[Deepin-Kernel-SIG] [linux 6.6-y] [Upstream] Update kernel base to 6.6.125,6.6.126

[opsiff]

Update kernel base to 6.6.126.

git log --oneline v6.6.125..v6.6.126
56865d9b7074c (tag: v6.6.126) Linux 6.6.126
673dafb9a8634 Revert "driver core: enforce device_lock for driver_match_device()"

Update kernel base to 6.6.125.

git log --oneline v6.6.124..v6.6.125 |wc
26 239 1954

Anil Gurumurthy (4):
scsi: qla2xxx: Validate sp before freeing associated memory
scsi: qla2xxx: Delay module unload while fabric scan in progress
scsi: qla2xxx: Free sp in error path to fix system crash
scsi: qla2xxx: Query FW again before proceeding with login

Bibo Mao (2):
crypto: virtio - Add spinlock protection with virtqueue notification
crypto: virtio - Remove duplicated virtqueue_kick in
virtio_crypto_skcipher_crypt_req

Bitterblue Smith (1):
wifi: rtw88: Fix alignment fault in rtw_core_enable_beacon()

Danilo Krummrich (1):
gpio: omap: do not register driver in probe()

Edward Adam Davis (1):
nilfs2: Fix potential block overflow that cause system hang

Eric Dumazet (1):
mptcp: fix race in mptcp_pm_nl_flush_addrs_doit()

Greg Kroah-Hartman (1):
Linux 6.6.125

Gui-Dong Han (1):
driver core: enforce device_lock for driver_match_device()

Henrique Carvalho (2):
smb: client: split cached_fid bitfields to avoid shared-byte RMW races
smb: server: fix leak of active_num_conn in ksmbd_tcp_new_connection()

Jeff Layton (1):
nfsd: don't ignore the return code of svc_proc_register()

Kees Cook (1):
crypto: omap - Allocate OMAP_CRYPTO_FORCE_COPY scatterlists correctly

Khairul Anuar Romli (1):
spi: cadence-quadspi: Implement refcount to handle unbind during busy

Konstantin Shkolnyy (1):
vsock/test: verify socket options after setting them

Marek Behún (1):
net: sfp: Fix quirk for Ubiquiti U-Fiber Instant SFP module

Matthieu Baerts (NGI0) (1):
selftests: mptcp: pm: ensure unknown flags are ignored

Namjae Jeon (1):
ksmbd: fix infinite loop caused by next_smb2_rcv_hdr_off reset in
error paths

Pablo Neira Ayuso (2):
netfilter: nf_tables: missing objects with no memcg accounting
netfilter: nft_set_pipapo: prevent overflow in lookup table allocation

Shreyas Deodhar (1):
scsi: qla2xxx: Allow recovery for tape devices

Thorsten Blum (1):
crypto: octeontx - Fix length check to avoid truncation in
ucode_load_store

Zenm Chen (1):
Bluetooth: btusb: Add USB ID 7392:e611 for Edimax EW-7611UXB

Makefile | 2 +-
drivers/base/base.h | 9 ++
drivers/base/bus.c | 2 +-
drivers/base/dd.c | 2 +-
drivers/bluetooth/btusb.c | 2 +
.../crypto/marvell/octeontx/otx_cptpf_ucode.c | 2 +-
drivers/crypto/omap-crypto.c | 2 +-
drivers/crypto/virtio/virtio_crypto_core.c | 5 +
.../virtio/virtio_crypto_skcipher_algs.c | 2 -
drivers/gpio/gpio-omap.c | 22 ++-
drivers/net/phy/sfp.c | 2 +
drivers/net/wireless/realtek/rtw88/main.c | 4 +-
drivers/scsi/qla2xxx/qla_gs.c | 41 +++--
drivers/scsi/qla2xxx/qla_init.c | 28 ++--
drivers/scsi/qla2xxx/qla_isr.c | 19 ++-
drivers/scsi/qla2xxx/qla_os.c | 3 +-
drivers/spi/spi-cadence-quadspi.c | 34 +++++
fs/nfsd/nfsctl.c | 9 +-
fs/nfsd/stats.c | 4 +-
fs/nfsd/stats.h | 2 +-
fs/nilfs2/sufile.c | 4 +
fs/smb/client/cached_dir.h | 8 +-
fs/smb/server/server.c | 6 +-
fs/smb/server/transport_tcp.c | 3 +-
net/mptcp/pm_netlink.c | 16 +-
net/netfilter/nf_tables_api.c | 2 +-
net/netfilter/nft_compat.c | 6 +-
net/netfilter/nft_log.c | 2 +-
net/netfilter/nft_meta.c | 2 +-
net/netfilter/nft_numgen.c | 2 +-
net/netfilter/nft_set_pipapo.c | 64 +++++---
net/netfilter/nft_tunnel.c | 5 +-
.../testing/selftests/net/mptcp/pm_netlink.sh | 4 +
tools/testing/selftests/net/mptcp/pm_nl_ctl.c | 11 ++
tools/testing/vsock/control.c | 9 +-
tools/testing/vsock/util.c | 143 ++++++++++++++++++
tools/testing/vsock/util.h | 7 +
tools/testing/vsock/vsock_test.c | 29 ++--
38 files changed, 405 insertions(+), 114 deletions(-)

Summary by Sourcery

Update to Linux 6.6.125 and integrate upstream fixes across networking, storage, filesystems, driver core, and test utilities.

Bug Fixes:

• Prevent overflows and allocation issues in nftables pipapo lookup tables and NILFS2 filesystem trimming.
• Fix qla2xxx SCSI error paths, fabric scan handling, and logout processing to avoid crashes and improve recovery, including during HBA unload.
• Ensure virtio-crypto queue completion handling is properly synchronized and remove redundant virtqueue kicks.
• Correct rtw88 beacon enable/disable TX pause handling to avoid alignment faults.
• Avoid use-after-unbind and busy removal issues in Cadence QSPI by adding refcounting and inflight operation tracking.
• Fix ksmbd server connection teardown and status handling, and eliminate a cached_fid bitfield race by using full booleans.
• Ensure nfsd correctly handles failures when registering proc stats and does not ignore svc_proc_register return codes.
• Fix mptcp address flush race conditions by safely detaching and freeing address lists under RCU.
• Prevent buffer overflows and length mischecks in OcteonTX ucode loader and OMAP crypto scatterlist allocation, and enforce memcg-aware allocations in several netfilter paths.
• Ensure OMAP GPIO/MPUIO drivers are registered and unregistered in a safe order and only once, avoiding probe-time driver registration issues.

Enhancements:

• Enforce device locking when matching drivers to devices in the core driver model.
• Extend sfp, btusb, and qla2xxx behaviors to support additional hardware quirks and tape device recovery.
• Improve virtio vsock and MPTCP selftests by adding socket option verification helpers, unknown-flag handling, and new test coverage for address flags.
• Refine ksmbd server error handling to abort processing on invalid or unauthorized commands instead of continuing.

[sourcery-ai]

Reviewer's Guide

Rebases the Deepin 6.6.y kernel to upstream 6.6.125 and pulls in the corresponding upstream fixes across storage (qla2xxx, NILFS2), networking (netfilter/nftables, MPTCP, vsock tests, rtw88, SFP, Bluetooth), crypto (virtio, omap, octeontx), SMB (client and ksmbd), GPIO/driver core, SPI cadence-quadspi, and NFSD proc stats.

Sequence diagram for SPI cadence-quadspi refcounted memory operations and remove

sequenceDiagram
    actor user
    participant cqspi_probe
    participant cqspi_st
    participant cqspi_exec_mem_op
    participant cqspi_indirect_read_execute
    participant cqspi_indirect_write_execute
    participant cqspi_remove

    user ->> cqspi_probe: probe()
    cqspi_probe ->> cqspi_st: refcount_set(refcount, 1)
    cqspi_probe ->> cqspi_st: refcount_set(inflight_ops, 1)

    user ->> cqspi_exec_mem_op: exec_mem_op(mem, op)
    cqspi_exec_mem_op ->> cqspi_st: refcount_read(inflight_ops)
    cqspi_exec_mem_op ->> cqspi_st: refcount_read(refcount)
    alt refcount == 0
        cqspi_exec_mem_op -->> user: return -EBUSY
    else refcount > 0
        cqspi_exec_mem_op ->> cqspi_st: refcount_inc(inflight_ops)
        cqspi_exec_mem_op ->> cqspi_st: refcount_read(refcount)
        alt refcount dropped to 0
            cqspi_exec_mem_op ->> cqspi_st: refcount_dec(inflight_ops)
            cqspi_exec_mem_op -->> user: return -EBUSY
        else refcount still > 0
            alt read operation
                cqspi_exec_mem_op ->> cqspi_indirect_read_execute: cqspi_indirect_read_execute()
                cqspi_indirect_read_execute ->> cqspi_st: refcount_read(refcount)
                alt refcount == 0
                    cqspi_indirect_read_execute -->> cqspi_exec_mem_op: -ENODEV
                else
                    cqspi_indirect_read_execute -->> cqspi_exec_mem_op: ret
                end
            else write operation
                cqspi_exec_mem_op ->> cqspi_indirect_write_execute: cqspi_indirect_write_execute()
                cqspi_indirect_write_execute ->> cqspi_st: refcount_read(refcount)
                alt refcount == 0
                    cqspi_indirect_write_execute -->> cqspi_exec_mem_op: -ENODEV
                else
                    cqspi_indirect_write_execute -->> cqspi_exec_mem_op: ret
                end
            end
            cqspi_exec_mem_op ->> cqspi_st: refcount_read(inflight_ops)
            alt inflight_ops > 1
                cqspi_exec_mem_op ->> cqspi_st: refcount_dec(inflight_ops)
            end
            cqspi_exec_mem_op -->> user: return ret
        end
    end

    user ->> cqspi_remove: remove(pdev)
    cqspi_remove ->> cqspi_st: refcount_set(refcount, 0)
    cqspi_remove ->> cqspi_st: refcount_dec_and_test(inflight_ops)
    alt inflight_ops == 0
        cqspi_remove ->> cqspi_remove: skip cqspi_wait_idle()
    else inflight_ops > 0
        cqspi_remove ->> cqspi_st: cqspi_wait_idle(cqspi)
    end
    cqspi_remove -->> user: unregister and disable controller

Loading

Class diagram for cadence-quadspi driver refcounting and related operations

classDiagram
    class cqspi_st {
        bool apb_ahb_hazard
        bool is_jh7110
        refcount_t refcount
        refcount_t inflight_ops
    }

    class cqspi_driver_platdata {
        <<struct>>
    }

    class spi_mem {
        <<struct>>
    }

    class cqspi_driver {
        <<module>>
        +int cqspi_probe(platform_device *pdev)
        +void cqspi_remove(platform_device *pdev)
        +int cqspi_exec_mem_op(spi_mem *mem, spi_mem_op const *op)
        +int cqspi_mem_process(spi_mem *mem, spi_mem_op const *op)
        +int cqspi_indirect_read_execute(cqspi_flash_pdata *f_pdata, u8 *rxbuf, u32 from_addr, size_t n_rx)
        +int cqspi_indirect_write_execute(cqspi_flash_pdata *f_pdata, const u8 *txbuf, u32 to_addr, size_t n_tx)
    }

    cqspi_driver *-- cqspi_st : owns
    cqspi_driver_platdata --> cqspi_st : configures
    cqspi_driver --> spi_mem : executes_ops_on
    cqspi_exec_mem_op ..> cqspi_indirect_read_execute : may_call
    cqspi_exec_mem_op ..> cqspi_indirect_write_execute : may_call
    cqspi_probe ..> cqspi_st : refcount_set(refcount,1)\nrefcount_set(inflight_ops,1)
    cqspi_remove ..> cqspi_st : refcount_set(refcount,0)\nrefcount_dec_and_test(inflight_ops)

Loading

File-Level Changes

Change	Details	Files
Update kernel version to 6.6.125 and sync with upstream release	Bump SUBLEVEL from 124 to 125 in top-level Makefile Implicitly pull in all upstream 6.6.125 commits listed in the PR description	Makefile
Harden netfilter nf_tables pipapo lookup table sizing and accounting	Introduce lt_calculate_size() helper to compute lookup-table sizes with overflow checking Use lt_calculate_size() from pipapo_resize(), pipapo_lt_bits_adjust(), and pipapo_clone() instead of open-coded arithmetic Switch kvzalloc/kmalloc/kzalloc/kvmalloc users in pipapo paths to GFP_KERNEL_ACCOUNT and reuse lt_size where appropriate Ensure scratch map allocations also use GFP_KERNEL_ACCOUNT	net/netfilter/nft_set_pipapo.c
Tighten memory accounting and allocations in various nftables helpers	Allocate nft_compat match private info, expression ops, target ops, nf_tables catchall sets, nft_log prefix, nft_secmark context, and nft_numgen counter with GFP_KERNEL_ACCOUNT Use metadata_dst_alloc and dst_cache_init with GFP_KERNEL_ACCOUNT in nft_tunnel object init	net/netfilter/nft_compat.c net/netfilter/nf_tables_api.c net/netfilter/nft_log.c net/netfilter/nft_meta.c net/netfilter/nft_numgen.c net/netfilter/nft_tunnel.c
Fix MPTCP netlink PM address flushing and extend selftests to cover unknown flags	Rework mptcp_nl_cmd_flush_addrs to snapshot local_addr_list, re-initialize it under lock, and safely free entries after synchronize_rcu Guard against flushing an empty list and correctly relink free_list before deletion Add MPTCP_PM_ADDR_FLAG_UNKNOWN constant and support an "unknown" flag in pm_nl_ctl add_addr and dump output Extend pm_netlink.sh selftest to verify unknown flags are ignored while preserving output format	net/mptcp/pm_netlink.c tools/testing/selftests/net/mptcp/pm_nl_ctl.c tools/testing/selftests/net/mptcp/pm_netlink.sh
Refine vsock selftest utilities to verify setsockopt/getsockopt symmetry and enable SO_ZEROCOPY helper	Include string.h and add generic helpers to set and then read back socket options for unsigned long long, int, and struct timeval types with size/value validation and error reporting Add enable_so_zerocopy_check() to enable SO_ZEROCOPY via the int helper Replace direct setsockopt calls in vsock_test and control.c with the new *_check helpers for buffer sizes, timeouts, RCVLOWAT, and SO_REUSEADDR	tools/testing/vsock/util.c tools/testing/vsock/util.h tools/testing/vsock/vsock_test.c tools/testing/vsock/control.c
Fix QLogic qla2xxx fabric scan, login state handling, and async logout handling	Ensure async fabric scans free sp and drop reference even when a scan is already active, guarding null sp and cleaning DMA buffers before kref_put Remove special-case avoidance of FCP2 device deletion during fabric scan and FC2-target handling in RSCN, so sessions can be properly torn down Improve plogi-done handling by re-checking FW login state when DSC_LS_PLOGI_PEND, forcing state to DSC_LS_PRLI_COMP and posting GPDB work outside N2N topology Enhance async event logout processing to handle missing fcport, non-online states, and schedule session deletion, adding debug logs Make HBA-wait path also block while SF_SCANNING is set to avoid module unload/reset races	drivers/scsi/qla2xxx/qla_gs.c drivers/scsi/qla2xxx/qla_init.c drivers/scsi/qla2xxx/qla_isr.c drivers/scsi/qla2xxx/qla_os.c
Prevent races and refcount issues in Cadence QuadSPI during unbind/removal	Add refcount and inflight_ops fields to cqspi state structure Gate indirect read/write execution and mem ops on refcount/inflight_ops, returning -ENODEV/-EBUSY when device is going away Increment inflight_ops on exec, decrement on completion, and wait for idle in remove() if there are outstanding operations Initialize refcount and inflight_ops in probe() and clear refcount on remove()	drivers/spi/spi-cadence-quadspi.c
Ensure proper ordering and single registration of OMAP GPIO and MPU I/O drivers	Change omap_mpuio_init to only register the platform device once using a static registered flag and no longer register the driver there Register omap_mpuio_driver and omap_gpio_driver together in an initcall, with rollback if the second registration fails Unregister omap_mpuio_driver alongside omap_gpio_driver in module exit	drivers/gpio/gpio-omap.c
Enforce device locking when matching drivers to devices in the driver core	Add driver_match_device_locked() wrapper that guards driver_match_device() with device_lock via guard(device) and asserts the lock in driver_match_device() Use driver_match_device_locked() from bus bind_store() and __driver_attach() instead of calling driver_match_device() directly	drivers/base/base.h drivers/base/bus.c drivers/base/dd.c
Fix nfsd per-net statistics proc entry initialization and error handling	Change nfsd_proc_stat_init() to return the proc_dir_entry * from svc_proc_register In nfsd_net_init, treat failure of nfsd_proc_stat_init as -ENOMEM and properly unwind via new out_proc_error path Update header declaration accordingly	fs/nfsd/nfsctl.c fs/nfsd/stats.c fs/nfsd/stats.h
Prevent NILFS2 fstrim from overflowing below first data block	Clamp nilfs_sufile_trim_fs() so that if the computed end_block is before ns_first_data_block it exits early without trimming Ensure range->len is still updated based on discarded segments before returning	fs/nilfs2/sufile.c
Fix SMB client bitfield races and ksmbd server error handling and connection accounting	Replace bitfield booleans in cached_fid with plain bools to avoid shared-byte RMW races on multi-bit bitfields Change ksmbd server request handler to return SERVER_HANDLER_ABORT instead of CONTINUE on invalid/unsupported commands or signature failures Ensure ksmbd_tcp_new_connection cleans up via ksmbd_tcp_disconnect() instead of free_transport() on thread creation failure Fix ksmbd_tcp_new_connection leak of active_num_conn counters via updated error path (pulled from upstream patch)	fs/smb/client/cached_dir.h fs/smb/server/server.c fs/smb/server/transport_tcp.c
Fix virtio-crypto completion callback concurrency and remove redundant queue kick	Wrap virtcrypto_done_task virtqueue processing in a spinlock, dropping the lock while invoking the algorithm callback and re-acquiring afterward, to serialize access to the virtqueue and data_vq Remove an unnecessary virtqueue_kick() from virtio_crypto_skcipher_crypt_req now that queue notification is handled elsewhere Add a spinlock-protected virtqueue notification path to avoid races with the callback handler	drivers/crypto/virtio/virtio_crypto_core.c drivers/crypto/virtio/virtio_crypto_skcipher_algs.c
Fix scatterlist allocation bug in omap-crypto and prevent ucode name truncation in octeontx	Allocate scatterlist array with sizeof(*new_sg) instead of sizeof(*sg) in omap_crypto_copy_sg_lists Change ucode_load_store to validate input length using count against OTX_CPT_UCODE_NAME_LENGTH to avoid off-by-one issues with strlen	drivers/crypto/omap-crypto.c drivers/crypto/marvell/octeontx/otx_cptpf_ucode.c
Miscellaneous driver and net fixes (rtw88 beacon, SFP quirk, Bluetooth ID, MPTCP test, vsock control, nf_tables memcg, etc.)	Use 8-bit TXPAUSE register accessors in rtw_core_enable_beacon to avoid alignment faults Extend sfp_quirk_ubnt_uf_instant to also restrict supported PHY interface modes to 1000BASEX Add new Edimax EW-7611UXB (7392:e611) USB ID to btusb quirks table as a Realtek device with wideband speech Ensure nf_tables objects without memcg accounting are not skipped (pulled from upstream patch) Adjust selftests for vsock and MPTCP to reflect new helpers and behavior	drivers/net/wireless/realtek/rtw88/main.c drivers/net/phy/sfp.c drivers/bluetooth/btusb.c net/netfilter/nf_tables_api.c tools/testing/vsock/control.c tools/testing/vsock/vsock_test.c tools/testing/vsock/util.c tools/testing/selftests/net/mptcp/pm_netlink.sh

---

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.

[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.

Details

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

[sourcery-ai]

Hey - I've reviewed your changes and they look great!

---

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 updates the Linux kernel from version 6.6.124 to 6.6.126, incorporating upstream stable fixes from the 6.6.125 and 6.6.126 releases. The update includes 26 commits addressing various bug fixes, security improvements, and driver enhancements across multiple subsystems including networking (MPTCP, netfilter), filesystems (SMB, NFSD, NILFS2), drivers (SCSI, SPI, GPIO, crypto, wireless, Bluetooth), and testing infrastructure (vsock, MPTCP selftests).

Changes:

• Integrated upstream stable fixes for kernel 6.6.125 and 6.6.126
• Fixed race conditions and memory accounting issues in netfilter subsystem with GFP_KERNEL_ACCOUNT and overflow checking
• Improved error handling in SMB server, NFSD, and SCSI qla2xxx driver
• Enhanced test coverage with vsock socket option verification helpers and MPTCP unknown flag handling
• Added hardware support for new Bluetooth devices and fixed various driver bugs

Reviewed changes

Copilot reviewed 35 out of 35 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
Makefile	Version bump from 6.6.124 to 6.6.126
tools/testing/vsock/util.c	Added socket option verification helpers with get-after-set validation
tools/testing/vsock/util.h	Declared new helper function prototypes
tools/testing/vsock/vsock_test.c	Refactored to use new setsockopt helper functions
tools/testing/vsock/control.c	Updated to use setsockopt_int_check helper
tools/testing/selftests/net/mptcp/pm_nl_ctl.c	Added support for unknown address flags in testing
tools/testing/selftests/net/mptcp/pm_netlink.sh	Added test case for unknown flag handling
net/mptcp/pm_netlink.c	Fixed race condition in address flush operation using proper list detachment
net/netfilter/nft_tunnel.c	Changed to GFP_KERNEL_ACCOUNT for memcg accounting
net/netfilter/nft_set_pipapo.c	Added overflow checking for lookup table allocation and GFP_KERNEL_ACCOUNT
net/netfilter/nft_numgen.c	Changed to GFP_KERNEL_ACCOUNT
net/netfilter/nft_meta.c	Changed to GFP_KERNEL_ACCOUNT
net/netfilter/nft_log.c	Changed to GFP_KERNEL_ACCOUNT
net/netfilter/nft_compat.c	Changed to GFP_KERNEL_ACCOUNT
net/netfilter/nf_tables_api.c	Changed to GFP_KERNEL_ACCOUNT
fs/smb/server/transport_tcp.c	Fixed connection leak by using proper disconnect function
fs/smb/server/server.c	Changed error handling to abort instead of continue on auth/validation failures
fs/smb/client/cached_dir.h	Split bitfields to separate booleans to avoid RMW races
fs/nilfs2/sufile.c	Added early exit to prevent block overflow in trim operation
fs/nfsd/stats.h	Changed nfsd_proc_stat_init to return proc_dir_entry pointer
fs/nfsd/stats.c	Updated to return svc_proc_register result
fs/nfsd/nfsctl.c	Added error checking for proc stat initialization
drivers/spi/spi-cadence-quadspi.c	Implemented refcount mechanism to handle unbind during operations
drivers/scsi/qla2xxx/qla_os.c	Added fabric scan flag check to module unload wait condition
drivers/scsi/qla2xxx/qla_isr.c	Enhanced port logout event handling with additional state checks
drivers/scsi/qla2xxx/qla_init.c	Modified PLOGI handling to query FW state and removed FCP2 special case
drivers/scsi/qla2xxx/qla_gs.c	Fixed fabric scan error path to free sp when already scanning
drivers/net/wireless/realtek/rtw88/main.c	Fixed alignment fault by using write8 instead of write32 for register access
drivers/net/phy/sfp.c	Added missing interface mode setting for Ubiquiti quirk
drivers/gpio/gpio-omap.c	Fixed driver registration order to prevent probe-time registration
drivers/crypto/virtio/virtio_crypto_skcipher_algs.c	Removed duplicate virtqueue_kick call
drivers/crypto/virtio/virtio_crypto_core.c	Added spinlock protection for virtqueue completion handling
drivers/crypto/omap-crypto.c	Fixed scatterlist allocation to use correct sizeof operand
drivers/crypto/marvell/octeontx/otx_cptpf_ucode.c	Fixed buffer length check to use count instead of strlen
drivers/bluetooth/btusb.c	Added USB device ID for Edimax EW-7611UXB adapter

---

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

[Copilot]

There is an extraneous semicolon after the closing brace. This creates an empty statement and should be removed.

[Copilot]

Missing space before the newline escape character. The comment should have a space between "State:0x%x" and "\n" for consistency with standard formatting.

[Copilot]

Missing space before the newline escape character. The comment should have a space between "deletion" and "\n" for consistency with standard formatting.

Suggested change

	"Scheduling session for deletion \n");
	"Scheduling session for deletion\n");

[Copilot]

The refcount_dec operation should be unconditional, not conditional on refcount > 1. Since refcount_inc was called unconditionally at line 1431, the decrement should also be unconditional. The current logic could cause the inflight_ops counter to become unbalanced, leading to resource leaks or preventing future operations. The check at line 1443 should be removed, and refcount_dec should always be called to match the increment.

Suggested change

	if (refcount_read(&cqspi->inflight_ops) > 1)
	refcount_dec(&cqspi->inflight_ops);
	refcount_dec(&cqspi->inflight_ops);

[opsiff]

refcount_set(&cqspi->inflight_ops, 1); will be run in cqspi_probe, need to be keep it.

[Copilot]

The refcount_dec_and_test logic appears inverted. Since inflight_ops is initialized to 1 and decremented here, the function should wait for idle when the test returns true (i.e., when the counter reaches 0), not when it returns false. The current logic will wait when there are still operations in flight but not wait when the counter has reached zero, which is backwards.

Suggested change

	if (!refcount_dec_and_test(&cqspi->inflight_ops))
	if (refcount_dec_and_test(&cqspi->inflight_ops))

[opsiff]

refcount_set(&cqspi->inflight_ops, 1); will be run in cqspi_probe, and if not wait "when the counter has reached zero" is not probed, so not need to cqspi_wait_idle