Mark BTF as reserved in DEBUGCTL on AMD, as KVM doesn't actually support
BTF, and fully enabling BTF virtualization is non-trivial due to
interactions with the emulator, guest_debug, #DB interception, nested SVM,
etc.
Don't inject #GP if the guest attempts to set BTF, as there's no way to
communicate lack of support to the guest, and instead suppress the flag
and treat the WRMSR as (partially) unsupported.
In short, make KVM behave the same on AMD and Intel (VMX already squashes
BTF).
Note, due to other bugs in KVM's handling of DEBUGCTL, the only way BTF
has "worked" in any capacity is if the guest simultaneously enables LBRs.
Reported-by: Ravi Bangoria <ravi.bangoria@amd.com> Cc: stable@vger.kernel.org Reviewed-and-tested-by: Ravi Bangoria <ravi.bangoria@amd.com> Link: https://lore.kernel.org/r/20250227222411.3490595-3-seanjc@google.com Signed-off-by: Sean Christopherson <seanjc@google.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Drop bits 5:2 from the guest's effective DEBUGCTL value, as AMD changed
the architectural behavior of the bits and broke backwards compatibility.
On CPUs without BusLockTrap (or at least, in APMs from before ~2023),
bits 5:2 controlled the behavior of external pins:
Performance-Monitoring/Breakpoint Pin-Control (PBi)—Bits 5:2, read/write.
Software uses thesebits to control the type of information reported by
the four external performance-monitoring/breakpoint pins on the
processor. When a PBi bit is cleared to 0, the corresponding external pin
(BPi) reports performance-monitor information. When a PBi bit is set to
1, the corresponding external pin (BPi) reports breakpoint information.
With the introduction of BusLockTrap, presumably to be compatible with
Intel CPUs, AMD redefined bit 2 to be BLCKDB:
Bus Lock #DB Trap (BLCKDB)—Bit 2, read/write. Software sets this bit to
enable generation of a #DB trap following successful execution of a bus
lock when CPL is > 0.
and redefined bits 5:3 (and bit 6) as "6:3 Reserved MBZ".
Ideally, KVM would treat bits 5:2 as reserved. Defer that change to a
feature cleanup to avoid breaking existing guest in LTS kernels. For now,
drop the bits to retain backwards compatibility (of a sort).
Note, dropping bits 5:2 is still a guest-visible change, e.g. if the guest
is enabling LBRs *and* the legacy PBi bits, then the state of the PBi bits
is visible to the guest, whereas now the guest will always see '0'.
Reported-by: Ravi Bangoria <ravi.bangoria@amd.com> Cc: stable@vger.kernel.org Reviewed-and-tested-by: Ravi Bangoria <ravi.bangoria@amd.com> Link: https://lore.kernel.org/r/20250227222411.3490595-2-seanjc@google.com Signed-off-by: Sean Christopherson <seanjc@google.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When running SEV-SNP guests on a CPU that supports DebugSwap, always save
the host's DR0..DR3 mask MSR values irrespective of whether or not
DebugSwap is enabled, to ensure the host values aren't clobbered by the
CPU. And for now, also save DR0..DR3, even though doing so isn't
necessary (see below).
SVM_VMGEXIT_AP_CREATE is deeply flawed in that it allows the *guest* to
create a VMSA with guest-controlled SEV_FEATURES. A well behaved guest
can inform the hypervisor, i.e. KVM, of its "requested" features, but on
CPUs without ALLOWED_SEV_FEATURES support, nothing prevents the guest from
lying about which SEV features are being enabled (or not!).
If a misbehaving guest enables DebugSwap in a secondary vCPU's VMSA, the
CPU will load the DR0..DR3 mask MSRs on #VMEXIT, i.e. will clobber the
MSRs with '0' if KVM doesn't save its desired value.
Note, DR0..DR3 themselves are "ok", as DR7 is reset on #VMEXIT, and KVM
restores all DRs in common x86 code as needed via hw_breakpoint_restore().
I.e. there is no risk of host DR0..DR3 being clobbered (when it matters).
However, there is a flaw in the opposite direction; because the guest can
lie about enabling DebugSwap, i.e. can *disable* DebugSwap without KVM's
knowledge, KVM must not rely on the CPU to restore DRs. Defer fixing
that wart, as it's more of a documentation issue than a bug in the code.
Note, KVM added support for DebugSwap on commit d1f85fbe836e ("KVM: SEV:
Enable data breakpoints in SEV-ES"), but that is not an appropriate Fixes,
as the underlying flaw exists in hardware, not in KVM. I.e. all kernels
that support SEV-SNP need to be patched, not just kernels with KVM's full
support for DebugSwap (ignoring that DebugSwap support landed first).
Opportunistically fix an incorrect statement in the comment; on CPUs
without DebugSwap, the CPU does NOT save or load debug registers, i.e.
Fixes: e366f92ea99e ("KVM: SEV: Support SEV-SNP AP Creation NAE event") Cc: stable@vger.kernel.org Cc: Naveen N Rao <naveen@kernel.org> Cc: Kim Phillips <kim.phillips@amd.com> Cc: Tom Lendacky <thomas.lendacky@amd.com> Cc: Alexey Kardashevskiy <aik@amd.com> Reviewed-by: Tom Lendacky <thomas.lendacky@amd.com> Link: https://lore.kernel.org/r/20250227012541.3234589-2-seanjc@google.com Signed-off-by: Sean Christopherson <seanjc@google.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Enable/disable local IRQs, i.e. set/clear RFLAGS.IF, in the common
svm_vcpu_enter_exit() just after/before guest_state_{enter,exit}_irqoff()
so that VMRUN is not executed in an STI shadow. AMD CPUs have a quirk
(some would say "bug"), where the STI shadow bleeds into the guest's
intr_state field if a #VMEXIT occurs during injection of an event, i.e. if
the VMRUN doesn't complete before the subsequent #VMEXIT.
The spurious "interrupts masked" state is relatively benign, as it only
occurs during event injection and is transient. Because KVM is already
injecting an event, the guest can't be in HLT, and if KVM is querying IRQ
blocking for injection, then KVM would need to force an immediate exit
anyways since injecting multiple events is impossible.
However, because KVM copies int_state verbatim from vmcb02 to vmcb12, the
spurious STI shadow is visible to L1 when running a nested VM, which can
trip sanity checks, e.g. in VMware's VMM.
Hoist the STI+CLI all the way to C code, as the aforementioned calls to
guest_state_{enter,exit}_irqoff() already inform lockdep that IRQs are
enabled/disabled, and taking a fault on VMRUN with RFLAGS.IF=1 is already
possible. I.e. if there's kernel code that is confused by running with
RFLAGS.IF=1, then it's already a problem. In practice, since GIF=0 also
blocks NMIs, the only change in exposure to non-KVM code (relative to
surrounding VMRUN with STI+CLI) is exception handling code, and except for
the kvm_rebooting=1 case, all exception in the core VM-Enter/VM-Exit path
are fatal.
Use the "raw" variants to enable/disable IRQs to avoid tracing in the
"no instrumentation" code; the guest state helpers also take care of
tracing IRQ state.
Oppurtunstically document why KVM needs to do STI in the first place.
Reported-by: Doug Covelli <doug.covelli@broadcom.com> Closes: https://lore.kernel.org/all/CADH9ctBs1YPmE4aCfGPNBwA10cA8RuAk2gO7542DjMZgs4uzJQ@mail.gmail.com Fixes: f14eec0a3203 ("KVM: SVM: move more vmentry code to assembly") Cc: stable@vger.kernel.org Reviewed-by: Jim Mattson <jmattson@google.com> Link: https://lore.kernel.org/r/20250224165442.2338294-2-seanjc@google.com Signed-off-by: Sean Christopherson <seanjc@google.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Raspberry Pi is a major user of those chips and they discovered a bug -
when the end of a transfer ring segment is reached, up to four TRBs can
be prefetched from the next page even if the segment ends with link TRB
and on page boundary (the chip claims to support standard 4KB pages).
It also appears that if the prefetched TRBs belong to a different ring
whose doorbell is later rung, they may be used without refreshing from
system RAM and the endpoint will stay idle if their cycle bit is stale.
Other users complain about IOMMU faults on x86 systems, unsurprisingly.
Deal with it by using existing quirk which allocates a dummy page after
each transfer ring segment. This was seen to resolve both problems. RPi
came up with a more efficient solution, shortening each segment by four
TRBs, but it complicated the driver and they ditched it for this quirk.
Also rename the quirk and add VL805 device ID macro.
The following FFI types are replaced compared to `core::ffi`:
1. `char` type is now always mapped to `u8`, since kernel uses
`-funsigned-char` on the C code. `core::ffi` maps it to platform
default ABI, which can be either signed or unsigned.
2. `long` is now always mapped to `isize`. It's very common in the
kernel to use `long` to represent a pointer-sized integer, and in
fact `intptr_t` is a typedef of `long` in the kernel. Enforce this
mapping rather than mapping to `i32/i64` depending on platform can
save us a lot of unnecessary casts.
Signed-off-by: Gary Guo <gary@garyguo.net> Reviewed-by: Alice Ryhl <aliceryhl@google.com> Link: https://lore.kernel.org/r/20240913213041.395655-5-gary@garyguo.net
[ Moved `uaccess` changes from the next commit, since they were
irrefutable patterns that Rust >= 1.82.0 warns about. Reworded
slightly and reformatted a few documentation comments. Rebased on
top of `rust-next`. Added the removal of two casts to avoid Clippy
warnings. - Miguel ] Signed-off-by: Miguel Ojeda <ojeda@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In the last kernel cycle we migrated most of the `core::ffi` cases in
commit d072acda4862 ("rust: use custom FFI integer types"):
Currently FFI integer types are defined in libcore. This commit
creates the `ffi` crate and asks bindgen to use that crate for FFI
integer types instead of `core::ffi`.
This commit is preparatory and no type changes are made in this
commit yet.
Finish now the few remaining/new cases so that we perform the actual
remapping in the next commit as planned.
For the ACPI backend of UCSI the UCSI "registers" are just a memory copy
of the register values in an opregion. The ACPI implementation in the
BIOS ensures that the opregion contents are synced to the embedded
controller and it ensures that the registers (in particular CCI) are
synced back to the opregion on notifications. While there is an ACPI call
that syncs the actual registers to the opregion there is rarely a need to
do this and on some ACPI implementations it actually breaks in various
interesting ways.
The only reason to force a sync from the embedded controller is to poll
CCI while notifications are disabled. Only the ucsi core knows if this
is the case and guessing based on the current command is suboptimal, i.e.
leading to the following spurious assertion splat:
Thus introduce a ->poll_cci() method that works like ->read_cci() with an
additional forced sync and document that this should be used when polling
with notifications disabled. For all other backends that presumably don't
have this issue use the same implementation for both methods.
Fixes: fa48d7e81624 ("usb: typec: ucsi: Do not call ACPI _DSM method for UCSI read operations") Cc: stable <stable@kernel.org> Signed-off-by: Christian A. Ehrhardt <lk@c--e.de> Tested-by: Fedor Pchelkin <boddah8794@gmail.com> Signed-off-by: Fedor Pchelkin <boddah8794@gmail.com> Reviewed-by: Heikki Krogerus <heikki.krogerus@linux.intel.com> Link: https://lore.kernel.org/r/20250217105442.113486-2-boddah8794@gmail.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The userprog infrastructure links objects files through $(CC).
Either explicitly by manually calling $(CC) on multiple object files or
implicitly by directly compiling a source file to an executable.
The documentation at Documentation/kbuild/llvm.rst indicates that ld.lld
would be used for linking if LLVM=1 is specified.
However clang instead will use either a globally installed cross linker
from $PATH called ${target}-ld or fall back to the system linker, which
probably does not support crosslinking.
For the normal kernel build this is not an issue because the linker is
always executed directly, without the compiler being involved.
Explicitly pass --ld-path to clang so $(LD) is respected.
As clang 13.0.1 is required to build the kernel, this option is available.
Fixes: 7f3a59db274c ("kbuild: add infrastructure to build userspace programs") Cc: stable@vger.kernel.org # needs wrapping in $(cc-option) for < 6.9 Signed-off-by: Thomas Weißschuh <thomas.weissschuh@linutronix.de> Reviewed-by: Nathan Chancellor <nathan@kernel.org> Signed-off-by: Masahiro Yamada <masahiroy@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If the USB configuration is not valid, then avoid checking for
bmAttributes to prevent null pointer deference.
Cc: stable <stable@kernel.org> Fixes: 40e89ff5750f ("usb: gadget: Set self-powered based on MaxPower and bmAttributes") Signed-off-by: Prashanth K <prashanth.k@oss.qualcomm.com> Link: https://lore.kernel.org/r/20250224085604.417327-1-prashanth.k@oss.qualcomm.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
cdev->config might be NULL, so check it before dereferencing.
CC: stable <stable@kernel.org> Fixes: 40e89ff5750f ("usb: gadget: Set self-powered based on MaxPower and bmAttributes") Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com> Link: https://lore.kernel.org/r/20250220120314.3614330-1-m.szyprowski@samsung.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently the USB gadget will be set as bus-powered based solely
on whether its bMaxPower is greater than 100mA, but this may miss
devices that may legitimately draw less than 100mA but still want
to report as bus-powered. Similarly during suspend & resume, USB
gadget is incorrectly marked as bus/self powered without checking
the bmAttributes field. Fix these by configuring the USB gadget
as self or bus powered based on bmAttributes, and explicitly set
it as bus-powered if it draws more than 100mA.
During probe, the TCPC alert interrupts are getting masked to
avoid unwanted interrupts during chip setup: this is ok to do
but there is no unmasking happening at any later time, which
means that the chip will not raise any interrupt, essentially
making it not functional as, while internally it does perform
all of the intended functions, it won't signal anything to the
outside.
Still, increasing the timeout value, albeit being the most straightforward
solution, eliminates the problem: the initial PPM reset may take up to
~8000-10000ms on some Lenovo laptops. When it is reset after the above
period of time (or even if ucsi_reset_ppm() is not called overall), UCSI
works as expected.
Moreover, if the ucsi_acpi module is loaded/unloaded manually after the
system has booted, reading the CCI values and resetting the PPM works
perfectly, without any timeout. Thus it's only a boot-time issue.
The reason for this behavior is not clear but it may be the consequence
of some tricks that the firmware performs or be an actual firmware bug.
As a workaround, increase the timeout to avoid failing the UCSI
initialization prematurely.
While commit d325a1de49d6 ("usb: dwc3: gadget: Prevent losing events in
event cache") makes sure that top half(TH) does not end up overwriting the
cached events before processing them when the TH gets invoked more than one
time, returning IRQ_HANDLED results in occasional irq storm where the TH
hogs the CPU. The irq storm can be prevented by the flag before event
handler busy is cleared. Default enable interrupt moderation in all
versions which support them.
After phy initialization, some phy operations can only be executed while
in lower P states. Ensure GUSB3PIPECTL.SUSPENDENABLE and
GUSB2PHYCFG.SUSPHY are set soon after initialization to avoid blocking
phy ops.
Previously the SUSPENDENABLE bits are only set after the controller
initialization, which may not happen right away if there's no gadget
driver or xhci driver bound. Revise this to clear SUSPENDENABLE bits
only when there's mode switching (change in GCTL.PRTCAPDIR).
Syzbot once again identified a flaw in usb endpoint checking, see [1].
This time the issue stems from a commit authored by me (2eabb655a968
("usb: atm: cxacru: fix endpoint checking in cxacru_bind()")).
While using usb_find_common_endpoints() may usually be enough to
discard devices with wrong endpoints, in this case one needs more
than just finding and identifying the sufficient number of endpoints
of correct types - one needs to check the endpoint's address as well.
Since cxacru_bind() fills URBs with CXACRU_EP_CMD address in mind,
switch the endpoint verification approach to usb_check_XXX_endpoints()
instead to fix incomplete ep testing.
Currently while UDC suspends, u_ether attempts to remote wakeup
the host if there are any pending transfers. However, if remote
wakeup fails, the UDC remains suspended but the is_suspend flag
is not set. And since is_suspend flag isn't set, the subsequent
eth_start_xmit() would queue USB requests to suspended UDC.
To fix this, bail out from gether_suspend() only if remote wakeup
operation is successful.
When performing continuous unbind/bind operations on the USB drivers
available on the Renesas RZ/G2L SoC, a kernel crash with the message
"Unable to handle kernel NULL pointer dereference at virtual address"
may occur. This issue points to the usbhsc_notify_hotplug() function.
Flush the delayed work to avoid its execution when driver resources are
unavailable.
Resources should be released only after all threads that utilize them
have been destroyed.
This commit ensures that resources are not released prematurely by waiting
for the associated workqueue to complete before deallocating them.
Cc: stable <stable@kernel.org> Fixes: b9aa02ca39a4 ("usb: typec: ucsi: Add polling mechanism for partner tasks like alt mode checking") Signed-off-by: Andrei Kuchynski <akuchynski@chromium.org> Reviewed-by: Heikki Krogerus <heikki.krogerus@linux.intel.com> Link: https://lore.kernel.org/r/20250305111739.1489003-2-akuchynski@chromium.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When used on Huawei hisi platforms, Prolific Mass Storage Card Reader
which the VID:PID is in 067b:2731 might fail to enumerate at boot time
and doesn't work well with LPM enabled, combination quirks:
USB_QUIRK_DELAY_INIT + USB_QUIRK_NO_LPM
fixed the problems.
The xHC resources allocated for USB devices are not released in correct
order after resuming in case when while suspend device was reconnected.
This issue has been detected during the fallowing scenario:
- connect hub HS to root port
- connect LS/FS device to hub port
- wait for enumeration to finish
- force host to suspend
- reconnect hub attached to root port
- wake host
For this scenario during enumeration of USB LS/FS device the Cadence xHC
reports completion error code for xHC commands because the xHC resources
used for devices has not been properly released.
XHCI specification doesn't mention that device can be reset in any order
so, we should not treat this issue as Cadence xHC controller bug.
Similar as during disconnecting in this case the device resources should
be cleared starting form the last usb device in tree toward the root hub.
To fix this issue usbcore driver should call hcd->driver->reset_device
for all USB devices connected to hub which was reconnected while
suspending.
When adding support for USB3-over-USB4 tunnelling detection, a check
for an Intel-specific capability was added. This capability, which
goes by ID 206, is used without any check that we are actually
dealing with an Intel host.
As it turns out, the Cadence XHCI controller *also* exposes an
extended capability numbered 206 (for unknown purposes), but of
course doesn't have the Intel-specific registers that the tunnelling
code is trying to access. Fun follows.
The core of the problems is that the tunnelling code blindly uses
vendor-specific capabilities without any check (the Intel-provided
documentation I have at hand indicates that 192-255 are indeed
vendor-specific).
Restrict the detection code to Intel HW for real, preventing any
further explosion on my (non-Intel) HW.
Cc: stable <stable@kernel.org> Fixes: 948ce83fbb7df ("xhci: Add USB4 tunnel detection for USB3 devices on Intel hosts") Signed-off-by: Marc Zyngier <maz@kernel.org> Acked-by: Mathias Nyman <mathias.nyman@linux.intel.com> Link: https://lore.kernel.org/r/20250227194529.2288718-1-maz@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This commit was found responsible for issues with SD card recognition,
as users had to re-insert their cards in the readers and wait for a
while. As for some people the SD card was involved in the boot process
it also caused boot failures.
of_parse_phandle_with_fixed_args() requires its caller to
call into of_node_put() on the node pointer from the output
structure, but such a call is currently missing.
This patch follows commit 92191dd10730 ("net: ipv6: fix dst ref loops in
rpl, seg6 and ioam6 lwtunnels") and, on a second thought, the same patch
is also needed for ila (even though the config that triggered the issue
was pathological, but still, we don't want that to happen).
Fixes: 79ff2fc31e0f ("ila: Cache a route to translated address") Cc: Tom Herbert <tom@herbertland.com> Signed-off-by: Justin Iurman <justin.iurman@uliege.be> Link: https://patch.msgid.link/20250304181039.35951-1-justin.iurman@uliege.be Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
On MMIO devices (e.g. MT7988 or EN7581) unicast traffic received on lanX
port is flooded on all other user ports if the DSA switch is configured
without VLAN support since PORT_MATRIX in PCR regs contains all user
ports. Similar to MDIO devices (e.g. MT7530 and MT7531) fix the issue
defining default VLAN-ID 0 for MT7530 MMIO devices.
Fixes: 110c18bfed414 ("net: dsa: mt7530: introduce driver for MT7988 built-in switch") Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org> Reviewed-by: Chester A. Unal <chester.a.unal@arinc9.com> Link: https://patch.msgid.link/20250304-mt7988-flooding-fix-v1-1-905523ae83e9@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The kernel_recvmsg() function returns an int which could be either
negative error codes or the number of bytes received. The problem is
that the condition:
if (ret < sizeof(*icresp)) {
is type promoted to type unsigned long and negative values are treated
as high positive values which is success, when they should be treated as
failure. Handle invalid positive returns separately from negative
error codes to avoid this problem.
child_cfs_rq_on_list attempts to convert a 'prev' pointer to a cfs_rq.
This 'prev' pointer can originate from struct rq's leaf_cfs_rq_list,
making the conversion invalid and potentially leading to memory
corruption. Depending on the relative positions of leaf_cfs_rq_list and
the task group (tg) pointer within the struct, this can cause a memory
fault or access garbage data.
The issue arises in list_add_leaf_cfs_rq, where both
cfs_rq->leaf_cfs_rq_list and rq->leaf_cfs_rq_list are added to the same
leaf list. Also, rq->tmp_alone_branch can be set to rq->leaf_cfs_rq_list.
This adds a check `if (prev == &rq->leaf_cfs_rq_list)` after the main
conditional in child_cfs_rq_on_list. This ensures that the container_of
operation will convert a correct cfs_rq struct.
This check is sufficient because only cfs_rqs on the same CPU are added
to the list, so verifying the 'prev' pointer against the current rq's list
head is enough.
Fixes a potential memory corruption issue that due to current struct
layout might not be manifesting as a crash but could lead to unpredictable
behavior when the layout changes.
Fixes: fdaba61ef8a2 ("sched/fair: Ensure that the CFS parent is added after unthrottling") Signed-off-by: Zecheng Li <zecheng@google.com> Reviewed-and-tested-by: K Prateek Nayak <kprateek.nayak@amd.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org> Link: https://lore.kernel.org/r/20250304214031.2882646-1-zecheng@google.com Signed-off-by: Sasha Levin <sashal@kernel.org>
The parameters set by the set_params call are only applied to the block
device in the start_dev call. So if a device has already been started, a
subsequently issued set_params on that device will not have the desired
effect, and should return an error. There is an existing check for this
- set_params fails on devices in the LIVE state. But this check is not
sufficient to cover the recovery case. In this case, the device will be
in the QUIESCED or FAIL_IO states, so set_params will succeed. But this
success is misleading, because the parameters will not be applied, since
the device has already been started (by a previous ublk server). The bit
UB_STATE_USED is set on completion of the start_dev; use it to detect
and fail set_params commands which arrive too late to be applied (after
start_dev).
Signed-off-by: Uday Shankar <ushankar@purestorage.com> Fixes: 0aa73170eba5 ("ublk_drv: add SET_PARAMS/GET_PARAMS control command") Reviewed-by: Ming Lei <ming.lei@redhat.com> Link: https://lore.kernel.org/r/20250304-set_params-v1-1-17b5e0887606@purestorage.com Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Sasha Levin <sashal@kernel.org>
When I read through the TSO codes, I found out that we probably
miss initializing the tx_flags of last seg when TSO is turned
off, which means at the following points no more timestamp
(for this last one) will be generated. There are three flags
to be handled in this patch:
1. SKBTX_HW_TSTAMP
2. SKBTX_BPF
3. SKBTX_SCHED_TSTAMP
Note that SKBTX_BPF[1] was added in 6.14.0-rc2 by commit 6b98ec7e882af ("bpf: Add BPF_SOCK_OPS_TSTAMP_SCHED_CB callback")
and only belongs to net-next branch material for now. The common
issue of the above three flags can be fixed by this single patch.
This patch initializes the tx_flags to SKBTX_ANY_TSTAMP like what
the UDP GSO does to make the newly segmented last skb inherit the
tx_flags so that requested timestamp will be generated in each
certain layer, or else that last one has zero value of tx_flags
which leads to no timestamp at all.
Fixes: 4ed2d765dfacc ("net-timestamp: TCP timestamping") Signed-off-by: Jason Xing <kerneljasonxing@gmail.com> Reviewed-by: Willem de Bruijn <willemb@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
When generic_write_checks() returns zero, it means that
iov_iter_count() is zero, and there is no work to do.
Simply return success like all other filesystems do, rather than
proceeding down the write path, which today yields an -EFAULT in
generic_perform_write() via the
(fault_in_iov_iter_readable(i, bytes) == bytes) check when bytes
== 0.
Fixes: 11a347fb6cef ("exfat: change to get file size from DataLength") Reported-by: Noah <kernel-org-10@maxgrass.eu> Signed-off-by: Eric Sandeen <sandeen@redhat.com> Reviewed-by: Yuezhang Mo <Yuezhang.Mo@sony.com> Signed-off-by: Namjae Jeon <linkinjeon@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
bitmap clear loop will take long time in __exfat_free_cluster()
if data size of file/dir enty is invalid.
If cluster bit in bitmap is already clear, stop clearing bitmap go to
out of loop.
Fixes: 31023864e67a ("exfat: add fat entry operations") Reported-by: Kun Hu <huk23@m.fudan.edu.cn>, Jiaji Qin <jjtan24@m.fudan.edu.cn> Reviewed-by: Sungjong Seo <sj1557.seo@samsung.com> Signed-off-by: Namjae Jeon <linkinjeon@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
This commit fixes the condition for allocating cluster to parent
directory to avoid allocating new cluster to parent directory when
there are just enough empty directory entries at the end of the
parent directory.
Fixes: af02c72d0b62 ("exfat: convert exfat_find_empty_entry() to use dentry cache") Signed-off-by: Yuezhang Mo <Yuezhang.Mo@sony.com> Signed-off-by: Namjae Jeon <linkinjeon@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The total size calculated for EPC can overflow u64 given the added up page
for SECS. Further, the total size calculated for shmem can overflow even
when the EPC size stays within limits of u64, given that it adds the extra
space for 128 byte PCMD structures (one for each page).
Address this by pre-evaluating the micro-architectural requirement of
SGX: the address space size must be power of two. This is eventually
checked up by ECREATE but the pre-check has the additional benefit of
making sure that there is some space for additional data.
Fixes: 888d24911787 ("x86/sgx: Add SGX_IOC_ENCLAVE_CREATE") Reported-by: Dan Carpenter <dan.carpenter@linaro.org> Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org> Signed-off-by: Ingo Molnar <mingo@kernel.org> Acked-by: Dave Hansen <dave.hansen@intel.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: "H. Peter Anvin" <hpa@zytor.com> Link: https://lore.kernel.org/r/20250305050006.43896-1-jarkko@kernel.org Closes: https://lore.kernel.org/linux-sgx/c87e01a0-e7dd-4749-a348-0980d3444f04@stanley.mountain/ Signed-off-by: Sasha Levin <sashal@kernel.org>
Introduce support for standardized PHY statistics reporting in ethtool
by extending the PHYLIB framework. Add the functions
phy_ethtool_get_phy_stats() and phy_ethtool_get_link_ext_stats() to
provide a consistent interface for retrieving PHY-level and
link-specific statistics. These functions are used within the ethtool
implementation to avoid direct access to the phy_device structure
outside of the PHYLIB framework.
A new structure, ethtool_phy_stats, is introduced to standardize PHY
statistics such as packet counts, byte counts, and error counters.
Drivers are updated to include callbacks for retrieving PHY and
link-specific statistics, ensuring values are explicitly set only for
supported fields, initialized with ETHTOOL_STAT_NOT_SET to avoid
ambiguity.
Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de> Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Stable-dep-of: 637399bf7e77 ("net: ethtool: netlink: Allow NULL nlattrs when getting a phy_device") Signed-off-by: Sasha Levin <sashal@kernel.org>
Adapt linkstate_get_sqi() and linkstate_get_sqi_max() to take a
phy_device argument directly, enabling support for setups with
multiple PHYs. The previous assumption of a single PHY attached to
a net_device no longer holds.
Use ethnl_req_get_phydev() to identify the appropriate PHY device
for the operation. Update linkstate_prepare_data() and related
logic to accommodate this change, ensuring compatibility with
multi-PHY configurations.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de> Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Stable-dep-of: 637399bf7e77 ("net: ethtool: netlink: Allow NULL nlattrs when getting a phy_device") Signed-off-by: Sasha Levin <sashal@kernel.org>
Syzbot caught an "KMSAN: uninit-value" warning [1], which is caused by the
ppp driver not initializing a 2-byte header when using socket filter.
The following code can generate a PPP filter BPF program:
'''
struct bpf_program fp;
pcap_t *handle;
handle = pcap_open_dead(DLT_PPP_PPPD, 65535);
pcap_compile(handle, &fp, "ip and outbound", 0, 0);
bpf_dump(&fp, 1);
'''
Its output is:
'''
(000) ldh [2]
(001) jeq #0x21 jt 2 jf 5
(002) ldb [0]
(003) jeq #0x1 jt 4 jf 5
(004) ret #65535
(005) ret #0
'''
Wen can find similar code at the following link:
https://github.com/ppp-project/ppp/blob/master/pppd/options.c#L1680
The maintainer of this code repository is also the original maintainer
of the ppp driver.
As you can see the BPF program skips 2 bytes of data and then reads the
'Protocol' field to determine if it's an IP packet. Then it read the first
byte of the first 2 bytes to determine the direction.
The issue is that only the first byte indicating direction is initialized
in current ppp driver code while the second byte is not initialized.
For normal BPF programs generated by libpcap, uninitialized data won't be
used, so it's not a problem. However, for carefully crafted BPF programs,
such as those generated by syzkaller [2], which start reading from offset
0, the uninitialized data will be used and caught by KMSAN.
Enable the checksum option for these two endpoints in order to allow
mobile data to actually work. Without this, no packets seem to make it
through the IPA.
When a hid-steam device is removed it must clean up the client_hdev used for
intercepting hidraw access. This can lead to scheduling deferred work to
reattach the input device. Though the cleanup cancels the deferred work, this
was done before the client_hdev itself is cleaned up, so it gets rescheduled.
This patch fixes the ordering to make sure the deferred work is properly
canceled.
We need to be able to do both MMIO and DSB based pipe/plane
programming. To that end plumb the 'dsb' all way from the top
into the plane commit hooks.
The compiler appears smart enough to combine the branches from
all the back-to-back register writes into a single branch.
So the generated asm ends up looking more or less like this:
plane_hook()
{
if (dsb) {
intel_dsb_reg_write();
intel_dsb_reg_write();
...
} else {
intel_de_write_fw();
intel_de_write_fw();
...
}
}
which seems like a reasonably efficient way to do this.
An alternative I was also considering is some kind of closure
(register write function + display vs. dsb pointer passed to it).
That does result is smaller code as there are no branches anymore,
but having each register access go via function pointer sounds
less efficient.
Not that I actually measured the overhead of either approach yet.
Also the reg_rw tracepoint seems to be making a huge mess of the
generated code for the mmio path. And additionally there's some
kind of IS_GSI_REG() hack in __raw_uncore_read() which ends up
generating a pointless branch for every mmio register access.
So looks like there might be quite a bit of room for improvement
in the mmio path still.
During the initialization of ptp, hclge_ptp_get_cycle might return an error
and returned directly without unregister clock and free it. To avoid that,
call hclge_ptp_destroy_clock to unregist and free clock if
hclge_ptp_get_cycle failed.
Fixes: 8373cd38a888 ("net: hns3: change the method of obtaining default ptp cycle") Signed-off-by: Peiyang Wang <wangpeiyang1@huawei.com> Signed-off-by: Jijie Shao <shaojijie@huawei.com> Reviewed-by: Simon Horman <horms@kernel.org> Link: https://patch.msgid.link/20250228105258.1243461-1-shaojijie@huawei.com Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Partially revert commit b71724147e73 ("be2net: replace polling with
sleeping in the FW completion path") w.r.t mcc mutex it introduces and the
use of usleep_range. The be2net be_ndo_bridge_getlink() callback is
called with rcu_read_lock, so this code has been broken for a long time.
Both the mutex_lock and the usleep_range can cause the issue Ian Kumlien
reported[1]. The call path is:
be_ndo_bridge_getlink -> be_cmd_get_hsw_config -> be_mcc_notify_wait ->
be_mcc_wait_compl -> usleep_range()
Tested-by: Ian Kumlien <ian.kumlien@gmail.com> Fixes: b71724147e73 ("be2net: replace polling with sleeping in the FW completion path") Signed-off-by: Nikolay Aleksandrov <razor@blackwall.org> Link: https://patch.msgid.link/20250227164129.1201164-1-razor@blackwall.org Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Reported-by: syzbot+da65c993ae113742a25f@syzkaller.appspotmail.com Closes: https://lore.kernel.org/netdev/67c020c0.050a0220.222324.0011.GAE@google.com/T/#u Signed-off-by: Eric Dumazet <edumazet@google.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
The module parameter defines number of iso packets per one URB. User is
allowed to set any value to the parameter of type int, which can lead to
various kinds of weird and incorrect behavior like integer overflows,
truncations, etc. Number of packets should be a small non-negative number.
Since this parameter is read-only, its value can be validated on driver
probe.
When firmware traces are enabled, the firmware dumps 48-bit timestamps
for each trace as two 32-bit values, highest 32 bits (of which only 16
useful) first.
The driver was reassembling them the other way round i.e. interpreting
the first value in memory as the lowest 32 bits, and the second value
as the highest 32 bits (then truncated to 16 bits).
Due to this, firmware trace dumps showed very large timestamps even for
traces recorded shortly after GPU boot. The timestamps in these dumps
would also sometimes jump backwards because of the truncation.
Example trace dumped after loading the powervr module and enabling
firmware traces, where each line is commented with the timestamp value
in hexadecimal to better show both issues:
Leading zero bits are sent on the bus before the temperature value is
transmitted. If any of these bits are high, the connection might be
unstable or there could be no AD7314 / ADT730x (or compatible) at all.
Return -EIO in that case.
I could not find a single table that has the values currently present in
the table, change it to the actual values that can be found in [1]/[2]
and [3] (page 15 column 2)
The `pmbus_identify()` function fails to correctly determine the number
of supported pages on PMBus devices. This occurs because `info->pages`
is implicitly zero-initialised, and `pmbus_set_page()` does not perform
writes to the page register if `info->pages` is not yet initialised.
Without this patch, `info->pages` is always set to the maximum after
scanning.
This patch initialises `info->pages` to `PMBUS_PAGES` before the probing
loop, enabling `pmbus_set_page()` writes to make it out onto the bus
correctly identifying the number of pages. `PMBUS_PAGES` seemed like a
reasonable non-zero number because that's the current result of the
identification process.
Commit a63fbed776c7 ("perf/tracing/cpuhotplug: Fix locking order")
placed pmus_lock inside pmus_srcu, this makes perf_pmu_unregister()
trip lockdep.
Move the locking about such that only pmu_idr and pmus (list) are
modified while holding pmus_lock. This avoids doing synchronize_srcu()
while holding pmus_lock and all is well again.
del_vqs() frees virtqueues, therefore cfv->vq_tx pointer should be checked
for NULL before calling it, not cfv->vdev. Also the current implementation
is redundant because the pointer cfv->vdev is dereferenced before it is
checked for NULL.
Fix this by checking cfv->vq_tx for NULL instead of cfv->vdev before
calling del_vqs().
Fixes: 0d2e1a2926b1 ("caif_virtio: Introduce caif over virtio") Signed-off-by: Vitaliy Shevtsov <v.shevtsov@mt-integration.ru> Reviewed-by: Gerhard Engleder <gerhard@engleder-embedded.com> Link: https://patch.msgid.link/20250227184716.4715-1-v.shevtsov@mt-integration.ru Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
In __udp_gso_segment the skb destructor is removed before segmenting the
skb but the socket reference is kept as-is. This is an issue if the
original skb is later orphaned as we can hit the following bug:
The above can happen following a sequence of events when using
OpenVSwitch, when an OVS_ACTION_ATTR_USERSPACE action precedes an
OVS_ACTION_ATTR_OUTPUT action:
1. OVS_ACTION_ATTR_USERSPACE is handled (in do_execute_actions): the skb
goes through queue_gso_packets and then __udp_gso_segment, where its
destructor is removed.
2. The segments' data are copied and sent to userspace.
3. OVS_ACTION_ATTR_OUTPUT is handled (in do_execute_actions) and the
same original skb is sent to its path.
4. If it later hits skb_orphan, we hit the bug.
Fix this by also removing the reference to the socket in
__udp_gso_segment.
Fixes: ad405857b174 ("udp: better wmem accounting on gso") Signed-off-by: Antoine Tenart <atenart@kernel.org> Link: https://patch.msgid.link/20250226171352.258045-1-atenart@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
In commit 1e9c708dc3ae ("ALSA: hda/tas2781: Add new quirk for Lenovo,
ASUS, Dell projects") Baojun adds a bunch of projects to the file,
including for the Ally X. Turns out the initial Ally X was not sorted
properly, so the kernel had 2 quirks for it.
The previous quirk overrode the new one due to being earlier and they
are different. When AB testing, the normal pin fixup seems to work ok
but causes a bit of a minor popping. Given the other config is more
complicated and may cause undefined behavior, revert it.
The order in which queue->cmd and rcv_state are updated is crucial.
If these assignments are reordered by the compiler, the worker might not
get queued in nvmet_tcp_queue_response(), hanging the IO. to enforce the
the correct reordering, set rcv_state using smp_store_release().
Fixes: bdaf13279192 ("nvmet-tcp: fix a segmentation fault during io parsing error") Signed-off-by: Meir Elisha <meir.elisha@volumez.com> Reviewed-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Keith Busch <kbusch@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
nvme_tcp_recv_pdu() doesn't check the validity of the header length.
When header digests are enabled, a target might send a packet with an
invalid header length (e.g. 255), causing nvme_tcp_verify_hdgst()
to access memory outside the allocated area and cause memory corruptions
by overwriting it with the calculated digest.
Fix this by rejecting packets with an unexpected header length.
Previously, the NVMe/TCP host driver did not handle the C2HTermReq PDU,
instead printing "unsupported pdu type (3)" when received. This patch adds
support for processing the C2HTermReq PDU, allowing the driver
to print the Fatal Error Status field.
Example of output:
nvme nvme4: Received C2HTermReq (FES = Invalid PDU Header Field)
Initialize .owner field of force_poll_sync_fops to THIS_MODULE in order to
prevent btusb from being unloaded while its operations are in use.
Fixes: 800fe5ec302e ("Bluetooth: btusb: Add support for queuing during polling interval") Signed-off-by: Salah Triki <salah.triki@gmail.com> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
nouveau tries to load some firmware during suspend that it loaded
earlier, but with fw caching disabled it hangs suspend, so just rely on
FW cache enabling instead of working around it in the driver.
Call drm_client_setup() to run the kernel's default client setup
for DRM. Set fbdev_probe in struct drm_driver, so that the client
setup can start the common fbdev client.
The nouveau driver specifies a preferred color mode depending on
the available video memory, with a default of 32. Adapt this for
the new client interface.
Rework fbdev probing to support fbdev_probe in struct drm_driver
and reimplement the old fb_probe callback on top of it. Provide an
initializer macro for struct drm_driver that sets the callback
according to the kernel configuration.
This change allows the common fbdev client to run on top of TTM-
based DRM drivers.
DRM may support multiple in-kernel clients that run as soon as a DRM
driver has been registered. To select the client(s) in a single place,
introduce drm_client_setup().
Drivers that call the new helper automatically instantiate the kernel's
configured default clients. Only fbdev emulation is currently supported.
Later versions can add support for DRM-based logging, a boot logo or even
a console.
Some drivers handle the color mode for clients internally. Provide the
helper drm_client_setup_with_color_mode() for them.
Using the new interface requires the driver to select
DRM_CLIENT_SELECTION in its Kconfig. For now this only enables the
client-setup helpers if the fbdev client has been configured by the
user. A future patchset will further modularize client support and
rework DRM_CLIENT_SELECTION to select the correct dependencies for
all its clients.
v5:
- add CONFIG_DRM_CLIENT_SELECTION und DRM_CLIENT_SETUP
v4:
- fix docs for drm_client_setup_with_fourcc() (Geert)
v3:
- fix build error
v2:
- add drm_client_setup_with_fourcc() (Laurent)
- push default-format handling into actual clients
Add an fbdev client that can work with any memory manager. The
client implementation is the same as existing code in fbdev-dma or
fbdev-shmem.
Provide struct drm_driver.fbdev_probe for the new client to allocate
the surface GEM buffer. The new callback replaces fb_probe of struct
drm_fb_helper_funcs, which does the same.
To use the new client, DRM drivers set fbdev_probe in their struct
drm_driver instance and call drm_fbdev_client_setup(). Probing and
creating the fbdev surface buffer is now independent from the other
operations in struct drm_fb_helper. For the pixel format, the fbdev
client either uses a specified format, the value in preferred_depth
or 32-bit RGB.
v2:
- test for struct drm_fb_helper.funcs for NULL (Sui)
- respect struct drm_mode_config.preferred_depth for default format
The color mode as specified on the kernel command line gives the user's
preferred color depth and number of bits per pixel. Move the
color-mode-to-format conversion from fbdev helpers into a 4CC helper,
so that it can be shared among DRM clients.
If there's any vendor-specific element in the subelements
then the outer element parsing must not parse any vendor
element at all. This isn't implemented correctly now due
to parsing into the pointers and then overriding them, so
explicitly skip vendor elements if any exist in the sub-
elements (non-transmitted profile or per-STA profile).
The code is erroneously applying the non-inheritance element
to the inner elements rather than the outer, which is clearly
completely wrong. Fix it by finding the MLE basic element at
the beginning, and then applying the non-inheritance for the
outer parsing.
While at it, do some general cleanups such as not allowing
callers to try looking for a specific non-transmitted BSS
and link at the same time.
Fixes: 45ebac4f059b ("wifi: mac80211: Parse station profile from association response") Reviewed-by: Ilan Peer <ilan.peer@intel.com> Reviewed-by: Miriam Rachel Korenblit <miriam.rachel.korenblit@intel.com> Link: https://patch.msgid.link/20250221112451.b46d42f45b66.If5b95dc3c80208e0c62d8895fb6152aa54b6620b@changeid Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Add support for parsing an ML element of type EPCS priority
access, which can optionally be included in EHT protected action
frames used to configure EPCS.
All the callers assume nvme_map_user_request() frees the request on a
failure. This wasn't happening on invalid metadata or io_uring command
flags, so we've been leaking those requests.
Fixes: 23fd22e55b767b ("nvme: wire up fixed buffer support for nvme passthrough") Fixes: 7c2fd76048e95d ("nvme: fix metadata handling in nvme-passthrough") Reviewed-by: Damien Le Moal <dlemoal@kernel.org> Reviewed-by: Kanchan Joshi <joshi.k@samsung.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Keith Busch <kbusch@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
If the device supports SGLs, use these for all user requests. This
format encodes the expected transfer length so it can catch short buffer
errors in a user command, whether it occurred accidently or maliciously.
For controllers that support SGL data mode, this is a viable mitigation
to CVE-2023-6238. For controllers that don't support SGLs, log a warning
in the passthrough path since not having the capability can corrupt
data if the interface is not used correctly.
Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Keith Busch <kbusch@kernel.org>
Stable-dep-of: 00817f0f1c45 ("nvme-ioctl: fix leaked requests on mapping error") Signed-off-by: Sasha Levin <sashal@kernel.org>
Supporting this mode allows creating and merging multi-segment metadata
requests that wouldn't be possible otherwise. It also allows directly
using user space requests that straddle physically discontiguous pages.
Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Keith Busch <kbusch@kernel.org>
Stable-dep-of: 00817f0f1c45 ("nvme-ioctl: fix leaked requests on mapping error") Signed-off-by: Sasha Levin <sashal@kernel.org>
The sorting of VMAs by size in commit 7d442a33bfe8 ("binfmt_elf: Dump
smaller VMAs first in ELF cores") breaks elfutils[1]. Instead, sort
based on the setting of the new sysctl, core_sort_vma, which defaults
to 0, no sorting.
Reported-by: Michael Stapelberg <michael@stapelberg.ch> Closes: https://lore.kernel.org/all/20250218085407.61126-1-michael@stapelberg.de/ [1] Fixes: 7d442a33bfe8 ("binfmt_elf: Dump smaller VMAs first in ELF cores") Signed-off-by: Kees Cook <kees@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The system can experience a random crash a few minutes after the driver is
removed. This issue occurs due to improper handling of memory freeing in
the ishtp_hid_remove() function.
The function currently frees the `driver_data` directly within the loop
that destroys the HID devices, which can lead to accessing freed memory.
Specifically, `hid_destroy_device()` uses `driver_data` when it calls
`hid_ishtp_set_feature()` to power off the sensor, so freeing
`driver_data` beforehand can result in accessing invalid memory.
This patch resolves the issue by storing the `driver_data` in a temporary
variable before calling `hid_destroy_device()`, and then freeing the
`driver_data` after the device is destroyed.
During the `rmmod` operation for the `intel_ishtp_hid` driver, a
use-after-free issue can occur in the hid_ishtp_cl_remove() function.
The function hid_ishtp_cl_deinit() is called before ishtp_hid_remove(),
which can lead to accessing freed memory or resources during the
removal process.
Additionally, ishtp_hid_remove() is a HID level power off, which should
occur before the ISHTP level disconnect.
This patch resolves the issue by reordering the calls in
hid_ishtp_cl_remove(). The function ishtp_hid_remove() is now
called before hid_ishtp_cl_deinit().
As reported by the kernel test robot, the following warning occurs:
>> drivers/hid/hid-google-hammer.c:261:36: warning: 'cbas_ec_acpi_ids' defined but not used [-Wunused-const-variable=]
261 | static const struct acpi_device_id cbas_ec_acpi_ids[] = {
| ^~~~~~~~~~~~~~~~
The 'cbas_ec_acpi_ids' array is only used when CONFIG_ACPI is enabled.
Wrapping its definition and 'MODULE_DEVICE_TABLE' in '#ifdef CONFIG_ACPI'
prevents a compiler warning when ACPI is disabled.
Fixes: eb1aac4c8744f75 ("HID: google: add support tablet mode switch for Whiskers") Reported-by: kernel test robot <lkp@intel.com> Closes: https://lore.kernel.org/oe-kbuild-all/202501201141.jctFH5eB-lkp@intel.com/ Signed-off-by: Yu-Chun Lin <eleanor15x@gmail.com> Signed-off-by: Jiri Kosina <jkosina@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The TSO preparation assumed that the skb head contained the headers
while the rest of the data was in the fragments. Since this is not
always true, e.g., it is possible that the data was linearised, modify
the TSO preparation to start the data processing after the network
headers.
Fixes: 7f5e3038f029 ("wifi: iwlwifi: map entire SKB when sending AMSDUs") Signed-off-by: Ilan Peer <ilan.peer@intel.com> Reviewed-by: Benjamin Berg <benjamin.berg@intel.com> Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com> Link: https://patch.msgid.link/20250209143303.75769a4769bf.Iaf79e8538093cdf8c446c292cc96164ad6498f61@changeid Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
There's no guarantee here that the file is always with a
NUL-termination, so reading the string may read beyond the
end of the TLV. If that's the last TLV in the file, it can
perhaps even read beyond the end of the file buffer.
Fix that by limiting the print format to the size of the
buffer we have.
If the firmware fails to start the session protection, then we
do call iwl_mvm_roc_finished() here, but that won't do anything
at all because IWL_MVM_STATUS_ROC_P2P_RUNNING was never set.
Set IWL_MVM_STATUS_ROC_P2P_RUNNING in the failure/stop path.
If it started successfully before, it's already set, so that
doesn't matter, and if it didn't start it needs to be set to
clean up.
Not doing so will lead to a WARN_ON() later on a fresh remain-
on-channel, since the link is already active when activated as
it was never deactivated.
If a folio has an increased reference count, folio_try_get() will acquire
it, perform necessary operations, and then release it. In the case of a
poisoned folio without an elevated reference count (which is unlikely for
memory-failure), folio_try_get() will simply bypass it.
Therefore, relocate the folio_try_get() function, responsible for checking
and acquiring this reference count at first.
Link: https://lkml.kernel.org/r/20250217014329.3610326-3-mawupeng1@huawei.com Signed-off-by: Ma Wupeng <mawupeng1@huawei.com> Acked-by: David Hildenbrand <david@redhat.com> Acked-by: Miaohe Lin <linmiaohe@huawei.com> Cc: Michal Hocko <mhocko@suse.com> Cc: Naoya Horiguchi <nao.horiguchi@gmail.com> Cc: Oscar Salvador <osalvador@suse.de> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit b15c87263a69 ("hwpoison, memory_hotplug: allow hwpoisoned pages to
be offlined) add page poison checks in do_migrate_range in order to make
offline hwpoisoned page possible by introducing isolate_lru_page and
try_to_unmap for hwpoisoned page. However folio lock must be held before
calling try_to_unmap. Add it to fix this problem.
Warning will be produced if folio is not locked during unmap:
When handling faults for anon shmem finish_fault() will attempt to install
ptes for the entire folio. Unfortunately if it encounters a single
non-pte_none entry in that range it will bail, even if the pte that
triggered the fault is still pte_none. When this situation happens the
fault will be retried endlessly never making forward progress.
This patch fixes this behavior and if it detects that a pte in the range
is not pte_none it will fall back to setting a single pte.
[bgeffon@google.com: tweak whitespace] Link: https://lkml.kernel.org/r/20250227133236.1296853-1-bgeffon@google.com Link: https://lkml.kernel.org/r/20250226162341.915535-1-bgeffon@google.com Fixes: 43e027e41423 ("mm: memory: extend finish_fault() to support large folio") Signed-off-by: Brian Geffon <bgeffon@google.com> Suggested-by: Baolin Wang <baolin.wang@linux.alibaba.com> Reported-by: Marek Maslanka <mmaslanka@google.com> Cc: Hugh Dickins <hughd@google.com> Cc: David Hildenbrand <david@redhat.com> Cc: Hugh Dickens <hughd@google.com> Cc: Kefeng Wang <wangkefeng.wang@huawei.com> Cc: Matthew Wilcow (Oracle) <willy@infradead.org> Cc: Suren Baghdasaryan <surenb@google.com> Cc: Zi Yan <ziy@nvidia.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix callers that previously skipped calling arch_sync_kernel_mappings() if
an error occurred during a pgtable update. The call is still required to
sync any pgtable updates that may have occurred prior to hitting the error
condition.
These are theoretical bugs discovered during code review.
Link: https://lkml.kernel.org/r/20250226121610.2401743-1-ryan.roberts@arm.com Fixes: 2ba3e6947aed ("mm/vmalloc: track which page-table levels were modified") Fixes: 0c95cba49255 ("mm: apply_to_pte_range warn and fail if a large pte is encountered") Signed-off-by: Ryan Roberts <ryan.roberts@arm.com> Reviewed-by: Anshuman Khandual <anshuman.khandual@arm.com> Reviewed-by: Catalin Marinas <catalin.marinas@arm.com> Cc: Christop Hellwig <hch@infradead.org> Cc: "Uladzislau Rezki (Sony)" <urezki@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Patch series "mm: memory_failure: unmap poisoned folio during migrate
properly", v3.
Fix two bugs during folio migration if the folio is poisoned.
This patch (of 3):
Commit 6da6b1d4a7df ("mm/hwpoison: convert TTU_IGNORE_HWPOISON to
TTU_HWPOISON") introduce TTU_HWPOISON to replace TTU_IGNORE_HWPOISON in
order to stop send SIGBUS signal when accessing an error page after a
memory error on a clean folio. However during page migration, anon folio
must be set with TTU_HWPOISON during unmap_*(). For pagecache we need
some policy just like the one in hwpoison_user_mappings to set this flag.
So move this policy from hwpoison_user_mappings to unmap_poisoned_folio to
handle this warning properly.
Warning will be produced during unamp poison folio with the following log:
The remainder of vma_modify() relies upon the vmg state remaining pristine
after a merge attempt.
Usually this is the case, however in the one edge case scenario of a merge
attempt failing not due to the specified range being unmergeable, but
rather due to an out of memory error arising when attempting to commit the
merge, this assumption becomes untrue.
This results in vmg->start, end being modified, and thus the proceeding
attempts to split the VMA will be done with invalid start/end values.
Thankfully, it is likely practically impossible for us to hit this in
reality, as it would require a maple tree node pre-allocation failure that
would likely never happen due to it being 'too small to fail', i.e. the
kernel would simply keep retrying reclaim until it succeeded.
However, this scenario remains theoretically possible, and what we are
doing here is wrong so we must correct it.
The safest option is, when this scenario occurs, to simply give up the
operation. If we cannot allocate memory to merge, then we cannot allocate
memory to split either (perhaps moreso!).
Any scenario where this would be happening would be under very extreme
(likely fatal) memory pressure, so it's best we give up early.
So there is no doubt it is appropriate to simply bail out in this
scenario.
However, in general we must if at all possible never assume VMG state is
stable after a merge attempt, since merge operations update VMG fields.
As a result, additionally also make this clear by storing start, end in
local variables.
The issue was reported originally by syzkaller, and by Brad Spengler (via
an off-list discussion), and in both instances it manifested as a
triggering of the assert:
VM_WARN_ON_VMG(start >= end, vmg);
In vma_merge_existing_range().
It seems at least one scenario in which this is occurring is one in which
the merge being attempted is due to an madvise() across multiple VMAs
which looks like this:
start end
|<------>|
|----------|------|
| vma | next |
|----------|------|
When madvise_walk_vmas() is invoked, we first find vma in the above
(determining prev to be equal to vma as we are offset into vma), and then
enter the loop.
We determine the end of vma that forms part of the range we are
madvise()'ing by setting 'tmp' to this value:
Where the visit() function pointer in this instance is
madvise_vma_behavior().
As observed in syzkaller reports, it is ultimately madvise_update_vma()
that is invoked, calling vma_modify_flags_name() and vma_modify() in turn.
Then, in vma_modify(), we attempt the merge:
merged = vma_merge_existing_range(vmg);
if (merged)
return merged;
We invoke this with vmg->start, end set to start, tmp as such:
start tmp
|<--->|
|----------|------|
| vma | next |
|----------|------|
We find ourselves in the merge right scenario, but the one in which we
cannot remove the middle (we are offset into vma).
Here we have a special case where vmg->start, end get set to perhaps
unintuitive values - we intended to shrink the middle VMA and expand the
next.
This means vmg->start, end are set to... vma->vm_start, start.
Now the commit_merge() fails, and vmg->start, end are left like this.
This means we return to the rest of vma_modify() with vmg->start, end
(here denoted as start', end') set as:
start' end'
|<-->|
|----------|------|
| vma | next |
|----------|------|
So we now erroneously try to split accordingly. This is where the
unfortunate stuff begins.
We start with:
/* Split any preceding portion of the VMA. */
if (vma->vm_start < vmg->start) {
...
}
This doesn't trigger as we are no longer offset into vma at the start.
But then we invoke:
/* Split any trailing portion of the VMA. */
if (vma->vm_end > vmg->end) {
...
}
Which does get invoked. This leaves us with:
start' end'
|<-->|
|----|-----|------|
| vma| new | next |
|----|-----|------|
We then return ultimately to madvise_walk_vmas(). Here 'new' is unknown,
and putting back the values known in this function we are faced with:
start tmp end
| | |
|----|-----|------|
| vma| new | next |
|----|-----|------|
prev
Then:
start = tmp;
So:
start end
| |
|----|-----|------|
| vma| new | next |
|----|-----|------|
prev
The following code does not cause anything to happen:
if (prev && start < prev->vm_end)
start = prev->vm_end;
if (start >= end)
break;
And then we invoke:
if (prev)
vma = find_vma(mm, prev->vm_end);
Which is where a problem occurs - we don't know about 'new' so we
essentially look for the vma after prev, which is new, whereas we actually
intended to discover next!
So we end up with:
start end
| |
|----|-----|------|
|prev| vma | next |
|----|-----|------|
And we have successfully bypassed all of the checks madvise_walk_vmas()
has to ensure early exit should we end up moving out of range.
Where start == tmp. That is, a zero range. This is not good.
We invoke visit() which is madvise_vma_behavior() which does not check the
range (for good reason, it assumes all checks have been done before it was
called), which in turn finally calls madvise_update_vma().
The madvise_update_vma() function calls vma_modify_flags_name() in turn,
which ultimately invokes vma_modify() with... start == end.
vma_modify() calls vma_merge_existing_range() and finally we hit:
VM_WARN_ON_VMG(start >= end, vmg);
Which triggers, as start == end.
While it might be useful to add some CONFIG_DEBUG_VM asserts in these
instances to catch this kind of error, since we have just eliminated any
possibility of that happening, we will add such asserts separately as to
reduce churn and aid backporting.
Local variable compact_result created at:
__alloc_pages_slowpath+0x66/0x16c0 mm/page_alloc.c:4218
__alloc_frozen_pages_noprof+0xa4c/0xe00 mm/page_alloc.c:4752
The utf16_le_to_7bit function claims to, naively, convert a UTF-16
string to a 7-bit ASCII string. By naively, we mean that it:
* drops the first byte of every character in the original UTF-16 string
* checks if all characters are printable, and otherwise replaces them
by exclamation mark "!".
This means that theoretically, all characters outside the 7-bit ASCII
range should be replaced by another character. Examples:
* lower-case alpha (ɒ) 0x0252 becomes 0x52 (R)
* ligature OE (œ) 0x0153 becomes 0x53 (S)
* hangul letter pieup (ㅂ) 0x3142 becomes 0x42 (B)
* upper-case gamma (Ɣ) 0x0194 becomes 0x94 (not printable) so gets
replaced by "!"
The result of this conversion for the GPT partition name is passed to
user-space as PARTNAME via udev, which is confusing and feels questionable.
However, there is a flaw in the conversion function itself. By dropping
one byte of each character and using isprint() to check if the remaining
byte corresponds to a printable character, we do not actually guarantee
that the resulting character is 7-bit ASCII.
This happens because we pass 8-bit characters to isprint(), which
in the kernel returns 1 for many values > 0x7f - as defined in ctype.c.
This results in many values which should be replaced by "!" to be kept
as-is, despite not being valid 7-bit ASCII. Examples:
* e with acute accent (é) 0x00E9 becomes 0xE9 - kept as-is because
isprint(0xE9) returns 1.
* euro sign (€) 0x20AC becomes 0xAC - kept as-is because isprint(0xAC)
returns 1.
This way has broken pyudev utility[1], fixes it by using a mask of 7 bits
instead of 8 bits before calling isprint.
projects / thirdparty / kernel / stable.git / log
