Instead of kzalloc and kcalloc, we can use a flex array to reduce to a
single allocation.
Also added __counted_by() for extra possible analysis.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Reviewed-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Link: https://patch.msgid.link/20260304043420.14151-1-rosenp@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
SMB2 READ response validation in cifs_readv_receive() and
handle_read_data() checks data_offset + data_len against the received
buffer length. Both values are attacker-controlled fields from the
server response and are stored as unsigned int, so the addition can
wrap before the bounds check:
fs/smb/client/transport.c:1259
if (!use_rdma_mr && (data_offset + data_len > buflen))
fs/smb/client/smb2ops.c:4839
else if (buf_len >= data_offset + data_len)
A malicious SMB server can use this to bypass validation. In the
non-encrypted receive path the client attempts an oversized socket
read and stalls for the SMB response timeout (180 seconds) before
reconnecting. In the SMB3 encrypted path, runtime testing shows the
malformed length can reach copy_to_iter() in handle_read_data() with
attacker-controlled size, where usercopy hardening stops the oversized
copy before bytes reach userspace.
Guard both call sites with check_add_overflow(), which is already
used elsewhere in this subsystem (smb2pdu.c). On overflow, treat the
response as malformed and reject with -EIO.
Signed-off-by: Jeremy Erazo <mendozayt13@gmail.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
-----BEGIN PGP SIGNATURE-----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=bV7l
-----END PGP SIGNATURE-----
Merge tag 'audit-pr-20260513' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/audit
Pull audit fixes from Paul Moore:
- Correctly log the inheritable capabilities
- Honor AUDIT_LOCKED in the AUDIT_TRIM and AUDIT_MAKE_EQUIV commands
* tag 'audit-pr-20260513' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/audit:
audit: enforce AUDIT_LOCKED for AUDIT_TRIM and AUDIT_MAKE_EQUIV
audit: fix incorrect inheritable capability in CAPSET records
atcphy_probe_switch() and atcphy_probe_mux() discard the pointers
returned by typec_switch_register() and typec_mux_register(). The
platform driver has no .remove callback, so when the driver unbinds
(e.g. via sysfs unbind) neither typec_switch_unregister() nor
typec_mux_unregister() is called. The framework reference taken in
typec_switch_register() (device_initialize() + device_add() in
drivers/usb/typec/mux.c) is therefore never dropped and the
typec_switch_dev / typec_mux_dev objects stay live forever, with
their sysfs entries under the typec_mux class also left behind. A
subsequent rebind cannot recreate them with the same fwnode-derived
name.
Save the registered handles and unregister them through
devm_add_action_or_reset() so framework registration is torn down
in step with the driver's other devm-managed state. While here,
drop struct apple_atcphy::sw and ::mux: they were declared with the
consumer-side types (typec_switch *, typec_mux *) instead of the
provider-side types and were never assigned.
Scope of the fix
================
This patch fixes the registration leak only. It does not close the
use-after-free window that arises when a consumer that obtained a
reference via fwnode_typec_switch_get() / fwnode_typec_mux_get()
outlives the provider unbind: such consumers keep the underlying
typec_switch_dev / typec_mux_dev alive past device_unregister(),
and a later typec_switch_set() / typec_mux_set() still invokes the
registered atcphy_sw_set() / atcphy_mux_set(), which dereferences
the freed apple_atcphy through typec_{switch,mux}_get_drvdata().
On Apple Silicon the relevant consumers are the typec port and the
cd321x controller registered by drivers/usb/typec/tipd/core.c.
Cable plug / orientation events and alt-mode transitions trigger
the .set callbacks via:
tps6598x_interrupt() drivers/usb/typec/tipd/core.c
tps6598x_handle_plug_event()
tps6598x_connect()/_disconnect()
typec_set_orientation() drivers/usb/typec/class.c
typec_switch_set(port->sw) drivers/usb/typec/mux.c
atcphy_sw_set() drivers/phy/apple/atc.c
cd321x_update_work() drivers/usb/typec/tipd/core.c
cd321x_typec_update_mode()
typec_mux_set(cd321x->mux) drivers/usb/typec/mux.c
atcphy_mux_set() drivers/phy/apple/atc.c
Closing that window requires framework support for invalidating
consumer-held references on provider unbind. The same
consumer-survives-provider pattern has been discussed for the PHY
framework [1] and is out of scope here.
[1] https://lore.kernel.org/linux-phy/aZejMSJ9qqRWb2pX@google.com/
Fixes: 8e98ca1e74 ("phy: apple: Add Apple Type-C PHY")
Signed-off-by: David Carlier <devnexen@gmail.com>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Tested-by: Joshua Peisach <jpeisach@ubuntu.com>
Link: https://lkml.kernel.org/r/6ec1ed08328340db42655287afd5fa4067316b11.camel@perches.com
Link: https://patch.msgid.link/20260508201958.30060-1-devnexen@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Introduce support for EcoNet PCIe PHY controllers found in EN751221
and EN7528 SoCs, these SoCs are not identical but are similar, each
having one Gen1 port, and one Gen1/Gen2 port.
Co-developed-by: Ahmed Naseef <naseefkm@gmail.com>
Signed-off-by: Ahmed Naseef <naseefkm@gmail.com>
[cjd@cjdns.fr: add EN751221 support and refactor for clarity]
Signed-off-by: Caleb James DeLisle <cjd@cjdns.fr>
Link: https://patch.msgid.link/20260425173642.406089-3-cjd@cjdns.fr
Signed-off-by: Vinod Koul <vkoul@kernel.org>
EN751221 and EN7528 SoCs have two PCIe slots, and each one has a PHY
which behaves slightly differently because one slot is Gen1/Gen2 while
the other is Gen1 only.
Signed-off-by: Caleb James DeLisle <cjd@cjdns.fr>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
Link: https://patch.msgid.link/20260425173642.406089-2-cjd@cjdns.fr
Signed-off-by: Vinod Koul <vkoul@kernel.org>
When O_TMPFILE is set, the open mode needs to be passed to the kernel as
per the documentation. Currently this is not done.
Instead of checking for O_TMPFILE explicitly and making the conditionals
more complex, just always pass the mode to the kernel. If no value was
passed the mode will be garbage, but the kernel will ignore it anyways.
Fixes: a7604ba149 ("tools/nolibc/sys: make open() take a vararg on the 3rd argument")
Suggested-by: Willy Tarreau <w@1wt.eu>
Link: https://lore.kernel.org/lkml/afRfjdovT6pNtwtP@1wt.eu/
Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Acked-by: Willy Tarreau <w@1wt.eu>
Link: https://patch.msgid.link/20260514-nolibc-open-tmpfile-v2-3-b4c6c5efa266@weissschuh.net
This logic is duplicated and some upcoming extensions would require even
more duplicated logic.
Move it into a macro to avoid the duplication and allow cleaner changes.
Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Acked-by: Willy Tarreau <w@1wt.eu>
Link: https://patch.msgid.link/20260514-nolibc-open-tmpfile-v2-2-b4c6c5efa266@weissschuh.net
This logic is duplicated and its current form will be in the way of some
upcoming simplificiations.
Move it into a macro to avoid the duplication and enable some cleanups.
Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Acked-by: Willy Tarreau <w@1wt.eu>
Link: https://patch.msgid.link/20260514-nolibc-open-tmpfile-v2-1-b4c6c5efa266@weissschuh.net
The RK3528 has a single USB2PHY with a otg and host port.
Add support for the RK3528 variant of USB2PHY.
PHY tuning for RK3528:
- Turn off differential receiver in suspend mode to save power
consumption.
- Set HS eye-height to 400mV instead of default 450mV.
- Choose the Tx fs/ls data as linestate from TX driver for otg port
which uses dwc3 controller to improve fs/ls devices compatibility with
long cables.
Undocumented magic-values are based on the linux-stan-6.1-rkr5 tag of
the vendor-kernel.
Signed-off-by: Jianwei Zheng <jianwei.zheng@rock-chips.com>
Signed-off-by: Jonas Karlman <jonas@kwiboo.se>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
Link: https://patch.msgid.link/20260505170410.3265305-6-heiko@sntech.de
Signed-off-by: Vinod Koul <vkoul@kernel.org>
The 480m clk is controlled using regs in the PHY address space and not
in the USB GRF address space on e.g. RK3528 and RK3506.
Add a clkout_ctl_phy usb2phy_reg to handle enable/disable of the 480m
clk on these SoCs.
Signed-off-by: Jonas Karlman <jonas@kwiboo.se>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
Link: https://patch.msgid.link/20260505170410.3265305-5-heiko@sntech.de
Signed-off-by: Vinod Koul <vkoul@kernel.org>
The embedded USB2 PHY on RK3528 is very similar to the one in RK3568,
the main difference being that it only uses two clocks instead of three.
Add compatible to support the USB2 PHY in RK3528.
Signed-off-by: Jonas Karlman <jonas@kwiboo.se>
Reviewed-by: Rob Herring (Arm) <robh@kernel.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patch.msgid.link/20260505170410.3265305-4-heiko@sntech.de
Signed-off-by: Vinod Koul <vkoul@kernel.org>
The logic to decide if usbgrf or grf should be used is more complex than
it needs to be. For RK3568, RV1108 and soon RK3528 we can assign the
rockchip,usbgrf regmap directly to grf instead of doing a usbgrf and grf
dance.
Simplify the code to only use the grf regmap and handle the logic of
what regmap should be used in driver probe instead.
The only expected change from this is that RK3528 can be supported
because of an addition of a of_property_present() check.
Signed-off-by: Jonas Karlman <jonas@kwiboo.se>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
Link: https://patch.msgid.link/20260505170410.3265305-3-heiko@sntech.de
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Typically these Rockchip USB2 PHYs are fully contained within a single
GRF. However, for RK3568 and RV1108 regs to control the USB2 PHY is
located in a different GRF compared to the base address.
Update this binding to require rockchip,usbgrf for RK3568 and RV1108 to
properly reflect that the USB GRF is required to control the USB2 PHYs
on these variants. Also disable use of rockchip,usbgrf for variants
where it is not required.
This should not introduce any breakage as the affected usb2phy nodes for
RK3568 and RV1108 were added together with a rockchip,usbgrf phandle in
their initial commit.
Signed-off-by: Jonas Karlman <jonas@kwiboo.se>
Acked-by: Rob Herring (Arm) <robh@kernel.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patch.msgid.link/20260505170410.3265305-2-heiko@sntech.de
Signed-off-by: Vinod Koul <vkoul@kernel.org>
This file does not use the symbols from the legacy
<linux/gpio.h> header, so let's drop it.
Reviewed-by: Josua Mayer <josua@solid-run.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://patch.msgid.link/20260513220336.369628-7-andriy.shevchenko@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
The code like
int foo = X;
...
if (bar)
foo = Y;
is prone to subtle mistakes and hence harder to maintain as the foo value
may be changed inadvertently while code in '...' grown in lines. On top
it's harder to navigate to understand the possible values of foo when branch
is not taken (requires to look somewhere else in the code, far from the piece
at hand).
Besides that in case of taken branch the foo will be rewritten, which is
not a problem per se, just an unneeded operation.
Decouple assignment and definition to use if-else to address the inconveniences
described above.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://patch.msgid.link/20260513220336.369628-6-andriy.shevchenko@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Instead of checking for the specific error codes (that can be considered
a layering violation to some extent) check for the property existence first
and then either parse it, or apply a default value.
With that, return an error when parsing of the existing property fails.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://patch.msgid.link/20260513220336.369628-5-andriy.shevchenko@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
There is no code that uses ID table directly, except the
struct device_driver at the end of the file. Hence, move
table closer to its user. It's always possible to access
them via a pointer.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://patch.msgid.link/20260513220336.369628-4-andriy.shevchenko@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Use the generic firmware node interface for retrieving
device match data instead of the OF-specific one.
While at it, drop unneeded argument to devm_phy_create() which
extracts device node from the given device by default.
Reviewed-by: Josua Mayer <josua@solid-run.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://patch.msgid.link/20260513220336.369628-3-andriy.shevchenko@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Every platform driver can be forced to match a device that doesn't
match its list of device IDs because of device_match_driver_override()
so platform drivers that rely on the existence of a device's driver
data need to verify its presence.
Accordingly, add requisite match and driver data checks against NULL
to the driver where they are missing.
Fixes: a4a86d273f ("phy: phy-can-transceiver: Add support for generic CAN transceiver driver")
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://patch.msgid.link/20260513220336.369628-2-andriy.shevchenko@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
The 'dumpability' of a task is fundamentally about the memory image of
the task - the concept comes from whether it can core dump or not - and
makes no sense when you don't have an associated mm.
And almost all users do in fact use it only for the case where the task
has a mm pointer.
But we have one odd special case: ptrace_may_access() uses 'dumpable' to
check various other things entirely independently of the MM (typically
explicitly using flags like PTRACE_MODE_READ_FSCREDS). Including for
threads that no longer have a VM (and maybe never did, like most kernel
threads).
It's not what this flag was designed for, but it is what it is.
The ptrace code does check that the uid/gid matches, so you do have to
be uid-0 to see kernel thread details, but this means that the
traditional "drop capabilities" model doesn't make any difference for
this all.
Make it all make a *bit* more sense by saying that if you don't have a
MM pointer, we'll use a cached "last dumpability" flag if the thread
ever had a MM (it will be zero for kernel threads since it is never
set), and require a proper CAP_SYS_PTRACE capability to override.
Reported-by: Qualys Security Advisory <qsa@qualys.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Kees Cook <kees@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Add support for 25GBASE-R in the Lynx 28G SerDes PHY driver. This will
be used by the dpaa2-mac consumer on LX2160A with:
- phy_validate(phy, PHY_MODE_ETHERNET, PHY_INTERFACE_MODE_25GBASER) to
detect support.
- phy_set_mode_ext(phy, PHY_MODE_ETHERNET, PHY_INTERFACE_MODE_25GBASER)
to reconfigure the lane for this protocol.
The intended use case for dynamic protocol switching to 25GBase-R is
with SFP28 modules, and protocol switching is triggered by the SFP
module insertion. There also exists a 25GBase-KR use case, where the
protocol switching is covered by IEEE 802.3 clause 73 auto-negotiation.
However, that is not handled here; it merely needs the support added
here as basic ground work.
The lane frequency for 25GbE is sourced from a clock net frequency of
12.890625 GHz, as produced by PLLF or PLLS, further multiplied by the
lane by 2. The clock net frequencies produced by the PLLs are treated as
read-only by the driver, so the absence of a PLL provisioned for the
right clock net frequency implies absence of 25GbE support, even though
a lane might have the appropriate protocol converter for it.
In terms of implementation, the change consists of:
- determining at probe time if any PLL was preconfigured for the
required clock net frequency for 25GbE
- adding the default lane parameters for reconfiguring a lane to 25GbE
irrespective of the original protocol
- allowing this operating mode only on supported lanes, i.e. all lanes
of LX2162A SerDes #1, and LX2160A SerDes lanes 0-1, 4-7.
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Tested-by: Josua Mayer <josua@solid-run.com>
Link: https://patch.msgid.link/20260511150023.1903577-6-vladimir.oltean@nxp.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Add driver support for probing on the new, per-instance and per-SoC
bindings, which provide the main benefit that they allow rejecting
unsupported protocols per lane (10GbE on SerDes 2 lanes 0-5), but they
also allow avoiding the creation of PHYs for lanes that don't exist
(LX2162A lanes 0-3).
For old device trees with just "fsl,lynx-28g", the only things that
change are:
- a probe time warning/encouragement to update the device tree. This is
warranted by the fact that using "fsl,lynx-28g" may already provide
incorrect behaviour (undetected absent 10GbE support on LX2160A
SerDes 2 lanes 0-5). But we retain bug compatibility nonetheless.
- the feature set is frozen in time (e.g. no 25GbE). Since we cannot
guarantee that this protocol will work on a lane, just err on the safe
side and don't offer it (and require a device tree update to get it).
In terms of code, the lynx_28g_supports_lane_mode() function prototype
changes. It was a SerDes-global function and now becomes per lane, to
reflect the specific capabilities each instance may have. The
implementation goes through priv->info->lane_supports_mode().
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Josua Mayer <josua@solid-run.com>
Link: https://patch.msgid.link/20260511150023.1903577-5-vladimir.oltean@nxp.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
The driver will gain support for variants in an upcoming change, and
will use of_device_get_match_data() to deduce the running variant from
the compatible string.
Currently, the driver expects the schema at phy/fsl,lynx-28g.yaml, and
OF-based consumers, but doesn't enforce this. And it is possible for
user space to force-bind the driver to a device without OF node using
the driver_override sysfs.
To avoid future surprise crashes for an unsupported configuration,
explicitly test for the presence of an OF node and fail probing if
found.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Josua Mayer <josua@solid-run.com>
Link: https://patch.msgid.link/20260511150023.1903577-4-vladimir.oltean@nxp.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
The SerDes 1 of LX2162A has fewer lanes than all other instances, and
strangely, their indices are not 0-3, but 4-7.
This is a best-effort constraint, since we can only impose it when using
per-SoC compatible string and per-lane OF nodes.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Acked-by: Conor Dooley <conor.dooley@microchip.com>
Tested-by: Josua Mayer <josua@solid-run.com>
Link: https://patch.msgid.link/20260511150023.1903577-3-vladimir.oltean@nxp.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
The 28G Lynx SerDes is instantiated 3 times in the NXP LX2160A SoC and
twice in the NXP LX2162A. All these instances share the same register
map, but the number of lanes and the protocols supported by each lane
differs in a way that isn't detectable by the programming model.
For example, not all lanes of all SerDes block instantiations support
25GbE.
So, using a generic "fsl,lynx-28g" compatible string and expecting all
SerDes instantiations to use it was a mistake that needs to be fixed.
The option chosen is to encode the SoC and the SerDes instance in the
compatible string, with everything else being the responsibility of the
driver to derive.
An alternative considered but dismissed was to add sufficient device
tree properties to describe the per-lane differences (implying:
supported protocols), as well as the different lane count.
Any decision made for the 28G Lynx should be consistent with the
decisions taken for the yet-to-be-introduced 10G Lynx SerDes (older
generation for older SoCs), because of how similar they are.
I've seen the alternative at play in this unmerged patch set for the 10G
Lynx here, and I didn't like it:
https://lore.kernel.org/linux-phy/20230413160607.4128315-3-sean.anderson@seco.com/
This is because there, we have a higher degree of variability in the
PCCR register values that need to be written per protocol. This makes
that approach more drawn-out and more prone to errors, compared to the
compatible strings which are more succinct and obviously correct.
NXP SoC reference manuals clearly document the SerDes instantiations as
not identical, and refers to them as such (SerDes 1, 2, etc).
The per-SoC compatible string is prepended to the "fsl,lynx-28g" generic
compatible, which is left there for compatibility with old kernels. An
exception would be LX2160A SerDes #3, which at the time of writing is
not described in fsl-lx2160a.dtsi. As "fsl,lx2160a-serdes3" implies it
is a 28G Lynx SerDes, it makes "fsl,lynx-28g" redundant so we don't
accept it.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Rob Herring (Arm) <robh@kernel.org>
Tested-by: Josua Mayer <josua@solid-run.com>
Link: https://patch.msgid.link/20260511150023.1903577-2-vladimir.oltean@nxp.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
The residency multiplier conversion in get_residency_ms() used the
floating-point literal 1e6 as the divisor of mul_u64_u32_div(). While
the compiler constant-folds this to an integer, using float literals
in kernel code is bad practice since the kernel generally avoids
floating-point operations.
Replace 1e6 with the standard NSEC_PER_MSEC macro from <linux/time64.h>,
which is both self-documenting (ns to ms conversion) and unambiguously
integer. Add the corresponding include rather than relying on
transitive inclusion.
No functional change.
Assisted-by: Claude:claude-opus-4.6
Reviewed-by: Nitin Gote <nitin.r.gote@intel.com>
Link: https://patch.msgid.link/20260511153307.223435-1-shuicheng.lin@intel.com
Signed-off-by: Shuicheng Lin <shuicheng.lin@intel.com>
The error path in xe_gsc_init_post_hwconfig() explicitly frees a BO
allocated with xe_managed_bo_create_pin_map() via
xe_bo_unpin_map_no_vm(). Since the managed BO already has a devm
cleanup action registered, this causes a double-free when devm
unwinds during probe failure.
Remove the explicit free and let devm handle it, consistent with
all other xe_managed_bo_create_pin_map() callers.
Fixes: 2e5d47fe78 ("drm/xe/uc: Use managed bo for HuC and GSC objects")
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Assisted-by: Claude:claude-opus-4.6
Link: https://patch.msgid.link/20260511154134.223696-1-shuicheng.lin@intel.com
Signed-off-by: Shuicheng Lin <shuicheng.lin@intel.com>
percpu_array_map_ops.map_meta_equal points to the generic
bpf_map_meta_equal(), which does not compare max_entries. When a
percpu array serves as an inner map, replacing it with one that has
fewer max_entries bypasses the check. Since percpu_array_map_gen_lookup()
inlines the original template's index_mask as a JIT immediate, a lookup
on the replacement map can access pptrs[] out of bounds.
Point percpu_array_map_ops.map_meta_equal to array_map_meta_equal(),
which already enforces the max_entries equality check.
Add a selftest to verify that replacing a percpu array inner map with
a differently-sized one is rejected.
Fixes: db69718b8e ("bpf: inline bpf_map_lookup_elem() for PERCPU_ARRAY maps")
Signed-off-by: Guannan Wang <wgnbuaa@gmail.com>
Acked-by: Mykyta Yatsenko <yatsenko@meta.com>
Link: https://lore.kernel.org/r/20260514074454.77491-1-wgnbuaa@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Mark all open DAI runtimes as suspended in the component .suspend
callback instead of relying on SNDRV_PCM_TRIGGER_SUSPEND, which is
not delivered during PAUSE or xrun states.
If during system suspend a dai is open it means that it is in either in
SUSPENDED, PAUSED or STOPPED (due to xrun) state and they will need to be
re-initialized during resume (which is done in .prepare callback).
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Link: https://patch.msgid.link/20260508101755.1247039-1-yung-chuan.liao@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
smb3_reconfigure() moves strings out of cifs_sb->ctx before the
multichannel update, so a later failure can leave the live context
with NULL strings or options that do not match the session.
Stage the new ctx separately, commit it only on success, and restore
the snapshot on failure. Also make smb3_sync_session_ctx_passwords()
all-or-nothing.
Commit session passwords before channel updates so newly added channels
authenticate with the staged credentials.
Fixes: ef529f655a ("cifs: client: allow changing multichannel mount options on remount")
Reported-by: RAJASI MANDAL <rajasimandalos@gmail.com>
Closes: https://lore.kernel.org/lkml/CAEY6_V1+dzW3OD5zqXhsWyXwrDTrg5tAMGZ1AJ7_GAuRE+aevA@mail.gmail.com/
Link: https://lore.kernel.org/lkml/xkr2dlvgibq5j6gkcxd3yhhnj4atgxw2uy4eug2pxm7wy7nbms@iq6cf5taa65v/
Reviewed-by: Henrique Carvalho <henrique.carvalho@suse.com>
Signed-off-by: DaeMyung Kang <charsyam@gmail.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
Fixes a "duplicate tag error for tag 0" firmware crash during controller
reset while setting up a queue on Apple A11 / T8015 caused by stale
entries in the submission queue due to an invalid sq_tail offset after
reset.
Fixes: 04d8ecf37b ("nvme: apple: Add Apple A11 support")
Cc: stable@vger.kernel.org
Suggested-by: Yuriy Havrylyuk <yhavry@gmail.com>
Reviewed-by: Sven Peter <sven@kernel.org>
Signed-off-by: Nick Chan <towinchenmi@gmail.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
On 32-bit architectures, the infinite loop is as follows:
len = p->ErrorDataLength == 0xfffffff8
u8 *next = p->ErrorContextData + len
next == p
On 32-bit architectures, the out-of-bounds read is as follows:
len = p->ErrorDataLength == 0xfffffff0
u8 *next = p->ErrorContextData + len
next == (u8 *)p - 8
Reported-by: ChenXiaoSong <chenxiaosong@kylinos.cn>
Fixes: 76894f3e2f ("cifs: improve symlink handling for smb2+")
Cc: stable@vger.kernel.org
Signed-off-by: Ye Bin <yebin10@huawei.com>
Reviewed-by: ChenXiaoSong <chenxiaosong@kylinos.cn>
Signed-off-by: Steve French <stfrench@microsoft.com>
While deleting an existing ovpn interface, there is a very
narrow window where adding a new peer via netlink may cause
the netdevice to hang and prevent its unregistration.
It may happen during ovpn_dellink(), when all existing peers are
freed and the device is queued for deregistration, but a
CMD_PEER_NEW message comes in adding a new peer that takes again
a reference to the netdev.
At this point there is no way to release the device because we are
under the assumption that all peers were already released.
Fix the race condition by releasing all peers in ndo_uninit(),
when the netdevice has already been removed from the netdev
list.
Also ovpn_peer_add() has now an extra check that forces the
function to bail out if the device reg_state is not REGISTERED.
This way any incoming CMD_PEER_NEW racing with the interface
deletion routine will simply stop before adding the peer.
Note that the above check happens while holding the netdev_lock
to prevent racing netdev state changes.
ovpn_dellink() is now empty and can be removed.
Reported-by: Hyunwoo Kim <imv4bel@gmail.com>
Closes: https://lore.kernel.org/netdev/aaVgJ16edTfQkYbx@v4bel/
Suggested-by: Sabrina Dubroca <sd@queasysnail.net>
Fixes: 80747caef3 ("ovpn: introduce the ovpn_peer object")
Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
ovpn_nl_peer_new_doit()'s error path calls ovpn_peer_release() directly
rather than ovpn_peer_put(), bypassing the kref. The accompanying
comment ("peer was not yet hashed, thus it is not used in any context")
holds for UDP but not for TCP.
For UDP, the ovpn_socket union uses the .ovpn arm and never points back
at a peer; UDP encap_recv looks up peers via the not-yet-populated
hashtables, so the new peer is unreachable until ovpn_peer_add()
publishes it.
For TCP, ovpn_socket_new() sets ovpn_sock->peer and
ovpn_tcp_socket_attach() publishes ovpn_sock via rcu_assign_sk_user_data().
From that moment until ovpn_socket_release() detaches in the error path,
the TCP fd is fully wired: userspace recvmsg / sendmsg / close / poll
on the fd, as well as the strparser-driven ovpn_tcp_rcv() path, can
reach the peer through sk_user_data -> ovpn_sock->peer and bump its
refcount via ovpn_peer_hold().
ovpn_tcp_socket_wait_finish() (called inside ovpn_socket_release())
drains strparser and the tx work, but does not synchronize with
userspace syscall callers that already hold a peer reference. If
ovpn_nl_peer_modify() or ovpn_peer_add() returns an error while such
a caller is in flight - notably an ovpn_tcp_recvmsg() blocked in
__skb_recv_datagram() on peer->tcp.user_queue - the direct
ovpn_peer_release() destroys the peer while the caller still holds
the reference, and the eventual ovpn_peer_put() from that caller
operates on freed memory.
Replace the direct destructor call with ovpn_peer_put() so the kref
correctly defers destruction until the last reference is dropped.
In the common case where no concurrent user is present, behaviour is
unchanged: the kref hits zero immediately and ovpn_peer_release_kref()
runs the same destructor.
With this conversion ovpn_peer_release() has no callers outside peer.c
- ovpn_peer_release_kref() in the same translation unit is the only
remaining user - so make it static and drop its declaration from
peer.h.
Fixes: 11851cbd60 ("ovpn: implement TCP transport")
Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>
Assisted-by: Claude:claude-opus-4-7
Signed-off-by: David Carlier <devnexen@gmail.com>
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
ovpn_tcp_close() loads the ovpn_socket via rcu_dereference_sk_user_data()
under rcu_read_lock(), takes a reference on sock->peer, caches the peer
pointer in a local, and drops the read lock. It then passes sock->peer
(rather than the cached local) to ovpn_peer_del(), re-dereferencing the
ovpn_socket after the RCU read section has ended.
Unlike ovpn_tcp_sendmsg(), which uses the same "load under RCU, use
after unlock" pattern but is protected by lock_sock() held across the
function, ovpn_tcp_close() runs without the socket lock: inet_release()
invokes sk_prot->close() without taking lock_sock first.
ovpn_socket_release() can therefore complete its kref_put -> detach ->
synchronize_rcu -> kfree(sock) sequence concurrently, in the window
after ovpn_tcp_close() drops rcu_read_lock() but before it dereferences
sock->peer. The synchronize_rcu() in ovpn_socket_release() protects
readers that use the dereferenced pointer inside the RCU read section,
not those that escape the pointer to a local and use it afterwards.
A reproducer follows the pattern of commit 94560267d6 ("ovpn: tcp -
don't deref NULL sk_socket member after tcp_close()"): trigger a peer
removal (keepalive expiration or netlink OVPN_CMD_DEL_PEER) at the same
moment userspace closes the TCP fd. That commit fixed the detach-side
of the same race window; this one fixes the close-side at a different
victim.
Tighten the entry block to read sock->peer exactly once into the cached
peer local, and route all subsequent uses (the hold check, the
ovpn_peer_del() call, and the prot->close() invocation) through that
local. sock->peer is only ever written once in ovpn_socket_new() under
lock_sock(), before rcu_assign_sk_user_data() publishes the ovpn_socket,
and is never reassigned afterwards - but the previous multi-read pattern
made that invariant implicit rather than explicit. The same multi-read
shape exists in ovpn_tcp_recvmsg(), ovpn_tcp_sendmsg(),
ovpn_tcp_data_ready() and ovpn_tcp_write_space(); those will be cleaned
up via a dedicated helper in a follow-up net-next series.
Fixes: 11851cbd60 ("ovpn: implement TCP transport")
Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>
Assisted-by: Claude:claude-opus-4-7
Signed-off-by: David Carlier <devnexen@gmail.com>
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
Commit 201ba70631 ("selftests: ovpn: reduce ping count in test.sh")
lowered the baseline traffic flood ping count to avoid flakes on slower
CI instances, however some instances were left out.
Apply the same limit to the remaining ovpn selftest flood pings that
still request 500 packets.
Fixes: 201ba70631 ("selftests: ovpn: reduce ping count in test.sh")
Signed-off-by: Ralf Lici <ralf@mandelbit.com>
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
There's no real reason to have a limit, as the memory is accounted by
the lockmem limits anyway, if any exist. io_pin_pages() will still
restrict the maximum allowed limit per buffer, which is INT_MAX
number of pages. Cap it a bit lower than that, at 1TB for a 64-bit
system. Surely that should be enough for everyone. For now.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
In preparation for supporting bigger individual buffers, bump the length
field to a full 8-bytes with size_t rather than an unsigned int.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Track huge page references in a per-ring xarray to prevent double
accounting when the same huge page is used by multiple registered
buffers, either within the same ring or across cloned rings.
When registering buffers backed by huge pages, we need to account for
RLIMIT_MEMLOCK. But if multiple buffers share the same huge page (common
with cloned buffers), we must not account for the same page multiple
times. Similarly, we must only unaccount when the last reference to a
huge page is released.
Maintain a per-ring xarray (hpage_acct) that tracks reference counts for
each huge page. When registering a buffer, for each unique huge page,
increment its accounting reference count, and only account pages that
are newly added.
When unregistering a buffer, for each unique huge page, decrement its
refcount. Once the refcount hits zero, the page is unaccounted.
Note: any account is done against the ctx->user that was assigned when
the ring was setup. As before, if root is running the operation, no
accounting is done.
With these changes, any use of imu->acct_pages is also dead, hence kill
it from struct io_mapped_ubuf. This shrinks it from 56b to 48b on a
64-bit arch. Additionally, hpage_already_acct() is gone, which was an
O(M*M) scan over current + previous registrations.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Since the timer uses jiffies as its unit rather than ms, the timeout value
must be converted from ms to jiffies when configuring the timer. Otherwise,
the intended 8s timeout is incorrectly set to approximately 33s.
To improve readability, embed msecs_to_jiffies() directly in the macro
definitions and drop the _MS suffix from macros that now yield jiffies
values: MEMDUMP_TIMEOUT, FW_DOWNLOAD_TIMEOUT, IBS_DISABLE_SSR_TIMEOUT,
CMD_TRANS_TIMEOUT, and IBS_BTSOC_TX_IDLE_TIMEOUT.
IBS_WAKE_RETRANS_TIMEOUT_MS and IBS_HOST_TX_IDLE_TIMEOUT_MS are
intentionally left unchanged. Their values are stored in the struct fields
wake_retrans and tx_idle_delay, which hold ms values at runtime and can be
modified via debugfs. The msecs_to_jiffies() conversion happens at each
call site against the field value, so it cannot be embedded in the macro.
Wake timer depends on commit c347ca17d6
Cc: stable@vger.kernel.org
Fixes: d841502c79 ("Bluetooth: hci_qca: Collect controller memory dump during SSR")
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Acked-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Signed-off-by: Shuai Zhang <shuai.zhang@oss.qualcomm.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Commit 1c08108f30 ("Bluetooth: L2CAP: Avoid -Wflex-array-member-not-at-end
warnings") converted the on-stack request PDU in l2cap_ecred_reconfigure()
from an explicit packed struct to DEFINE_RAW_FLEX(), but did not adjust the
size and source-pointer arguments to l2cap_send_cmd():
- struct {
- struct l2cap_ecred_reconf_req req;
- __le16 scid;
- } pdu;
+ DEFINE_RAW_FLEX(struct l2cap_ecred_reconf_req, pdu, scid, 1);
...
l2cap_send_cmd(conn, chan->ident, L2CAP_ECRED_RECONF_REQ,
sizeof(pdu), &pdu);
After the conversion, DEFINE_RAW_FLEX() expands to declare an anonymous
union pdu_u plus a local pointer "pdu" pointing at it. Therefore:
- sizeof(pdu) is now sizeof(struct l2cap_ecred_reconf_req *) = 8 on
64-bit (4 on 32-bit), not the 6 bytes of (mtu, mps, scid[1]).
- &pdu is the address of the local pointer's stack storage, not the
address of the request payload.
l2cap_send_cmd() forwards (data, count) to l2cap_build_cmd(), which calls
skb_put_data(skb, data, count). The L2CAP_ECRED_RECONFIGURE_REQ packet
body therefore contains 8 bytes copied from the kernel stack starting at
&pdu -- the 8 bytes overlap the pdu pointer's value, leaking a kernel
stack address to the paired Bluetooth peer. The intended (mtu, mps, scid)
fields are not transmitted at all, so the peer rejects the request as
malformed and the L2CAP_ECRED_RECONFIGURE feature itself has been broken
for the local-side initiator since the introducing commit landed.
The sibling site l2cap_ecred_conn_req() in the same commit was converted
correctly (sizeof(*pdu) + len, pdu); only this site was missed.
Restore the original semantics: pass the full flex-struct size via
struct_size(pdu, scid, 1) and the pdu pointer (the struct address) as
the source.
Validated on a stock 7.0-based host kernel via the real call path:
setsockopt(SOL_BLUETOOTH, BT_RCVMTU, ...) on a BT_CONNECTED
L2CAP_MODE_EXT_FLOWCTL socket emits an L2CAP_ECRED_RECONFIGURE_REQ
whose body is 8 bytes (the on-stack pdu local's value) rather than
the expected 6. Three captures from fresh socket / fresh hciemu peer
on the same host -- low bytes vary per call, high 0xffff confirms a
kernel virtual address (KASLR-randomised stack slot, not a fixed
string):
RECONF_REQ body (ident=0x02 len=8): 42 fb 54 af 0e ca ff ff
RECONF_REQ body (ident=0x02 len=8): 52 3d 2e af 0e ca ff ff
RECONF_REQ body (ident=0x02 len=8): b2 fc 5b af 0e ca ff ff
After this patch the body is 6 bytes carrying the expected
little-endian (mtu, mps, scid).
Cc: stable@vger.kernel.org
Fixes: 1c08108f30 ("Bluetooth: L2CAP: Avoid -Wflex-array-member-not-at-end warnings")
Assisted-by: Claude:claude-opus-4-7
Signed-off-by: Michael Bommarito <michael.bommarito@gmail.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
MT7925 (USB ID 0e8d:e025) on fw version 20260106153314 sends WMT
FUNC_CTRL events that are missing the status field.
Prior to commit 006b9943b982 ("Bluetooth: btmtk: validate WMT event SKB
length before struct access") the status was read from out-of-bounds of
SKB data, which usually would result to success with
BTMTK_WMT_ON_UNDONE, although I don't know the intent here. The bounds
check added in that commit returns with error instead, producing
"Bluetooth: hci0: Failed to send wmt func ctrl (-22)" and makes the
device unusable.
Fix the regression by interpreting too short packet as status
BTMTK_WMT_ON_UNDONE, which makes the device work normally again.
Fixes: 634a4408c0 ("Bluetooth: btmtk: validate WMT event SKB length before struct access")
Signed-off-by: Pauli Virtanen <pav@iki.fi>
Tested-by: Mikhail Gavrilov <mikhail.v.gavrilov@gmail.com> # MT7922 (0489:e0e2)
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
bt_sock_poll() walks the accept queue without synchronization, while
child teardown can unlink the same socket and drop its last reference.
The unsynchronized accept queue walk has existed since the initial
Bluetooth import.
Protect accept_q with a dedicated lock for queue updates and polling.
Also rework bt_accept_dequeue() to take temporary child references under
the queue lock before dropping it and locking the child socket.
Fixes: 1da177e4c3 ("Linux-2.6.12-rc2")
Cc: stable@vger.kernel.org
Reported-by: Jann Horn <jannh@google.com>
Reported-by: Yuan Tan <yuantan098@gmail.com>
Reported-by: Yifan Wu <yifanwucs@gmail.com>
Reported-by: Juefei Pu <tomapufckgml@gmail.com>
Reported-by: Xin Liu <bird@lzu.edu.cn>
Signed-off-by: Jiexun Wang <wangjiexun2025@gmail.com>
Signed-off-by: Ren Wei <n05ec@lzu.edu.cn>
Signed-off-by: Jiexun Wang <wangjiexun2025@gmail.com>
Reviewed-by: Jann Horn <jannh@google.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
There are some performance issues being identified by dynamic EPP
and we don't want to have distributions turning it on by default
exposing them to users at this time.
Drop the kconfig option, and require an explicit opt in from kernel
command line or runtime sysfs option to turn it on.
Reported-by: Viktor Jägersküpper <viktor_jaegerskuepper@freenet.de>
Closes: https://lore.kernel.org/linux-pm/14a87c99-785c-4b16-bfce-35ecbf053448@freenet.de/
Reported-by: Stuart Meckle <stuartmeckle@gmail.com>
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=221473
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Reviewed-by: K Prateek Nayak <kprateek.nayak@amd.com>
Link: https://lore.kernel.org/r/20260512221947.1652988-1-mario.limonciello@amd.com
(fix sysfs file path)
Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
Apply the same fix as b2ed01e7ad ("drm/ttm: Fix ttm_bo_swapout()
infinite LRU walk on swapout failure") to the ttm_bo_shrink() path.
Move del_bulk_move from before the backup to after success only,
using ttm_resource_del_bulk_move_unevictable() since the resource
is now unevictable once fully backed up.
Fixes: 70d645deac ("drm/ttm: Add helpers for shrinking")
Cc: Christian König <christian.koenig@amd.com>
Cc: Huang Rui <ray.huang@amd.com>
Cc: Matthew Auld <matthew.auld@intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Dave Airlie <airlied@redhat.com>
Cc: dri-devel@lists.freedesktop.org
Cc: stable@vger.kernel.org # v6.15+
Assisted-by: GitHub_Copilot:claude-opus-4.6
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patch.msgid.link/20260511162443.24352-1-thomas.hellstrom@linux.intel.com
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
This adds custom filtering for IORING_OP_CONNECT, where the target
family is always exposed, and (for AF_INET / AF_INET6) port and
address are exposed. port and v4_addr are in network byte order so
filter authors can compare against on-wire constants.
Skip population unless addr_len covers the populated fields, to
avoid leaking stale io_async_msghdr data on short connects.
Signed-off-by: Shouvik Kar <auxcorelabs@gmail.com>
Link: https://patch.msgid.link/20260512110242.26219-1-auxcorelabs@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
