By experiments, a single queue representor netdev consumes kernel
memory around 2.8MB, and 1.8MB out of the 2.8MB is due to page
pool for the RXQ. Scaling to a thousand representors consumes 2.8GB,
which becomes a memory pressure issue for embedded devices such as
BlueField-2 16GB / BlueField-3 32GB memory.
Since representor netdevs mostly handles miss traffic, and ideally,
most of the traffic will be offloaded, reduce the default non-uplink
rep netdev's RXQ default depth from 1024 to 256 if mdev is ecpf eswitch
manager. This saves around 1MB of memory per regular RQ,
(1024 - 256) * 2KB, allocated from page pool.
With rxq depth of 256, the netlink page pool tool reports
$./tools/net/ynl/cli.py --spec Documentation/netlink/specs/netdev.yaml \
--dump page-pool-get
{'id': 277,
'ifindex': 9,
'inflight': 128,
'inflight-mem': 786432,
'napi-id': 775}]
This is due to mtu 1500 + headroom consumes half pages, so 256 rxq
entries consumes around 128 pages (thus create a page pool with
size 128), shown above at inflight.
Note that each netdev has multiple types of RQs, including
Regular RQ, XSK, PTP, Drop, Trap RQ. Since non-uplink representor
only supports regular rq, this patch only changes the regular RQ's
default depth.
Signed-off-by: William Tu <witu@nvidia.com> Reviewed-by: Bodong Wang <bodong@nvidia.com> Reviewed-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Tariq Toukan <tariqt@nvidia.com> Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com> Link: https://patch.msgid.link/20250209101716.112774-8-tariqt@nvidia.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
By default, the mq netdev creates a pfifo_fast qdisc. On a
system with 16 core, the pfifo_fast with 3 bands consumes
16 * 3 * 8 (size of pointer) * 1024 (default tx queue len)
= 393KB. The patch sets the tx qlen to representor default
value, 128 (1<<MLX5E_REP_PARAMS_DEF_LOG_SQ_SIZE), which
consumes 16 * 3 * 8 * 128 = 49KB, saving 344KB for each
representor at ECPF.
Signed-off-by: William Tu <witu@nvidia.com> Reviewed-by: Daniel Jurgens <danielj@nvidia.com> Signed-off-by: Tariq Toukan <tariqt@nvidia.com> Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com> Link: https://patch.msgid.link/20250209101716.112774-9-tariqt@nvidia.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Current loopback test validation ignores non-linear SKB case in
the SKB access, which can lead to failures in scenarios such as
when HW GRO is enabled.
Linearize the SKB so both cases will be handled.
[Why & How]
The initial setting for psr_version is not correct while
create a virtual link.
The default psr_version should be DC_PSR_VERSION_UNSUPPORTED.
Reviewed-by: Roman Li <roman.li@amd.com> Signed-off-by: Tom Chung <chiahsuan.chung@amd.com> Signed-off-by: Zaeem Mohamed <zaeem.mohamed@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
As reported by Damon Ding, the phy_get_mode() call doesn't work as
expected unless the PHY driver has a .set_mode() call. This prompts PHY
drivers to have empty stubs for .set_mode() for the sake of being able
to get the mode.
Make .set_mode() callback truly optional and update PHY's mode even if
it there is none.
The Renesas RZ/G3S supports a power saving mode where power to most of the
SoC components is turned off. When returning from this power saving mode,
SoC components need to be re-configured.
The SCIFs on the Renesas RZ/G3S need to be re-configured as well when
returning from this power saving mode. The sh-sci code already configures
the SCIF clocks, power domain and registers by calling uart_resume_port()
in sci_resume(). On suspend path the SCIF UART ports are suspended
accordingly (by calling uart_suspend_port() in sci_suspend()). The only
missing setting is the reset signal. For this assert/de-assert the reset
signal on driver suspend/resume.
In case the no_console_suspend is specified by the user, the registers need
to be saved on suspend path and restore on resume path. To do this the
sci_console_save()/sci_console_restore() functions were added. There is no
need to cache/restore the status or FIFO registers. Only the control
registers. The registers that will be saved/restored on suspend/resume are
specified by the struct sci_suspend_regs data structure.
If regmap_read() fails, random stack value was used in calculating new
frequency in recalc_rate() callbacks. Such failure is really not
expected as these are all MMIO reads, however code should be here
correct and bail out. This also avoids possible warning on
uninitialized value.
GCC can see that the value range for "order" is capped, but this leads
it to consider that it might be negative, leading to a false positive
warning (with GCC 15 with -Warray-bounds -fdiagnostics-details):
../drivers/net/ethernet/mellanox/mlx4/alloc.c:691:47: error: array subscript -1 is below array bounds of 'long unsigned int *[2]' [-Werror=array-bounds=]
691 | i = find_first_bit(pgdir->bits[o], MLX4_DB_PER_PAGE >> o);
| ~~~~~~~~~~~^~~
'mlx4_alloc_db_from_pgdir': events 1-2
691 | i = find_first_bit(pgdir->bits[o], MLX4_DB_PER_PAGE >> o); | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| | | | | (2) out of array bounds here
| (1) when the condition is evaluated to true In file included from ../drivers/net/ethernet/mellanox/mlx4/mlx4.h:53,
from ../drivers/net/ethernet/mellanox/mlx4/alloc.c:42:
../include/linux/mlx4/device.h:664:33: note: while referencing 'bits'
664 | unsigned long *bits[2];
| ^~~~
Switch the argument to unsigned int, which removes the compiler needing
to consider negative values.
1) security/smack/smackfs.c`smk_set_cipso
does not clear NETLBL_SECATTR_MLS_CAT
from (struct smack_known *)skp->smk_netlabel.flags
on assigning CIPSO w/o categories:
2) security/smack/smack_lsm.c`smack_from_secattr
can not match skp->smk_netlabel with input packet's
struct netlbl_lsm_secattr *sap
because sap->flags have not NETLBL_SECATTR_MLS_CAT (what is correct)
but skp->smk_netlabel.flags have (what is incorrect):
| if ((sap->flags & NETLBL_SECATTR_MLS_CAT) == 0) {
| if ((skp->smk_netlabel.flags &
| NETLBL_SECATTR_MLS_CAT) == 0)
| found = 1;
| break;
| }
This commit sets/clears NETLBL_SECATTR_MLS_CAT in
skp->smk_netlabel.flags according to the presense of CIPSO categories.
The update of smk_netlabel is not atomic, so input packets processing
still may be incorrect during short time while update proceeds.
Cross case in pinctrl framework make impossible to an hogged pin and
another, not hogged, used within the same device-tree node. For example
with this simplified device-tree :
"pinctrl_pin_1" configuration is never set. This produces this path in
the code:
really_probe()
pinctrl_bind_pins()
| devm_pinctrl_get()
| pinctrl_get()
| create_pinctrl()
| pinctrl_dt_to_map()
| // Hog pin create an abort for all pins of the node
| ret = dt_to_map_one_config()
| | /* Do not defer probing of hogs (circular loop) */
| | if (np_pctldev == p->dev->of_node)
| | return -ENODEV;
| if (ret)
| goto err
|
call_driver_probe()
stm32_rtc_probe()
pinctrl_enable()
pinctrl_claim_hogs()
create_pinctrl()
for_each_maps(maps_node, i, map)
// Not hog pin is skipped
if (pctldev && strcmp(dev_name(pctldev->dev),
map->ctrl_dev_name))
continue;
At the first call of create_pinctrl() the hogged pin produces an abort to
avoid a defer of hogged pins. All other pin configurations are trashed.
At the second call, create_pinctrl is now called with pctldev parameter to
get hogs, but in this context only hogs are set. And other pins are
skipped.
To handle this, do not produce an abort in the first call of
create_pinctrl(). Classic pin configuration will be set in
pinctrl_bind_pins() context. And the hogged pin configuration will be set
in pinctrl_claim_hogs() context.
The ASoC convention is that clocks are removed after codec mute, and
power up/down is more about top level power management. For these chips,
the "mute" state still expects a TDM clock, and yanking the clock in
this state will trigger clock errors. So, do the full
shutdown<->mute<->active transition on the mute operation, so the amp is
in software shutdown by the time the clocks are removed.
This fixes TDM clock errors when streams are stopped.
Lower the volume if it is violating the platform maximum at its initial
value (i.e. at the time of the 'snd_soc_limit_volume' call).
Signed-off-by: Martin Povišer <povik+lin@cutebit.org>
[Cherry picked from the Asahi kernel with fixups -- broonie] Signed-off-by: Mark Brown <broonie@kernel.org> Link: https://patch.msgid.link/20250208-asoc-volume-limit-v1-1-b98fcf4cdbad@kernel.org Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Wrap the high temperature warning in a temperature event with
a call to net_ratelimit() to prevent flooding the kernel log
with repeated warning messages when temperature exceeds the
threshold multiple times within a short duration.
In the sensor_count field of the MTEWE register, bits 1-62 are
supported only for unmanaged switches, not for NICs, and bit 63
is reserved for internal use.
To prevent confusing output that may include set bits that are
not relevant to NIC sensors, we update the bitmask to retain only
the first bit, which corresponds to the sensor ASIC.
When a vxlan netdevice is brought up, if its default remote is a multicast
address, the device joins the indicated group.
Therefore when the multicast remote address changes, the device should
leave the current group and subscribe to the new one. Similarly when the
interface used for endpoint communication is changed in a situation when
multicast remote is configured. This is currently not done.
Both vxlan_igmp_join() and vxlan_igmp_leave() can however fail. So it is
possible that with such fix, the netdevice will end up in an inconsistent
situation where the old group is not joined anymore, but joining the new
group fails. Should we join the new group first, and leave the old one
second, we might end up in the opposite situation, where both groups are
joined. Undoing any of this during rollback is going to be similarly
problematic.
One solution would be to just forbid the change when the netdevice is up.
However in vnifilter mode, changing the group address is allowed, and these
problems are simply ignored (see vxlan_vni_update_group()):
# ip link add name br up type bridge vlan_filtering 1
# ip link add vx1 up master br type vxlan external vnifilter local 192.0.2.1 dev lo dstport 4789
# bridge vni add dev vx1 vni 200 group 224.0.0.1
# tcpdump -i lo &
# bridge vni add dev vx1 vni 200 group 224.0.0.2
18:55:46.523438 IP 0.0.0.0 > 224.0.0.22: igmp v3 report, 1 group record(s)
18:55:46.943447 IP 0.0.0.0 > 224.0.0.22: igmp v3 report, 1 group record(s)
# bridge vni
dev vni group/remote
vx1 200 224.0.0.2
Having two different modes of operation for conceptually the same interface
is silly, so in this patch, just do what the vnifilter code does and deal
with the errors by crossing fingers real hard.
The vnifilter code leaves old before joining new, and in case of join /
leave failures does not roll back the configuration changes that have
already been applied, but bails out of joining if it could not leave. Do
the same here: leave before join, apply changes unconditionally and do not
attempt to join if we couldn't leave.
Signed-off-by: Petr Machata <petrm@nvidia.com> Reviewed-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org> Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When the HED driver is built-in, it initializes after evged because they
both are at the same initcall level, so the initialization ordering
depends on the Makefile order. However, this prevents RAS records
coming in between the evged driver initialization and the HED driver
initialization from being handled.
If the number of such RAS records is above the APEI HEST error source
number, the HEST resources may be exhausted, and that may affect
subsequent RAS error reporting.
To fix this issue, change the initcall level of HED to subsys_initcall
and prevent the driver from being built as a module by changing ACPI_HED
in Kconfig from "tristate" to "bool".
Commit 903534fa7d30 ("PCI: Fix resource double counting on remove &
rescan") fixed double counting of mem resources because of old_size being
applied too early.
Fix a similar counting bug on the io resource side.
mlx4 doesn't support ndo_xdp_xmit / XDP_REDIRECT and wasn't
using page pool until now, so it could run XDP completions
in netpoll (NAPI budget == 0) just fine. Page pool has calling
context requirements, make sure we don't try to call it from
what is potentially HW IRQ context.
Starting with commit c141ecc3cecd7 ("of: Warn when of_property_read_bool()
is used on non-boolean properties"), probing the gpcv2 device on i.MX8M
SoCs leads to warnings when LOCKDEP is enabled.
Fix this by checking property presence with of_property_present as
intended.
A single scatter-gather entry is limited by a 32 bits "length" field
that is practically 4GB - PAGE_SIZE. This means that even when the
memory is physically contiguous, we might need more than one entry to
represent it. Additionally when using dmabuf, the sg_table might be
originated outside the subsystem and optimized for other needs.
For instance an SGT of 16GB GPU continuous memory might look like this:
(a real life example)
Since ib_umem_find_best_pgsz works within SG entries, in the above case
we will result with the worst possible 4KB page size.
Fix this by taking into consideration only the alignment of addresses of
real discontinuity points rather than treating SG entries as such, and
adjust the page iterator to correctly handle cross SG entry pages.
There is currently an assumption that drivers do not ask for pages
bigger than maximal DMA size supported by their devices.
The following splat has been observed on a SAMA5D27 platform using
atmel_serial:
BUG: sleeping function called from invalid context at kernel/irq/manage.c:738
in_atomic(): 1, irqs_disabled(): 128, non_block: 0, pid: 27, name: kworker/u5:0
preempt_count: 1, expected: 0
INFO: lockdep is turned off.
irq event stamp: 0
hardirqs last enabled at (0): [<00000000>] 0x0
hardirqs last disabled at (0): [<c01588f0>] copy_process+0x1c4c/0x7bec
softirqs last enabled at (0): [<c0158944>] copy_process+0x1ca0/0x7bec
softirqs last disabled at (0): [<00000000>] 0x0
CPU: 0 UID: 0 PID: 27 Comm: kworker/u5:0 Not tainted 6.13.0-rc7+ #74
Hardware name: Atmel SAMA5
Workqueue: hci0 hci_power_on [bluetooth]
Call trace:
unwind_backtrace from show_stack+0x18/0x1c
show_stack from dump_stack_lvl+0x44/0x70
dump_stack_lvl from __might_resched+0x38c/0x598
__might_resched from disable_irq+0x1c/0x48
disable_irq from mctrl_gpio_disable_ms+0x74/0xc0
mctrl_gpio_disable_ms from atmel_disable_ms.part.0+0x80/0x1f4
atmel_disable_ms.part.0 from atmel_set_termios+0x764/0x11e8
atmel_set_termios from uart_change_line_settings+0x15c/0x994
uart_change_line_settings from uart_set_termios+0x2b0/0x668
uart_set_termios from tty_set_termios+0x600/0x8ec
tty_set_termios from ttyport_set_flow_control+0x188/0x1e0
ttyport_set_flow_control from wilc_setup+0xd0/0x524 [hci_wilc]
wilc_setup [hci_wilc] from hci_dev_open_sync+0x330/0x203c [bluetooth]
hci_dev_open_sync [bluetooth] from hci_dev_do_open+0x40/0xb0 [bluetooth]
hci_dev_do_open [bluetooth] from hci_power_on+0x12c/0x664 [bluetooth]
hci_power_on [bluetooth] from process_one_work+0x998/0x1a38
process_one_work from worker_thread+0x6e0/0xfb4
worker_thread from kthread+0x3d4/0x484
kthread from ret_from_fork+0x14/0x28
This warning is emitted when trying to toggle, at the highest level,
some flow control (with serdev_device_set_flow_control) in a device
driver. At the lowest level, the atmel_serial driver is using
serial_mctrl_gpio lib to enable/disable the corresponding IRQs
accordingly. The warning emitted by CONFIG_DEBUG_ATOMIC_SLEEP is due to
disable_irq (called in mctrl_gpio_disable_ms) being possibly called in
some atomic context (some tty drivers perform modem lines configuration
in regions protected by port lock).
Split mctrl_gpio_disable_ms into two differents APIs, a non-blocking one
and a blocking one. Replace mctrl_gpio_disable_ms calls with the
relevant version depending on whether the call is protected by some port
lock.
Ensure the FIFO is empty before issuing the DAA command to prevent
incorrect command data from being sent. Align with other data transfers,
such as svc_i3c_master_start_xfer_locked(), which flushes the FIFO before
sending a command.
Signed-off-by: Frank Li <Frank.Li@nxp.com> Reviewed-by: Miquel Raynal <miquel.raynal@bootlin.com> Link: https://lore.kernel.org/r/20250129162250.3629189-1-Frank.Li@nxp.com Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
gcc-14 produces a bogus warning in some configurations:
drivers/edac/ie31200_edac.c: In function 'ie31200_probe1.isra':
drivers/edac/ie31200_edac.c:412:26: error: 'dimm_info' is used uninitialized [-Werror=uninitialized]
412 | struct dimm_data dimm_info[IE31200_CHANNELS][IE31200_DIMMS_PER_CHANNEL];
| ^~~~~~~~~
drivers/edac/ie31200_edac.c:412:26: note: 'dimm_info' declared here
412 | struct dimm_data dimm_info[IE31200_CHANNELS][IE31200_DIMMS_PER_CHANNEL];
| ^~~~~~~~~
I don't see any way the unintialized access could really happen here,
but I can see why the compiler gets confused by the two loops.
Instead, rework the two nested loops to only read the addr_decode
registers and then keep only one instance of the dimm info structure.
[Tony: Qiuxu pointed out that the "populate DIMM info" comment was left
behind in the refactor and suggested moving it. I deleted the comment
as unnecessry in front os a call to populate_dimm_info(). That seems
pretty self-describing.]
This function translates the rate number reported by the hardware into
something mac80211 can understand. It was ignoring the 3SS and 4SS HT
rates. Translate them too.
Also set *nss to 0 for the HT rates, just to make sure it's
initialised.
When an IOCTL times out and driver issues a target reset, if firmware
fails the task management elevate the recovery by issuing a diag reset to
controller.
In multi-cluster MIPS I6500 systems there is a GIC in each cluster,
each with its own counter. When a cluster powers up the counter will
be stopped, with the COUNTSTOP bit set in the GIC_CONFIG register.
In single cluster systems, it has been fine to clear COUNTSTOP once
in gic_clocksource_of_init() to start the counter. In multi-cluster
systems, this will only have started the counter in the boot cluster,
and any CPUs in other clusters will find their counter stopped which
will break the GIC clock_event_device.
Resolve this by having CPUs clear the COUNTSTOP bit when they come
online, using the existing gic_starting_cpu() CPU hotplug callback. This
will allow CPUs in secondary clusters to ensure that the cluster's GIC
counter is running as expected.
Signed-off-by: Paul Burton <paulburton@kernel.org> Signed-off-by: Chao-ying Fu <cfu@wavecomp.com> Signed-off-by: Dragan Mladjenovic <dragan.mladjenovic@syrmia.com> Signed-off-by: Aleksandar Rikalo <arikalo@gmail.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Tested-by: Serge Semin <fancer.lancer@gmail.com> Tested-by: Gregory CLEMENT <gregory.clement@bootlin.com> Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org> Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
The pm-cps code has up until now used per-CPU variables indexed by core,
rather than CPU number, in order to share data amongst sibling CPUs (ie.
VPs/threads in a core). This works fine for single cluster systems, but
with multi-cluster systems a core number is no longer unique in the
system, leading to sharing between CPUs that are not actually siblings.
Avoid this issue by using per-CPU variables as they are more generally
used - ie. access them using CPU numbers rather than core numbers.
Sharing between siblings is then accomplished by:
- Assigning the same pointer to entries for each sibling CPU for the
nc_asm_enter & ready_count variables, which allow this by virtue of
being per-CPU pointers.
- Indexing by the first CPU set in a CPUs cpu_sibling_map in the case
of pm_barrier, for which we can't use the previous approach because
the per-CPU variable is not a pointer.
Signed-off-by: Paul Burton <paulburton@kernel.org> Signed-off-by: Dragan Mladjenovic <dragan.mladjenovic@syrmia.com> Signed-off-by: Aleksandar Rikalo <arikalo@gmail.com> Tested-by: Serge Semin <fancer.lancer@gmail.com> Tested-by: Gregory CLEMENT <gregory.clement@bootlin.com> Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
The IOMMU translation for MSI message addresses has been a 2-step process,
separated in time:
1) iommu_dma_prepare_msi(): A cookie pointer containing the IOVA address
is stored in the MSI descriptor when an MSI interrupt is allocated.
2) iommu_dma_compose_msi_msg(): this cookie pointer is used to compute a
translated message address.
This has an inherent lifetime problem for the pointer stored in the cookie
that must remain valid between the two steps. However, there is no locking
at the irq layer that helps protect the lifetime. Today, this works under
the assumption that the iommu domain is not changed while MSI interrupts
being programmed. This is true for normal DMA API users within the kernel,
as the iommu domain is attached before the driver is probed and cannot be
changed while a driver is attached.
Classic VFIO type1 also prevented changing the iommu domain while VFIO was
running as it does not support changing the "container" after starting up.
However, iommufd has improved this so that the iommu domain can be changed
during VFIO operation. This potentially allows userspace to directly race
VFIO_DEVICE_ATTACH_IOMMUFD_PT (which calls iommu_attach_group()) and
VFIO_DEVICE_SET_IRQS (which calls into iommu_dma_compose_msi_msg()).
This potentially causes both the cookie pointer and the unlocked call to
iommu_get_domain_for_dev() on the MSI translation path to become UAFs.
Fix the MSI cookie UAF by removing the cookie pointer. The translated IOVA
address is already known during iommu_dma_prepare_msi() and cannot change.
Thus, it can simply be stored as an integer in the MSI descriptor.
The other UAF related to iommu_get_domain_for_dev() will be addressed in
patch "iommu: Make iommu_dma_prepare_msi() into a generic operation" by
using the IOMMU group mutex.
On MIPS system, most of the syscall function name begin with prefix
sys_. Some syscalls are special such as clone/fork, function name of
these begin with __sys_. Since scratch registers need be saved in
stack when these system calls happens.
With ftrace system call method, system call functions are declared with
SYSCALL_DEFINEx, metadata of the system call symbol name begins with
sys_. Here mips specific function arch_syscall_match_sym_name is used to
compare function name between sys_call_table[] and metadata of syscall
symbol.
Signed-off-by: Bibo Mao <maobibo@loongson.cn> Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
When CONFIG_PCI_P2PDMA=y (which is basically enabled on all
large x86 distros), it maps the PFN's via a ZONE_DEVICE
mapping using devm_memremap_pages(). The mapped virtual
address range corresponds to the pci_resource_start()
of the BAR address and size corresponding to the BAR length.
When KASLR is enabled, the direct map range of the kernel is
reduced to the size of physical memory plus additional padding.
If the BAR address is beyond this limit, PCI peer to peer DMA
mappings fail.
Fix this by not shrinking the size of the direct map when
CONFIG_PCI_P2PDMA=y.
This reduces the total available entropy, but it's better than
the current work around of having to disable KASLR completely.
[ mingo: Clarified the changelog to point out the broad impact ... ]
When a bio with REQ_PREFLUSH is submitted to dm, __send_empty_flush()
generates a flush_bio with REQ_OP_WRITE | REQ_PREFLUSH | REQ_SYNC,
which causes the flush_bio to be throttled by wbt_wait().
An example from v5.4, similar problem also exists in upstream:
After commit 2def2845cc33 ("xfs: don't allow log IO to be throttled"),
the metadata submitted by xlog_write_iclog() should not be throttled.
But due to the existence of the dm layer, throttling flush_bio indirectly
causes the metadata bio to be throttled.
Fix this by conditionally adding REQ_IDLE to flush_bio.bi_opf, which makes
wbt_should_throttle() return false to avoid wbt_wait().
In `set_kcfg_value_str`, an untrusted string is accessed with the assumption
that it will be at least two characters long due to the presence of checks for
opening and closing quotes. But the check for the closing quote
(value[len - 1] != '"') misses the fact that it could be checking the opening
quote itself in case of an invalid input that consists of just the opening
quote.
This commit adds an explicit check to make sure the string is at least two
characters long.
The register that enables selecting a test-pattern to be outputted in
free-run mode (FREE_RUN_PAT_SEL[2:0]) is only available on adv7280 based
devices, not the adv7180 based ones.
Add a flag to mark devices that are capable of generating test-patterns,
and those that are not. And only register the control on supported
devices.
When giving up on making a high-confidence prediction,
get_typical_interval() always returns UINT_MAX which means that the
next idle interval prediction will be based entirely on the time till
the next timer. However, the information represented by the most
recent intervals may not be completely useless in those cases.
Namely, the largest recent idle interval is an upper bound on the
recently observed idle duration, so it is reasonable to assume that
the next idle duration is unlikely to exceed it. Moreover, this is
still true after eliminating the suspected outliers if the sample
set still under consideration is at least as large as 50% of the
maximum sample set size.
Accordingly, make get_typical_interval() return the current maximum
recent interval value in that case instead of UINT_MAX.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Reported-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Tested-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Reviewed-by: Christian Loehle <christian.loehle@arm.com> Tested-by: Christian Loehle <christian.loehle@arm.com> Tested-by: Aboorva Devarajan <aboorvad@linux.ibm.com> Link: https://patch.msgid.link/7770672.EvYhyI6sBW@rjwysocki.net Signed-off-by: Sasha Levin <sashal@kernel.org>
Depending on the type of panics, it was found that the
__register_nmi_handler() function can be called in NMI context from
nmi_shootdown_cpus() leading to a lockdep splat:
WARNING: inconsistent lock state
inconsistent {INITIAL USE} -> {IN-NMI} usage.
In this particular case, the following panic message was printed before:
Kernel panic - not syncing: Fatal hardware error!
This message seemed to be given out from __ghes_panic() running in
NMI context.
The __register_nmi_handler() function which takes the nmi_desc lock
with irq disabled shouldn't be called from NMI context as this can
lead to deadlock.
The nmi_shootdown_cpus() function can only be invoked once. After the
first invocation, all other CPUs should be stuck in the newly added
crash_nmi_callback() and cannot respond to a second NMI.
Fix it by adding a new emergency NMI handler to the nmi_desc
structure and provide a new set_emergency_nmi_handler() helper to set
crash_nmi_callback() in any context. The new emergency handler will
preempt other handlers in the linked list. That will eliminate the need
to take any lock and serve the panic in NMI use case.
Signed-off-by: Waiman Long <longman@redhat.com> Signed-off-by: Ingo Molnar <mingo@kernel.org> Acked-by: Rik van Riel <riel@surriel.com> Cc: Thomas Gleixner <tglx@linutronix.de> Link: https://lore.kernel.org/r/20250206191844.131700-1-longman@redhat.com Signed-off-by: Sasha Levin <sashal@kernel.org>
[WHY]
If max_downscale_src_width check fails, we exit early from TAP calculation and left a NULL
value to the scaling data structure to cause the zero divide in the DML validation.
[HOW]
Call set default TAP calculation before early exit in get_optimal_number_of_taps due to
max downscale limit exceed.
Reviewed-by: Samson Tam <samson.tam@amd.com> Signed-off-by: Yihan Zhu <Yihan.Zhu@amd.com> Signed-off-by: Zaeem Mohamed <zaeem.mohamed@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The syscon helper device_node_to_regmap() is used to fetch a regmap
registered to a device node. It also currently creates this regmap
if the node did not already have a regmap associated with it. This
should only be used on "syscon" nodes. This driver is not such a
device and instead uses device_node_to_regmap() on its own node as
a hacky way to create a regmap for itself.
This will not work going forward and so we should create our regmap
the normal way by defining our regmap_config, fetching our memory
resource, then using the normal regmap_init_mmio() function.
Normally, a bond uses the MAC address of the first added slave as the bond’s
MAC address. And the bond will set active slave’s MAC address to bond’s
address if fail_over_mac is set to none (0) or follow (2).
When the first slave is removed, the bond will still use the removed slave’s
MAC address, which can lead to a duplicate MAC address and potentially cause
issues with the switch. To avoid confusion, let's warn the user in all
situations, including when fail_over_mac is set to 2 or not in active-backup
mode.
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com> Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org> Link: https://patch.msgid.link/20250225033914.18617-1-liuhangbin@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
mtk_foe_entry_set_vlan() in mtk_ppe.c already supports double vlan
tagging, but mtk_flow_offload_replace() in mtk_ppe_offload.c only allows
for 1 vlan tag, optionally in combination with pppoe and dsa tags.
However, mtk_foe_entry_set_vlan() only allows for setting the vlan id.
The protocol cannot be set, it is always ETH_P_8021Q, for inner and outer
tag. This patch adds QinQ support to mtk_flow_offload_replace(), only in
the case that both inner and outer tags are ETH_P_8021Q.
Only PPPoE-in-Q (as before) and Q-in-Q are allowed. A combination
of PPPoE and Q-in-Q is not allowed.
To fix this, KFD create process move flush release work outside
kfd_process_mutex.
Signed-off-by: Philip Yang <Philip.Yang@amd.com> Reviewed-by: Felix Kuehling <felix.kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Modify gro.sh to return a useful exit code when the -t flag is used. It
formerly returned 0 no matter what.
Tested: Ran `gro.sh -t large` and verified that test failures return 1. Signed-off-by: Kevin Krakauer <krakauer@google.com> Reviewed-by: Willem de Bruijn <willemb@google.com> Link: https://patch.msgid.link/20250226192725.621969-2-krakauer@google.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
In case health counter has not increased for few polling intervals, miss
counter will reach max misses threshold and health report will be
triggered for FW health reporter. In case syndrome found on same health
poll another health report will be triggered.
Avoid two health reports on same syndrome by marking this syndrome as
already known.
Signed-off-by: Moshe Shemesh <moshe@nvidia.com> Reviewed-by: Shahar Shitrit <shshitrit@nvidia.com> Signed-off-by: Tariq Toukan <tariqt@nvidia.com> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
Apparently nobody can figure out where the old logic came from, but it
seems like it has never been actually used on any supported firmware to
this day. OSLog buffers were apparently never requested.
But starting with 13.3, we actually need this implemented properly for
MTP (and later AOP) to work, so let's actually do that.
rtkit messages as communication with the DCP firmware for framebuffer
swaps or input events are time critical so use WQ_HIGHPRI to prevent
user space CPU load to increase latency.
With kwin_wayland 6's explicit sync mode user space load was able to
delay the IOMFB rtkit communication enough to miss vsync for surface
swaps. Minimal test scenario is constantly resizing a glxgears
Xwayland window.
Increase the timeout for SDM (Secure device manager) data credits from
20ms to 40ms. Internal stress tests running at 500 loops failed with the
current timeout of 20ms. At the start of a FPGA configuration, the CVP
host driver reads the transmit credits from SDM. It then sends bitstream
FPGA data to SDM based on the total credits. Each credit allows the
CVP host driver to send 4kBytes of data. There are situations whereby,
the SDM did not respond in time during testing.
In preparation for adding support for newer DPI instances which
do support direct-pin but do not have any H_FRE_CON register,
like the one found in MT8195 and MT8188, add a branch to check
if the reg_h_fre_con variable was declared in the mtk_dpi_conf
structure for the probed SoC DPI version.
As a note, this is useful specifically only for cases in which
the support_direct_pin variable is true, so mt8195-dpintf is
not affected by any issue.
For sama7g5 and sama7d65 backup mode, we encountered a "ZQ calibrate error"
during recalibrating the impedance in BootStrap.
We found that the impedance value saved in at91_suspend_finish() before
the DDR entered self-refresh mode did not match the resistor values. The
ZDATA field in the DDR3PHY_ZQ0CR0 register uses a modified gray code to
select the different impedance setting.
But these gray code are incorrect, a workaournd from design team fixed the
bug in the calibration logic. The ZDATA contains four independent impedance
elements, but the algorithm combined the four elements into one. The elements
were fixed using properly shifted offsets.
Signed-off-by: Li Bin <bin.li@microchip.com>
[nicolas.ferre@microchip.com: fix indentation and combine 2 patches] Signed-off-by: Nicolas Ferre <nicolas.ferre@microchip.com> Tested-by: Ryan Wanner <Ryan.Wanner@microchip.com> Tested-by: Durai Manickam KR <durai.manickamkr@microchip.com> Tested-by: Andrei Simion <andrei.simion@microchip.com> Signed-off-by: Ryan Wanner <Ryan.Wanner@microchip.com> Link: https://lore.kernel.org/r/28b33f9bcd0ca60ceba032969fe054d38f2b9577.1740671156.git.Ryan.Wanner@microchip.com Signed-off-by: Claudiu Beznea <claudiu.beznea@tuxon.dev> Signed-off-by: Sasha Levin <sashal@kernel.org>
Change the default value of spectre v2 in user mode to respect the
CONFIG_MITIGATION_SPECTRE_V2 config option.
Currently, user mode spectre v2 is set to auto
(SPECTRE_V2_USER_CMD_AUTO) by default, even if
CONFIG_MITIGATION_SPECTRE_V2 is disabled.
Set the spectre_v2 value to auto (SPECTRE_V2_USER_CMD_AUTO) if the
Spectre v2 config (CONFIG_MITIGATION_SPECTRE_V2) is enabled, otherwise
set the value to none (SPECTRE_V2_USER_CMD_NONE).
Important to say the command line argument "spectre_v2_user" overwrites
the default value in both cases.
When CONFIG_MITIGATION_SPECTRE_V2 is not set, users have the flexibility
to opt-in for specific mitigations independently. In this scenario,
setting spectre_v2= will not enable spectre_v2_user=, and command line
options spectre_v2_user and spectre_v2 are independent when
CONFIG_MITIGATION_SPECTRE_V2=n.
The IMX8MPCEC datasheet lists maximum frequencies allowed for different
modules. Some of these limits are universal, but some depend on
whether the SoC is operating in nominal or in overdrive mode.
The imx8mp.dtsi currently assumes overdrive mode and configures some
clocks in accordance with this. Boards wishing to make use of nominal
mode will need to override some of the clock rates manually.
As operating the clocks outside of their allowed range can lead to
difficult to debug issues, it makes sense to register the maximum rates
allowed in the driver, so the CCF can take them into account.
Tested-by: Yunke Cao <yunkec@google.com> Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Ricardo Ribalda <ribalda@chromium.org> Link: https://lore.kernel.org/r/20250203-uvc-roi-v17-15-5900a9fed613@chromium.org Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl> Signed-off-by: Sasha Levin <sashal@kernel.org>
In the first ioctl() path, rtentry_to_fib_config() checks the prefix
length with bad_mask(). Also, fib_magic() always passes the correct
prefix: 32 or ifa->ifa_prefixlen, which is already validated.
Let's move fib_valid_key_len() to the rtnetlink path, rtm_to_fib_config().
While at it, 2 direct returns in rtm_to_fib_config() are changed to
goto to match other places in the same function
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com> Reviewed-by: Eric Dumazet <edumazet@google.com> Reviewed-by: David Ahern <dsahern@kernel.org> Link: https://patch.msgid.link/20250228042328.96624-12-kuniyu@amazon.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
PC speaker works well on this platform in BIOS and in Linux until sound
card drivers are loaded. Then it stops working.
There seems to be a beep generator node at 0x1a in this CODEC
(ALC269_TYPE_ALC215) but it seems to be only connected to capture mixers
at nodes 0x22 and 0x23.
If I unmute the mixer input for 0x1a at node 0x23 and start recording
from its "ALC285 Analog" capture device I can clearly hear beeps in that
recording.
So the beep generator is indeed working properly, however I wasn't able to
figure out any way to connect it to speakers.
However, the bits in the "Passthrough Control" register (0x36) seems to
work at least partially: by zeroing "B" and "h" and setting "S" I can at
least make the PIT PC speaker output appear either in this laptop speakers
or headphones (depending on whether they are connected or not).
There are some caveats, however:
* If the CODEC gets runtime-suspended the beeps stop so it needs HDA beep
device for keeping it awake during beeping.
* If the beep generator node is generating any beep the PC beep passthrough
seems to be temporarily inhibited, so the HDA beep device has to be
prevented from using the actual beep generator node - but the beep device
is still necessary due to the previous point.
* In contrast with other platforms here beep amplification has to be
disabled otherwise the beeps output are WAY louder than they were on pure
BIOS setup.
Unless someone (from Realtek probably) knows how to make the beep generator
node output appear in speakers / headphones using PC beep passthrough seems
to be the only way to make PC speaker beeping actually work on this
platform.
Currently, __reserve_bp_slot() returns -ENOSPC for unsupported
breakpoint types on the architecture. For example, powerpc
does not support hardware instruction breakpoints. This causes
the perf_skip BPF selftest to fail, as neither ENOENT nor
EOPNOTSUPP is returned by perf_event_open for unsupported
breakpoint types. As a result, the test that should be skipped
for this arch is not correctly identified.
To resolve this, hw_breakpoint_event_init() should exit early by
checking for unsupported breakpoint types using
hw_breakpoint_slots_cached() and return the appropriate error
(-EOPNOTSUPP).
SRIOV VF does not have write access to AGP BAR regs.
Skip the writes to avoid a dmesg warning.
Signed-off-by: Victor Lu <victorchengchi.lu@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Both MSM8974 and MSM8226 have only CX as power domain with MX & PX being
handled as regulators. Handle this case by reodering pd_names to have CX
first, and handling that the driver core will already attach a single
power domain internally.
Signed-off-by: Matti Lehtimäki <matti.lehtimaki@gmail.com>
[luca: minor changes] Signed-off-by: Luca Weiss <luca@lucaweiss.eu> Link: https://lore.kernel.org/r/20250206-wcnss-singlepd-v2-2-9a53ee953dee@lucaweiss.eu
[bjorn: Added missing braces to else after multi-statement if] Signed-off-by: Bjorn Andersson <andersson@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The phylink_expects_phy() function allows MAC drivers to check if they are
expecting a PHY to attach. The checking condition in phylink_expects_phy()
aims to achieve the same result as the checking condition in
phylink_attach_phy().
However, the checking condition in phylink_expects_phy() uses
pl->link_config.interface, while phylink_attach_phy() uses
pl->link_interface.
Initially, both pl->link_interface and pl->link_config.interface are set
to SGMII, and pl->cfg_link_an_mode is set to MLO_AN_INBAND.
When the interface switches from SGMII to 2500BASE-X,
pl->link_config.interface is updated by phylink_major_config().
At this point, pl->cfg_link_an_mode remains MLO_AN_INBAND, and
pl->link_config.interface is set to 2500BASE-X.
Subsequently, when the STMMAC interface is taken down
administratively and brought back up, it is blocked by
phylink_expects_phy().
Since phylink_expects_phy() and phylink_attach_phy() aim to achieve the
same result, phylink_expects_phy() should check pl->link_interface,
which never changes, instead of pl->link_config.interface, which is
updated by phylink_major_config().
A cache device failing to resume due to mapping errors should not be
retried, as the failure leaves a partially initialized policy object.
Repeating the resume operation risks triggering BUG_ON when reloading
cache mappings into the incomplete policy object.
Reproduce steps:
1. create a cache metadata consisting of 512 or more cache blocks,
with some mappings stored in the first array block of the mapping
array. Here we use cache_restore v1.0 to build the metadata.
An of_node_put(i2c_bus) call was immediately used after a pointer check
for an of_find_i2c_adapter_by_node() call in this function implementation.
Thus call such a function only once instead directly before the check.
This issue was transformed by using the Coccinelle software.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net> Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl> Signed-off-by: Sasha Levin <sashal@kernel.org>
snd_seq_poll() calls snd_seq_write_pool_allocated() that reads out a
field in client->pool object, while it can be updated concurrently via
ioctls, as reported by syzbot. The data race itself is harmless, as
it's merely a poll() call, and the state is volatile. OTOH, the read
out of poll object info from the caller side is fragile, and we can
leave it better in snd_seq_pool_poll_wait() alone.
A similar pattern is seen in snd_seq_kernel_client_write_poll(), too,
which is called from the OSS sequencer.
This patch drops the pool checks from the caller side and add the
pool->lock in snd_seq_pool_poll_wait() for better data consistency.
According to the board schematics the enable pin of this regulator is
connected to gpio line #9 of the first instance of the TCA9539
GPIO expander, so adjust it.
Unlike the decompression code, the compression code in LZO never
checked for output overruns. It instead assumes that the caller
always provides enough buffer space, disregarding the buffer length
provided by the caller.
Add a safe compression interface that checks for the end of buffer
before each write. Use the safe interface in crypto/lzo.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Sasha Levin <sashal@kernel.org>
Commit 2545c1c948a6 ("auxdisplay: Move hwidth and bwidth to struct
hd44780_common") makes charlcd_alloc() argument-less effectively dropping
the single allocation for the struct charlcd_priv object along with
the driver specific one. Restore that behaviour here.
[WHY]
In some cases the remain de-tile buffer segments will be greater
than zero if we don't add the non-top pipe to calculate, at
this time the override de-tile buffer size will be valid and used.
But it makes the de-tile buffer segments used finally for all of pipes
exceed the maximum.
[HOW]
Add the non-top pipe to calculate the remain de-tile buffer segments.
Don't set override size to use the average according to pipe count
if the value exceed the maximum.
Reviewed-by: Charlene Liu <charlene.liu@amd.com> Signed-off-by: Zhikai Zhai <zhikai.zhai@amd.com> Signed-off-by: Tom Chung <chiahsuan.chung@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When spanning datagram construction over multiple send calls using
MSG_MORE, per datagram settings are configured on the first send.
That is when ip(6)_setup_cork stores these settings for subsequent use
in __ip(6)_append_data and others.
The only flag that escaped this was dontfrag. As a result, a datagram
could be constructed with df=0 on the first sendmsg, but df=1 on a
next. Which is what cmsg_ip.sh does in an upcoming MSG_MORE test in
the "diff" scenario.
Changing datagram conditions in the middle of constructing an skb
makes this already complex code path even more convoluted. It is here
unintentional. Bring this flag in line with expected sockopt/cmsg
behavior.
And stop passing ipc6 to __ip6_append_data, to avoid such issues
in the future. This is already the case for __ip_append_data.
inet6_cork had a 6 byte hole, so the 1B flag has no impact.
projects / thirdparty / kernel / stable.git / log
