| Age | Commit message (Collapse) | Author |
|
IEEE 754 lets an architecture determine tininess of a floating-point
result either before or after rounding, but requires the same choice for
every operation. Alpha determines it after rounding, and stdlib's
tst-tininess confirms the hardware does so for the results it produces
itself.
The soft-fp emulation has no notion of the distinction and always
determines tininess before rounding, so a result that the hardware would
not consider tiny is reported as underflowing whenever the instruction
happens to trap for software completion. The two paths then disagree on
the same machine. A multiply of the largest subnormal double by
1 + 2^-52 rounds up to the smallest normal and raises no underflow when
the operands are normal, but raises it when an operand is subnormal and
the instruction traps:
glibc math testsuite, test-float32x-float64-mul:
Failure: mul_double (0x3.ffffffffffffcp-1024, 0x1.0000000000001p+0):
Exception "Underflow" set
Add _FP_TININESS_AFTER_ROUNDING, as glibc's copy of soft-fp has, and set
it for alpha. It determines tininess by rounding a copy of the result as
if the exponent range were unbounded, which is what the definition asks
for. A plain check of whether the rounded result came out normal is not
equivalent: it differs for values that are still tiny once rounded with
an unbounded exponent range but reach the smallest normal on the coarser
subnormal grid. That plain check is what the hardware itself does (it
decides tininess from the delivered result), and it is why a divide such
as -0x4p-128 / 0x1.000002p+0 in FE_DOWNWARD delivers -0x1p-126 and raises
nothing although the result is tiny after rounding. Such cases do not
trap, so the emulation never sees them; where it does get control, it
follows the definition rather than reproducing that deviation, which is
what glibc's soft-fp does as well.
The macro defaults to zero, so powerpc, sh and sparc keep determining
tininess before rounding as they do now.
Cc: stable@vger.kernel.org # 5.15+
Reviewed-by: Magnus Lindholm <linmag7@gmail.com>
Tested-by: Magnus Lindholm <linmag7@gmail.com>
Link: https://lore.kernel.org/r/20260803-alpha-fp-exceptions-v1-3-c99d75608e60@gmail.com
Signed-off-by: Matt Turner <mattst88@gmail.com>
|
|
Carlo Szelinsky says:
====================
net: pse-pd: si3474: probe error path cleanups
Two small changes to the error paths in si3474 probe. They do not
depend on the PSE fixes for net ("net: pse-pd: assorted lifetime and
error-path fixes" v2) and apply to net-next as they are.
1 The error message after a failed devm_add_action_or_reset() reads
the ancillary I2C client, which was just unregistered.
2 Use dev_err_probe() for a failed controller registration, as
pd692x0 and tps23881 do.
1 was found by code inspection and not reproduced. Both are
build-tested only, with W=1.
[1] https://lore.kernel.org/20261018082314.48950-1-github@szelinsky.de
[2] https://lore.kernel.org/20261018081853.47484-1-github@szelinsky.de
====================
Link: https://patch.msgid.link/20261020171134.852878-1-github@szelinsky.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
si3474 prints the error of a failed controller registration as a hex
number, such as 0xfffffff4 for -ENOMEM. Use dev_err_probe(), as
pd692x0 and tps23881 already do. It prints the error by name, and if
the registration ever returns -EPROBE_DEFER, it does not log that as
an error.
Only the log message changes.
Signed-off-by: Carlo Szelinsky <github@szelinsky.de>
Link: https://patch.msgid.link/20261020171134.852878-3-github@szelinsky.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
If devm_add_action_or_reset() fails, it has already unregistered the
ancillary I2C client, and the error message then reads that client's
device. It can only fail with -ENOMEM, which is already reported, so
drop the message.
Found by code inspection. Not reproduced.
Fixes: a2317231df4b ("net: pse-pd: Add Si3474 PSE controller driver")
Reviewed-by: Kory Maincent <kory.maincent@bootlin.com>
Signed-off-by: Carlo Szelinsky <github@szelinsky.de>
Link: https://patch.msgid.link/20261020171134.852878-2-github@szelinsky.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
My old email address doesn't work anymore. Update it to the current
working one.
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://patch.msgid.link/20261020225358.1130051-1-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Setting FILE_ATTRIBUTE_TEMPORARY on a directory with SMB2 SET_INFO
(FileBasicInformation) succeeds: ksmbd keeps the attribute and, with
"store dos attributes" enabled, also writes it to the DOS attribute
xattr.
MS-FSA (FileBasicInformation) says the request must fail with
STATUS_INVALID_PARAMETER when FILE_ATTRIBUTE_TEMPORARY is set on a
directory, and Samba rejects it too. smbtorture's
smb2.create.dosattr_tmp_dir checks this and fails against ksmbd.
Return -EINVAL, which is sent as STATUS_INVALID_PARAMETER, when a
directory is asked to become TEMPORARY. Do the check before any
timestamp is changed, so a rejected request leaves the open file and
the inode unchanged.
Fixes: e2f34481b24d ("cifsd: add server-side procedures for SMB3")
Assisted-by: LLM
Signed-off-by: Michail Andreas Michalopoulos <mike.michalo@gmail.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
smb_check_perm_dacl() returns the access mask it computes to the SMB2
CREATE handler, which uses it as the granted access for the open. For a
security descriptor whose DACL is present but contains no ACEs, the
function is supposed to deny access: an empty (non-NULL) DACL grants no
access to anyone except the object owner's implicit READ_CONTROL and
WRITE_DAC (MS-DTYP 2.4.5).
The empty-DACL branch only denies when both
- the DACL has no trailing bytes (pdacl_size == sizeof(struct smb_acl)),
and
- the request contains bits beyond READ_CONTROL / WRITE_DAC.
In every other case it takes the shared "goto err_out" with rc still 0,
which returns success without modifying *pdaccess, so the caller grants
the full requested access. Three problems follow:
- The READ_CONTROL / WRITE_DAC carve-out models the owner's implicit
rights, but the check runs before is_owner is computed, so those
rights are handed to *every* caller, not just the owner. For
READ_CONTROL the SMB2 CREATE path does not fall back to
inode_permission(), so any user can read the security descriptor of
an object whose empty DACL is meant to deny them.
- A request of exactly READ_CONTROL and/or WRITE_DAC is granted to any
caller. WRITE_DAC lets the handle rewrite the DACL.
- A DACL that declares zero ACEs but carries trailing bytes
(pdacl_size > sizeof(struct smb_acl)) skips the deny entirely and is
granted whatever was requested.
On shares where the NT ACL is the authoritative access control -- e.g.
"force user" shares, where every client maps to one POSIX identity so
POSIX cannot distinguish users -- this is an access-control bypass.
Compute is_owner before the empty-DACL branch and deny access from an
empty DACL unless the caller is the owner requesting no more than
READ_CONTROL / WRITE_DAC.
Fixes: e2f34481b24d ("cifsd: add server-side procedures for SMB3")
Cc: stable@vger.kernel.org
Assisted-by: LLM
Signed-off-by: Aamir Ahmed <elb12345@hotmail.co.uk>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
https://gitlab.freedesktop.org/drm/misc/kernel into drm-next
drm-misc-next for v7.5:
UAPI Changes:
- Add Apple Interchange fourcc.
- Expose amdxdna busy time via DRM fdinfo.
Cross-subsystem Changes:
- Improve DT bindings for display/ste,mcde
- Add kunit test for dma_fence_dedup_array() to dma-buf/test.
Core Changes:
- Documentation updates to drm/scheduler, gpu/buddy.
- Fail kmapping a bo without pages.
Driver Changes:
- Bugfixes and improvements to panel/Raystar RFF500F-AWH-DNN,
accel/ivpu, vc4, accel/amdxdna, panthor, nouveau, panfrost,
panthor, accel/qaic, accel/rocket, bridge/lontium-lt8713sx,
appletbdrm.
- Add support for BOE Novatek NT51021, Retroid Pocket 6,
Raydium RM69220 DDIC panels, Sharp LQ138P1JX61 eDP,
BOE NV140FHD-N43, AUO B140HAN04.0, AUO 0x0db3,
panels.
- Add support for Mali-G68 to panfrost.
- Add Chrontel CH7218A to bridge/simple.
- Support ARGB8888 for virtio primary planes.
- Improve panfrost gpu reset handling and initialization.
- Add HDMI 2.0 support to rockchip/dw-hdmi-qp and improve HPD handling.
- Advertise HDMI 1.2 capabilities in sun4i, bridge/adv7511, it66121 & inno-hdmi.
- Advertise HDMI 1.3 capabilities in bridge/ite-6263.
- Advertise HDMI 1.4 capabilities in bridge/lt9611 & rockchip/rk3066_hdmi
- Switch many drivers to drm_atomic_helper_create_state in preparation
for readout.
Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: https://patch.msgid.link/75e7de61-048c-4fdd-9ad2-6b47b29a4863@linux.intel.com
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/chrome-platform/linux
Pull chrome platform firmware loader updates from Tzung-Bi Shih:
"New:
- Rename drivers/firmware/google/ -> drivers/firmware/coreboot/ and
corresponding Kconfig names
- Separate firmware_attributes_class from drivers/platform/x86/ to
drivers/firmware/ so that it is available for non-x86 platforms
- Add CFR (coreboot form representation) driver to expose the
attributes via sysfs
Improvements:
- Use named initializers for acpi_device_id"
* tag 'chrome-platform-firmware-v7.4' of git://git.kernel.org/pub/scm/linux/kernel/git/chrome-platform/linux:
firmware: coreboot: Add CFR firmware attributes driver
firmware: Move firmware attributes class helper
firmware: coreboot: Use named initializers for acpi_device_id
firmware: Rename google firmware directory to coreboot
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/chrome-platform/linux
Pull chrome platform updates from Tzung-Bi Shih:
"Improvements:
- Poll for role swap completion in cros_ec_typec
Fixes:
- Fix wrong returned data and length for deferred response in
transport protocol
- Check boundary for EC-returned data length in cros_ec_typec
- Fix a regression for older ChromeOS devices after applying
db65a06d10b3 ("ACPI: EC: Convert the driver to a platform one")
Cleanups:
- Fix typos"
* tag 'chrome-platform-v7.4' of git://git.kernel.org/pub/scm/linux/kernel/git/chrome-platform/linux:
platform/chrome: cros_ec_typec: Poll for role swap completion
platform/chrome: cros_usbpd_notify: Only use drvdata when parent is GOOG0004
platform/chrome: cros_ec_typec: Validate SVID and mode counts in discovery data
platform/chrome: cros_ec_rpmsg: Fix repeated word 'from' in comment
platform/chrome: Fix typos in comments
platform/chrome: cros_ec_proto: Simplify cros_ec_wait_until_complete()
platform/chrome: cros_ec_proto: Fix deferred response
platform/chrome: cros_ec: Sync host command header with EC
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu
Pull two-byte cmpxchg emulation from Paul McKenney:
"These commits add a cmpxchg_emu_u16() function that emulates two-byte
cmpxchg() in terms of four-byte cmpxchg() for architectures not having
native instructions for this purpose. They also make use of this
cmpxchg_emu_u16() function in those architectures, namely sh, arc,
csky, and xtensa. This permits use of two-byte cmpxchg() in common
code.
However, it sadly does not permit easy implementation of all
algorithms needing on two-byte atomic operations because there are
some systems still supported by Linux that do not have native two-byte
store instructions.
There is reason to believe that these systems will soon no longer be
supported by Linux [1]"
Link: https://lore.kernel.org/all/20260908152808.3928630-4-arnd@kernel.org/ [1]
* tag 'cmpxchg-emu.2026.10.12b' of git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu:
xtensa: Emulate two-byte cmpxchg
csky: Emulate two-byte cmpxchg
ARC: Emulate two-byte cmpxchg
sh: Emulate two-byte cmpxchg
lib: Add two-byte cmpxchg emulation function
|
|
Pull documentation updates from Jonathan Corbet:
"It was a reasonably busy cycle for docs; changes include:
- Add the AGENTS.md symbolic link into the top-level directory in the
hopes that LLMs will read it and produce better work
- The removal of the Assisted-by requirement, as concluded at the
Maintainer Summit
- A new "testdocs" make target to run the documentation unit tests
- Lots of work on the Portuguese, and Chinese translations; there is
now maintenance for the traditional Chinese translations as well
...and the usual pile of minor fixes, documentation updates, etc"
* tag 'docs-7.4' of git://git.kernel.org/pub/scm/linux/kernel/git/docs/linux: (125 commits)
docs: add AGENTS.md as a symlink to README
docs: kernel-parameters: adjust reserve_mem build option
docs: kernel-parameters: don't mix cmdline options & build options
docs/ja_JP: translate submitting-patches.rst (review and report tags)
docs: remove Assisted-by references
Revert "docs: landlock: fix ENOMSG condition in create_ruleset kernel-doc"
docs: kdoc: parse context_lock_struct() as struct declaration
docs: timers: hpet: Document userspace API and ioctl commands
docs: timers: hpet: Modernize file references
Documentation: sched-preemption: Add a SPDX license identifier
docs: ABI: sysfs-driver-xdata: use literal blocks for console commands
docs: ABI: sysfs-class-firmware-attributes: use literal blocks for console commands
ABI: sysfs-bus-usb: use literal block for console commands
docs/ABI: testing: aspeed-uart-routing: use literal block for console commands
docs: kernel-parameters: fix a truncated sentence for preempt=lazy
sched/doc: add a preemption model overview
docs: ABI: TMR manager: use literal blocks for console I/O
Documentation/ABI: IFS: use literal blocks for commands
ABI: sysfs-bus-pci: use literal blocks for commands
docs: translations: pt_BR: add debugging index
...
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/nolibc/linux-nolibc
Pull nolibc updates from Thomas Weißschuh:
- New architectures: Hexagon, 64-bit parisc
- New library functionality: sendfile(), pread(), pwrite()
- Various bugfixes and cleanups
* tag 'nolibc-20261018-for-7.4-1' of git://git.kernel.org/pub/scm/linux/kernel/git/nolibc/linux-nolibc: (28 commits)
selftests/nolibc: explicitly disable position independent executables
selftests/nolibc: use the configured kernel header directory
tools/nolibc: fix the function argument order of calloc()
tools/nolibc: check for overflow in malloc()
tools/nolibc: stop rounding malloc() sizes up to 4096 bytes
tools/nolibc: add support for pread() and pwrite()
tools/nolibc: introduce __NOLIBC_BITS_PER_SYSCALL_ARG
tools/nolibc: move __NOLIBC_LLARGPART() to compiler.h
tools/nolibc: make 64-bit systemcall argument padding generic
selftests/nolibc: make the broken MIPS N32 argument detection reusable
tools/nolibc: add 64-bit parisc support
selftests/nolibc: test the FD_* macros
tools/nolibc: fix FD_SET(), FD_CLR() and FD_ISSET() on 64-bit
tools/nolibc: fix readdir_r() with 64-bit directory offsets
tools/nolibc: fix verrx() and errx() message formatting
selftests/nolibc: add tests for the wait status macros
tools/nolibc: fix WIFSIGNALED() for status 0
tools/nolibc/parisc: consistently use %dp to refer to register %r27
selftests/nolibc: validate ENOTDIR return values from opendir()/fdopendir()
tools/nolibc: verify that a directory is opened
...
|
|
vxlan_vnifilter_process() parses the message with vni_filter_policy,
whose VXLAN_VNIFILTER_ENTRY is a bare NLA_NESTED with no nested policy
attached. The entry attributes are therefore validated only later, one
entry at a time, by the nla_parse_nested() inside
vxlan_process_vni_filter(). Entries are parsed as they are dispatched:
in a message whose first entry is valid and whose second is not, the
first entry is applied and its RTM_NEWTUNNEL notification sent before
the second is rejected.
Link the nest to vni_filter_entry_policy with NLA_POLICY_NESTED() so
every entry is checked against the entry policy up front, before any
entry is acted on. A message whose entry the policy rejects -- a
malformed attribute, or, with the next patch, an out-of-range endpoint
-- is now refused as a unit and installs nothing. Faults the policy does
not cover, such as a device or handler error raised on a later entry,
are unchanged: entries are still dispatched and applied one at a time.
This reorders two faults. The nest is validated before the device is
looked up and before the vnifilter flag is checked, so a request
rejected by the entry policy, aimed at a missing or non-vnifilter
device, now returns that policy error where it previously returned
-ENODEV or -EOPNOTSUPP. The request was invalid either way; only the
errno an operator sees changes.
The nla_parse_nested() in vxlan_process_vni_filter() stays: it is what
fills the per-entry attribute table the handler reads. It can no longer
fail on a message that has reached it.
Suggested-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/netdev/20260921150904.65a704eb@kernel.org/
Signed-off-by: Ali Firas <alishmery18@gmail.com>
Link: https://patch.msgid.link/20261014123814.388557-2-alishmery18@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip
Pull xen updates from Juergen Gross:
"This time only small fixes and cleanups all over Xen related drivers.
Nothing really stands out here"
* tag 'for-linus-7.4-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
xen-blkfront: Fix IO race during unplug
xen-pciback: Fix pcistub_device ref leak in pcistub_get_gsi_from_sbdf()
xen/pcifront: check that an AER callback exists before calling it
xen/gntdev: Fix gntdev_dmabuf ref leak in dmabuf_exp_wait_released()
x86/PVH: don't truncate initrd address/size
xen/privcmd: Only claim ioreqs matching an ioeventfd
xen/scsifront: check for a NULL shadow entry on a backend response
xen: fix typos in comments
xen/gntdev: prevent private mappings from becoming writable
xen: remove unused hvm_vcpu.h header
xen-blkfront: unbind irq before tearing down ring and shadow requests
xen/blkback: Prevent missed completion when draining I/O
xen/pcifront: Fix PCI device reference leak in AER handling
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/lindholm/alpha
Pull alpha updates from Magnus Lindholm:
"This completes Alpha's conversion to GENERIC_ENTRY and enables a range
of kernel features through Kconfig updates, including UBSAN, gcov,
kmemleak, page-table checking, fast GUP and sparse IRQs.
It includes TLB and MMU context fixes prompted by glibc test-suite
failures, HugeTLB support for 64KB, 512KB and 4MB pages, and fixes for
poisoned-page faults and page-table checking.
Machine-check handling is improved on Tsunami platforms, with bounded
EV6 logout decoding, Clipper environmental event reporting and safer
firmware console-log scanning.
Other changes add self-extracting compressed kernels with support for
all kernel compression formats, Alpha-specific XZ tuning, perf
register and user-stack sampling, and Alpha support in the seccomp
selftests. The series also fixes user stack-pointer reporting and IRQ
entry ordering, adds 128-bit shift helpers, cleans up linker sections,
and updates Kconfig help"
* tag 'alpha-for-v7.4-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/lindholm/alpha: (57 commits)
alpha: add bounded Clipper console-log scanning
alpha: decode Clipper environmental events
alpha: decode and acknowledge Tsunami system machine checks
alpha: bound EV6 logout decoding by the processor area
alpha: describe the Tsunami system machine check frame
alpha: move Tsunami machine check handling to err_tsunami.c
alpha: fill in pc for software event samples
alpha: add HAVE_PERF_REGS and HAVE_PERF_USER_STACK_DUMP support
alpha: use user_stack_pointer() in KSTK_ESP() and dump_elf_thread()
alpha: stop storing the user stack pointer in pt_regs
alpha: fix stale user_stack_pointer()
selftests/seccomp: add Alpha support to seccomp_bpf
alpha: do not skip TLB shootdown IPIs for kthread-borrowed mms
alpha: do not clear remote MMU contexts in migrate_flush_tlb_page()
alpha: invalidate the local context in flush_icache_user_page()
alpha: invalidate the local context in flush_tlb_mm()
alpha: fix the local TLB invalidate in the UP flush_tlb_page()
alpha: fix the local TLB invalidate in flush_tlb_page()
alpha: invalidate the local context in flush_tlb_page()
alpha: only use a targeted tbi() when the target mm is really current
...
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux
Pull s390 updates from Heiko Carstens:
- Add support for sclp-vt220 terminal resize events e.g. for QEMU
- Warn if kernel command line contains non-printable EBCDIC characters
- Improve show_code() disassembly output by using reliable markers for
the instruction address and faulting instruction
- Rework exception handling to require only a single exception table
entry per faulting instruction: convert all inline assemblies from
EX_TABLE to the new EX_TABLE_INSN macros. The previous exception
handling was error prone since it usually required two EX_TABLE
entries: one pointing at the instruction, and one pointing behind it
- Disallow kprobes on instructions with exception table entries to
avoid exception handlers decoding a breakpoint instead of the
original instruction
- Preemptible percpu ops changes:
- Simplify percpu_entry() and percpu_exit() by encoding the percpu
base, offset, and pointer registers into the lowcore marker,
allowing percpu_exit() to update the percpu pointer on CPU
migration without comparing the interrupted instruction
- Convert this_cpu_cmpxchg(), this_cpu_xchg(), and
this_cpu_cmpxchg128() to preemptible percpu ops; provide
this_cpu_try_cmpxchg()
- Emulate virtual timer in appldata via delayed work and remove the
s390 specific vtimer infrastructure
- Add lock contention tracepoints to rwlock slowpath
- Fix incorrect return address in signal frame setup
- Fix cpum_cf CPU hotplug handling
- Convert various __get_free_pages() / get_zeroed_page() users to
kmalloc() / kzalloc()
- PCI:
- Improve RPCIT instruction tracing
- Fix double-free, use-after-free, and resource leaks in zPCI MSI
and IRQ setup / cleanup paths
- Crypto and zcrypt:
- Improve reply message and length verification checks for CCA,
EP11, and ICA messages to prevent out-of-bounds reads
- Add missing array_index_nospec() for domain index uses in zcrypt
to prevent speculative execution bypass
- Fix out-of-bounds read in cca_query_crypto_facility()
- Various other small fixes and improvements
* tag 's390-7.4-1' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux: (102 commits)
s390/signal: Cleanup coding style
s390/signal: Fix incorrect return address in signal frame
s390/probes: Fix wrong mnemonic comment for clgib
s390/probes: Add missing fixup for bctgr and bic
s390/probes: Fix wrong opcode for cgrb fixup
s390: Select ARCH_MEMORY_ORDER_TSO
s390/mm: Cleanup pXXp_flush_lazy() routines
s390/pci: Add function handle to RPCIT insn trace
s390/pci: Use 32-bit fh outside of inline asm
s390/pci: Adjust alignment in insn trace
s390/sclp: Remove redundant memset after memblock_alloc_low()
s390/entry: Save lowcore on invalid kernel stack
s390/entry: Last breaking event vs invalid kernel stack
s390/percpu: ImplemenHeiko Carstenst this_cpu_try_cmpxchg()
s390/percpu: Drop unused output register from arch_this_cpu_to_op()
s390/percpu: Drop unused output register from arch_this_cpu_add()
s390/percpu: Use asm_inline for arch_this_cpu_add()
s390/extable: Remove EX_TABLE macros and fallback extable search
s390/bpf: Simplify exception table entry handling
s390/sclp: Convert inline assembly to EX_TABLE_INSN
...
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/dinguyen/linux
Pull NIOS2 update from Dinh Nguyen
- Declare __mulsi3 export
* tag 'nios2_update_for_v7.4_part_2' of git://git.kernel.org/pub/scm/linux/kernel/git/dinguyen/linux:
nios2: ksyms: Declare __mulsi3 export
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k
Pull m68k updates from Geert Uytterhoeven:
- Add and enable support for the SVEthLANa Ethernet interface on the
Atari SuperVidel FPGA expansion card
- Enable support for the IDE interface on the Amiga CS-Lab Warp
expension card
- defconfig updates
- Miscellaneous fixes and improvements
* tag 'm68k-for-v7.4-tag1' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k:
m68k: defconfig: Enable ATARI_SVETHLANA and ETHOC
m68k: defconfig: Update defconfigs for v7.3-rc1
zorro: Use assign_bit() where applicable
m68k: apollo: Fix interrupt setup
m68k: defconfig: Enable PATA_CSWARP
m68k: Add support for SVEthLANa
m68k: Add include guard to asm/serial.h
m68k: Use generic BASE_BAUD in asm/serial.h
|
|
RPC pipe handles on IPC$ are allocated from a single server-global IDA,
and ksmbd.mountd looks a pipe up by handle alone, with no session
information. smb2_read_pipe() and smb2_write_pipe() take the handle
straight from the client request (VolatileFileId) and pass it to the
daemon without checking that the requesting session opened it. An
authenticated user can therefore issue an SMB2 READ or WRITE on IPC$
with another session's handle and receive that session's in-flight
DCERPC response, or write attacker-chosen bytes into its pipe stream.
A tree connect to IPC$, which every RPC client already has, is the only
requirement.
Validate ownership before the transfer, the way the sibling pipe paths
already do: under sess->rpc_lock, ksmbd_session_rpc_method() returns 0
for a handle the session did not open. smb2_get_info_file_pipe() and
the FSCTL_PIPE_TRANSCEIVE ioctl already reject such handles; read and
write were the remaining gap. Reject them with STATUS_FILE_CLOSED, as
the ioctl path does.
Fixes: e2f34481b24d ("cifsd: add server-side procedures for SMB3")
Assisted-by: LLM
Signed-off-by: Mohammed Alquraini <mohammed.saad.alquraini@gmail.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
When a send wraps the end of the send buffer, smc_tx_sendmsg() copies
the data in two chunks. If memcpy_from_msg() faults on the second one,
smc_tx_sendmsg() returns send_done, which already counts the first
chunk, but leaves tx_curs_prep and sndbuf_space as they were. send()
reports bytes the peer never receives, and the next send() overwrites
them. Any caller can hit this on its own stream with a buffer whose
tail is unreadable.
On a fault in the second chunk, queue the first one. The next pass
retries the rest and returns send_done if it faults again, so send()
reports only what it queued, as TCP does.
Fixes: e6727f39004b ("smc: send data (through RDMA)")
Cc: stable@vger.kernel.org
Signed-off-by: Bryam Vargas <hexlabsecurity@proton.me>
Link: https://patch.msgid.link/20261013-b4-disp-872a3689-v1-1-2a4c63b04761@proton.me
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
A disconnected durable handle keeps its ksmbd_file and the fd_limit
reservation taken at its initial open. Reconnecting allocates another
volatile ID and takes another reservation for the same file. Final close
returns only one reservation, so repeated reconnects permanently reduce
the server-wide open-file budget.
Move reservation into ksmbd_open_fd() and release it on every creation
failure. Keep the reservation until final close, including while a durable
handle is disconnected. Make ID allocation independent of fd_limit so
reconnects neither consume nor return slots. This also avoids returning an
unreserved slot when persistent ID allocation fails.
New opens continue to return -EMFILE when the budget is exhausted, keeping
the existing STATUS_INSUFFICIENT_RESOURCES mapping.
Fixes: c8efcc786146 ("ksmbd: add support for durable handles v1/v2")
Reported-by: Mohammed <mohammed.saad.alquraini@gmail.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
Session registration failure and connection cleanup can drop the table
reference while the per-session procfs entry is still accessible. If a
reader holds the last remaining reference, its final put calls
proc_remove() from its own callback and waits for that callback to
finish.
Remove the procfs entry before dropping the table reference in the
registration failure path. Queue sessions removed during connection
cleanup and drain their procfs readers after releasing both session
locks, then return their table references. Keep the global hash node
unhashed while a session is queued so a concurrent unregister cannot
drop the reference again.
Previous-session replacement also destroys the file table before the
session is finally freed. Detach the idr under the lock used by procfs
file counters, and report zero files after it has been detached. This
prevents a reader from accessing the idr during or after its
destruction.
Fixes: d7fd1f98607f ("ksmbd: follow SMB2 session expiration semantics")
Cc: stable@vger.kernel.org
Reported-by: Abdifatah Suruur <suruurism@gmail.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
LOGOFF and previous-session replacement leave sessions Expired with a
zero credential expiry. Ended sessions can retain their memory and
connection slots when no further setup arrives on the owning connection.
Unpublish a logged-off session while the request reference keeps its
response signing and encryption keys alive. Remove its procfs entry
before dropping the table reference so a reader cannot remove its own
entry on the last put. Clear proc_entry after removal to avoid removing
it twice.
Reclaim ended sessions from the global table in the existing expiration
worker before it acquires conn_list_lock. Only select Expired sessions
with zero credential expiry and no references beyond the table
reference. Keep sessions whose Kerberos credentials expired available
for reauthentication.
Unlink the session from its setup connection and all channel
connections, so SMB2 sessions and sessions with only secondary SMB3
channels are covered. The session already holds connection references,
and sessions_table_lock prevents concurrent channel removal. Queue
sessions on a separate disposal list, leaving their hash nodes unhashed,
and remove procfs entries and drop table references outside the locks.
Keep the pending setup check before allocating a new session so
abandoned InProgress sessions can release their slots even at the
connection limit. Retain the timeout and reference checks and dispose of
those sessions outside the locks. Leave ended Expired sessions to the
expiration worker.
Fixes: d7fd1f98607f ("ksmbd: follow SMB2 session expiration semantics")
Cc: stable@vger.kernel.org
Reported-by: Abdifatah Suruur <suruurism@gmail.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
The pending setup limit only counts InProgress sessions. A client can
complete authentication repeatedly and retain an unbounded number of Valid
sessions on one connection.
Reserve a connection slot before allocating a session and retain a
connection reference while authentication is in progress. The first SMB3
channel takes over that reservation; binding to another connection reserves
one slot there. Reauthentication and authentication continuation reuse the
existing reservation. Return slots on allocation or insertion failure,
channel removal, and final session destruction.
Accept max_sessions_per_connection in reserved startup-request space and
default to 64 when older tools leave it zero. Keep the request size and
existing field offsets unchanged. Expose the configured limit in procfs.
The existing connection limits bound the aggregate number of sessions
without a separate global session allocation limit.
Fixes: d7fd1f98607f ("ksmbd: follow SMB2 session expiration semantics")
Cc: stable@vger.kernel.org
Reported-by: Abdifatah Suruur <suruurism@gmail.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
ksmbd_session_unregister() removes a session from the global and per-
connection tables before dropping the table reference. A concurrent
reference holder can keep the session and its channel list alive after
those removals.
Connection cleanup can no longer find the unpublished session to remove its
channels. Since a channel only stores a raw conn pointer, the connection
can be freed while another session reference holder is still accessing it,
for example during SMB3.1.1 binding or a procfs read.
Take a connection reference when registering a channel and release it on
insertion failure, channel removal, or final session destruction. Keep the
channels and signing keys until the session is destroyed so a response can
still be signed after the session has been unregistered.
Fixes: d7fd1f98607f ("ksmbd: follow SMB2 session expiration semantics")
Cc: stable@vger.kernel.org
Reported-by: Abdifatah Suruur <suruurism@gmail.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
The expiration worker holds conn_list_lock for reading while checking
for authenticated sessions. Without this change, the check can wait for
conn->session_lock or sessions_table_lock.
In destroy_previous_session(), the status updates after
ksmbd_conn_wait_idle_sess() run with both session locks held for
writing. Both the KSMBD_SESS_GOOD error path and the
KSMBD_SESS_NEED_SETUP success path call ksmbd_all_conn_set_status(),
which takes conn_list_lock for reading.
A writer queued on conn_list_lock can block these new read acquisitions.
The worker then waits for a session lock held by
destroy_previous_session(), which waits behind the writer. The writer
cannot proceed until the worker releases conn_list_lock.
Try both session locks instead of waiting in the scavenging check. Keep
the connection when either lock is busy and let the periodic worker
retry later.
Fixes: d7fd1f98607f ("ksmbd: follow SMB2 session expiration semantics")
Cc: stable@vger.kernel.org
Reported-by: Abdifatah Suruur <suruurism@gmail.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
ksmbd-tools releases before 3.5.2 do not have the payload_sz field in
struct ksmbd_share_config_response; the field occupies the last word of
their reserved[] area, so the kernel always sees payload_sz == 0 from
them. ipc_validate_msg() and share_config_request() used to treat
payload_sz == 0 as "legacy userspace" and skip the payload size checks.
That case was removed by
commit f25e93768fcc ("ksmbd: prevent out-of-bounds reads in share config
responses").
With payload_sz == 0:
- the IPC$ (pipe) response fails the final message size comparison:
these tools always send a one-byte payload, so msg_sz is
sizeof(*resp) + 1, but the expected size is now computed as
sizeof(*resp) + 0;
- every other share fails "payload_sz <= veto_list_sz".
In addition, ksmbd-tools before 3.5.3 copy a global "veto files" setting
into the IPC$ response, so with such a configuration IPC$ also fails the
now unconditional "veto_list_sz > payload_sz" check, even though pipe
shares never use a veto list.
ksmbd_ipc_share_config_request() then returns NULL and every tree
connect fails with STATUS_BAD_NETWORK_NAME, so ksmbd cannot serve any
share when used with these ksmbd-tools releases, for example the 3.5.1
package shipped by Ubuntu 24.04. Login still works, which makes the
failure look like a configuration problem:
CIFS: VFS: BAD_NETWORK_NAME: \\127.0.0.1\IPC$
CIFS: VFS: \\127.0.0.1 failed to connect to IPC (rc=-2)
When payload_sz is zero, derive it from the size of the message actually
received and store it back in the response, so that the checks in
ipc_validate_msg() and the path and veto list parsing in
share_config_request() all use the same bound. The response is the
kernel's own copy made by handle_response(), and msg_sz has already been
checked to be at least sizeof(*resp), so this cannot underflow. All
checks added by that commit then apply unchanged, bounded by the bytes
that really arrived, and newer ksmbd-tools that fill in payload_sz are
unaffected. Also skip the veto list size check for pipe shares:
share_config_request() reads neither the path nor the veto list for
them.
Tested in QEMU with ksmbd-tools 3.5.1 and 3.5.7 and the in-kernel SMB
client over loopback, with and without a global "veto files" setting.
Fixes: f25e93768fcc ("ksmbd: prevent out-of-bounds reads in share config responses")
Cc: stable@vger.kernel.org
Assisted-by: LLM
Signed-off-by: Michail Andreas Michalopoulos <mike.michalo@gmail.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
SetAllocation rounds requests to 512-byte blocks, although
FS_SIZE_INFORMATION advertises the filesystem's f_bsize as its
allocation unit. For example, on a filesystem with a 4 KiB unit, a
1000-byte request can truncate an 8 KiB file to 1024 bytes instead of
4 KiB.
Round the request to the advertised unit and use the same result for
the block-count decision, fallocate, truncate, and the size stored on
the open handle. Validate the unit and rounded size before either
operation. Use DIV_ROUND_UP_ULL() so the calculation works on 32-bit
kernels.
This does not change the existing block-count decision for whether to
call fallocate. In particular, success does not yet guarantee
reservation of holes within the requested range on a sparse file.
Fixes: e2f34481b24d ("cifsd: add server-side procedures for SMB3")
Signed-off-by: DaeMyung Kang <charsyam@gmail.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
When SetAllocation requests a size below the allocated block count but
above EOF, ksmbd truncates the file to that size and restores only the
in-memory i_size. The larger EOF remains on disk and reappears after a
reboot. A 3-byte file preallocated to 64 KiB can acquire a 4 KiB EOF
after reducing its allocation to 4 KiB. Extending EOF this way is not
a reliable way to reclaim the preallocation beyond the old EOF either.
Only use this truncate path when the requested allocation is at or below
the observed EOF. A truncate to the existing EOF still lets filesystems
release preallocated blocks beyond EOF. Remove the in-memory-only i_size
restoration. The size check still precedes vfs_truncate() taking the
inode lock, so a concurrent shrink can make the truncate extend EOF;
that race also existed in the original path.
Fixes: e2f34481b24d ("cifsd: add server-side procedures for SMB3")
Cc: stable@vger.kernel.org
Signed-off-by: DaeMyung Kang <charsyam@gmail.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
MS-SMB2 specifies that subsequent related operations use the FileId from
the preceding operation. Track FileIds from handle-bearing requests so
related requests that use the all-ones placeholder resolve to the
inherited handle.
Return STATUS_INVALID_HANDLE when a related request refers to a FileId
but the preceding operation neither contained nor generated one.
Fixes: e2f34481b24d ("cifsd: add server-side procedures for SMB3")
Reported-by: Ronnie Sahlberg <ronniesahlberg@gmail.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
smb_strtoUTF16() always writes a terminating zero UTF-16 unit after the
converted characters, but get_file_stream_info() budgets only
sizeof(struct smb2_file_stream_info) + streamlen * 2 for a stream entry
and sizeof(struct smb2_file_stream_info) + 7 * 2 for the trailing
"::$DATA" entry. When an entry exactly fills the 448-byte small response
buffer, the terminator is written two bytes past the allocation.
Reserve one extra UTF-16 unit in both budget checks. The terminator is
not part of the wire format, so the reported lengths are unchanged.
Cc: stable@vger.kernel.org
Signed-off-by: Abdifatah Suruur <suruurism@gmail.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
An authenticated client with a writable tree connection can open a
directory with FILE_WRITE_DATA and send FSCTL_SET_ZERO_DATA on that
handle. The handler only checks fp->daccess & FILE_WRITE_DATA_LE, but on
a directory FILE_WRITE_DATA is the legitimate add-file right, so the
check passes. ksmbd_vfs_zero_data() then calls kernel_write() on the
directory file, which has no FMODE_WRITE, and __kernel_write_iter()
hits WARN_ON_ONCE(!(file->f_mode & FMODE_WRITE)), panicking a
panic_on_warn kernel.
Before commit 25e414db7033 ("ksmbd: report holes in allocated range
queries") the dense-file path used vfs_fallocate(), which rejects
directories with -EISDIR. The kernel_write() loop that replaced it has
no such check. Reject a directory handle in ksmbd_vfs_zero_data() before
writing, as ksmbd_vfs_write() already does for SMB2_WRITE.
WARNING: fs/read_write.c:608 at __kernel_write_iter+0x74b/0x8c0
Workqueue: ksmbd-io handle_ksmbd_work
Call Trace:
kernel_write+0x1ef/0x4d0
ksmbd_vfs_zero_data+0x413/0x590
smb2_ioctl+0x21fa/0x66a0
handle_ksmbd_work+0x61b/0x1ff0
Fixes: 25e414db7033 ("ksmbd: report holes in allocated range queries")
Signed-off-by: Lee Seong Hyeon <ihopenre@gmail.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
POSIX filesystem information uses f_bsize for BlockSize even when the
filesystem reports a distinct fragment size for its block counts.
Use f_frsize when available, falling back to f_bsize. Keep f_bsize as
the optimal transfer size and derive the response length from the
information structure instead of a hard-coded value.
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
POSIX filesystem queries use the share root for statfs and do not
validate the requested file handle. This can report the wrong filesystem
for files below a mount point within the share.
Require SMB3.1.1 POSIX extensions and a valid POSIX open. Resolve the
requested handle, including compound file IDs, and use its referenced
path for the filesystem query.
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
POSIX file information queries only check the tree's extension flag,
while directory queries do not verify that the directory was opened with
a POSIX create context.
Require SMB3.1.1, negotiated POSIX tree extensions and a POSIX open for
both file and directory POSIX information requests. Return
STATUS_INVALID_INFO_CLASS when these conditions are not met.
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
The POSIX file query response ends after the owner and group SIDs,
omitting the filename length and filename.
Append the path relative to the share in UTF-16, without a leading
separator, and include its byte length in the response. Check both the
fixed fields and the converted filename against the available output
buffer before writing the response.
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
POSIX responses omit special permission bits, and the create response
reports a file identifier as its reparse tag without encoding the file
type. Directory and file queries also report rdev as the filesystem
device identifier.
Report the POSIX mode and file type in create responses, include
special permission bits in all three response paths, and use the
special-file reparse tag consistently. Set the corresponding DOS
attribute in directory and file queries, report the filesystem device
identifier and initialize reserved fields.
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
Explicit modes carried by NFS mode ACEs can be overwritten by the DACL
permission fallback, and security updates discard special mode bits.
Track modes supplied by ACCESS_ALLOWED NFS mode ACEs and bypass the
permission fallback for them. Apply both permission and special bits
when setting an explicit POSIX mode.
For SET_INFO, require a POSIX open for explicit mode changes and apply
the mode only when DACL security information is requested.
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
The share mode helpers treat a requested POSIX mode of zero as an
unspecified mode, replacing it with the share's default permissions.
Use the requested mode, including zero, before applying the share masks
and forced modes. Pass an explicit default-mode sentinel for non-POSIX
creates to keep their existing mode selection.
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
Using the tree's POSIX extension flag to skip byte-range lock checks
also skips them for ordinary opens on a POSIX-enabled tree.
Base the read and write lock checks on each file's POSIX create context
instead.
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
POSIX append-only opens do not set O_APPEND, and the SMB append sentinel
is rejected as a negative write offset.
Track append-only POSIX opens, set O_APPEND and require the U64_MAX
offset sentinel for their writes. Accept FILE_WRITE_DATA or
FILE_APPEND_DATA for WRITE requests, preserving ordinary SMB write
permissions. Use the POSIX append flag and offset check to prevent
POSIX append-only opens from overwriting existing data.
For synchronous writes, derive the fsync range from the completed write
position so append writes flush the range actually written.
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
POSIX create contexts are only examined when the tree has POSIX
extensions enabled, and IPC requests bypass this validation.
Examine POSIX contexts before selecting the IPC create path. Reject
duplicate contexts and require an SMB3.1.1 connection, a POSIX-enabled
tree and a non-IPC share for POSIX creates.
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
The POSIX negotiate context currently enables extensions without
checking its payload, and duplicate contexts are accepted.
Reject duplicate contexts and invalid payload lengths. Enable POSIX
extensions only for the published tag, and use the same tag constant
when building the negotiate response.
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
Arm a lease break acknowledgment timer once the common notification path
accepts an ACK-required notification for transmission. Keep the deadline
on the lease across channel retries and scan every lease in the global
lease tables when the timer expires. Set timed-out leases and all their
opens to NONE, then rearm for the next deadline.
Serialize client acknowledgments, notification state and timer expiry
under the lease lock, taking each open's state lock after it. Update every
open before publishing break completion so a late acknowledgment cannot
restore an expired lease. Wake blocked opens on timeout, send failure and
global lease-table teardown, and cancel the timer at teardown.
Keep the existing connection-independent lease-table lifetime and use the
notifier's opinfo reference to retain the lease. Stop after selecting one
open during lease-key lookup.
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
Arm the oplock break acknowledgment timer once the common notification
path accepts a classic break for transmission. Preserve its shared retry
deadline when another channel is tried. Store a deadline per open and scan
the global open table when the earliest deadline expires. Drop expired
breaks to NONE and restart the timer while other breaks remain.
Serialize timer expiry with client acknowledgments and close using the
oplock state lock. Wake blocked requests when a break expires or sending
the notification fails, and invalidate durable reconnect for the open.
Use wake_up_all() for acknowledgment, timeout, close and send failure.
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
SMB2 lock ranges use unsigned 64-bit offsets, but file_lock stores signed
loff_t values. Track valid ranges above OFFSET_MAX in ksmbd's inode-based
lock index and arbitrate conflicts there. Keep these offsets out of
file_lock, and handle blocking requests, unlocks, rollback, close cleanup
and cancellation in the SMB layer.
Reserve large ranges under the index spinlock until the full request
commits. Keep lookup and ownership independent of the connection so
channel removal and hard-link opens use the same conflict checks.
Finish large-range unlocks in the SMB layer after unlocking any VFS
portion. Wake large-range lock waiters when normal locks are removed
by unlock or close.
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
A completed DurableHandleReqV2 CREATE can be replayed through another
channel while its original channel is exiting. The existing replay path
returns the open but leaves its connection and oplock references pointing
at the old channel. Its open count also stays on that channel, and a
later CLOSE through the surviving channel decrements the wrong counter.
After checking the owner, lease key and persistent flag, verify that an
active open is still published in the requesting session's file table.
Do this even when no oplock was granted, and before processing a consumed
replay. Do not read a borrowed opinfo session pointer for this check.
Keep a healthy original binding. If the original channel is exiting or
releasing, move the open and its own oplock references to the requesting
channel. Serialize this with CLOSE using the session file-table lock,
and with detach, reconnect and notification snapshots using m_lock.
The earlier lock index and lease-table changes make their lifetimes
independent of this connection change. No old-channel r_count wait is
needed, including when an unrelated request remains active there.
Account final close against fp->conn in one place, balance durable detach
and reconnect, and transfer the count with a replay binding change. A
final global lookup put may run after CLOSE has removed the volatile id;
unlink the open from the inode list on that final close as well.
A live completed CREATE replay may follow a break that lowered the
durable open's batch oplock. Apply the batch eligibility check only to
detached reconnects, while retaining the owner, lease key and persistent
flag checks for active replays.
Fixes: bb8bf7eb13b5 ("ksmbd: add SMB3 request replay support")
Reported-by: Pan Guihua <pan.guihua@datatom.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
SMB locks belong to an open, but the lock conflict and UNLOCK scan only
visits connections in conn_list. When a channel is removed, its locks
remain attached to live opens and become invisible to another channel.
An exact UNLOCK returns STATUS_RANGE_NOT_LOCKED, and zero-byte lock
conflicts are no longer detected. A request finishing on the old channel
after it is removed can publish another invisible lock.
Index granted SMB locks by VFS inode, with the open's lock_list retaining
ownership. Protect lookup, publication and removal with one spinlock.
Use the inode rather than the dentry-based ksmbd_inode so hard-link opens
still participate in the same conflict checks.
Keep these entries through durable detach and reconnect. Publish a batch
only after the existing LOCK handler commits it, and remove every entry
before final close drops the file reference. Drop the per-lock connection
references and connection-local lists; channel changes no longer require
moving lock entries or waiting for unrelated requests to finish.
Prefer an exclusive lock when several locks on an open match an UNLOCK.
This preserves stacked zero-byte lock behavior despite the inode hash
inserting newer locks before older entries.
Fixes: d63528eb0d43 ("ksmbd: free ksmbd_lock when file is closed")
Reported-by: Pan Guihua <pan.guihua@datatom.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|
|
Disconnecting one channel destroys the lease table for its ClientGuid,
even when another channel still uses the same opens. A lease break ACK
on the surviving channel can no longer find the lease. Detached durable
opens also lose their lease lookup entries before reconnect or expiry.
Tie a client table to its lease entries instead of the connection which
created it. Remove the table when its last lease is removed, and release
any remaining entries at server shutdown. Do not destroy client tables
from connection teardown or keep a connection reference in each table.
For a detached lease, collect break candidates using the table's
ClientGuid directly. The existing notification retry path takes its own
references on matching live connections.
Fixes: 0626e6641f6b ("cifsd: add server handler for central processing and tranport layers")
Reported-by: Pan Guihua <pan.guihua@datatom.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
|