[Deepin-Kernel-SIG] [linux 6.6-y] [Upstream] Update kernel base to 6.6.117-part2

[opsiff]

git log --oneline v6.6.117429..v6.6.117329 | wc
100 886 7223

Handle:
drm/amdgpu: Respect max pixel clock for HDMI and DVI-D (v2)

• riscv: bpf: Fix uninitialized symbol 'retval_off'
• bpf: Clear pfmemalloc flag when freeing all fragments
• nvme: Use non zero KATO for persistent discovery connections
• uprobe: Do not emulate/sstep original instruction when ip is changed
• hwmon: (asus-ec-sensors) increase timeout for locking ACPI mutex
• hwmon: (dell-smm) Add support for Dell OptiPlex 7040
• tools/cpupower: Fix incorrect size in cpuidle_state_disable()
• tools/power x86_energy_perf_policy: Fix incorrect fopen mode usage
• tools/power x86_energy_perf_policy: Enhance HWP enable
• tools/power x86_energy_perf_policy: Prefer driver HWP limits
• mfd: stmpe: Remove IRQ domain upon removal
• mfd: stmpe-i2c: Add missing MODULE_LICENSE
• mfd: madera: Work around false-positive -Wininitialized warning
• mfd: da9063: Split chip variant reading in two bus transactions
• drm/amd/display: ensure committing streams is seamless
• drm/amd/display: Increase AUX Intra-Hop Done Max Wait Duration
• drm/amd/display: add more cyan skillfish devices
• drm/amd/display: update dpp/disp clock from smu clock table
• drm/amd/pm: Use cached metrics data on aldebaran
• drm/amd/pm: Use cached metrics data on arcturus
• drm/amdgpu/jpeg: Hold pg_lock before jpeg poweroff
• drm/nouveau: replace snprintf() with scnprintf() in nvkm_snprintbf()
• PCI: Disable MSI on RDC PCI to PCIe bridges
• selftests/net: Replace non-standard __WORDSIZE with sizeof(long) * 8
• selftests/net: Ensure assert() triggers in psock_tpacket.c
• wifi: rtw88: sdio: use indirect IO for device registers before power-on
• drm/amdkfd: return -ENOTTY for unsupported IOCTLs
• media: pci: ivtv: Don't create fake v4l2_fh
• media: amphion: Delete v4l2_fh synchronously in .release()
• drm/tidss: Use the crtc_ timings when programming the HW*
• drm/bridge: cdns-dsi: Fix REG_WAKEUP_TIME value
• drm/bridge: cdns-dsi: Don't fail on MIPI_DSI_MODE_VIDEO_BURST
• drm/tidss: Set crtc modesetting parameters with adjusted mode
• media: i2c: Kconfig: Ensure a dependency on HAVE_CLK for VIDEO_CAMERA_SENSOR
• x86/vsyscall: Do not require X86_PF_INSTR to emulate vsyscall
• net: stmmac: Check stmmac_hw_setup() in stmmac_resume()
• ice: Don't use %pK through printk or tracepoints
• thunderbolt: Use is_pciehp instead of is_hotplug_bridge
• powerpc/eeh: Use result of error_detected() in uevent
• s390/pci: Use pci_uevent_ers() in PCI recovery
• bridge: Redirect to backup port when port is administratively down
• scsi: ufs: host: mediatek: Fix auto-hibern8 timer configuration
• scsi: ufs: host: mediatek: Assign power mode userdata before FASTAUTO mode change
• scsi: ufs: host: mediatek: Change reset sequence for improved stability
• scsi: ufs: host: mediatek: Fix invalid access in vccqx handling
• net: ipv6: fix field-spanning memcpy warning in AH output
• media: imon: make send_packet() more robust
• drm/bridge: display-connector: don't set OP_DETECT for DisplayPorts
• drm/amdkfd: Handle lack of READ permissions in SVM mapping
• iio: adc: spear_adc: mask SPEAR_ADC_STATUS channel and avg sample before setting register
• iio: adc: imx93_adc: load calibrated values even calibration failed
• usb: gadget: f_ncm: Fix MAC assignment NCM ethernet
• char: misc: Make misc_register() reentry for miscdevice who wants dynamic minor
• char: misc: Does not request module for miscdevice with dynamic minor
• net: When removing nexthops, don't call synchronize_net if it is not necessary
• net: stmmac: Correctly handle Rx checksum offload errors
• net: Call trace_sock_exceed_buf_limit() for memcg failure with SK_MEM_RECV.
• PCI/P2PDMA: Fix incorrect pointer usage in devm_kfree() call
• ALSA: usb-audio: Add validation of UAC2/UAC3 effect units
• rds: Fix endianness annotation for RDS_MPATH_HASH
• scsi: mpi3mr: Fix controller init failure on fault during queue creation
• scsi: pm80xx: Fix race condition caused by static variables
• extcon: adc-jack: Fix wakeup source leaks on device unbind
• remoteproc: wkup_m3: Use devm_pm_runtime_enable() helper
• net: phy: fixed_phy: let fixed_phy_unregister free the phy_device
• fuse: zero initialize inode private data
• drm/amdkfd: fix vram allocation failure for a special case
• drm/amdkfd: Tie UNMAP_LATENCY to queue_preemption
• media: fix uninitialized symbol warnings
• drm/amdgpu: Respect max pixel clock for HDMI and DVI-D (v2)
• mips: lantiq: danube: add missing properties to cpu node
• mips: lantiq: danube: add model to EASY50712 dts
• mips: lantiq: danube: add missing device_type in pci node
• mips: lantiq: xway: sysctrl: rename stp clock
• mips: lantiq: danube: rename stp node on EASY50712 reference board
• crypto: qat - use kcalloc() in qat_uclo_map_objs_from_mof()
• scsi: pm8001: Use int instead of u32 to store error codes
• ptp: Limit time setting of PTP clocks
• dmaengine: sh: setup_xref error handling
• dmaengine: mv_xor: match alloc_wc and free_wc
• dmaengine: dw-edma: Set status for callback_result
• drm/msm/dsi/phy: Toggle back buffer resync after preparing PLL
• drm/msm/dsi/phy_7nm: Fix missing initial VCO rate
• drm/amdgpu: Allow kfd CRIU with no buffer objects
• ipv6: Add sanity checks on ipv6_devconf.rpl_seg_enabled
• net: nfc: nci: Increase NCI_DATA_TIMEOUT to 3000 ms
• media: adv7180: Add missing lock in suspend callback
• media: adv7180: Do not write format to device in set_fmt
• media: adv7180: Only validate format in querystd
• media: verisilicon: Explicitly disable selection api ioctls for decoders
• ALSA: usb-audio: apply quirk for MOONDROP Quark2
• net: call cond_resched() less often in __release_sock()
• smsc911x: add second read of EEPROM mac when possible corruption seen
• iommu/amd: Skip enabling command/event buffers for kdump
• iommu/apple-dart: Clear stream error indicator bits for T8110 DARTs
• drm/amd: add more cyan skillfish PCI ids
• drm/amdgpu: don't enable SMU on cyan skillfish
• drm/amdgpu: add support for cyan skillfish gpu_info
• usb: gadget: f_hid: Fix zero length packet transfer
• usb: cdns3: gadget: Use-after-free during failed initialization and exit of cdnsp gadget

Summary by Sourcery

Rebase the Deepin 6.6.y kernel to a newer 6.6.117 upstream point, picking up a broad set of driver, networking, media, and tooling fixes and small feature updates.

Bug Fixes:

• Correct UFS Mediatek auto-hibernation timers, clock gating delays, UniPro timeout configuration, reset timing, and voltage regulator handling to improve stability.
• Harden multiple media drivers (imon, adv7180, ivtv, amphion, verisilicon, IR I2C tuners, XC4000/5000) against timeouts, early callbacks, lockless access, uninitialized variables, and improper V4L2 handle lifetimes.
• Enforce HDMI/DVI TMDS pixel clock limits on AMDGPU connectors and fix display timing / clock table handling in AMD DC and TI DSS pipelines to avoid invalid modes and flicker.
• Fix IPv6 AH header handling, vsyscall emulation checks, PTP clock set/adjust validation, nexthop removal synchronization, socket backlog scheduling, and RDS multipath hashing endianness issues.
• Resolve DMA engine and IOMMU corner cases (error paths, callback results, kdump reuse) and add timeout/error propagation in SH/MV/DW EDMA and Apple DART handlers.
• Improve robustness of several storage and SCSI drivers (NVMe discovery keep-alive, UFS Mediatek calibration and power modes, mpi3mr queue bring-up, pm8001 logging, pm80xx race) under fault conditions.
• Address USB gadget issues in HID and NCM functions and Cadence CDNSP, including zero-length packet handling, MAC assignment, teardown ordering, and endpoint freeing.
• Tighten hwmon, PCI, Thunderbolt, EEH, and s390 PCI error reporting and MSI quirks, including new Dell OptiPlex support and RDC bridge MSI disablement.
• Fix miscdevice registration/deregistration corner cases, FUSE inode private data initialization, and cpupower/cpuidle/x86_energy_perf_policy userspace tool bugs (file modes, sizes, HWP handling).
• Repair various networking driver issues in stmmac, smsc911x, ice tracing, DSA loop, NCI timeout handling, fixed PHY lifetime, bridge backup-port redirection, and rtw88 SDIO power-on access.
• Ensure AMDGPU JPEG power-gating, KFD CRIU/ioctl handling, SVM mapping permissions, SDMA unmap timeout calculation, and GPU metrics paths behave correctly for new cyan_skillfish and other ASICs.
• Refresh validation and quirks in ALSA USB audio for UAC2/3 effect units and specific devices (e.g., MOONDROP Quark2) to avoid parsing errors and mis-calibrated volumes.
• Prevent use-after-free and race conditions in STPME MFD IRQ domain, ADC drivers (spear, imx93), extcon-adc-jack wakeup, wkup_m3 PM, and pm8001 log access.
• Increase robustness of tests and selftests (psock_tpacket) by making assertions effective and replacing non-portable word-size macros.

Enhancements:

• Tune auto-hibern8 and clock gating behavior for specific UFS vendor devices and expose more precise DP training / intra-hop wait times.
• Prefer updating Intel HWP performance limits via cpufreq sysfs when available, log full MSR changes, and keep bandwidth tables aligned to maximum display clocks.
• Adjust DP and DSI bridge behavior (Cadence DSI wakeup time, display-connector detection rules) and make DC commit transitions more seamless across topology changes.
• Improve AMD and Qualcomm display PLL and bandwidth setup for DCN3.0.1 and SDM845 7nm PHY, including proper VCO rate storage and buffer resync controls.
• Refine AMD DC ASIC ID handling and PCI ID tables to support additional cyan_skillfish variants and their SMU enablement constraints.
• Clarify and harden various error paths and logging in SCSI (pm8001/pm80xx, mpi3mr), PCI P2PDMA, and XDP frag handling to aid diagnosis and recovery.
• Use mutex guards and modern helper macros (guard(mutex), FIELD_PREP, timespec64 helpers) across subsystems to simplify state management and avoid subtle races.

Tests:

• Update networking selftests to avoid glibc-specific __WORDSIZE usage and ensure assertions are active during psock_tpacket runs.

[sourcery-ai]

Reviewer's Guide

Kernel rebase to 6.6.117 (part 2) with a large set of upstream backports focused on storage (UFS Mediatek, NVMe), AMD GPU/AMDKFD, media (V4L2 tuners/codecs), DMA/IOMMU, networking (stmmac, bridge, IPv4/IPv6, NFC), USB gadget, hwmon and misc core subsystems, mainly fixing bugs, tightening validation, and improving power/error handling; local changes are mechanical backports of these upstream patches with minimal adaptation.

Sequence diagram for UFS Mediatek link setup and Auto-Hibernate/clock-gating handling

sequenceDiagram
    participant Probe as ufs_mtk_probe
    participant HBA as ufs_hba
    participant Variant as ufs_mtk_host

    Probe->>HBA: ufshcd_alloc_host()
    Probe->>HBA: ufshcd_init()
    Probe->>Variant: ufs_mtk_fixup_dev_quirks(hba)
    activate Variant
    Variant->>Variant: ufs_mtk_vreg_fix_vcc(hba)
    Variant->>Variant: ufs_mtk_vreg_fix_vccqx(hba)
    Variant->>Variant: ufs_mtk_fix_ahit(hba)
    activate Variant
    alt Auto-hibern8 supported
        Variant->>HBA: check dev_info.wmanufacturerid
        alt Samsung
            Variant->>Variant: us = 3500
        else Micron
            Variant->>Variant: us = 2000
        else Other
            Variant->>Variant: us = 1000
        end
        Variant->>Variant: ufs_mtk_us_to_ahit(us)
        Variant->>HBA: hba->ahit = encoded_timer
    end
    Variant->>Variant: ufs_mtk_setup_clk_gating(hba)
    activate Variant
    Variant->>HBA: ufshcd_is_clkgating_allowed(hba)?
    alt Clock gating allowed
        Variant->>HBA: if auto_hibern8 supported and hba->ahit != 0
        Variant->>Variant: decode ah_scale, ah_timer from hba->ahit
        Variant->>Variant: ah_ms = timer * scale / 1000
        Variant->>HBA: clk_gating.delay_ms = max(ah_ms, 10)
    end
    deactivate Variant
    deactivate Variant
    deactivate Variant

    Note over HBA,Variant: Later during power change

    HBA->>Variant: ufs_mtk_pre_pwr_change(hba, ...) 
    activate Variant
    alt FASTAUTO mode and !UFSHCD_QUIRK_SKIP_DEF_UNIPRO_TIMEOUT_SETTING
        Variant->>HBA: ufshcd_dme_set(PA_PWRMODEUSERDATA0..5,
        Variant->>HBA: ufshcd_dme_set(DME_LocalFC0/TC0/AFC0, default_vals)
    end
    Variant->>HBA: ufshcd_uic_change_pwr_mode(... FASTAUTO ...)
    deactivate Variant

    Note over HBA: Device reset timing change

    HBA->>Variant: ufs_mtk_device_reset(hba)
    activate Variant
    Variant->>Variant: ufs_mtk_device_reset_ctrl(0, res)
    Variant->>HBA: ufshcd_hba_stop(hba)  %% now in middle of reset
    Variant->>Variant: ufs_mtk_device_reset_ctrl(1, res)
    deactivate Variant

Loading

File-Level Changes

Change	Details	Files
Adjust UFS Mediatek host driver power management, auto-hibernate timer and clock-gating behaviour for better stability and vendor-specific tuning	Introduce ufs_mtk_fix_ahit() to program auto-hibernate idle timer (AHIT) per UFS vendor and convert microseconds to register encoding via ufs_mtk_us_to_ahit() Rework ufs_mtk_setup_clk_gating() to derive clk_gating.delay_ms from AHIT under host lock instead of using ufshcd_clkgate_delay_set() directly Set UniPro timeout user data registers in ufs_mtk_pre_pwr_change() unless UFSHCD_QUIRK_SKIP_DEF_UNIPRO_TIMEOUT_SETTING is set Change device reset flow to stop the HBA after asserting the reset line rather than before, and guard vccqx LPM updates when no vccq/vccq2 regulators are present Invoke the new AHIT fixup from ufs_mtk_fixup_dev_quirks()	drivers/ufs/host/ufs-mediatek.c
Improve robustness and error handling in several media and input USB drivers (imon, adv7180, amphion, ivtv, verisilicon, ir-kbd, tuners)	In imon, add timeout to send_packet() completion wait, capture error vs timeout in tx.status, relax early dev_present checks in RX callbacks, and treat some USB errors as terminal without resubmitting URBs In adv7180, rework querystd() to avoid changing standard while streaming, simplify pad format handling, and guard suspend/resume power transitions with the state mutex In amphion V4L2, ensure v4l2_fh_del/exit are called on open error and in close(), not in the generic release path In ivtv, replace v4l2_fh-based stream ownership tracking with ivtv_open_id *id, adjust claim/release/read/irq paths accordingly, and avoid creating fake v4l2_fh for ALSA PCM capture Disable selection ioctls (G/S_SELECTION) for the verisilicon decoder node at registration time Initialize various debug variables to safe defaults in xc4000/xc5000 tuners and ir-kbd-i2c to silence uninitialized warnings	drivers/media/rc/imon.c drivers/media/i2c/adv7180.c drivers/media/platform/amphion/vpu_v4l2.c drivers/media/pci/ivtv/ivtv-fileops.c drivers/media/pci/ivtv/ivtv-irq.c drivers/media/pci/ivtv/ivtv-driver.h drivers/media/pci/ivtv/ivtv-alsa-pcm.c drivers/media/platform/verisilicon/hantro_v4l2.c drivers/media/platform/verisilicon/hantro_drv.c drivers/media/i2c/ir-kbd-i2c.c drivers/media/tuners/xc5000.c drivers/media/tuners/xc4000.c
Tighten HDMI/DVI mode validation and clock handling, add support for new AMD "cyan skillfish" variants and improve AMD display/PM logic	Introduce amdgpu_max_hdmi_pixel_clock() and use it in amdgpu_connector_dvi_mode_valid() to enforce per-ASIC HDMI TMDS limits and correct DVI single/dual-link bandwidth checks, including EDID HDMI displays Make DC commit_streams() optionally insert a minimal intermediate state if the topology transition is not seamless according to hwss Compute dispclk/dppclk from max values across voltage levels in dcn301 bounding-box and vg clock manager, rather than per-state limits Extend AMDGPU PCI ID tables and DC ASIC parsing with additional cyan skillfish device IDs and a cyan_skillfish_gpu_info firmware, and only enable SMU for cyan skillfish v2 APU variants Use cached metrics table instead of live SMU queries for Arcturus and Aldebaran get_gpu_metrics() to avoid stalling paths In JPEG PG idle work, protect power-gating with jpeg_pg_lock mutex Let AMDKFD allow CRIU with no buffer objects and treat out-of-range IOCTL numbers as -ENOTTY	drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c drivers/gpu/drm/amd/display/dc/core/dc.c drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.c drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/vg_clk_mgr.c drivers/gpu/drm/amd/display/dc/core/dc_resource.c drivers/gpu/drm/amd/display/include/dal_asic_id.h drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c drivers/gpu/drm/amd/amdgpu/amdgpu_device.c drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c drivers/gpu/drm/amd/amdkfd/kfd_chardev.c drivers/gpu/drm/amd/amdkfd/kfd_priv.h drivers/gpu/drm/amd/amdkfd/kfd_svm.c
Fix various networking correctness and performance issues (IPv4/v6, TCP backlog processing, bridge, NFC, STMMAC, RDS, nexthop, PTP)	Refactor AH IPv6 header saving/restoring in ah6_output/ah6_output_done() using helpers to avoid field-spanning memcpy and handle MIP6 correctly In __release_sock(), reduce cond_resched() frequency by only calling it every 16 backlog packets and keep processing under separate loop; always emit trace_sock_exceed_buf_limit() even for failed allocations Have the bridge redirect to backup port when primary is down or administratively down (not running) Increase NCI_DATA_TIMEOUT to 3000 ms in NFC core and lengthen IPv6 devconf rpl_seg_enabled sysctl with min/max validation (0..1) For STMMAC, treat Rx descriptors with csum_none error as checksum-none and make resume path check stmmac_hw_setup() return value Avoid unnecessary synchronize_net() when removing nexthops with empty grp_list Fix RDS multipath hash to use ntohs(rs_bound_port) for correct endianness In PTP clock, validate settime and offset operations using timespec64_valid_settod and reject invalid times before calling driver ops	net/ipv6/ah6.c net/core/sock.c net/bridge/br_forward.c include/net/nfc/nci_core.h net/ipv6/addrconf.c drivers/net/ethernet/stmicro/stmmac/stmmac_main.c net/ipv4/nexthop.c net/rds/rds.h drivers/ptp/ptp_clock.c
Improve DMA, IOMMU and PCI robustness (SH/MV DMA, DW eDMA, AMD/Apple IOMMU, MPI3MR, P2PDMA, MSI quirks, Thunderbolt)	Change shdma_ops.setup_xfer() and sh_dmae_setup_xfer() to return int and propagate errors from DMAE register programming; in shdma_tx_submit(), on failure, move queued chunks back to free list, reset PM state, and drop runtime PM reference In MV XOR DMA, pair alloc_wc/free_wc by switching the free call to dma_free_wc Add DW eDMA helper to populate callback_result with DMA_TRANS_NOERROR or DMA_TRANS_ABORTED plus residue when transfers complete or abort Skip enabling AMD IOMMU command/event MMIO buffers when running as kdump kernel, while still resetting internal buffers, and for Apple DART T8110, clear per-stream error indicator bits on IRQ In MPI3MR, after creating operational queues, check IOC status and state (READY, no RESET_HISTORY) before declaring success Fix pci_p2pdma_add_resource() to devm_kfree the right pgmap object and add RDC bridge to MSI-disable quirk list; use pdev->is_pciehp instead of is_hotplug_bridge in Thunderbolt and MSI hotplug quirks Extend pci_uevent_ers/pci_uevent_ers() support to s390 and use it in powerpc and s390 EEH paths to signal recovery results	drivers/dma/sh/shdma-base.c drivers/dma/sh/shdmac.c drivers/dma/mv_xor.c drivers/dma/dw-edma/dw-edma-core.c drivers/iommu/amd/init.c drivers/iommu/apple-dart.c drivers/scsi/mpi3mr/mpi3mr_fw.c drivers/pci/p2pdma.c drivers/pci/quirks.c drivers/thunderbolt/tb.c drivers/pci/pci-driver.c include/linux/pci.h arch/powerpc/kernel/eeh_driver.c arch/s390/pci/pci_event.c
Tighten USB gadget function behaviour and MAC address handling, and improve misc and hwmon drivers	In f_hid and f_ncm gadget functions, set req->zero for HID writes and EP0 replies to ensure proper ZLP handling and bind the NCM MAC string at bind time to pick the configured address In cdnsp gadget init/exit, use usb_del_gadget/usb_put_gadget and free endpoints reliably on init failure and shutdown; fix use-after-free by ensuring gadget lifetime is managed Make asus-ec-sensors ACPI mutex timeout more conservative (800 ms) and register Dell OptiPlex 7040 in dell-smm-hwmon DMI table In misc_open(), only request modules for fixed minors (< MISC_DYNAMIC_MINOR) and avoid module loading for fully dynamic minors; in misc_deregister(), reset misc->minor to MISC_DYNAMIC_MINOR after destroying the device In extcon-adc-jack, disable wakeup on remove to avoid wakeup source leaks; ensure stmpe-i2c has a MODULE_LICENSE and stmpe core removes IRQ domain on removal Have smsc911x read MAC ADDRH twice when zero is seen to work around possible invalid first read	drivers/usb/gadget/function/f_hid.c drivers/usb/gadget/function/f_ncm.c drivers/usb/cdns3/cdnsp-gadget.c drivers/hwmon/asus-ec-sensors.c drivers/hwmon/dell-smm-hwmon.c drivers/char/misc.c drivers/extcon/extcon-adc-jack.c drivers/mfd/stmpe.c drivers/mfd/stmpe-i2c.c drivers/net/ethernet/smsc/smsc911x.c
Miscellaneous tooling, BPF, NVMe, vsyscall and other subsystem fixes	In cpupower cpuidle library, write the correct number of bytes to the sysfs disable file based on snprintf length, and in x86_energy_perf_policy, fix fopen modes and honor driver-selected HWP min/max when sysfs is used; enable HWP only if not already enabled Clear XDP frag pfmemalloc flag when shrinking fragments to zero and add helper xdp_buff_clear_frag_pfmemalloc() Fix riscv BPF trampoline retval_off initialization to avoid using an uninitialized offset and make MIPS/OF fixed PHY unregister paths not double-free phy_device For NVMe, when reconnecting persistent discovery controllers without a KATO, set a default KATO and restart keep-alive before sending rediscover uevent Relax vsyscall emulation to not rely solely on X86_PF_INSTR by comparing fault address with regs->ip and only WARN if NX is supported and X86_PF_INSTR is not set In uprobes, skip emulating the original instruction if the handler changed instruction_pointer(), avoiding mis-execution after probes that divert control flow Apply audio-class descriptor validation for UAC2/UAC3 effect units in usb-audio validate.c and add a MOONDROP Quark2 volume quirk in mixer.c Adjust dmaengine shdma API and Hantro selection ioctls, and several small cleanups like using scnprintf in nouveau nvkm_snprintbf and initializing madera mfd_devs only when available	tools/power/cpupower/lib/cpuidle.c tools/power/x86/x86_energy_perf_policy/x86_energy_perf_policy.c include/net/xdp.h net/core/filter.c arch/riscv/net/bpf_jit_comp64.c drivers/net/phy/fixed_phy.c drivers/net/mdio/of_mdio.c drivers/nvme/host/core.c arch/x86/entry/vsyscall/vsyscall_64.c kernel/events/uprobes.c sound/usb/validate.c sound/usb/mixer.c drivers/gpu/drm/nouveau/nvkm/core/enum.c drivers/mfd/madera-core.c

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.

[github-actions]

WARNING: Prefer a maximum 75 chars per line (possible unwrapped commit description?)

[github-actions]

WARNING: Prefer a maximum 75 chars per line (possible unwrapped commit description?)

[github-actions]

WARNING: Prefer a maximum 75 chars per line (possible unwrapped commit description?)

[github-actions]

WARNING: Prefer a maximum 75 chars per line (possible unwrapped commit description?)

[github-actions]

WARNING: Prefer a maximum 75 chars per line (possible unwrapped commit description?)

[deepin-ci-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by:
Once this PR has been reviewed and has the lgtm label, please ask for approval from opsiff. For more information see the Code Review Process.

The full list of commands accepted by this bot can be found here.

Needs approval from an approver in each of these files:

• deepin/OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[deepin-ci-robot]

deepin pr auto review

我来帮你审查这个 git diff，从语法逻辑、代码质量、性能和安全等方面进行分析。

1. 语法逻辑方面：

• 大部分修改都遵循了内核代码的规范，没有明显的语法错误
• 变量命名和函数命名符合规范
• 错误处理逻辑完整

2. 代码质量改进：

• 在 arch/mips/boot/dts/lantiq/danube.dtsi 中添加了必要的设备树属性(#address-cells, #size-cells等)，提高了设备树的完整性
• 在 drivers/gpu/drm/amd/amdgpu/ 相关文件中添加了对新硬件的支持，代码结构清晰
• 在 drivers/net/ethernet/stmicro/stmmac/stmmac_main.c 中改进了错误处理逻辑

3. 性能优化：

• 在 net/core/sock.c 中优化了 __release_sock 函数，通过批量处理和条件调度提高了性能
• 在 drivers/net/dma 相关驱动中优化了内存管理和中断处理

4. 安全性改进：

• 在 drivers/iio/adc/imx93_adc.c 中改进了校准失败的处理逻辑，避免系统崩溃
• 在 drivers/usb/gadget/function/f_hid.c 中修复了 zero flag 的设置，提高了安全性
• 在 kernel/events/uprobes.c 中增加了对指令指针的检查，提高了安全性

5. 具体建议：

• 在 arch/mips/boot/dts/lantiq/danube.dtsi 中，建议添加更多的注释说明新增属性的作用
• 在 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c 中，建议对新增的 GPU_INFO 固件添加版本检查
• 在 net/core/sock.c 中，建议对批量处理的数量(15)添加宏定义，提高代码可维护性
• 在 drivers/usb/gadget/function/f_hid.c 中，建议对 zero flag 的修改添加更详细的注释说明原因

6. 其他改进建议：

• 建议在修改硬件相关代码时，添加更多的调试信息，方便问题定位
• 对于新添加的设备ID，建议添加对应的文档说明
• 在性能相关的修改中，建议添加性能测试数据作为注释

总体来说，这个 diff 的修改质量较高，涵盖了多个子系统的改进，包括硬件支持、性能优化和安全性增强。建议在提交前进行更充分的测试，特别是新硬件支持部分。

[sourcery-ai]

Hey there - I've reviewed your changes and found some issues that need to be addressed.

• In drivers/char/misc.c, the logic resetting misc->minor in misc_deregister() only triggers when minor > MISC_DYNAMIC_MINOR, which will never reset dynamically allocated minors (== MISC_DYNAMIC_MINOR); consider using >= to ensure dynamic minors are correctly reset after deregistration.
• In f_ncm.c, ncm_bind() now hard-codes index 1 when assigning ncm_string_defs[1].s = ncm->ethaddr; it would be more robust to use the existing STRING_MAC_IDX constant (as done previously) instead of a literal index to avoid breakage if the enum layout changes.

Prompt for AI Agents

Please address the comments from this code review:

## Overall Comments
- In drivers/char/misc.c, the logic resetting misc->minor in misc_deregister() only triggers when minor > MISC_DYNAMIC_MINOR, which will never reset dynamically allocated minors (== MISC_DYNAMIC_MINOR); consider using >= to ensure dynamic minors are correctly reset after deregistration.
- In f_ncm.c, ncm_bind() now hard-codes index 1 when assigning ncm_string_defs[1].s = ncm->ethaddr; it would be more robust to use the existing STRING_MAC_IDX constant (as done previously) instead of a literal index to avoid breakage if the enum layout changes.

## Individual Comments

### Comment 1
<location> `drivers/remoteproc/wkup_m3_rproc.c:151-152` </location>
<code_context>
 	}

-	pm_runtime_enable(&pdev->dev);
+	ret = devm_pm_runtime_enable(dev);
+	if (ret < 0)
+		return dev_err_probe(dev, ret, "Failed to enable runtime PM\n");
 	ret = pm_runtime_get_sync(&pdev->dev);
</code_context>

<issue_to_address>
**issue (bug_risk):** devm_pm_runtime_enable returns void, assigning to `ret` will not compile

`devm_pm_runtime_enable()` returns `void`, so assigning it to `ret` and checking `ret < 0` is a compile error. If you need an error code, use `pm_runtime_enable()` (which returns `int`) and handle disable in error paths, or stop checking the return value and rely on the devm helper semantics.
</issue_to_address>

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}

[Copilot]

Pull request overview

This pull request backports 100 upstream kernel patches from v6.6.117429 to v6.6.117329 to the Deepin 6.6.y kernel branch. The changes span multiple subsystems with bug fixes, hardware support additions, and stability improvements.

Key changes:

• Fix critical bugs in networking (IPv6 AH, RDS endianness, socket backlog, bridge forwarding), USB gadget (NCM MAC, HID zero-length packets, cdnsp teardown), and media drivers (imon timeout, ivtv fake v4l2_fh, adv7180 locking)
• Add hardware support for Dell OptiPlex 7040, MOONDROP Quark2 USB audio, additional AMD cyan_skillfish GPU variants, and RDC PCI bridge MSI quirks
• Improve stability in SCSI/UFS (Mediatek auto-hibernation, calibration, reset timing), AMDGPU (JPEG power-gating, KFD CRIU/IOCTL handling, display clock tables), and DMA engines (error propagation, callback results)

Reviewed changes

Copilot reviewed 100 out of 100 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
tools/testing/selftests/net/psock_tpacket.c	Remove non-portable __WORDSIZE macro, enable assertions
tools/power/x86/x86_energy_perf_policy/x86_energy_perf_policy.c	Fix fopen mode, enhance HWP handling, prefer driver limits
tools/power/cpupower/lib/cpuidle.c	Fix incorrect size calculation in cpuidle_state_disable
sound/usb/validate.c	Add validation for UAC2/UAC3 effect units
sound/usb/mixer.c	Add volume quirk for MOONDROP Quark2
net/rds/rds.h	Fix endianness in RDS_MPATH_HASH
net/ipv6/ah6.c	Fix field-spanning memcpy with helper functions
net/ipv6/addrconf.c	Add sanity checks on rpl_seg_enabled sysctl
net/ipv4/nexthop.c	Skip synchronize_net when grp_list is empty
net/core/sock.c	Call cond_resched less often, fix trace call for memcg
net/core/filter.c	Clear pfmemalloc flag when freeing XDP fragments
net/bridge/br_forward.c	Check administratively down ports for backup redirection
kernel/events/uprobes.c	Skip emulation when IP changed by handler
include/net/xdp.h	Add xdp_buff_clear_frag_pfmemalloc helper
include/net/nfc/nci_core.h	Increase NCI_DATA_TIMEOUT to 3000ms
include/linux/shdma-base.h	Change setup_xfer return type to int
include/linux/pci.h	Add S390 to pci_uevent_ers availability
fs/fuse/inode.c	Zero initialize all inode private data
drivers/usb/gadget/function/f_ncm.c	Fix MAC address assignment timing
drivers/usb/gadget/function/f_hid.c	Enable zero-length packet transfers
drivers/usb/cdns3/cdnsp-gadget.c	Fix use-after-free in gadget teardown
drivers/ufs/host/ufs-mediatek.c	Fix auto-hibernate timer, power mode settings, reset sequence, vccqx handling
drivers/thunderbolt/tb.c	Use is_pciehp instead of is_hotplug_bridge
drivers/scsi/pm8001/*	Fix race condition with static variables, use int for error codes
drivers/scsi/mpi3mr/mpi3mr_fw.c	Check controller state after queue creation
drivers/remoteproc/wkup_m3_rproc.c	Use devm_pm_runtime_enable helper
drivers/ptp/ptp_clock.c	Validate time settings for PTP clocks
drivers/pci/quirks.c	Disable MSI on RDC bridges, use is_pciehp
drivers/pci/pci-driver.c	Add S390 support for pci_uevent_ers
drivers/pci/p2pdma.c	Fix incorrect pointer in devm_kfree
drivers/nvme/host/core.c	Set non-zero KATO for persistent discovery
drivers/net/wireless/realtek/rtw88/sdio.c	Use indirect IO before power-on
drivers/net/phy/fixed_phy.c	Let fixed_phy_unregister free phy_device
drivers/net/mdio/of_mdio.c	Remove duplicate phy_device_free call
drivers/net/ethernet/stmicro/stmmac/stmmac_main.c	Check stmmac_hw_setup result, fix Rx checksum handling
drivers/net/ethernet/smsc/smsc911x.c	Re-read MAC on corruption
drivers/net/ethernet/intel/ice/*	Replace %pK with %p in traces/debug
drivers/net/dsa/dsa_loop.c	Remove duplicate phy_device_free
drivers/mfd/stmpe.c	Remove IRQ domain on removal
drivers/mfd/stmpe-i2c.c	Add missing MODULE_LICENSE
drivers/mfd/madera-core.c	Fix false-positive uninitialized warning
drivers/mfd/da9063-i2c.c	Split chip variant reading into two transactions
drivers/media/tuners/*	Initialize variables to fix uninitialized warnings
drivers/media/rc/imon.c	Add timeout to wait_for_completion, handle errors
drivers/media/platform/verisilicon/*	Disable selection ioctls for decoders
drivers/media/platform/amphion/vpu_v4l2.c	Delete v4l2_fh synchronously in release
drivers/media/pci/ivtv/*	Replace fake v4l2_fh with ivtv_open_id
drivers/media/i2c/ir-kbd-i2c.c	Initialize variables in ir_key_poll
drivers/media/i2c/adv7180.c	Add locking, fix format handling, validate querystd
drivers/media/i2c/Kconfig	Add HAVE_CLK dependency for camera sensors
drivers/iommu/apple-dart.c	Clear stream error indicator for T8110
drivers/iommu/amd/init.c	Skip buffer setup in kdump kernel
drivers/iio/adc/spear_adc.c	Mask channel and avg sample bits properly
drivers/iio/adc/imx93_adc.c	Load calibration even on failure
drivers/hwmon/dell-smm-hwmon.c	Add Dell OptiPlex 7040 support
drivers/hwmon/asus-ec-sensors.c	Increase ACPI mutex timeout
drivers/gpu/drm/tidss/*	Use crtc_* timing fields
drivers/gpu/drm/nouveau/nvkm/core/enum.c	Replace snprintf with scnprintf
drivers/gpu/drm/msm/dsi/phy/dsi_phy_7nm.c	Toggle buffer resync, store VCO rate
drivers/gpu/drm/bridge/display-connector.c	Don't set detect for DisplayPort
drivers/gpu/drm/bridge/cadns-dsi-core.c	Fix REG_WAKEUP_TIME, support burst mode
drivers/gpu/drm/amd/pm/swsmu/*	Use cached metrics on aldebaran/arcturus
drivers/gpu/drm/amd/display/*	Add cyan_skillfish IDs, fix clock tables, ensure seamless commits
drivers/gpu/drm/amd/amdkfd/*	Fix CRIU, SVM permissions, IOCTL errors, unmap latency
drivers/gpu/drm/amd/amdgpu/*	Hold pg_lock for JPEG, add cyan_skillfish support, fix HDMI/DVI limits
drivers/extcon/extcon-adc-jack.c	Fix wakeup source leak
drivers/dma/sh/*	Add error handling for setup_xfer
drivers/dma/mv_xor.c	Match alloc_wc/free_wc calls
drivers/dma/dw-edma/dw-edma-core.c	Set status for callback_result
drivers/crypto/intel/qat/qat_common/qat_uclo.c	Use kcalloc for safety
drivers/char/misc.c	Fix reentry for dynamic minor devices
arch/x86/entry/vsyscall/vsyscall_64.c	Don't require X86_PF_INSTR for emulation
arch/s390/pci/pci_event.c	Use pci_uevent_ers in recovery
arch/riscv/net/bpf_jit_comp64.c	Fix uninitialized retval_off
arch/powerpc/kernel/eeh_driver.c	Use error_detected result in uevent
arch/mips/*	Fix Lantiq Danube device tree properties

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

[nitpick] The variable initialization int j = 0, max_dispclk_mhz = 0, max_dppclk_mhz = 0; sets j = 0, but j is later initialized in line 352 to dcn3_01_soc.num_states - 1 and used in a backward loop. The initial j = 0 is unnecessary since it's overwritten before use and could be misleading.

Suggested change

	int j = 0, max_dispclk_mhz = 0, max_dppclk_mhz = 0;
	int j, max_dispclk_mhz = 0, max_dppclk_mhz = 0;

[Copilot]

The error path returns the result of pm_runtime_put() instead of the original setup_xfer error code. If pm_runtime_put() succeeds (returns 0), this masks the original setup_xfer failure (which was < 0). The function should return the original setup_xfer error code, not the pm_runtime_put result. Consider: int setup_ret = ret; ret = pm_runtime_put(schan->dev); ... return setup_ret;

Suggested change

	ret = pm_runtime_put(schan->dev);
	spin_unlock_irq(&schan->chan_lock);
	return ret;
	{
	int setup_ret = ret;
	pm_runtime_put(schan->dev);
	spin_unlock_irq(&schan->chan_lock);
	return setup_ret;
	}

[Copilot]

Typo in comment: "resault" should be "result".

[Copilot]

Typo in comment: "precidence" should be "precedence".

[Copilot]

[nitpick] The condition check here is inverted. The code should skip the special case when args->va_addr is 0 AND the VRAM flag is set. The current double negative !(!args->va_addr && ...) is correct but unnecessarily complex. Consider simplifying to: if ((args->va_addr || !(flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM)) && interval_tree_iter_first(...)) for better readability.

Suggested change

	if (!(!args->va_addr && (flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM)) &&
	if ((args->va_addr || !(flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM)) &&