8B_10B;
+	uint64_t effective_kbps = (uint64_t)kbps;
 
-	link_coding_cap = dc_link_dp_mst_decide_link_encoding_format(dc_link);
-	if (link_coding_cap == DP_128b_132b_ENCODING)
-		fec_overhead_multiplier_x1000 = PBN_FEC_OVERHEAD_MULTIPLIER_128B_132B;
+	if (is_peak_pbn) {	// add 0.6% (1006/1000) overhead into effective kbps
+		effective_kbps *= 1006;
+		effective_kbps = div_u64(effective_kbps, 1000);
+	}
 
-	return fec_overhead_multiplier_x1000;
+	return (uint64_t) DIV64_U64_ROUND_UP(effective_kbps * 64, (54 * 8 * 1000));
 }
 
-static int kbps_to_peak_pbn(int kbps, uint16_t fec_overhead_multiplier_x1000)
+static uint32_t pbn_to_kbps(unsigned int pbn, bool with_margin)
 {
-	u64 peak_kbps = kbps;
+	uint64_t pbn_effective = (uint64_t)pbn;
+
+	if (with_margin)	// deduct 0.6% (994/1000) overhead from effective pbn
+		pbn_effective *= (1000000 / PEAK_FACTOR_X1000);
+	else
+		pbn_effective *= 1000;
 
-	peak_kbps *= 1006;
-	peak_kbps *= fec_overhead_multiplier_x1000;
-	peak_kbps = div_u64(peak_kbps, 1000 * 1000);
-	return (int) DIV64_U64_ROUND_UP(peak_kbps * 64, (54 * 8 * 1000));
+	return DIV_U64_ROUND_UP(pbn_effective * 8 * 54, 64);
 }
 
 static void set_dsc_configs_from_fairness_vars(struct dsc_mst_fairness_params *params,

@@ -936,7 +938,7 @@ static int bpp_x16_from_pbn(struct dsc_m
 	dc_dsc_get_default_config_option(param.sink->ctx->dc, &dsc_options);
 	dsc_options.max_target_bpp_limit_override_x16 = drm_connector->display_info.max_dsc_bpp * 16;
 
-	kbps = div_u64((u64)pbn * 994 * 8 * 54, 64);
+	kbps = pbn_to_kbps(pbn, false);
 	dc_dsc_compute_config(
 			param.sink->ctx->dc->res_pool->dscs[0],
 			&param.sink->dsc_caps.dsc_dec_caps,

@@ -965,12 +967,11 @@ static int increase_dsc_bpp(struct drm_a
 	int link_timeslots_used;
 	int fair_pbn_alloc;
 	int ret = 0;
-	uint16_t fec_overhead_multiplier_x1000 = get_fec_overhead_multiplier(dc_link);
 
 	for (i = 0; i < count; i++) {
 		if (vars[i + k].dsc_enabled) {
 			initial_slack[i] =
-			kbps_to_peak_pbn(params[i].bw_range.max_kbps, fec_overhead_multiplier_x1000) - vars[i + k].pbn;
+			kbps_to_pbn(params[i].bw_range.max_kbps, false) - vars[i + k].pbn;
 			bpp_increased[i] = false;
 			remaining_to_increase += 1;
 		} else {

@@ -1066,7 +1067,6 @@ static int try_disable_dsc(struct drm_at
 	int next_index;
 	int remaining_to_try = 0;
 	int ret;
-	uint16_t fec_overhead_multiplier_x1000 = get_fec_overhead_multiplier(dc_link);
 	int var_pbn;
 
 	for (i = 0; i < count; i++) {

@@ -1099,7 +1099,7 @@ static int try_disable_dsc(struct drm_at
 
 		DRM_DEBUG_DRIVER("MST_DSC index #%d, try no compression\n", next_index);
 		var_pbn = vars[next_index].pbn;
-		vars[next_index].pbn = kbps_to_peak_pbn(params[next_index].bw_range.stream_kbps, fec_overhead_multiplier_x1000);
+		vars[next_index].pbn = kbps_to_pbn(params[next_index].bw_range.stream_kbps, true);
 		ret = drm_dp_atomic_find_time_slots(state,
 						    params[next_index].port->mgr,
 						    params[next_index].port,

@@ -1159,7 +1159,6 @@ static int compute_mst_dsc_configs_for_l
 	int count = 0;
 	int i, k, ret;
 	bool debugfs_overwrite = false;
-	uint16_t fec_overhead_multiplier_x1000 = get_fec_overhead_multiplier(dc_link);
 	struct drm_connector_state *new_conn_state;
 
 	memset(params, 0, sizeof(params));

@@ -1240,7 +1239,7 @@ static int compute_mst_dsc_configs_for_l
 	DRM_DEBUG_DRIVER("MST_DSC Try no compression\n");
 	for (i = 0; i < count; i++) {
 		vars[i + k].aconnector = params[i].aconnector;
-		vars[i + k].pbn = kbps_to_peak_pbn(params[i].bw_range.stream_kbps, fec_overhead_multiplier_x1000);
+		vars[i + k].pbn = kbps_to_pbn(params[i].bw_range.stream_kbps, false);
 		vars[i + k].dsc_enabled = false;
 		vars[i + k].bpp_x16 = 0;
 		ret = drm_dp_atomic_find_time_slots(state, params[i].port->mgr, params[i].port,

@@ -1262,7 +1261,7 @@ static int compute_mst_dsc_configs_for_l
 	DRM_DEBUG_DRIVER("MST_DSC Try max compression\n");
 	for (i = 0; i < count; i++) {
 		if (params[i].compression_possible && params[i].clock_force_enable != DSC_CLK_FORCE_DISABLE) {
-			vars[i + k].pbn = kbps_to_peak_pbn(params[i].bw_range.min_kbps, fec_overhead_multiplier_x1000);
+			vars[i + k].pbn = kbps_to_pbn(params[i].bw_range.min_kbps, false);
 			vars[i + k].dsc_enabled = true;
 			vars[i + k].bpp_x16 = params[i].bw_range.min_target_bpp_x16;
 			ret = drm_dp_atomic_find_time_slots(state, params[i].port->mgr,

@@ -1270,7 +1269,7 @@ static int compute_mst_dsc_configs_for_l
 			if (ret < 0)
 				return ret;
 		} else {
-			vars[i + k].pbn = kbps_to_peak_pbn(params[i].bw_range.stream_kbps, fec_overhead_multiplier_x1000);
+			vars[i + k].pbn = kbps_to_pbn(params[i].bw_range.stream_kbps, false);
 			vars[i + k].dsc_enabled = false;
 			vars[i + k].bpp_x16 = 0;
 			ret = drm_dp_atomic_find_time_slots(state, params[i].port->mgr,

@@ -1722,18 +1721,6 @@ clean_exit:
 	return ret;
 }
 
-static uint32_t kbps_from_pbn(unsigned int pbn)
-{
-	uint64_t kbps = (uint64_t)pbn;
-
-	kbps *= (1000000 / PEAK_FACTOR_X1000);
-	kbps *= 8;
-	kbps *= 54;
-	kbps /= 64;
-
-	return (uint32_t)kbps;
-}
-
 static bool is_dsc_common_config_possible(struct dc_stream_state *stream,
 					  struct dc_dsc_bw_range *bw_range)
 {

@@ -1825,7 +1812,7 @@ enum dc_status dm_dp_mst_is_port_support
 			dc_link_get_highest_encoding_format(stream->link));
 	cur_link_settings = stream->link->verified_link_cap;
 	root_link_bw_in_kbps = dc_link_bandwidth_kbps(aconnector->dc_link, &cur_link_settings);
-	virtual_channel_bw_in_kbps = kbps_from_pbn(aconnector->mst_output_port->full_pbn);
+	virtual_channel_bw_in_kbps = pbn_to_kbps(aconnector->mst_output_port->full_pbn, true);
 
 	/* pick the end to end bw bottleneck */
 	end_to_end_bw_in_kbps = min(root_link_bw_in_kbps, virtual_channel_bw_in_kbps);

@@ -1876,7 +1863,7 @@ enum dc_status dm_dp_mst_is_port_support
 				immediate_upstream_port = aconnector->mst_output_port->parent->port_parent;
 
 			if (immediate_upstream_port) {
-				virtual_channel_bw_in_kbps = kbps_from_pbn(immediate_upstream_port->full_pbn);
+				virtual_channel_bw_in_kbps = pbn_to_kbps(immediate_upstream_port->full_pbn, true);
 				virtual_channel_bw_in_kbps = min(root_link_bw_in_kbps, virtual_channel_bw_in_kbps);
 			} else {
 				/* For topology LCT 1 case - only one mstb*/

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Ivan Lipski <[email protected]>

commit 5bab4c89390f32b2f491f49a151948cd226dd909 upstream.

[Why]
On DCN20 & DCN30, the 6th DPP's & HUBP's are powered on permanently and cannot be power gated. Thus, when dpp_reset() is invoked for the DPP5, while it's still powered on, the cached cursor_state (dpp_base->pos.cur0_ctl.bits.cur0_enable)
and the actual state (CUR0_ENABLE) bit are unsycned. This can cause a double cursor in full screen with non-native scaling.

[How]
Force disable cursor on DPP5 on plane powerdown for ASICs w/ 6 DPPs/HUBPs.

---
 drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c |    8 ++++++++
 1 file changed, 8 insertions(+)

@@ -604,6 +604,14 @@ void dcn20_dpp_pg_control(
 		 *		DOMAIN11_PGFSM_PWR_STATUS, pwr_status,
 		 * 		1, 1000);
 		 */
+
+		/* Force disable cursor on plane powerdown on DPP 5 using dpp_force_disable_cursor */
+		if (!power_on) {
+			struct dpp *dpp5 = hws->ctx->dc->res_pool->dpps[dpp_inst];
+			if (dpp5 && dpp5->funcs->dpp_force_disable_cursor)
+				dpp5->funcs->dpp_force_disable_cursor(dpp5);
+		}
+
 		break;
 	default:
 		BREAK_TO_DEBUGGER();

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Sabrina Dubroca <[email protected]>

[ Upstream commit 8d2a2a49c30f67a480fa9ed25e08436a446f057e ]

We're not updating x1, but we still need to put() it.

---
 net/xfrm/xfrm_state.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

index 1e2f5ecd63248..f8cb033f102ed 100644
@@ -2117,14 +2117,18 @@ int xfrm_state_update(struct xfrm_state *x)
 	}
 
 	if (x1->km.state == XFRM_STATE_ACQ) {
-		if (x->dir && x1->dir != x->dir)
+		if (x->dir && x1->dir != x->dir) {
+			to_put = x1;
 			goto out;
+		}
 
 		__xfrm_state_insert(x);
 		x = NULL;
 	} else {
-		if (x1->dir != x->dir)
+		if (x1->dir != x->dir) {
+			to_put = x1;
 			goto out;
+		}
 	}
 	err = 0;
 

2.51.0

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Sabrina Dubroca <[email protected]>

[ Upstream commit 1dcf617bec5cb85f68ca19969e7537ef6f6931d3 ]

xfrm_state_construct can fail without setting an error if the requested pcpu_num value is too big. Set err and add an extack message to avoid confusing userspace.

---
 net/xfrm/xfrm_user.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

index d41e5642625e3..3d0fdeebaf3c8 100644
@@ -893,8 +893,11 @@ static struct xfrm_state *xfrm_state_construct(struct net *net,
 
 	if (attrs[XFRMA_SA_PCPU]) {
 		x->pcpu_num = nla_get_u32(attrs[XFRMA_SA_PCPU]);
-		if (x->pcpu_num >= num_possible_cpus())
+		if (x->pcpu_num >= num_possible_cpus()) {
+			err = -ERANGE;
+			NL_SET_ERR_MSG(extack, "pCPU number too big");
 			goto error;
+		}
 	}
 
 	err = __xfrm_init_state(x, false, attrs[XFRMA_OFFLOAD_DEV], extack);

2.51.0

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Yu-Chun Lin <[email protected]>

[ Upstream commit 369f772299821f93f872bf1b4d7d7ed2fc50243b ]

The pinctrl-rtd driver uses 'devm_regmap_init_mmio', which requires 'REGMAP_MMIO' to be enabled.

Without this selection, the build fails with an undefined reference:
aarch64-none-linux-gnu-ld: drivers/pinctrl/realtek/pinctrl-rtd.o: in function rtd_pinctrl_probe': pinctrl-rtd.c:(.text+0x5a0): undefined reference to __devm_regmap_init_mmio_clk'

Fix this by selecting 'REGMAP_MMIO' in the Kconfig.

---
 drivers/pinctrl/realtek/Kconfig | 1 +
 1 file changed, 1 insertion(+)

index 0fc6bd4fcb7ec..400c9e5b16ada 100644
@@ -6,6 +6,7 @@ config PINCTRL_RTD
 	default y
 	select PINMUX
 	select GENERIC_PINCONF
+	select REGMAP_MMIO
 
 config PINCTRL_RTD1619B
 	tristate "Realtek DHC 1619B pin controller driver"

2.51.0

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Jianbo Liu <[email protected]>

[ Upstream commit 61fafbee6cfed283c02a320896089f658fa67e56 ]

The GSO segmentation functions for ESP tunnel mode (xfrm4_tunnel_gso_segment and xfrm6_tunnel_gso_segment) were determining the inner packet's L2 protocol type by checking the static x->inner_mode.family field from the xfrm state.

This is unreliable. In tunnel mode, the state's actual inner family could be defined by x->inner_mode.family or by x->inner_mode_iaf.family. Checking only the former can lead to a mismatch with the actual packet being processed, causing GSO to create segments with the wrong L2 header type.

This patch fixes the bug by deriving the inner mode directly from the packet's inner protocol stored in XFRM_MODE_SKB_CB(skb)->protocol.

Instead of replicating the code, this patch modifies the xfrm_ip2inner_mode helper function. It now correctly returns &x->inner_mode if the selector family (x->sel.family) is already specified, thereby handling both specific and AF_UNSPEC cases appropriately.

With this change, ESP GSO can use xfrm_ip2inner_mode to get the correct inner mode. It doesn't affect existing callers, as the updated logic now mirrors the checks they were already performing externally.

---
 include/net/xfrm.h      | 3 ++-
 net/ipv4/esp4_offload.c | 6 ++++--
 net/ipv6/esp6_offload.c | 6 ++++--
 3 files changed, 10 insertions(+), 5 deletions(-)

index 1484dd15a3694..caaff61601a07 100644
@@ -472,7 +472,8 @@ static inline int xfrm_af2proto(unsigned int family)
 
 static inline const struct xfrm_mode *xfrm_ip2inner_mode(struct xfrm_state *x, int ipproto)
 {
-	if ((ipproto == IPPROTO_IPIP && x->props.family == AF_INET) ||
+	if ((x->sel.family != AF_UNSPEC) ||
+	    (ipproto == IPPROTO_IPIP && x->props.family == AF_INET) ||
 	    (ipproto == IPPROTO_IPV6 && x->props.family == AF_INET6))
 		return &x->inner_mode;
 	else

index e0d94270da28a..05828d4cb6cdb 100644
@@ -122,8 +122,10 @@ static struct sk_buff *xfrm4_tunnel_gso_segment(struct xfrm_state *x,
 						struct sk_buff *skb,
 						netdev_features_t features)
 {
-	__be16 type = x->inner_mode.family == AF_INET6 ? htons(ETH_P_IPV6)
-						       : htons(ETH_P_IP);
+	const struct xfrm_mode *inner_mode = xfrm_ip2inner_mode(x,
+					XFRM_MODE_SKB_CB(skb)->protocol);
+	__be16 type = inner_mode->family == AF_INET6 ? htons(ETH_P_IPV6)
+						     : htons(ETH_P_IP);
 
 	return skb_eth_gso_segment(skb, features, type);
 }

index 7b41fb4f00b58..22410243ebe88 100644
@@ -158,8 +158,10 @@ static struct sk_buff *xfrm6_tunnel_gso_segment(struct xfrm_state *x,
 						struct sk_buff *skb,
 						netdev_features_t features)
 {
-	__be16 type = x->inner_mode.family == AF_INET ? htons(ETH_P_IP)
-						      : htons(ETH_P_IPV6);
+	const struct xfrm_mode *inner_mode = xfrm_ip2inner_mode(x,
+					XFRM_MODE_SKB_CB(skb)->protocol);
+	__be16 type = inner_mode->family == AF_INET ? htons(ETH_P_IP)
+						    : htons(ETH_P_IPV6);
 
 	return skb_eth_gso_segment(skb, features, type);
 }

2.51.0

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Jianbo Liu <[email protected]>

[ Upstream commit 59630e2ccd728703cc826e3a3515d70f8c7a766c ]

Add a check to ensure locally generated packets (skb->sk != NULL) do not use direct output in tunnel mode, as these packets require proper L2 header setup that is handled by the normal XFRM processing path.

---
 net/xfrm/xfrm_output.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

index a30538a980cc7..9277dd4ed541a 100644
@@ -766,8 +766,12 @@ int xfrm_output(struct sock *sk, struct sk_buff *skb)
 		/* Exclusive direct xmit for tunnel mode, as
 		 * some filtering or matching rules may apply
 		 * in transport mode.
+		 * Locally generated packets also require
+		 * the normal XFRM path for L2 header setup,
+		 * as the hardware needs the L2 header to match
+		 * for encryption, so skip direct output as well.
 		 */
-		if (x->props.mode == XFRM_MODE_TUNNEL)
+		if (x->props.mode == XFRM_MODE_TUNNEL && !skb->sk)
 			return xfrm_dev_direct_output(sk, x, skb);
 
 		return xfrm_output_resume(sk, skb, 0);

2.51.0

6.12-stable review patch. If anyone has any objections, please let me know.

------------------

From: Haotian Zhang <[email protected]>

[ Upstream commit 9b07cdf86a0b90556f5b68a6b20b35833b558df3 ]

The driver calls fwnode_get_named_child_node() which takes a reference on the child node, but never releases it, which causes a reference leak.

Fix by using devm_add_action_or_reset() to automatically release the reference when the device is removed.