Skip to contentSkip to conversation

6.18.32-rc1 review

PATCH In review Greg Kroah-Hartman 12 participants 206 messages May 15 – May 19 Page 1 of 2

Subsystems (3)
lore
Share & tools
lore

Greg Kroah-Hartman

author maintainer
#0 lore raw

cover 6.18.32-rc1 review

This is the start of the stable review cycle for the 6.18.32 release. There are 188 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 Sun, 17 May 2026 15:46:37 +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-6.18.32-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-6.18.y
and the diffstat can be found below.

thanks,

greg k-h

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

Greg Kroah-Hartman <[email protected]>

    Linux 6.18.32-rc1

Benjamin Cheng <[email protected]>
drm/amdgpu/vcn4: Avoid overflow on msg bound check

Benjamin Cheng <[email protected]>
drm/amdgpu/vcn3: Avoid overflow on msg bound check

Dudu Lu <[email protected]>
vsock/virtio: fix accept queue count leak on transport mismatch

Eric Dumazet <[email protected]>
vsock/virtio: fix potential unbounded skb queue

Stefano Garzarella <[email protected]>
vsock/virtio: fix empty payload in tap skb for non-linear buffers

Stefano Garzarella <[email protected]>
vsock/virtio: fix length and offset in tap skb for split packets

Norbert Szetei <[email protected]>
vsock: fix buffer size clamping order

Sven Eckelmann <[email protected]>
batman-adv: tp_meter: fix tp_num leak on kmalloc failure

Masami Hiramatsu (Google) <[email protected]>
tracing/fprobe: Remove fprobe from hash in failure path

Masami Hiramatsu (Google) <[email protected]>
tracing/fprobe: Unregister fprobe even if memory allocation fails

Menglong Dong <[email protected]>
tracing: fprobe: optimization for entry only case

Menglong Dong <[email protected]>
tracing: fprobe: use rhltable for fprobe_ip_table

Yochai Eisenrich <[email protected]>
btrfs: fix btrfs_ioctl_space_info() slot_count TOCTOU which can lead to info-leak

Guangshuo Li <[email protected]>
btrfs: fix double free in create_space_info_sub_group() error path

Filipe Manana <[email protected]>
btrfs: remove fs_info argument from btrfs_sysfs_add_space_info_type()

Selvarasu Ganesan <[email protected]>
usb: dwc3: Move GUID programming after PHY initialization

Prashanth K <[email protected]>
usb: dwc3: Add dwc pointer to dwc3_readl/writel

Prashanth K <[email protected]>

    usb: dwc3: Remove of dep->regs

Tejun Heo <[email protected]>
sched_ext: Read scx_root under scx_cgroup_ops_rwsem in cgroup setters

zhidao su <[email protected]>
sched/ext: Implement cgroup_set_idle() callback

David Carlier <[email protected]>
Bluetooth: hci_conn: fix potential UAF in create_big_sync

Johan Hovold <[email protected]>
spi: zynq-qspi: fix controller deregistration

Pei Xiao <[email protected]>
spi: zynq-qspi: Simplify clock handling with devm_clk_get_enabled()

Johan Hovold <[email protected]>
spi: tegra114: fix controller deregistration

Johan Hovold <[email protected]>
spi: tegra20-sflash: fix controller deregistration

Johan Hovold <[email protected]>
spi: uniphier: fix controller deregistration

Pei Xiao <[email protected]>
spi: uniphier: Simplify clock handling with devm_clk_get_enabled()

Ritesh Harjani (IBM) <[email protected]>
pseries/papr-hvpipe: Fix race with interrupt handler

Christian Brauner <[email protected]>
papr-hvpipe: convert papr_hvpipe_dev_create_handle() to FD_PREPARE()

Prasanna Kumar T S M <[email protected]>
EDAC/versalnet: Fix device name memory leak

Damien Le Moal <[email protected]>
block: fix zone write plug removal

Thomas Zimmermann <[email protected]>
fbcon: Avoid OOB font access if console rotation fails

Thomas Zimmermann <[email protected]>
fbcon: Rename struct fbcon_ops to struct fbcon_par

Alex Deucher <[email protected]>
drm/amdgpu: rework how we handle TLB fences

Prike Liang <[email protected]>
Revert "drm/amdgpu: don't attach the tlb fence for SI"

Timur Kristóf <[email protected]>
drm/amdgpu: Fix validating flush_gpu_tlb_pasid()

Prike Liang <[email protected]>
drm/amdgpu: validate the flush_gpu_tlb_pasid()

SeongJae Park <[email protected]>
mm/damon/core: disallow time-quota setting zero esz

Amit Sunil Dhamne <[email protected]>
usb: typec: tcpm: reset internal port states on soft reset AMS

SeongJae Park <[email protected]>
mm/damon/reclaim: detect and use fresh enabled and kdamond_pid values

SeongJae Park <[email protected]>
mm/damon/lru_sort: detect and use fresh enabled and kdamond_pid values

SeongJae Park <[email protected]>
mm/damon/core: implement damon_kdamond_pid()

Xianglai Li <[email protected]>
LoongArch: KVM: Compile switch.S directly into the kernel

Pavel Begunkov <[email protected]>
io_uring/zcrx: warn on freelist violations

Pavel Begunkov <[email protected]>
io_uring/zcrx: use guards for locking

Andrea Righi <[email protected]>
sched_ext: Use HK_TYPE_DOMAIN_BOOT to detect isolcpus= domain isolation

Sven Eckelmann <[email protected]>
batman-adv: bla: put backbone reference on failed claim hash insert

Sven Eckelmann <[email protected]>
batman-adv: bla: only purge non-released claims

Sven Eckelmann <[email protected]>
batman-adv: bla: prevent use-after-free when deleting claims

Jiexun Wang <[email protected]>
batman-adv: stop caching unowned originator pointers in BAT IV

Jiexun Wang <[email protected]>
batman-adv: stop tp_meter sessions during mesh teardown

Jiexun Wang <[email protected]>
batman-adv: reject new tp_meter sessions during teardown

Lyes Bourennani <[email protected]>
batman-adv: fix integer overflow on buff_pos

Ben Morris <[email protected]>
sctp: revalidate list cursor after sctp_sendmsg_to_asoc() in SCTP_SENDALL

Siddharth Vadapalli <[email protected]>
arm64: dts: ti: k3-am62a7-sk: Fix pin name in comment from M19 to N22

Viken Dadhaniya <[email protected]>
arm64: dts: qcom: lemans: Correct QUP interrupt numbers

Alex Deucher <[email protected]>
drm/amdgpu/pm: align Hawaii mclk workaround with radeon

Alex Deucher <[email protected]>
drm/amdgpu/pm: add missing revision check for CI

John B. Moore <[email protected]>
drm/amdgpu/sdma4: replace BUG_ON with WARN_ON in fence emission

Felix Kuehling <[email protected]>
drm/amdkfd: Make all TLB-flushes heavy-weight

Icenowy Zheng <[email protected]>
drm/panel: boe-tv101wum-nl6: restore MODE_LPM after sending disable cmds

Kory Maincent (TI) <[email protected]>
drm/bridge: tda998x: Use __be32 for audio port OF property pointer

John B. Moore <[email protected]>
drm/amdgpu/gfx9: drop unnecessary 64-bit fence flag check in KIQ

Icenowy Zheng <[email protected]>
drm/panel: himax-hx83102: restore MODE_LPM after sending disable cmds

Osama Abdelkader <[email protected]>
drm/exynos: remove bridge when component_add fails

Philip Yang <[email protected]>
drm/amdgpu: zero-initialize GART table on allocation

Alex Deucher <[email protected]>
drm/radeon: add missing revision check for CI

Francis, David <[email protected]>
drm: Set old handle to NULL before prime swap in change_handle

Jia Yao <[email protected]>
drm/xe/uapi: Reject coh_none PAT index for CPU cached memory in madvise

Shuicheng Lin <[email protected]>
drm/xe/bo: Fix bo leak on unaligned size validation in xe_bo_init_locked()

Shuicheng Lin <[email protected]>
drm/xe: Fix dma-buf attachment leak in xe_gem_prime_import()

Shuicheng Lin <[email protected]>
drm/xe/bo: Fix bo leak on GGTT flag validation in xe_bo_init_locked()

Shuicheng Lin <[email protected]>
drm/xe: Fix bo leak in xe_dma_buf_init_obj() on allocation failure

Shixiong Ou <[email protected]>
drm/udl: Increase GET_URB_TIMEOUT

Alysa Liu <[email protected]>
drm/amdkfd: validate SVM ioctl nattr against buffer size

Sasha Finkelstein <[email protected]>
drm/appletbdrm: Use kvzalloc for big allocations

Ashutosh Desai <[email protected]>
drm/gem: Fix inconsistent plane dimension calculation in drm_gem_fb_init_with_funcs()

Mario Kleiner <[email protected]>
drm/amd/display: Change dither policy for 10 bpc output back to dithering

Benjamin Cheng <[email protected]>
drm/amdgpu/vcn3: Prevent OOB reads when parsing dec msg

Benjamin Cheng <[email protected]>
drm/amdgpu/vcn4: Prevent OOB reads when parsing dec msg

Benjamin Cheng <[email protected]>
drm/amdgpu/vce: Prevent partial address patches

Benjamin Cheng <[email protected]>
drm/amdgpu/vcn4: Prevent OOB reads when parsing IB

Benjamin Cheng <[email protected]>
drm/amdgpu: Add bounds checking to ib_{get,set}_value

Alysa Liu <[email protected]>
drm/amdkfd: Add upper bound check for num_of_nodes

Yang Wang <[email protected]>
drm/amd/pm: fix incorrect FeatureCtrlMask setting on smu v14.0.x

Chenglei Xie <[email protected]>
drm/amdgpu: gate VM CPU HDP flush on reset lock

Ramalingeswara Reddy, Kanala <[email protected]>
drm/amdgpu: Use SMUIO 15.0.0 offsets for TSC upper and lower count.

Amir Shetaia <[email protected]>
drm/amdkfd: Clear VRAM on allocation to prevent stale data exposure

Jouni Högander <[email protected]>
drm/i915/psr: Init variable to avoid early exit from et alignment loop

Anna Maniscalco <[email protected]>
drm/msm: always recover the gpu

Marek Vasut <[email protected]>
drm/imx: parallel-display: Prefer bus format set via legacy "interface-pix-fmt" DT property

Yasuaki Torimaru <[email protected]>
drm/msm/gem: fix error handling in msm_ioctl_gem_info_get_metadata()

Johan Hovold <[email protected]>
spi: cadence: fix clock imbalance on probe failure

Johan Hovold <[email protected]>
spi: cadence: fix unclocked access on unbind

Johan Hovold <[email protected]>
spi: cadence: fix controller deregistration

Johan Hovold <[email protected]>
spi: mpc52xx: fix use-after-free on unbind

Johan Hovold <[email protected]>
spi: mpc52xx: fix controller deregistration

Johan Hovold <[email protected]>
spi: mpc52xx: fix use-after-free on registration failure

Johan Hovold <[email protected]>
spi: orion: fix clock imbalance on registration failure

Johan Hovold <[email protected]>
spi: orion: fix runtime pm leak on unbind

Johan Hovold <[email protected]>
spi: orion: fix controller deregistration

Johan Hovold <[email protected]>
spi: mxic: fix controller deregistration

Johan Hovold <[email protected]>
spi: imx: fix runtime pm leak on probe deferral

Johan Hovold <[email protected]>
spi: mpfs: fix controller deregistration

Johan Hovold <[email protected]>
spi: img-spfi: fix controller deregistration

Johan Hovold <[email protected]>
spi: slave-mt27xx: fix controller deregistration

Johan Hovold <[email protected]>
spi: sh-msiof: fix controller deregistration

Johan Hovold <[email protected]>
spi: rspi: fix controller deregistration

Johan Hovold <[email protected]>
spi: sprd: fix controller deregistration

Johan Hovold <[email protected]>
spi: pic32-sqi: fix controller deregistration

Johan Hovold <[email protected]>
spi: cavium-thunderx: fix controller deregistration

Johan Hovold <[email protected]>
spi: npcm-pspi: fix controller deregistration

Johan Hovold <[email protected]>
spi: coldfire-qspi: fix controller deregistration

Johan Hovold <[email protected]>
spi: bcmbca-hsspi: fix controller deregistration

Johan Hovold <[email protected]>
spi: fsl: fix controller deregistration

Johan Hovold <[email protected]>
spi: sh-hspi: fix controller deregistration

Johan Hovold <[email protected]>
spi: ch341: fix devres lifetime

Johan Hovold <[email protected]>
spi: pl022: fix controller deregistration

Johan Hovold <[email protected]>
spi: mtk-nor: fix controller deregistration

Johan Hovold <[email protected]>
spi: pic32: fix controller deregistration

Johan Hovold <[email protected]>
spi: omap2-mcspi: fix controller deregistration

Johan Hovold <[email protected]>
spi: fsl-espi: fix controller deregistration

Johan Hovold <[email protected]>
spi: s3c64xx: fix controller deregistration

Johan Hovold <[email protected]>
spi: dln2: fix controller deregistration

Johan Hovold <[email protected]>
spi: mt65xx: fix controller deregistration

Johan Hovold <[email protected]>
spi: mxs: fix controller deregistration

Wenmeng Liu <[email protected]>
media: qcom: camss: Add missing clocks for VFE lite on sa8775p

Thomas Fourier <[email protected]>
media: iris: Fix dma_free_attrs() size in iris_hfi_queues_init()

Arnd Bergmann <[email protected]>
media: venus: fix QCOM_MDT_LOADER dependency

Dmitry Baryshkov <[email protected]>
media: qcom: iris: increase H265D_MAX_SLICE to fix H.265 decoding on SC7280

Wenmeng Liu <[email protected]>
media: qcom: camss: Fix csid IRQ offset for sa8775p

Wenmeng Liu <[email protected]>
media: qcom: camss: Fix csid clock configuration for sa8775p

Dikshita Agarwal <[email protected]>
media: iris: Fix use-after-free in iris_release_internal_buffers()

Arnd Bergmann <[email protected]>
media: iris: fix QCOM_MDT_LOADER dependency

Haoxiang Li <[email protected]>
media: omap3isp: drop the use count of v4l2 pipeline

Matthias Fend <[email protected]>
media: i2c: ov08d10: fix runtime PM handling in probe

Matthias Fend <[email protected]>
media: i2c: ov08d10: fix image vertical start setting

Michael Tretter <[email protected]>
media: staging: imx: request mbus_config in csi_start

Wenmeng Liu <[email protected]>
media: i2c: imx412: Assert reset GPIO during probe

Sergey Shtylyov <[email protected]>

    media: dib8000: avoid division by 0 in dib8000_set_dds()

Abdun Nihaal <[email protected]>
media: pci: zoran: fix potential memory leak in zoran_probe()

Gregor Herburger <[email protected]>

    arm64: dts: broadcom: bcm2712-d-rpi-5-b: update uart10 interrupt

Gregor Herburger <[email protected]>
arm64: dts: broadcom: bcm2712-d-rpi-5-b: add fixes for pinctrl/pinctrl_aon

Luigi Leonardi <[email protected]>
vsock/virtio: fix MSG_PEEK ignoring skb offset when calculating bytes to copy

Krishna Chomal <[email protected]>
platform/x86: hp-wmi: Ignore backlight and FnLock events

Johan Hovold <[email protected]>
spi: aspeed-smc: fix controller deregistration

Johan Hovold <[email protected]>
spi: amlogic-spisg: fix controller deregistration

Wang Jun <[email protected]>
media: saa7164: add ioremap return checks and cleanups

Johan Hovold <[email protected]>
spi: at91-usart: fix controller deregistration

Johan Hovold <[email protected]>
spi: qup: fix controller deregistration

Johan Hovold <[email protected]>
spi: meson-spicc: fix controller deregistration

Johan Hovold <[email protected]>
spi: lantiq-ssc: fix controller deregistration

Johan Hovold <[email protected]>
regulator: bd9571mwv: fix OF node reference imbalance

Johan Hovold <[email protected]>
regulator: s2dos05: fix OF node reference imbalance

Johan Hovold <[email protected]>
regulator: act8945a: fix OF node reference imbalance

Jai Luthra <[email protected]>
media: i2c: imx283: Fix hang when going from large to small resolution

Ethan Tidmore <[email protected]>
media: intel/ipu6: fix error pointer dereference

Janne Grunau <[email protected]>
media: videobuf2: Set vma_flags in vb2_dma_sg_mmap

Johan Hovold <[email protected]>
regulator: rk808: fix OF node reference imbalance

Johan Hovold <[email protected]>
regulator: bq257xx: fix OF node reference imbalance

Jai Luthra <[email protected]>
media: i2c: imx283: Enter full standby when stopping streaming

Oliver Neukum <[email protected]>
media: rc: streamzap: Error handling in probe

Oliver Neukum <[email protected]>
media: rc: xbox_remote: heed DMA restrictions

Johan Hovold <[email protected]>
regulator: max77650: fix OF node reference imbalance

Johan Hovold <[email protected]>
spi: st-ssc4: fix controller deregistration

Johan Hovold <[email protected]>
regulator: mt6357: fix OF node reference imbalance

Sakari Ailus <[email protected]>
staging: media: atomisp: Disallow all private IOCTLs

Josua Mayer <[email protected]>

    arm64: dts: lx2160a-cex7/lx2162a-sr-som: fix usd-cd & gpio pinmux

Johan Hovold <[email protected]>
spi: atmel: fix controller deregistration

Johan Hovold <[email protected]>
spi: bcm63xx: fix controller deregistration

Matthew Brost <[email protected]>
drm/gpusvm: Force unmapping on error in drm_gpusvm_get_pages

Ziyi Guo <[email protected]>
media: chips-media: wave5: add missing spinlock protection for handle_dynamic_resolution_change()

Ziyi Guo <[email protected]>
media: chips-media: wave5: add missing spinlock protection for send_eos_event()

Haoxiang Li <[email protected]>
media: chips-media: wave5: fix a potential memory leak in wave5_vdi_init()

Cristian Ciocaltea <[email protected]>

    media: dt-bindings: rockchip,vdec: Mark reg-names required for RK35{76,88}

Cristian Ciocaltea <[email protected]>

    media: dt-bindings: rockchip,vdec: Add alternative reg-names order for RK35{76,88}

Alexander Koskovich <[email protected]>
media: i2c: ov8856: free control handler on error in ov8856_init_controls()

Tomi Valkeinen <[email protected]>
media: renesas: vin: Fix RAW8 (again)

Tomi Valkeinen <[email protected]>
media: renesas: vsp1: Fix NULL pointer deref on module unload

Guoniu Zhou <[email protected]>
media: nxp: imx8-isi: Reduce minimum queued buffers from 2 to 0

Krzysztof Kozlowski <[email protected]>
drm/msm/hdmi: Fix wrong CTRL1 register used in writing info frames

Ricardo Ribalda <[email protected]>
media: uvcvideo: Enable VB2_DMABUF for metadata stream

Tomasz Pakuła <[email protected]>
HID: pidff: Fix integer overflow in pidff_rescale

Benjamin Tissoires <[email protected]>
HID: core: introduce hid_safe_input_report()

Benjamin Tissoires <[email protected]>
HID: pass the buffer size to hid_report_raw_event

Sangyun Kim <[email protected]>
HID: appletb-kbd: run inactivity autodim from workqueues

Sangyun Kim <[email protected]>
HID: appletb-kbd: fix UAF in inactivity-timer cleanup path

T.J. Mercier <[email protected]>
HID: playstation: Clamp num_touch_reports

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

Diffstat:

 .../devicetree/bindings/media/rockchip,vdec.yaml   |  22 +-
 Makefile                                           |   4 +-
 arch/arm64/boot/dts/broadcom/bcm2712-d-rpi-5-b.dts |  14 +
 .../arm64/boot/dts/freescale/fsl-lx2160a-cex7.dtsi |   7 +
 .../dts/freescale/fsl-lx2160a-clearfog-itx.dtsi    |   2 +
 arch/arm64/boot/dts/freescale/fsl-lx2160a.dtsi     |  24 ++
 .../boot/dts/freescale/fsl-lx2162a-clearfog.dts    |   2 +
 .../boot/dts/freescale/fsl-lx2162a-sr-som.dtsi     |   7 +
 arch/arm64/boot/dts/qcom/lemans.dtsi               |   8 +-
 arch/arm64/boot/dts/ti/k3-am62a7-sk.dts            |   2 +-
 arch/loongarch/Kbuild                              |   2 +-
 arch/loongarch/include/asm/asm-prototypes.h        |  20 +
 arch/loongarch/include/asm/kvm_host.h              |   3 -
 arch/loongarch/kvm/Makefile                        |   3 +-
 arch/loongarch/kvm/main.c                          |  35 +-
 arch/loongarch/kvm/switch.S                        |  20 +-
 arch/powerpc/platforms/pseries/papr-hvpipe.c       |  59 +--
 block/blk-zoned.c                                  | 149 +++-----
 drivers/edac/versalnet_edac.c                      |   3 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c   |   3 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c           |  13 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c            |   6 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h           |  11 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c            |   3 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c             |   9 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h             |   2 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm_cpu.c         |  12 +-
 drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c             |  31 +-
 drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c              |   3 -
 drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c             |   4 +-
 drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c              |  25 +-
 drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c              |  46 ++-
 drivers/gpu/drm/amd/amdkfd/kfd_chardev.c           |  33 +-
 .../gpu/drm/amd/amdkfd/kfd_device_queue_manager.c  |   6 +-
 drivers/gpu/drm/amd/amdkfd/kfd_priv.h              |  10 +-
 drivers/gpu/drm/amd/amdkfd/kfd_svm.c               |   4 +-
 drivers/gpu/drm/amd/amdkfd/kfd_topology.c          |  11 +
 drivers/gpu/drm/amd/display/dc/core/dc_resource.c  |   2 +-
 .../gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c    |  13 +-
 .../gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c   |  10 +-
 drivers/gpu/drm/bridge/tda998x_drv.c               |   2 +-
 drivers/gpu/drm/drm_gem.c                          |  25 +-
 drivers/gpu/drm/drm_gem_framebuffer_helper.c       |   4 +-
 drivers/gpu/drm/drm_gpusvm.c                       |   1 +
 drivers/gpu/drm/exynos/exynos_drm_mic.c            |   8 +-
 drivers/gpu/drm/i915/display/intel_psr.c           |   2 +-
 drivers/gpu/drm/imx/ipuv3/parallel-display.c       |  15 +-
 drivers/gpu/drm/msm/hdmi/hdmi_bridge.c             |   4 +-
 drivers/gpu/drm/msm/msm_drv.c                      |   7 +-
 drivers/gpu/drm/msm/msm_gpu.c                      |  42 +--
 drivers/gpu/drm/panel/panel-boe-tv101wum-nl6.c     |   2 +
 drivers/gpu/drm/panel/panel-himax-hx83102.c        |   2 +
 drivers/gpu/drm/radeon/ci_dpm.c                    |   9 +-
 drivers/gpu/drm/tiny/appletbdrm.c                  |   4 +-
 drivers/gpu/drm/udl/udl_main.c                     |   3 +-
 drivers/gpu/drm/udl/udl_modeset.c                  |   5 +-
 drivers/gpu/drm/xe/xe_bo.c                         |   8 +-
 drivers/gpu/drm/xe/xe_dma_buf.c                    |  23 +-
 drivers/gpu/drm/xe/xe_vm_madvise.c                 |  47 +++
 drivers/hid/bpf/hid_bpf_dispatch.c                 |   6 +-
 drivers/hid/hid-appletb-kbd.c                      |  58 +--
 drivers/hid/hid-core.c                             |  67 +++-
 drivers/hid/hid-gfrm.c                             |   4 +-
 drivers/hid/hid-logitech-hidpp.c                   |   2 +-
 drivers/hid/hid-multitouch.c                       |   2 +-
 drivers/hid/hid-playstation.c                      |   6 +-
 drivers/hid/hid-primax.c                           |   2 +-
 drivers/hid/hid-vivaldi-common.c                   |   2 +-
 drivers/hid/i2c-hid/i2c-hid-core.c                 |   7 +-
 drivers/hid/usbhid/hid-core.c                      |  11 +-
 drivers/hid/usbhid/hid-pidff.c                     |   7 +-
 drivers/hid/wacom_sys.c                            |   6 +-
 drivers/media/common/videobuf2/videobuf2-dma-sg.c  |   1 +
 drivers/media/dvb-frontends/dib8000.c              |   4 +-
 drivers/media/i2c/imx283.c                         |  15 +-
 drivers/media/i2c/imx412.c                         |   2 +-
 drivers/media/i2c/ov08d10.c                        |  21 +-
 drivers/media/i2c/ov8856.c                         |  10 +-
 drivers/media/pci/intel/ipu6/ipu6.c                |   2 +-
 drivers/media/pci/saa7164/saa7164-core.c           |  47 ++-
 drivers/media/pci/zoran/zoran_card.c               |   2 +-
 .../media/platform/chips-media/wave5/wave5-vdi.c   |   1 +
 .../platform/chips-media/wave5/wave5-vpu-dec.c     |  14 +-
 .../media/platform/nxp/imx8-isi/imx8-isi-video.c   |   2 +-
 .../media/platform/qcom/camss/camss-csid-gen3.c    |   6 +-
 drivers/media/platform/qcom/camss/camss.c          |  80 ++--
 drivers/media/platform/qcom/iris/Kconfig           |   2 +-
 drivers/media/platform/qcom/iris/iris_buffer.c     |   8 +-
 drivers/media/platform/qcom/iris/iris_hfi_queue.c  |   2 +-
 drivers/media/platform/qcom/iris/iris_vpu_buffer.h |   2 +-
 drivers/media/platform/qcom/venus/Kconfig          |   2 +-
 drivers/media/platform/renesas/rcar-vin/rcar-dma.c |  22 ++
 .../media/platform/renesas/rcar-vin/rcar-v4l2.c    |  12 +
 drivers/media/platform/renesas/vsp1/vsp1_drv.c     |   8 +-
 drivers/media/platform/ti/omap3isp/ispvideo.c      |   1 +
 drivers/media/rc/streamzap.c                       |  12 +-
 drivers/media/rc/xbox_remote.c                     |   9 +-
 drivers/media/usb/uvc/uvc_queue.c                  |   3 +-
 drivers/platform/x86/hp/hp-wmi.c                   |   5 +
 drivers/regulator/act8945a-regulator.c             |   3 +-
 drivers/regulator/bd9571mwv-regulator.c            |   3 +-
 drivers/regulator/bq257xx-regulator.c              |   3 +-
 drivers/regulator/max77650-regulator.c             |   2 +-
 drivers/regulator/mt6357-regulator.c               |   2 +-
 drivers/regulator/rk808-regulator.c                |   3 +-
 drivers/regulator/s2dos05-regulator.c              |   2 +-
 drivers/spi/spi-amlogic-spisg.c                    |   4 +-
 drivers/spi/spi-aspeed-smc.c                       |   9 +-
 drivers/spi/spi-at91-usart.c                       |   8 +-
 drivers/spi/spi-atmel.c                            |   8 +-
 drivers/spi/spi-bcm63xx.c                          |   8 +-
 drivers/spi/spi-bcmbca-hsspi.c                     |   4 +-
 drivers/spi/spi-cadence.c                          |  21 +-
 drivers/spi/spi-cavium-thunderx.c                  |   8 +-
 drivers/spi/spi-ch341.c                            |   7 +-
 drivers/spi/spi-coldfire-qspi.c                    |  10 +-
 drivers/spi/spi-dln2.c                             |   8 +-
 drivers/spi/spi-fsl-espi.c                         |  10 +-
 drivers/spi/spi-fsl-spi.c                          |  14 +-
 drivers/spi/spi-img-spfi.c                         |   8 +-
 drivers/spi/spi-imx.c                              |   1 +
 drivers/spi/spi-lantiq-ssc.c                       |   8 +-
 drivers/spi/spi-meson-spicc.c                      |   8 +-
 drivers/spi/spi-mpc52xx.c                          |   9 +-
 drivers/spi/spi-mpfs.c                             |   4 +-
 drivers/spi/spi-mt65xx.c                           |   4 +-
 drivers/spi/spi-mtk-nor.c                          |   4 +-
 drivers/spi/spi-mxic.c                             |   3 +-
 drivers/spi/spi-mxs.c                              |   8 +-
 drivers/spi/spi-npcm-pspi.c                        |   8 +-
 drivers/spi/spi-omap2-mcspi.c                      |   8 +-
 drivers/spi/spi-orion.c                            |  16 +-
 drivers/spi/spi-pic32-sqi.c                        |   8 +-
 drivers/spi/spi-pic32.c                            |  11 +-
 drivers/spi/spi-pl022.c                            |   8 +-
 drivers/spi/spi-qup.c                              |   8 +-
 drivers/spi/spi-rspi.c                             |  10 +-
 drivers/spi/spi-s3c64xx.c                          |   4 +-
 drivers/spi/spi-sh-hspi.c                          |  10 +-
 drivers/spi/spi-sh-msiof.c                         |  10 +-
 drivers/spi/spi-slave-mt27xx.c                     |  10 +-
 drivers/spi/spi-sprd.c                             |   8 +-
 drivers/spi/spi-st-ssc4.c                          |   8 +-
 drivers/spi/spi-tegra114.c                         |   8 +-
 drivers/spi/spi-tegra20-sflash.c                   |   8 +-
 drivers/spi/spi-uniphier.c                         |  24 +-
 drivers/spi/spi-zynq-qspi.c                        |  55 +--
 drivers/staging/greybus/hid.c                      |   2 +-
 drivers/staging/media/atomisp/pci/atomisp_ioctl.c  |   4 +
 drivers/staging/media/imx/imx-media-csi.c          |  40 +-
 drivers/usb/dwc3/core.c                            | 204 +++++-----
 drivers/usb/dwc3/core.h                            |  10 +-
 drivers/usb/dwc3/debugfs.c                         |  44 +--
 drivers/usb/dwc3/drd.c                             |  76 ++--
 drivers/usb/dwc3/ep0.c                             |  20 +-
 drivers/usb/dwc3/gadget.c                          | 162 ++++----
 drivers/usb/dwc3/gadget.h                          |   4 +-
 drivers/usb/dwc3/io.h                              |   7 +-
 drivers/usb/dwc3/ulpi.c                            |  10 +-
 drivers/usb/typec/tcpm/tcpm.c                      |   2 +
 drivers/video/fbdev/core/bitblit.c                 | 120 +++---
 drivers/video/fbdev/core/fbcon.c                   | 419 ++++++++++-----------
 drivers/video/fbdev/core/fbcon.h                   |   6 +-
 drivers/video/fbdev/core/fbcon_ccw.c               | 146 +++----
 drivers/video/fbdev/core/fbcon_cw.c                | 146 +++----
 drivers/video/fbdev/core/fbcon_rotate.c            |  43 ++-
 drivers/video/fbdev/core/fbcon_rotate.h            |   6 +-
 drivers/video/fbdev/core/fbcon_ud.c                | 162 ++++----
 drivers/video/fbdev/core/softcursor.c              |  18 +-
 drivers/video/fbdev/core/tileblit.c                |  28 +-
 fs/btrfs/ioctl.c                                   |   5 +-
 fs/btrfs/space-info.c                              |   8 +-
 fs/btrfs/sysfs.c                                   |   5 +-
 fs/btrfs/sysfs.h                                   |   3 +-
 include/linux/damon.h                              |   1 +
 include/linux/fprobe.h                             |   3 +-
 include/linux/hid.h                                |   6 +-
 include/linux/hid_bpf.h                            |  14 +-
 include/linux/sched/ext.h                          |   1 +
 io_uring/zcrx.c                                    |  17 +-
 kernel/sched/ext.c                                 |  26 +-
 kernel/sched/ext_internal.h                        |  13 +-
 kernel/trace/fprobe.c                              | 380 +++++++++++++------
 mm/damon/core.c                                    |  25 +-
 mm/damon/lru_sort.c                                |  83 ++--
 mm/damon/reclaim.c                                 |  83 ++--
 net/batman-adv/bat_iv_ogm.c                        |  85 +++--
 net/batman-adv/bridge_loop_avoidance.c             |  11 +-
 net/batman-adv/main.c                              |   1 +
 net/batman-adv/tp_meter.c                          | 116 +++++-
 net/batman-adv/tp_meter.h                          |   1 +
 net/batman-adv/types.h                             |   4 +
 net/bluetooth/hci_conn.c                           |  19 +-
 net/sctp/socket.c                                  |   9 +
 net/vmw_vsock/af_vsock.c                           |   6 +-
 net/vmw_vsock/virtio_transport_common.c            |  57 ++-
 196 files changed, 2699 insertions(+), 1709 deletions(-)
author maintainer
#1 lore raw

PATCH 1/188 HID: playstation: Clamp num_touch_reports

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

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

From: T.J. Mercier <[email protected]>

commit cac61b58a3b6340c52afa06bb15eac033158db2f upstream.

A device would never lie about the number of touch reports would it?

If it does the loop in dualshock4_parse_report will read off the end of the touch_reports array, up to about 2 KiB for the maximum number of 256 loop iteraions. The data that is read is emitted via evdev if the DS4_TOUCH_POINT_INACTIVE bit happens to be set. Protect against this by clamping the num_touch_reports value provided by the device to the maximum size of the touch_reports array.

Fixes 752038248808 ("HID: playstation: add DualShock4 touchpad support.")
---
 drivers/hid/hid-playstation.c |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/hid/hid-playstation.c +4 -2
@@ -2378,7 +2378,8 @@ static int dualshock4_parse_report(struc
 			(struct dualshock4_input_report_usb *)data;
 
 		ds4_report = &usb->common;
-		num_touch_reports = usb->num_touch_reports;
+		num_touch_reports = min_t(u8, usb->num_touch_reports,
+					  ARRAY_SIZE(usb->touch_reports));
 		touch_reports = usb->touch_reports;
 	} else if (hdev->bus == BUS_BLUETOOTH && report->id == DS4_INPUT_REPORT_BT &&
 		   size == DS4_INPUT_REPORT_BT_SIZE) {
@@ -2392,7 +2393,8 @@ static int dualshock4_parse_report(struc
 		}
 
 		ds4_report = &bt->common;
-		num_touch_reports = bt->num_touch_reports;
+		num_touch_reports = min_t(u8, bt->num_touch_reports,
+					  ARRAY_SIZE(bt->touch_reports));
 		touch_reports = bt->touch_reports;
 	} else if (hdev->bus == BUS_BLUETOOTH &&
 		   report->id == DS4_INPUT_REPORT_BT_MINIMAL &&
author maintainer
#2 lore raw

PATCH 2/188 HID: appletb-kbd: fix UAF in inactivity-timer cleanup path

In reply to Greg Kroah-Hartman

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

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

From: Sangyun Kim <[email protected]>

commit 4db2af929279c799b5653a39eb0795c72baffca4 upstream.

Commit 38224c472a03 ("HID: appletb-kbd: fix slab use-after-free bug in appletb_kbd_probe") added timer_delete_sync(&kbd->inactivity_timer) to both the probe close_hw error path and appletb_kbd_remove(), but the way it was wired in left the inactivity timer reachable during driver tear-down via two distinct windows.

Window A -- put_device() before timer_delete_sync():

	put_device(&kbd->backlight_dev->dev);
	timer_delete_sync(&kbd->inactivity_timer);

The inactivity_timer softirq reads kbd->backlight_dev and calls backlight_device_set_brightness() -> mutex_lock(&ops_lock). If a concurrent hid_appletb_bl unbind drops the last devm reference between these two calls, the backlight_device is freed and the mutex_lock() touches freed memory.

Window B -- backlight cleanup before hid_hw_stop():

	if (kbd->backlight_dev) {
		timer_delete_sync(...);
		put_device(...);
	}
	hid_hw_close(hdev);
	hid_hw_stop(hdev);

Even after Window A is closed, hid_hw_close()/hid_hw_stop() still run afterwards, so a late ".event" callback from the HID core (USB URB completion on real Apple hardware) can arrive after timer_delete_sync() drained the softirq but before put_device() drops the reference. That callback reaches reset_inactivity_timer(), which calls mod_timer() and re-arms the timer. The freshly re-armed timer can then fire on the about-to-be-freed backlight_device.

Both windows produce the same KASAN slab-use-after-free:

  BUG: KASAN: slab-use-after-free in __mutex_lock+0x1aab/0x21c0
  Read of size 8 at addr ffff88803ee9a108 by task swapper/0/0
  Call Trace:
   <IRQ>
   __mutex_lock
   backlight_device_set_brightness
   appletb_inactivity_timer
   call_timer_fn
   run_timer_softirq
   handle_softirqs
  Allocated by task N:
   devm_backlight_device_register
   appletb_bl_probe
  Freed by task M:
   (concurrent hid_appletb_bl unbind path)

Close both windows at once by reworking the tear-down in appletb_kbd_remove() and in the probe close_hw error path so that

1) hid_hw_close()/hid_hw_stop() run before the backlight cleanup, guaranteeing no further .event callback can fire and re-arm the timer, and
2) inside the "if (kbd->backlight_dev)" block, timer_delete_sync() runs before put_device(), so the softirq is drained before the final reference is dropped.

Fixes 38224c472a03 ("HID: appletb-kbd: fix slab use-after-free bug in appletb_kbd_probe")
---
 drivers/hid/hid-appletb-kbd.c |   16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)
drivers/hid/hid-appletb-kbd.c +8 -8
@@ -440,13 +440,13 @@ static int appletb_kbd_probe(struct hid_
 unregister_handler:
 	input_unregister_handler(&kbd->inp_handler);
 close_hw:
-	if (kbd->backlight_dev) {
-		put_device(&kbd->backlight_dev->dev);
-		timer_delete_sync(&kbd->inactivity_timer);
-	}
 	hid_hw_close(hdev);
 stop_hw:
 	hid_hw_stop(hdev);
+	if (kbd->backlight_dev) {
+		timer_delete_sync(&kbd->inactivity_timer);
+		put_device(&kbd->backlight_dev->dev);
+	}
 	return ret;
 }
 
@@ -457,13 +457,13 @@ static void appletb_kbd_remove(struct hi
 	appletb_kbd_set_mode(kbd, APPLETB_KBD_MODE_OFF);
 
 	input_unregister_handler(&kbd->inp_handler);
+	hid_hw_close(hdev);
+	hid_hw_stop(hdev);
+
 	if (kbd->backlight_dev) {
-		put_device(&kbd->backlight_dev->dev);
 		timer_delete_sync(&kbd->inactivity_timer);
+		put_device(&kbd->backlight_dev->dev);
 	}
-
-	hid_hw_close(hdev);
-	hid_hw_stop(hdev);
 }
 
 #ifdef CONFIG_PM
author maintainer
#3 lore raw

PATCH 3/188 HID: appletb-kbd: run inactivity autodim from workqueues

In reply to Greg Kroah-Hartman

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

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

From: Sangyun Kim <[email protected]>

commit 1654e53349d4e657b331de354313461f401f5063 upstream.

The autodim code in hid-appletb-kbd takes backlight_device->ops_lock via backlight_device_set_brightness() -> mutex_lock() from two different atomic contexts:

* appletb_inactivity_timer() is a struct timer_list callback, so it runs in softirq context. Every expiry triggers

     BUG: sleeping function called from invalid context at kernel/locking/mutex.c:591
     Call Trace:
      <IRQ>
      __might_resched
      __mutex_lock
      backlight_device_set_brightness
      appletb_inactivity_timer
      call_timer_fn
      run_timer_softirq

* reset_inactivity_timer() is called from appletb_kbd_hid_event() and appletb_kbd_inp_event(). On real USB hardware these run in softirq/IRQ context (URB completion and input-event dispatch). When the Touch Bar has already been dimmed or turned off, the reset path calls backlight_device_set_brightness() directly to restore brightness, producing the same warning.

Both call sites hit the same mutex_lock()-from-atomic bug. Fix them together by moving the blocking work onto the system workqueue:

* Convert the inactivity timer from struct timer_list to struct delayed_work; the callback (appletb_inactivity_work) now runs in process context where mutex_lock() is legal.
* Add a dedicated struct work_struct restore_brightness_work and have reset_inactivity_timer() schedule it instead of calling backlight_device_set_brightness() directly.

Cancel both works synchronously during driver tear-down alongside the existing backlight reference drop.

The semantics are unchanged (same delays, same state transitions on dim, turn-off and user activity); only the execution context of the sleeping call changes. The timer field and callback are renamed to match their new type; reset_inactivity_timer() keeps its name because it is invoked from input event paths that read naturally as "reset the inactivity timer".

Fixes 93a0fc489481 ("HID: hid-appletb-kbd: add support for automatic brightness control while using the touchbar")
---
 drivers/hid/hid-appletb-kbd.c |   44 ++++++++++++++++++++++++++++--------------
 1 file changed, 30 insertions(+), 14 deletions(-)
drivers/hid/hid-appletb-kbd.c +30 -14
@@ -17,7 +17,7 @@
 #include <linux/module.h>
 #include <linux/string.h>
 #include <linux/backlight.h>
-#include <linux/timer.h>
+#include <linux/workqueue.h>
 #include <linux/input/sparse-keymap.h>
 
 #include "hid-ids.h"
@@ -62,7 +62,8 @@ struct appletb_kbd {
 	struct input_handle kbd_handle;
 	struct input_handle tpd_handle;
 	struct backlight_device *backlight_dev;
-	struct timer_list inactivity_timer;
+	struct delayed_work inactivity_work;
+	struct work_struct restore_brightness_work;
 	bool has_dimmed;
 	bool has_turned_off;
 	u8 saved_mode;
@@ -164,16 +165,18 @@ static int appletb_tb_key_to_slot(unsign
 	}
 }
 
-static void appletb_inactivity_timer(struct timer_list *t)
+static void appletb_inactivity_work(struct work_struct *work)
 {
-	struct appletb_kbd *kbd = timer_container_of(kbd, t, inactivity_timer);
+	struct appletb_kbd *kbd = container_of(to_delayed_work(work),
+					       struct appletb_kbd,
+					       inactivity_work);
 
 	if (kbd->backlight_dev && appletb_tb_autodim) {
 		if (!kbd->has_dimmed) {
 			backlight_device_set_brightness(kbd->backlight_dev, 1);
 			kbd->has_dimmed = true;
-			mod_timer(&kbd->inactivity_timer,
-				jiffies + secs_to_jiffies(appletb_tb_idle_timeout));
+			mod_delayed_work(system_wq, &kbd->inactivity_work,
+					 secs_to_jiffies(appletb_tb_idle_timeout));
 		} else if (!kbd->has_turned_off) {
 			backlight_device_set_brightness(kbd->backlight_dev, 0);
 			kbd->has_turned_off = true;
@@ -181,16 +184,25 @@ static void appletb_inactivity_timer(str
 	}
 }
 
+static void appletb_restore_brightness_work(struct work_struct *work)
+{
+	struct appletb_kbd *kbd = container_of(work, struct appletb_kbd,
+					       restore_brightness_work);
+
+	if (kbd->backlight_dev)
+		backlight_device_set_brightness(kbd->backlight_dev, 2);
+}
+
 static void reset_inactivity_timer(struct appletb_kbd *kbd)
 {
 	if (kbd->backlight_dev && appletb_tb_autodim) {
 		if (kbd->has_dimmed || kbd->has_turned_off) {
-			backlight_device_set_brightness(kbd->backlight_dev, 2);
 			kbd->has_dimmed = false;
 			kbd->has_turned_off = false;
+			schedule_work(&kbd->restore_brightness_work);
 		}
-		mod_timer(&kbd->inactivity_timer,
-			jiffies + secs_to_jiffies(appletb_tb_dim_timeout));
+		mod_delayed_work(system_wq, &kbd->inactivity_work,
+				 secs_to_jiffies(appletb_tb_dim_timeout));
 	}
 }
 
@@ -408,9 +420,11 @@ static int appletb_kbd_probe(struct hid_
 		dev_err_probe(dev, -ENODEV, "Failed to get backlight device\n");
 	} else {
 		backlight_device_set_brightness(kbd->backlight_dev, 2);
-		timer_setup(&kbd->inactivity_timer, appletb_inactivity_timer, 0);
-		mod_timer(&kbd->inactivity_timer,
-			jiffies + secs_to_jiffies(appletb_tb_dim_timeout));
+		INIT_DELAYED_WORK(&kbd->inactivity_work, appletb_inactivity_work);
+		INIT_WORK(&kbd->restore_brightness_work,
+			  appletb_restore_brightness_work);
+		mod_delayed_work(system_wq, &kbd->inactivity_work,
+				 secs_to_jiffies(appletb_tb_dim_timeout));
 	}
 
 	kbd->inp_handler.event = appletb_kbd_inp_event;
@@ -444,7 +458,8 @@ close_hw:
 stop_hw:
 	hid_hw_stop(hdev);
 	if (kbd->backlight_dev) {
-		timer_delete_sync(&kbd->inactivity_timer);
+		cancel_delayed_work_sync(&kbd->inactivity_work);
+		cancel_work_sync(&kbd->restore_brightness_work);
 		put_device(&kbd->backlight_dev->dev);
 	}
 	return ret;
@@ -461,7 +476,8 @@ static void appletb_kbd_remove(struct hi
 	hid_hw_stop(hdev);
 
 	if (kbd->backlight_dev) {
-		timer_delete_sync(&kbd->inactivity_timer);
+		cancel_delayed_work_sync(&kbd->inactivity_work);
+		cancel_work_sync(&kbd->restore_brightness_work);
 		put_device(&kbd->backlight_dev->dev);
 	}
 }
author maintainer
#4 lore raw

PATCH 4/188 HID: pass the buffer size to hid_report_raw_event

In reply to Greg Kroah-Hartman

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

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

From: Benjamin Tissoires <[email protected]>

commit 2c85c61d1332e1e16f020d76951baf167dcb6f7a upstream.

commit 0a3fe972a7cb ("HID: core: Mitigate potential OOB by removing bogus memset()") enforced the provided data to be at least the size of the declared buffer in the report descriptor to prevent a buffer overflow. However, we can try to be smarter by providing both the buffer size and the data size, meaning that hid_report_raw_event() can make better decision whether we should plaining reject the buffer (buffer overflow attempt) or if we can safely memset it to 0 and pass it to the rest of the stack.

Fixes 0a3fe972a7cb ("HID: core: Mitigate potential OOB by removing bogus memset()")
---
 drivers/hid/bpf/hid_bpf_dispatch.c |    6 +++--
 drivers/hid/hid-core.c             |   42 ++++++++++++++++++++++++-------------
 drivers/hid/hid-gfrm.c             |    4 +--
 drivers/hid/hid-logitech-hidpp.c   |    2 -
 drivers/hid/hid-multitouch.c       |    2 -
 drivers/hid/hid-primax.c           |    2 -
 drivers/hid/hid-vivaldi-common.c   |    2 -
 drivers/hid/wacom_sys.c            |    6 ++---
 drivers/staging/greybus/hid.c      |    2 -
 include/linux/hid.h                |    4 +--
 include/linux/hid_bpf.h            |   14 +++++++-----
 11 files changed, 53 insertions(+), 33 deletions(-)
drivers/hid/bpf/hid_bpf_dispatch.c +4 -2
@@ -24,7 +24,8 @@ EXPORT_SYMBOL(hid_ops);
 
 u8 *
 dispatch_hid_bpf_device_event(struct hid_device *hdev, enum hid_report_type type, u8 *data,
-			      u32 *size, int interrupt, u64 source, bool from_bpf)
+			      size_t *buf_size, u32 *size, int interrupt, u64 source,
+			      bool from_bpf)
 {
 	struct hid_bpf_ctx_kern ctx_kern = {
 		.ctx = {
@@ -74,6 +75,7 @@ dispatch_hid_bpf_device_event(struct hid
 		*size = ret;
 	}
 
+	*buf_size = ctx_kern.ctx.allocated_size;
 	return ctx_kern.data;
 }
 EXPORT_SYMBOL_GPL(dispatch_hid_bpf_device_event);
@@ -508,7 +510,7 @@ __hid_bpf_input_report(struct hid_bpf_ct
 	if (ret)
 		return ret;
 
-	return hid_ops->hid_input_report(ctx->hid, type, buf, size, 0, (u64)(long)ctx, true,
+	return hid_ops->hid_input_report(ctx->hid, type, buf, size, size, 0, (u64)(long)ctx, true,
 					 lock_already_taken);
 }
 
drivers/hid/hid-core.c +28 -14

106 lines not on this page · raw message

drivers/hid/hid-gfrm.c +2 -2
@@ -66,7 +66,7 @@ static int gfrm_raw_event(struct hid_dev
 	switch (data[1]) {
 	case GFRM100_SEARCH_KEY_DOWN:
 		ret = hid_report_raw_event(hdev, HID_INPUT_REPORT, search_key_dn,
-					   sizeof(search_key_dn), 1);
+					   sizeof(search_key_dn), sizeof(search_key_dn), 1);
 		break;
 
 	case GFRM100_SEARCH_KEY_AUDIO_DATA:
@@ -74,7 +74,7 @@ static int gfrm_raw_event(struct hid_dev
 
 	case GFRM100_SEARCH_KEY_UP:
 		ret = hid_report_raw_event(hdev, HID_INPUT_REPORT, search_key_up,
-					   sizeof(search_key_up), 1);
+					   sizeof(search_key_up), sizeof(search_key_up), 1);
 		break;
 
 	default:
drivers/hid/hid-logitech-hidpp.c +1 -1
@@ -3664,7 +3664,7 @@ static int hidpp10_consumer_keys_raw_eve
 	memcpy(&consumer_report[1], &data[3], 4);
 	/* We are called from atomic context */
 	hid_report_raw_event(hidpp->hid_dev, HID_INPUT_REPORT,
-			     consumer_report, 5, 1);
+			     consumer_report, sizeof(consumer_report), 5, 1);
 
 	return 1;
 }
drivers/hid/hid-multitouch.c +1 -1
@@ -531,7 +531,7 @@ static void mt_get_feature(struct hid_de
 		}
 
 		ret = hid_report_raw_event(hdev, HID_FEATURE_REPORT, buf,
-					   size, 0);
+					   size, size, 0);
 		if (ret)
 			dev_warn(&hdev->dev, "failed to report feature\n");
 	}
drivers/hid/hid-primax.c +1 -1
@@ -44,7 +44,7 @@ static int px_raw_event(struct hid_devic
 			data[0] |= (1 << (data[idx] - 0xE0));
 			data[idx] = 0;
 		}
-		hid_report_raw_event(hid, HID_INPUT_REPORT, data, size, 0);
+		hid_report_raw_event(hid, HID_INPUT_REPORT, data, size, size, 0);
 		return 1;
 
 	default:	/* unknown report */
drivers/hid/hid-vivaldi-common.c +1 -1
@@ -85,7 +85,7 @@ void vivaldi_feature_mapping(struct hid_
 	}
 
 	ret = hid_report_raw_event(hdev, HID_FEATURE_REPORT, report_data,
-				   report_len, 0);
+				   report_len, report_len, 0);
 	if (ret) {
 		dev_warn(&hdev->dev, "failed to report feature %d\n",
 			 field->report->id);
drivers/hid/wacom_sys.c +3 -3
@@ -90,7 +90,7 @@ static void wacom_wac_queue_flush(struct
 			kfree(buf);
 			continue;
 		}
-		err = hid_report_raw_event(hdev, HID_INPUT_REPORT, buf, size, false);
+		err = hid_report_raw_event(hdev, HID_INPUT_REPORT, buf, size, size, false);
 		if (err) {
 			hid_warn(hdev, "%s: unable to flush event due to error %d\n",
 				 __func__, err);
@@ -334,7 +334,7 @@ static void wacom_feature_mapping(struct
 					       data, n, WAC_CMD_RETRIES);
 			if (ret == n && features->type == HID_GENERIC) {
 				ret = hid_report_raw_event(hdev,
-					HID_FEATURE_REPORT, data, n, 0);
+					HID_FEATURE_REPORT, data, n, n, 0);
 			} else if (ret == 2 && features->type != HID_GENERIC) {
 				features->touch_max = data[1];
 			} else {
@@ -395,7 +395,7 @@ static void wacom_feature_mapping(struct
 					data, n, WAC_CMD_RETRIES);
 		if (ret == n) {
 			ret = hid_report_raw_event(hdev, HID_FEATURE_REPORT,
-						   data, n, 0);
+						   data, n, n, 0);
 		} else {
 			hid_warn(hdev, "%s: could not retrieve sensor offsets\n",
 				 __func__);
drivers/staging/greybus/hid.c +1 -1
@@ -201,7 +201,7 @@ static void gb_hid_init_report(struct gb
 	 * we just need to setup the input fields, so using
 	 * hid_report_raw_event is safe.
 	 */
-	hid_report_raw_event(ghid->hid, report->type, ghid->inbuf, size, 1);
+	hid_report_raw_event(ghid->hid, report->type, ghid->inbuf, ghid->bufsize, size, 1);
 }
 
 static void gb_hid_init_reports(struct gb_hid *ghid)
include/linux/hid.h +2 -2
@@ -1258,8 +1258,8 @@ static inline u32 hid_report_len(struct
 	return DIV_ROUND_UP(report->size, 8) + (report->id > 0);
 }
 
-int hid_report_raw_event(struct hid_device *hid, enum hid_report_type type, u8 *data, u32 size,
-			 int interrupt);
+int hid_report_raw_event(struct hid_device *hid, enum hid_report_type type, u8 *data,
+			 size_t bufsize, u32 size, int interrupt);
 
 /* HID quirks API */
 unsigned long hid_lookup_quirk(const struct hid_device *hdev);
include/linux/hid_bpf.h +9 -5
@@ -72,8 +72,8 @@ struct hid_ops {
 	int (*hid_hw_output_report)(struct hid_device *hdev, __u8 *buf, size_t len,
 				    u64 source, bool from_bpf);
 	int (*hid_input_report)(struct hid_device *hid, enum hid_report_type type,
-				u8 *data, u32 size, int interrupt, u64 source, bool from_bpf,
-				bool lock_already_taken);
+				u8 *data, size_t bufsize, u32 size, int interrupt, u64 source,
+				bool from_bpf, bool lock_already_taken);
 	struct module *owner;
 	const struct bus_type *bus_type;
 };
@@ -200,7 +200,8 @@ struct hid_bpf {
 
 #ifdef CONFIG_HID_BPF
 u8 *dispatch_hid_bpf_device_event(struct hid_device *hid, enum hid_report_type type, u8 *data,
-				  u32 *size, int interrupt, u64 source, bool from_bpf);
+				  size_t *buf_size, u32 *size, int interrupt, u64 source,
+				  bool from_bpf);
 int dispatch_hid_bpf_raw_requests(struct hid_device *hdev,
 				  unsigned char reportnum, __u8 *buf,
 				  u32 size, enum hid_report_type rtype,
@@ -215,8 +216,11 @@ int hid_bpf_device_init(struct hid_devic
 const u8 *call_hid_bpf_rdesc_fixup(struct hid_device *hdev, const u8 *rdesc, unsigned int *size);
 #else /* CONFIG_HID_BPF */
 static inline u8 *dispatch_hid_bpf_device_event(struct hid_device *hid, enum hid_report_type type,
-						u8 *data, u32 *size, int interrupt,
-						u64 source, bool from_bpf) { return data; }
+						u8 *data, size_t *buf_size, u32 *size,
+						int interrupt, u64 source, bool from_bpf)
+{
+	return data;
+}
 static inline int dispatch_hid_bpf_raw_requests(struct hid_device *hdev,
 						unsigned char reportnum, u8 *buf,
 						u32 size, enum hid_report_type rtype,
author maintainer
#5 lore raw

PATCH 5/188 HID: core: introduce hid_safe_input_report()

In reply to Greg Kroah-Hartman

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

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

From: Benjamin Tissoires <[email protected]>

commit 206342541fc887ae919774a43942dc883161fece upstream.

hid_input_report() is used in too many places to have a commit that doesn't cross subsystem borders. Instead of changing the API, introduce a new one when things matters in the transport layers:
- usbhid
- i2chid

This effectively revert to the old behavior for those two transport layers.

Fixes 0a3fe972a7cb ("HID: core: Mitigate potential OOB by removing bogus memset()")
---
 drivers/hid/hid-core.c             |   25 +++++++++++++++++++++++++
 drivers/hid/i2c-hid/i2c-hid-core.c |    7 ++++---
 drivers/hid/usbhid/hid-core.c      |   11 ++++++-----
 include/linux/hid.h                |    2 ++
 4 files changed, 37 insertions(+), 8 deletions(-)
drivers/hid/hid-core.c +25 -0
@@ -2177,6 +2177,7 @@ unlock:
  * @interrupt: distinguish between interrupt and control transfers
  *
  * This is data entry for lower layers.
+ * Legacy, please use hid_safe_input_report() instead.
  */
 int hid_input_report(struct hid_device *hid, enum hid_report_type type, u8 *data, u32 size,
 		     int interrupt)
@@ -2187,6 +2188,30 @@ int hid_input_report(struct hid_device *
 }
 EXPORT_SYMBOL_GPL(hid_input_report);
 
+/**
+ * hid_safe_input_report - report data from lower layer (usb, bt...)
+ *
+ * @hid: hid device
+ * @type: HID report type (HID_*_REPORT)
+ * @data: report contents
+ * @bufsize: allocated size of the data buffer
+ * @size: useful size of data parameter
+ * @interrupt: distinguish between interrupt and control transfers
+ *
+ * This is data entry for lower layers.
+ * Please use this function instead of the non safe version because we provide
+ * here the size of the buffer, allowing hid-core to make smarter decisions
+ * regarding the incoming buffer.
+ */
+int hid_safe_input_report(struct hid_device *hid, enum hid_report_type type, u8 *data,
+			  size_t bufsize, u32 size, int interrupt)
+{
+	return __hid_input_report(hid, type, data, bufsize, size, interrupt, 0,
+				  false, /* from_bpf */
+				  false /* lock_already_taken */);
+}
+EXPORT_SYMBOL_GPL(hid_safe_input_report);
+
 bool hid_match_one_id(const struct hid_device *hdev,
 		      const struct hid_device_id *id)
 {
drivers/hid/i2c-hid/i2c-hid-core.c +4 -3
@@ -574,9 +574,10 @@ static void i2c_hid_get_input(struct i2c
 		if (ihid->hid->group != HID_GROUP_RMI)
 			pm_wakeup_event(&ihid->client->dev, 0);
 
-		hid_input_report(ihid->hid, HID_INPUT_REPORT,
-				ihid->inbuf + sizeof(__le16),
-				ret_size - sizeof(__le16), 1);
+		hid_safe_input_report(ihid->hid, HID_INPUT_REPORT,
+				      ihid->inbuf + sizeof(__le16),
+				      ihid->bufsize - sizeof(__le16),
+				      ret_size - sizeof(__le16), 1);
 	}
 
 	return;
drivers/hid/usbhid/hid-core.c +6 -5
@@ -283,9 +283,9 @@ static void hid_irq_in(struct urb *urb)
 			break;
 		usbhid_mark_busy(usbhid);
 		if (!test_bit(HID_RESUME_RUNNING, &usbhid->iofl)) {
-			hid_input_report(urb->context, HID_INPUT_REPORT,
-					 urb->transfer_buffer,
-					 urb->actual_length, 1);
+			hid_safe_input_report(urb->context, HID_INPUT_REPORT,
+					      urb->transfer_buffer, urb->transfer_buffer_length,
+					      urb->actual_length, 1);
 			/*
 			 * autosuspend refused while keys are pressed
 			 * because most keyboards don't wake up when
@@ -482,9 +482,10 @@ static void hid_ctrl(struct urb *urb)
 	switch (status) {
 	case 0:			/* success */
 		if (usbhid->ctrl[usbhid->ctrltail].dir == USB_DIR_IN)
-			hid_input_report(urb->context,
+			hid_safe_input_report(urb->context,
 				usbhid->ctrl[usbhid->ctrltail].report->type,
-				urb->transfer_buffer, urb->actual_length, 0);
+				urb->transfer_buffer, urb->transfer_buffer_length,
+				urb->actual_length, 0);
 		break;
 	case -ESHUTDOWN:	/* unplug */
 		unplug = 1;
include/linux/hid.h +2 -0
@@ -990,6 +990,8 @@ struct hid_field *hid_find_field(struct
 int hid_set_field(struct hid_field *, unsigned, __s32);
 int hid_input_report(struct hid_device *hid, enum hid_report_type type, u8 *data, u32 size,
 		     int interrupt);
+int hid_safe_input_report(struct hid_device *hid, enum hid_report_type type, u8 *data,
+			  size_t bufsize, u32 size, int interrupt);
 struct hid_field *hidinput_get_led_field(struct hid_device *hid);
 unsigned int hidinput_count_leds(struct hid_device *hid);
 __s32 hidinput_calc_abs_res(const struct hid_field *field, __u16 code);
author maintainer
#6 lore raw

PATCH 6/188 HID: pidff: Fix integer overflow in pidff_rescale

In reply to Greg Kroah-Hartman

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

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

From: Tomasz Pakuła <[email protected]>

commit 48d1677779ad6816978ad4a4f7588aec5ec960fe upstream.

Rescaling values close to the max (U16_MAX) temporarily creates values that exceed the s32 range. This caused value overflow in case when, for example, a periodic effect phase was higer than 180 degrees. In turn, rescale function could return values outised of the logical range of the HID field.

Fix by using 64 bit signed integer to store the value during calculation but still return only 32 bit integer.

Fixes 224ee88fe395 ("Input: add force feedback driver for PID devices")
---
 drivers/hid/usbhid/hid-pidff.c |    7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
drivers/hid/usbhid/hid-pidff.c +5 -2
@@ -11,6 +11,7 @@
 #include "hid-pidff.h"
 #include <linux/hid.h>
 #include <linux/input.h>
+#include <linux/math64.h>
 #include <linux/minmax.h>
 #include <linux/slab.h>
 #include <linux/usb.h>
@@ -325,8 +326,10 @@ static s32 pidff_clamp(s32 i, struct hid
  */
 static int pidff_rescale(int i, int max, struct hid_field *field)
 {
-	return i * (field->logical_maximum - field->logical_minimum) / max +
-	       field->logical_minimum;
+	/* 64 bits needed for big values during rescale */
+	s64 result = field->logical_maximum - field->logical_minimum;
+
+	return div_s64(result * i, max) + field->logical_minimum;
 }
 
 /*
author maintainer
#7 lore raw

PATCH 7/188 media: uvcvideo: Enable VB2_DMABUF for metadata stream

In reply to Greg Kroah-Hartman

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

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

From: Ricardo Ribalda <[email protected]>

commit fbac03467e53d8d72e5099c03df26d9adae11416 upstream.

The UVC driver has two video streams, one for the frames and another one for the metadata. Both streams share most of the codebase, but only the data stream declares support for DMABUF transfer mode.

I have tried the DMABUF transfer mode with CONFIG_DMABUF_HEAPS_SYSTEM and the frames looked correct.

This patch announces the support for DMABUF for the metadata stream. This is useful for apps/HALs that only want to support DMABUF.

---
 drivers/media/usb/uvc/uvc_queue.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
drivers/media/usb/uvc/uvc_queue.c +1 -2
@@ -243,7 +243,7 @@ int uvc_queue_init(struct uvc_video_queu
 	int ret;
 
 	queue->queue.type = type;
-	queue->queue.io_modes = VB2_MMAP | VB2_USERPTR;
+	queue->queue.io_modes = VB2_MMAP | VB2_USERPTR | VB2_DMABUF;
 	queue->queue.drv_priv = queue;
 	queue->queue.buf_struct_size = sizeof(struct uvc_buffer);
 	queue->queue.mem_ops = &vb2_vmalloc_memops;
@@ -256,7 +256,6 @@ int uvc_queue_init(struct uvc_video_queu
 		queue->queue.ops = &uvc_meta_queue_qops;
 		break;
 	default:
-		queue->queue.io_modes |= VB2_DMABUF;
 		queue->queue.ops = &uvc_queue_qops;
 		break;
 	}
author maintainer
#8 lore raw

PATCH 8/188 drm/msm/hdmi: Fix wrong CTRL1 register used in writing info frames

In reply to Greg Kroah-Hartman

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

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

From: Krzysztof Kozlowski <[email protected]>

commit 8c6c93b7db42d15c6e8c2540a648d32986a04b1a upstream.

Commit 384d2b03d0a1 ("drm/msm/hdmi: make use of the drm_connector_hdmi framework") changed the unconditional register writes in few places to
updates: read, apply mask, write. The new code reads REG_HDMI_INFOFRAME_CTRL1 register, applies fields/mask for HDMI_INFOFRAME_CTRL0 register and finally writes to HDMI_INFOFRAME_CTRL0. This difference between CTRL1 and CTRL0 looks unintended and may result in wrong data being written to HDMI bridge registers.

Fixes 384d2b03d0a1 ("drm/msm/hdmi: make use of the drm_connector_hdmi framework")

Patchwork: https://patchwork.freedesktop.org/patch/711156/

---
 drivers/gpu/drm/msm/hdmi/hdmi_bridge.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/msm/hdmi/hdmi_bridge.c +2 -2
@@ -80,7 +80,7 @@ static int msm_hdmi_config_avi_infoframe
 	for (i = 0; i < ARRAY_SIZE(buf); i++)
 		hdmi_write(hdmi, REG_HDMI_AVI_INFO(i), buf[i]);
 
-	val = hdmi_read(hdmi, REG_HDMI_INFOFRAME_CTRL1);
+	val = hdmi_read(hdmi, REG_HDMI_INFOFRAME_CTRL0);
 	val |= HDMI_INFOFRAME_CTRL0_AVI_SEND |
 		HDMI_INFOFRAME_CTRL0_AVI_CONT;
 	hdmi_write(hdmi, REG_HDMI_INFOFRAME_CTRL0, val);
@@ -116,7 +116,7 @@ static int msm_hdmi_config_audio_infofra
 		   buffer[9] << 16 |
 		   buffer[10] << 24);
 
-	val = hdmi_read(hdmi, REG_HDMI_INFOFRAME_CTRL1);
+	val = hdmi_read(hdmi, REG_HDMI_INFOFRAME_CTRL0);
 	val |= HDMI_INFOFRAME_CTRL0_AUDIO_INFO_SEND |
 		HDMI_INFOFRAME_CTRL0_AUDIO_INFO_CONT |
 		HDMI_INFOFRAME_CTRL0_AUDIO_INFO_SOURCE |
author maintainer
#9 lore raw

PATCH 9/188 media: nxp: imx8-isi: Reduce minimum queued buffers from 2 to 0

In reply to Greg Kroah-Hartman

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

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

From: Guoniu Zhou <[email protected]>

commit 2f38622d0f85f317be9e6b131da6cd511db94fd2 upstream.

Fix a hang issue when capturing a single frame with applications like cam in libcamera. It would hang waiting for the driver to complete the buffer, but streaming never starts because min_queued_buffers was set to 2.

The ISI module uses a ping-pong buffer mechanism that requires two buffers to be programmed at all times. However, when fewer than 2 user buffers are available, the driver use internal discard buffers to fill the remaining slot(s). Reduce minimum queued buffers from 2 to 0 allows streaming to start without any queued buffers.

---
 drivers/media/platform/nxp/imx8-isi/imx8-isi-video.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/media/platform/nxp/imx8-isi/imx8-isi-video.c +1 -1
@@ -1410,7 +1410,7 @@ int mxc_isi_video_register(struct mxc_is
 	q->mem_ops = &vb2_dma_contig_memops;
 	q->buf_struct_size = sizeof(struct mxc_isi_buffer);
 	q->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_MONOTONIC;
-	q->min_queued_buffers = 2;
+	q->min_queued_buffers = 0;
 	q->lock = &video->lock;
 	q->dev = pipe->isi->dev;
 
author maintainer
#10 lore raw

PATCH 10/188 media: renesas: vsp1: Fix NULL pointer deref on module unload

In reply to Greg Kroah-Hartman

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

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

From: Tomi Valkeinen <[email protected]>

commit 58b1e9664d8f74d55d8411cc7a7b275a76a6f24f upstream.

When unloading the module on gen 4, we hit a NULL pointer dereference. This is caused by the cleanup code calling vsp1_drm_cleanup() where it should be calling vsp1_vspx_cleanup().

Fix this by checking the IP version and calling the drm or vspx function accordingly, the same way as the init code does.

---
 drivers/media/platform/renesas/vsp1/vsp1_drv.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
drivers/media/platform/renesas/vsp1/vsp1_drv.c +6 -2
index 2de515c497eb..627b5046fa80 100644
@@ -240,8 +240,12 @@ static void vsp1_destroy_entities(struct vsp1_device *vsp1)
 		media_device_unregister(&vsp1->media_dev);
 	media_device_cleanup(&vsp1->media_dev);
 
-	if (!vsp1->info->uapi)
-		vsp1_drm_cleanup(vsp1);
+	if (!vsp1->info->uapi) {
+		if (vsp1->info->version == VI6_IP_VERSION_MODEL_VSPX_GEN4)
+			vsp1_vspx_cleanup(vsp1);
+		else
+			vsp1_drm_cleanup(vsp1);
+	}
 }
 
 static int vsp1_create_entities(struct vsp1_device *vsp1)
Signature · 1 line
2.54.0
author maintainer
#11 lore raw

PATCH 11/188 media: renesas: vin: Fix RAW8 (again)

In reply to Greg Kroah-Hartman

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

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

From: Tomi Valkeinen <[email protected]>

commit 40c6da8a9c0f897f99a439330584d93ca7d41226 upstream.

Commit e7376745ad5c ("media: rcar-vin: Fix stride setting for RAW8 formats") removed dividing the stride by two for RAW8 formats. It is unclear how this was tested, but in any of the recent tests this does not seem to work and produces quite distorted images.

However, reverting the patch fixes the issues only partially. VNIS_REG requires alignment to 16 bytes, and when dividing the stride by 2, in some cases we end up with a non-aligned stride, producing a tilted image. This issue has to be fixed in rvin_format_bytesperline() where we do the alignment for bytesperline.

Adding back the stride division and increasing the alignment for RAW8 formats to 0x20 fixes the problems related to RAW8.

Fixes e7376745ad5c ("media: rcar-vin: Fix stride setting for RAW8 formats")
---
 drivers/media/platform/renesas/rcar-vin/rcar-dma.c  |   22 ++++++++++++++++++++
 drivers/media/platform/renesas/rcar-vin/rcar-v4l2.c |   12 ++++++++++
 2 files changed, 34 insertions(+)
drivers/media/platform/renesas/rcar-vin/rcar-dma.c +22 -0
@@ -676,8 +676,30 @@ void rvin_crop_scale_comp(struct rvin_de
 	if (vin->scaler)
 		vin->scaler(vin);
 
+	/*
+	 * VNIS_REG has four lowest bits always 0, i.e. the stride has to be
+	 * aligned to 16 bytes. This is done in rvin_format_bytesperline().
+	 */
+
 	fmt = rvin_format_from_pixel(vin, vin->format.pixelformat);
 	stride = vin->format.bytesperline / fmt->bpp;
+
+	/*
+	 * RAW8 format bpp is 1, but the hardware process RAW8 format in 2 pixel
+	 * units, so we need to divide the stride by 2.
+	 */
+	switch (vin->format.pixelformat) {
+	case V4L2_PIX_FMT_SBGGR8:
+	case V4L2_PIX_FMT_SGBRG8:
+	case V4L2_PIX_FMT_SGRBG8:
+	case V4L2_PIX_FMT_SRGGB8:
+	case V4L2_PIX_FMT_GREY:
+		stride /= 2;
+		break;
+	default:
+		break;
+	}
+
 	rvin_write(vin, stride, VNIS_REG);
 }
 
drivers/media/platform/renesas/rcar-vin/rcar-v4l2.c +12 -0
@@ -155,6 +155,18 @@ static u32 rvin_format_bytesperline(stru
 	case V4L2_PIX_FMT_NV16:
 		align = 0x20;
 		break;
+	case V4L2_PIX_FMT_SBGGR8:
+	case V4L2_PIX_FMT_SGBRG8:
+	case V4L2_PIX_FMT_SGRBG8:
+	case V4L2_PIX_FMT_SRGGB8:
+	case V4L2_PIX_FMT_GREY:
+		/*
+		 * RAW8 format bpp is 1, but the hardware process RAW8 format in
+		 * 2 pixel units, and we need to align to 32 bytes. See
+		 * rvin_crop_scale_comp().
+		 */
+		align = 0x20;
+		break;
 	default:
 		align = 0x10;
 		break;
author maintainer
#12 lore raw

PATCH 12/188 media: i2c: ov8856: free control handler on error in ov8856_init_controls()

In reply to Greg Kroah-Hartman

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

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

From: Alexander Koskovich <[email protected]>

commit f75e160745663ce9b13362ae6e90bd439c58df69 upstream.

The control handler wasn't freed if adding controls failed, add an error exit label and convert the existing error return to use it.

Fixes 879347f0c258 ("media: ov8856: Add support for OV8856 sensor")
---
 drivers/media/i2c/ov8856.c |   10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/media/i2c/ov8856.c +8 -2
@@ -1951,12 +1951,18 @@ static int ov8856_init_controls(struct o
 			  V4L2_CID_HFLIP, 0, 1, 1, 0);
 	v4l2_ctrl_new_std(ctrl_hdlr, &ov8856_ctrl_ops,
 			  V4L2_CID_VFLIP, 0, 1, 1, 0);
-	if (ctrl_hdlr->error)
-		return ctrl_hdlr->error;
+	if (ctrl_hdlr->error) {
+		ret = ctrl_hdlr->error;
+		goto err_ctrl_handler_free;
+	}
 
 	ov8856->sd.ctrl_handler = ctrl_hdlr;
 
 	return 0;
+
+err_ctrl_handler_free:
+	v4l2_ctrl_handler_free(ctrl_hdlr);
+	return ret;
 }
 
 static void ov8856_update_pad_format(struct ov8856 *ov8856,
author maintainer
#13 lore raw

PATCH 13/188 media: dt-bindings: rockchip,vdec: Add alternative reg-names order for RK35{76,88}

In reply to Greg Kroah-Hartman

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

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

From: Cristian Ciocaltea <[email protected]>

commit 35c8178ed2bd9821a75a406d762b2f2e161f9c70 upstream.

With the introduction of the RK3588 SoC, and RK3576 afterwards, three register blocks have been provided for the video decoder unit instead of just one, which are further referenced in vendor's datasheet by 'link table', 'function' and 'cache'. The former is present at the top of the listing, starting at video decoder unit base address.

However, while documenting RK3588, the binding broke the convention expecting the unit address to indicate the start of the primary register range, i.e. the 'function' block got listed before the 'link' one.

Since the binding changes have been already released and a fix would bring up an ABI break, mark the current 'reg-names' ordering as deprecated and introduce an alternative 'link,function,cache' listing which follows the address-based ordering according to the TRM.

Additionally, drop the 'reg' description items as the order is not fixed anymore, while the information they offer is not very relevant anyway.

It's worth noting there are currently no (known) users impacted by these binding changes, since the video decoder support for the aforementioned SoCs in mainline driver and devicetrees hasn't been released yet - it landed in v7.0-rc1 while all DTS updates resulting from this will be handled before v7.0 is out.

Fixes c6ffb7e1fb90 ("media: dt-bindings: rockchip: Document RK3588 Video Decoder bindings")
Fixes a5c4a6526476 ("media: dt-bindings: rockchip: Add RK3576 Video Decoder bindings")
---
 Documentation/devicetree/bindings/media/rockchip,vdec.yaml |   20 +++++++------
 1 file changed, 12 insertions(+), 8 deletions(-)
Documentation/devicetree/bindings/media/rockchip,vdec.yaml +12 -8
@@ -27,16 +27,20 @@ properties:
 
   reg:
     minItems: 1
-    items:
-      - description: The function configuration registers base
-      - description: The link table configuration registers base
-      - description: The cache configuration registers base
+    maxItems: 3
 
   reg-names:
-    items:
-      - const: function
-      - const: link
-      - const: cache
+    oneOf:
+      - items:
+          - const: link
+          - const: function
+          - const: cache
+      - items:
+          - const: function
+          - const: link
+          - const: cache
+        deprecated: true
+        description: Use link,function,cache block order instead.
 
   interrupts:
     maxItems: 1
author maintainer
#14 lore raw

PATCH 14/188 media: dt-bindings: rockchip,vdec: Mark reg-names required for RK35{76,88}

In reply to Greg Kroah-Hartman

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

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

From: Cristian Ciocaltea <[email protected]>

commit a11db8d8b403eba1f82728f440727128e9997edd upstream.

The Rockchip Video Decoder driver expects reg-names to be mandatory for RK3576 and RK3588 SoCs, however the binding does not currently require the use of them.

As a consequence, driver would fail to probe with a hypothetical devicetree that doesn't provide the reg-names for these SoCs, but which is otherwise a perfectly valid DT from the binding perspective.

Update the binding and make reg-names required for the aforementioned SoCs. While this change introduces an ABI break, the expected impact on potential users would be minimal, if any, since the old SoCs are unaffected, while the video decoder support for these newer variants in mainline driver and devicetrees hasn't been released yet.

Moreover, this is also a prerequisite for a subsequent binding update introducing an alternative reg-names order, according to the address-based listing in the vendor's datasheet.

Fixes c6ffb7e1fb90 ("media: dt-bindings: rockchip: Document RK3588 Video Decoder bindings")
Fixes a5c4a6526476 ("media: dt-bindings: rockchip: Add RK3576 Video Decoder bindings")
---
 Documentation/devicetree/bindings/media/rockchip,vdec.yaml |    2 ++
 1 file changed, 2 insertions(+)
Documentation/devicetree/bindings/media/rockchip,vdec.yaml +2 -0
@@ -126,6 +126,8 @@ allOf:
           minItems: 5
         reset-names:
           minItems: 5
+      required:
+        - reg-names
     else:
       properties:
         reg:
author maintainer
#15 lore raw

PATCH 15/188 media: chips-media: wave5: fix a potential memory leak in wave5_vdi_init()

In reply to Greg Kroah-Hartman

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

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

From: Haoxiang Li <[email protected]>

commit 95bd174a453f77b09ea66e1e22834680754ba501 upstream.

Add wave5_vdi_free_dma_memory() in the error path of wave5_vdi_init() to prevent a potential memory leak.

Fixes 45d1a2b93277 ("media: chips-media: wave5: Add vpuapi layer")
---
 drivers/media/platform/chips-media/wave5/wave5-vdi.c |    1 +
 1 file changed, 1 insertion(+)
drivers/media/platform/chips-media/wave5/wave5-vdi.c +1 -0
@@ -49,6 +49,7 @@ int wave5_vdi_init(struct device *dev)
 
 	if (!PRODUCT_CODE_W_SERIES(vpu_dev->product_code)) {
 		WARN_ONCE(1, "unsupported product code: 0x%x\n", vpu_dev->product_code);
+		wave5_vdi_free_dma_memory(vpu_dev, &vpu_dev->common_mem);
 		return -EOPNOTSUPP;
 	}
 
author maintainer
#16 lore raw

PATCH 16/188 media: chips-media: wave5: add missing spinlock protection for send_eos_event()

In reply to Greg Kroah-Hartman

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

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

From: Ziyi Guo <[email protected]>

commit f48050436746be75227fbc90066a8658cbe94d17 upstream.

Add spin_lock_irqsave()/spin_unlock_irqrestore() around send_eos_event() calls in the VB2 buffer queue and streamoff callbacks to fix the missing lock protection.

wave5_vpu_dec_buf_queue_dst() and streamoff_output() call send_eos_event() without holding inst->state_spinlock. However, send_eos_event() has lockdep_assert_held(&inst->state_spinlock) indicating that callers must hold this lock.

Other callers of send_eos_event() properly acquire the spinlock:
- wave5_vpu_dec_finish_decode() acquires lock at line 431
- wave5_vpu_dec_encoder_cmd() acquires lock at line 821
- wave5_vpu_dec_device_run() acquires lock at line 1592

Fixes 9707a6254a8a6b ("media: chips-media: wave5: Add the v4l2 layer")
---
 drivers/media/platform/chips-media/wave5/wave5-vpu-dec.c |   11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)
drivers/media/platform/chips-media/wave5/wave5-vpu-dec.c +10 -1
@@ -1265,13 +1265,17 @@ static void wave5_vpu_dec_buf_queue_dst(
 
 	if (vb2_is_streaming(vb->vb2_queue) && v4l2_m2m_dst_buf_is_last(m2m_ctx)) {
 		unsigned int i;
+		unsigned long flags;
 
 		for (i = 0; i < vb->num_planes; i++)
 			vb2_set_plane_payload(vb, i, 0);
 
 		vbuf->field = V4L2_FIELD_NONE;
 
+		spin_lock_irqsave(&inst->state_spinlock, flags);
 		send_eos_event(inst);
+		spin_unlock_irqrestore(&inst->state_spinlock, flags);
+
 		v4l2_m2m_last_buffer_done(m2m_ctx, vbuf);
 	} else {
 		v4l2_m2m_buf_queue(m2m_ctx, vbuf);
@@ -1415,8 +1419,13 @@ static int streamoff_output(struct vb2_q
 	inst->codec_info->dec_info.stream_rd_ptr = new_rd_ptr;
 	inst->codec_info->dec_info.stream_wr_ptr = new_rd_ptr;
 
-	if (v4l2_m2m_has_stopped(m2m_ctx))
+	if (v4l2_m2m_has_stopped(m2m_ctx)) {
+		unsigned long flags;
+
+		spin_lock_irqsave(&inst->state_spinlock, flags);
 		send_eos_event(inst);
+		spin_unlock_irqrestore(&inst->state_spinlock, flags);
+	}
 
 	/* streamoff on output cancels any draining operation */
 	inst->eos = false;
author maintainer
#17 lore raw

PATCH 17/188 media: chips-media: wave5: add missing spinlock protection for handle_dynamic_resolution_change()

In reply to Greg Kroah-Hartman

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

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

From: Ziyi Guo <[email protected]>

commit cb8bdd3ffca280d014311ab395651d33f58a8708 upstream.

Add spin_lock_irqsave()/spin_unlock_irqrestore() around the handle_dynamic_resolution_change() call in initialize_sequence() to fix the missing lock protection.

initialize_sequence() calls handle_dynamic_resolution_change() without holding inst->state_spinlock. However, handle_dynamic_resolution_change() has lockdep_assert_held(&inst->state_spinlock) indicating that callers must hold this lock.

Other callers of handle_dynamic_resolution_change() properly acquire the spinlock:
- wave5_vpu_dec_finish_decode()
- wave5_vpu_dec_device_run()

Fixes 9707a6254a8a6b ("media: chips-media: wave5: Add the v4l2 layer")
---
 drivers/media/platform/chips-media/wave5/wave5-vpu-dec.c |    3 +++
 1 file changed, 3 insertions(+)
drivers/media/platform/chips-media/wave5/wave5-vpu-dec.c +3 -0
@@ -1544,6 +1544,7 @@ static int initialize_sequence(struct vp
 {
 	struct dec_initial_info initial_info;
 	int ret = 0;
+	unsigned long flags;
 
 	memset(&initial_info, 0, sizeof(struct dec_initial_info));
 
@@ -1565,7 +1566,9 @@ static int initialize_sequence(struct vp
 		return ret;
 	}
 
+	spin_lock_irqsave(&inst->state_spinlock, flags);
 	handle_dynamic_resolution_change(inst);
+	spin_unlock_irqrestore(&inst->state_spinlock, flags);
 
 	return 0;
 }
author maintainer
#18 lore raw

PATCH 18/188 drm/gpusvm: Force unmapping on error in drm_gpusvm_get_pages

In reply to Greg Kroah-Hartman

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

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

From: Matthew Brost <[email protected]>

commit 556dba95473900073a6c03121361c11f646dc551 upstream.

drm_gpusvm_get_pages() only sets the local flags prior to committing the pages. If an error occurs mid-mapping, has_dma_mapping will be clear, causing the unmap function to skip unmapping pages that were successfully mapped before the error. Fix this by forcibly setting has_dma_mapping in the error path to ensure all previously mapped pages are properly unmapped.

Fixes 99624bdff867 ("drm/gpusvm: Add support for GPU Shared Virtual Memory")
---
 drivers/gpu/drm/drm_gpusvm.c |    1 +
 1 file changed, 1 insertion(+)
drivers/gpu/drm/drm_gpusvm.c +1 -0
@@ -1443,6 +1443,7 @@ set_seqno:
 	return 0;
 
 err_unmap:
+	svm_pages->flags.has_dma_mapping = true;
 	__drm_gpusvm_unmap_pages(gpusvm, svm_pages, num_dma_mapped);
 	drm_gpusvm_notifier_unlock(gpusvm);
 err_free:
author maintainer
#19 lore raw

PATCH 19/188 spi: bcm63xx: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit c39e65a4e3b8e764efed0b2f5152a1a8547b80fd upstream.

Make sure to deregister the controller before disabling underlying resources like clocks during driver unbind.

Fixes b42dfed83d95 ("spi: add Broadcom BCM63xx SPI controller driver")
Cc Florian Fainelli <[email protected]>
---
 drivers/spi/spi-bcm63xx.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/spi/spi-bcm63xx.c +7 -1
@@ -603,7 +603,7 @@ static int bcm63xx_spi_probe(struct plat
 		goto out_clk_disable;
 
 	/* register and we are done */
-	ret = devm_spi_register_controller(dev, host);
+	ret = spi_register_controller(host);
 	if (ret) {
 		dev_err(dev, "spi register failed\n");
 		goto out_clk_disable;
@@ -626,11 +626,17 @@ static void bcm63xx_spi_remove(struct pl
 	struct spi_controller *host = platform_get_drvdata(pdev);
 	struct bcm63xx_spi *bs = spi_controller_get_devdata(host);
 
+	spi_controller_get(host);
+
+	spi_unregister_controller(host);
+
 	/* reset spi block */
 	bcm_spi_writeb(bs, 0, SPI_INT_MASK);
 
 	/* HW shutdown */
 	clk_disable_unprepare(bs->clk);
+
+	spi_controller_put(host);
 }
 
 static int bcm63xx_spi_suspend(struct device *dev)
author maintainer
#20 lore raw

PATCH 20/188 spi: atmel: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 8d4de97e83520be89d0ff40610ca633b3963a7de upstream.

Make sure to deregister the controller before disabling underlying resources like clocks during driver unbind.

---
 drivers/spi/spi-atmel.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/spi/spi-atmel.c +7 -1
@@ -1655,7 +1655,7 @@ static int atmel_spi_probe(struct platfo
 	pm_runtime_set_active(&pdev->dev);
 	pm_runtime_enable(&pdev->dev);
 
-	ret = devm_spi_register_controller(&pdev->dev, host);
+	ret = spi_register_controller(host);
 	if (ret)
 		goto out_free_dma;
 
@@ -1689,8 +1689,12 @@ static void atmel_spi_remove(struct plat
 	struct spi_controller	*host = platform_get_drvdata(pdev);
 	struct atmel_spi	*as = spi_controller_get_devdata(host);
 
+	spi_controller_get(host);
+
 	pm_runtime_get_sync(&pdev->dev);
 
+	spi_unregister_controller(host);
+
 	/* reset the hardware and block queue progress */
 	if (as->use_dma) {
 		atmel_spi_stop_dma(host);
@@ -1717,6 +1721,8 @@ static void atmel_spi_remove(struct plat
 
 	pm_runtime_put_noidle(&pdev->dev);
 	pm_runtime_disable(&pdev->dev);
+
+	spi_controller_put(host);
 }
 
 static int atmel_spi_runtime_suspend(struct device *dev)
author maintainer
#21 lore raw

PATCH 21/188 arm64: dts: lx2160a-cex7/lx2162a-sr-som: fix usd-cd & gpio pinmux

In reply to Greg Kroah-Hartman

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

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

From: Josua Mayer <[email protected]>

commit 70008aee892bbb5c2969bbe9e5778fc081b14bd2 upstream.

Commit 8a1365c7bbc1 ("arm64: dts: lx2160a: add pinmux and i2c gpio to support bus recovery") introduced pinmux nodes for lx2160 i2c interfaces, allowing runtime change between i2c and gpio functions implementing bus recovery.

However, the dynamic configuration area (overwrite MUX) used by the pinctrl-single driver initially reads as zero and does not reflect the actual hardware state set by the Reset Configuration Word (RCW) at power-on.

Because multiple groups of pins are configured from a single 32-bit register, the first write from the pinctrl driver unintentionally clears all other bits to zero.

For example, on the LX2162A Clearfog, RCWSR12 is initialized to 0x08000006. When any i2c pinmux is applied, it clears all other fields. This inadvertently disables SD card-detect (IIC2_PMUX) and some GPIOs (SDHC1_DIR_PMUX):

LX2162-CF RCWSR12: 0b0000100000000000 0000000000000110

IIC2_PMUX              |||   |||   || |   |||   |||XXX : I2C/GPIO/CD-WP
SDHC1_DIR_PMUX         XXX   |||   || |   |||   |||    : SDHC/GPIO/SPI

Reverting the commit in question was considered but bus recovery is an important feature.

Instead add pinmux nodes for those pins that were unintentionally reconfigured on SolidRun LX2160A Clearfog-CX and LX2162A Clearfog boards.

Fixes 8a1365c7bbc1 ("arm64: dts: lx2160a: add pinmux and i2c gpio to support bus recovery")
---
 arch/arm64/boot/dts/freescale/fsl-lx2160a-cex7.dtsi         |    7 +++
 arch/arm64/boot/dts/freescale/fsl-lx2160a-clearfog-itx.dtsi |    2 +
 arch/arm64/boot/dts/freescale/fsl-lx2160a.dtsi              |   24 ++++++++++++
 arch/arm64/boot/dts/freescale/fsl-lx2162a-clearfog.dts      |    2 +
 arch/arm64/boot/dts/freescale/fsl-lx2162a-sr-som.dtsi       |    7 +++
 5 files changed, 42 insertions(+)
arch/arm64/boot/dts/freescale/fsl-lx2160a-cex7.dtsi +7 -0
@@ -162,6 +162,8 @@
 };
 
 &fspi {
+	pinctrl-names = "default";
+	pinctrl-0 = <&fspi_data74_pins>, <&fspi_data30_pins>, <&fspi_dqs_sck_cs10_pins>;
 	status = "okay";
 
 	flash@0 {
@@ -177,6 +179,11 @@
 	};
 };
 
+&pinmux_i2crv {
+	pinctrl-names = "default";
+	pinctrl-0 = <&gpio0_14_12_pins>;
+};
+
 &usb0 {
 	status = "okay";
 };
arch/arm64/boot/dts/freescale/fsl-lx2160a-clearfog-itx.dtsi +2 -0
@@ -89,6 +89,8 @@
 };
 
 &esdhc0 {
+	pinctrl-names = "default";
+	pinctrl-0 = <&esdhc0_cd_wp_pins>, <&esdhc0_cmd_data30_clk_vsel_pins>;
 	sd-uhs-sdr104;
 	sd-uhs-sdr50;
 	sd-uhs-sdr25;
arch/arm64/boot/dts/freescale/fsl-lx2160a.dtsi +24 -0
@@ -1717,6 +1717,10 @@
 				pinctrl-single,bits = <0x0 0x1 0x7>;
 			};
 
+			esdhc0_cd_wp_pins: iic2-sdhc-pins {
+				pinctrl-single,bits = <0x0 0x6 0x7>;
+			};
+
 			i2c2_scl: i2c2-scl-pins {
 				pinctrl-single,bits = <0x0 0 (0x7 << 3)>;
 			};
@@ -1749,6 +1753,26 @@
 				pinctrl-single,bits = <0x0 (0x1 << 12) (0x7 << 12)>;
 			};
 
+			fspi_data74_pins: xspi1-data74-pins {
+				pinctrl-single,bits = <0x0 0x0 (0x7 << 15)>;
+			};
+
+			fspi_data30_pins: xspi1-data30-pins {
+				pinctrl-single,bits = <0x0 0x0 (0x7 << 18)>;
+			};
+
+			fspi_dqs_sck_cs10_pins: xspi1-base-pins {
+				pinctrl-single,bits = <0x0 0x0 (0x7 << 21)>;
+			};
+
+			esdhc0_cmd_data30_clk_vsel_pins: sdhc1-base-sdhc-vsel-pins {
+				pinctrl-single,bits = <0x0 0x0 (0x7 << 24)>;
+			};
+
+			gpio0_14_12_pins: sdhc1-dir-gpio-pins {
+				pinctrl-single,bits = <0x0 (0x1 << 27) (0x7 << 27)>;
+			};
+
 			i2c6_scl: i2c6-scl-pins {
 				pinctrl-single,bits = <0x4 0x2 0x7>;
 			};
arch/arm64/boot/dts/freescale/fsl-lx2162a-clearfog.dts +2 -0
@@ -223,6 +223,8 @@
 };
 
 &esdhc0 {
+	pinctrl-names = "default";
+	pinctrl-0 = <&esdhc0_cd_wp_pins>, <&esdhc0_cmd_data30_clk_vsel_pins>;
 	sd-uhs-sdr104;
 	sd-uhs-sdr50;
 	sd-uhs-sdr25;
arch/arm64/boot/dts/freescale/fsl-lx2162a-sr-som.dtsi +7 -0
@@ -30,6 +30,8 @@
 };
 
 &fspi {
+	pinctrl-names = "default";
+	pinctrl-0 = <&fspi_data74_pins>, <&fspi_data30_pins>, <&fspi_dqs_sck_cs10_pins>;
 	status = "okay";
 
 	flash@0 {
@@ -80,3 +82,8 @@
 		reg = <0x6f>;
 	};
 };
+
+&pinmux_i2crv {
+	pinctrl-names = "default";
+	pinctrl-0 = <&gpio0_14_12_pins>;
+};
author maintainer
#22 lore raw

PATCH 22/188 staging: media: atomisp: Disallow all private IOCTLs

In reply to Greg Kroah-Hartman

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

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

From: Sakari Ailus <[email protected]>

commit 2b7eb2c5dc72f0fc954ac4aa155f9e285e937f7c upstream.

Disallow all private IOCTLs. These aren't quite as safe as one could assume of IOCTL handlers; disable them for now. Instead of removing the code, return in the beginning of the function if cmd is non-zero in order to keep static checkers happy.

Fixes a49d25364dfb ("staging/atomisp: Add support for the Intel IPU v2")
Fixes ad85094b293e ("Revert "media: staging: atomisp: Remove driver"")
---
 drivers/staging/media/atomisp/pci/atomisp_ioctl.c |    4 ++++
 1 file changed, 4 insertions(+)
drivers/staging/media/atomisp/pci/atomisp_ioctl.c +4 -0
@@ -1357,6 +1357,10 @@ static int atomisp_s_parm(struct file *f
 static long atomisp_vidioc_default(struct file *file, void *fh,
 				   bool valid_prio, unsigned int cmd, void *arg)
 {
+	/* Disable all private IOCTLs for now! */
+	if (cmd)
+		return -EINVAL;
+
 	struct video_device *vdev = video_devdata(file);
 	struct atomisp_sub_device *asd = atomisp_to_video_pipe(vdev)->asd;
 	int err;
author maintainer
#23 lore raw

PATCH 23/188 regulator: mt6357: fix OF node reference imbalance

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 2f38e96c273e15f5e9f5d1fc2c0cbba703751602 upstream.

The driver reuses the OF node of the parent multi-function device but fails to take another reference to balance the one dropped by the platform bus code when unbinding the MFD and deregistering the child devices.

Fix this by using the intended helper for reusing OF nodes.

---
 drivers/regulator/mt6357-regulator.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/regulator/mt6357-regulator.c +1 -1
@@ -410,7 +410,7 @@ static int mt6357_regulator_probe(struct
 	struct regulator_dev *rdev;
 	int i;
 
-	pdev->dev.of_node = pdev->dev.parent->of_node;
+	device_set_of_node_from_dev(&pdev->dev, pdev->dev.parent);
 
 	for (i = 0; i < MT6357_MAX_REGULATOR; i++) {
 		config.dev = &pdev->dev;
author maintainer
#24 lore raw

PATCH 24/188 spi: st-ssc4: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 19857374010d06ca6a2f7c2c53464122eb804df0 upstream.

Make sure to deregister the controller before disabling underlying resources like clocks during driver unbind.

Fixes 9e862375c542 ("spi: Add new driver for STMicroelectronics' SPI Controller")
Cc Lee Jones <[email protected]>
---
 drivers/spi/spi-st-ssc4.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/spi/spi-st-ssc4.c +7 -1
@@ -349,7 +349,7 @@ static int spi_st_probe(struct platform_
 
 	platform_set_drvdata(pdev, host);
 
-	ret = devm_spi_register_controller(&pdev->dev, host);
+	ret = spi_register_controller(host);
 	if (ret) {
 		dev_err(&pdev->dev, "Failed to register host\n");
 		goto rpm_disable;
@@ -371,10 +371,16 @@ static void spi_st_remove(struct platfor
 	struct spi_controller *host = platform_get_drvdata(pdev);
 	struct spi_st *spi_st = spi_controller_get_devdata(host);
 
+	spi_controller_get(host);
+
+	spi_unregister_controller(host);
+
 	pm_runtime_disable(&pdev->dev);
 
 	clk_disable_unprepare(spi_st->clk);
 
+	spi_controller_put(host);
+
 	pinctrl_pm_select_sleep_state(&pdev->dev);
 }
 
author maintainer
#25 lore raw

PATCH 25/188 regulator: max77650: fix OF node reference imbalance

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 2edaf5f7ada0ab5c9ec1f0836bd19779a8d85262 upstream.

The driver reuses the OF node of the parent multi-function device but fails to take another reference to balance the one dropped by the platform bus code when unbinding the MFD and deregistering the child devices.

Fix this by using the intended helper for reusing OF nodes.

---
 drivers/regulator/max77650-regulator.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/regulator/max77650-regulator.c +1 -1
@@ -337,7 +337,7 @@ static int max77650_regulator_probe(stru
 	parent = dev->parent;
 
 	if (!dev->of_node)
-		dev->of_node = parent->of_node;
+		device_set_of_node_from_dev(dev, parent);
 
 	rdescs = devm_kcalloc(dev, MAX77650_REGULATOR_NUM_REGULATORS,
 			      sizeof(*rdescs), GFP_KERNEL);
author maintainer
#26 lore raw

PATCH 26/188 media: rc: xbox_remote: heed DMA restrictions

In reply to Greg Kroah-Hartman

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

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

From: Oliver Neukum <[email protected]>

commit e280d1e5e3f2595bbb43fe6e1bce00c59a43c0ff upstream.

The buffer for IO must not be part of the device structure because that violates the DMA coherency rules.

Fixes 02d32bdad3123 ("media: rc: add driver for Xbox DVD Movie Playback Kit")
---
 drivers/media/rc/xbox_remote.c |    9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
drivers/media/rc/xbox_remote.c +8 -1
@@ -55,7 +55,7 @@ struct xbox_remote {
 	struct usb_interface *interface;
 
 	struct urb *irq_urb;
-	unsigned char inbuf[DATA_BUFSIZE] __aligned(sizeof(u16));
+	u8 *inbuf;
 
 	char rc_name[NAME_BUFSIZE];
 	char rc_phys[NAME_BUFSIZE];
@@ -218,6 +218,10 @@ static int xbox_remote_probe(struct usb_
 	if (!xbox_remote || !rc_dev)
 		goto exit_free_dev_rdev;
 
+	xbox_remote->inbuf = kzalloc(DATA_BUFSIZE, GFP_KERNEL);
+	if (!xbox_remote->inbuf)
+		goto exit_free_inbuf;
+
 	/* Allocate URB buffer */
 	xbox_remote->irq_urb = usb_alloc_urb(0, GFP_KERNEL);
 	if (!xbox_remote->irq_urb)
@@ -262,6 +266,8 @@ exit_kill_urbs:
 	usb_kill_urb(xbox_remote->irq_urb);
 exit_free_buffers:
 	usb_free_urb(xbox_remote->irq_urb);
+exit_free_inbuf:
+	kfree(xbox_remote->inbuf);
 exit_free_dev_rdev:
 	rc_free_device(rc_dev);
 	kfree(xbox_remote);
@@ -286,6 +292,7 @@ static void xbox_remote_disconnect(struc
 	usb_kill_urb(xbox_remote->irq_urb);
 	rc_unregister_device(xbox_remote->rdev);
 	usb_free_urb(xbox_remote->irq_urb);
+	kfree(xbox_remote->inbuf);
 	kfree(xbox_remote);
 }
 
author maintainer
#27 lore raw

PATCH 27/188 media: rc: streamzap: Error handling in probe

In reply to Greg Kroah-Hartman

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

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

From: Oliver Neukum <[email protected]>

commit 42844992664f03ef9f930e64f7370fa481e9c267 upstream.

If submitting the URB fails, the device will be unusable. Probe() must fail.

Fixes 7a569f524dd36 ("V4L/DVB: IR/streamzap: functional in-kernel decoding")
---
 drivers/media/rc/streamzap.c |   12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)
drivers/media/rc/streamzap.c +8 -4
@@ -219,9 +219,8 @@ static void streamzap_callback(struct ur
 	case -ESHUTDOWN:
 		/*
 		 * this urb is terminated, clean up.
-		 * sz might already be invalid at this point
 		 */
-		dev_err(sz->dev, "urb terminated, status: %d\n", urb->status);
+		dev_dbg(sz->dev, "urb terminated, status: %d\n", urb->status);
 		return;
 	default:
 		break;
@@ -358,11 +357,16 @@ static int streamzap_probe(struct usb_in
 
 	usb_set_intfdata(intf, sz);
 
-	if (usb_submit_urb(sz->urb_in, GFP_ATOMIC))
+	retval = usb_submit_urb(sz->urb_in, GFP_ATOMIC);
+	if (retval < 0) {
 		dev_err(sz->dev, "urb submit failed\n");
+		goto rc_submit_fail;
+	}
 
 	return 0;
-
+rc_submit_fail:
+	rc_free_device(sz->rdev);
+	usb_set_intfdata(intf, NULL);
 rc_dev_fail:
 	usb_free_urb(sz->urb_in);
 free_buf_in:
author maintainer
#28 lore raw

PATCH 28/188 media: i2c: imx283: Enter full standby when stopping streaming

In reply to Greg Kroah-Hartman

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

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

From: Jai Luthra <[email protected]>

commit bce1349dbf6348ddee47308e2ed08878356de317 upstream.

Use IMX283_STANDBY (bit 0) instead of IMX283_STBLOGIC (bit 1) when stopping streaming. STBLOGIC only puts the sensor logic into standby but leaves the MIPI interface (along with other components) in an indeterminate state.

This (presumably) causes the CSI receiver (e.g. Raspberry Pi's CFE) to miss the LP-11 to HS transition when streaming restarts, resulting in a hang of 10+ seconds. The issue is most visible when immediately restarting a full-resolution stream after stopping a 3x3 binned one, so that runtime suspend hasn't yet been triggered.

Writing IMX283_STANDBY puts the entire sensor into standby. The imx283_standby_cancel() sequence already handles the full wakeup from this suspended state.

---
 drivers/media/i2c/imx283.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/media/i2c/imx283.c +1 -1
@@ -1158,7 +1158,7 @@ static int imx283_disable_streams(struct
 	if (pad != IMAGE_PAD)
 		return -EINVAL;
 
-	ret = cci_write(imx283->cci, IMX283_REG_STANDBY, IMX283_STBLOGIC, NULL);
+	ret = cci_write(imx283->cci, IMX283_REG_STANDBY, IMX283_STANDBY, NULL);
 	if (ret)
 		dev_err(imx283->dev, "Failed to stop stream\n");
 
author maintainer
#29 lore raw

PATCH 29/188 regulator: bq257xx: fix OF node reference imbalance

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 7ea07bc030d8d6395524dec22ff3267441a28c0d upstream.

The driver reuses the OF node of the parent multi-function device but fails to take another reference to balance the one dropped by the platform bus code when unbinding the MFD and deregistering the child devices.

Fix this by using the intended helper for reusing OF nodes.

Fixes 981dd162b635 ("regulator: bq257xx: Add bq257xx boost regulator driver")
Cc Chris Morgan <[email protected]>
---
 drivers/regulator/bq257xx-regulator.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
drivers/regulator/bq257xx-regulator.c +1 -2
@@ -142,8 +142,7 @@ static int bq257xx_regulator_probe(struc
 	struct device_node *np = dev->of_node;
 	struct regulator_config cfg = {};
 
-	pdev->dev.of_node = pdev->dev.parent->of_node;
-	pdev->dev.of_node_reused = true;
+	device_set_of_node_from_dev(&pdev->dev, pdev->dev.parent);
 
 	pdata = devm_kzalloc(&pdev->dev, sizeof(struct bq257xx_reg_data), GFP_KERNEL);
 	if (!pdata)
author maintainer
#30 lore raw

PATCH 30/188 regulator: rk808: fix OF node reference imbalance

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 65290b24d8a5f0b8cd065201e653db824c4a4da6 upstream.

The driver reuses the OF node of the parent multi-function device but fails to take another reference to balance the one dropped by the platform bus code when unbinding the MFD and deregistering the child devices.

Fix this by using the intended helper for reusing OF nodes.

---
 drivers/regulator/rk808-regulator.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
drivers/regulator/rk808-regulator.c +1 -2
@@ -1875,8 +1875,7 @@ static int rk808_regulator_probe(struct
 	struct regmap *regmap;
 	int ret, i, nregulators;
 
-	pdev->dev.of_node = pdev->dev.parent->of_node;
-	pdev->dev.of_node_reused = true;
+	device_set_of_node_from_dev(&pdev->dev, pdev->dev.parent);
 
 	regmap = dev_get_regmap(pdev->dev.parent, NULL);
 	if (!regmap)
author maintainer
#31 lore raw

PATCH 31/188 media: videobuf2: Set vma_flags in vb2_dma_sg_mmap

In reply to Greg Kroah-Hartman

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

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

From: Janne Grunau <[email protected]>

commit 7254b31a13aaa0c2c0f9ffbc335b718656117ff4 upstream.

vb2_dma_contig sets VMA flags VM_DONTEXPAND and VM_DONTDUMP and I do not see a reason why vb2_dma_sg should behave differently. This avoids hitting WARN_ON(!(vma->vm_flags & VM_DONTEXPAND)); in drm_gem_mmap_obj() during mmap() of an imported dma-buf from the out of tree Apple ISP camera capture driver which uses vb2_dma_sg_memops.

gst-launch-1.0 v4l2src ! gtk4paintablesink

[   38.201528] ------------[ cut here ]------------
[   38.202135] WARNING: CPU: 7 PID: 2362 at drivers/gpu/drm/drm_gem.c:1144 drm_gem_mmap_obj+0x1f8/0x210
[   38.203278] Modules linked in: rfcomm snd_seq_dummy snd_hrtimer

snd_seq snd_seq_device uinput nf_conntrack_netbios_ns nf_conntrack_broadcast nft_fib_inet nft_fib_ipv4 nft_fib_ipv6 nft_fib nft_reject_inet nf_reject_ipv6 nft_reject nft_ct nft_chain_nat nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 nf_tables qrtr bnep nls_ascii i2c_dev loop fuse dm_multipath nfnetlink brcmfmac_wcc hid_magicmouse hci_bcm4377 brcmfmac brcmutil bluetooth ecdh_generic cfg80211 ecc btrfs xor xor_neon rfkill hid_apple raid6_pq joydev aop_als apple_nvmem_spmi industrialio snd_soc_aop apple_z2 snd_soc_cs42l84 tps6598x snd_soc_tas2764 macsmc_reboot spi_nor macsmc_hwmon rtc_macsmc gpio_macsmc macsmc_power regmap_spmi macsmc_input dockchannel_hid panel_summit appledrm nvme_apple dwc3 snd_soc_macaudio drm_client_lib nvme_core phy_apple_atc hwmon apple_sart apple_dockchannel macsmc apple_rtkit_helper spmi_apple_controller aop apple_wdt mfd_core nvmem_apple_efuses pinctrl_apple_gpio apple_isp apple_dcp videobuf2_dma_sg mux_core spi_apple

[   38.203300]  videobuf2_memops i2c_pasemi_platform snd_soc_apple_mca videobuf2_v4l2 videodev clk_apple_nco videobuf2_common snd_pcm_dmaengine adpdrm asahi apple_admac adpdrm_mipi drm_dma_helper pwm_apple i2c_pasemi_core drm_display_helper mc cec apple_dart ofpart apple_soc_cpufreq leds_pwm phram
[   38.217677] CPU: 7 UID: 1000 PID: 2362 Comm: gst-launch-1.0 Tainted: G        W           6.17.6+ #asahi-dev PREEMPT(full)
[   38.219040] Tainted: [W]=WARN
[   38.219398] Hardware name: Apple MacBook Pro (13-inch, M2, 2022) (DT)
[   38.220213] pstate: 21400005 (nzCv daif +PAN -UAO -TCO +DIT -SSBS BTYPE=--)
[   38.221088] pc : drm_gem_mmap_obj+0x1f8/0x210
[   38.221643] lr : drm_gem_mmap_obj+0x78/0x210
[   38.222178] sp : ffffc0008dc678e0
[   38.222579] x29: ffffc0008dc678e0 x28: 0000000000042a97 x27: ffff8000b701b480
[   38.223465] x26: 00000000000000fb x25: ffffc0008dc67d20 x24: ffffc0008dc67968
[   38.224402] x23: ffff8000e3ca5600 x22: ffff8000265b7800 x21: ffff80003000c0c0
[   38.225279] x20: 0000000000000000 x19: ffff8000b68c5200 x18: ffffc0008dc67968
[   38.226151] x17: 0000000000000000 x16: 0000000000000000 x15: ffffc000810a30a8
[   38.227042] x14: 00007fff637effff x13: 00005555de91ffff x12: 00007fff63293fff
[   38.227942] x11: 0000000000000000 x10: ffff8000184ecf08 x9 : ffffc0007a1900c8
[   38.228824] x8 : ffffc0008dc67968 x7 : 0000000000000012 x6 : ffffc0015cf1c000
[   38.229703] x5 : ffffc0008dc676a0 x4 : ffffc00081a27dc0 x3 : 0000000000000038
[   38.230607] x2 : 0000000000000003 x1 : 0000000000000003 x0 : 00000000100000fb
[   38.231488] Call trace:
[   38.231806]  drm_gem_mmap_obj+0x1f8/0x210 (P)
[   38.232342]  drm_gem_mmap+0x140/0x260
[   38.232813]  __mmap_region+0x488/0x9a0
[   38.233277]  mmap_region+0xd0/0x148
[   38.233703]  do_mmap+0x350/0x5c0
[   38.234148]  vm_mmap_pgoff+0x14c/0x200
[   38.234612]  ksys_mmap_pgoff+0x150/0x208
[   38.235107]  __arm64_sys_mmap+0x34/0x50
[   38.235611]  invoke_syscall+0x50/0x120
[   38.236075]  el0_svc_common.constprop.0+0x48/0xf0
[   38.236680]  do_el0_svc+0x24/0x38
[   38.237113]  el0_svc+0x38/0x168
[   38.237507]  el0t_64_sync_handler+0xa0/0xe8
[   38.238034]  el0t_64_sync+0x198/0x1a0
[   38.238491] ---[ end trace 0000000000000000 ]---

There were discussions in [1] at the end of 2023 that mmap() on imported dma-bufs should not be supported but as of v6.17 drm_gem_shmem_mmap() in drm_gem_shmem_helper.c still supports it.
This might affect all gpu or accel drivers using drm_gem_shmem_mmap() or the wrapper drm_gem_shmem_object_mmap().

[1] https://lore.kernel.org/dri-devel/[email protected]/

Fixes 5ba3f757f059 ("[media] v4l: videobuf2: add DMA scatter/gather allocator")
---
 drivers/media/common/videobuf2/videobuf2-dma-sg.c |    1 +
 1 file changed, 1 insertion(+)
drivers/media/common/videobuf2/videobuf2-dma-sg.c +1 -0
@@ -345,6 +345,7 @@ static int vb2_dma_sg_mmap(void *buf_pri
 		return err;
 	}
 
+	vm_flags_set(vma, VM_DONTEXPAND | VM_DONTDUMP);
 	/*
 	 * Use common vm_area operations to track buffer refcount.
 	 */
author maintainer
#32 lore raw

PATCH 32/188 media: intel/ipu6: fix error pointer dereference

In reply to Greg Kroah-Hartman

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

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

From: Ethan Tidmore <[email protected]>

commit 8dd088b8b106f7b119664f965b691785998edcfb upstream.

In a error path isp->psys is confirmed to be an error pointer not NULL so this condition is true and the error pointer is dereferenced. So isp-psys should be set to NULL before going to out_ipu6_bus_del_devices.

Detected by Smatch:
drivers/media/pci/intel/ipu6/ipu6.c:690 ipu6_pci_probe() error:
'isp->psys' dereferencing possible ERR_PTR()

Fixes 25fedc021985a ("media: intel/ipu6: add Intel IPU6 PCI device driver")
[Sakari Ailus: Fix commit message.]
---
 drivers/media/pci/intel/ipu6/ipu6.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/media/pci/intel/ipu6/ipu6.c +1 -1
@@ -686,7 +686,7 @@ out_free_irq:
 out_ipu6_rpm_put:
 	pm_runtime_put_sync(&isp->psys->auxdev.dev);
 out_ipu6_bus_del_devices:
-	if (isp->psys) {
+	if (!IS_ERR_OR_NULL(isp->psys)) {
 		ipu6_cpd_free_pkg_dir(isp->psys);
 		ipu6_buttress_unmap_fw_image(isp->psys, &isp->psys->fw_sgt);
 	}
author maintainer
#33 lore raw

PATCH 33/188 media: i2c: imx283: Fix hang when going from large to small resolution

In reply to Greg Kroah-Hartman

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

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

From: Jai Luthra <[email protected]>

commit 9206359b2c396ff594adf39bc7daaadab0fcb367 upstream.

When switching between modes (e.g. full resolution to binned), standby_cancel() previously cleared XMSTA (starting master mode data output) before the new mode's MDSEL, crop, and timing registers were programmed in start_streaming(). This caused the sensor to briefly output MIPI data using the previous mode's configuration.

On receivers like imx-mipi-csis, this leads to FIFO overflow errors when switching from a higher to a lower resolution, as the receiver is configured for the new smaller frame size but receives stale full-resolution data.

Fix this by moving the XMSTA and SYNCDRV register writes from standby_cancel() to the end of start_streaming(), after all mode, crop, and timing registers have been configured. Also explicitly stop master mode (XMSTA=1) when stopping the stream, matching the pattern used by other Sony sensor drivers (imx290, imx415).

Use named macros IMX283_XMSTA_START/STOP instead of raw 0/BIT(0) for readability.

---
 drivers/media/i2c/imx283.c |   15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)
drivers/media/i2c/imx283.c +10 -5
@@ -129,7 +129,8 @@
 
 /* Master Mode Operation Control */
 #define IMX283_REG_XMSTA		CCI_REG8(0x3105)
-#define   IMX283_XMSTA			BIT(0)
+#define   IMX283_XMSTA_START		0
+#define   IMX283_XMSTA_STOP		BIT(0)
 
 #define IMX283_REG_SYNCDRV		CCI_REG8(0x3107)
 #define   IMX283_SYNCDRV_XHS_XVS	(0xa0 | 0x02)
@@ -1023,8 +1024,6 @@ static int imx283_standby_cancel(struct
 	usleep_range(19000, 20000);
 
 	cci_write(imx283->cci, IMX283_REG_CLAMP, IMX283_CLPSQRST, &ret);
-	cci_write(imx283->cci, IMX283_REG_XMSTA, 0, &ret);
-	cci_write(imx283->cci, IMX283_REG_SYNCDRV, IMX283_SYNCDRV_XHS_XVS, &ret);
 
 	return ret;
 }
@@ -1117,6 +1116,10 @@ static int imx283_start_streaming(struct
 	/* Apply customized values from controls (HMAX/VMAX/SHR) */
 	ret =  __v4l2_ctrl_handler_setup(imx283->sd.ctrl_handler);
 
+	/* Start master mode */
+	cci_write(imx283->cci, IMX283_REG_XMSTA, IMX283_XMSTA_START, &ret);
+	cci_write(imx283->cci, IMX283_REG_SYNCDRV, IMX283_SYNCDRV_XHS_XVS, &ret);
+
 	return ret;
 }
 
@@ -1153,12 +1156,14 @@ static int imx283_disable_streams(struct
 				  u64 streams_mask)
 {
 	struct imx283 *imx283 = to_imx283(sd);
-	int ret;
+	int ret = 0;
 
 	if (pad != IMAGE_PAD)
 		return -EINVAL;
 
-	ret = cci_write(imx283->cci, IMX283_REG_STANDBY, IMX283_STANDBY, NULL);
+	cci_write(imx283->cci, IMX283_REG_XMSTA, IMX283_XMSTA_STOP, &ret);
+	cci_write(imx283->cci, IMX283_REG_STANDBY, IMX283_STANDBY, &ret);
+
 	if (ret)
 		dev_err(imx283->dev, "Failed to stop stream\n");
 
author maintainer
#34 lore raw

PATCH 34/188 regulator: act8945a: fix OF node reference imbalance

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 0d15ce31375ccef4162f960b34547a821b7619d2 upstream.

The driver reuses the OF node of the parent multi-function device but fails to take another reference to balance the one dropped by the platform bus code when unbinding the MFD and deregistering the child devices.

Fix this by using the intended helper for reusing OF nodes.

Fixes 38c09961048b ("regulator: act8945a: add regulator driver for ACT8945A")
Cc Wenyou Yang <[email protected]>
---
 drivers/regulator/act8945a-regulator.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/regulator/act8945a-regulator.c +2 -1
@@ -302,8 +302,9 @@ static int act8945a_pmic_probe(struct pl
 		num_regulators = ARRAY_SIZE(act8945a_regulators);
 	}
 
+	device_set_of_node_from_dev(&pdev->dev, pdev->dev.parent);
+
 	config.dev = &pdev->dev;
-	config.dev->of_node = pdev->dev.parent->of_node;
 	config.driver_data = act8945a;
 	for (i = 0; i < num_regulators; i++) {
 		rdev = devm_regulator_register(&pdev->dev, &regulators[i],
author maintainer
#35 lore raw

PATCH 35/188 regulator: s2dos05: fix OF node reference imbalance

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit ebe694d67f159899b063eee61bacda4cb825ed7b upstream.

The driver reuses the OF node of the parent multi-function device but fails to take another reference to balance the one dropped by the platform bus code when unbinding the MFD and deregistering the child devices.

Fix this by using the intended helper for reusing OF nodes.

Fixes bb2441402392 ("regulator: add s2dos05 regulator support")
Cc Dzmitry Sankouski <[email protected]>
---
 drivers/regulator/s2dos05-regulator.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/regulator/s2dos05-regulator.c +1 -1
@@ -126,7 +126,7 @@ static int s2dos05_pmic_probe(struct pla
 	s2dos05->regmap = iodev->regmap_pmic;
 	s2dos05->dev = dev;
 	if (!dev->of_node)
-		dev->of_node = dev->parent->of_node;
+		device_set_of_node_from_dev(dev, dev->parent);
 
 	config.dev = dev;
 	config.driver_data = s2dos05;
author maintainer
#36 lore raw

PATCH 36/188 regulator: bd9571mwv: fix OF node reference imbalance

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 8498100ee1d00422b8c5b161b3e332278b92a59a upstream.

The driver reuses the OF node of the parent multi-function device but fails to take another reference to balance the one dropped by the platform bus code when unbinding the MFD and deregistering the child devices.

Fix this by using the intended helper for reusing OF nodes.

Fixes e85c5a153fe2 ("regulator: Add ROHM BD9571MWV-M PMIC regulator driver")
Cc Marek Vasut <[email protected]>
---
 drivers/regulator/bd9571mwv-regulator.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/regulator/bd9571mwv-regulator.c +2 -1
@@ -287,8 +287,9 @@ static int bd9571mwv_regulator_probe(str
 
 	platform_set_drvdata(pdev, bdreg);
 
+	device_set_of_node_from_dev(&pdev->dev, pdev->dev.parent);
+
 	config.dev = &pdev->dev;
-	config.dev->of_node = pdev->dev.parent->of_node;
 	config.driver_data = bdreg;
 	config.regmap = bdreg->regmap;
 
author maintainer
#37 lore raw

PATCH 37/188 spi: lantiq-ssc: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit b99206710d032c16b7f8b75e4bc18414d8e4b9f4 upstream.

Make sure to deregister the controller before releasing underlying resources like clocks during driver unbind.

Fixes 17f84b793c01 ("spi: lantiq-ssc: add support for Lantiq SSC SPI controller")
Cc Hauke Mehrtens <[email protected]>
---
 drivers/spi/spi-lantiq-ssc.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/spi/spi-lantiq-ssc.c +7 -1
@@ -995,7 +995,7 @@ static int lantiq_ssc_probe(struct platf
 		"Lantiq SSC SPI controller (Rev %i, TXFS %u, RXFS %u, DMA %u)\n",
 		revision, spi->tx_fifo_size, spi->rx_fifo_size, supports_dma);
 
-	err = devm_spi_register_controller(dev, host);
+	err = spi_register_controller(host);
 	if (err) {
 		dev_err(dev, "failed to register spi host\n");
 		goto err_wq_destroy;
@@ -1017,6 +1017,10 @@ static void lantiq_ssc_remove(struct pla
 {
 	struct lantiq_ssc_spi *spi = platform_get_drvdata(pdev);
 
+	spi_controller_get(spi->host);
+
+	spi_unregister_controller(spi->host);
+
 	lantiq_ssc_writel(spi, 0, LTQ_SPI_IRNEN);
 	lantiq_ssc_writel(spi, 0, LTQ_SPI_CLC);
 	rx_fifo_flush(spi);
@@ -1025,6 +1029,8 @@ static void lantiq_ssc_remove(struct pla
 
 	destroy_workqueue(spi->wq);
 	clk_put(spi->fpi_clk);
+
+	spi_controller_put(spi->host);
 }
 
 static struct platform_driver lantiq_ssc_driver = {
author maintainer
#38 lore raw

PATCH 38/188 spi: meson-spicc: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 77953c76bec9af4191f8692a10225dd816208718 upstream.

Make sure to deregister the controller before disabling it to allow SPI device drivers to do I/O during deregistration.

Fixes 454fa271bc4e ("spi: Add Meson SPICC driver")
Cc Neil Armstrong <[email protected]>
---
 drivers/spi/spi-meson-spicc.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/spi/spi-meson-spicc.c +7 -1
@@ -1082,7 +1082,7 @@ static int meson_spicc_probe(struct plat
 		}
 	}
 
-	ret = devm_spi_register_controller(&pdev->dev, host);
+	ret = spi_register_controller(host);
 	if (ret) {
 		dev_err(&pdev->dev, "spi registration failed\n");
 		goto out_host;
@@ -1100,8 +1100,14 @@ static void meson_spicc_remove(struct pl
 {
 	struct meson_spicc_device *spicc = platform_get_drvdata(pdev);
 
+	spi_controller_get(spicc->host);
+
+	spi_unregister_controller(spicc->host);
+
 	/* Disable SPI */
 	writel(0, spicc->base + SPICC_CONREG);
+
+	spi_controller_put(spicc->host);
 }
 
 static const struct meson_spicc_data meson_spicc_gx_data = {
author maintainer
#39 lore raw

PATCH 39/188 spi: qup: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 443e3a0005a4342b218b6dbd4c6387d3c7fed85a upstream.

Make sure to deregister the controller before disabling underlying resources like clocks during driver unbind.

Fixes 64ff247a978f ("spi: Add Qualcomm QUP SPI controller support")
---
 drivers/spi/spi-qup.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/spi/spi-qup.c +7 -1
@@ -1194,7 +1194,7 @@ static int spi_qup_probe(struct platform
 	pm_runtime_set_active(dev);
 	pm_runtime_enable(dev);
 
-	ret = devm_spi_register_controller(dev, host);
+	ret = spi_register_controller(host);
 	if (ret)
 		goto disable_pm;
 
@@ -1321,6 +1321,10 @@ static void spi_qup_remove(struct platfo
 	struct spi_qup *controller = spi_controller_get_devdata(host);
 	int ret;
 
+	spi_controller_get(host);
+
+	spi_unregister_controller(host);
+
 	ret = pm_runtime_get_sync(&pdev->dev);
 
 	if (ret >= 0) {
@@ -1340,6 +1344,8 @@ static void spi_qup_remove(struct platfo
 
 	pm_runtime_put_noidle(&pdev->dev);
 	pm_runtime_disable(&pdev->dev);
+
+	spi_controller_put(host);
 }
 
 static const struct of_device_id spi_qup_dt_match[] = {
author maintainer
#40 lore raw

PATCH 40/188 spi: at91-usart: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 9acecc9bcff058eaef40fd7a4c3650e88b06b220 upstream.

Make sure to deregister the controller before disabling and releasing underlying resources like clocks and DMA during driver unbind.

Fixes e1892546ff66 ("spi: at91-usart: Add driver for at91-usart as SPI")
Cc Radu Pirea <[email protected]>
---
 drivers/spi/spi-at91-usart.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/spi/spi-at91-usart.c +7 -1
@@ -556,7 +556,7 @@ static int at91_usart_spi_probe(struct p
 	spin_lock_init(&aus->lock);
 	init_completion(&aus->xfer_completion);
 
-	ret = devm_spi_register_controller(&pdev->dev, controller);
+	ret = spi_register_controller(controller);
 	if (ret)
 		goto at91_usart_fail_register_controller;
 
@@ -634,8 +634,14 @@ static void at91_usart_spi_remove(struct
 	struct spi_controller *ctlr = platform_get_drvdata(pdev);
 	struct at91_usart_spi *aus = spi_controller_get_devdata(ctlr);
 
+	spi_controller_get(ctlr);
+
+	spi_unregister_controller(ctlr);
+
 	at91_usart_spi_release_dma(ctlr);
 	clk_disable_unprepare(aus->clk);
+
+	spi_controller_put(ctlr);
 }
 
 static const struct dev_pm_ops at91_usart_spi_pm_ops = {
author maintainer
#41 lore raw

PATCH 41/188 media: saa7164: add ioremap return checks and cleanups

In reply to Greg Kroah-Hartman

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

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

From: Wang Jun <[email protected]>

commit d51c60a498e83c9a79884c8e420f97e3885c9583 upstream.

Add checks for ioremap return values in saa7164_dev_setup(). If ioremap for BAR0 or BAR2 fails, release the already allocated PCI memory regions, remove the device from the global list, decrement the device count, and return -ENODEV.

This prevents potential null pointer dereferences and ensures proper cleanup on memory mapping failures.

Fixes 443c1228d505 ("V4L/DVB (12923): SAA7164: Add support for the NXP SAA7164 silicon")
---
 drivers/media/pci/saa7164/saa7164-core.c |   47 +++++++++++++++++++++++--------
 1 file changed, 35 insertions(+), 12 deletions(-)
drivers/media/pci/saa7164/saa7164-core.c +35 -12
@@ -888,6 +888,15 @@ static int get_resources(struct saa7164_
 	return -EBUSY;
 }
 
+static void release_resources(struct saa7164_dev *dev)
+{
+	release_mem_region(pci_resource_start(dev->pci, 0),
+			   pci_resource_len(dev->pci, 0));
+
+	release_mem_region(pci_resource_start(dev->pci, 2),
+			   pci_resource_len(dev->pci, 2));
+}
+
 static int saa7164_port_init(struct saa7164_dev *dev, int portnr)
 {
 	struct saa7164_port *port = NULL;
@@ -947,9 +956,9 @@ static int saa7164_dev_setup(struct saa7
 
 	snprintf(dev->name, sizeof(dev->name), "saa7164[%d]", dev->nr);
 
-	mutex_lock(&devlist);
-	list_add_tail(&dev->devlist, &saa7164_devlist);
-	mutex_unlock(&devlist);
+	scoped_guard(mutex, &devlist) {
+		list_add_tail(&dev->devlist, &saa7164_devlist);
+	}
 
 	/* board config */
 	dev->board = UNSET;
@@ -996,11 +1005,17 @@ static int saa7164_dev_setup(struct saa7
 	}
 
 	/* PCI/e allocations */
-	dev->lmmio = ioremap(pci_resource_start(dev->pci, 0),
-			     pci_resource_len(dev->pci, 0));
+	dev->lmmio = pci_ioremap_bar(dev->pci, 0);
+	if (!dev->lmmio) {
+		dev_err(&dev->pci->dev, "Failed to remap MMIO BAR 0\n");
+		goto err_ioremap_bar0;
+	}
 
-	dev->lmmio2 = ioremap(pci_resource_start(dev->pci, 2),
-			     pci_resource_len(dev->pci, 2));
+	dev->lmmio2 = pci_ioremap_bar(dev->pci, 2);
+	if (!dev->lmmio2) {
+		dev_err(&dev->pci->dev, "Failed to remap MMIO BAR 2\n");
+		goto err_ioremap_bar2;
+	}
 
 	dev->bmmio = (u8 __iomem *)dev->lmmio;
 	dev->bmmio2 = (u8 __iomem *)dev->lmmio2;
@@ -1019,17 +1034,25 @@ static int saa7164_dev_setup(struct saa7
 	saa7164_pci_quirks(dev);
 
 	return 0;
+
+err_ioremap_bar2:
+	iounmap(dev->lmmio);
+err_ioremap_bar0:
+	release_resources(dev);
+
+	scoped_guard(mutex, &devlist) {
+		list_del(&dev->devlist);
+	}
+	saa7164_devcount--;
+
+	return -ENODEV;
 }
 
 static void saa7164_dev_unregister(struct saa7164_dev *dev)
 {
 	dprintk(1, "%s()\n", __func__);
 
-	release_mem_region(pci_resource_start(dev->pci, 0),
-		pci_resource_len(dev->pci, 0));
-
-	release_mem_region(pci_resource_start(dev->pci, 2),
-		pci_resource_len(dev->pci, 2));
+	release_resources(dev);
 
 	if (!atomic_dec_and_test(&dev->refcount))
 		return;
author maintainer
#42 lore raw

PATCH 42/188 spi: amlogic-spisg: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 84d31bb1f6256eea0db6cf64a3c7a53145f92bb9 upstream.

Make sure to deregister the controller before disabling underlying resources like clocks during driver unbind.

Fixes cef9991e04ae ("spi: Add Amlogic SPISG driver")
Cc [email protected] # 6.17: b8db95529979
Cc Sunny Luo <[email protected]>
---
 drivers/spi/spi-amlogic-spisg.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/spi/spi-amlogic-spisg.c +3 -1
@@ -801,7 +801,7 @@ static int aml_spisg_probe(struct platfo
 		goto out_clk;
 	}
 
-	ret = devm_spi_register_controller(dev, ctlr);
+	ret = spi_register_controller(ctlr);
 	if (ret) {
 		dev_err(&pdev->dev, "spi controller registration failed\n");
 		goto out_clk;
@@ -824,6 +824,8 @@ static void aml_spisg_remove(struct plat
 {
 	struct spisg_device *spisg = platform_get_drvdata(pdev);
 
+	spi_unregister_controller(spisg->controller);
+
 	if (!pm_runtime_suspended(&pdev->dev)) {
 		pinctrl_pm_select_sleep_state(&spisg->pdev->dev);
 		clk_disable_unprepare(spisg->core);
author maintainer
#43 lore raw

PATCH 43/188 spi: aspeed-smc: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 1044e5a4ccd57bf5a64f90100a321b498e0267a2 upstream.

Make sure to deregister the controller before disabling it to allow SPI device drivers to do I/O during deregistration.

Fixes e3228ed92893 ("spi: spi-mem: Convert Aspeed SMC driver to spi-mem")
Cc Cédric Le Goater <[email protected]>
---
 drivers/spi/spi-aspeed-smc.c |    9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)
drivers/spi/spi-aspeed-smc.c +6 -3
@@ -726,7 +726,7 @@ static int aspeed_spi_probe(struct platf
 		return -ENOMEM;
 
 	aspi = spi_controller_get_devdata(ctlr);
-	platform_set_drvdata(pdev, aspi);
+	platform_set_drvdata(pdev, ctlr);
 	aspi->data = data;
 	aspi->dev = dev;
 
@@ -765,7 +765,7 @@ static int aspeed_spi_probe(struct platf
 	ctlr->num_chipselect = data->max_cs;
 	ctlr->dev.of_node = dev->of_node;
 
-	ret = devm_spi_register_controller(dev, ctlr);
+	ret = spi_register_controller(ctlr);
 	if (ret)
 		dev_err(&pdev->dev, "spi_register_controller failed\n");
 
@@ -774,7 +774,10 @@ static int aspeed_spi_probe(struct platf
 
 static void aspeed_spi_remove(struct platform_device *pdev)
 {
-	struct aspeed_spi *aspi = platform_get_drvdata(pdev);
+	struct spi_controller *ctlr = platform_get_drvdata(pdev);
+	struct aspeed_spi *aspi = spi_controller_get_devdata(ctlr);
+
+	spi_unregister_controller(ctlr);
 
 	aspeed_spi_enable(aspi, false);
 }
author maintainer
#44 lore raw

PATCH 44/188 platform/x86: hp-wmi: Ignore backlight and FnLock events

In reply to Greg Kroah-Hartman

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

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

From: Krishna Chomal <[email protected]>

commit e8c597368b8500a824c639bfb5ed0044068c6870 upstream.

On HP OmniBook 7 the keyboard backlight and FnLock keys are handled directly by the firmware. However, they still trigger WMI events which results in "Unknown key code" warnings in dmesg.

Add these key codes to the keymap with KE_IGNORE to silence the warnings since no software action is needed.

---
 drivers/platform/x86/hp/hp-wmi.c |    5 +++++
 1 file changed, 5 insertions(+)
drivers/platform/x86/hp/hp-wmi.c +5 -0
@@ -362,6 +362,11 @@ static const struct key_entry hp_wmi_key
 	{ KE_KEY, 0x21a9,  { KEY_TOUCHPAD_OFF } },
 	{ KE_KEY, 0x121a9, { KEY_TOUCHPAD_ON } },
 	{ KE_KEY, 0x231b,  { KEY_HELP } },
+	{ KE_IGNORE, 0x21ab, }, /* FnLock on */
+	{ KE_IGNORE, 0x121ab, }, /* FnLock off */
+	{ KE_IGNORE, 0x30021aa, }, /* kbd backlight: level 2 -> off */
+	{ KE_IGNORE, 0x33221aa, }, /* kbd backlight: off -> level 1 */
+	{ KE_IGNORE, 0x36421aa, }, /* kbd backlight: level 1 -> level 2*/
 	{ KE_END, 0 }
 };
 
author maintainer
#45 lore raw

PATCH 45/188 vsock/virtio: fix MSG_PEEK ignoring skb offset when calculating bytes to copy

In reply to Greg Kroah-Hartman

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

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

From: Luigi Leonardi <[email protected]>

commit 080f22f5d30233faf3d83be3098f35b8be9b7a00 upstream.

virtio_transport_stream_do_peek() does not account for the skb offset when computing the number of bytes to copy.

This means that, after a partial recv() that advances the offset, a peek requesting more bytes than are available in the sk_buff causes skb_copy_datagram_iter() to go past the valid payload, resulting in a -EFAULT.

The dequeue path already handles this correctly.
Apply the same logic to the peek path.

---
 net/vmw_vsock/virtio_transport_common.c |    5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
net/vmw_vsock/virtio_transport_common.c +2 -3
@@ -546,9 +546,8 @@ virtio_transport_stream_do_peek(struct v
 	skb_queue_walk(&vvs->rx_queue, skb) {
 		size_t bytes;
 
-		bytes = len - total;
-		if (bytes > skb->len)
-			bytes = skb->len;
+		bytes = min_t(size_t, len - total,
+			      skb->len - VIRTIO_VSOCK_SKB_CB(skb)->offset);
 
 		spin_unlock_bh(&vvs->rx_lock);
 
author maintainer
#46 lore raw

PATCH 46/188 arm64: dts: broadcom: bcm2712-d-rpi-5-b: add fixes for pinctrl/pinctrl_aon

In reply to Greg Kroah-Hartman

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

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

From: Gregor Herburger <[email protected]>

commit aeb078cebc40d421f61a8f07b0e7919aeb44d751 upstream.

On the -d revision of the bcm2712 the pinctrl differs from the c0 revision. The driver already supports both and distinguishes the two with the compatible string.

Update the compatible string and reg length to reflect the different pinctrl.

---
 arch/arm64/boot/dts/broadcom/bcm2712-d-rpi-5-b.dts |   10 ++++++++++
 1 file changed, 10 insertions(+)
arch/arm64/boot/dts/broadcom/bcm2712-d-rpi-5-b.dts +10 -0
@@ -35,3 +35,13 @@
 		"PMIC_SCL", // AON_SGPIO_04
 		"PMIC_SDA"; // AON_SGPIO_05
 };
+
+&pinctrl {
+	compatible = "brcm,bcm2712d0-pinctrl";
+	reg = <0x7d504100 0x20>;
+};
+
+&pinctrl_aon {
+	compatible = "brcm,bcm2712d0-aon-pinctrl";
+	reg = <0x7d510700 0x1c>;
+};
author maintainer
#47 lore raw

PATCH 47/188 arm64: dts: broadcom: bcm2712-d-rpi-5-b: update uart10 interrupt

In reply to Greg Kroah-Hartman

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

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

From: Gregor Herburger <[email protected]>

commit 18d4a06e10051681de074a9250e54afc1f3ee312 upstream.

On the -d revision of bcm2712 the uart interrupt is on 120. Update it accordingly.

---
 arch/arm64/boot/dts/broadcom/bcm2712-d-rpi-5-b.dts |    4 ++++
 1 file changed, 4 insertions(+)
arch/arm64/boot/dts/broadcom/bcm2712-d-rpi-5-b.dts +4 -0
@@ -45,3 +45,7 @@
 	compatible = "brcm,bcm2712d0-aon-pinctrl";
 	reg = <0x7d510700 0x1c>;
 };
+
+&uart10 {
+	interrupts = <GIC_SPI 120 IRQ_TYPE_LEVEL_HIGH>;
+};
author maintainer
#48 lore raw

PATCH 48/188 media: pci: zoran: fix potential memory leak in zoran_probe()

In reply to Greg Kroah-Hartman

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

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

From: Abdun Nihaal <[email protected]>

commit 8ea21435fe36fb853706f4935d78bc11beb63fb4 upstream.

The memory allocated for codec in videocodec_attach() is not freed in one of the error paths, due to an incorrect goto label. Fix the label to free it on error.

Fixes 8f7cc5c0b0eb ("media: staging: media: zoran: introduce zoran_i2c_init")
---
 drivers/media/pci/zoran/zoran_card.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/media/pci/zoran/zoran_card.c +1 -1
@@ -1373,7 +1373,7 @@ static int zoran_probe(struct pci_dev *p
 		}
 		if (zr->codec->type != zr->card.video_codec) {
 			pci_err(pdev, "%s - wrong codec\n", __func__);
-			goto zr_unreg_videocodec;
+			goto zr_detach_codec;
 		}
 	}
 	if (zr->card.video_vfe != 0) {
author maintainer
#49 lore raw

PATCH 49/188 media: dib8000: avoid division by 0 in dib8000_set_dds()

In reply to Greg Kroah-Hartman

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

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

From: Sergey Shtylyov <[email protected]>

commit dde3c37af95cd6fa301c4906f33d627bc9dd874c upstream.

In dib8000_set_dds(), 1 << 26 (67108864) divided by e.g. 1 apparently can't fit into 16-bit variable unit_khz_dds_val, being truncated to 0; this will cause division by 0 while calling dprintk() with debugging enabled (via the module parameter). Use s32 instead of s16 to declare the variable, getting rid of the cast to u16 in the *else* branch as well...

Found by Linux Verification Center (linuxtesting.org) with the Svace static analysis tool.

Fixes 173a64cb3fcf ("[media] dib8000: enhancement")
---
 drivers/media/dvb-frontends/dib8000.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/media/dvb-frontends/dib8000.c +2 -2
@@ -2695,7 +2695,7 @@ static void dib8000_viterbi_state(struct
 
 static void dib8000_set_dds(struct dib8000_state *state, s32 offset_khz)
 {
-	s16 unit_khz_dds_val;
+	s32 unit_khz_dds_val;
 	u32 abs_offset_khz = abs(offset_khz);
 	u32 dds = state->cfg.pll->ifreq & 0x1ffffff;
 	u8 invert = !!(state->cfg.pll->ifreq & (1 << 25));
@@ -2716,7 +2716,7 @@ static void dib8000_set_dds(struct dib80
 			dds = (1<<26) - dds;
 	} else {
 		ratio = 2;
-		unit_khz_dds_val = (u16) (67108864 / state->cfg.pll->internal);
+		unit_khz_dds_val = 67108864 / state->cfg.pll->internal;
 
 		if (offset_khz < 0)
 			unit_khz_dds_val *= -1;
author maintainer
#50 lore raw

PATCH 50/188 media: i2c: imx412: Assert reset GPIO during probe

In reply to Greg Kroah-Hartman

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

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

From: Wenmeng Liu <[email protected]>

commit 8467c5ff5acae28513bc1e0af535e06b41b04344 upstream.

Assert the reset GPIO before first power up. This avoids a mismatch where the first power up (when the reset GPIO defaults deasserted) differs from subsequent cycles.

Fixes 9214e86c0cc1 ("media: i2c: Add imx412 camera sensor driver")
---
 drivers/media/i2c/imx412.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/media/i2c/imx412.c +1 -1
@@ -925,7 +925,7 @@ static int imx412_parse_hw_config(struct
 
 	/* Request optional reset pin */
 	imx412->reset_gpio = devm_gpiod_get_optional(imx412->dev, "reset",
-						     GPIOD_OUT_LOW);
+						     GPIOD_OUT_HIGH);
 	if (IS_ERR(imx412->reset_gpio)) {
 		dev_err(imx412->dev, "failed to get reset gpio %ld\n",
 			PTR_ERR(imx412->reset_gpio));
author maintainer
#51 lore raw

PATCH 51/188 media: staging: imx: request mbus_config in csi_start

In reply to Greg Kroah-Hartman

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

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

From: Michael Tretter <[email protected]>

commit 9df2aaa64890c0b6226057eb6fcb6352bd2df432 upstream.

Request the upstream mbus_config in csi_start, which starts the stream, instead of caching it in link_validate.

This allows to get rid of the mbus_cfg field in the struct csi_priv and avoids state in the driver.

Fixes 4a34ec8e470c ("[media] media: imx: Add CSI subdev driver")
---
 drivers/staging/media/imx/imx-media-csi.c |   40 ++++++++++++++++++------------
 1 file changed, 24 insertions(+), 16 deletions(-)
drivers/staging/media/imx/imx-media-csi.c +24 -16
@@ -97,9 +97,6 @@ struct csi_priv {
 	/* the mipi virtual channel number at link validate */
 	int vc_num;
 
-	/* media bus config of the upstream subdevice CSI is receiving from */
-	struct v4l2_mbus_config mbus_cfg;
-
 	spinlock_t irqlock; /* protect eof_irq handler */
 	struct timer_list eof_timeout_timer;
 	int eof_irq;
@@ -403,7 +400,8 @@ static void csi_idmac_unsetup_vb2_buf(st
 }
 
 /* init the SMFC IDMAC channel */
-static int csi_idmac_setup_channel(struct csi_priv *priv)
+static int csi_idmac_setup_channel(struct csi_priv *priv,
+				   struct v4l2_mbus_config *mbus_cfg)
 {
 	struct imx_media_video_dev *vdev = priv->vdev;
 	const struct imx_media_pixfmt *incc;
@@ -432,7 +430,7 @@ static int csi_idmac_setup_channel(struc
 	image.phys0 = phys[0];
 	image.phys1 = phys[1];
 
-	passthrough = requires_passthrough(&priv->mbus_cfg, infmt, incc);
+	passthrough = requires_passthrough(mbus_cfg, infmt, incc);
 	passthrough_cycles = 1;
 
 	/*
@@ -572,11 +570,12 @@ static void csi_idmac_unsetup(struct csi
 	csi_idmac_unsetup_vb2_buf(priv, state);
 }
 
-static int csi_idmac_setup(struct csi_priv *priv)
+static int csi_idmac_setup(struct csi_priv *priv,
+			   struct v4l2_mbus_config *mbus_cfg)
 {
 	int ret;
 
-	ret = csi_idmac_setup_channel(priv);
+	ret = csi_idmac_setup_channel(priv, mbus_cfg);
 	if (ret)
 		return ret;
 
@@ -595,7 +594,8 @@ static int csi_idmac_setup(struct csi_pr
 	return 0;
 }
 
-static int csi_idmac_start(struct csi_priv *priv)
+static int csi_idmac_start(struct csi_priv *priv,
+			   struct v4l2_mbus_config *mbus_cfg)
 {
 	struct imx_media_video_dev *vdev = priv->vdev;
 	int ret;
@@ -619,7 +619,7 @@ static int csi_idmac_start(struct csi_pr
 	priv->last_eof = false;
 	priv->nfb4eof = false;
 
-	ret = csi_idmac_setup(priv);
+	ret = csi_idmac_setup(priv, mbus_cfg);
 	if (ret) {
 		v4l2_err(&priv->sd, "csi_idmac_setup failed: %d\n", ret);
 		goto out_free_dma_buf;
@@ -701,7 +701,8 @@ static void csi_idmac_stop(struct csi_pr
 }
 
 /* Update the CSI whole sensor and active windows */
-static int csi_setup(struct csi_priv *priv)
+static int csi_setup(struct csi_priv *priv,
+		     struct v4l2_mbus_config *mbus_cfg)
 {
 	struct v4l2_mbus_framefmt *infmt, *outfmt;
 	const struct imx_media_pixfmt *incc;
@@ -719,7 +720,7 @@ static int csi_setup(struct csi_priv *pr
 	 * if cycles is set, we need to handle this over multiple cycles as
 	 * generic/bayer data
 	 */
-	if (is_parallel_bus(&priv->mbus_cfg) && incc->cycles) {
+	if (is_parallel_bus(mbus_cfg) && incc->cycles) {
 		if_fmt.width *= incc->cycles;
 		crop.width *= incc->cycles;
 	}
@@ -730,7 +731,7 @@ static int csi_setup(struct csi_priv *pr
 			     priv->crop.width == 2 * priv->compose.width,
 			     priv->crop.height == 2 * priv->compose.height);
 
-	ipu_csi_init_interface(priv->csi, &priv->mbus_cfg, &if_fmt, outfmt);
+	ipu_csi_init_interface(priv->csi, mbus_cfg, &if_fmt, outfmt);
 
 	ipu_csi_set_dest(priv->csi, priv->dest);
 
@@ -745,9 +746,17 @@ static int csi_setup(struct csi_priv *pr
 
 static int csi_start(struct csi_priv *priv)
 {
+	struct v4l2_mbus_config mbus_cfg = { .type = 0 };
 	struct v4l2_fract *input_fi, *output_fi;
 	int ret;
 
+	ret = csi_get_upstream_mbus_config(priv, &mbus_cfg);
+	if (ret) {
+		v4l2_err(&priv->sd,
+			 "failed to get upstream media bus configuration\n");
+		return ret;
+	}
+
 	input_fi = &priv->frame_interval[CSI_SINK_PAD];
 	output_fi = &priv->frame_interval[priv->active_output_pad];
 
@@ -758,7 +767,7 @@ static int csi_start(struct csi_priv *pr
 		return ret;
 
 	/* Skip first few frames from a BT.656 source */
-	if (priv->mbus_cfg.type == V4L2_MBUS_BT656) {
+	if (mbus_cfg.type == V4L2_MBUS_BT656) {
 		u32 delay_usec, bad_frames = 20;
 
 		delay_usec = DIV_ROUND_UP_ULL((u64)USEC_PER_SEC *
@@ -769,12 +778,12 @@ static int csi_start(struct csi_priv *pr
 	}
 
 	if (priv->dest == IPU_CSI_DEST_IDMAC) {
-		ret = csi_idmac_start(priv);
+		ret = csi_idmac_start(priv, &mbus_cfg);
 		if (ret)
 			goto stop_upstream;
 	}
 
-	ret = csi_setup(priv);
+	ret = csi_setup(priv, &mbus_cfg);
 	if (ret)
 		goto idmac_stop;
 
@@ -1138,7 +1147,6 @@ static int csi_link_validate(struct v4l2
 
 	mutex_lock(&priv->lock);
 
-	priv->mbus_cfg = mbus_cfg;
 	is_csi2 = !is_parallel_bus(&mbus_cfg);
 	if (is_csi2) {
 		/*
author maintainer
#52 lore raw

PATCH 52/188 media: i2c: ov08d10: fix image vertical start setting

In reply to Greg Kroah-Hartman

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

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

From: Matthias Fend <[email protected]>

commit 5d150fa0f16096d736bd24d13e04495da5116fab upstream.

The current settings for the "image vertical start" register appear to be incorrect. While this only results in an incorrect start line for native modes, this faulty setting causes actual problems in binning mode. At least on an i.MX8MP test system, only corrupted frames could be received. To correct this, the recommended settings from the reference register sets are used for all modes. Since this shifts the start by one line, the Bayer pattern also changes, which has also been corrected.

Fixes 7be91e02ed57 ("media: i2c: Add ov08d10 camera sensor driver")
---
 drivers/media/i2c/ov08d10.c |   10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)
drivers/media/i2c/ov08d10.c +5 -5
@@ -217,7 +217,7 @@ static const struct ov08d10_reg lane_2_m
 	{0x9a, 0x30},
 	{0xa8, 0x02},
 	{0xfd, 0x02},
-	{0xa1, 0x01},
+	{0xa1, 0x00},
 	{0xa2, 0x09},
 	{0xa3, 0x9c},
 	{0xa5, 0x00},
@@ -335,7 +335,7 @@ static const struct ov08d10_reg lane_2_m
 	{0x9a, 0x30},
 	{0xa8, 0x02},
 	{0xfd, 0x02},
-	{0xa1, 0x09},
+	{0xa1, 0x08},
 	{0xa2, 0x09},
 	{0xa3, 0x90},
 	{0xa5, 0x08},
@@ -467,7 +467,7 @@ static const struct ov08d10_reg lane_2_m
 	{0xaa, 0xd0},
 	{0xab, 0x06},
 	{0xac, 0x68},
-	{0xa1, 0x09},
+	{0xa1, 0x04},
 	{0xa2, 0x04},
 	{0xa3, 0xc8},
 	{0xa5, 0x04},
@@ -613,8 +613,8 @@ static const struct ov08d10_lane_cfg lan
 static u32 ov08d10_get_format_code(struct ov08d10 *ov08d10)
 {
 	static const u32 codes[2][2] = {
-		{ MEDIA_BUS_FMT_SGRBG10_1X10, MEDIA_BUS_FMT_SRGGB10_1X10},
-		{ MEDIA_BUS_FMT_SBGGR10_1X10, MEDIA_BUS_FMT_SGBRG10_1X10},
+		{ MEDIA_BUS_FMT_SBGGR10_1X10, MEDIA_BUS_FMT_SGBRG10_1X10 },
+		{ MEDIA_BUS_FMT_SGRBG10_1X10, MEDIA_BUS_FMT_SRGGB10_1X10 },
 	};
 
 	return codes[ov08d10->vflip->val][ov08d10->hflip->val];
author maintainer
#53 lore raw

PATCH 53/188 media: i2c: ov08d10: fix runtime PM handling in probe

In reply to Greg Kroah-Hartman

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

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

From: Matthias Fend <[email protected]>

commit 35c7046be2be5e60be8128facb359a47f39e99cd upstream.

Set the device's runtime PM status and enable runtime PM before registering the async sub-device. This is needed to avoid the case where the device is runtime PM resumed while runtime PM has not been enabled yet.

Remove the related, non-driver-specific comment while at it.

Fixes 7be91e02ed57 ("media: i2c: Add ov08d10 camera sensor driver")
---
 drivers/media/i2c/ov08d10.c |   11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)
drivers/media/i2c/ov08d10.c +5 -6
@@ -1430,6 +1430,9 @@ static int ov08d10_probe(struct i2c_clie
 		goto probe_error_v4l2_ctrl_handler_free;
 	}
 
+	pm_runtime_set_active(ov08d10->dev);
+	pm_runtime_enable(ov08d10->dev);
+
 	ret = v4l2_async_register_subdev_sensor(&ov08d10->sd);
 	if (ret < 0) {
 		dev_err(ov08d10->dev, "failed to register V4L2 subdev: %d",
@@ -1437,17 +1440,13 @@ static int ov08d10_probe(struct i2c_clie
 		goto probe_error_media_entity_cleanup;
 	}
 
-	/*
-	 * Device is already turned on by i2c-core with ACPI domain PM.
-	 * Enable runtime PM and turn off the device.
-	 */
-	pm_runtime_set_active(ov08d10->dev);
-	pm_runtime_enable(ov08d10->dev);
 	pm_runtime_idle(ov08d10->dev);
 
 	return 0;
 
 probe_error_media_entity_cleanup:
+	pm_runtime_disable(ov08d10->dev);
+	pm_runtime_set_suspended(ov08d10->dev);
 	media_entity_cleanup(&ov08d10->sd.entity);
 
 probe_error_v4l2_ctrl_handler_free:
author maintainer
#54 lore raw

PATCH 54/188 media: omap3isp: drop the use count of v4l2 pipeline

In reply to Greg Kroah-Hartman

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

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

From: Haoxiang Li <[email protected]>

commit 9da49bd9d4224035cff39b40d7395310abb10201 upstream.

In isp_video_open(), drop the use count of v4l2
pipeline if vb2_queue_init() fails.

Fixes 8fd390b89cc8 ("media: Split v4l2_pipeline_pm_use into v4l2_pipeline_pm_{get, put}")
---
 drivers/media/platform/ti/omap3isp/ispvideo.c |    1 +
 1 file changed, 1 insertion(+)
drivers/media/platform/ti/omap3isp/ispvideo.c +1 -0
@@ -1325,6 +1325,7 @@ static int isp_video_open(struct file *f
 
 	ret = vb2_queue_init(&handle->queue);
 	if (ret < 0) {
+		v4l2_pipeline_pm_put(&video->video.entity);
 		omap3isp_put(video->isp);
 		goto done;
 	}
author maintainer
#55 lore raw

PATCH 55/188 media: iris: fix QCOM_MDT_LOADER dependency

In reply to Greg Kroah-Hartman

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

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

From: Arnd Bergmann <[email protected]>

commit a297c5165f91366cbc3490e630aabd1c0f70efb8 upstream.

When build-testined with CONFIG_QCOM_MDT_LOADER=m and VIDEO_QCOM_IRIS=y, the kernel fails to link:

x86_64-linux-ld: drivers/media/platform/qcom/iris/iris_firmware.o: in function `iris_fw_load':
iris_firmware.c:(.text+0xb0): undefined reference to qcom_mdt_get_size' iris_firmware.c:(.text+0xfd): undefined reference to qcom_mdt_load'

The problem is the conditional 'select' statement. Change this to make the driver built-in here regardless of CONFIG_ARCH_QCOM.

---
 drivers/media/platform/qcom/iris/Kconfig |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/media/platform/qcom/iris/Kconfig +1 -1
@@ -3,7 +3,7 @@ config VIDEO_QCOM_IRIS
         depends on VIDEO_DEV
         depends on ARCH_QCOM || COMPILE_TEST
         select V4L2_MEM2MEM_DEV
-        select QCOM_MDT_LOADER if ARCH_QCOM
+        select QCOM_MDT_LOADER
         select QCOM_SCM
         select VIDEOBUF2_DMA_CONTIG
         help
author maintainer
#56 lore raw

PATCH 56/188 media: iris: Fix use-after-free in iris_release_internal_buffers()

In reply to Greg Kroah-Hartman

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

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

From: Dikshita Agarwal <[email protected]>

commit f27cfdcfc916bb59297825805f4c3499f89f9e76 upstream.

The recent change in commit 1dabf00ee206 ("media: iris: gen1: Destroy internal buffers after FW releases") introduced a regression where session_release_buf() may free the buffer. The caller, iris_release_internal_buffers(), continued to access buffer after the call, leading to a potential use-after-free.

Fix this by setting BUF_ATTR_PENDING_RELEASE before calling session_release_buf(), and reverting the flag if the call fails. This ensures no dereference occurs after potential freeing.

---
 drivers/media/platform/qcom/iris/iris_buffer.c |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/media/platform/qcom/iris/iris_buffer.c +4 -2
@@ -571,10 +571,12 @@ static int iris_release_internal_buffers
 			continue;
 		if (!(buffer->attr & BUF_ATTR_QUEUED))
 			continue;
+		buffer->attr |= BUF_ATTR_PENDING_RELEASE;
 		ret = hfi_ops->session_release_buf(inst, buffer);
-		if (ret)
+		if (ret) {
+			buffer->attr &= ~BUF_ATTR_PENDING_RELEASE;
 			return ret;
-		buffer->attr |= BUF_ATTR_PENDING_RELEASE;
+		}
 	}
 
 	return 0;
author maintainer
#57 lore raw

PATCH 57/188 media: qcom: camss: Fix csid clock configuration for sa8775p

In reply to Greg Kroah-Hartman

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

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

From: Wenmeng Liu <[email protected]>

commit fe56c674118aa46da1a3e65aa22ca709ebd7d812 upstream.

Fix the mismatch between clock list and clock rate table for CSID lite instances. The current implementation has 5 clocks defined but only 2 are actually needed (vfe_lite_csid and vfe_lite_cphy_rx), while the clock rate table doesn't match this configuration.

Update both clock list and rate table to maintain consistency:
- Remove unused clocks: cpas_vfe_lite, vfe_lite_ahb, vfe_lite
- Update clock rate table to match the remaining two clocks

Fixes ed03e99de0fa ("media: qcom: camss: Add support for CSID 690")
---
 drivers/media/platform/qcom/camss/camss.c |   40 +++++++++++-------------------
 1 file changed, 15 insertions(+), 25 deletions(-)
drivers/media/platform/qcom/camss/camss.c +15 -25
@@ -2782,12 +2782,10 @@ static const struct camss_subdev_resourc
 	/* CSID2 (lite) */
 	{
 		.regulators = {},
-		.clock = { "cpas_vfe_lite", "vfe_lite_ahb",
-			   "vfe_lite_csid", "vfe_lite_cphy_rx",
-			   "vfe_lite"},
+		.clock = { "vfe_lite_csid", "vfe_lite_cphy_rx" },
 		.clock_rate = {
-			{ 0, 0, 400000000, 400000000, 0},
-			{ 0, 0, 400000000, 480000000, 0}
+			{ 400000000, 480000000 },
+			{ 400000000, 480000000 }
 		},
 		.reg = { "csid_lite0" },
 		.interrupt = { "csid_lite0" },
@@ -2801,12 +2799,10 @@ static const struct camss_subdev_resourc
 	/* CSID3 (lite) */
 	{
 		.regulators = {},
-		.clock = { "cpas_vfe_lite", "vfe_lite_ahb",
-			   "vfe_lite_csid", "vfe_lite_cphy_rx",
-			   "vfe_lite"},
+		.clock = { "vfe_lite_csid", "vfe_lite_cphy_rx" },
 		.clock_rate = {
-			{ 0, 0, 400000000, 400000000, 0},
-			{ 0, 0, 400000000, 480000000, 0}
+			{ 400000000, 480000000 },
+			{ 400000000, 480000000 }
 		},
 		.reg = { "csid_lite1" },
 		.interrupt = { "csid_lite1" },
@@ -2820,12 +2816,10 @@ static const struct camss_subdev_resourc
 	/* CSID4 (lite) */
 	{
 		.regulators = {},
-		.clock = { "cpas_vfe_lite", "vfe_lite_ahb",
-			   "vfe_lite_csid", "vfe_lite_cphy_rx",
-			   "vfe_lite"},
+		.clock = { "vfe_lite_csid", "vfe_lite_cphy_rx" },
 		.clock_rate = {
-			{ 0, 0, 400000000, 400000000, 0},
-			{ 0, 0, 400000000, 480000000, 0}
+			{ 400000000, 480000000 },
+			{ 400000000, 480000000 }
 		},
 		.reg = { "csid_lite2" },
 		.interrupt = { "csid_lite2" },
@@ -2839,12 +2833,10 @@ static const struct camss_subdev_resourc
 	/* CSID5 (lite) */
 	{
 		.regulators = {},
-		.clock = { "cpas_vfe_lite", "vfe_lite_ahb",
-			   "vfe_lite_csid", "vfe_lite_cphy_rx",
-			   "vfe_lite"},
+		.clock = { "vfe_lite_csid", "vfe_lite_cphy_rx" },
 		.clock_rate = {
-			{ 0, 0, 400000000, 400000000, 0},
-			{ 0, 0, 400000000, 480000000, 0}
+			{ 400000000, 480000000 },
+			{ 400000000, 480000000 }
 		},
 		.reg = { "csid_lite3" },
 		.interrupt = { "csid_lite3" },
@@ -2858,12 +2850,10 @@ static const struct camss_subdev_resourc
 	/* CSID6 (lite) */
 	{
 		.regulators = {},
-		.clock = { "cpas_vfe_lite", "vfe_lite_ahb",
-			   "vfe_lite_csid", "vfe_lite_cphy_rx",
-			   "vfe_lite"},
+		.clock = { "vfe_lite_csid", "vfe_lite_cphy_rx" },
 		.clock_rate = {
-			{ 0, 0, 400000000, 400000000, 0},
-			{ 0, 0, 400000000, 480000000, 0}
+			{ 400000000, 480000000 },
+			{ 400000000, 480000000 }
 		},
 		.reg = { "csid_lite4" },
 		.interrupt = { "csid_lite4" },
author maintainer
#58 lore raw

PATCH 58/188 media: qcom: camss: Fix csid IRQ offset for sa8775p

In reply to Greg Kroah-Hartman

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

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

From: Wenmeng Liu <[email protected]>

commit dd1b373941079cc102cc18bc68884e18245f5912 upstream.

Fix BUF_DONE_IRQ_STATUS_RDI_OFFSET calculation for csid lite on sa8775p platform. The offset should be 0 for csid lite on sa8775p,

Fixes ed03e99de0fa ("media: qcom: camss: Add support for CSID 690")
---
 drivers/media/platform/qcom/camss/camss-csid-gen3.c |    6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/media/platform/qcom/camss/camss-csid-gen3.c +3 -3
@@ -48,9 +48,9 @@
 #define IS_CSID_690(csid)	((csid->camss->res->version == CAMSS_8775P) \
 				 || (csid->camss->res->version == CAMSS_8300))
 #define CSID_BUF_DONE_IRQ_STATUS	0x8C
-#define BUF_DONE_IRQ_STATUS_RDI_OFFSET  (csid_is_lite(csid) ?\
-						1 : (IS_CSID_690(csid) ?\
-						13 : 14))
+#define BUF_DONE_IRQ_STATUS_RDI_OFFSET  (csid_is_lite(csid) ? \
+						((IS_CSID_690(csid) ? 0 : 1)) : \
+						((IS_CSID_690(csid) ? 13 : 14)))
 #define CSID_BUF_DONE_IRQ_MASK		0x90
 #define CSID_BUF_DONE_IRQ_CLEAR		0x94
 #define CSID_BUF_DONE_IRQ_SET		0x98
author maintainer
#59 lore raw

PATCH 59/188 media: qcom: iris: increase H265D_MAX_SLICE to fix H.265 decoding on SC7280

In reply to Greg Kroah-Hartman

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

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

From: Dmitry Baryshkov <[email protected]>

commit 3e0b2053751657ed2924adfe3ff25b1450231e33 upstream.

Follow the commit bfe1326573ff ("venus: Fix for H265 decoding failure.") and increase H265D_MAX_SLICE following firmware requirements on that platform. Otherwise decoding of the H.265 streams fails with the "insufficient scratch_1 buffer size" from the firmware.

[bod: Fixed commit log withthe => with the]
Fixes e1f5d32608ec ("media: iris: Add internal buffer calculation for HEVC and VP9 decoders")
---
 drivers/media/platform/qcom/iris/iris_vpu_buffer.h |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/media/platform/qcom/iris/iris_vpu_buffer.h +1 -1
@@ -61,7 +61,7 @@ struct iris_inst;
 #define SIZE_DOLBY_RPU_METADATA (41 * 1024)
 #define H264_CABAC_HDR_RATIO_HD_TOT	1
 #define H264_CABAC_RES_RATIO_HD_TOT	3
-#define H265D_MAX_SLICE	1200
+#define H265D_MAX_SLICE	3600
 #define SIZE_H265D_HW_PIC_T SIZE_H264D_HW_PIC_T
 #define H265_CABAC_HDR_RATIO_HD_TOT 2
 #define H265_CABAC_RES_RATIO_HD_TOT 2
author maintainer
#60 lore raw

PATCH 60/188 media: venus: fix QCOM_MDT_LOADER dependency

In reply to Greg Kroah-Hartman

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

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

From: Arnd Bergmann <[email protected]>

commit aa23c94cc433b145d1ce93820ecdfe16d8940e28 upstream.

When build-testined with CONFIG_QCOM_MDT_LOADER=m and VIDEO_QCOM_VENUS=y, the kernel fails to link:

x86_64-linux-ld: drivers/media/platform/qcom/venus/firmware.o: in function `venus_boot':
firmware.c:(.text+0x1e3): undefined reference to qcom_mdt_get_size' firmware.c:(.text+0x25a): undefined reference to qcom_mdt_load' firmware.c:(.text+0x272): undefined reference to `qcom_mdt_load_no_init'

The problem is the conditional 'select' statement. Change this to make the driver built-in here regardless of CONFIG_ARCH_QCOM, same as for the similar IRIS driver.

Fixes 0399b696f7f4 ("media: venus: fix compile-test build on non-qcom ARM platform")
---
 drivers/media/platform/qcom/venus/Kconfig |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/media/platform/qcom/venus/Kconfig +1 -1
@@ -4,7 +4,7 @@ config VIDEO_QCOM_VENUS
 	depends on VIDEO_DEV && QCOM_SMEM
 	depends on (ARCH_QCOM && ARM64 && IOMMU_API) || COMPILE_TEST
 	select OF_DYNAMIC if ARCH_QCOM
-	select QCOM_MDT_LOADER if ARCH_QCOM
+	select QCOM_MDT_LOADER
 	select QCOM_SCM
 	select VIDEOBUF2_DMA_CONTIG
 	select V4L2_MEM2MEM_DEV
author maintainer
#61 lore raw

PATCH 61/188 media: iris: Fix dma_free_attrs() size in iris_hfi_queues_init()

In reply to Greg Kroah-Hartman

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

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

From: Thomas Fourier <[email protected]>

commit 4a49ae56b0e4268d48fd96babe0cc68596bc301a upstream.

The core->iface_q_table_vaddr buffer is alloc'd with size queue_size but freed with sizeof(*q_tbl_hdr) which is different.

Change the dma_free_attrs() size.

Fixes d7378f84e94e ("media: iris: introduce iris core state management with shared queues")
---
 drivers/media/platform/qcom/iris/iris_hfi_queue.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/media/platform/qcom/iris/iris_hfi_queue.c +1 -1
@@ -263,7 +263,7 @@ int iris_hfi_queues_init(struct iris_cor
 					  GFP_KERNEL, DMA_ATTR_WRITE_COMBINE);
 	if (!core->sfr_vaddr) {
 		dev_err(core->dev, "sfr alloc and map failed\n");
-		dma_free_attrs(core->dev, sizeof(*q_tbl_hdr), core->iface_q_table_vaddr,
+		dma_free_attrs(core->dev, queue_size, core->iface_q_table_vaddr,
 			       core->iface_q_table_daddr, DMA_ATTR_WRITE_COMBINE);
 		return -ENOMEM;
 	}
author maintainer
#62 lore raw

PATCH 62/188 media: qcom: camss: Add missing clocks for VFE lite on sa8775p

In reply to Greg Kroah-Hartman

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

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

From: Wenmeng Liu <[email protected]>

commit d31fac47b39f5e1ed85a587688ca70b793e421b4 upstream.

Add missing required clocks (cpas_ahb and camnoc_axi) for VFE lite instances on sa8775p platform. These clocks are necessary for proper VFE lite operation:

Fixes e7b59e1d06fb ("media: qcom: camss: Add support for VFE 690")
---
 drivers/media/platform/qcom/camss/camss.c |   40 ++++++++++++++++++------------
 1 file changed, 25 insertions(+), 15 deletions(-)
drivers/media/platform/qcom/camss/camss.c +25 -15
@@ -2926,15 +2926,17 @@ static const struct camss_subdev_resourc
 	/* VFE2 (lite) */
 	{
 		.regulators = {},
-		.clock = { "cpas_vfe_lite", "vfe_lite_ahb",
+		.clock = { "cpas_ahb", "cpas_vfe_lite", "vfe_lite_ahb",
 			   "vfe_lite_csid", "vfe_lite_cphy_rx",
-			   "vfe_lite"},
+			   "vfe_lite", "camnoc_axi"},
 		.clock_rate = {
-			{ 0, 0, 0, 0  },
+			{ 0 },
+			{ 0 },
 			{ 300000000, 400000000, 400000000, 400000000 },
 			{ 400000000, 400000000, 400000000, 400000000 },
 			{ 400000000, 400000000, 400000000, 400000000 },
 			{ 480000000, 600000000, 600000000, 600000000 },
+			{ 400000000 },
 		},
 		.reg = { "vfe_lite0" },
 		.interrupt = { "vfe_lite0" },
@@ -2949,15 +2951,17 @@ static const struct camss_subdev_resourc
 	/* VFE3 (lite) */
 	{
 		.regulators = {},
-		.clock = { "cpas_vfe_lite", "vfe_lite_ahb",
+		.clock = { "cpas_ahb", "cpas_vfe_lite", "vfe_lite_ahb",
 			   "vfe_lite_csid", "vfe_lite_cphy_rx",
-			   "vfe_lite"},
+			   "vfe_lite", "camnoc_axi"},
 		.clock_rate = {
-			{ 0, 0, 0, 0  },
+			{ 0 },
+			{ 0 },
 			{ 300000000, 400000000, 400000000, 400000000 },
 			{ 400000000, 400000000, 400000000, 400000000 },
 			{ 400000000, 400000000, 400000000, 400000000 },
 			{ 480000000, 600000000, 600000000, 600000000 },
+			{ 400000000 },
 		},
 		.reg = { "vfe_lite1" },
 		.interrupt = { "vfe_lite1" },
@@ -2972,15 +2976,17 @@ static const struct camss_subdev_resourc
 	/* VFE4 (lite) */
 	{
 		.regulators = {},
-		.clock = { "cpas_vfe_lite", "vfe_lite_ahb",
+		.clock = { "cpas_ahb", "cpas_vfe_lite", "vfe_lite_ahb",
 			   "vfe_lite_csid", "vfe_lite_cphy_rx",
-			   "vfe_lite"},
+			   "vfe_lite", "camnoc_axi"},
 		.clock_rate = {
-			{ 0, 0, 0, 0  },
+			{ 0 },
+			{ 0 },
 			{ 300000000, 400000000, 400000000, 400000000 },
 			{ 400000000, 400000000, 400000000, 400000000 },
 			{ 400000000, 400000000, 400000000, 400000000 },
 			{ 480000000, 600000000, 600000000, 600000000 },
+			{ 400000000 },
 		},
 		.reg = { "vfe_lite2" },
 		.interrupt = { "vfe_lite2" },
@@ -2995,15 +3001,17 @@ static const struct camss_subdev_resourc
 	/* VFE5 (lite) */
 	{
 		.regulators = {},
-		.clock = { "cpas_vfe_lite", "vfe_lite_ahb",
+		.clock = { "cpas_ahb", "cpas_vfe_lite", "vfe_lite_ahb",
 			   "vfe_lite_csid", "vfe_lite_cphy_rx",
-			   "vfe_lite"},
+			   "vfe_lite", "camnoc_axi"},
 		.clock_rate = {
-			{ 0, 0, 0, 0  },
+			{ 0 },
+			{ 0 },
 			{ 300000000, 400000000, 400000000, 400000000 },
 			{ 400000000, 400000000, 400000000, 400000000 },
 			{ 400000000, 400000000, 400000000, 400000000 },
 			{ 480000000, 600000000, 600000000, 600000000 },
+			{ 400000000 },
 		},
 		.reg = { "vfe_lite3" },
 		.interrupt = { "vfe_lite3" },
@@ -3018,15 +3026,17 @@ static const struct camss_subdev_resourc
 	/* VFE6 (lite) */
 	{
 		.regulators = {},
-		.clock = { "cpas_vfe_lite", "vfe_lite_ahb",
+		.clock = { "cpas_ahb", "cpas_vfe_lite", "vfe_lite_ahb",
 			   "vfe_lite_csid", "vfe_lite_cphy_rx",
-			   "vfe_lite"},
+			   "vfe_lite", "camnoc_axi"},
 		.clock_rate = {
-			{ 0, 0, 0, 0  },
+			{ 0 },
+			{ 0 },
 			{ 300000000, 400000000, 400000000, 400000000 },
 			{ 400000000, 400000000, 400000000, 400000000 },
 			{ 400000000, 400000000, 400000000, 400000000 },
 			{ 480000000, 600000000, 600000000, 600000000 },
+			{ 400000000 },
 		},
 		.reg = { "vfe_lite4" },
 		.interrupt = { "vfe_lite4" },
author maintainer
#63 lore raw

PATCH 63/188 spi: mxs: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 8b0d0011af20fb547aa67a1cefbf320992fd5e92 upstream.

Make sure to deregister the controller before releasing underlying resources like DMA during driver unbind.

Fixes 33e195acf268 ("spi: mxs: use devm_spi_register_master()")
Cc Jingoo Han <[email protected]>
---
 drivers/spi/spi-mxs.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/spi/spi-mxs.c +7 -1
@@ -619,7 +619,7 @@ static int mxs_spi_probe(struct platform
 	if (ret)
 		goto out_pm_runtime_put;
 
-	ret = devm_spi_register_controller(&pdev->dev, host);
+	ret = spi_register_controller(host);
 	if (ret) {
 		dev_err(&pdev->dev, "Cannot register SPI host, %d\n", ret);
 		goto out_pm_runtime_put;
@@ -650,11 +650,17 @@ static void mxs_spi_remove(struct platfo
 	spi = spi_controller_get_devdata(host);
 	ssp = &spi->ssp;
 
+	spi_controller_get(host);
+
+	spi_unregister_controller(host);
+
 	pm_runtime_disable(&pdev->dev);
 	if (!pm_runtime_status_suspended(&pdev->dev))
 		mxs_spi_runtime_suspend(&pdev->dev);
 
 	dma_release_channel(ssp->dmach);
+
+	spi_controller_put(host);
 }
 
 static struct platform_driver mxs_spi_driver = {
author maintainer
#64 lore raw

PATCH 64/188 spi: mt65xx: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 2ad30599cccc572ba2fc11010670eb6e01ea6bfc upstream.

Make sure to deregister the controller before disabling underlying resources like clocks during driver unbind.

Fixes a568231f4632 ("spi: mediatek: Add spi bus for Mediatek MT8173")
Cc [email protected] # 4.3: ace145802350
Cc Leilk Liu <[email protected]>
---
 drivers/spi/spi-mt65xx.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/spi/spi-mt65xx.c +3 -1
@@ -1326,7 +1326,7 @@ static int mtk_spi_probe(struct platform
 
 	pm_runtime_enable(dev);
 
-	ret = devm_spi_register_controller(dev, host);
+	ret = spi_register_controller(host);
 	if (ret) {
 		pm_runtime_disable(dev);
 		return dev_err_probe(dev, ret, "failed to register host\n");
@@ -1341,6 +1341,8 @@ static void mtk_spi_remove(struct platfo
 	struct mtk_spi *mdata = spi_controller_get_devdata(host);
 	int ret;
 
+	spi_unregister_controller(host);
+
 	cpu_latency_qos_remove_request(&mdata->qos_request);
 	if (mdata->use_spimem && !completion_done(&mdata->spimem_done))
 		complete(&mdata->spimem_done);
author maintainer
#65 lore raw

PATCH 65/188 spi: dln2: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit c353020fbfa8514ee91a6de2d88de4e5edca5803 upstream.

Make sure to deregister the controller before disabling it to allow SPI device drivers to do I/O during deregistration.

Fixes 3d8c0d749da3 ("spi: add support for DLN-2 USB-SPI adapter")
Cc Laurentiu Palcu <[email protected]>
---
 drivers/spi/spi-dln2.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/spi/spi-dln2.c +7 -1
@@ -761,7 +761,7 @@ static int dln2_spi_probe(struct platfor
 	pm_runtime_set_active(&pdev->dev);
 	pm_runtime_enable(&pdev->dev);
 
-	ret = devm_spi_register_controller(&pdev->dev, host);
+	ret = spi_register_controller(host);
 	if (ret < 0) {
 		dev_err(&pdev->dev, "Failed to register host\n");
 		goto exit_register;
@@ -786,10 +786,16 @@ static void dln2_spi_remove(struct platf
 	struct spi_controller *host = platform_get_drvdata(pdev);
 	struct dln2_spi *dln2 = spi_controller_get_devdata(host);
 
+	spi_controller_get(host);
+
+	spi_unregister_controller(host);
+
 	pm_runtime_disable(&pdev->dev);
 
 	if (dln2_spi_enable(dln2, false) < 0)
 		dev_err(&pdev->dev, "Failed to disable SPI module\n");
+
+	spi_controller_put(host);
 }
 
 #ifdef CONFIG_PM_SLEEP
author maintainer
#66 lore raw

PATCH 66/188 spi: s3c64xx: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit c1446b61e472da24d1547525193467b4bea4a7cb upstream.

Make sure to deregister the controller before releasing underlying resources like DMA during driver unbind.

Fixes 91800f0e9005 ("spi/s3c64xx: Use managed registration")
Cc [email protected] # 3.13: 76fbad410c0f
---
 drivers/spi/spi-s3c64xx.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/spi/spi-s3c64xx.c +3 -1
@@ -1370,7 +1370,7 @@ static int s3c64xx_spi_probe(struct plat
 	       S3C64XX_SPI_INT_TX_OVERRUN_EN | S3C64XX_SPI_INT_TX_UNDERRUN_EN,
 	       sdd->regs + S3C64XX_SPI_INT_EN);
 
-	ret = devm_spi_register_controller(&pdev->dev, host);
+	ret = spi_register_controller(host);
 	if (ret != 0) {
 		dev_err(&pdev->dev, "cannot register SPI host: %d\n", ret);
 		goto err_pm_put;
@@ -1400,6 +1400,8 @@ static void s3c64xx_spi_remove(struct pl
 
 	pm_runtime_get_sync(&pdev->dev);
 
+	spi_unregister_controller(host);
+
 	writel(0, sdd->regs + S3C64XX_SPI_INT_EN);
 
 	pm_runtime_put_noidle(&pdev->dev);
author maintainer
#67 lore raw

PATCH 67/188 spi: fsl-espi: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit e506a700a7ad229f5c8f01f4b8350119cccb4158 upstream.

Make sure to deregister the controller before disabling runtime PM (which can leave the controller disabled) to allow SPI device drivers to do I/O during deregistration.

Fixes e9abb4db8d10 ("spi: fsl-espi: add runtime PM")
Cc Heiner Kallweit <[email protected]>
---
 drivers/spi/spi-fsl-espi.c |   10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)
drivers/spi/spi-fsl-espi.c +9 -1
@@ -719,7 +719,7 @@ static int fsl_espi_probe(struct device
 	pm_runtime_enable(dev);
 	pm_runtime_get_sync(dev);
 
-	ret = devm_spi_register_controller(dev, host);
+	ret = spi_register_controller(host);
 	if (ret < 0)
 		goto err_pm;
 
@@ -783,7 +783,15 @@ static int of_fsl_espi_probe(struct plat
 
 static void of_fsl_espi_remove(struct platform_device *dev)
 {
+	struct spi_controller *host = platform_get_drvdata(dev);
+
+	spi_controller_get(host);
+
+	spi_unregister_controller(host);
+
 	pm_runtime_disable(&dev->dev);
+
+	spi_controller_put(host);
 }
 
 #ifdef CONFIG_PM_SLEEP
author maintainer
#68 lore raw

PATCH 68/188 spi: omap2-mcspi: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit fb45f95c377e4a4bdece2c5e17643b459c9c13e7 upstream.

Make sure to deregister the controller before releasing underlying resources like DMA during driver unbind.

---
 drivers/spi/spi-omap2-mcspi.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/spi/spi-omap2-mcspi.c +7 -1
@@ -1585,7 +1585,7 @@ static int omap2_mcspi_probe(struct plat
 	if (status < 0)
 		goto disable_pm;
 
-	status = devm_spi_register_controller(&pdev->dev, ctlr);
+	status = spi_register_controller(ctlr);
 	if (status < 0)
 		goto disable_pm;
 
@@ -1606,11 +1606,17 @@ static void omap2_mcspi_remove(struct pl
 	struct spi_controller *ctlr = platform_get_drvdata(pdev);
 	struct omap2_mcspi *mcspi = spi_controller_get_devdata(ctlr);
 
+	spi_controller_get(ctlr);
+
+	spi_unregister_controller(ctlr);
+
 	omap2_mcspi_release_dma(ctlr);
 
 	pm_runtime_dont_use_autosuspend(mcspi->dev);
 	pm_runtime_put_sync(mcspi->dev);
 	pm_runtime_disable(&pdev->dev);
+
+	spi_controller_put(ctlr);
 }
 
 /* work with hotplug and coldplug */
author maintainer
#69 lore raw

PATCH 69/188 spi: pic32: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 6b627bfe0c44e064aba464839e430dc1ca2b0bb8 upstream.

Make sure to deregister the controller before releasing underlying resources like DMA during driver unbind.

Fixes 1bcb9f8ceb67 ("spi: spi-pic32: Add PIC32 SPI master driver")
Cc Purna Chandra Mandal <[email protected]>
---
 drivers/spi/spi-pic32.c |   11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)
drivers/spi/spi-pic32.c +8 -3
@@ -821,7 +821,7 @@ static int pic32_spi_probe(struct platfo
 	}
 
 	/* register host */
-	ret = devm_spi_register_controller(&pdev->dev, host);
+	ret = spi_register_controller(host);
 	if (ret) {
 		dev_err(&host->dev, "failed registering spi host\n");
 		goto err_bailout;
@@ -840,11 +840,16 @@ err_host:
 
 static void pic32_spi_remove(struct platform_device *pdev)
 {
-	struct pic32_spi *pic32s;
+	struct pic32_spi *pic32s = platform_get_drvdata(pdev);
+
+	spi_controller_get(pic32s->host);
+
+	spi_unregister_controller(pic32s->host);
 
-	pic32s = platform_get_drvdata(pdev);
 	pic32_spi_disable(pic32s);
 	pic32_spi_dma_unprep(pic32s);
+
+	spi_controller_put(pic32s->host);
 }
 
 static const struct of_device_id pic32_spi_of_match[] = {
author maintainer
#70 lore raw

PATCH 70/188 spi: mtk-nor: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 76336f24934621db286cabb20b483773ee01dcaa upstream.

Make sure to deregister the controller before disabling underlying resources like clocks during driver unbind.

Fixes 881d1ee9fe81 ("spi: add support for mediatek spi-nor controller")
Cc Chuanhong Guo <[email protected]>
---
 drivers/spi/spi-mtk-nor.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/spi/spi-mtk-nor.c +3 -1
@@ -914,7 +914,7 @@ static int mtk_nor_probe(struct platform
 	pm_runtime_enable(&pdev->dev);
 	pm_runtime_get_noresume(&pdev->dev);
 
-	ret = devm_spi_register_controller(&pdev->dev, ctlr);
+	ret = spi_register_controller(ctlr);
 	if (ret < 0)
 		goto err_probe;
 
@@ -939,6 +939,8 @@ static void mtk_nor_remove(struct platfo
 	struct spi_controller *ctlr = dev_get_drvdata(&pdev->dev);
 	struct mtk_nor *sp = spi_controller_get_devdata(ctlr);
 
+	spi_unregister_controller(ctlr);
+
 	pm_runtime_disable(&pdev->dev);
 	pm_runtime_set_suspended(&pdev->dev);
 	pm_runtime_dont_use_autosuspend(&pdev->dev);
author maintainer
#71 lore raw

PATCH 71/188 spi: pl022: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 994b33366be9148240690e3e94bffe17c4d89458 upstream.

Make sure to deregister the controller before releasing underlying resources like DMA during driver unbind.

Fixes b43d65f7e818 ("[ARM] 5546/1: ARM PL022 SSP/SPI driver v3")
Cc [email protected] # 2.6.31
Cc Linus Walleij <[email protected]>
---
 drivers/spi/spi-pl022.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/spi/spi-pl022.c +7 -1
@@ -1957,7 +1957,7 @@ static int pl022_probe(struct amba_devic
 
 	/* Register with the SPI framework */
 	amba_set_drvdata(adev, pl022);
-	status = devm_spi_register_controller(&adev->dev, host);
+	status = spi_register_controller(host);
 	if (status != 0) {
 		dev_err_probe(&adev->dev, status,
 			      "problem registering spi host\n");
@@ -1998,6 +1998,10 @@ pl022_remove(struct amba_device *adev)
 	if (!pl022)
 		return;
 
+	spi_controller_get(pl022->host);
+
+	spi_unregister_controller(pl022->host);
+
 	/*
 	 * undo pm_runtime_put() in probe.  I assume that we're not
 	 * accessing the primecell here.
@@ -2009,6 +2013,8 @@ pl022_remove(struct amba_device *adev)
 		pl022_dma_remove(pl022);
 
 	amba_release_regions(adev);
+
+	spi_controller_put(pl022->host);
 }
 
 #ifdef CONFIG_PM_SLEEP
author maintainer
#72 lore raw

PATCH 72/188 spi: ch341: fix devres lifetime

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit abe572f630bc1f0e77041012ab075869036ede4f upstream.

USB drivers bind to USB interfaces and any device managed resources should have their lifetime tied to the interface rather than parent USB device. This avoids issues like memory leaks when drivers are unbound without their devices being physically disconnected (e.g. on probe deferral or configuration changes).

Fix the controller and driver data lifetime so that they are released on driver unbind.

Note that this also makes sure that the SPI controller is placed correctly under the USB interface in the device tree.

Fixes 8846739f52af ("spi: add ch341a usb2spi driver")
Cc Johannes Thumshirn <[email protected]>
---
 drivers/spi/spi-ch341.c |    7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)
drivers/spi/spi-ch341.c +3 -4
@@ -152,7 +152,7 @@ static int ch341_probe(struct usb_interf
 	if (ret)
 		return ret;
 
-	ctrl = devm_spi_alloc_host(&udev->dev, sizeof(struct ch341_spi_dev));
+	ctrl = devm_spi_alloc_host(&intf->dev, sizeof(struct ch341_spi_dev));
 	if (!ctrl)
 		return -ENOMEM;
 
@@ -163,7 +163,7 @@ static int ch341_probe(struct usb_interf
 	ch341->read_pipe = usb_rcvbulkpipe(udev, usb_endpoint_num(in));
 
 	ch341->rx_len = usb_endpoint_maxp(in);
-	ch341->rx_buf = devm_kzalloc(&udev->dev, ch341->rx_len, GFP_KERNEL);
+	ch341->rx_buf = devm_kzalloc(&intf->dev, ch341->rx_len, GFP_KERNEL);
 	if (!ch341->rx_buf)
 		return -ENOMEM;
 
@@ -171,8 +171,7 @@ static int ch341_probe(struct usb_interf
 	if (!ch341->rx_urb)
 		return -ENOMEM;
 
-	ch341->tx_buf =
-		devm_kzalloc(&udev->dev, CH341_PACKET_LENGTH, GFP_KERNEL);
+	ch341->tx_buf = devm_kzalloc(&intf->dev, CH341_PACKET_LENGTH, GFP_KERNEL);
 	if (!ch341->tx_buf) {
 		ret = -ENOMEM;
 		goto err_free_urb;
author maintainer
#73 lore raw

PATCH 73/188 spi: sh-hspi: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit e63982e6392e45a6ecd68d6c317a081cc8e70143 upstream.

Make sure to deregister the controller before releasing underlying resources like clocks during driver unbind.

Fixes 49e599b8595f ("spi: sh-hspi: control spi clock more correctly")
---
 drivers/spi/spi-sh-hspi.c |   10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/spi/spi-sh-hspi.c +8 -2
@@ -258,9 +258,9 @@ static int hspi_probe(struct platform_de
 	ctlr->transfer_one_message = hspi_transfer_one_message;
 	ctlr->bits_per_word_mask = SPI_BPW_MASK(8);
 
-	ret = devm_spi_register_controller(&pdev->dev, ctlr);
+	ret = spi_register_controller(ctlr);
 	if (ret < 0) {
-		dev_err(&pdev->dev, "devm_spi_register_controller error.\n");
+		dev_err(&pdev->dev, "failed to register controller\n");
 		goto error2;
 	}
 
@@ -280,9 +280,15 @@ static void hspi_remove(struct platform_
 {
 	struct hspi_priv *hspi = platform_get_drvdata(pdev);
 
+	spi_controller_get(hspi->ctlr);
+
+	spi_unregister_controller(hspi->ctlr);
+
 	pm_runtime_disable(&pdev->dev);
 
 	clk_put(hspi->clk);
+
+	spi_controller_put(hspi->ctlr);
 }
 
 static const struct of_device_id hspi_of_match[] = {
author maintainer
#74 lore raw

PATCH 74/188 spi: fsl: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 9b7abfed4c3754062d1f3ffd452e65a38667f586 upstream.

Make sure to deregister the controller before releasing underlying resources like DMA during driver unbind.

Fixes 4178b6b1b595 ("spi: fsl-(e)spi: migrate to using devm_ functions to simplify cleanup")
Cc Heiner Kallweit <[email protected]>
---
 drivers/spi/spi-fsl-spi.c |   14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)
drivers/spi/spi-fsl-spi.c +13 -1
@@ -614,7 +614,7 @@ static struct spi_controller *fsl_spi_pr
 
 	mpc8xxx_spi_write_reg(&reg_base->mode, regval);
 
-	ret = devm_spi_register_controller(dev, host);
+	ret = spi_register_controller(host);
 	if (ret < 0)
 		goto err_probe;
 
@@ -705,7 +705,13 @@ static void of_fsl_spi_remove(struct pla
 	struct spi_controller *host = platform_get_drvdata(ofdev);
 	struct mpc8xxx_spi *mpc8xxx_spi = spi_controller_get_devdata(host);
 
+	spi_controller_get(host);
+
+	spi_unregister_controller(host);
+
 	fsl_spi_cpm_free(mpc8xxx_spi);
+
+	spi_controller_put(host);
 }
 
 static struct platform_driver of_fsl_spi_driver = {
@@ -751,7 +757,13 @@ static void plat_mpc8xxx_spi_remove(stru
 	struct spi_controller *host = platform_get_drvdata(pdev);
 	struct mpc8xxx_spi *mpc8xxx_spi = spi_controller_get_devdata(host);
 
+	spi_controller_get(host);
+
+	spi_unregister_controller(host);
+
 	fsl_spi_cpm_free(mpc8xxx_spi);
+
+	spi_controller_put(host);
 }
 
 MODULE_ALIAS("platform:mpc8xxx_spi");
author maintainer
#75 lore raw

PATCH 75/188 spi: bcmbca-hsspi: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit c3d97c3320b9a1ebbd6119857341be034f7b3efc upstream.

Make sure to deregister the controller before disabling underlying resources like interrupts during driver unbind to allow SPI drivers to do I/O during deregistration.

Note that clocks were also disabled before the recent commit e532e21a246d ("spi: bcm63xx-hsspi: Simplify clock handling with devm_clk_get_enabled()").

Fixes a38a2233f23b ("spi: bcmbca-hsspi: Add driver for newer HSSPI controller")
Cc [email protected] # 6.3: deb269e0394f
Cc William Zhang <[email protected]>
---
 drivers/spi/spi-bcmbca-hsspi.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/spi/spi-bcmbca-hsspi.c +3 -1
@@ -550,7 +550,7 @@ static int bcmbca_hsspi_probe(struct pla
 	}
 
 	/* register and we are done */
-	ret = devm_spi_register_controller(dev, host);
+	ret = spi_register_controller(host);
 	if (ret)
 		goto out_sysgroup_disable;
 
@@ -572,6 +572,8 @@ static void bcmbca_hsspi_remove(struct p
 	struct spi_controller *host = platform_get_drvdata(pdev);
 	struct bcmbca_hsspi *bs = spi_controller_get_devdata(host);
 
+	spi_unregister_controller(host);
+
 	/* reset the hardware and block queue progress */
 	__raw_writel(0, bs->regs + HSSPI_INT_MASK_REG);
 	clk_disable_unprepare(bs->pll_clk);
author maintainer
#76 lore raw

PATCH 76/188 spi: coldfire-qspi: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit e7c510e192ff2a1264d999575eea39a506424264 upstream.

Make sure to deregister the controller before disabling underlying resources like clocks (via runtime pm) during driver unbind.

Fixes 34b8c6617366 ("spi: Add Freescale/Motorola Coldfire QSPI driver")
Cc [email protected] # 2.6.34
Cc Steven King <[email protected]>
---
 drivers/spi/spi-coldfire-qspi.c |   10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/spi/spi-coldfire-qspi.c +8 -2
@@ -410,9 +410,9 @@ static int mcfqspi_probe(struct platform
 	platform_set_drvdata(pdev, host);
 	pm_runtime_enable(&pdev->dev);
 
-	status = devm_spi_register_controller(&pdev->dev, host);
+	status = spi_register_controller(host);
 	if (status) {
-		dev_dbg(&pdev->dev, "devm_spi_register_controller failed\n");
+		dev_dbg(&pdev->dev, "failed to register controller\n");
 		goto fail1;
 	}
 
@@ -436,11 +436,17 @@ static void mcfqspi_remove(struct platfo
 	struct spi_controller *host = platform_get_drvdata(pdev);
 	struct mcfqspi *mcfqspi = spi_controller_get_devdata(host);
 
+	spi_controller_get(host);
+
+	spi_unregister_controller(host);
+
 	pm_runtime_disable(&pdev->dev);
 	/* disable the hardware (set the baud rate to 0) */
 	mcfqspi_wr_qmr(mcfqspi, MCFQSPI_QMR_MSTR);
 
 	mcfqspi_cs_teardown(mcfqspi);
+
+	spi_controller_put(host);
 }
 
 #ifdef CONFIG_PM_SLEEP
author maintainer
#77 lore raw

PATCH 77/188 spi: npcm-pspi: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit ebd81199e00e107980bf8c4d2c747ae50158f797 upstream.

Make sure to deregister the controller before disabling underlying resources like clocks during driver unbind.

Fixes 2a22f1b30cee ("spi: npcm: add NPCM PSPI controller driver")
Cc Tomer Maimon <[email protected]>
---
 drivers/spi/spi-npcm-pspi.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/spi/spi-npcm-pspi.c +7 -1
@@ -414,7 +414,7 @@ static int npcm_pspi_probe(struct platfo
 	/* set to default clock rate */
 	npcm_pspi_set_baudrate(priv, NPCM_PSPI_DEFAULT_CLK);
 
-	ret = devm_spi_register_controller(&pdev->dev, host);
+	ret = spi_register_controller(host);
 	if (ret)
 		goto out_disable_clk;
 
@@ -435,8 +435,14 @@ static void npcm_pspi_remove(struct plat
 	struct spi_controller *host = platform_get_drvdata(pdev);
 	struct npcm_pspi *priv = spi_controller_get_devdata(host);
 
+	spi_controller_get(host);
+
+	spi_unregister_controller(host);
+
 	npcm_pspi_reset_hw(priv);
 	clk_disable_unprepare(priv->clk);
+
+	spi_controller_put(host);
 }
 
 static const struct of_device_id npcm_pspi_match[] = {
author maintainer
#78 lore raw

PATCH 78/188 spi: cavium-thunderx: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit dbb6b01267c0c866eaac4019cec19f414beec61d upstream.

Make sure to deregister the controller before disabling it to avoid hanging or leaking resources associated with the queue when the queue non-empty.

Fixes 7347a6c7af8d ("spi: octeon: Add ThunderX driver")
Cc Jan Glauber <[email protected]>
---
 drivers/spi/spi-cavium-thunderx.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/spi/spi-cavium-thunderx.c +7 -1
@@ -71,7 +71,7 @@ static int thunderx_spi_probe(struct pci
 
 	pci_set_drvdata(pdev, host);
 
-	ret = devm_spi_register_controller(dev, host);
+	ret = spi_register_controller(host);
 	if (ret)
 		goto error;
 
@@ -91,8 +91,14 @@ static void thunderx_spi_remove(struct p
 	if (!p)
 		return;
 
+	spi_controller_get(host);
+
+	spi_unregister_controller(host);
+
 	/* Put everything in a known state. */
 	writeq(0, p->register_base + OCTEON_SPI_CFG(p));
+
+	spi_controller_put(host);
 }
 
 static const struct pci_device_id thunderx_spi_pci_id_table[] = {
author maintainer
#79 lore raw

PATCH 79/188 spi: pic32-sqi: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 420df79d1a618951eb0eb4331df95c9f4f763b8b upstream.

Make sure to deregister the controller before releasing underlying resources like DMA during driver unbind.

Fixes 3270ac230f66 ("spi: pic32-sqi: add SPI driver for PIC32 SQI controller.")
Cc Purna Chandra Mandal <[email protected]>
---
 drivers/spi/spi-pic32-sqi.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/spi/spi-pic32-sqi.c +7 -1
@@ -642,7 +642,7 @@ static int pic32_sqi_probe(struct platfo
 	host->prepare_transfer_hardware	= pic32_sqi_prepare_hardware;
 	host->unprepare_transfer_hardware	= pic32_sqi_unprepare_hardware;
 
-	ret = devm_spi_register_controller(&pdev->dev, host);
+	ret = spi_register_controller(host);
 	if (ret) {
 		dev_err(&host->dev, "failed registering spi host\n");
 		free_irq(sqi->irq, sqi);
@@ -665,9 +665,15 @@ static void pic32_sqi_remove(struct plat
 {
 	struct pic32_sqi *sqi = platform_get_drvdata(pdev);
 
+	spi_controller_get(sqi->host);
+
+	spi_unregister_controller(sqi->host);
+
 	/* release resources */
 	free_irq(sqi->irq, sqi);
 	ring_desc_ring_free(sqi);
+
+	spi_controller_put(sqi->host);
 }
 
 static const struct of_device_id pic32_sqi_of_ids[] = {
author maintainer
#80 lore raw

PATCH 80/188 spi: sprd: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 123d17dbc5f07059752fa5e616385ca29a8f935a upstream.

Make sure to deregister the controller before disabling underlying resources like clocks during driver unbind.

Note that the controller is suspended before disabling and releasing resources since commit de082d866cce ("spi: sprd: Add the SPI irq function for the SPI DMA mode") which avoids issues like unclocked accesses but prevents SPI device drivers from doing I/O during deregistration.

Fixes e7d973a31c24 ("spi: sprd: Add SPI driver for Spreadtrum SC9860")
Cc Lanqing Liu <[email protected]>
---
 drivers/spi/spi-sprd.c |    8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
drivers/spi/spi-sprd.c +6 -2
@@ -978,7 +978,7 @@ static int sprd_spi_probe(struct platfor
 		goto err_rpm_put;
 	}
 
-	ret = devm_spi_register_controller(&pdev->dev, sctlr);
+	ret = spi_register_controller(sctlr);
 	if (ret)
 		goto err_rpm_put;
 
@@ -1009,7 +1009,9 @@ static void sprd_spi_remove(struct platf
 	if (ret < 0)
 		dev_err(ss->dev, "failed to resume SPI controller\n");
 
-	spi_controller_suspend(sctlr);
+	spi_controller_get(sctlr);
+
+	spi_unregister_controller(sctlr);
 
 	if (ret >= 0) {
 		if (ss->dma.enable)
@@ -1018,6 +1020,8 @@ static void sprd_spi_remove(struct platf
 	}
 	pm_runtime_put_noidle(&pdev->dev);
 	pm_runtime_disable(&pdev->dev);
+
+	spi_controller_put(sctlr);
 }
 
 static int __maybe_unused sprd_spi_runtime_suspend(struct device *dev)
author maintainer
#81 lore raw

PATCH 81/188 spi: rspi: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 9944fa6726afb1e6eb7e2212764e7da0c97f2dcc upstream.

Make sure to deregister the controller before releasing underlying resources like DMA during driver unbind.

Fixes 9e03d05eee4c ("spi: rcar: Use devm_spi_register_master()")
Cc Jingoo Han <[email protected]>
---
 drivers/spi/spi-rspi.c |   10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/spi/spi-rspi.c +8 -2
@@ -1171,8 +1171,14 @@ static void rspi_remove(struct platform_
 {
 	struct rspi_data *rspi = platform_get_drvdata(pdev);
 
+	spi_controller_get(rspi->ctlr);
+
+	spi_unregister_controller(rspi->ctlr);
+
 	rspi_release_dma(rspi->ctlr);
 	pm_runtime_disable(&pdev->dev);
+
+	spi_controller_put(rspi->ctlr);
 }
 
 static const struct spi_ops rspi_ops = {
@@ -1377,9 +1383,9 @@ static int rspi_probe(struct platform_de
 	if (ret < 0)
 		dev_warn(&pdev->dev, "DMA not available, using PIO\n");
 
-	ret = devm_spi_register_controller(&pdev->dev, ctlr);
+	ret = spi_register_controller(ctlr);
 	if (ret < 0) {
-		dev_err(&pdev->dev, "devm_spi_register_controller error.\n");
+		dev_err(&pdev->dev, "failed to register controller\n");
 		goto error3;
 	}
 
author maintainer
#82 lore raw

PATCH 82/188 spi: sh-msiof: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 45170f67a08b912ead6ccc387ba06954d1d4e53a upstream.

Make sure to deregister the controller before releasing underlying resources like DMA during driver unbind.

Fixes 1bd6363bc0c6 ("spi: sh-msiof: Use core message handling instead of spi-bitbang")
Cc Geert Uytterhoeven <[email protected]>
---
 drivers/spi/spi-sh-msiof.c |   10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/spi/spi-sh-msiof.c +8 -2
@@ -1290,9 +1290,9 @@ static int sh_msiof_spi_probe(struct pla
 	if (ret < 0)
 		dev_warn(dev, "DMA not available, using PIO\n");
 
-	ret = devm_spi_register_controller(dev, ctlr);
+	ret = spi_register_controller(ctlr);
 	if (ret < 0) {
-		dev_err(dev, "devm_spi_register_controller error.\n");
+		dev_err(dev, "failed to register controller\n");
 		goto err2;
 	}
 
@@ -1310,8 +1310,14 @@ static void sh_msiof_spi_remove(struct p
 {
 	struct sh_msiof_spi_priv *p = platform_get_drvdata(pdev);
 
+	spi_controller_get(p->ctlr);
+
+	spi_unregister_controller(p->ctlr);
+
 	sh_msiof_release_dma(p);
 	pm_runtime_disable(&pdev->dev);
+
+	spi_controller_put(p->ctlr);
 }
 
 static const struct platform_device_id spi_driver_ids[] = {
author maintainer
#83 lore raw

PATCH 83/188 spi: slave-mt27xx: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit ab840cbda4fe6c40e52f6415c47056797c663bb2 upstream.

Make sure to deregister the controller before disabling underlying resources like clocks (by disabling runtime PM) during driver unbind.

Fixes 805be7ddf367 ("spi: mediatek: add spi slave for Mediatek MT2712")
Cc Leilk Liu <[email protected]>
---
 drivers/spi/spi-slave-mt27xx.c |   10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)
drivers/spi/spi-slave-mt27xx.c +9 -1
@@ -454,7 +454,7 @@ static int mtk_spi_slave_probe(struct pl
 
 	pm_runtime_enable(&pdev->dev);
 
-	ret = devm_spi_register_controller(&pdev->dev, ctlr);
+	ret = spi_register_controller(ctlr);
 	clk_disable_unprepare(mdata->spi_clk);
 	if (ret) {
 		dev_err(&pdev->dev,
@@ -474,7 +474,15 @@ err_put_ctlr:
 
 static void mtk_spi_slave_remove(struct platform_device *pdev)
 {
+	struct spi_controller *ctlr = platform_get_drvdata(pdev);
+
+	spi_controller_get(ctlr);
+
+	spi_unregister_controller(ctlr);
+
 	pm_runtime_disable(&pdev->dev);
+
+	spi_controller_put(ctlr);
 }
 
 #ifdef CONFIG_PM_SLEEP
author maintainer
#84 lore raw

PATCH 84/188 spi: img-spfi: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit fc3a83b0d9c16b941c9028f5a8db9541dce4ddf2 upstream.

Make sure to deregister the controller before disabling and releasing underlying resources like clocks and DMA during driver unbind.

Fixes deba25800a12 ("spi: Add driver for IMG SPFI controller")
Cc Andrew Bresticker <[email protected]>
---
 drivers/spi/spi-img-spfi.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/spi/spi-img-spfi.c +7 -1
@@ -644,7 +644,7 @@ static int img_spfi_probe(struct platfor
 	pm_runtime_set_active(spfi->dev);
 	pm_runtime_enable(spfi->dev);
 
-	ret = devm_spi_register_controller(spfi->dev, host);
+	ret = spi_register_controller(host);
 	if (ret)
 		goto disable_pm;
 
@@ -670,6 +670,10 @@ static void img_spfi_remove(struct platf
 	struct spi_controller *host = platform_get_drvdata(pdev);
 	struct img_spfi *spfi = spi_controller_get_devdata(host);
 
+	spi_controller_get(host);
+
+	spi_unregister_controller(host);
+
 	if (spfi->tx_ch)
 		dma_release_channel(spfi->tx_ch);
 	if (spfi->rx_ch)
@@ -680,6 +684,8 @@ static void img_spfi_remove(struct platf
 		clk_disable_unprepare(spfi->spfi_clk);
 		clk_disable_unprepare(spfi->sys_clk);
 	}
+
+	spi_controller_put(host);
 }
 
 #ifdef CONFIG_PM
author maintainer
#85 lore raw

PATCH 85/188 spi: mpfs: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 573c7db8fce91a1b07dd64a260bb44b9e6d05943 upstream.

Make sure to deregister the controller before disabling underlying resources like interrupts during driver unbind.

Fixes 9ac8d17694b6 ("spi: add support for microchip fpga spi controllers")
Cc Conor Dooley <[email protected]>
---
 drivers/spi/spi-mpfs.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/spi/spi-mpfs.c +3 -1
@@ -575,7 +575,7 @@ static int mpfs_spi_probe(struct platfor
 
 	mpfs_spi_init(host, spi);
 
-	ret = devm_spi_register_controller(&pdev->dev, host);
+	ret = spi_register_controller(host);
 	if (ret) {
 		mpfs_spi_disable_ints(spi);
 		mpfs_spi_disable(spi);
@@ -593,6 +593,8 @@ static void mpfs_spi_remove(struct platf
 	struct spi_controller *host  = platform_get_drvdata(pdev);
 	struct mpfs_spi *spi = spi_controller_get_devdata(host);
 
+	spi_unregister_controller(host);
+
 	mpfs_spi_disable_ints(spi);
 	mpfs_spi_disable(spi);
 }
author maintainer
#86 lore raw

PATCH 86/188 spi: imx: fix runtime pm leak on probe deferral

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit a1d50a37d3b1df84f536a982f692371039df4a48 upstream.

Make sure to balance the runtime PM usage count before returning on probe failure (e.g. probe deferral) so that the controller can be suspended when a driver is later bound.

Fixes 43b6bf406cd0 ("spi: imx: fix runtime pm support for !CONFIG_PM")
Cc Sascha Hauer <[email protected]>
---
 drivers/spi/spi-imx.c |    1 +
 1 file changed, 1 insertion(+)
drivers/spi/spi-imx.c +1 -0
@@ -1948,6 +1948,7 @@ out_register_controller:
 out_runtime_pm_put:
 	pm_runtime_dont_use_autosuspend(spi_imx->dev);
 	pm_runtime_disable(spi_imx->dev);
+	pm_runtime_put_noidle(spi_imx->dev);
 	pm_runtime_set_suspended(&pdev->dev);
 
 	clk_disable_unprepare(spi_imx->clk_ipg);
author maintainer
#87 lore raw

PATCH 87/188 spi: mxic: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit adbc595e272052181d40ec307a4c5ba98571b0fe upstream.

Make sure to deregister the controller before disabling underlying resources like clocks (via runtime pm) during driver unbind.

Fixes b942d80b0a39 ("spi: Add MXIC controller driver")
Cc [email protected] # 5.0: cc53711b2191
Cc Mason Yang <[email protected]>
---
 drivers/spi/spi-mxic.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/spi/spi-mxic.c +2 -1
@@ -833,9 +833,10 @@ static void mxic_spi_remove(struct platf
 	struct spi_controller *host = platform_get_drvdata(pdev);
 	struct mxic_spi *mxic = spi_controller_get_devdata(host);
 
+	spi_unregister_controller(host);
+
 	pm_runtime_disable(&pdev->dev);
 	mxic_spi_mem_ecc_remove(mxic);
-	spi_unregister_controller(host);
 }
 
 static const struct of_device_id mxic_spi_of_ids[] = {
author maintainer
#88 lore raw

PATCH 88/188 spi: orion: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 220f4f11104a7f83b71543ef0e48dde1da2bc5d3 upstream.

Make sure to deregister the controller before disabling underlying resources like clocks during driver unbind.

---
 drivers/spi/spi-orion.c |    7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
drivers/spi/spi-orion.c +6 -1
@@ -802,10 +802,15 @@ static void orion_spi_remove(struct plat
 	struct spi_controller *host = platform_get_drvdata(pdev);
 	struct orion_spi *spi = spi_controller_get_devdata(host);
 
+	spi_controller_get(host);
+
+	spi_unregister_controller(host);
+
 	pm_runtime_get_sync(&pdev->dev);
 	clk_disable_unprepare(spi->axi_clk);
 
-	spi_unregister_controller(host);
+	spi_controller_put(host);
+
 	pm_runtime_disable(&pdev->dev);
 }
 
author maintainer
#89 lore raw

PATCH 89/188 spi: orion: fix runtime pm leak on unbind

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 97b17dd8266d2e26d9ee3c75a0fa34ecde6944f0 upstream.

Make sure to balance the runtime PM usage count on driver unbind so that the controller can be suspended when a driver is rebound.

Also restore the autosuspend setting.

This issue was flagged by Sashiko when reviewing a controller deregistration fix.

---
 drivers/spi/spi-orion.c |    3 +++
 1 file changed, 3 insertions(+)
drivers/spi/spi-orion.c +3 -0
@@ -812,6 +812,9 @@ static void orion_spi_remove(struct plat
 	spi_controller_put(host);
 
 	pm_runtime_disable(&pdev->dev);
+	pm_runtime_put_noidle(&pdev->dev);
+	pm_runtime_set_suspended(&pdev->dev);
+	pm_runtime_dont_use_autosuspend(&pdev->dev);
 }
 
 MODULE_ALIAS("platform:" DRIVER_NAME);
author maintainer
#90 lore raw

PATCH 90/188 spi: orion: fix clock imbalance on registration failure

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 443cde0dc59c5d154156ac9f27a7dadef8ebc0c2 upstream.

Make sure that the controller is not runtime suspended before disabling clocks on probe failure.

Also restore the autosuspend setting.

Fixes 5c6786945b4e ("spi: spi-orion: add runtime PM support")
Cc Russell King <[email protected]>
---
 drivers/spi/spi-orion.c |    6 ++++++
 1 file changed, 6 insertions(+)
drivers/spi/spi-orion.c +6 -0
@@ -774,6 +774,7 @@ static int orion_spi_probe(struct platfo
 	pm_runtime_set_active(&pdev->dev);
 	pm_runtime_use_autosuspend(&pdev->dev);
 	pm_runtime_set_autosuspend_delay(&pdev->dev, SPI_AUTOSUSPEND_TIMEOUT);
+	pm_runtime_get_noresume(&pdev->dev);
 	pm_runtime_enable(&pdev->dev);
 
 	status = orion_spi_reset(spi);
@@ -785,10 +786,15 @@ static int orion_spi_probe(struct platfo
 	if (status < 0)
 		goto out_rel_pm;
 
+	pm_runtime_put_autosuspend(&pdev->dev);
+
 	return status;
 
 out_rel_pm:
 	pm_runtime_disable(&pdev->dev);
+	pm_runtime_put_noidle(&pdev->dev);
+	pm_runtime_set_suspended(&pdev->dev);
+	pm_runtime_dont_use_autosuspend(&pdev->dev);
 out_rel_axi_clk:
 	clk_disable_unprepare(spi->axi_clk);
 out:
author maintainer
#91 lore raw

PATCH 91/188 spi: mpc52xx: fix use-after-free on registration failure

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit f62c060272b9d7423b1650b844e8e4e7b8f9f925 upstream.

Make sure to disable and free the interrupts in case controller registration fails to avoid a potential use-after-free and resource leak.

This issue was flagged by Sashiko when reviewing a controller deregistration fix.

---
 drivers/spi/spi-mpc52xx.c |    3 +++
 1 file changed, 3 insertions(+)
drivers/spi/spi-mpc52xx.c +3 -0
@@ -501,6 +501,9 @@ static int mpc52xx_spi_probe(struct plat
 
  err_register:
 	dev_err(&ms->host->dev, "initialization failed\n");
+	free_irq(ms->irq0, ms);
+	free_irq(ms->irq1, ms);
+	cancel_work_sync(&ms->work);
  err_gpio:
 	while (i-- > 0)
 		gpiod_put(ms->gpio_cs[i]);
author maintainer
#92 lore raw

PATCH 92/188 spi: mpc52xx: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 0f997fdae819a8c2cc83bd4ff7d935ad76c727c9 upstream.

Make sure to deregister the controller before disabling and releasing underlying resources like interrupts and gpios during driver unbind.

Fixes 42bbb70980f3 ("powerpc/5200: Add mpc5200-spi (non-PSC) device driver")
Fixes b8d4e2ce60b6 ("mpc52xx_spi: add gpio chipselect")
Cc [email protected] # 2.6.33
Cc Grant Likely <[email protected]>
Cc Luotao Fu <[email protected]>
---
 drivers/spi/spi-mpc52xx.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/spi/spi-mpc52xx.c +2 -1
@@ -523,6 +523,8 @@ static void mpc52xx_spi_remove(struct pl
 	struct mpc52xx_spi *ms = spi_controller_get_devdata(host);
 	int i;
 
+	spi_unregister_controller(host);
+
 	cancel_work_sync(&ms->work);
 	free_irq(ms->irq0, ms);
 	free_irq(ms->irq1, ms);
@@ -531,7 +533,6 @@ static void mpc52xx_spi_remove(struct pl
 		gpiod_put(ms->gpio_cs[i]);
 
 	kfree(ms->gpio_cs);
-	spi_unregister_controller(host);
 	iounmap(ms->regs);
 	spi_controller_put(host);
 }
author maintainer
#93 lore raw

PATCH 93/188 spi: mpc52xx: fix use-after-free on unbind

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 706b3dc2ac7a998c55e14b3fd2e8f934c367e6e0 upstream.

The state machine work is scheduled by the interrupt handler and therefore needs to be cancelled after disabling interrupts to avoid a potential use-after-free.

Fixes 984836621aad ("spi: mpc52xx: Add cancel_work_sync before module remove")
Cc Pei Xiao <[email protected]>
---
 drivers/spi/spi-mpc52xx.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/spi/spi-mpc52xx.c +2 -1
@@ -525,10 +525,11 @@ static void mpc52xx_spi_remove(struct pl
 
 	spi_unregister_controller(host);
 
-	cancel_work_sync(&ms->work);
 	free_irq(ms->irq0, ms);
 	free_irq(ms->irq1, ms);
 
+	cancel_work_sync(&ms->work);
+
 	for (i = 0; i < ms->gpio_cs_count; i++)
 		gpiod_put(ms->gpio_cs[i]);
 
author maintainer
#94 lore raw

PATCH 94/188 spi: cadence: fix controller deregistration

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 666fa7e9ca98e71c880086ca24147ae843f1ed6e upstream.

Make sure to deregister the controller before disabling underlying resources like clocks during driver unbind.

Fixes c474b3866546 ("spi: Add driver for Cadence SPI controller")
Cc Harini Katakam <[email protected]>
---
 drivers/spi/spi-cadence.c |    6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/spi/spi-cadence.c +5 -1
@@ -698,6 +698,10 @@ static void cdns_spi_remove(struct platf
 	struct spi_controller *ctlr = platform_get_drvdata(pdev);
 	struct cdns_spi *xspi = spi_controller_get_devdata(ctlr);
 
+	spi_controller_get(ctlr);
+
+	spi_unregister_controller(ctlr);
+
 	cdns_spi_write(xspi, CDNS_SPI_ER, CDNS_SPI_ER_DISABLE);
 
 	if (!spi_controller_is_target(ctlr)) {
@@ -705,7 +709,7 @@ static void cdns_spi_remove(struct platf
 		pm_runtime_set_suspended(&pdev->dev);
 	}
 
-	spi_unregister_controller(ctlr);
+	spi_controller_put(ctlr);
 }
 
 /**
author maintainer
#95 lore raw

PATCH 95/188 spi: cadence: fix unclocked access on unbind

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit 5b1689a41f02955c5361944f748a4812a6ff9307 upstream.

Make sure that the controller is runtime resumed before disabling it during driver unbind to avoid unclocked register access and unbalanced clock disable.

Also restore the autosuspend setting.

This issue was flagged by Sashiko when reviewing a controller deregistration fix.

---
 drivers/spi/spi-cadence.c |    9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
drivers/spi/spi-cadence.c +8 -1
@@ -697,16 +697,23 @@ static void cdns_spi_remove(struct platf
 {
 	struct spi_controller *ctlr = platform_get_drvdata(pdev);
 	struct cdns_spi *xspi = spi_controller_get_devdata(ctlr);
+	int ret = 0;
+
+	if (!spi_controller_is_target(ctlr))
+		ret = pm_runtime_get_sync(&pdev->dev);
 
 	spi_controller_get(ctlr);
 
 	spi_unregister_controller(ctlr);
 
-	cdns_spi_write(xspi, CDNS_SPI_ER, CDNS_SPI_ER_DISABLE);
+	if (ret >= 0)
+		cdns_spi_write(xspi, CDNS_SPI_ER, CDNS_SPI_ER_DISABLE);
 
 	if (!spi_controller_is_target(ctlr)) {
 		pm_runtime_disable(&pdev->dev);
 		pm_runtime_set_suspended(&pdev->dev);
+		pm_runtime_put_noidle(&pdev->dev);
+		pm_runtime_dont_use_autosuspend(&pdev->dev);
 	}
 
 	spi_controller_put(ctlr);
author maintainer
#96 lore raw

PATCH 96/188 spi: cadence: fix clock imbalance on probe failure

In reply to Greg Kroah-Hartman

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

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

From: Johan Hovold <[email protected]>

commit ecea4f0e9db2fb6ab4a68a59c5aba0d8f59a9566 upstream.

Make sure that the controller is active before disabling clocks on probe failure to avoid unbalanced clock disable.

Also drop the usage count before returning (so that the controller can be suspended after a probe deferral) and restore the autosuspend setting.

Fixes d36ccd9f7ea4 ("spi: cadence: Runtime pm adaptation")
Cc Shubhrajyoti Datta <[email protected]>
---
 drivers/spi/spi-cadence.c |    6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/spi/spi-cadence.c +5 -1
@@ -662,7 +662,6 @@ static int cdns_spi_probe(struct platfor
 		/* Set to default valid value */
 		ctlr->max_speed_hz = xspi->clk_rate / 4;
 		xspi->speed_hz = ctlr->max_speed_hz;
-		pm_runtime_put_autosuspend(&pdev->dev);
 	} else {
 		ctlr->mode_bits |= SPI_NO_CS;
 		ctlr->target_abort = cdns_target_abort;
@@ -673,12 +672,17 @@ static int cdns_spi_probe(struct platfor
 		goto clk_dis_all;
 	}
 
+	if (!spi_controller_is_target(ctlr))
+		pm_runtime_put_autosuspend(&pdev->dev);
+
 	return ret;
 
 clk_dis_all:
 	if (!spi_controller_is_target(ctlr)) {
 		pm_runtime_disable(&pdev->dev);
 		pm_runtime_set_suspended(&pdev->dev);
+		pm_runtime_put_noidle(&pdev->dev);
+		pm_runtime_dont_use_autosuspend(&pdev->dev);
 	}
 remove_ctlr:
 	spi_controller_put(ctlr);
author maintainer
#97 lore raw

PATCH 97/188 drm/msm/gem: fix error handling in msm_ioctl_gem_info_get_metadata()

In reply to Greg Kroah-Hartman

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

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

From: Yasuaki Torimaru <[email protected]>

commit 47cbfe2608314b833ad61a65827d8fb363bc2d2d upstream.

msm_ioctl_gem_info_get_metadata() always returns 0 regardless of errors. When copy_to_user() fails or the user buffer is too small, the error code stored in ret is ignored because the function unconditionally returns 0. This causes userspace to believe the ioctl succeeded when it did not.

Additionally, kmemdup() can return NULL on allocation failure, but the return value is not checked. This leads to a NULL pointer dereference in the subsequent copy_to_user() call.

Add the missing NULL check for kmemdup() and return ret instead of 0.

Note that the SET counterpart (msm_ioctl_gem_info_set_metadata) correctly returns ret.

Fixes 9902cb999e4e ("drm/msm/gem: Add metadata")

Patchwork: https://patchwork.freedesktop.org/patch/714478/
Message-ID: <[email protected]>

---
 drivers/gpu/drm/msm/msm_drv.c |    7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
drivers/gpu/drm/msm/msm_drv.c +6 -1
@@ -536,6 +536,11 @@ static int msm_ioctl_gem_info_get_metada
 	len = msm_obj->metadata_size;
 	buf = kmemdup(msm_obj->metadata, len, GFP_KERNEL);
 
+	if (!buf) {
+		msm_gem_unlock(obj);
+		return -ENOMEM;
+	}
+
 	msm_gem_unlock(obj);
 
 	if (*metadata_size < len) {
@@ -548,7 +553,7 @@ static int msm_ioctl_gem_info_get_metada
 
 	kfree(buf);
 
-	return 0;
+	return ret;
 }
 
 static int msm_ioctl_gem_info(struct drm_device *dev, void *data,
author maintainer
#98 lore raw

PATCH 98/188 drm/imx: parallel-display: Prefer bus format set via legacy "interface-pix-fmt" DT property

In reply to Greg Kroah-Hartman

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

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

From: Marek Vasut <[email protected]>

commit cdf26e1462c220629bb79d487263b66f8b679eab upstream.

Prefer bus format set via legacy "interface-pix-fmt" DT property over panel bus format. This is necessary to retain support for DTs which configure the IPUv3 parallel output as 24bit DPI, but connect 18bit DPI panels to it with hardware swizzling.

This used to work up to Linux 6.12, but stopped working in 6.13, reinstate the behavior to support old DTs.

---
 drivers/gpu/drm/imx/ipuv3/parallel-display.c |   15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)
drivers/gpu/drm/imx/ipuv3/parallel-display.c +12 -3
@@ -110,8 +110,7 @@ imx_pd_bridge_atomic_get_input_bus_fmts(
 		output_fmt = imxpd->bus_format ? : MEDIA_BUS_FMT_RGB888_1X24;
 
 	/* Now make sure the requested output format is supported. */
-	if ((imxpd->bus_format && imxpd->bus_format != output_fmt) ||
-	    !imx_pd_format_supported(output_fmt)) {
+	if (!imx_pd_format_supported(output_fmt)) {
 		*num_input_fmts = 0;
 		return NULL;
 	}
@@ -121,7 +120,17 @@ imx_pd_bridge_atomic_get_input_bus_fmts(
 	if (!input_fmts)
 		return NULL;
 
-	input_fmts[0] = output_fmt;
+	/*
+	 * Prefer bus format set via legacy "interface-pix-fmt" DT property
+	 * over panel bus format. This is necessary to retain support for
+	 * DTs which configure the IPUv3 parallel output as 24bit, but
+	 * connect 18bit DPI panels to it with hardware swizzling.
+	 */
+	if (imxpd->bus_format && imxpd->bus_format != output_fmt)
+		input_fmts[0] = imxpd->bus_format;
+	else
+		input_fmts[0] = output_fmt;
+
 	return input_fmts;
 }
 
author maintainer
#99 lore raw

PATCH 99/188 drm/msm: always recover the gpu

In reply to Greg Kroah-Hartman

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

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

From: Anna Maniscalco <[email protected]>

commit 01a0d6cd7032e9993feea19fadb03ef9d5b488f2 upstream.

Previously, in case there was no more work to do, recover worker wouldn't trigger recovery and would instead rely on the gpu going to sleep and then resuming when more work is submitted.

Recover_worker will first increment the fence of the hung ring so, if there's only one job submitted to a ring and that causes an hang, it will early out.

There's no guarantee that the gpu will suspend and resume before more work is submitted and if the gpu is in a hung state it will stay in that state and probably trigger a timeout again.

Just stop checking and always recover the gpu.

Patchwork: https://patchwork.freedesktop.org/patch/704066/
Message-ID: <[email protected]>

---
 drivers/gpu/drm/msm/msm_gpu.c |   42 ++++++++++++++++++++----------------------
 1 file changed, 20 insertions(+), 22 deletions(-)
drivers/gpu/drm/msm/msm_gpu.c +20 -22
@@ -548,32 +548,30 @@ static void recover_worker(struct kthrea
 		msm_update_fence(ring->fctx, fence);
 	}
 
-	if (msm_gpu_active(gpu)) {
-		/* retire completed submits, plus the one that hung: */
-		retire_submits(gpu);
+	/* retire completed submits, plus the one that hung: */
+	retire_submits(gpu);
 
-		gpu->funcs->recover(gpu);
+	gpu->funcs->recover(gpu);
 
-		/*
-		 * Replay all remaining submits starting with highest priority
-		 * ring
-		 */
-		for (i = 0; i < gpu->nr_rings; i++) {
-			struct msm_ringbuffer *ring = gpu->rb[i];
-			unsigned long flags;
+	/*
+	 * Replay all remaining submits starting with highest priority
+	 * ring
+	 */
+	for (i = 0; i < gpu->nr_rings; i++) {
+		struct msm_ringbuffer *ring = gpu->rb[i];
+		unsigned long flags;
 
-			spin_lock_irqsave(&ring->submit_lock, flags);
-			list_for_each_entry(submit, &ring->submits, node) {
-				/*
-				 * If the submit uses an unusable vm make sure
-				 * we don't actually run it
-				 */
-				if (to_msm_vm(submit->vm)->unusable)
-					submit->nr_cmds = 0;
-				gpu->funcs->submit(gpu, submit);
-			}
-			spin_unlock_irqrestore(&ring->submit_lock, flags);
+		spin_lock_irqsave(&ring->submit_lock, flags);
+		list_for_each_entry(submit, &ring->submits, node) {
+			/*
+			 * If the submit uses an unusable vm make sure
+			 * we don't actually run it
+			 */
+			if (to_msm_vm(submit->vm)->unusable)
+				submit->nr_cmds = 0;
+			gpu->funcs->submit(gpu, submit);
 		}
+		spin_unlock_irqrestore(&ring->submit_lock, flags);
 	}
 
 	pm_runtime_put(&gpu->pdev->dev);
author maintainer
#100 lore raw

PATCH 100/188 drm/i915/psr: Init variable to avoid early exit from et alignment loop

In reply to Greg Kroah-Hartman

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

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

From: Jouni Högander <[email protected]>

commit 314f6179e370988ac00dadf373a4f6166eb3db15 upstream.

Uninitialized boolean variable may cause unwanted exit from et alignment loop. Fix this by initializing it as false.

Fixes 1be2fca84f52 ("drm/i915/psr: Repeat Selective Update area alignment")
Cc <[email protected]> # v6.9+
(cherry picked from commit 289678a90b8cf81e3514c9d6c667235cd39c7acf)
---
 drivers/gpu/drm/i915/display/intel_psr.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/i915/display/intel_psr.c +1 -1
@@ -2840,7 +2840,7 @@ int intel_psr2_sel_fetch_update(struct i
 		return ret;
 
 	do {
-		bool cursor_in_su_area;
+		bool cursor_in_su_area = false;
 
 		/*
 		 * Adjust su area to cover cursor fully as necessary
author maintainer
#101 lore raw

PATCH 101/188 drm/amdkfd: Clear VRAM on allocation to prevent stale data exposure

In reply to Greg Kroah-Hartman

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

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

From: Amir Shetaia <[email protected]>

commit ad52d61d82181dbdb7f05826de38352d5e550cc2 upstream.

KFD VRAM allocations set AMDGPU_GEM_CREATE_VRAM_WIPE_ON_RELEASE but not AMDGPU_GEM_CREATE_VRAM_CLEARED, leaving freshly allocated VRAM with stale data from prior use observable by compute kernels.

The GEM ioctl path already sets VRAM_CLEARED for all userspace allocations via amdgpu_gem_create_ioctl() and amdgpu_mode_dumb_create(). The KFD path was missing this flag, allowing stale page table remnants to leak into user buffers.

This causes crashes in RCCL P2P transport where non-zero data in ptrExchange/head/tail fields corrupts the protocol handshake.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +2 -1
@@ -1717,7 +1717,8 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_
 			alloc_domain = AMDGPU_GEM_DOMAIN_GTT;
 			alloc_flags = 0;
 		} else {
-			alloc_flags = AMDGPU_GEM_CREATE_VRAM_WIPE_ON_RELEASE;
+			alloc_flags = AMDGPU_GEM_CREATE_VRAM_WIPE_ON_RELEASE |
+				AMDGPU_GEM_CREATE_VRAM_CLEARED;
 			alloc_flags |= (flags & KFD_IOC_ALLOC_MEM_FLAGS_PUBLIC) ?
 			AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED : 0;
 
author maintainer
#102 lore raw

PATCH 102/188 drm/amdgpu: Use SMUIO 15.0.0 offsets for TSC upper and lower count.

In reply to Greg Kroah-Hartman

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

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

From: Ramalingeswara Reddy, Kanala <[email protected]>

commit 574b3b14f7d1b329fc6e67b79328f0e6f4d4b3d4 upstream.

Define and use regGOLDEN_TSC_COUNT_UPPER_smu_15_0_0 and regGOLDEN_TSC_COUNT_LOWER_smu_15_0_0 for TSC upper and lower count.

---
 drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c |   31 ++++++++++++++++++++++++++-----
 1 file changed, 26 insertions(+), 5 deletions(-)
drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c +26 -5
@@ -64,6 +64,11 @@
 #define regPC_CONFIG_CNTL_1		0x194d
 #define regPC_CONFIG_CNTL_1_BASE_IDX	1
 
+#define regGOLDEN_TSC_COUNT_UPPER_smu_15_0_0               0x0030
+#define regGOLDEN_TSC_COUNT_UPPER_smu_15_0_0_BASE_IDX      1
+#define regGOLDEN_TSC_COUNT_LOWER_smu_15_0_0               0x0031
+#define regGOLDEN_TSC_COUNT_LOWER_smu_15_0_0_BASE_IDX      1
+
 #define regCP_GFX_MQD_CONTROL_DEFAULT                                             0x00000100
 #define regCP_GFX_HQD_VMID_DEFAULT                                                0x00000000
 #define regCP_GFX_HQD_QUEUE_PRIORITY_DEFAULT                                      0x00000000
@@ -5174,11 +5179,27 @@ static uint64_t gfx_v11_0_get_gpu_clock_
 		amdgpu_gfx_off_ctrl(adev, true);
 	} else {
 		preempt_disable();
-		clock_counter_hi_pre = (uint64_t)RREG32_SOC15(SMUIO, 0, regGOLDEN_TSC_COUNT_UPPER);
-		clock_counter_lo = (uint64_t)RREG32_SOC15(SMUIO, 0, regGOLDEN_TSC_COUNT_LOWER);
-		clock_counter_hi_after = (uint64_t)RREG32_SOC15(SMUIO, 0, regGOLDEN_TSC_COUNT_UPPER);
-		if (clock_counter_hi_pre != clock_counter_hi_after)
-			clock_counter_lo = (uint64_t)RREG32_SOC15(SMUIO, 0, regGOLDEN_TSC_COUNT_LOWER);
+		if (amdgpu_ip_version(adev, SMUIO_HWIP, 0) < IP_VERSION(15, 0, 0)) {
+			clock_counter_hi_pre = (uint64_t)RREG32_SOC15(SMUIO, 0,
+					regGOLDEN_TSC_COUNT_UPPER);
+			clock_counter_lo = (uint64_t)RREG32_SOC15(SMUIO, 0,
+					regGOLDEN_TSC_COUNT_LOWER);
+			clock_counter_hi_after = (uint64_t)RREG32_SOC15(SMUIO, 0,
+					regGOLDEN_TSC_COUNT_UPPER);
+			if (clock_counter_hi_pre != clock_counter_hi_after)
+				clock_counter_lo = (uint64_t)RREG32_SOC15(SMUIO, 0,
+						regGOLDEN_TSC_COUNT_LOWER);
+		} else {
+			clock_counter_hi_pre = (uint64_t)RREG32_SOC15(SMUIO, 0,
+					regGOLDEN_TSC_COUNT_UPPER_smu_15_0_0);
+			clock_counter_lo = (uint64_t)RREG32_SOC15(SMUIO, 0,
+					regGOLDEN_TSC_COUNT_LOWER_smu_15_0_0);
+			clock_counter_hi_after = (uint64_t)RREG32_SOC15(SMUIO, 0,
+					regGOLDEN_TSC_COUNT_UPPER_smu_15_0_0);
+			if (clock_counter_hi_pre != clock_counter_hi_after)
+				clock_counter_lo = (uint64_t)RREG32_SOC15(SMUIO, 0,
+						regGOLDEN_TSC_COUNT_LOWER_smu_15_0_0);
+		}
 		preempt_enable();
 	}
 	clock = clock_counter_lo | (clock_counter_hi_after << 32ULL);
author maintainer
#103 lore raw

PATCH 103/188 drm/amdgpu: gate VM CPU HDP flush on reset lock

In reply to Greg Kroah-Hartman

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

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

From: Chenglei Xie <[email protected]>

commit ddda81c4d7e71e41b1be91d921fd85747eddbd12 upstream.

During GPU reset, the application could still run CPU page table updates. Each commit called amdgpu_device_flush_hdp(), which on SR-IOV sends work through the KIQ ring. That can advance sync_seq while the GPU is being reset,
leaving fence writeback out of sync and causing amdgpu_fence_emit_polling() to time out on later KIQ use.

Fix:
amdgpu_vm_cpu_commit():
Reset will flush HDP anyway, the HDP flush in amdgpu_vm_cpu_commit() can be skipped when a reset is ongoging.
Take reset_domain->sem with down_read_trylock() before amdgpu_device_flush_hdp(). If the reset path holds the write lock, skip the HDP flush so no HDP-related HW access (including KIQ) runs during reset; state is re-established after reset.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm_cpu.c |   12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_vm_cpu.c +11 -1
@@ -21,6 +21,8 @@
  */
 
 #include "amdgpu_vm.h"
+#include "amdgpu.h"
+#include "amdgpu_reset.h"
 #include "amdgpu_object.h"
 #include "amdgpu_trace.h"
 
@@ -108,11 +110,19 @@ static int amdgpu_vm_cpu_update(struct a
 static int amdgpu_vm_cpu_commit(struct amdgpu_vm_update_params *p,
 				struct dma_fence **fence)
 {
+	struct amdgpu_device *adev = p->adev;
+
 	if (p->needs_flush)
 		atomic64_inc(&p->vm->tlb_seq);
 
 	mb();
-	amdgpu_device_flush_hdp(p->adev, NULL);
+	/* A reset flushed the HDP anyway, so that here can be skipped when a reset is ongoing */
+	if (!down_read_trylock(&adev->reset_domain->sem))
+		return 0;
+
+	amdgpu_device_flush_hdp(adev, NULL);
+	up_read(&adev->reset_domain->sem);
+
 	return 0;
 }
 
author maintainer
#104 lore raw

PATCH 104/188 drm/amd/pm: fix incorrect FeatureCtrlMask setting on smu v14.0.x

In reply to Greg Kroah-Hartman

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

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

From: Yang Wang <[email protected]>

commit 504f0098ebd074ac8c0ce3471795d79f68e3d265 upstream.

OverDriveTable.FanMinimumPwm and FeatureCtrlMask.PP_OD_FEATURE_FAN_LEGACY_BIT have a hard dependency.
Invalid handling of this dependency leads to disabled thermal monitoring and temperature boundary validation.

v2: squash in typo fix (Yang)

Fixes 9710b84e2a6a ("drm/amd/pm: add overdrive support on smu v14.0.2/3")
---
 drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c |   10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c +7 -3
@@ -2442,6 +2442,7 @@ static int smu_v14_0_2_od_restore_table_
 		}
 		od_table->OverDriveTable.FanMode = FAN_MODE_AUTO;
 		od_table->OverDriveTable.FeatureCtrlMask |= BIT(PP_OD_FEATURE_FAN_CURVE_BIT);
+		od_table->OverDriveTable.FeatureCtrlMask &= ~BIT(PP_OD_FEATURE_FAN_LEGACY_BIT);
 		break;
 	case PP_OD_EDIT_FAN_ZERO_RPM_ENABLE:
 		od_table->OverDriveTable.FanZeroRpmEnable =
@@ -2470,7 +2471,8 @@ static int smu_v14_0_2_od_restore_table_
 		od_table->OverDriveTable.FanMinimumPwm =
 					boot_overdrive_table->OverDriveTable.FanMinimumPwm;
 		od_table->OverDriveTable.FanMode = FAN_MODE_AUTO;
-		od_table->OverDriveTable.FeatureCtrlMask |= BIT(PP_OD_FEATURE_FAN_CURVE_BIT);
+		od_table->OverDriveTable.FeatureCtrlMask |= BIT(PP_OD_FEATURE_FAN_LEGACY_BIT);
+		od_table->OverDriveTable.FeatureCtrlMask &= ~BIT(PP_OD_FEATURE_FAN_CURVE_BIT);
 		break;
 	default:
 		dev_info(adev->dev, "Invalid table index: %ld\n", input);
@@ -2640,6 +2642,7 @@ static int smu_v14_0_2_od_edit_dpm_table
 		od_table->OverDriveTable.FanLinearPwmPoints[input[0]] = input[2];
 		od_table->OverDriveTable.FanMode = FAN_MODE_MANUAL_LINEAR;
 		od_table->OverDriveTable.FeatureCtrlMask |= BIT(PP_OD_FEATURE_FAN_CURVE_BIT);
+		od_table->OverDriveTable.FeatureCtrlMask &= ~BIT(PP_OD_FEATURE_FAN_LEGACY_BIT);
 		break;
 
 	case PP_OD_EDIT_ACOUSTIC_LIMIT:
@@ -2709,7 +2712,7 @@ static int smu_v14_0_2_od_edit_dpm_table
 		break;
 
 	case PP_OD_EDIT_FAN_MINIMUM_PWM:
-		if (!smu_v14_0_2_is_od_feature_supported(smu, PP_OD_FEATURE_FAN_CURVE_BIT)) {
+		if (!smu_v14_0_2_is_od_feature_supported(smu, PP_OD_FEATURE_FAN_LEGACY_BIT)) {
 			dev_warn(adev->dev, "Fan curve setting not supported!\n");
 			return -ENOTSUPP;
 		}
@@ -2727,7 +2730,8 @@ static int smu_v14_0_2_od_edit_dpm_table
 
 		od_table->OverDriveTable.FanMinimumPwm = input[0];
 		od_table->OverDriveTable.FanMode = FAN_MODE_AUTO;
-		od_table->OverDriveTable.FeatureCtrlMask |= BIT(PP_OD_FEATURE_FAN_CURVE_BIT);
+		od_table->OverDriveTable.FeatureCtrlMask |= BIT(PP_OD_FEATURE_FAN_LEGACY_BIT);
+		od_table->OverDriveTable.FeatureCtrlMask &= ~BIT(PP_OD_FEATURE_FAN_CURVE_BIT);
 		break;
 
 	case PP_OD_EDIT_FAN_ZERO_RPM_ENABLE:
author maintainer
#105 lore raw

PATCH 105/188 drm/amdkfd: Add upper bound check for num_of_nodes

In reply to Greg Kroah-Hartman

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

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

From: Alysa Liu <[email protected]>

commit 74b73fa56a395d46745e4f245225963e9f8be7f1 upstream.

drm/amdkfd: Add upper bound check for num_of_nodes in kfd_ioctl_get_process_apertures_new.

(cherry picked from commit 98ff46a5ea090c14d2cdb4f5b993b05d74f3949f)
---
 drivers/gpu/drm/amd/amdkfd/kfd_chardev.c  |    3 +++
 drivers/gpu/drm/amd/amdkfd/kfd_priv.h     |    1 +
 drivers/gpu/drm/amd/amdkfd/kfd_topology.c |   11 +++++++++++
 3 files changed, 15 insertions(+)
drivers/gpu/drm/amd/amdkfd/kfd_chardev.c +3 -0
@@ -763,6 +763,9 @@ static int kfd_ioctl_get_process_apertur
 		goto out_unlock;
 	}
 
+	if (args->num_of_nodes > kfd_topology_get_num_devices())
+		return -EINVAL;
+
 	/* Fill in process-aperture information for all available
 	 * nodes, but not more than args->num_of_nodes as that is
 	 * the amount of memory allocated by user
drivers/gpu/drm/amd/amdkfd/kfd_priv.h +1 -0
@@ -1175,6 +1175,7 @@ static inline struct kfd_node *kfd_node_
 	return NULL;
 }
 int kfd_topology_enum_kfd_devices(uint8_t idx, struct kfd_node **kdev);
+uint32_t kfd_topology_get_num_devices(void);
 int kfd_numa_node_to_apic_id(int numa_node_id);
 
 /* Interrupts */
drivers/gpu/drm/amd/amdkfd/kfd_topology.c +11 -0
@@ -2295,6 +2295,17 @@ int kfd_topology_remove_device(struct kf
 	return res;
 }
 
+uint32_t kfd_topology_get_num_devices(void)
+{
+	uint32_t num_devices;
+
+	down_read(&topology_lock);
+	num_devices = sys_props.num_devices;
+	up_read(&topology_lock);
+
+	return num_devices;
+}
+
 /* kfd_topology_enum_kfd_devices - Enumerate through all devices in KFD
  *	topology. If GPU device is found @idx, then valid kfd_dev pointer is
  *	returned through @kdev
author maintainer
#106 lore raw

PATCH 106/188 drm/amdgpu: Add bounds checking to ib_{get,set}_value

In reply to Greg Kroah-Hartman

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

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

From: Benjamin Cheng <[email protected]>

commit 66085e206431ef88ce36f53c1f53d570790ccc9e upstream.

The uvd/vce/vcn code accesses the IB at predefined offsets without checking that the IB is large enough. Check the bounds here. The caller is responsible for making sure it can handle arbitrary return values.

Also make the idx a uint32_t to prevent overflows causing the condition to fail.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h |   11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h +7 -4
@@ -547,15 +547,18 @@ void amdgpu_debugfs_ring_init(struct amd
 
 int amdgpu_ring_init_mqd(struct amdgpu_ring *ring);
 
-static inline u32 amdgpu_ib_get_value(struct amdgpu_ib *ib, int idx)
+static inline u32 amdgpu_ib_get_value(struct amdgpu_ib *ib, uint32_t idx)
 {
-	return ib->ptr[idx];
+	if (idx < ib->length_dw)
+		return ib->ptr[idx];
+	return 0;
 }
 
-static inline void amdgpu_ib_set_value(struct amdgpu_ib *ib, int idx,
+static inline void amdgpu_ib_set_value(struct amdgpu_ib *ib, uint32_t idx,
 				       uint32_t value)
 {
-	ib->ptr[idx] = value;
+	if (idx < ib->length_dw)
+		ib->ptr[idx] = value;
 }
 
 int amdgpu_ib_get(struct amdgpu_device *adev, struct amdgpu_vm *vm,
author maintainer
#107 lore raw

PATCH 107/188 drm/amdgpu/vcn4: Prevent OOB reads when parsing IB

In reply to Greg Kroah-Hartman

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

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

From: Benjamin Cheng <[email protected]>

commit 2444eb0ec8283f4a3845eb7febad378476e1ba3c upstream.

Rewrite the IB parsing to use amdgpu_ib_get_value() which handles the bounds checks.

---
 drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c |   23 ++++++++++++-----------
 1 file changed, 12 insertions(+), 11 deletions(-)
drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c +12 -11
@@ -1912,9 +1912,10 @@ out:
 static int vcn_v4_0_enc_find_ib_param(struct amdgpu_ib *ib, uint32_t id, int start)
 {
 	int i;
+	uint32_t len;
 
-	for (i = start; i < ib->length_dw && ib->ptr[i] >= 8; i += ib->ptr[i] / 4) {
-		if (ib->ptr[i + 1] == id)
+	for (i = start; (len = amdgpu_ib_get_value(ib, i)) >= 8; i += len / 4) {
+		if (amdgpu_ib_get_value(ib, i + 1) == id)
 			return i;
 	}
 	return -1;
@@ -1925,8 +1926,6 @@ static int vcn_v4_0_ring_patch_cs_in_pla
 					   struct amdgpu_ib *ib)
 {
 	struct amdgpu_ring *ring = amdgpu_job_ring(job);
-	struct amdgpu_vcn_decode_buffer *decode_buffer;
-	uint64_t addr;
 	uint32_t val;
 	int idx = 0, sidx;
 
@@ -1937,20 +1936,22 @@ static int vcn_v4_0_ring_patch_cs_in_pla
 	while ((idx = vcn_v4_0_enc_find_ib_param(ib, RADEON_VCN_ENGINE_INFO, idx)) >= 0) {
 		val = amdgpu_ib_get_value(ib, idx + 2); /* RADEON_VCN_ENGINE_TYPE */
 		if (val == RADEON_VCN_ENGINE_TYPE_DECODE) {
-			decode_buffer = (struct amdgpu_vcn_decode_buffer *)&ib->ptr[idx + 6];
+			uint32_t valid_buf_flag = amdgpu_ib_get_value(ib, idx + 6);
+			uint64_t msg_buffer_addr;
 
-			if (!(decode_buffer->valid_buf_flag & 0x1))
+			if (!(valid_buf_flag & 0x1))
 				return 0;
 
-			addr = ((u64)decode_buffer->msg_buffer_address_hi) << 32 |
-				decode_buffer->msg_buffer_address_lo;
-			return vcn_v4_0_dec_msg(p, job, addr);
+			msg_buffer_addr = ((u64)amdgpu_ib_get_value(ib, idx + 7)) << 32 |
+				amdgpu_ib_get_value(ib, idx + 8);
+			return vcn_v4_0_dec_msg(p, job, msg_buffer_addr);
 		} else if (val == RADEON_VCN_ENGINE_TYPE_ENCODE) {
 			sidx = vcn_v4_0_enc_find_ib_param(ib, RENCODE_IB_PARAM_SESSION_INIT, idx);
-			if (sidx >= 0 && ib->ptr[sidx + 2] == RENCODE_ENCODE_STANDARD_AV1)
+			if (sidx >= 0 &&
+			    amdgpu_ib_get_value(ib, sidx + 2) == RENCODE_ENCODE_STANDARD_AV1)
 				return vcn_v4_0_limit_sched(p, job);
 		}
-		idx += ib->ptr[idx] / 4;
+		idx += amdgpu_ib_get_value(ib, idx) / 4;
 	}
 	return 0;
 }
author maintainer
#108 lore raw

PATCH 108/188 drm/amdgpu/vce: Prevent partial address patches

In reply to Greg Kroah-Hartman

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

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

From: Benjamin Cheng <[email protected]>

commit de2a02cc28d6d5d37db07d00a9a684c754a5fd74 upstream.

In the case that only one of lo/hi is valid, the patching could result in a bad address written to in FW.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c |    3 +++
 1 file changed, 3 insertions(+)
drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c +3 -0
@@ -656,6 +656,9 @@ static int amdgpu_vce_cs_reloc(struct am
 	uint64_t addr;
 	int r;
 
+	if (lo >= ib->length_dw || hi >= ib->length_dw)
+		return -EINVAL;
+
 	if (index == 0xffffffff)
 		index = 0;
 
author maintainer
#109 lore raw

PATCH 109/188 drm/amdgpu/vcn4: Prevent OOB reads when parsing dec msg

In reply to Greg Kroah-Hartman

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

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

From: Benjamin Cheng <[email protected]>

commit 0a78f2bac1424deb7c9d5e09c6b8e849d8e8b648 upstream.

Check bounds against the end of the BO whenever we access the msg.

---
 drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c |   21 ++++++++++++++++++---
 1 file changed, 18 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c +18 -3
@@ -1825,7 +1825,7 @@ static int vcn_v4_0_dec_msg(struct amdgp
 {
 	struct ttm_operation_ctx ctx = { false, false };
 	struct amdgpu_bo_va_mapping *map;
-	uint32_t *msg, num_buffers;
+	uint32_t *msg, num_buffers, len_dw;
 	struct amdgpu_bo *bo;
 	uint64_t start, end;
 	unsigned int i;
@@ -1846,6 +1846,11 @@ static int vcn_v4_0_dec_msg(struct amdgp
 		return -EINVAL;
 	}
 
+	if (end - addr < 16) {
+		DRM_ERROR("VCN messages must be at least 4 DWORDs!\n");
+		return -EINVAL;
+	}
+
 	bo->flags |= AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED;
 	amdgpu_bo_placement_from_domain(bo, bo->allowed_domains);
 	r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
@@ -1862,8 +1867,8 @@ static int vcn_v4_0_dec_msg(struct amdgp
 
 	msg = ptr + addr - start;
 
-	/* Check length */
 	if (msg[1] > end - addr) {
+		DRM_ERROR("VCN message header does not fit in BO!\n");
 		r = -EINVAL;
 		goto out;
 	}
@@ -1871,7 +1876,16 @@ static int vcn_v4_0_dec_msg(struct amdgp
 	if (msg[3] != RDECODE_MSG_CREATE)
 		goto out;
 
+	len_dw = msg[1] / 4;
 	num_buffers = msg[2];
+
+	/* Verify that all indices fit within the claimed length. Each index is 4 DWORDs */
+	if (num_buffers > len_dw || 6 + num_buffers * 4 > len_dw) {
+		DRM_ERROR("VCN message has too many buffers!\n");
+		r = -EINVAL;
+		goto out;
+	}
+
 	for (i = 0, msg = &msg[6]; i < num_buffers; ++i, msg += 4) {
 		uint32_t offset, size, *create;
 
@@ -1881,7 +1895,8 @@ static int vcn_v4_0_dec_msg(struct amdgp
 		offset = msg[1];
 		size = msg[2];
 
-		if (offset + size > end) {
+		if (size < 4 || offset + size > end - addr) {
+			DRM_ERROR("VCN message buffer exceeds BO bounds!\n");
 			r = -EINVAL;
 			goto out;
 		}
author maintainer
#110 lore raw

PATCH 110/188 drm/amdgpu/vcn3: Prevent OOB reads when parsing dec msg

In reply to Greg Kroah-Hartman

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

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

From: Benjamin Cheng <[email protected]>

commit b193019860d61e92da395eae2011f2f6716b182f upstream.

Check bounds against the end of the BO whenever we access the msg.

---
 drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c |   23 +++++++++++++++++++----
 1 file changed, 19 insertions(+), 4 deletions(-)
drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c +19 -4
@@ -1908,7 +1908,7 @@ static int vcn_v3_0_dec_msg(struct amdgp
 {
 	struct ttm_operation_ctx ctx = { false, false };
 	struct amdgpu_bo_va_mapping *map;
-	uint32_t *msg, num_buffers;
+	uint32_t *msg, num_buffers, len_dw;
 	struct amdgpu_bo *bo;
 	uint64_t start, end;
 	unsigned int i;
@@ -1929,6 +1929,11 @@ static int vcn_v3_0_dec_msg(struct amdgp
 		return -EINVAL;
 	}
 
+	if (end - addr < 16) {
+		DRM_ERROR("VCN messages must be at least 4 DWORDs!\n");
+		return -EINVAL;
+	}
+
 	bo->flags |= AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED;
 	amdgpu_bo_placement_from_domain(bo, bo->allowed_domains);
 	r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
@@ -1945,8 +1950,8 @@ static int vcn_v3_0_dec_msg(struct amdgp
 
 	msg = ptr + addr - start;
 
-	/* Check length */
 	if (msg[1] > end - addr) {
+		DRM_ERROR("VCN message header does not fit in BO!\n");
 		r = -EINVAL;
 		goto out;
 	}
@@ -1954,7 +1959,16 @@ static int vcn_v3_0_dec_msg(struct amdgp
 	if (msg[3] != RDECODE_MSG_CREATE)
 		goto out;
 
+	len_dw = msg[1] / 4;
 	num_buffers = msg[2];
+
+	/* Verify that all indices fit within the claimed length. Each index is 4 DWORDs */
+	if (num_buffers > len_dw || 6 + num_buffers * 4 > len_dw) {
+		DRM_ERROR("VCN message has too many buffers!\n");
+		r = -EINVAL;
+		goto out;
+	}
+
 	for (i = 0, msg = &msg[6]; i < num_buffers; ++i, msg += 4) {
 		uint32_t offset, size, *create;
 
@@ -1964,14 +1978,15 @@ static int vcn_v3_0_dec_msg(struct amdgp
 		offset = msg[1];
 		size = msg[2];
 
-		if (offset + size > end) {
+		if (size < 4 || offset + size > end - addr) {
+			DRM_ERROR("VCN message buffer exceeds BO bounds!\n");
 			r = -EINVAL;
 			goto out;
 		}
 
 		create = ptr + addr + offset - start;
 
-		/* H246, HEVC and VP9 can run on any instance */
+		/* H264, HEVC and VP9 can run on any instance */
 		if (create[0] == 0x7 || create[0] == 0x10 || create[0] == 0x11)
 			continue;
 
author maintainer
#111 lore raw

PATCH 111/188 drm/amd/display: Change dither policy for 10 bpc output back to dithering

In reply to Greg Kroah-Hartman

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

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

From: Mario Kleiner <[email protected]>

commit d65bfb1782304b03862c8c725fac608015dffd36 upstream.

Commit d5df648ec830 ("drm/amd/display: Change dither policy for 10bpc to round") degraded display of 12 bpc color precision output to 10 bpc sinks by switching 10 bpc output from dithering to "truncate to 10 bpc".

I don't find the argumentation in that commit convincing, but the consequences highly unfortunate, especially for applications that require effective > 10 bpc precision output of > 10 bpc framebuffers.

The argument wasn't something strong like "there are hardware design defects or limitations which require us to work around broken dithering to 10 bpc", or "there are some special use cases which do require truncation to 10 bpc", but essentially "at some point in the past we used truncation in Polaris/Vega times and it looks like it got inadvertently changed for Navi, so let's do that again". I couldn't find evidence for that in the git commit logs for this. The commit message also acknowledges that using dithering "...makes some sense for FP16... ...but not for ARGB2101010 surfaces..."

The problem with this is that it makes fp16 surfaces, and especially rgba16 fixed point surfaces, less useful. These are now well supported by Mesa 25.3 and later via OpenGL + EGL, Vulkan/WSI, and by OSS AMDVLK Vulkan/WSI/display, and also by GNOME 50 mutter under Wayland, and they used to provide more than 10 bpc effective precision at the output.

Even for 8 or 10 bpc surfaces, the color pipeline behind the framebuffer, e.g., gamma tables, CTM, can be used for color correction and will benefit from an effective > 10 bpc output precision via dithering, retaining some precision that would get lost on the way through the pipeline, e.g., due to non-linear gamma functions.

Scientific apps rely on this for > 10 bpc display precision. Truncating to 10 bpc, instead of dithering the pipeline internal 12 bpc precision down to 10 bpc, causes a serious loss of precision. This also creates the undesirable and slightly absurd situation that using a cheap monitor with only 8 bpc input and display panel will yield roughly 12 bpc precision via dithering from 12 -> 8 bpc, whereas investment into a more expensive monitor with 10 bpc input and native 10 bpc display will only yield 10 bpc, even if a fp16 or rgb16 framebuffer and/or a properly set up color pipeline (gamma tables, CTM's etc. with more than 10 bpc out precision) would allow effective 12 bpc precision output.

Therefore this patch proposes reverting that commit and going back to dithering down to 10 bpc, consistent with the behaviour for 6 bpc or 8 bpc output.

Successfully tested on AMD Polaris DCE 11.2 and Raven Ridge DCN 1.0 with a native 10 bpc capable monitor, outputting a RGBA16 unorm framebuffer and measuring resulting color precision with a photometer. No apparent visual artifacts or problems were observed, and effective precision was measured to be 12 bpc again, as expected.

Fixes d5df648ec830 ("drm/amd/display: Change dither policy for 10bpc to round")
Cc Aric Cyr <[email protected]>
Cc Anthony Koo <[email protected]>
Cc Rodrigo Siqueira <[email protected]>
Cc Krunoslav Kovac <[email protected]>
Cc Alex Deucher <[email protected]>
---
 drivers/gpu/drm/amd/display/dc/core/dc_resource.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/core/dc_resource.c +1 -1
@@ -4954,7 +4954,7 @@ void resource_build_bit_depth_reduction_
 			option = DITHER_OPTION_SPATIAL8;
 			break;
 		case COLOR_DEPTH_101010:
-			option = DITHER_OPTION_TRUN10;
+			option = DITHER_OPTION_SPATIAL10;
 			break;
 		default:
 			option = DITHER_OPTION_DISABLE;
author maintainer
#112 lore raw

PATCH 112/188 drm/gem: Fix inconsistent plane dimension calculation in drm_gem_fb_init_with_funcs()

In reply to Greg Kroah-Hartman

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

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

From: Ashutosh Desai <[email protected]>

commit 3d4c2268bd7243c3780fe32bf24ff876da272acf upstream.

drm_gem_fb_init_with_funcs() computes sub-sampled plane dimensions using plain integer division:

  unsigned int width  = mode_cmd->width  / (i ? info->hsub : 1);
  unsigned int height = mode_cmd->height / (i ? info->vsub : 1);

However, the ioctl-level framebuffer_check() in drm_framebuffer.c uses drm_format_info_plane_width/height() which round up dimensions via DIV_ROUND_UP(). This inconsistency corrupts the subsequent GEM object size check for certain pixel format and dimension combinations.

For example, with NV12 (vsub=2) and a 1-pixel-tall framebuffer the GEM size validation path sees height=0 instead of height=1. The expression (height - 1) then wraps to UINT_MAX as an unsigned int, causing min_size to overflow and wrap back to a small value. A tiny GEM object therefore passes the size guard, yet when the GPU accesses the chroma plane it will read or write memory beyond the object's bounds.

Fix by replacing the open-coded divisions with drm_format_info_plane_width() and drm_format_info_plane_height(), which use DIV_ROUND_UP() and match the calculation already used in framebuffer_check().

---
 drivers/gpu/drm/drm_gem_framebuffer_helper.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/drm_gem_framebuffer_helper.c +2 -2
@@ -171,8 +171,8 @@ int drm_gem_fb_init_with_funcs(struct dr
 	}
 
 	for (i = 0; i < info->num_planes; i++) {
-		unsigned int width = mode_cmd->width / (i ? info->hsub : 1);
-		unsigned int height = mode_cmd->height / (i ? info->vsub : 1);
+		unsigned int width = drm_format_info_plane_width(info, mode_cmd->width, i);
+		unsigned int height = drm_format_info_plane_height(info, mode_cmd->height, i);
 		unsigned int min_size;
 
 		objs[i] = drm_gem_object_lookup(file, mode_cmd->handles[i]);
author maintainer
#113 lore raw

PATCH 113/188 drm/appletbdrm: Use kvzalloc for big allocations

In reply to Greg Kroah-Hartman

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

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

From: Sasha Finkelstein <[email protected]>

commit aaaa684bab1f6d9ecfc49db328facb1771fd0eb2 upstream.

This driver is attached to a ~2000x80 screen, which is a lot more than a single page. This causes out of memory errors in some rare cases.

---
 drivers/gpu/drm/tiny/appletbdrm.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/tiny/appletbdrm.c +2 -2
@@ -353,7 +353,7 @@ static int appletbdrm_primary_plane_help
 		       frames_size +
 		       sizeof(struct appletbdrm_fb_request_footer), 16);
 
-	appletbdrm_state->request = kzalloc(request_size, GFP_KERNEL);
+	appletbdrm_state->request = kvzalloc(request_size, GFP_KERNEL);
 
 	if (!appletbdrm_state->request)
 		return -ENOMEM;
@@ -543,7 +543,7 @@ static void appletbdrm_primary_plane_des
 {
 	struct appletbdrm_plane_state *appletbdrm_state = to_appletbdrm_plane_state(state);
 
-	kfree(appletbdrm_state->request);
+	kvfree(appletbdrm_state->request);
 	kfree(appletbdrm_state->response);
 
 	__drm_gem_destroy_shadow_plane_state(&appletbdrm_state->base);
author maintainer
#114 lore raw

PATCH 114/188 drm/amdkfd: validate SVM ioctl nattr against buffer size

In reply to Greg Kroah-Hartman

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

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

From: Alysa Liu <[email protected]>

commit 045e0ff208f0838a246c10204105126611b267a1 upstream.

Validate nattr field against the buffer size, preventing out-of-bounds buffer access via user-controlled attribute count.

(cherry picked from commit 5eca8bfdfa456c3304ca77523718fe24254c172f)
---
 drivers/gpu/drm/amd/amdkfd/kfd_chardev.c |   26 ++++++++++++++++++++++++--
 drivers/gpu/drm/amd/amdkfd/kfd_priv.h    |    3 +++
 2 files changed, 27 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/amdkfd/kfd_chardev.c +24 -2
@@ -25,6 +25,7 @@
 #include <linux/err.h>
 #include <linux/fs.h>
 #include <linux/file.h>
+#include <linux/overflow.h>
 #include <linux/sched.h>
 #include <linux/slab.h>
 #include <linux/uaccess.h>
@@ -1676,6 +1677,16 @@ static int kfd_ioctl_smi_events(struct f
 	return kfd_smi_event_open(pdd->dev, &args->anon_fd);
 }
 
+static int kfd_ioctl_svm_validate(void *kdata, unsigned int usize)
+{
+	struct kfd_ioctl_svm_args *args = kdata;
+	size_t expected = struct_size(args, attrs, args->nattr);
+
+	if (expected == SIZE_MAX || usize < expected)
+		return -EINVAL;
+	return 0;
+}
+
 #if IS_ENABLED(CONFIG_HSA_AMD_SVM)
 
 static int kfd_ioctl_set_xnack_mode(struct file *filep,
@@ -3122,7 +3133,11 @@ out:
 
 #define AMDKFD_IOCTL_DEF(ioctl, _func, _flags) \
 	[_IOC_NR(ioctl)] = {.cmd = ioctl, .func = _func, .flags = _flags, \
-			    .cmd_drv = 0, .name = #ioctl}
+			    .validate = NULL, .cmd_drv = 0, .name = #ioctl}
+
+#define AMDKFD_IOCTL_DEF_V(ioctl, _func, _validate, _flags) \
+	[_IOC_NR(ioctl)] = {.cmd = ioctl, .func = _func, .flags = _flags, \
+			    .validate = _validate, .cmd_drv = 0, .name = #ioctl}
 
 /** Ioctl table */
 static const struct amdkfd_ioctl_desc amdkfd_ioctls[] = {
@@ -3219,7 +3234,8 @@ static const struct amdkfd_ioctl_desc am
 	AMDKFD_IOCTL_DEF(AMDKFD_IOC_SMI_EVENTS,
 			kfd_ioctl_smi_events, 0),
 
-	AMDKFD_IOCTL_DEF(AMDKFD_IOC_SVM, kfd_ioctl_svm, 0),
+	AMDKFD_IOCTL_DEF_V(AMDKFD_IOC_SVM, kfd_ioctl_svm,
+			   kfd_ioctl_svm_validate, 0),
 
 	AMDKFD_IOCTL_DEF(AMDKFD_IOC_SET_XNACK_MODE,
 			kfd_ioctl_set_xnack_mode, 0),
@@ -3341,6 +3357,12 @@ static long kfd_ioctl(struct file *filep
 		memset(kdata, 0, usize);
 	}
 
+	if (ioctl->validate) {
+		retcode = ioctl->validate(kdata, usize);
+		if (retcode)
+			goto err_i1;
+	}
+
 	retcode = func(filep, process, kdata);
 
 	if (cmd & IOC_OUT)
drivers/gpu/drm/amd/amdkfd/kfd_priv.h +3 -0
@@ -1036,10 +1036,13 @@ extern struct srcu_struct kfd_processes_
 typedef int amdkfd_ioctl_t(struct file *filep, struct kfd_process *p,
 				void *data);
 
+typedef int amdkfd_ioctl_validate_t(void *kdata, unsigned int usize);
+
 struct amdkfd_ioctl_desc {
 	unsigned int cmd;
 	int flags;
 	amdkfd_ioctl_t *func;
+	amdkfd_ioctl_validate_t *validate;
 	unsigned int cmd_drv;
 	const char *name;
 };
author maintainer
#115 lore raw

PATCH 115/188 drm/udl: Increase GET_URB_TIMEOUT

In reply to Greg Kroah-Hartman

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

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

From: Shixiong Ou <[email protected]>

commit ac2c996675755c725a0065dbe3e2ebffded9080b upstream.

[WHY]
A situation has occurred where udl_handle_damage() executed successfully and the kernel log appears normal, but the display fails to show any output. This is because the call to udl_get_urb() in udl_crtc_helper_atomic_enable() failed without generating any error message.

[HOW]
1. Increase timeout of getting urb.
2. Add error messages when calling udl_get_urb() failed in udl_crtc_helper_atomic_enable().

---
 drivers/gpu/drm/udl/udl_main.c    |    3 +--
 drivers/gpu/drm/udl/udl_modeset.c |    5 ++++-
 2 files changed, 5 insertions(+), 3 deletions(-)
drivers/gpu/drm/udl/udl_main.c +1 -2
@@ -285,13 +285,12 @@ static struct urb *udl_get_urb_locked(st
 	return unode->urb;
 }
 
-#define GET_URB_TIMEOUT	HZ
 struct urb *udl_get_urb(struct udl_device *udl)
 {
 	struct urb *urb;
 
 	spin_lock_irq(&udl->urbs.lock);
-	urb = udl_get_urb_locked(udl, GET_URB_TIMEOUT);
+	urb = udl_get_urb_locked(udl, HZ * 2);
 	spin_unlock_irq(&udl->urbs.lock);
 	return urb;
 }
drivers/gpu/drm/udl/udl_modeset.c +4 -1
@@ -21,6 +21,7 @@
 #include <drm/drm_gem_framebuffer_helper.h>
 #include <drm/drm_gem_shmem_helper.h>
 #include <drm/drm_modeset_helper_vtables.h>
+#include <drm/drm_print.h>
 #include <drm/drm_probe_helper.h>
 #include <drm/drm_vblank.h>
 
@@ -342,8 +343,10 @@ static void udl_crtc_helper_atomic_enabl
 		return;
 
 	urb = udl_get_urb(udl);
-	if (!urb)
+	if (!urb) {
+		drm_err_ratelimited(dev, "get urb failed when enabling crtc\n");
 		goto out;
+	}
 
 	buf = (char *)urb->transfer_buffer;
 	buf = udl_vidreg_lock(buf);
author maintainer
#116 lore raw

PATCH 116/188 drm/xe: Fix bo leak in xe_dma_buf_init_obj() on allocation failure

In reply to Greg Kroah-Hartman

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

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

From: Shuicheng Lin <[email protected]>

commit 93a528f67ce5095bcab46a69839eca97f43dd352 upstream.

When drm_gpuvm_resv_object_alloc() fails, the pre-allocated storage bo is not freed. Add xe_bo_free(storage) before returning the error.

xe_dma_buf_init_obj() calls xe_bo_init_locked(), which frees the bo on error. Therefore, xe_dma_buf_init_obj() must also free the bo on its own error paths. Otherwise, since xe_gem_prime_import() cannot distinguish whether the failure originated from xe_dma_buf_init_obj() or from xe_bo_init_locked(), it cannot safely decide whether the bo should be freed.

Add comments documenting the ownership semantics: on success, ownership of storage is transferred to the returned drm_gem_object; on failure, storage is freed before returning.

v2: Add comments to explain the free logic.

Fixes eb289a5f6cc6 ("drm/xe: Convert xe_dma_buf.c for exhaustive eviction")
Assisted-by Claude:claude-opus-4.6
(cherry picked from commit 78a6c5f899f22338bbf48b44fb8950409c5a69b9)
---
 drivers/gpu/drm/xe/xe_dma_buf.c |   12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)
drivers/gpu/drm/xe/xe_dma_buf.c +11 -1
@@ -227,6 +227,13 @@ struct dma_buf *xe_gem_prime_export(stru
 	return buf;
 }
 
+/*
+ * Takes ownership of @storage: on success it is transferred to the returned
+ * drm_gem_object; on failure it is freed before returning the error.
+ * This matches the contract of xe_bo_init_locked() which frees @storage on
+ * its error paths, so callers need not (and must not) free @storage after
+ * this call.
+ */
 static struct drm_gem_object *
 xe_dma_buf_init_obj(struct drm_device *dev, struct xe_bo *storage,
 		    struct dma_buf *dma_buf)
@@ -240,8 +247,10 @@ xe_dma_buf_init_obj(struct drm_device *d
 	int ret = 0;
 
 	dummy_obj = drm_gpuvm_resv_object_alloc(&xe->drm);
-	if (!dummy_obj)
+	if (!dummy_obj) {
+		xe_bo_free(storage);
 		return ERR_PTR(-ENOMEM);
+	}
 
 	dummy_obj->resv = resv;
 	xe_validation_guard(&ctx, &xe->val, &exec, (struct xe_val_flags) {}, ret) {
@@ -250,6 +259,7 @@ xe_dma_buf_init_obj(struct drm_device *d
 		if (ret)
 			break;
 
+		/* xe_bo_init_locked() frees storage on error */
 		bo = xe_bo_init_locked(xe, storage, NULL, resv, NULL, dma_buf->size,
 				       0, /* Will require 1way or 2way for vm_bind */
 				       ttm_bo_type_sg, XE_BO_FLAG_SYSTEM, &exec);
author maintainer
#117 lore raw

PATCH 117/188 drm/xe/bo: Fix bo leak on GGTT flag validation in xe_bo_init_locked()

In reply to Greg Kroah-Hartman

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

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

From: Shuicheng Lin <[email protected]>

commit 1d0adf2fd94fb0c0037c643fadd8f2cf3cffc009 upstream.

When XE_BO_FLAG_GGTT_ALL is set without XE_BO_FLAG_GGTT, the function returns an error without freeing a caller-provided bo, violating the documented contract that bo is freed on failure.

Add xe_bo_free(bo) before returning the error.

Fixes 5a3b0df25d6a ("drm/xe: Allow bo mapping on multiple ggtts")
Assisted-by Claude:claude-opus-4.6
(cherry picked from commit 3fbd6cf43cac7b60757f3ce3d95195d3843a902c)
---
 drivers/gpu/drm/xe/xe_bo.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/gpu/drm/xe/xe_bo.c +3 -1
@@ -2112,8 +2112,10 @@ struct xe_bo *xe_bo_init_locked(struct x
 	}
 
 	/* XE_BO_FLAG_GGTTx requires XE_BO_FLAG_GGTT also be set */
-	if ((flags & XE_BO_FLAG_GGTT_ALL) && !(flags & XE_BO_FLAG_GGTT))
+	if ((flags & XE_BO_FLAG_GGTT_ALL) && !(flags & XE_BO_FLAG_GGTT)) {
+		xe_bo_free(bo);
 		return ERR_PTR(-EINVAL);
+	}
 
 	if (flags & (XE_BO_FLAG_VRAM_MASK | XE_BO_FLAG_STOLEN) &&
 	    !(flags & XE_BO_FLAG_IGNORE_MIN_PAGE_SIZE) &&
author maintainer
#118 lore raw

PATCH 118/188 drm/xe: Fix dma-buf attachment leak in xe_gem_prime_import()

In reply to Greg Kroah-Hartman

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

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

From: Shuicheng Lin <[email protected]>

commit 111ab678471bf1f90d078d5513bb086b70596c3c upstream.

When xe_dma_buf_init_obj() fails, the attachment from dma_buf_dynamic_attach() is not detached. Add dma_buf_detach() before returning the error. Note: we cannot use goto out_err here because xe_dma_buf_init_obj() already frees bo on failure, and out_err would double-free it.

Fixes dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs")
Assisted-by Claude:claude-opus-4.6
(cherry picked from commit a828eb185aac41800df8eae4b60501ccc0dbbe51)
---
 drivers/gpu/drm/xe/xe_dma_buf.c |   11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)
drivers/gpu/drm/xe/xe_dma_buf.c +7 -4
@@ -347,12 +347,15 @@ struct drm_gem_object *xe_gem_prime_impo
 		goto out_err;
 	}
 
-	/* Errors here will take care of freeing the bo. */
+	/*
+	 * xe_dma_buf_init_obj() takes ownership of bo on both success
+	 * and failure, so we must not touch bo after this call.
+	 */
 	obj = xe_dma_buf_init_obj(dev, bo, dma_buf);
-	if (IS_ERR(obj))
+	if (IS_ERR(obj)) {
+		dma_buf_detach(dma_buf, attach);
 		return obj;
-
-
+	}
 	get_dma_buf(dma_buf);
 	obj->import_attach = attach;
 	return obj;
author maintainer
#119 lore raw

PATCH 119/188 drm/xe/bo: Fix bo leak on unaligned size validation in xe_bo_init_locked()

In reply to Greg Kroah-Hartman

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

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

From: Shuicheng Lin <[email protected]>

commit 09a8f3c1c11977a6e10c167f26dd298790b31c32 upstream.

When type is ttm_bo_type_device and aligned_size != size, the function returns an error without freeing a caller-provided bo, violating the documented contract that bo is freed on failure.

Add xe_bo_free(bo) before returning the error.

Fixes 4e03b584143e ("drm/xe/uapi: Reject bo creation of unaligned size")
Assisted-by Claude:claude-opus-4.6
(cherry picked from commit 601c2aa087b6f21014300a3f107a08ee4dde7bdf)
---
 drivers/gpu/drm/xe/xe_bo.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/gpu/drm/xe/xe_bo.c +3 -1
@@ -2134,8 +2134,10 @@ struct xe_bo *xe_bo_init_locked(struct x
 		alignment = SZ_4K >> PAGE_SHIFT;
 	}
 
-	if (type == ttm_bo_type_device && aligned_size != size)
+	if (type == ttm_bo_type_device && aligned_size != size) {
+		xe_bo_free(bo);
 		return ERR_PTR(-EINVAL);
+	}
 
 	if (!bo) {
 		bo = xe_bo_alloc();
author maintainer
#120 lore raw

PATCH 120/188 drm/xe/uapi: Reject coh_none PAT index for CPU cached memory in madvise

In reply to Greg Kroah-Hartman

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

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

From: Jia Yao <[email protected]>

commit 4e5591c2fc1b30f4ea5e2eab4c3a695acc404e39 upstream.

Add validation in xe_vm_madvise_ioctl() to reject PAT indices with XE_COH_NONE coherency mode when applied to CPU cached memory.

Using coh_none with CPU cached buffers is a security issue. When the kernel clears pages before reallocation, the clear operation stays in CPU cache (dirty). GPU with coh_none can bypass CPU caches and read stale sensitive data directly from DRAM, potentially leaking data from previously freed pages of other processes.

This aligns with the existing validation in vm_bind path (xe_vm_bind_ioctl_validate_bo).

v2(Matthew brost)
- Add fixes
- Move one debug print to better place

v3(Matthew Auld)
- Should be drm/xe/uapi
- More Cc

v4(Shuicheng Lin)
- Fix kmem leak issues by the way

v5
- Remove kmem leak because it has been merged by another patch

v6
- Remove the fix which is not related to current fix

v7
- No change

v8
- Rebase

v9
- Limit the restrictions to iGPU

v10
- No change

Fixes ada7486c5668 ("drm/xe: Implement madvise ioctl for xe")
Cc <[email protected]> # v6.18+
Cc Shuicheng Lin <[email protected]>
Cc Mathew Alwin <[email protected]>
Cc Michal Mrozek <[email protected]>
Cc Matthew Brost <[email protected]>
Cc Matthew Auld <[email protected]>
(cherry picked from commit 016ccdb674b8c899940b3944952c96a6a490d10a)
---
 drivers/gpu/drm/xe/xe_vm_madvise.c |   47 +++++++++++++++++++++++++++++++++++++
 1 file changed, 47 insertions(+)
drivers/gpu/drm/xe/xe_vm_madvise.c +47 -0
@@ -299,6 +299,45 @@ static bool madvise_args_are_sane(struct
 	return true;
 }
 
+static bool check_pat_args_are_sane(struct xe_device *xe,
+				    struct xe_vmas_in_madvise_range *madvise_range,
+				    u16 pat_index)
+{
+	u16 coh_mode = xe_pat_index_get_coh_mode(xe, pat_index);
+	int i;
+
+	/*
+	 * Using coh_none with CPU cached buffers is not allowed on iGPU.
+	 * On iGPU the GPU shares the LLC with the CPU, so with coh_none
+	 * the GPU bypasses CPU caches and reads directly from DRAM,
+	 * potentially seeing stale sensitive data from previously freed
+	 * pages. On dGPU this restriction does not apply, because the
+	 * platform does not provide a non-coherent system memory access
+	 * path that would violate the DMA coherency contract.
+	 */
+	if (coh_mode != XE_COH_NONE || IS_DGFX(xe))
+		return true;
+
+	for (i = 0; i < madvise_range->num_vmas; i++) {
+		struct xe_vma *vma = madvise_range->vmas[i];
+		struct xe_bo *bo = xe_vma_bo(vma);
+
+		if (bo) {
+			/* BO with WB caching + COH_NONE is not allowed */
+			if (XE_IOCTL_DBG(xe, bo->cpu_caching == DRM_XE_GEM_CPU_CACHING_WB))
+				return false;
+			/* Imported dma-buf without caching info, assume cached */
+			if (XE_IOCTL_DBG(xe, !bo->cpu_caching))
+				return false;
+		} else if (XE_IOCTL_DBG(xe, xe_vma_is_cpu_addr_mirror(vma) ||
+					    xe_vma_is_userptr(vma)))
+			/* System memory (userptr/SVM) is always CPU cached */
+			return false;
+	}
+
+	return true;
+}
+
 static bool check_bo_args_are_sane(struct xe_vm *vm, struct xe_vma **vmas,
 				   int num_vmas, u32 atomic_val)
 {
@@ -384,6 +423,14 @@ int xe_vm_madvise_ioctl(struct drm_devic
 	if (err || !madvise_range.num_vmas)
 		goto unlock_vm;
 
+	if (args->type == DRM_XE_MEM_RANGE_ATTR_PAT) {
+		if (!check_pat_args_are_sane(xe, &madvise_range,
+					     args->pat_index.val)) {
+			err = -EINVAL;
+			goto free_vmas;
+		}
+	}
+
 	if (madvise_range.has_bo_vmas) {
 		if (args->type == DRM_XE_MEM_RANGE_ATTR_ATOMIC) {
 			if (!check_bo_args_are_sane(vm, madvise_range.vmas,
author maintainer
#121 lore raw

PATCH 121/188 drm: Set old handle to NULL before prime swap in change_handle

In reply to Greg Kroah-Hartman

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

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

From: Francis, David <[email protected]>

commit 5e28b7b94408897e41c63477aabc9e1db439bc8c upstream.

There was a potential race condition in change_handle. The ioctl briefly had a single object with two idr entries; a concurrent gem_close could delete the object and remove one of the handles while leaving the other one dangling, which could subsequently be dereferenced for a use-after-free.

To fix this, do the same dance that gem_close itself does.
(f6cd7daecff5 drm: Release driver references to handle before making it available again) First idr_replace the old handle to NULL. Later, if the prime
operations are successful, actually close it.

create_tail required a similar dance to avoid a similar problem. (bd46cece51a3 drm/gem: Fix race in drm_gem_handle_create_tail()) It idr_allocs the new handle with NULL, then swaps in the correct object later to avoid races. We don't need to do that here, since the only operations that could race are drm_prime, and change_handle holds the prime lock for the entire duration.

v2: cleanups of error paths

Cc Simona Vetter <[email protected]>
Cc Christian Koenig <[email protected]>
Fixes 53096728b8910 ("drm: Add DRM prime interface to reassign GEM handle")
---
 drivers/gpu/drm/drm_gem.c |   25 ++++++++++++++++++++++++-
 1 file changed, 24 insertions(+), 1 deletion(-)
drivers/gpu/drm/drm_gem.c +24 -1
@@ -969,7 +969,7 @@ int drm_gem_change_handle_ioctl(struct d
 				struct drm_file *file_priv)
 {
 	struct drm_gem_change_handle *args = data;
-	struct drm_gem_object *obj;
+	struct drm_gem_object *obj, *idrobj;
 	int handle, ret;
 
 	if (!drm_core_check_feature(dev, DRIVER_GEM))
@@ -992,8 +992,29 @@ int drm_gem_change_handle_ioctl(struct d
 	mutex_lock(&file_priv->prime.lock);
 
 	spin_lock(&file_priv->table_lock);
+
+       /* When create_tail allocs an obj idr, it needs to first alloc as NULL,
+	* then later replace with the correct object. This is not necessary
+	* here, because the only operations that could race are drm_prime
+	* bookkeeping, and we hold the prime lock.
+	*/
 	ret = idr_alloc(&file_priv->object_idr, obj, handle, handle + 1,
 			GFP_NOWAIT);
+
+       if (ret < 0) {
+	       spin_unlock(&file_priv->table_lock);
+	       goto out_unlock;
+       }
+
+       idrobj = idr_replace(&file_priv->object_idr, NULL, handle);
+       if (idrobj != obj) {
+	       idr_replace(&file_priv->object_idr, idrobj, handle);
+	       idr_remove(&file_priv->object_idr, args->new_handle);
+	       spin_unlock(&file_priv->table_lock);
+	       ret = -ENOENT;
+	       goto out_unlock;
+       }
+
 	spin_unlock(&file_priv->table_lock);
 
 	if (ret < 0)
@@ -1005,6 +1026,8 @@ int drm_gem_change_handle_ioctl(struct d
 		if (ret < 0) {
 			spin_lock(&file_priv->table_lock);
 			idr_remove(&file_priv->object_idr, handle);
+			idrobj = idr_replace(&file_priv->object_idr, obj, handle);
+			WARN_ON(idrobj != NULL);
 			spin_unlock(&file_priv->table_lock);
 			goto out_unlock;
 		}
author maintainer
#122 lore raw

PATCH 122/188 drm/radeon: add missing revision check for CI

In reply to Greg Kroah-Hartman

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

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

From: Alex Deucher <[email protected]>

commit 17223816498f7b117d138d18eb0eba63604dc74e upstream.

The memory level workarounds only apply to revision 0 SKUs.

Fixes 127e056e2a82 ("drm/radeon: fix mclk vddc configuration for cards for hawaii")
Fixes 21b8a369046f ("drm/radeon: fix dram timing for certain hawaii boards")
Fixes 90b2fee35cb9 ("drm/radeon: fix dpm mc init for certain hawaii boards")
(cherry picked from commit 4d8dcc14311515077062b5740f39f427075de5c9)
---
 drivers/gpu/drm/radeon/ci_dpm.c |    9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)
drivers/gpu/drm/radeon/ci_dpm.c +6 -3
@@ -2461,7 +2461,8 @@ static void ci_register_patching_mc_arb(
 
 	if (patch &&
 	    ((rdev->pdev->device == 0x67B0) ||
-	     (rdev->pdev->device == 0x67B1))) {
+	     (rdev->pdev->device == 0x67B1)) &&
+	    (rdev->pdev->revision == 0)) {
 		if ((memory_clock > 100000) && (memory_clock <= 125000)) {
 			tmp2 = (((0x31 * engine_clock) / 125000) - 1) & 0xff;
 			*dram_timimg2 &= ~0x00ff0000;
@@ -3304,7 +3305,8 @@ static int ci_populate_all_memory_levels
 	pi->smc_state_table.MemoryLevel[0].EnabledForActivity = 1;
 
 	if ((dpm_table->mclk_table.count >= 2) &&
-	    ((rdev->pdev->device == 0x67B0) || (rdev->pdev->device == 0x67B1))) {
+	    ((rdev->pdev->device == 0x67B0) || (rdev->pdev->device == 0x67B1)) &&
+	    (rdev->pdev->revision == 0)) {
 		pi->smc_state_table.MemoryLevel[1].MinVddc =
 			pi->smc_state_table.MemoryLevel[0].MinVddc;
 		pi->smc_state_table.MemoryLevel[1].MinVddcPhases =
@@ -4493,7 +4495,8 @@ static int ci_register_patching_mc_seq(s
 
 	if (patch &&
 	    ((rdev->pdev->device == 0x67B0) ||
-	     (rdev->pdev->device == 0x67B1))) {
+	     (rdev->pdev->device == 0x67B1)) &&
+	    (rdev->pdev->revision == 0)) {
 		for (i = 0; i < table->last; i++) {
 			if (table->last >= SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE)
 				return -EINVAL;
author maintainer
#123 lore raw

PATCH 123/188 drm/amdgpu: zero-initialize GART table on allocation

In reply to Greg Kroah-Hartman

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

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

From: Philip Yang <[email protected]>

commit e6c2e6c2e1fa066968a16aca1cb66cd1bdde7741 upstream.

GART TLB is flushed after unmapping but not after mapping. Since amdgpu_bo_create_kernel() does not zero-initialize the buffer, when a single PTE is written the TLB may speculatively load other uninitialized entries from the same cacheline. Those garbage entries can appear valid, and a subsequent write to another PTE in the same cacheline may cause the GPU to use a stale garbage PTE from the TLB.

Fix this by calling memset_io() to zero-initialize the GART table with gart_pte_flags immediately after allocation.

Using AMDGPU_GEM_CREATE_VRAM_CLEARED, SDMA-based clear will not work since SDMA needs GART to be initialized to work.

(cherry picked from commit d9af8263b82b6eaa60c5718e0c6631c5037e4b24)
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c |   13 ++++++++++---
 1 file changed, 10 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c +10 -3
@@ -262,12 +262,19 @@ void amdgpu_gart_table_ram_free(struct a
  */
 int amdgpu_gart_table_vram_alloc(struct amdgpu_device *adev)
 {
+	int r;
+
 	if (adev->gart.bo != NULL)
 		return 0;
 
-	return amdgpu_bo_create_kernel(adev,  adev->gart.table_size, PAGE_SIZE,
-				       AMDGPU_GEM_DOMAIN_VRAM, &adev->gart.bo,
-				       NULL, (void *)&adev->gart.ptr);
+	r = amdgpu_bo_create_kernel(adev,  adev->gart.table_size, PAGE_SIZE,
+				    AMDGPU_GEM_DOMAIN_VRAM, &adev->gart.bo,
+				    NULL, (void *)&adev->gart.ptr);
+	if (r)
+		return r;
+
+	memset_io(adev->gart.ptr, adev->gart.gart_pte_flags, adev->gart.table_size);
+	return 0;
 }
 
 /**
author maintainer
#124 lore raw

PATCH 124/188 drm/exynos: remove bridge when component_add fails

In reply to Greg Kroah-Hartman

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

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

From: Osama Abdelkader <[email protected]>

commit 26f6654a9a60eb4d241f42a0ec85412e8821480b upstream.

Use devm_drm_bridge_add() so the bridge is released if probe fails after registration, and drop the manual drm_bridge_remove() in remove().

Check the return value of devm_drm_bridge_add().

---
 drivers/gpu/drm/exynos/exynos_drm_mic.c |    8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)
drivers/gpu/drm/exynos/exynos_drm_mic.c +3 -5
@@ -423,7 +423,9 @@ static int exynos_mic_probe(struct platf
 
 	mic->bridge.of_node = dev->of_node;
 
-	drm_bridge_add(&mic->bridge);
+	ret = devm_drm_bridge_add(dev, &mic->bridge);
+	if (ret)
+		goto err;
 
 	pm_runtime_enable(dev);
 
@@ -443,12 +445,8 @@ err:
 
 static void exynos_mic_remove(struct platform_device *pdev)
 {
-	struct exynos_mic *mic = platform_get_drvdata(pdev);
-
 	component_del(&pdev->dev, &exynos_mic_component_ops);
 	pm_runtime_disable(&pdev->dev);
-
-	drm_bridge_remove(&mic->bridge);
 }
 
 static const struct of_device_id exynos_mic_of_match[] = {
author maintainer
#125 lore raw

PATCH 125/188 drm/panel: himax-hx83102: restore MODE_LPM after sending disable cmds

In reply to Greg Kroah-Hartman

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

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

From: Icenowy Zheng <[email protected]>

commit 2d4e80271f784aa0c7b17676e9762c7e8156be1c upstream.

When preparing the panel, it seems that it always expects commands to be transferred in LP mode. However, the disable function removes the MIPI_DSI_MODE_LPM flag, and no other function re-adds it.

As the unprepare function contains no DSI commands, re-adding the flag just after disabling the panel should be safe. Add the code re-adding the flag after the two commands for disabling the panel are sent.

This fixes screen unblanking (after blanking once) on mt8188-geralt-ciri-sku1 device.

---
 drivers/gpu/drm/panel/panel-himax-hx83102.c |    2 ++
 1 file changed, 2 insertions(+)
drivers/gpu/drm/panel/panel-himax-hx83102.c +2 -0
@@ -850,6 +850,8 @@ static int hx83102_disable(struct drm_pa
 	mipi_dsi_dcs_set_display_off_multi(&dsi_ctx);
 	mipi_dsi_dcs_enter_sleep_mode_multi(&dsi_ctx);
 
+	dsi->mode_flags |= MIPI_DSI_MODE_LPM;
+
 	mipi_dsi_msleep(&dsi_ctx, 150);
 
 	return dsi_ctx.accum_err;
author maintainer
#126 lore raw

PATCH 126/188 drm/amdgpu/gfx9: drop unnecessary 64-bit fence flag check in KIQ

In reply to Greg Kroah-Hartman

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

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

From: John B. Moore <[email protected]>

commit 7bbfb2559bcec39d1a4e1182d931a2046112c352 upstream.

Remove the BUG_ON(flags & AMDGPU_FENCE_FLAG_64BIT) assertion from gfx_v9_0_ring_emit_fence_kiq(). The KIQ hardware supports 64-bit fence writes; the 32-bit writeback address constraint is an upper-layer convention, not a hardware limitation. The check serves no purpose and should not be present.

Found by code inspection while investigating related BUG_ON assertions in the GFX and compute ring emission paths.

(cherry picked from commit 1b1101a46a426bb4328116bb5273c326a2780389)
---
 drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c |    3 ---
 1 file changed, 3 deletions(-)
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +0 -3
@@ -5640,9 +5640,6 @@ static void gfx_v9_0_ring_emit_fence_kiq
 {
 	struct amdgpu_device *adev = ring->adev;
 
-	/* we only allocate 32bit for each seq wb address */
-	BUG_ON(flags & AMDGPU_FENCE_FLAG_64BIT);
-
 	/* write fence seq to the "addr" */
 	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
 	amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
author maintainer
#127 lore raw

PATCH 127/188 drm/bridge: tda998x: Use __be32 for audio port OF property pointer

In reply to Greg Kroah-Hartman

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

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

From: Kory Maincent (TI) <[email protected]>

commit 2a46a9356ba7b1bdd741c8b41e5374edcd960557 upstream.

of_get_property() returns a pointer to big-endian (__be32) data, but port_data in tda998x_get_audio_ports() was declared as const u32 *, causing a sparse endianness type mismatch warning. Fix the declaration to use const __be32 *.

Fixes 7e567624dc5a4 ("drm/i2c: tda998x: Register ASoC hdmi-codec and add audio DT binding")
---
 drivers/gpu/drm/bridge/tda998x_drv.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/bridge/tda998x_drv.c +1 -1
@@ -1697,7 +1697,7 @@ static const struct drm_bridge_funcs tda
 static int tda998x_get_audio_ports(struct tda998x_priv *priv,
 				   struct device_node *np)
 {
-	const u32 *port_data;
+	const __be32 *port_data;
 	u32 size;
 	int i;
 
author maintainer
#128 lore raw

PATCH 128/188 drm/panel: boe-tv101wum-nl6: restore MODE_LPM after sending disable cmds

In reply to Greg Kroah-Hartman

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

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

From: Icenowy Zheng <[email protected]>

commit 570cf799e87ae805eacfab3b4ba66676b5fccdb6 upstream.

When preparing the panel, it seems that it always expects commands to be transferred in LP mode. However, the disable function removes the MIPI_DSI_MODE_LPM flag, and no other function re-adds it.

As the unprepare function contains no DSI commands, re-adding the flag just after disabling the panel should be safe. Add the code re-adding the flag after the two commands for disabling the panel are sent.

This fixes error messages shown in kernel log when unblanking on mt8183-kukui-kodama-sku32 device.

---
 drivers/gpu/drm/panel/panel-boe-tv101wum-nl6.c |    2 ++
 1 file changed, 2 insertions(+)
drivers/gpu/drm/panel/panel-boe-tv101wum-nl6.c +2 -0
@@ -1324,6 +1324,8 @@ static int boe_panel_disable(struct drm_
 	mipi_dsi_dcs_set_display_off_multi(&ctx);
 	mipi_dsi_dcs_enter_sleep_mode_multi(&ctx);
 
+	boe->dsi->mode_flags |= MIPI_DSI_MODE_LPM;
+
 	mipi_dsi_msleep(&ctx, 150);
 
 	return ctx.accum_err;
author maintainer
#129 lore raw

PATCH 129/188 drm/amdkfd: Make all TLB-flushes heavy-weight

In reply to Greg Kroah-Hartman

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

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

From: Felix Kuehling <[email protected]>

commit 9b4e3495d1bd2469bf94b74930c153c2d534ddb7 upstream.

With only one sequence number we cannot track the need for legacy vs heavy-weight flushes reliably. Always use heavy-weight.

(cherry picked from commit c1a3ff1d327820cd9a52bc1056b98681fc088949)
---
 drivers/gpu/drm/amd/amdkfd/kfd_chardev.c              |    4 ++--
 drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c |    6 +++---
 drivers/gpu/drm/amd/amdkfd/kfd_priv.h                 |    6 +++---
 drivers/gpu/drm/amd/amdkfd/kfd_svm.c                  |    4 ++--
 4 files changed, 10 insertions(+), 10 deletions(-)
drivers/gpu/drm/amd/amdkfd/kfd_chardev.c +2 -2
@@ -1341,7 +1341,7 @@ static int kfd_ioctl_map_memory_to_gpu(s
 		peer_pdd = kfd_process_device_data_by_id(p, devices_arr[i]);
 		if (WARN_ON_ONCE(!peer_pdd))
 			continue;
-		kfd_flush_tlb(peer_pdd, TLB_FLUSH_LEGACY);
+		kfd_flush_tlb(peer_pdd);
 	}
 	kfree(devices_arr);
 
@@ -1436,7 +1436,7 @@ static int kfd_ioctl_unmap_memory_from_g
 		if (WARN_ON_ONCE(!peer_pdd))
 			continue;
 		if (flush_tlb)
-			kfd_flush_tlb(peer_pdd, TLB_FLUSH_HEAVYWEIGHT);
+			kfd_flush_tlb(peer_pdd);
 
 		/* Remove dma mapping after tlb flush to avoid IO_PAGE_FAULT */
 		err = amdgpu_amdkfd_gpuvm_dmaunmap_mem(mem, peer_pdd->drm_priv);
drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c +3 -3
@@ -569,7 +569,7 @@ static int allocate_vmid(struct device_q
 			qpd->vmid,
 			qpd->page_table_base);
 	/* invalidate the VM context after pasid and vmid mapping is set up */
-	kfd_flush_tlb(qpd_to_pdd(qpd), TLB_FLUSH_LEGACY);
+	kfd_flush_tlb(qpd_to_pdd(qpd));
 
 	if (dqm->dev->kfd2kgd->set_scratch_backing_va)
 		dqm->dev->kfd2kgd->set_scratch_backing_va(dqm->dev->adev,
@@ -607,7 +607,7 @@ static void deallocate_vmid(struct devic
 		if (flush_texture_cache_nocpsch(q->device, qpd))
 			dev_err(dev, "Failed to flush TC\n");
 
-	kfd_flush_tlb(qpd_to_pdd(qpd), TLB_FLUSH_LEGACY);
+	kfd_flush_tlb(qpd_to_pdd(qpd));
 
 	/* Release the vmid mapping */
 	set_pasid_vmid_mapping(dqm, 0, qpd->vmid);
@@ -1282,7 +1282,7 @@ static int restore_process_queues_nocpsc
 				dqm->dev->adev,
 				qpd->vmid,
 				qpd->page_table_base);
-		kfd_flush_tlb(pdd, TLB_FLUSH_LEGACY);
+		kfd_flush_tlb(pdd);
 	}
 
 	/* Take a safe reference to the mm_struct, which may otherwise
drivers/gpu/drm/amd/amdkfd/kfd_priv.h +3 -3
@@ -1533,13 +1533,13 @@ void kfd_signal_reset_event(struct kfd_n
 
 void kfd_signal_poison_consumed_event(struct kfd_node *dev, u32 pasid);
 
-static inline void kfd_flush_tlb(struct kfd_process_device *pdd,
-				 enum TLB_FLUSH_TYPE type)
+static inline void kfd_flush_tlb(struct kfd_process_device *pdd)
 {
 	struct amdgpu_device *adev = pdd->dev->adev;
 	struct amdgpu_vm *vm = drm_priv_to_vm(pdd->drm_priv);
 
-	amdgpu_vm_flush_compute_tlb(adev, vm, type, pdd->dev->xcc_mask);
+	amdgpu_vm_flush_compute_tlb(adev, vm, TLB_FLUSH_HEAVYWEIGHT,
+				    pdd->dev->xcc_mask);
 }
 
 static inline bool kfd_flush_tlb_after_unmap(struct kfd_dev *dev)
drivers/gpu/drm/amd/amdkfd/kfd_svm.c +2 -2
@@ -1391,7 +1391,7 @@ svm_range_unmap_from_gpus(struct svm_ran
 			if (r)
 				break;
 		}
-		kfd_flush_tlb(pdd, TLB_FLUSH_HEAVYWEIGHT);
+		kfd_flush_tlb(pdd);
 	}
 
 	return r;
@@ -1525,7 +1525,7 @@ svm_range_map_to_gpus(struct svm_range *
 			}
 		}
 
-		kfd_flush_tlb(pdd, TLB_FLUSH_LEGACY);
+		kfd_flush_tlb(pdd);
 	}
 
 	return r;
author maintainer
#130 lore raw

PATCH 130/188 drm/amdgpu/sdma4: replace BUG_ON with WARN_ON in fence emission

In reply to Greg Kroah-Hartman

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

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

From: John B. Moore <[email protected]>

commit 78d2e624fa073c14970aa097adcf3ea31c157a66 upstream.

sdma_v4_0_ring_emit_fence() contains two BUG_ON(addr & 0x3) assertions that verify fence writeback addresses are dword-aligned. These assertions can be reached from unprivileged userspace via crafted DRM_IOCTL_AMDGPU_CS submissions, causing a fatal kernel panic in a scheduler worker thread.

Replace both BUG_ON() calls with WARN_ON() to log the condition without crashing the kernel. A misaligned fence address at this point indicates a driver bug, but crashing the kernel is never the correct response when the assertion is reachable from userspace.

The CS IOCTL path is the correct place to filter invalid submissions; the ring emission callback is too late to do anything about it.

Fixes 2130f89ced2c ("drm/amdgpu: add SDMA v4.0 implementation (v2)")
(cherry picked from commit b90250bd933afd1ba94d86d6b13821997b22b18e)
---
 drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c +2 -2
@@ -890,7 +890,7 @@ static void sdma_v4_0_ring_emit_fence(st
 	/* write the fence */
 	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_FENCE));
 	/* zero in first two bits */
-	BUG_ON(addr & 0x3);
+	WARN_ON(addr & 0x3);
 	amdgpu_ring_write(ring, lower_32_bits(addr));
 	amdgpu_ring_write(ring, upper_32_bits(addr));
 	amdgpu_ring_write(ring, lower_32_bits(seq));
@@ -900,7 +900,7 @@ static void sdma_v4_0_ring_emit_fence(st
 		addr += 4;
 		amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_FENCE));
 		/* zero in first two bits */
-		BUG_ON(addr & 0x3);
+		WARN_ON(addr & 0x3);
 		amdgpu_ring_write(ring, lower_32_bits(addr));
 		amdgpu_ring_write(ring, upper_32_bits(addr));
 		amdgpu_ring_write(ring, upper_32_bits(seq));
author maintainer
#131 lore raw

PATCH 131/188 drm/amdgpu/pm: add missing revision check for CI

In reply to Greg Kroah-Hartman

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

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

From: Alex Deucher <[email protected]>

commit 2a561b361b7681509710f3cfc3d95d54c87ac69f upstream.

The ci_populate_all_memory_levels() workaround only applies to revision 0 SKUs.

Fixes 9f4b35411cfe ("drm/amd/powerplay: add CI asics support to smumgr (v3)")
(cherry picked from commit 1db15ba8f72f400bbad8ae0ce24fafc43429d4bd)
---
 drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c |    5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c +3 -2
@@ -1326,8 +1326,9 @@ static int ci_populate_all_memory_levels
 
 	dev_id = adev->pdev->device;
 
-	if ((dpm_table->mclk_table.count >= 2)
-		&& ((dev_id == 0x67B0) ||  (dev_id == 0x67B1))) {
+	if ((dpm_table->mclk_table.count >= 2) &&
+	    ((dev_id == 0x67B0) ||  (dev_id == 0x67B1)) &&
+	    (adev->pdev->revision == 0)) {
 		smu_data->smc_state_table.MemoryLevel[1].MinVddci =
 				smu_data->smc_state_table.MemoryLevel[0].MinVddci;
 		smu_data->smc_state_table.MemoryLevel[1].MinMvdd =
author maintainer
#132 lore raw

PATCH 132/188 drm/amdgpu/pm: align Hawaii mclk workaround with radeon

In reply to Greg Kroah-Hartman

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

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

From: Alex Deucher <[email protected]>

commit 1987c79b4fe5789dfa14423e78b5c25f6acf3e9d upstream.

Align the hawaii mclk workaround with radeon and windows.

Fixes 9f4b35411cfe ("drm/amd/powerplay: add CI asics support to smumgr (v3)")
(cherry picked from commit 9649528b637f668c5af9f2b83ca4ad8576ae2121)
---
 drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c |    8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)
drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c +4 -4
@@ -1329,10 +1329,10 @@ static int ci_populate_all_memory_levels
 	if ((dpm_table->mclk_table.count >= 2) &&
 	    ((dev_id == 0x67B0) ||  (dev_id == 0x67B1)) &&
 	    (adev->pdev->revision == 0)) {
-		smu_data->smc_state_table.MemoryLevel[1].MinVddci =
-				smu_data->smc_state_table.MemoryLevel[0].MinVddci;
-		smu_data->smc_state_table.MemoryLevel[1].MinMvdd =
-				smu_data->smc_state_table.MemoryLevel[0].MinMvdd;
+		smu_data->smc_state_table.MemoryLevel[1].MinVddc =
+				smu_data->smc_state_table.MemoryLevel[0].MinVddc;
+		smu_data->smc_state_table.MemoryLevel[1].MinVddcPhases =
+				smu_data->smc_state_table.MemoryLevel[0].MinVddcPhases;
 	}
 	smu_data->smc_state_table.MemoryLevel[0].ActivityLevel = 0x1F;
 	CONVERT_FROM_HOST_TO_SMC_US(smu_data->smc_state_table.MemoryLevel[0].ActivityLevel);
author maintainer
#133 lore raw

PATCH 133/188 arm64: dts: qcom: lemans: Correct QUP interrupt numbers

In reply to Greg Kroah-Hartman

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

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

From: Viken Dadhaniya <[email protected]>

commit c5b22c88cc09b180e3a23010b29f4d02ec117a44 upstream.

Fix GIC_SPI interrupt numbers for QUPv3 SE6 nodes on Lemans SoC. Using incorrect interrupt lines can prevent IRQs from triggering and break I2C, SPI, and UART operation.

Fixes 34a407316b7d3 ("arm64: dts: qcom: sa8775p: Populate additional UART DT nodes")
Fixes 1b2d7ad5ac14d ("arm64: dts: qcom: sa8775p: add missing spi nodes")
Fixes ee2f5f906d69d ("arm64: dts: qcom: sa8775p: add missing i2c nodes")
---
 arch/arm64/boot/dts/qcom/lemans.dtsi |    8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)
arch/arm64/boot/dts/qcom/lemans.dtsi +4 -4
@@ -1563,7 +1563,7 @@
 				reg = <0x0 0x898000 0x0 0x4000>;
 				#address-cells = <1>;
 				#size-cells = <0>;
-				interrupts = <GIC_SPI 834 IRQ_TYPE_LEVEL_HIGH>;
+				interrupts = <GIC_SPI 833 IRQ_TYPE_LEVEL_HIGH>;
 				clocks = <&gcc GCC_QUPV3_WRAP2_S6_CLK>;
 				clock-names = "se";
 				pinctrl-0 = <&qup_i2c20_default>;
@@ -1590,7 +1590,7 @@
 				reg = <0x0 0x898000 0x0 0x4000>;
 				#address-cells = <1>;
 				#size-cells = <0>;
-				interrupts = <GIC_SPI 834 IRQ_TYPE_LEVEL_HIGH>;
+				interrupts = <GIC_SPI 833 IRQ_TYPE_LEVEL_HIGH>;
 				clocks = <&gcc GCC_QUPV3_WRAP2_S6_CLK>;
 				clock-names = "se";
 				pinctrl-0 = <&qup_spi20_default>;
@@ -1615,7 +1615,7 @@
 			uart20: serial@898000 {
 				compatible = "qcom,geni-uart";
 				reg = <0x0 0x00898000 0x0 0x4000>;
-				interrupts = <GIC_SPI 834 IRQ_TYPE_LEVEL_HIGH>;
+				interrupts = <GIC_SPI 833 IRQ_TYPE_LEVEL_HIGH>;
 				clocks = <&gcc GCC_QUPV3_WRAP2_S6_CLK>;
 				clock-names = "se";
 				pinctrl-0 = <&qup_uart20_default>;
@@ -2561,7 +2561,7 @@
 				reg = <0x0 0xa98000 0x0 0x4000>;
 				#address-cells = <1>;
 				#size-cells = <0>;
-				interrupts = <GIC_SPI 836 IRQ_TYPE_LEVEL_HIGH>;
+				interrupts = <GIC_SPI 835 IRQ_TYPE_LEVEL_HIGH>;
 				clocks = <&gcc GCC_QUPV3_WRAP1_S6_CLK>;
 				clock-names = "se";
 				pinctrl-0 = <&qup_i2c13_default>;
author maintainer
#134 lore raw

PATCH 134/188 arm64: dts: ti: k3-am62a7-sk: Fix pin name in comment from M19 to N22

In reply to Greg Kroah-Hartman

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

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

From: Siddharth Vadapalli <[email protected]>

commit 6ee0792d83d5c690205c350825a4c30746c0e0a2 upstream.

The pin for GPMC0_CLK.GPIO0_31 at address 0x000F407C is N22 and not M19. Hence, fix the pin name in the comment to avoid confusion.

---
 arch/arm64/boot/dts/ti/k3-am62a7-sk.dts |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm64/boot/dts/ti/k3-am62a7-sk.dts +1 -1
@@ -394,7 +394,7 @@
 
 	vddshv_sdio_pins_default: vddshv-sdio-default-pins {
 		pinctrl-single,pins = <
-			AM62AX_IOPAD(0x07c, PIN_OUTPUT, 7) /* (M19) GPMC0_CLK.GPIO0_31 */
+			AM62AX_IOPAD(0x07c, PIN_OUTPUT, 7) /* (N22) GPMC0_CLK.GPIO0_31 */
 		>;
 	};
 
author maintainer
#135 lore raw

PATCH 135/188 sctp: revalidate list cursor after sctp_sendmsg_to_asoc() in SCTP_SENDALL

In reply to Greg Kroah-Hartman

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

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

From: Ben Morris <[email protected]>

commit abb5f36771cc4c05899b34000829a787572a8817 upstream.

The SCTP_SENDALL path in sctp_sendmsg() iterates ep->asocs with list_for_each_entry_safe(), which caches the next entry in @tmp before the loop body runs. The body calls sctp_sendmsg_to_asoc(), which may drop the socket lock inside sctp_wait_for_sndbuf().

While the lock is dropped, another thread can SCTP_SOCKOPT_PEELOFF the association cached in @tmp, migrating it to a new endpoint via sctp_sock_migrate() (list_del_init() + list_add_tail() to newep->asocs), and optionally close the new socket which frees the association via kfree_rcu(). The cached @tmp can also be freed by a network ABORT for that association, processed in softirq while the lock is dropped.

sctp_wait_for_sndbuf() revalidates @asoc (the current entry) on re-lock via the "sk != asoc->base.sk" and "asoc->base.dead" checks, but nothing revalidates @tmp. After a successful return, the iterator advances to the stale @tmp, yielding either a use-after-free (if the peeled socket was closed) or a list-walk onto the new endpoint's list head (type confusion of &newep->asocs as a struct sctp_association *).

Both are reachable from CapEff=0; the type-confusion path gives controlled indirect call via the outqueue.sched->init_sid pointer.

Fix by re-deriving @tmp from @asoc after sctp_sendmsg_to_asoc() returns. @asoc is known to still be on ep->asocs at that point: the only callers that list_del an association from ep->asocs are sctp_association_free() (which sets asoc->base.dead) and sctp_assoc_migrate() (which changes asoc->base.sk), and sctp_wait_for_sndbuf() checks both under the lock before any successful return; a tripped check propagates as err < 0 and the loop bails before the re-derive.

The SCTP_ABORT path in sctp_sendmsg_check_sflags() returns 0 and the loop hits 'continue' before sctp_sendmsg_to_asoc() is ever called, so the @tmp cached by list_for_each_entry_safe() still covers the lock-held free that ba59fb027307 ("sctp: walk the list of asoc safely") was added for.

Fixes 4910280503f3 ("sctp: add support for snd flag SCTP_SENDALL process in sendmsg")
---
 net/sctp/socket.c |    9 +++++++++
 1 file changed, 9 insertions(+)
net/sctp/socket.c +9 -0
@@ -1985,6 +1985,15 @@ static int sctp_sendmsg(struct sock *sk,
 				goto out_unlock;
 
 			iov_iter_revert(&msg->msg_iter, err);
+
+			/* sctp_sendmsg_to_asoc() may have released the socket
+			 * lock (sctp_wait_for_sndbuf), during which other
+			 * associations on ep->asocs could have been peeled
+			 * off or freed.  @asoc itself is revalidated by the
+			 * base.dead and base.sk checks in sctp_wait_for_sndbuf,
+			 * so re-derive the cached cursor from it.
+			 */
+			tmp = list_next_entry(asoc, asocs);
 		}
 
 		goto out_unlock;
author maintainer
#136 lore raw

PATCH 136/188 batman-adv: fix integer overflow on buff_pos

In reply to Greg Kroah-Hartman

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

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

From: Lyes Bourennani <[email protected]>

commit 0799e5943611006b346b8813c7daf7dd5aa26bfd upstream.

Fixing an integer overflow present in batadv_iv_ogm_send_to_if. The size check is done using the int type in batadv_iv_ogm_aggr_packet whereas the buff_pos variable uses the s16 type. This could lead to an out-of-bound read.

Fixes c6c8fea29769 ("net: Add batman-adv meshing protocol")
---
 net/batman-adv/bat_iv_ogm.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/batman-adv/bat_iv_ogm.c +1 -1
@@ -335,7 +335,7 @@ static void batadv_iv_ogm_send_to_if(str
 	struct batadv_priv *bat_priv = netdev_priv(hard_iface->mesh_iface);
 	const char *fwd_str;
 	u8 packet_num;
-	s16 buff_pos;
+	int buff_pos;
 	struct batadv_ogm_packet *batadv_ogm_packet;
 	struct sk_buff *skb;
 	u8 *packet_pos;
author maintainer
#137 lore raw

PATCH 137/188 batman-adv: reject new tp_meter sessions during teardown

In reply to Greg Kroah-Hartman

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

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

From: Jiexun Wang <[email protected]>

commit 3243543592425beec83d453793e9d27caa0d8e66 upstream.

Prevent tp_meter from starting new sender or receiver sessions after mesh_state has left BATADV_MESH_ACTIVE.

---
 net/batman-adv/tp_meter.c |   17 ++++++++++++++++-
 1 file changed, 16 insertions(+), 1 deletion(-)
net/batman-adv/tp_meter.c +16 -1
@@ -947,6 +947,13 @@ void batadv_tp_start(struct batadv_priv
 
 	/* look for an already existing test towards this node */
 	spin_lock_bh(&bat_priv->tp_list_lock);
+	if (atomic_read(&bat_priv->mesh_state) != BATADV_MESH_ACTIVE) {
+		spin_unlock_bh(&bat_priv->tp_list_lock);
+		batadv_tp_batctl_error_notify(BATADV_TP_REASON_DST_UNREACHABLE,
+					      dst, bat_priv, session_cookie);
+		return;
+	}
+
 	tp_vars = batadv_tp_list_find(bat_priv, dst);
 	if (tp_vars) {
 		spin_unlock_bh(&bat_priv->tp_list_lock);
@@ -1329,9 +1336,12 @@ static struct batadv_tp_vars *
 batadv_tp_init_recv(struct batadv_priv *bat_priv,
 		    const struct batadv_icmp_tp_packet *icmp)
 {
-	struct batadv_tp_vars *tp_vars;
+	struct batadv_tp_vars *tp_vars = NULL;
 
 	spin_lock_bh(&bat_priv->tp_list_lock);
+	if (atomic_read(&bat_priv->mesh_state) != BATADV_MESH_ACTIVE)
+		goto out_unlock;
+
 	tp_vars = batadv_tp_list_find_session(bat_priv, icmp->orig,
 					      icmp->session);
 	if (tp_vars)
@@ -1464,6 +1474,9 @@ void batadv_tp_meter_recv(struct batadv_
 {
 	struct batadv_icmp_tp_packet *icmp;
 
+	if (atomic_read(&bat_priv->mesh_state) != BATADV_MESH_ACTIVE)
+		goto out;
+
 	icmp = (struct batadv_icmp_tp_packet *)skb->data;
 
 	switch (icmp->subtype) {
@@ -1478,6 +1491,8 @@ void batadv_tp_meter_recv(struct batadv_
 			   "Received unknown TP Metric packet type %u\n",
 			   icmp->subtype);
 	}
+
+out:
 	consume_skb(skb);
 }
 
author maintainer
#138 lore raw

PATCH 138/188 batman-adv: stop tp_meter sessions during mesh teardown

In reply to Greg Kroah-Hartman

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

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

From: Jiexun Wang <[email protected]>

commit 3d3cf6a7314aca4df0a6dde28ce784a2a30d0166 upstream.

TP meter sessions remain linked on bat_priv->tp_list after the netlink request has already finished. When the mesh interface is removed, batadv_mesh_free() currently tears down the mesh without first draining these sessions.

A running sender thread or a late incoming tp_meter packet can then keep processing against a mesh instance which is already shutting down. Synchronize tp_meter with the mesh lifetime by stopping all active sessions from batadv_mesh_free() and waiting for sender threads to exit before teardown continues.

---
 net/batman-adv/main.c     |    1 
 net/batman-adv/tp_meter.c |   94 +++++++++++++++++++++++++++++++++++++---------
 net/batman-adv/tp_meter.h |    1 
 net/batman-adv/types.h    |    4 +
 4 files changed, 82 insertions(+), 18 deletions(-)
net/batman-adv/main.c +1 -0
@@ -249,6 +249,7 @@ void batadv_mesh_free(struct net_device
 	atomic_set(&bat_priv->mesh_state, BATADV_MESH_DEACTIVATING);
 
 	batadv_purge_outstanding_packets(bat_priv, NULL);
+	batadv_tp_stop_all(bat_priv);
 
 	batadv_gw_node_free(bat_priv);
 
net/batman-adv/tp_meter.c +76 -18

145 lines not on this page · raw message

net/batman-adv/tp_meter.h +1 -0
@@ -17,6 +17,7 @@ void batadv_tp_start(struct batadv_priv
 		     u32 test_length, u32 *cookie);
 void batadv_tp_stop(struct batadv_priv *bat_priv, const u8 *dst,
 		    u8 return_value);
+void batadv_tp_stop_all(struct batadv_priv *bat_priv);
 void batadv_tp_meter_recv(struct batadv_priv *bat_priv, struct sk_buff *skb);
 
 #endif /* _NET_BATMAN_ADV_TP_METER_H_ */
net/batman-adv/types.h +4 -0
@@ -14,6 +14,7 @@
 #include <linux/average.h>
 #include <linux/bitops.h>
 #include <linux/compiler.h>
+#include <linux/completion.h>
 #include <linux/if.h>
 #include <linux/if_ether.h>
 #include <linux/kref.h>
@@ -1328,6 +1329,9 @@ struct batadv_tp_vars {
 	/** @finish_work: work item for the finishing procedure */
 	struct delayed_work finish_work;
 
+	/** @finished: completion signaled when a sender thread exits */
+	struct completion finished;
+
 	/** @test_length: test length in milliseconds */
 	u32 test_length;
 
author maintainer
#139 lore raw

PATCH 139/188 batman-adv: stop caching unowned originator pointers in BAT IV

In reply to Greg Kroah-Hartman

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

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

From: Jiexun Wang <[email protected]>

commit f03e8583532941b07761c5429de7d50766fa3110 upstream.

BAT IV keeps the last-hop neighbor address in each neigh_node, but some paths also cache an originator pointer derived from a temporary lookup. That pointer is not owned by the neigh_node and may no longer refer to a live originator entry after purge handling runs.

Stop storing the auxiliary originator pointer in the BAT IV neighbor state. When BAT IV needs the neighbor originator data, resolve it from the stored neighbor address and drop the reference again after use.

Fixes c6c8fea29769 ("net: Add batman-adv meshing protocol")
[sven: avoid bonding logic for outgoing OGM]
---
 net/batman-adv/bat_iv_ogm.c |   83 +++++++++++++++++++++++++++++++-------------
 1 file changed, 59 insertions(+), 24 deletions(-)
net/batman-adv/bat_iv_ogm.c +59 -24
@@ -173,19 +173,12 @@ free_orig_node_hash:
 static struct batadv_neigh_node *
 batadv_iv_ogm_neigh_new(struct batadv_hard_iface *hard_iface,
 			const u8 *neigh_addr,
-			struct batadv_orig_node *orig_node,
-			struct batadv_orig_node *orig_neigh)
+			struct batadv_orig_node *orig_node)
 {
 	struct batadv_neigh_node *neigh_node;
 
 	neigh_node = batadv_neigh_node_get_or_create(orig_node,
 						     hard_iface, neigh_addr);
-	if (!neigh_node)
-		goto out;
-
-	neigh_node->orig_node = orig_neigh;
-
-out:
 	return neigh_node;
 }
 
@@ -907,6 +900,31 @@ static u8 batadv_iv_orig_ifinfo_sum(stru
 }
 
 /**
+ * batadv_iv_ogm_neigh_ifinfo_sum() - Get bcast_own sum for a last-hop neighbor
+ * @bat_priv: the bat priv with all the mesh interface information
+ * @neigh_node: last-hop neighbor of an originator
+ *
+ * Return: Number of replied (rebroadcasted) OGMs for the originator currently
+ * announced by the neighbor. Returns 0 if the neighbor's originator entry is
+ * not available anymore.
+ */
+static u8 batadv_iv_ogm_neigh_ifinfo_sum(struct batadv_priv *bat_priv,
+					 const struct batadv_neigh_node *neigh_node)
+{
+	struct batadv_orig_node *orig_neigh;
+	u8 sum;
+
+	orig_neigh = batadv_orig_hash_find(bat_priv, neigh_node->addr);
+	if (!orig_neigh)
+		return 0;
+
+	sum = batadv_iv_orig_ifinfo_sum(orig_neigh, neigh_node->if_incoming);
+	batadv_orig_node_put(orig_neigh);
+
+	return sum;
+}
+
+/**
  * batadv_iv_ogm_orig_update() - use OGM to update corresponding data in an
  *  originator
  * @bat_priv: the bat priv with all the mesh interface information
@@ -975,17 +993,9 @@ batadv_iv_ogm_orig_update(struct batadv_
 	}
 
 	if (!neigh_node) {
-		struct batadv_orig_node *orig_tmp;
-
-		orig_tmp = batadv_iv_ogm_orig_get(bat_priv, ethhdr->h_source);
-		if (!orig_tmp)
-			goto unlock;
-
 		neigh_node = batadv_iv_ogm_neigh_new(if_incoming,
 						     ethhdr->h_source,
-						     orig_node, orig_tmp);
-
-		batadv_orig_node_put(orig_tmp);
+						     orig_node);
 		if (!neigh_node)
 			goto unlock;
 	} else {
@@ -1037,10 +1047,9 @@ batadv_iv_ogm_orig_update(struct batadv_
 	 */
 	if (router_ifinfo &&
 	    neigh_ifinfo->bat_iv.tq_avg == router_ifinfo->bat_iv.tq_avg) {
-		sum_orig = batadv_iv_orig_ifinfo_sum(router->orig_node,
-						     router->if_incoming);
-		sum_neigh = batadv_iv_orig_ifinfo_sum(neigh_node->orig_node,
-						      neigh_node->if_incoming);
+		sum_orig = batadv_iv_ogm_neigh_ifinfo_sum(bat_priv, router);
+		sum_neigh = batadv_iv_ogm_neigh_ifinfo_sum(bat_priv,
+							   neigh_node);
 		if (sum_orig >= sum_neigh)
 			goto out;
 	}
@@ -1106,7 +1115,6 @@ static bool batadv_iv_ogm_calc_tq(struct
 	if (!neigh_node)
 		neigh_node = batadv_iv_ogm_neigh_new(if_incoming,
 						     orig_neigh_node->orig,
-						     orig_neigh_node,
 						     orig_neigh_node);
 
 	if (!neigh_node)
@@ -1303,6 +1311,32 @@ out:
 }
 
 /**
+ * batadv_orig_to_direct_router() - get direct next hop neighbor to an orig address
+ * @bat_priv: the bat priv with all the mesh interface information
+ * @orig_addr: the originator MAC address to search the best next hop router for
+ * @if_outgoing: the interface where the OGM should be sent to
+ *
+ * Return: A neighbor node which is the best router towards the given originator
+ * address. Bonding candidates are ignored.
+ */
+static struct batadv_neigh_node *
+batadv_orig_to_direct_router(struct batadv_priv *bat_priv, u8 *orig_addr,
+			     struct batadv_hard_iface *if_outgoing)
+{
+	struct batadv_neigh_node *neigh_node;
+	struct batadv_orig_node *orig_node;
+
+	orig_node = batadv_orig_hash_find(bat_priv, orig_addr);
+	if (!orig_node)
+		return NULL;
+
+	neigh_node = batadv_orig_router_get(orig_node, if_outgoing);
+	batadv_orig_node_put(orig_node);
+
+	return neigh_node;
+}
+
+/**
  * batadv_iv_ogm_process_per_outif() - process a batman iv OGM for an outgoing
  *  interface
  * @skb: the skb containing the OGM
@@ -1372,8 +1406,9 @@ batadv_iv_ogm_process_per_outif(const st
 
 	router = batadv_orig_router_get(orig_node, if_outgoing);
 	if (router) {
-		router_router = batadv_orig_router_get(router->orig_node,
-						       if_outgoing);
+		router_router = batadv_orig_to_direct_router(bat_priv,
+							     router->addr,
+							     if_outgoing);
 		router_ifinfo = batadv_neigh_ifinfo_get(router, if_outgoing);
 	}
 
author maintainer
#140 lore raw

PATCH 140/188 batman-adv: bla: prevent use-after-free when deleting claims

In reply to Greg Kroah-Hartman

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

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

From: Sven Eckelmann <[email protected]>

commit 4ae1709a314060a196981b344610d023ea841e57 upstream.

When batadv_bla_del_backbone_claims() removes all claims for a backbone, it does this by dropping the link entry in the hash list. This list entry itself was one of the references which need to be dropped at the same time via batadv_claim_put().

But the batadv_claim_put() must not be done before the last access to the claim object in this function. Otherwise the claim might be freed already by the batadv_claim_release() function before the list entry was dropped.

Fixes 23721387c409 ("batman-adv: add basic bridge loop avoidance code")
---
 net/batman-adv/bridge_loop_avoidance.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/batman-adv/bridge_loop_avoidance.c +1 -1
@@ -319,8 +319,8 @@ batadv_bla_del_backbone_claims(struct ba
 			if (claim->backbone_gw != backbone_gw)
 				continue;
 
-			batadv_claim_put(claim);
 			hlist_del_rcu(&claim->hash_entry);
+			batadv_claim_put(claim);
 		}
 		spin_unlock_bh(list_lock);
 	}
author maintainer
#141 lore raw

PATCH 141/188 batman-adv: bla: only purge non-released claims

In reply to Greg Kroah-Hartman

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

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

From: Sven Eckelmann <[email protected]>

commit cf6b604011591865ae39ac82de8978c1120d17af upstream.

When batadv_bla_purge_claims() goes through the list of claims, it is only traversing the hash list with an rcu_read_lock(). Due to a potential parallel batadv_claim_put(), it can happen that it encounters a claim which was actually in the process of being released+freed by batadv_claim_release(). In this case, backbone_gw is set to NULL before the delayed RCU kfree is started. Calling batadv_bla_claim_get_backbone_gw() is then no longer allowed because it would cause a NULL-ptr derefence.

To avoid this, only claims with a valid reference counter must be purged. All others are already taken care of.

Fixes 23721387c409 ("batman-adv: add basic bridge loop avoidance code")
---
 net/batman-adv/bridge_loop_avoidance.c |    8 ++++++++
 1 file changed, 8 insertions(+)
net/batman-adv/bridge_loop_avoidance.c +8 -0
@@ -1289,6 +1289,13 @@ static void batadv_bla_purge_claims(stru
 
 		rcu_read_lock();
 		hlist_for_each_entry_rcu(claim, head, hash_entry) {
+			/* only purge claims not currently in the process of being released.
+			 * Such claims could otherwise have a NULL-ptr backbone_gw set because
+			 * they already went through batadv_claim_release()
+			 */
+			if (!kref_get_unless_zero(&claim->refcount))
+				continue;
+
 			backbone_gw = batadv_bla_claim_get_backbone_gw(claim);
 			if (now)
 				goto purge_now;
@@ -1314,6 +1321,7 @@ purge_now:
 					      claim->addr, claim->vid);
 skip:
 			batadv_backbone_gw_put(backbone_gw);
+			batadv_claim_put(claim);
 		}
 		rcu_read_unlock();
 	}
author maintainer
#142 lore raw

PATCH 142/188 batman-adv: bla: put backbone reference on failed claim hash insert

In reply to Greg Kroah-Hartman

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

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

From: Sven Eckelmann <[email protected]>

commit ba9d20ee9076dac32c371116bacbe72480eb356c upstream.

When batadv_bla_add_claim() fails to insert a new claim into the hash, it leaked a reference to the backbone_gw for which the claim was intended. Call batadv_backbone_gw_put() on the error path to release the reference and avoid leaking the backbone_gw object.

Fixes 3db0decf1185 ("batman-adv: Fix non-atomic bla_claim::backbone_gw access")
---
 net/batman-adv/bridge_loop_avoidance.c |    1 +
 1 file changed, 1 insertion(+)
net/batman-adv/bridge_loop_avoidance.c +1 -0
@@ -724,6 +724,7 @@ static void batadv_bla_add_claim(struct
 
 		if (unlikely(hash_added != 0)) {
 			/* only local changes happened. */
+			batadv_backbone_gw_put(backbone_gw);
 			kfree(claim);
 			return;
 		}
author maintainer
#143 lore raw

PATCH 143/188 sched_ext: Use HK_TYPE_DOMAIN_BOOT to detect isolcpus= domain isolation

In reply to Greg Kroah-Hartman

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

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

From: Andrea Righi <[email protected]>

commit 6ae315d37924435516d697ea7dde0b799a5928e0 upstream.

scx_enable() refuses to attach a BPF scheduler when isolcpus=domain is in effect by comparing housekeeping_cpumask(HK_TYPE_DOMAIN) against cpu_possible_mask.

Since commit 27c3a5967f05 ("sched/isolation: Convert housekeeping cpumasks to rcu pointers"), HK_TYPE_DOMAIN's cpumask is RCU protected and dereferencing it requires either RCU read lock, the cpu_hotplug write lock, or the cpuset lock; scx_enable() holds none of these, so booting with isolcpus=domain and attaching any BPF scheduler triggers the following lockdep splat:

=============================
WARNING: suspicious RCU usage
-----------------------------
kernel/sched/isolation.c:60 suspicious rcu_dereference_check() usage!

1 lock held by scx_flash/281:
#0: ffffffff8379fce0 (update_mutex){+.+.}-{4:4}, at:

       bpf_struct_ops_link_create+0x134/0x1c0

  Call Trace:
   dump_stack_lvl+0x6f/0xb0
   lockdep_rcu_suspicious.cold+0x37/0x70
   housekeeping_cpumask+0xcd/0xe0
   scx_enable.isra.0+0x17/0x120
   bpf_scx_reg+0x5e/0x80
   bpf_struct_ops_link_create+0x151/0x1c0
   __sys_bpf+0x1e4b/0x33c0
   __x64_sys_bpf+0x21/0x30
   do_syscall_64+0x117/0xf80
   entry_SYSCALL_64_after_hwframe+0x77/0x7f

In addition, commit 03ff73510169 ("cpuset: Update HK_TYPE_DOMAIN cpumask from cpuset") made HK_TYPE_DOMAIN include cpuset isolated partitions as well, which means the current check also rejects BPF schedulers when a cpuset partition is active. That contradicts the original intent of commit 9f391f94a173 ("sched_ext: Disallow loading BPF scheduler if isolcpus= domain isolation is in effect"), which explicitly noted that cpuset partitions are honored through per-task cpumasks and should not be rejected.

Switch to housekeeping_enabled(HK_TYPE_DOMAIN_BOOT), which reads only the housekeeping flag bit (no RCU dereference) and reflects exactly the boot-time isolcpus= configuration that the error message refers to.

Fixes 27c3a5967f05 ("sched/isolation: Convert housekeeping cpumasks to rcu pointers")
---
 kernel/sched/ext.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
kernel/sched/ext.c +1 -2
@@ -4906,8 +4906,7 @@ static int scx_enable(struct sched_ext_o
 	static DEFINE_MUTEX(helper_mutex);
 	struct scx_enable_cmd cmd;
 
-	if (!cpumask_equal(housekeeping_cpumask(HK_TYPE_DOMAIN),
-			   cpu_possible_mask)) {
+	if (housekeeping_enabled(HK_TYPE_DOMAIN_BOOT)) {
 		pr_err("sched_ext: Not compatible with \"isolcpus=\" domain isolation\n");
 		return -EINVAL;
 	}
Tested-by on v1 143/188
#144 lore raw

Re: [PATCH 6.18 143/188] sched_ext: Use HK_TYPE_DOMAIN_BOOT to detect isolcpus= domain isolation

Hi Greg,

Am 15.05.2026 um 17:49 schrieb Greg Kroah-Hartman:

Quoted 68 lines

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

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

From: Andrea Righi <[email protected]>

commit 6ae315d37924435516d697ea7dde0b799a5928e0 upstream.

scx_enable() refuses to attach a BPF scheduler when isolcpus=domain is in effect by comparing housekeeping_cpumask(HK_TYPE_DOMAIN) against cpu_possible_mask.

Since commit 27c3a5967f05 ("sched/isolation: Convert housekeeping cpumasks to rcu pointers"), HK_TYPE_DOMAIN's cpumask is RCU protected and dereferencing it requires either RCU read lock, the cpu_hotplug write lock, or the cpuset lock; scx_enable() holds none of these, so booting with isolcpus=domain and attaching any BPF scheduler triggers the following lockdep splat:

   =============================
   WARNING: suspicious RCU usage
   -----------------------------
   kernel/sched/isolation.c:60 suspicious rcu_dereference_check() usage!

   1 lock held by scx_flash/281:
    #0: ffffffff8379fce0 (update_mutex){+.+.}-{4:4}, at:
        bpf_struct_ops_link_create+0x134/0x1c0

   Call Trace:
    dump_stack_lvl+0x6f/0xb0
    lockdep_rcu_suspicious.cold+0x37/0x70
    housekeeping_cpumask+0xcd/0xe0
    scx_enable.isra.0+0x17/0x120
    bpf_scx_reg+0x5e/0x80
    bpf_struct_ops_link_create+0x151/0x1c0
    __sys_bpf+0x1e4b/0x33c0
    __x64_sys_bpf+0x21/0x30
    do_syscall_64+0x117/0xf80
    entry_SYSCALL_64_after_hwframe+0x77/0x7f

In addition, commit 03ff73510169 ("cpuset: Update HK_TYPE_DOMAIN cpumask from cpuset") made HK_TYPE_DOMAIN include cpuset isolated partitions as well, which means the current check also rejects BPF schedulers when a cpuset partition is active. That contradicts the original intent of commit 9f391f94a173 ("sched_ext: Disallow loading BPF scheduler if isolcpus= domain isolation is in effect"), which explicitly noted that cpuset partitions are honored through per-task cpumasks and should not be rejected.

Switch to housekeeping_enabled(HK_TYPE_DOMAIN_BOOT), which reads only the housekeeping flag bit (no RCU dereference) and reflects exactly the boot-time isolcpus= configuration that the error message refers to.

Fixes 27c3a5967f05 ("sched/isolation: Convert housekeeping cpumasks to rcu pointers")
---
  kernel/sched/ext.c |    3 +--
  1 file changed, 1 insertion(+), 2 deletions(-)

kernel/sched/ext.c static int scx_enable(struct sched_ext_o · line 4912

Quoted 9 lines
kernel/sched/ext.c +1 -2
@@ -4906,8 +4906,7 @@ static int scx_enable(struct sched_ext_o
  	static DEFINE_MUTEX(helper_mutex);
  	struct scx_enable_cmd cmd;
 
-	if (!cpumask_equal(housekeeping_cpumask(HK_TYPE_DOMAIN),
-			   cpu_possible_mask)) {
+	if (housekeeping_enabled(HK_TYPE_DOMAIN_BOOT)) {
  		pr_err("sched_ext: Not compatible with \"isolcpus=\" domain isolation\n");
  		return -EINVAL;
  	}

This patch causes a build failure for me:

CC kernel/sched/build_policy.o
In file included from kernel/sched/build_policy.c:62:
kernel/sched/ext.c: In function ‘scx_enable’:
kernel/sched/ext.c:4924:34: error: ‘HK_TYPE_DOMAIN_BOOT’ undeclared (first use in this function); did you mean ‘HK_TYPE_DOMAIN’?

 4924 |         if (housekeeping_enabled(HK_TYPE_DOMAIN_BOOT)) {
      |                                  ^~~~~~~~~~~~~~~~~~~
      |                                  HK_TYPE_DOMAIN

kernel/sched/ext.c:4924:34: note: each undeclared identifier is reported only once for each function it appears in make[4]: *** [scripts/Makefile.build:287: kernel/sched/build_policy.o] Fehler 1
make[3]: *** [scripts/Makefile.build:544: kernel/sched] Fehler 2
make[2]: *** [scripts/Makefile.build:544: kernel] Fehler 2
make[1]: *** [/usr/src/linux-stable-rc/Makefile:2024: .] Fehler 2
make: *** [Makefile:248: __sub-make] Fehler 2
root@linus:/usr/src/linux-stable-rc#

If I revert this patch, the build succeeds, and the kernel boots and seems to work fine without any observable regressions.

Beste Grüße,
Peter Schneider

Signature · 8 lines
Climb the mountain not to plant your flag, but to embrace the challenge,
enjoy the air and behold the view. Climb it so you can see the world,
not so the world can see you.                    -- David McCullough Jr.

OpenPGP:  0xA3828BD796CCE11A8CADE8866E3A92C92C3FF244
Download: https://www.peters-netzplatz.de/download/pschneider1968_pub.asc
https://keys.mailvelope.com/pks/lookup?op=get&[email protected]
https://keys.mailvelope.com/pks/lookup?op=get&[email protected]
author maintainer
#145 lore raw

Re: [PATCH 6.18 143/188] sched_ext: Use HK_TYPE_DOMAIN_BOOT to detect isolcpus= domain isolation

Quoted 99 lines · Peter Schneider

Hi Greg,

Am 15.05.2026 um 17:49 schrieb Greg Kroah-Hartman:

Quoted 67 lines

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

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

From: Andrea Righi <[email protected]>

commit 6ae315d37924435516d697ea7dde0b799a5928e0 upstream.

scx_enable() refuses to attach a BPF scheduler when isolcpus=domain is in effect by comparing housekeeping_cpumask(HK_TYPE_DOMAIN) against cpu_possible_mask.

Since commit 27c3a5967f05 ("sched/isolation: Convert housekeeping cpumasks to rcu pointers"), HK_TYPE_DOMAIN's cpumask is RCU protected and dereferencing it requires either RCU read lock, the cpu_hotplug write lock, or the cpuset lock; scx_enable() holds none of these, so booting with isolcpus=domain and attaching any BPF scheduler triggers the following lockdep splat:

   =============================
   WARNING: suspicious RCU usage
   -----------------------------
   kernel/sched/isolation.c:60 suspicious rcu_dereference_check() usage!

   1 lock held by scx_flash/281:
    #0: ffffffff8379fce0 (update_mutex){+.+.}-{4:4}, at:
        bpf_struct_ops_link_create+0x134/0x1c0

   Call Trace:
    dump_stack_lvl+0x6f/0xb0
    lockdep_rcu_suspicious.cold+0x37/0x70
    housekeeping_cpumask+0xcd/0xe0
    scx_enable.isra.0+0x17/0x120
    bpf_scx_reg+0x5e/0x80
    bpf_struct_ops_link_create+0x151/0x1c0
    __sys_bpf+0x1e4b/0x33c0
    __x64_sys_bpf+0x21/0x30
    do_syscall_64+0x117/0xf80
    entry_SYSCALL_64_after_hwframe+0x77/0x7f

In addition, commit 03ff73510169 ("cpuset: Update HK_TYPE_DOMAIN cpumask from cpuset") made HK_TYPE_DOMAIN include cpuset isolated partitions as well, which means the current check also rejects BPF schedulers when a cpuset partition is active. That contradicts the original intent of commit 9f391f94a173 ("sched_ext: Disallow loading BPF scheduler if isolcpus= domain isolation is in effect"), which explicitly noted that cpuset partitions are honored through per-task cpumasks and should not be rejected.

Switch to housekeeping_enabled(HK_TYPE_DOMAIN_BOOT), which reads only the housekeeping flag bit (no RCU dereference) and reflects exactly the boot-time isolcpus= configuration that the error message refers to.

Fixes 27c3a5967f05 ("sched/isolation: Convert housekeeping cpumasks to rcu pointers")
---
  kernel/sched/ext.c |    3 +--
  1 file changed, 1 insertion(+), 2 deletions(-)

kernel/sched/ext.c static int scx_enable(struct sched_ext_o · line 4912

Quoted 9 lines
kernel/sched/ext.c +1 -2
@@ -4906,8 +4906,7 @@ static int scx_enable(struct sched_ext_o
  	static DEFINE_MUTEX(helper_mutex);
  	struct scx_enable_cmd cmd;
-	if (!cpumask_equal(housekeeping_cpumask(HK_TYPE_DOMAIN),
-			   cpu_possible_mask)) {
+	if (housekeeping_enabled(HK_TYPE_DOMAIN_BOOT)) {
  		pr_err("sched_ext: Not compatible with \"isolcpus=\" domain isolation\n");
  		return -EINVAL;
  	}
 

This patch causes a build failure for me:

CC kernel/sched/build_policy.o
In file included from kernel/sched/build_policy.c:62:
kernel/sched/ext.c: In function ‘scx_enable’:
kernel/sched/ext.c:4924:34: error: ‘HK_TYPE_DOMAIN_BOOT’ undeclared (first use in this function); did you mean ‘HK_TYPE_DOMAIN’?

 4924 |         if (housekeeping_enabled(HK_TYPE_DOMAIN_BOOT)) {
      |                                  ^~~~~~~~~~~~~~~~~~~
      |                                  HK_TYPE_DOMAIN

kernel/sched/ext.c:4924:34: note: each undeclared identifier is reported only once for each function it appears in make[4]: *** [scripts/Makefile.build:287: kernel/sched/build_policy.o] Fehler 1
make[3]: *** [scripts/Makefile.build:544: kernel/sched] Fehler 2
make[2]: *** [scripts/Makefile.build:544: kernel] Fehler 2
make[1]: *** [/usr/src/linux-stable-rc/Makefile:2024: .] Fehler 2
make: *** [Makefile:248: __sub-make] Fehler 2
root@linus:/usr/src/linux-stable-rc#

Quoted 3 lines

If I revert this patch, the build succeeds, and the kernel boots and seems to work fine without any observable regressions.

Thanks for this, I'll go drop this commit and push out a -rc2

greg k-h

maintainer
#146 lore raw

Re: [PATCH 6.18 143/188] sched_ext: Use HK_TYPE_DOMAIN_BOOT to detect isolcpus= domain isolation

In reply to Peter Schneider
Quoted 4 lines

kernel/sched/ext.c:4924:34: error: 'HK_TYPE_DOMAIN_BOOT' undeclared (first use in this function)

If I revert this patch, the build succeeds, and the kernel boots and seems to work fine without any observable regressions.

Dropped from the 6.18 and 6.12 queues. The prerequisite enum HK_TYPE_DOMAIN_BOOT (added by 4fca0e550d50 "sched/isolation: Save boot defined domain flags") only exists in 7.0+, matching the author's "Cc: [email protected] # v7.0+" tag. The 7.0 queue retains the patch.

Thanks for the report and the Tested-by.

--
Thanks,
Sasha

author maintainer
#147 lore raw

PATCH 144/188 io_uring/zcrx: use guards for locking

In reply to Greg Kroah-Hartman

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

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

From: Pavel Begunkov <[email protected]>

commit 898ad80d1207cbdb22b21bafb6de4adfd7627bd0 upstream.

Convert last several places using manual locking to guards to simplify the code.

---
 io_uring/zcrx.c |   15 +++++++--------
 1 file changed, 7 insertions(+), 8 deletions(-)
io_uring/zcrx.c +7 -8
@@ -696,9 +696,8 @@ static void io_zcrx_return_niov_freelist
 {
 	struct io_zcrx_area *area = io_zcrx_iov_to_area(niov);
 
-	spin_lock_bh(&area->freelist_lock);
+	guard(spinlock_bh)(&area->freelist_lock);
 	area->freelist[area->free_count++] = net_iov_idx(niov);
-	spin_unlock_bh(&area->freelist_lock);
 }
 
 static void io_zcrx_return_niov(struct net_iov *niov)
@@ -829,7 +828,8 @@ static void io_zcrx_refill_slow(struct p
 {
 	struct io_zcrx_area *area = ifq->area;
 
-	spin_lock_bh(&area->freelist_lock);
+	guard(spinlock_bh)(&area->freelist_lock);
+
 	while (area->free_count && pp->alloc.count < PP_ALLOC_CACHE_REFILL) {
 		struct net_iov *niov = __io_zcrx_get_free_niov(area);
 		netmem_ref netmem = net_iov_to_netmem(niov);
@@ -838,7 +838,6 @@ static void io_zcrx_refill_slow(struct p
 		io_zcrx_sync_for_device(pp, niov);
 		net_mp_netmem_place_in_cache(pp, netmem);
 	}
-	spin_unlock_bh(&area->freelist_lock);
 }
 
 static netmem_ref io_pp_zc_alloc_netmems(struct page_pool *pp, gfp_t gfp)
@@ -975,10 +974,10 @@ static struct net_iov *io_alloc_fallback
 	if (area->mem.is_dmabuf)
 		return NULL;
 
-	spin_lock_bh(&area->freelist_lock);
-	if (area->free_count)
-		niov = __io_zcrx_get_free_niov(area);
-	spin_unlock_bh(&area->freelist_lock);
+	scoped_guard(spinlock_bh, &area->freelist_lock) {
+		if (area->free_count)
+			niov = __io_zcrx_get_free_niov(area);
+	}
 
 	if (niov)
 		page_pool_fragment_netmem(net_iov_to_netmem(niov), 1);
author maintainer
#148 lore raw

PATCH 145/188 io_uring/zcrx: warn on freelist violations

In reply to Greg Kroah-Hartman

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

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

From: Pavel Begunkov <[email protected]>

commit 770594e78c3964cf23cf5287f849437cdde9b7d0 upstream.

The freelist is appropriately sized to always be able to take a free niov, but let's be more defensive and check the invariant with a warning. That should help to catch any double-free issues.

---
 io_uring/zcrx.c |    2 ++
 1 file changed, 2 insertions(+)
io_uring/zcrx.c +2 -0
@@ -697,6 +697,8 @@ static void io_zcrx_return_niov_freelist
 	struct io_zcrx_area *area = io_zcrx_iov_to_area(niov);
 
 	guard(spinlock_bh)(&area->freelist_lock);
+	if (WARN_ON_ONCE(area->free_count >= area->nia.num_niovs))
+		return;
 	area->freelist[area->free_count++] = net_iov_idx(niov);
 }
 
author maintainer
#149 lore raw

PATCH 146/188 LoongArch: KVM: Compile switch.S directly into the kernel

In reply to Greg Kroah-Hartman

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

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

From: Xianglai Li <[email protected]>

commit 5203012fa6045aac4b69d4e7c212e16dcf38ef10 upstream.

If we directly compile the switch.S file into the kernel, the address of the kvm_exc_entry function will definitely be within the DMW memory area. Therefore, we will no longer need to perform a copy relocation of the kvm_exc_entry.

So this patch compiles switch.S directly into the kernel, and then remove the copy relocation execution logic for the kvm_exc_entry function.

---
 arch/loongarch/Kbuild                       |    2 -
 arch/loongarch/include/asm/asm-prototypes.h |   20 ++++++++++++++++
 arch/loongarch/include/asm/kvm_host.h       |    3 --
 arch/loongarch/kvm/Makefile                 |    3 +-
 arch/loongarch/kvm/main.c                   |   35 ++--------------------------
 arch/loongarch/kvm/switch.S                 |   20 ++++++++++++----
 6 files changed, 41 insertions(+), 42 deletions(-)
arch/loongarch/Kbuild +1 -1
@@ -3,7 +3,7 @@ obj-y += mm/
 obj-y += net/
 obj-y += vdso/
 
-obj-$(CONFIG_KVM) += kvm/
+obj-$(subst m,y,$(CONFIG_KVM)) += kvm/
 
 # for cleaning
 subdir- += boot
arch/loongarch/include/asm/asm-prototypes.h +20 -0
@@ -20,3 +20,23 @@ asmlinkage void noinstr __no_stack_prote
 								    struct pt_regs *regs,
 								    int (*fn)(void *),
 								    void *fn_arg);
+
+struct kvm_run;
+struct kvm_vcpu;
+struct loongarch_fpu;
+
+void kvm_exc_entry(void);
+int  kvm_enter_guest(struct kvm_run *run, struct kvm_vcpu *vcpu);
+
+void kvm_save_fpu(struct loongarch_fpu *fpu);
+void kvm_restore_fpu(struct loongarch_fpu *fpu);
+
+#ifdef CONFIG_CPU_HAS_LSX
+void kvm_save_lsx(struct loongarch_fpu *fpu);
+void kvm_restore_lsx(struct loongarch_fpu *fpu);
+#endif
+
+#ifdef CONFIG_CPU_HAS_LASX
+void kvm_save_lasx(struct loongarch_fpu *fpu);
+void kvm_restore_lasx(struct loongarch_fpu *fpu);
+#endif
arch/loongarch/include/asm/kvm_host.h +0 -3
@@ -85,7 +85,6 @@ struct kvm_context {
 struct kvm_world_switch {
 	int (*exc_entry)(void);
 	int (*enter_guest)(struct kvm_run *run, struct kvm_vcpu *vcpu);
-	unsigned long page_order;
 };
 
 #define MAX_PGTABLE_LEVELS	4
@@ -339,8 +338,6 @@ void kvm_exc_entry(void);
 int  kvm_enter_guest(struct kvm_run *run, struct kvm_vcpu *vcpu);
 
 extern unsigned long vpid_mask;
-extern const unsigned long kvm_exception_size;
-extern const unsigned long kvm_enter_guest_size;
 extern struct kvm_world_switch *kvm_loongarch_ops;
 
 #define SW_GCSR		(1 << 0)
arch/loongarch/kvm/Makefile +2 -1
@@ -7,11 +7,12 @@ include $(srctree)/virt/kvm/Makefile.kvm
 
 obj-$(CONFIG_KVM) += kvm.o
 
+obj-y += switch.o
+
 kvm-y += exit.o
 kvm-y += interrupt.o
 kvm-y += main.o
 kvm-y += mmu.o
-kvm-y += switch.o
 kvm-y += timer.o
 kvm-y += tlb.o
 kvm-y += vcpu.o
arch/loongarch/kvm/main.c +3 -32
@@ -340,8 +340,7 @@ void kvm_arch_disable_virtualization_cpu
 
 static int kvm_loongarch_env_init(void)
 {
-	int cpu, order, ret;
-	void *addr;
+	int cpu, ret;
 	struct kvm_context *context;
 
 	vmcs = alloc_percpu(struct kvm_context);
@@ -357,30 +356,8 @@ static int kvm_loongarch_env_init(void)
 		return -ENOMEM;
 	}
 
-	/*
-	 * PGD register is shared between root kernel and kvm hypervisor.
-	 * So world switch entry should be in DMW area rather than TLB area
-	 * to avoid page fault reenter.
-	 *
-	 * In future if hardware pagetable walking is supported, we won't
-	 * need to copy world switch code to DMW area.
-	 */
-	order = get_order(kvm_exception_size + kvm_enter_guest_size);
-	addr = (void *)__get_free_pages(GFP_KERNEL, order);
-	if (!addr) {
-		free_percpu(vmcs);
-		vmcs = NULL;
-		kfree(kvm_loongarch_ops);
-		kvm_loongarch_ops = NULL;
-		return -ENOMEM;
-	}
-
-	memcpy(addr, kvm_exc_entry, kvm_exception_size);
-	memcpy(addr + kvm_exception_size, kvm_enter_guest, kvm_enter_guest_size);
-	flush_icache_range((unsigned long)addr, (unsigned long)addr + kvm_exception_size + kvm_enter_guest_size);
-	kvm_loongarch_ops->exc_entry = addr;
-	kvm_loongarch_ops->enter_guest = addr + kvm_exception_size;
-	kvm_loongarch_ops->page_order = order;
+	kvm_loongarch_ops->exc_entry = (void *)kvm_exc_entry;
+	kvm_loongarch_ops->enter_guest = (void *)kvm_enter_guest;
 
 	vpid_mask = read_csr_gstat();
 	vpid_mask = (vpid_mask & CSR_GSTAT_GIDBIT) >> CSR_GSTAT_GIDBIT_SHIFT;
@@ -414,16 +391,10 @@ static int kvm_loongarch_env_init(void)
 
 static void kvm_loongarch_env_exit(void)
 {
-	unsigned long addr;
-
 	if (vmcs)
 		free_percpu(vmcs);
 
 	if (kvm_loongarch_ops) {
-		if (kvm_loongarch_ops->exc_entry) {
-			addr = (unsigned long)kvm_loongarch_ops->exc_entry;
-			free_pages(addr, kvm_loongarch_ops->page_order);
-		}
 		kfree(kvm_loongarch_ops);
 	}
 
arch/loongarch/kvm/switch.S +15 -5
@@ -4,9 +4,11 @@
  */
 
 #include <linux/linkage.h>
+#include <linux/kvm_types.h>
 #include <asm/asm.h>
 #include <asm/asmmacro.h>
 #include <asm/loongarch.h>
+#include <asm/page.h>
 #include <asm/regdef.h>
 #include <asm/unwind_hints.h>
 
@@ -100,8 +102,13 @@
 	 *  -        is still in guest mode, such as pgd table/vmid registers etc,
 	 *  -        will fix with hw page walk enabled in future
 	 * load kvm_vcpu from reserved CSR KVM_VCPU_KS, and save a2 to KVM_TEMP_KS
+	 *
+	 * PGD register is shared between root kernel and kvm hypervisor.
+	 * So world switch entry should be in DMW area rather than TLB area
+	 * to avoid page fault re-enter.
 	 */
 	.text
+	.p2align PAGE_SHIFT
 	.cfi_sections	.debug_frame
 SYM_CODE_START(kvm_exc_entry)
 	UNWIND_HINT_END_OF_STACK
@@ -190,8 +197,8 @@ ret_to_host:
 	kvm_restore_host_gpr    a2
 	jr      ra
 
-SYM_INNER_LABEL(kvm_exc_entry_end, SYM_L_LOCAL)
 SYM_CODE_END(kvm_exc_entry)
+EXPORT_SYMBOL_GPL(kvm_exc_entry)
 
 /*
  * int kvm_enter_guest(struct kvm_run *run, struct kvm_vcpu *vcpu)
@@ -215,8 +222,8 @@ SYM_FUNC_START(kvm_enter_guest)
 	/* Save kvm_vcpu to kscratch */
 	csrwr	a1, KVM_VCPU_KS
 	kvm_switch_to_guest
-SYM_INNER_LABEL(kvm_enter_guest_end, SYM_L_LOCAL)
 SYM_FUNC_END(kvm_enter_guest)
+EXPORT_SYMBOL_GPL(kvm_enter_guest)
 
 SYM_FUNC_START(kvm_save_fpu)
 	fpu_save_csr	a0 t1
@@ -224,6 +231,7 @@ SYM_FUNC_START(kvm_save_fpu)
 	fpu_save_cc	a0 t1 t2
 	jr              ra
 SYM_FUNC_END(kvm_save_fpu)
+EXPORT_SYMBOL_GPL(kvm_save_fpu)
 
 SYM_FUNC_START(kvm_restore_fpu)
 	fpu_restore_double a0 t1
@@ -231,6 +239,7 @@ SYM_FUNC_START(kvm_restore_fpu)
 	fpu_restore_cc	   a0 t1 t2
 	jr                 ra
 SYM_FUNC_END(kvm_restore_fpu)
+EXPORT_SYMBOL_GPL(kvm_restore_fpu)
 
 #ifdef CONFIG_CPU_HAS_LSX
 SYM_FUNC_START(kvm_save_lsx)
@@ -239,6 +248,7 @@ SYM_FUNC_START(kvm_save_lsx)
 	lsx_save_data   a0 t1
 	jr              ra
 SYM_FUNC_END(kvm_save_lsx)
+EXPORT_SYMBOL_GPL(kvm_save_lsx)
 
 SYM_FUNC_START(kvm_restore_lsx)
 	lsx_restore_data a0 t1
@@ -246,6 +256,7 @@ SYM_FUNC_START(kvm_restore_lsx)
 	fpu_restore_csr  a0 t1 t2
 	jr               ra
 SYM_FUNC_END(kvm_restore_lsx)
+EXPORT_SYMBOL_GPL(kvm_restore_lsx)
 #endif
 
 #ifdef CONFIG_CPU_HAS_LASX
@@ -255,6 +266,7 @@ SYM_FUNC_START(kvm_save_lasx)
 	lasx_save_data  a0 t1
 	jr              ra
 SYM_FUNC_END(kvm_save_lasx)
+EXPORT_SYMBOL_GPL(kvm_save_lasx)
 
 SYM_FUNC_START(kvm_restore_lasx)
 	lasx_restore_data a0 t1
@@ -262,10 +274,8 @@ SYM_FUNC_START(kvm_restore_lasx)
 	fpu_restore_csr   a0 t1 t2
 	jr                ra
 SYM_FUNC_END(kvm_restore_lasx)
+EXPORT_SYMBOL_GPL(kvm_restore_lasx)
 #endif
-	.section ".rodata"
-SYM_DATA(kvm_exception_size, .quad kvm_exc_entry_end - kvm_exc_entry)
-SYM_DATA(kvm_enter_guest_size, .quad kvm_enter_guest_end - kvm_enter_guest)
 
 #ifdef CONFIG_CPU_HAS_LBT
 STACK_FRAME_NON_STANDARD kvm_restore_fpu
author maintainer
#150 lore raw

PATCH 147/188 mm/damon/core: implement damon_kdamond_pid()

In reply to Greg Kroah-Hartman

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

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

From: SeongJae Park <[email protected]>

commit 4262c53236977de3ceaa3bf2aefdf772c9b874dd upstream.

Patch series "mm/damon: hide kdamond and kdamond_lock from API callers".

'kdamond' and 'kdamond_lock' fields initially exposed to DAMON API callers for flexible synchronization and use cases. As DAMON API became somewhat complicated compared to the early days, Keeping those exposed could only encourage the API callers to invent more creative but complicated and difficult-to-debug use cases.

Fortunately DAMON API callers didn't invent that many creative use cases. There exist only two use cases of 'kdamond' and 'kdamond_lock'. Finding whether the kdamond is actively running, and getting the pid of the kdamond. For the first use case, a dedicated API function, namely 'damon_is_running()' is provided, and all DAMON API callers are using the function for the use case. Hence only the second use case is where the fields are directly being used by DAMON API callers.

To prevent future invention of complicated and erroneous use cases of the fields, hide the fields from the API callers. For that, provide new dedicated DAMON API functions for the remaining use case, namely damon_kdamond_pid(), migrate DAMON API callers to use the new function, and mark the fields as private fields.

This patch (of 5):

'kdamond' and 'kdamond_lock' are directly being used by DAMON API callers for getting the pid of the corresponding kdamond. To discourage invention of creative but complicated and erroneous new usages of the fields that require careful synchronization, implement a new API function that can simply be used without the manual synchronizations.

---
 include/linux/damon.h |    1 +
 mm/damon/core.c       |   17 +++++++++++++++++
 2 files changed, 18 insertions(+)
include/linux/damon.h +1 -0
@@ -949,6 +949,7 @@ bool damon_initialized(void);
 int damon_start(struct damon_ctx **ctxs, int nr_ctxs, bool exclusive);
 int damon_stop(struct damon_ctx **ctxs, int nr_ctxs);
 bool damon_is_running(struct damon_ctx *ctx);
+int damon_kdamond_pid(struct damon_ctx *ctx);
 
 int damon_call(struct damon_ctx *ctx, struct damon_call_control *control);
 int damos_walk(struct damon_ctx *ctx, struct damos_walk_control *control);
mm/damon/core.c +17 -0
@@ -1437,6 +1437,23 @@ bool damon_is_running(struct damon_ctx *
 }
 
 /**
+ * damon_kdamond_pid() - Return pid of a given DAMON context's worker thread.
+ * @ctx:	The DAMON context of the question.
+ *
+ * Return: pid if @ctx is running, negative error code otherwise.
+ */
+int damon_kdamond_pid(struct damon_ctx *ctx)
+{
+	int pid = -EINVAL;
+
+	mutex_lock(&ctx->kdamond_lock);
+	if (ctx->kdamond)
+		pid = ctx->kdamond->pid;
+	mutex_unlock(&ctx->kdamond_lock);
+	return pid;
+}
+
+/**
  * damon_call() - Invoke a given function on DAMON worker thread (kdamond).
  * @ctx:	DAMON context to call the function for.
  * @control:	Control variable of the call request.
author maintainer
#151 lore raw

PATCH 148/188 mm/damon/lru_sort: detect and use fresh enabled and kdamond_pid values

In reply to Greg Kroah-Hartman

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

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

From: SeongJae Park <[email protected]>

commit b98b7ff6025ae82570d4915e083f0cbd8d48b3cf upstream.

DAMON_LRU_SORT updates 'enabled' and 'kdamond_pid' parameter values, which represents the running status of its kdamond, when the user explicitly requests start/stop of the kdamond. The kdamond can, however, be stopped in events other than the explicit user request in the following three events.

1. ctx->regions_score_histogram allocation failure at beginning of the execution,
2. damon_commit_ctx() failure due to invalid user input, and
3. damon_commit_ctx() failure due to its internal allocation failures.

Hence, if the kdamond is stopped by the above three events, the values of the status parameters can be stale. Users could show the stale values and be confused. This is already bad, but the real consequence is worse. DAMON_LRU_SORT avoids unnecessary damon_start() and damon_stop() calls based on the 'enabled' parameter value. And the update of 'enabled' parameter value depends on the damon_start() and damon_stop() call results. Hence, once the kdamond has stopped by the unintentional events, the user cannot restart the kdamond before the system reboot. For example, the issue can be reproduced via below steps.

    # cd /sys/module/damon_lru_sort/parameters
    #
    # # start DAMON_LRU_SORT
    # echo Y > enabled
    # ps -ef | grep kdamond
    root         806       2  0 17:53 ?        00:00:00 [kdamond.0]
    root         808     803  0 17:53 pts/4    00:00:00 grep kdamond
    #
    # # commit wrong input to stop kdamond withou explicit stop request
    # echo 3 > addr_unit
    # echo Y > commit_inputs
    bash: echo: write error: Invalid argument
    #
    # # confirm kdamond is stopped
    # ps -ef | grep kdamond
    root         811     803  0 17:53 pts/4    00:00:00 grep kdamond
    #
    # # users casn now show stable status
    # cat enabled
    Y
    # cat kdamond_pid
    806
    #
    # # even after fixing the wrong parameter,
    # # kdamond cannot be restarted.
    # echo 1 > addr_unit
    # echo Y > enabled
    # ps -ef | grep kdamond
    root         815     803  0 17:54 pts/4    00:00:00 grep kdamond

The problem will only rarely happen in real and common setups for the following reasons. The allocation failures are unlikely in such setups since those allocations are arguably too small to fail. Also sane users on real production environments may not commit wrong input parameters. But once it happens, the consequence is quite bad. And the bug is a bug.

The issue stems from the fact that there are multiple events that can change the status, and following all the events is challenging. Dynamically detect and use the fresh status for the parameters when those are requested.

Fixes 40e983cca927 ("mm/damon: introduce DAMON-based LRU-lists Sorting")
Cc <[email protected]> # 6.0.x
---
 mm/damon/lru_sort.c |   83 ++++++++++++++++++++++++++++++++++------------------
 1 file changed, 55 insertions(+), 28 deletions(-)
mm/damon/lru_sort.c +55 -28
@@ -118,15 +118,6 @@ module_param(monitor_region_end, ulong,
  */
 static unsigned long addr_unit __read_mostly = 1;
 
-/*
- * PID of the DAMON thread
- *
- * If DAMON_LRU_SORT is enabled, this becomes the PID of the worker thread.
- * Else, -1.
- */
-static int kdamond_pid __read_mostly = -1;
-module_param(kdamond_pid, int, 0400);
-
 static struct damos_stat damon_lru_sort_hot_stat;
 DEFINE_DAMON_MODULES_DAMOS_STATS_PARAMS(damon_lru_sort_hot_stat,
 		lru_sort_tried_hot_regions, lru_sorted_hot_regions,
@@ -288,12 +279,8 @@ static int damon_lru_sort_turn(bool on)
 {
 	int err;
 
-	if (!on) {
-		err = damon_stop(&ctx, 1);
-		if (!err)
-			kdamond_pid = -1;
-		return err;
-	}
+	if (!on)
+		return damon_stop(&ctx, 1);
 
 	err = damon_lru_sort_apply_parameters();
 	if (err)
@@ -302,7 +289,6 @@ static int damon_lru_sort_turn(bool on)
 	err = damon_start(&ctx, 1, true);
 	if (err)
 		return err;
-	kdamond_pid = ctx->kdamond->pid;
 	return damon_call(ctx, &call_control);
 }
 
@@ -330,42 +316,83 @@ module_param_cb(addr_unit, &addr_unit_pa
 MODULE_PARM_DESC(addr_unit,
 	"Scale factor for DAMON_LRU_SORT to ops address conversion (default: 1)");
 
+static bool damon_lru_sort_enabled(void)
+{
+	if (!ctx)
+		return false;
+	return damon_is_running(ctx);
+}
+
 static int damon_lru_sort_enabled_store(const char *val,
 		const struct kernel_param *kp)
 {
-	bool is_enabled = enabled;
-	bool enable;
 	int err;
 
-	err = kstrtobool(val, &enable);
+	err = kstrtobool(val, &enabled);
 	if (err)
 		return err;
 
-	if (is_enabled == enable)
+	if (damon_lru_sort_enabled() == enabled)
 		return 0;
 
 	/* Called before init function.  The function will handle this. */
 	if (!damon_initialized())
-		goto set_param_out;
+		return 0;
 
-	err = damon_lru_sort_turn(enable);
-	if (err)
-		return err;
+	return damon_lru_sort_turn(enabled);
+}
 
-set_param_out:
-	enabled = enable;
-	return err;
+static int damon_lru_sort_enabled_load(char *buffer,
+		const struct kernel_param *kp)
+{
+	return sprintf(buffer, "%c\n", damon_lru_sort_enabled() ? 'Y' : 'N');
 }
 
 static const struct kernel_param_ops enabled_param_ops = {
 	.set = damon_lru_sort_enabled_store,
-	.get = param_get_bool,
+	.get = damon_lru_sort_enabled_load,
 };
 
 module_param_cb(enabled, &enabled_param_ops, &enabled, 0600);
 MODULE_PARM_DESC(enabled,
 	"Enable or disable DAMON_LRU_SORT (default: disabled)");
 
+static int damon_lru_sort_kdamond_pid_store(const char *val,
+		const struct kernel_param *kp)
+{
+	/*
+	 * kdamond_pid is read-only, but kernel command line could write it.
+	 * Do nothing here.
+	 */
+	return 0;
+}
+
+static int damon_lru_sort_kdamond_pid_load(char *buffer,
+		const struct kernel_param *kp)
+{
+	int kdamond_pid = -1;
+
+	if (ctx) {
+		kdamond_pid = damon_kdamond_pid(ctx);
+		if (kdamond_pid < 0)
+			kdamond_pid = -1;
+	}
+	return sprintf(buffer, "%d\n", kdamond_pid);
+}
+
+static const struct kernel_param_ops kdamond_pid_param_ops = {
+	.set = damon_lru_sort_kdamond_pid_store,
+	.get = damon_lru_sort_kdamond_pid_load,
+};
+
+/*
+ * PID of the DAMON thread
+ *
+ * If DAMON_LRU_SORT is enabled, this becomes the PID of the worker thread.
+ * Else, -1.
+ */
+module_param_cb(kdamond_pid, &kdamond_pid_param_ops, NULL, 0400);
+
 static int __init damon_lru_sort_init(void)
 {
 	int err;
author maintainer
#152 lore raw

PATCH 149/188 mm/damon/reclaim: detect and use fresh enabled and kdamond_pid values

In reply to Greg Kroah-Hartman

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

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

From: SeongJae Park <[email protected]>

commit 64a140afa5ed1c6f5ba6d451512cbdbbab1ba339 upstream.

Patch series "mm/damon/modules: detect and use fresh status", v3.

DAMON modules including DAMON_RECLAIM, DAMON_LRU_SORT and DAMON_STAT commonly expose the kdamond running status via their parameters. Under certain scenarios including wrong user inputs and memory allocation failures, those parameter values can be stale. It can confuse users. For DAMON_RECLAIM and DAMON_LRU_SORT, it even makes the kdamond unable to be restarted before the system reboot.

The problem comes from the fact that there are multiple events for the status changes and it is difficult to follow up all the scenarios. Fix the issue by detecting and using the status on demand, instead of using a cached status that is difficult to be updated.

Patches 1-3 fix the bugs in DAMON_RECLAIM, DAMON_LRU_SORT and DAMON_STAT in the order.

This patch (of 3):

DAMON_RECLAIM updates 'enabled' and 'kdamond_pid' parameter values, which represents the running status of its kdamond, when the user explicitly requests start/stop of the kdamond. The kdamond can, however, be stopped in events other than the explicit user request in the following three events.

1. ctx->regions_score_histogram allocation failure at beginning of the execution,
2. damon_commit_ctx() failure due to invalid user input, and
3. damon_commit_ctx() failure due to its internal allocation failures.

Hence, if the kdamond is stopped by the above three events, the values of the status parameters can be stale. Users could show the stale values and be confused. This is already bad, but the real consequence is worse. DAMON_RECLAIM avoids unnecessary damon_start() and damon_stop() calls based on the 'enabled' parameter value. And the update of 'enabled' parameter value depends on the damon_start() and damon_stop() call results. Hence, once the kdamond has stopped by the unintentional events, the user cannot restart the kdamond before the system reboot. For example, the issue can be reproduced via below steps.

    # cd /sys/module/damon_reclaim/parameters
    #
    # # start DAMON_RECLAIM
    # echo Y > enabled
    # ps -ef | grep kdamond
    root         806       2  0 17:53 ?        00:00:00 [kdamond.0]
    root         808     803  0 17:53 pts/4    00:00:00 grep kdamond
    #
    # # commit wrong input to stop kdamond withou explicit stop request
    # echo 3 > addr_unit
    # echo Y > commit_inputs
    bash: echo: write error: Invalid argument
    #
    # # confirm kdamond is stopped
    # ps -ef | grep kdamond
    root         811     803  0 17:53 pts/4    00:00:00 grep kdamond
    #
    # # users casn now show stable status
    # cat enabled
    Y
    # cat kdamond_pid
    806
    #
    # # even after fixing the wrong parameter,
    # # kdamond cannot be restarted.
    # echo 1 > addr_unit
    # echo Y > enabled
    # ps -ef | grep kdamond
    root         815     803  0 17:54 pts/4    00:00:00 grep kdamond

The problem will only rarely happen in real and common setups for the following reasons. The allocation failures are unlikely in such setups since those allocations are arguably too small to fail. Also sane users on real production environments may not commit wrong input parameters. But once it happens, the consequence is quite bad. And the bug is a bug.

The issue stems from the fact that there are multiple events that can change the status, and following all the events is challenging. Dynamically detect and use the fresh status for the parameters when those are requested.

---
 mm/damon/reclaim.c |   83 +++++++++++++++++++++++++++++++++++------------------
 1 file changed, 55 insertions(+), 28 deletions(-)
mm/damon/reclaim.c +55 -28
@@ -144,15 +144,6 @@ static unsigned long addr_unit __read_mo
 static bool skip_anon __read_mostly;
 module_param(skip_anon, bool, 0600);
 
-/*
- * PID of the DAMON thread
- *
- * If DAMON_RECLAIM is enabled, this becomes the PID of the worker thread.
- * Else, -1.
- */
-static int kdamond_pid __read_mostly = -1;
-module_param(kdamond_pid, int, 0400);
-
 static struct damos_stat damon_reclaim_stat;
 DEFINE_DAMON_MODULES_DAMOS_STATS_PARAMS(damon_reclaim_stat,
 		reclaim_tried_regions, reclaimed_regions, quota_exceeds);
@@ -292,12 +283,8 @@ static int damon_reclaim_turn(bool on)
 {
 	int err;
 
-	if (!on) {
-		err = damon_stop(&ctx, 1);
-		if (!err)
-			kdamond_pid = -1;
-		return err;
-	}
+	if (!on)
+		return damon_stop(&ctx, 1);
 
 	err = damon_reclaim_apply_parameters();
 	if (err)
@@ -306,7 +293,6 @@ static int damon_reclaim_turn(bool on)
 	err = damon_start(&ctx, 1, true);
 	if (err)
 		return err;
-	kdamond_pid = ctx->kdamond->pid;
 	return damon_call(ctx, &call_control);
 }
 
@@ -334,42 +320,83 @@ module_param_cb(addr_unit, &addr_unit_pa
 MODULE_PARM_DESC(addr_unit,
 	"Scale factor for DAMON_RECLAIM to ops address conversion (default: 1)");
 
+static bool damon_reclaim_enabled(void)
+{
+	if (!ctx)
+		return false;
+	return damon_is_running(ctx);
+}
+
 static int damon_reclaim_enabled_store(const char *val,
 		const struct kernel_param *kp)
 {
-	bool is_enabled = enabled;
-	bool enable;
 	int err;
 
-	err = kstrtobool(val, &enable);
+	err = kstrtobool(val, &enabled);
 	if (err)
 		return err;
 
-	if (is_enabled == enable)
+	if (damon_reclaim_enabled() == enabled)
 		return 0;
 
 	/* Called before init function.  The function will handle this. */
 	if (!damon_initialized())
-		goto set_param_out;
+		return 0;
 
-	err = damon_reclaim_turn(enable);
-	if (err)
-		return err;
+	return damon_reclaim_turn(enabled);
+}
 
-set_param_out:
-	enabled = enable;
-	return err;
+static int damon_reclaim_enabled_load(char *buffer,
+		const struct kernel_param *kp)
+{
+	return sprintf(buffer, "%c\n", damon_reclaim_enabled() ? 'Y' : 'N');
 }
 
 static const struct kernel_param_ops enabled_param_ops = {
 	.set = damon_reclaim_enabled_store,
-	.get = param_get_bool,
+	.get = damon_reclaim_enabled_load,
 };
 
 module_param_cb(enabled, &enabled_param_ops, &enabled, 0600);
 MODULE_PARM_DESC(enabled,
 	"Enable or disable DAMON_RECLAIM (default: disabled)");
 
+static int damon_reclaim_kdamond_pid_store(const char *val,
+		const struct kernel_param *kp)
+{
+	/*
+	 * kdamond_pid is read-only, but kernel command line could write it.
+	 * Do nothing here.
+	 */
+	return 0;
+}
+
+static int damon_reclaim_kdamond_pid_load(char *buffer,
+		const struct kernel_param *kp)
+{
+	int kdamond_pid = -1;
+
+	if (ctx) {
+		kdamond_pid = damon_kdamond_pid(ctx);
+		if (kdamond_pid < 0)
+			kdamond_pid = -1;
+	}
+	return sprintf(buffer, "%d\n", kdamond_pid);
+}
+
+static const struct kernel_param_ops kdamond_pid_param_ops = {
+	.set = damon_reclaim_kdamond_pid_store,
+	.get = damon_reclaim_kdamond_pid_load,
+};
+
+/*
+ * PID of the DAMON thread
+ *
+ * If DAMON_RECLAIM is enabled, this becomes the PID of the worker thread.
+ * Else, -1.
+ */
+module_param_cb(kdamond_pid, &kdamond_pid_param_ops, NULL, 0400);
+
 static int __init damon_reclaim_init(void)
 {
 	int err;
author maintainer
#153 lore raw

PATCH 150/188 usb: typec: tcpm: reset internal port states on soft reset AMS

In reply to Greg Kroah-Hartman

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

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

From: Amit Sunil Dhamne <[email protected]>

commit 2909f0d4994fb4306bf116df5ccee797791fce2c upstream.

Reset internal port states (such as vdm_sm_running and explicit_contract) on soft reset AMS as the port needs to negotiate a new contract. The consequence of leaving the states in as-is cond are as follows:
* port is in SRC power role and an explicit contract is negotiated with the port partner (in sink role)
* port partner sends a Soft Reset AMS while VDM State Machine is running
* port accepts the Soft Reset request and the port advertises src caps
* port partner sends a Request message but since the explicit_contract and vdm_sm_running are true from previous negotiation, the port ends up sending Soft Reset instead of Accept msg.

Stub Log:

[  203.653942] AMS DISCOVER_IDENTITY start
[  203.653947] PD TX, header: 0x176f
[  203.655901] PD TX complete, status: 0
[  203.657470] PD RX, header: 0x124f [1]
[  203.657477] Rx VDM cmd 0xff008081 type 2 cmd 1 len 1
[  203.657482] AMS DISCOVER_IDENTITY finished
[  203.657484] cc:=4
[  204.155698] PD RX, header: 0x144f [1]
[  204.155718] Rx VDM cmd 0xeeee8001 type 0 cmd 1 len 1
[  204.155741] PD TX, header: 0x196f
[  204.157622] PD TX complete, status: 0
[  204.160060] PD RX, header: 0x4d [1]
[  204.160066] state change SRC_READY -> SOFT_RESET [rev2 SOFT_RESET_AMS]
[  204.160076] PD TX, header: 0x163
[  204.162486] PD TX complete, status: 0
[  204.162832] AMS SOFT_RESET_AMS finished
[  204.162840] cc:=4
[  204.162891] AMS POWER_NEGOTIATION start
[  204.162896] state change SOFT_RESET -> AMS_START [rev2 POWER_NEGOTIATION]
[  204.162908] state change AMS_START -> SRC_SEND_CAPABILITIES [rev2 POWER_NEGOTIATION]
[  204.162913] PD TX, header: 0x1361
[  204.165529] PD TX complete, status: 0
[  204.165571] pending state change SRC_SEND_CAPABILITIES -> SRC_SEND_CAPABILITIES_TIMEOUT @ 60 ms [rev2 POWER_NEGOTIATION]
[  204.166996] PD RX, header: 0x1242 [1]
[  204.167009] state change SRC_SEND_CAPABILITIES -> SRC_SOFT_RESET_WAIT_SNK_TX [rev2 POWER_NEGOTIATION]
[  204.167019] AMS POWER_NEGOTIATION finished
[  204.167020] cc:=4
[  204.167083] AMS SOFT_RESET_AMS start
[  204.167086] state change SRC_SOFT_RESET_WAIT_SNK_TX -> SOFT_RESET_SEND [rev2 SOFT_RESET_AMS]
[  204.167092] PD TX, header: 0x16d
[  204.168824] PD TX complete, status: 0
[  204.168854] pending state change SOFT_RESET_SEND -> HARD_RESET_SEND @ 60 ms [rev2 SOFT_RESET_AMS]
[  204.171876] PD RX, header: 0x43 [1]
[  204.171879] AMS SOFT_RESET_AMS finished

This causes COMMON.PROC.PD.11.2 check failure for TEST.PD.VDM.SRC.2_Rev2Src test on the PD compliance tester.

Fixes 8d3a0578ad1a ("usb: typec: tcpm: Respond Wait if VDM state machine is running")
Fixes f0690a25a140 ("staging: typec: USB Type-C Port Manager (tcpm)")
Cc stable <[email protected]>
---
 drivers/usb/typec/tcpm/tcpm.c |    2 ++
 1 file changed, 2 insertions(+)
drivers/usb/typec/tcpm/tcpm.c +2 -0
@@ -5526,6 +5526,8 @@ static void run_state_machine(struct tcp
 		usb_power_delivery_unregister_capabilities(port->partner_source_caps);
 		port->partner_source_caps = NULL;
 		tcpm_pd_send_control(port, PD_CTRL_ACCEPT, TCPC_TX_SOP);
+		port->vdm_sm_running = false;
+		port->explicit_contract = false;
 		tcpm_ams_finish(port);
 		if (port->pwr_role == TYPEC_SOURCE) {
 			port->upcoming_state = SRC_SEND_CAPABILITIES;
author maintainer
#154 lore raw

PATCH 151/188 mm/damon/core: disallow time-quota setting zero esz

In reply to Greg Kroah-Hartman

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

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

From: SeongJae Park <[email protected]>

commit 8bbde987c2b84f80da0853f739f0a920386f8b99 upstream.

When the throughput of a DAMOS scheme is very slow, DAMOS time quota can make the effective size quota smaller than damon_ctx->min_region_sz. In the case, damos_apply_scheme() will skip applying the action, because the action is tried at region level, which requires >=min_region_sz size. That is, the quota is effectively exceeded for the quota charge window.

Because no action will be applied, the total_charged_sz and total_charged_ns are also not updated. damos_set_effective_quota() will try to update the effective size quota before starting the next charge window. However, because the total_charged_sz and total_charged_ns have not updated, the throughput and effective size quota are also not changed. Since effective size quota can only be decreased, other effective size quota update factors including DAMOS quota goals and size quota cannot make any change, either.

As a result, the scheme is unexpectedly deactivated until the user notices and mitigates the situation. The users can mitigate this situation by changing the time quota online or re-install the scheme. While the mitigation is somewhat straightforward, finding the situation would be challenging, because DAMON is not providing good observabilities for that. Even if such observability is provided, doing the additional monitoring and the mitigation is somewhat cumbersome and not aligned to the intention of the time quota. The time quota was intended to help reduce the user's administration overhead.

Fix the problem by setting time quota-modified effective size quota be at least min_region_sz always.

The issue was discovered [1] by sashiko.

---
 mm/damon/core.c |    8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)
mm/damon/core.c +5 -3
@@ -2135,7 +2135,8 @@ static unsigned long damos_quota_score(s
 /*
  * Called only if quota->ms, or quota->sz are set, or quota->goals is not empty
  */
-static void damos_set_effective_quota(struct damos_quota *quota)
+static void damos_set_effective_quota(struct damos_quota *quota,
+		struct damon_ctx *ctx)
 {
 	unsigned long throughput;
 	unsigned long esz = ULONG_MAX;
@@ -2161,6 +2162,7 @@ static void damos_set_effective_quota(st
 		else
 			throughput = PAGE_SIZE * 1024;
 		esz = min(throughput * quota->ms, esz);
+		esz = max(ctx->min_sz_region, esz);
 	}
 
 	if (quota->sz && quota->sz < esz)
@@ -2197,7 +2199,7 @@ static void damos_adjust_quota(struct da
 	/* First charge window */
 	if (!quota->total_charged_sz && !quota->charged_from) {
 		quota->charged_from = jiffies;
-		damos_set_effective_quota(quota);
+		damos_set_effective_quota(quota, c);
 	}
 
 	/* New charge window starts */
@@ -2211,7 +2213,7 @@ static void damos_adjust_quota(struct da
 		quota->charged_sz = 0;
 		if (trace_damos_esz_enabled())
 			cached_esz = quota->esz;
-		damos_set_effective_quota(quota);
+		damos_set_effective_quota(quota, c);
 		if (trace_damos_esz_enabled() && quota->esz != cached_esz)
 			damos_trace_esz(c, s, quota);
 	}
author maintainer
#155 lore raw

PATCH 152/188 drm/amdgpu: validate the flush_gpu_tlb_pasid()

In reply to Greg Kroah-Hartman

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

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

From: Prike Liang <[email protected]>

commit f4db9913e4d3dabe9ff3ea6178f2c1bc286012b8 upstream.

Validate flush_gpu_tlb_pasid() availability before flushing tlb.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c |    4 ++++
 1 file changed, 4 insertions(+)
drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c +4 -0
@@ -737,6 +737,10 @@ int amdgpu_gmc_flush_gpu_tlb_pasid(struc
 		return 0;
 
 	if (!adev->gmc.flush_pasid_uses_kiq || !ring->sched.ready) {
+
+		if (!adev->gmc.gmc_funcs->flush_gpu_tlb_pasid)
+			return 0;
+
 		if (adev->gmc.flush_tlb_needs_extra_type_2)
 			adev->gmc.gmc_funcs->flush_gpu_tlb_pasid(adev, pasid,
 								 2, all_hub,
author maintainer
#156 lore raw

PATCH 153/188 drm/amdgpu: Fix validating flush_gpu_tlb_pasid()

In reply to Greg Kroah-Hartman

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

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

From: Timur Kristóf <[email protected]>

commit e3a6eff92bbd960b471966d9afccb4d584546d17 upstream.

When a function holds a lock and we return without unlocking it, it deadlocks the kernel. We should always unlock before returning.

This commit fixes suspend/resume on SI. Tested on two Tahiti GPUs: FirePro W9000 and R9 280X.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c +4 -2
@@ -738,8 +738,10 @@ int amdgpu_gmc_flush_gpu_tlb_pasid(struc
 
 	if (!adev->gmc.flush_pasid_uses_kiq || !ring->sched.ready) {
 
-		if (!adev->gmc.gmc_funcs->flush_gpu_tlb_pasid)
-			return 0;
+		if (!adev->gmc.gmc_funcs->flush_gpu_tlb_pasid) {
+			r = 0;
+			goto error_unlock_reset;
+		}
 
 		if (adev->gmc.flush_tlb_needs_extra_type_2)
 			adev->gmc.gmc_funcs->flush_gpu_tlb_pasid(adev, pasid,
author maintainer
#157 lore raw

PATCH 154/188 Revert "drm/amdgpu: dont attach the tlb fence for SI"

In reply to Greg Kroah-Hartman

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

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

From: Prike Liang <[email protected]>

commit 9163fe4d790fb4e16d6b0e23f55b43cddd3d4a65 upstream.

This reverts commit 820b3d376e8a102c6aeab737ec6edebbbb710e04.

It’s better to validate VM TLB flushes in the flush‑TLB backend rather than in the generic VM layer.

Reverting this patch depends on
commit fa7c231fc2b0 ("drm/amdgpu: validate the flush_gpu_tlb_pasid()") being present in the tree.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c |    4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +1 -3
@@ -1066,9 +1066,7 @@ amdgpu_vm_tlb_flush(struct amdgpu_vm_upd
 	}
 
 	/* Prepare a TLB flush fence to be attached to PTs */
-	if (!params->unlocked &&
-	    /* SI doesn't support pasid or KIQ/MES */
-	    params->adev->family > AMDGPU_FAMILY_SI) {
+	if (!params->unlocked) {
 		amdgpu_vm_tlb_fence_create(params->adev, vm, fence);
 
 		/* Makes sure no PD/PT is freed before the flush */
author maintainer
#158 lore raw

PATCH 155/188 drm/amdgpu: rework how we handle TLB fences

In reply to Greg Kroah-Hartman

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

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

From: Alex Deucher <[email protected]>

commit 69c5fbd2b93b5ced77c6e79afe83371bca84c788 upstream.

Add a new VM flag to indicate whether or not we need a TLB fence. Userqs (KFD or KGD) require a TLB fence. A TLB fence is not strictly required for kernel queues, but it shouldn't hurt. That said, enabling this unconditionally should be fine, but it seems to tickle some issues in KIQ/MES. Only enable them for KFD, or when KGD userq queues are enabled (currently via module parameter).

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c |    7 ++++++-
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h |    2 ++
 2 files changed, 8 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +6 -1
@@ -1066,7 +1066,10 @@ amdgpu_vm_tlb_flush(struct amdgpu_vm_upd
 	}
 
 	/* Prepare a TLB flush fence to be attached to PTs */
-	if (!params->unlocked) {
+	/* The check for need_tlb_fence should be dropped once we
+	 * sort out the issues with KIQ/MES TLB invalidation timeouts.
+	 */
+	if (!params->unlocked && vm->need_tlb_fence) {
 		amdgpu_vm_tlb_fence_create(params->adev, vm, fence);
 
 		/* Makes sure no PD/PT is freed before the flush */
@@ -2584,6 +2587,7 @@ int amdgpu_vm_init(struct amdgpu_device
 	ttm_lru_bulk_move_init(&vm->lru_bulk_move);
 
 	vm->is_compute_context = false;
+	vm->need_tlb_fence = amdgpu_userq_enabled(&adev->ddev);
 
 	vm->use_cpu_for_update = !!(adev->vm_manager.vm_update_mode &
 				    AMDGPU_VM_USE_CPU_FOR_GFX);
@@ -2721,6 +2725,7 @@ int amdgpu_vm_make_compute(struct amdgpu
 	dma_fence_put(vm->last_update);
 	vm->last_update = dma_fence_get_stub();
 	vm->is_compute_context = true;
+	vm->need_tlb_fence = true;
 
 unreserve_bo:
 	amdgpu_bo_unreserve(vm->root.bo);
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h +2 -0
@@ -439,6 +439,8 @@ struct amdgpu_vm {
 	struct ttm_lru_bulk_move lru_bulk_move;
 	/* Flag to indicate if VM is used for compute */
 	bool			is_compute_context;
+	/* Flag to indicate if VM needs a TLB fence (KFD or KGD) */
+	bool			need_tlb_fence;
 
 	/* Memory partition number, -1 means any partition */
 	int8_t			mem_id;