Add a comment to explain why KVM treats vCPUs with pending interrupts as
in-kernel when a vCPU wants to yield to a vCPU that was preempted while
running in kernel mode.

Link: https://lore.kernel.org/r/20240110003938.490206-5-seanjc@google.comSigned-off-by: Sean Christopherson <seanjc@google.com>

Directly return the boolean result of whether or not a vCPU has a pending
interrupt instead of effectively doing:

  if (true)
	return true;

  return false;
Reviewed-by: Yuan Yao <yuan.yao@intel.com>
Link: https://lore.kernel.org/r/20240110003938.490206-4-seanjc@google.comSigned-off-by: Sean Christopherson <seanjc@google.com>

Snapshot preempted_in_kernel using kvm_arch_vcpu_in_kernel() so that the
flag is "accurate" (or rather, consistent and deterministic within KVM)
for guests with protected state, and explicitly use preempted_in_kernel
when checking if a vCPU was preempted in kernel mode instead of bouncing
through kvm_arch_vcpu_in_kernel().

Drop the gnarly logic in kvm_arch_vcpu_in_kernel() that redirects to
preempted_in_kernel if the target vCPU is not the "running", i.e. loaded,
vCPU, as the only reason that code existed was for the directed yield case
where KVM wants to check the CPL of a vCPU that may or may not be loaded
on the current pCPU.

Cc: Like Xu <like.xu.linux@gmail.com>
Reviewed-by: Yuan Yao <yuan.yao@intel.com>
Link: https://lore.kernel.org/r/20240110003938.490206-3-seanjc@google.comSigned-off-by: Sean Christopherson <seanjc@google.com>

Plumb in a dedicated hook for querying whether or not a vCPU was preempted
in-kernel.  Unlike literally every other architecture, x86's VMX can check
if a vCPU is in kernel context if and only if the vCPU is loaded on the
current pCPU.

x86's kvm_arch_vcpu_in_kernel() works around the limitation by querying
kvm_get_running_vcpu() and redirecting to vcpu->arch.preempted_in_kernel
as needed.  But that's unnecessary, confusing, and fragile, e.g. x86 has
had at least one bug where KVM incorrectly used a stale
preempted_in_kernel.

No functional change intended.
Reviewed-by: Yuan Yao <yuan.yao@intel.com>
Link: https://lore.kernel.org/r/20240110003938.490206-2-seanjc@google.comSigned-off-by: Sean Christopherson <seanjc@google.com>

WARN if kvm.ko is unloaded with an elevated kvm_has_noapic_vcpu to guard
against incorrect management of the key, e.g. to detect if KVM fails to
decrement the key in error paths.  Because kvm_has_noapic_vcpu is purely
an optimization, in all likelihood KVM could completely botch handling of
kvm_has_noapic_vcpu and no one would notice (which is a good argument for
deleting the key entirely, but that's a problem for another day).

Note, ideally the sanity check would be performance when kvm_usage_count
goes to zero, but adding an arch callback just for this sanity check isn't
at all worth doing.

Link: https://lore.kernel.org/r/20240209222047.394389-3-seanjc@google.comSigned-off-by: Sean Christopherson <seanjc@google.com>

Move incrementing and decrementing of kvm_has_noapic_vcpu into
kvm_create_lapic() and kvm_free_lapic() respectively to fix a benign bug
where KVM fails to decrement the count if vCPU creation ultimately fails,
e.g. due to a memory allocation failing.

Note, the bug is benign as kvm_has_noapic_vcpu is used purely to optimize
lapic_in_kernel() checks, and that optimization is quite dubious.  That,
and practically speaking no setup that cares at all about performance runs
with a userspace local APIC.
Reported-by: Li RongQing <lirongqing@baidu.com>
Cc: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Xu Yilun <yilun.xu@linux.intel.com>
Link: https://lore.kernel.org/r/20240209222047.394389-2-seanjc@google.comSigned-off-by: Sean Christopherson <seanjc@google.com>

Now that vmx->req_immediate_exit is used only in the scope of
vmx_vcpu_run(), use force_immediate_exit to detect that KVM should usurp
the VMX preemption to force a VM-Exit and let vendor code fully handle
forcing a VM-Exit.

Opportunsitically drop __kvm_request_immediate_exit() and just have
vendor code call smp_send_reschedule() directly.  SVM already does this
when injecting an event while also trying to single-step an IRET, i.e.
it's not exactly secret knowledge that KVM uses a reschedule IPI to force
an exit.

Link: https://lore.kernel.org/r/20240110012705.506918-7-seanjc@google.comSigned-off-by: Sean Christopherson <seanjc@google.com>

Eat VMX treemption timer exits in the fastpath regardless of whether L1 or
L2 is active.  The VM-Exit is 100% KVM-induced, i.e. there is nothing
directly related to the exit that KVM needs to do on behalf of the guest,
thus there is no reason to wait until the slow path to do nothing.

Opportunistically add comments explaining why preemption timer exits for
emulating the guest's APIC timer need to go down the slow path.

Link: https://lore.kernel.org/r/20240110012705.506918-6-seanjc@google.comSigned-off-by: Sean Christopherson <seanjc@google.com>

Let the fastpath code decide which exits can/can't be handled in the
fastpath when L2 is active, e.g. when KVM generates a VMX preemption
timer exit to forcefully regain control, there is no "work" to be done and
so such exits can be handled in the fastpath regardless of whether L1 or
L2 is active.

Moving the is_guest_mode() check into the fastpath code also makes it
easier to see that L2 isn't allowed to use the fastpath in most cases,
e.g. it's not immediately obvious why handle_fastpath_preemption_timer()
is called from the fastpath and the normal path.

Link: https://lore.kernel.org/r/20240110012705.506918-5-seanjc@google.comSigned-off-by: Sean Christopherson <seanjc@google.com>

Handle VMX preemption timer VM-Exits due to KVM forcing an exit in the
exit fastpath, i.e. avoid calling back into handle_preemption_timer() for
the same exit.  There is no work to be done for forced exits, as the name
suggests the goal is purely to get control back in KVM.

In addition to shaving a few cycles, this will allow cleanly separating
handle_fastpath_preemption_timer() from handle_preemption_timer(), e.g.
it's not immediately obvious why _apparently_ calling
handle_fastpath_preemption_timer() twice on a "slow" exit is necessary:
the "slow" call is necessary to handle exits from L2, which are excluded
from the fastpath by vmx_vcpu_run().

Link: https://lore.kernel.org/r/20240110012705.506918-4-seanjc@google.comSigned-off-by: Sean Christopherson <seanjc@google.com>

Re-enter the guest in the fast path if VMX preeemption timer VM-Exit was
"spurious", i.e. if KVM "soft disabled" the timer by writing -1u and by
some miracle the timer expired before any other VM-Exit occurred.  This is
just an intermediate step to cleaning up the preemption timer handling,
optimizing these types of spurious VM-Exits is not interesting as they are
extremely rare/infrequent.

Link: https://lore.kernel.org/r/20240110012705.506918-3-seanjc@google.comSigned-off-by: Sean Christopherson <seanjc@google.com>

Annotate the kvm_entry() tracepoint with "immediate exit" when KVM is
forcing a VM-Exit immediately after VM-Enter, e.g. when KVM wants to
inject an event but needs to first complete some other operation.
Knowing that KVM is (or isn't) forcing an exit is useful information when
debugging issues related to event injection.
Suggested-by: Maxim Levitsky <mlevitsk@redhat.com>
Link: https://lore.kernel.org/r/20240110012705.506918-2-seanjc@google.comSigned-off-by: Sean Christopherson <seanjc@google.com>

Bite the bullet, and open code all direct reads of DR6 and DR7.  KVM
currently has a mix of open coded accesses and calls to kvm_get_dr(),
which is confusing and ugly because there's no rhyme or reason as to why
any particular chunk of code uses kvm_get_dr().

The obvious alternative is to force all accesses through kvm_get_dr(),
but it's not at all clear that doing so would be a net positive, e.g. even
if KVM ends up wanting/needing to force all reads through a common helper,
e.g. to play caching games, the cost of reverting this change is likely
lower than the ongoing cost of maintaining weird, arbitrary code.

No functional change intended.

Cc: Mathias Krause <minipli@grsecurity.net>
Reviewed-by: Mathias Krause <minipli@grsecurity.net>
Link: https://lore.kernel.org/r/20240209220752.388160-3-seanjc@google.comSigned-off-by: Sean Christopherson <seanjc@google.com>

Convert kvm_get_dr()'s output parameter to a return value, and clean up
most of the mess that was created by forcing callers to provide a pointer.

No functional change intended.
Acked-by: Mathias Krause <minipli@grsecurity.net>
Reviewed-by: Mathias Krause <minipli@grsecurity.net>
Link: https://lore.kernel.org/r/20240209220752.388160-2-seanjc@google.comSigned-off-by: Sean Christopherson <seanjc@google.com>

push and emulate_pop are counterparts. Rename push to emulate_push and
harmonize its function signature with emulate_pop. This should remove
a bit of cognitive load when reading this code.
Signed-off-by: Julian Stecklina <julian.stecklina@cyberus-technology.de>
Link: https://lore.kernel.org/r/20231009092054.556935-2-julian.stecklina@cyberus-technology.deSigned-off-by: Sean Christopherson <seanjc@google.com>

Explicitly zero out variables passed to emulate_pop() as output params
to harden against consuming uninitialized data, and to make sanitizers
happy.  Many flows that use emulate_pop() pass an "unsigned long" so as
to be able to hold the largest possible operand, but the actual number
of bytes written is usually the word with of the vCPU.  E.g. if the vCPU
is in 16-bit or 32-bit mode (on a 64-bit host), the upper portion of the
output param will be uninitialized.

Passing around the uninitialized data is benign, as actual KVM usage of
the output is also tied to the word width, but passing around
uninitialized data makes some sanitizers rightly complain.

Note, initializing the data in emulate_pop() is not a safe alternative,
e.g. it would result in em_leave() clobbering RBP[31:16] if LEAVE were
emulated with a 16-bit stack.
Signed-off-by: Julian Stecklina <julian.stecklina@cyberus-technology.de>
Link: https://lore.kernel.org/r/20231009092054.556935-1-julian.stecklina@cyberus-technology.de
[sean: massage changelog, drop em_popa() variable size change]]
Signed-off-by: Sean Christopherson <seanjc@google.com>

The MOVBE instruction can come with an operand-size prefix (66h). In
this, case the x86 emulation code returns EMULATION_FAILED.

It turns out that em_movbe can already handle this case and all that
is missing is an entry in respective opcode tables to populate
gprefix->pfx_66.
Signed-off-by: Thomas Prescher <thomas.prescher@cyberus-technology.de>
Signed-off-by: Julian Stecklina <julian.stecklina@cyberus-technology.de>
Acked-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/r/20231212095938.26731-1-julian.stecklina@cyberus-technology.deSigned-off-by: Sean Christopherson <seanjc@google.com>

The ioctl()s to get and set KVM's debug registers are broken for 32 bit
kernels as they'd only copy half of the user register state because of a
UAPI and in-kernel type mismatch (__u64 vs. unsigned long; 8 vs. 4
bytes).

This makes it impossible for userland to set anything but DR0 without
resorting to bit folding tricks.

Switch to a loop for copying debug registers that'll implicitly do the
type conversion for us, if needed.

There are likely no users (left) for 32bit KVM, fix the bug nonetheless.

Fixes: a1efbe77 ("KVM: x86: Add support for saving&restoring debug registers")
Signed-off-by: Mathias Krause <minipli@grsecurity.net>
Link: https://lore.kernel.org/r/20240203124522.592778-4-minipli@grsecurity.netSigned-off-by: Sean Christopherson <seanjc@google.com>

Use the recently introduced guard(mutex) infrastructure acquire and
automatically release vendor_module_lock when the guard goes out of scope.
Drop the inner __kvm_x86_vendor_init(), its sole purpose was to simplify
releasing vendor_module_lock in error paths.

No functional change intended.
Signed-off-by: Nikolay Borisov <nik.borisov@suse.com>
Reviewed-by: Kai Huang <kai.huang@intel.com>
Link: https://lore.kernel.org/r/20231030141728.1406118-1-nik.borisov@suse.com
[sean: rewrite changelog]
Signed-off-by: Sean Christopherson <seanjc@google.com>

Pull cxl fixes from Dan Williams:
 "A build regression fix, a device compatibility fix, and an original
  bug preventing creation of large (16 device) interleave sets:

   - Fix unit test build regression fallout from global
     "missing-prototypes" change

   - Fix compatibility with devices that do not support interrupts

   - Fix overflow when calculating the capacity of large interleave sets"

* tag 'cxl-fixes-6.8-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/cxl/cxl:
  cxl/region：Fix overflow issue in alloc_hpa()
  cxl/pci: Skip irq features if MSI/MSI-X are not supported
  tools/testing/nvdimm: Disable "missing prototypes / declarations" warnings
  tools/testing/cxl: Disable "missing prototypes / declarations" warnings

Pull MIPS fixes from Thomas Bogendoerfer:

 - fix boot issue on single core Lantiq Danube devices

 - fix boot issue on Loongson64 platforms

 - fix improper FPU setup

 - fix missing prototypes issues

* tag 'mips-fixes_6.8_1' of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux:
  mips: Call lose_fpu(0) before initializing fcr31 in mips_set_personality_nan
  MIPS: loongson64: set nid for reserved memblock region
  Revert "MIPS: loongson64: set nid for reserved memblock region"
  MIPS: lantiq: register smp_ops on non-smp platforms
  MIPS: loongson64: set nid for reserved memblock region
  MIPS: reserve exception vector space ONLY ONCE
  MIPS: BCM63XX: Fix missing prototypes
  MIPS: sgi-ip32: Fix missing prototypes
  MIPS: sgi-ip30: Fix missing prototypes
  MIPS: fw arc: Fix missing prototypes
  MIPS: sgi-ip27: Fix missing prototypes
  MIPS: Alchemy: Fix missing prototypes
  MIPS: Cobalt: Fix missing prototypes

Pull locking fix from Borislav Petkov:

 - Prevent an inconsistent futex operation leading to stale state
   exposure

* tag 'locking_urgent_for_v6.8_rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  futex: Prevent the reuse of stale pi_state

Pull irq fix from Borislav Petkov:

 - Initialize the resend node of each IRQ descriptor, not only the first
   one

* tag 'irq_urgent_for_v6.8_rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  genirq: Initialize resend_node hlist for all interrupt descriptors

Pull timer fixes from Borislav Petkov:

 - Preserve the number of idle calls and sleep entries across CPU
   hotplug events in order to be able to compute correct averages

 - Limit the duration of the clocksource watchdog checking interval as
   too long intervals lead to wrongly marking the TSC as unstable

* tag 'timers_urgent_for_v6.8_rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  tick/sched: Preserve number of idle sleeps across CPU hotplug events
  clocksource: Skip watchdog check for large watchdog intervals

Pull x86 fixes from Borislav Petkov:

 - Make sure 32-bit syscall registers are properly sign-extended

 - Add detection for AMD's Zen5 generation CPUs and Intel's Clearwater
   Forest CPU model number

 - Make a stub function export non-GPL because it is part of the
   paravirt alternatives and that can be used by non-GPL code

* tag 'x86_urgent_for_v6.8_rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/CPU/AMD: Add more models to X86_FEATURE_ZEN5
  x86/entry/ia32: Ensure s32 is sign extended to s64
  x86/cpu: Add model number for Intel Clearwater Forest processor
  x86/CPU/AMD: Add X86_FEATURE_ZEN5
  x86/paravirt: Make BUG_func() usable by non-GPL modules

Pull memblock fix from Mike Rapoport:
 "Fix crash when reserved memory is not added to memory.

  When CONFIG_DEFERRED_STRUCT_PAGE_INIT is enabled, the initialization
  of reserved pages may cause access of NODE_DATA() with invalid nid and
  crash.

  Add a fall back to early_pfn_to_nid() in memmap_init_reserved_pages()
  to ensure a valid node id is always passed to init_reserved_page()"

* tag 'fixes-2024-01-28' of git://git.kernel.org/pub/scm/linux/kernel/git/rppt/memblock:
  memblock: fix crash when reserved memory is not added to memory

Merge tag 'platform-drivers-x86-v6.8-2' of git://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86

Pull x86 platform driver fixes from Hans de Goede:

 - WMI bus driver fixes

 - Second attempt (previously reverted) at P2SB PCI rescan deadlock fix

 - AMD PMF driver improvements

 - MAINTAINERS updates

 - Misc other small fixes and hw-id additions

* tag 'platform-drivers-x86-v6.8-2' of git://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86:
  platform/x86: touchscreen_dmi: Add info for the TECLAST X16 Plus tablet
  platform/x86/intel/ifs: Call release_firmware() when handling errors.
  platform/x86/amd/pmf: Fix memory leak in amd_pmf_get_pb_data()
  platform/x86/amd/pmf: Get ambient light information from AMD SFH driver
  platform/x86/amd/pmf: Get Human presence information from AMD SFH driver
  platform/mellanox: mlxbf-pmc: Fix offset calculation for crspace events
  platform/mellanox: mlxbf-tmfifo: Drop Tx network packet when Tx TmFIFO is full
  MAINTAINERS: remove defunct acpi4asus project info from asus notebooks section
  MAINTAINERS: add Luke Jones as maintainer for asus notebooks
  MAINTAINERS: Remove Perry Yuan as DELL WMI HARDWARE PRIVACY SUPPORT maintainer
  platform/x86: silicom-platform: Add missing "Description:" for power_cycle sysfs attr
  platform/x86: intel-wmi-sbl-fw-update: Fix function name in error message
  platform/x86: p2sb: Use pci_resource_n() in p2sb_read_bar0()
  platform/x86: p2sb: Allow p2sb_bar() calls during PCI device probe
  platform/x86: intel-uncore-freq: Fix types in sysfs callbacks
  platform/x86: wmi: Fix wmi_dev_probe()
  platform/x86: wmi: Fix notify callback locking
  platform/x86: wmi: Decouple legacy WMI notify handlers from wmi_block_list
  platform/x86: wmi: Return immediately if an suitable WMI event is found
  platform/x86: wmi: Fix error handling in legacy WMI notify handler functions

Merge tag 'loongarch-fixes-6.8-1' of git://git.kernel.org/pub/scm/linux/kernel/git/chenhuacai/linux-loongson

Pull LoongArch fixes from Huacai Chen:
 "Fix boot failure on machines with more than 8 nodes, and fix two build
  errors about KVM"

* tag 'loongarch-fixes-6.8-1' of git://git.kernel.org/pub/scm/linux/kernel/git/chenhuacai/linux-loongson:
  LoongArch: KVM: Add returns to SIMD stubs
  LoongArch: KVM: Fix build due to API changes
  LoongArch/smp: Call rcutree_report_cpu_starting() at tlb_init()

Pull xfs fix from Chandan Babu:

 - Fix read only mounts when using fsopen mount API

* tag 'xfs-6.8-fixes-1' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux:
  xfs: read only mounts with fsopen mount API are busted

Pull bcachefs fixes from Kent Overstreet:

 - fix for REQ_OP_FLUSH usage; this fixes filesystems going read only
   with -EOPNOTSUPP from the block layer.

   (this really should have gone in with the block layer patch causing
   the -EOPNOTSUPP, or should have gone in before).

 - fix an allocation in non-sleepable context

 - fix one source of srcu lock latency, on devices with terrible discard
   latency

 - fix a reattach_inode() issue in fsck

* tag 'bcachefs-2024-01-26' of https://evilpiepirate.org/git/bcachefs:
  bcachefs: __lookup_dirent() works in snapshot, not subvol
  bcachefs: discard path uses unlock_long()
  bcachefs: fix incorrect usage of REQ_OP_FLUSH
  bcachefs: Add gfp flags param to bch2_prt_task_backtrace()

Pull smb server fixes from Steve French:

 - Fix netlink OOB

 - Minor kernel doc fix

* tag '6.8-rc2-smb3-server-fixes' of git://git.samba.org/ksmbd:
  ksmbd: fix global oob in ksmbd_nl_policy
  smb: Fix some kernel-doc comments

Pull smb client fixes from Steve French:
 "Nine cifs/smb client fixes

   - Four network error fixes (three relating to replays of requests
     that need to be retried, and one fixing some places where we were
     returning the wrong rc up the stack on network errors)

   - Two multichannel fixes including locking fix and case where subset
     of channels need reconnect

   - netfs integration fixup: share remote i_size with netfslib

   - Two small cleanups (one for addressing a clang warning)"

* tag '6.8-rc1-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6:
  cifs: fix stray unlock in cifs_chan_skip_or_disable
  cifs: set replay flag for retries of write command
  cifs: commands that are retried should have replay flag set
  cifs: helper function to check replayable error codes
  cifs: translate network errors on send to -ECONNABORTED
  cifs: cifs_pick_channel should try selecting active channels
  cifs: Share server EOF pos with netfslib
  smb: Work around Clang __bdos() type confusion
  smb: client: delete "true", "false" defines

If we still own the FPU after initializing fcr31, when we are preempted
the dirty value in the FPU will be read out and stored into fcr31,
clobbering our setting.  This can cause an improper floating-point
environment after execve().  For example:

    zsh% cat measure.c
    #include <fenv.h>
    int main() { return fetestexcept(FE_INEXACT); }
    zsh% cc measure.c -o measure -lm
    zsh% echo $((1.0/3)) # raising FE_INEXACT
    0.33333333333333331
    zsh% while ./measure; do ; done
    (stopped in seconds)

Call lose_fpu(0) before setting fcr31 to prevent this.

Closes: https://lore.kernel.org/linux-mips/7a6aa1bbdbbe2e63ae96ff163fab0349f58f1b9e.camel@xry111.site/
Fixes: 9b26616c ("MIPS: Respect the ISA level in FCSR handling")
Cc: stable@vger.kernel.org
Signed-off-by: Xi Ruoyao <xry111@xry111.site>
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>

Commit 61167ad5("mm: pass nid to reserve_bootmem_region()") reveals
that reserved memblock regions have no valid node id set, just set it
right since loongson64 firmware makes it clear in memory layout info.

This works around booting failure on 3A1000+ since commit 61167ad5
("mm: pass nid to reserve_bootmem_region()") under
CONFIG_DEFERRED_STRUCT_PAGE_INIT.
Signed-off-by: Huang Pei <huangpei@loongson.cn>
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>

This reverts commit ce7b1b97.
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>

Pull ata updates from Niklas Cassel:

 - Fix an incorrect link_power_management_policy sysfs attribute value.

   We were previously using the same attribute value for two different
   LPM policies (me)

 - Add a ASMedia ASM1166 quirk.

   The SATA host controller always reports that it has 32 ports, even
   though it only has six ports. Add a quirk that overrides the value
   reported by the controller (Conrad)

 - Add a ASMedia ASM1061 quirk.

   The SATA host controller completely ignores the upper 21 bits of the
   DMA address. This causes IOMMU error events when a (valid) DMA
   address actually has any of the upper 21 bits set. Add a quirk that
   limits the dma_mask to 43-bits (Lennert)

* tag 'ata-6.8-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/libata/linux:
  ahci: add 43-bit DMA address quirk for ASMedia ASM1061 controllers
  ahci: asm1166: correct count of reported ports
  ata: libata-sata: improve sysfs description for ATA_LPM_UNKNOWN

Pull block fixes from Jens Axboe:

 - RCU warning fix for md (Mikulas)

 - Fix for an aoe issue that lockdep rightfully complained about
   (Maksim)

 - Fix for an error code change in partitioning that caused a regression
   with some tools (Li)

 - Fix for a data direction warning with bi-direction commands
   (Christian)

* tag 'block-6.8-2024-01-26' of git://git.kernel.dk/linux:
  md: fix a suspicious RCU usage warning
  aoe: avoid potential deadlock at set_capacity
  block: Fix WARNING in _copy_from_iter
  block: Move checking GENHD_FL_NO_PART to bdev_add_partition()

Pull io_uring fix from Jens Axboe:
 "Just a single tweak to the newly added IORING_OP_FIXED_FD_INSTALL from
  Paul, ensuring it goes via the audit path and playing it safe by
  excluding it from using registered creds"

* tag 'io_uring-6.8-2024-01-26' of git://git.kernel.dk/linux:
  io_uring: enable audit and restrict cred override for IORING_OP_FIXED_FD_INSTALL

Pull thermal control update from Rafael Wysocki:
 "Remove some dead code from the Intel powerclamp thermal control driver
  (Srinivas Pandruvada)"

* tag 'thermal-6.8-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  thermal: intel: powerclamp: Remove dead code for target mwait value