commit 3644f44117 upstream.
Clang warns (or errors with CONFIG_WERROR=y / W=e):
drivers/hid/intel-ish-hid/ipc/ipc.c:935:36: error: cast from 'void (*)(struct workqueue_struct *)' to 'void (*)(void *)' converts to incompatible function type [-Werror,-Wcast-function-type-strict]
935 | if (devm_add_action_or_reset(dev, (void (*)(void *))destroy_workqueue,
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
include/linux/device/devres.h:168:34: note: expanded from macro 'devm_add_action_or_reset'
168 | __devm_add_action_or_ireset(dev, action, data, #action)
| ^~~~~~
This warning is pointing out a kernel control flow integrity (kCFI /
CONFIG_CFI=y) violation will occur due to this function cast when the
destroy_workqueue() is indirectly called via devm_action_release()
because the prototype of destroy_workqueue() does not match the
prototype of (*action)().
Use a local function with the correct prototype to wrap
destroy_workqueue() to resolve the warning and CFI violation.
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202510190103.qTZvfdjj-lkp@intel.com/
Closes: https://github.com/ClangBuiltLinux/linux/issues/2139
Fixes: 0d30dae38f ("HID: intel-ish-hid: Use dedicated unbound workqueues to prevent resume blocking")
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Reviewed-by: Zhang Lixu <lixu.zhang@intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 0d30dae38f upstream.
During suspend/resume tests with S2IDLE, some ISH functional failures were
observed because of delay in executing ISH resume handler. Here
schedule_work() is used from resume handler to do actual work.
schedule_work() uses system_wq, which is a per CPU work queue. Although
the queuing is not bound to a CPU, but it prefers local CPU of the caller,
unless prohibited.
Users of this work queue are not supposed to queue long running work.
But in practice, there are scenarios where long running work items are
queued on other unbound workqueues, occupying the CPU. As a result, the
ISH resume handler may not get a chance to execute in a timely manner.
In one scenario, one of the ish_resume_handler() executions was delayed
nearly 1 second because another work item on an unbound workqueue occupied
the same CPU. This delay causes ISH functionality failures.
A similar issue was previously observed where the ISH HID driver timed out
while getting the HID descriptor during S4 resume in the recovery kernel,
likely caused by the same workqueue contention problem.
Create dedicated unbound workqueues for all ISH operations to allow work
items to execute on any available CPU, eliminating CPU-specific bottlenecks
and improving resume reliability under varying system loads. Also ISH has
three different components, a bus driver which implements ISH protocols, a
PCI interface layer and HID interface. Use one dedicated work queue for all
of them.
Signed-off-by: Zhang Lixu <lixu.zhang@intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 7e0d1cff12 ]
The ISH driver performs a clock sync with the firmware once at system
startup and then every 20 seconds. If a firmware reset occurs right
after a clock sync, the driver would wait 20 seconds before performing
another clock sync with the firmware. This is particularly problematic
with the introduction of the "load firmware from host" feature, where
the driver performs a clock sync with the bootloader and then has to
wait 20 seconds before syncing with the main firmware.
This patch clears prev_sync immediately upon receiving an IPC reset,
so that the main firmware and driver will perform a clock sync
immediately after completing the IPC handshake.
Signed-off-by: Zhang Lixu <lixu.zhang@intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 4b54ae6919 ]
The timestamps in the Firmware log and HID sensor samples are incorrect.
They show 1970-01-01 because the current IPC driver only uses the first
8 bytes of bootup time when synchronizing time with the firmware. The
firmware converts the bootup time to UTC time, which results in the
display of 1970-01-01.
In write_ipc_from_queue(), when sending the MNG_SYNC_FW_CLOCK message,
the clock is updated according to the definition of ipc_time_update_msg.
However, in _ish_sync_fw_clock(), the message length is specified as the
size of uint64_t when building the doorbell. As a result, the firmware
only receives the first 8 bytes of struct ipc_time_update_msg.
This patch corrects the length in the doorbell to ensure the entire
ipc_time_update_msg is sent, fixing the timestamp issue.
Signed-off-by: Zhang Lixu <lixu.zhang@intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
Enhance the firmware loader to support the loading of vendor-specific
customized firmware for the Intel Integrated Sensor Hub (ISH). The
loader now constructs firmware file names based on the DMI_SYS_VENDOR,
DMI_PRODUCT_NAME, and DMI_PRODUCT_SKU information in Desktop Management
Interface (DMI). The loader will attempt to load the firmware files
following a specific naming convention in sequence. If successful, it
will skip the remaining files.
For more details, please refer to Documentation/hid/intel-ish-hid.rst.
Signed-off-by: Zhang Lixu <lixu.zhang@intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.com>
ISH allows vendors to customize ISH firmware. To differentiate different
vendors and load the correct firmware, Intel defined a firmware file
name format. As part of the filename, there is a "generation" string. To
load correct format the driver must know the generation name to create
file name.
In the absence of any vendor specific firmware, default ISH firmware is
loaded.
Currently full ISH firmware file name is stored as part of driver data.
This file name already includes the generation name. For example, for
Lunar Lake, the name is ish_lnlm.bin, where "lnlm" is the generation.
So instead of storing both generation name and ISH default firmware file
name, just store generation name and create the default ISH firmware
file name string during initialization.
No functional changes are expected.
Signed-off-by: Zhang Lixu <lixu.zhang@intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.com>
Add a check for the return value of pci_alloc_irq_vectors() and return
error if it fails.
[jkosina@suse.com: reworded changelog based on Srinivas' suggestion]
Fixes: 74fbc7d371 ("HID: intel-ish-hid: add MSI interrupt support")
Signed-off-by: Chen Ni <nichen@iscas.ac.cn>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.com>
This patch enhances the firmware reset handler in the Intel Integrated
Sensor Hub (ISH) driver. Previously, the ISH firmware would send a
MNG_RESET_NOTIFY message in response to an empty IPC message from the
ish_wakeup function. With the introduction of the feature to load ISH
firmware from the host on the LunarLake platform, the ISH bootloader
now involves the IPC function. This results in an additional
MNG_RESET_NOTIFY message being sent by ISH bootloader after power on.
Consequently, the driver receives two MNG_RESET_NOTIFY messages during
system boot up. This can disrupt the dev->dev_state during the first
reset flow due to the subsequent reset notify message.
To address this, the patch modifies the fw_reset_work_fn function to skip
the execution of ishtp_reset_compl_handler during the first reset flow if
a reset is pending. The ishtp_reset_compl_handler will then be executed
during the second reset flow, ensuring the dev->dev_state is not disrupted.
Signed-off-by: Zhang Lixu <lixu.zhang@intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.com>
Introduces a new structure, ishtp_driver_data, to hold driver-specific
data, including the firmware filename for different hardware variants of
the Intel Integrated Sensor Hub (ISH).
Signed-off-by: Zhang Lixu <lixu.zhang@intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.com>
The variable dev->devc in ish_dev_init was utilized by dev_err before it
was properly assigned. To rectify this, the assignment of dev->devc has
been moved to immediately follow memory allocation.
Without this change "(NULL device *)" is printed for device information.
Fixes: 8ae2f2b0a2 ("HID: intel-ish-hid: ipc: Fix potential use-after-free in work function")
Fixes: ae02e5d40d ("HID: intel-ish-hid: ipc layer")
Signed-off-by: Zhang Lixu <lixu.zhang@intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.com>
Since PCI core and ACPI core already handles PCI PME wake and GPE wake
when the device has wakeup capability, use device_init_wakeup() to let
them do the wakeup setting work.
Also add a shutdown callback which uses pci_prepare_to_sleep() to let
PCI and ACPI set OOB wakeup for S5.
Cc: Jian Hui Lee <jianhui.lee@canonical.com>
Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.com>
The EHL (Elkhart Lake) based platforms provide a OOB (Out of band)
service, which allows to wakup device when the system is in S5 (Soft-Off
state). This OOB service can be enabled/disabled from BIOS settings. When
enabled, the ISH device gets PME wake capability. To enable PME wakeup,
driver also needs to enable ACPI GPE bit.
On resume, BIOS will clear the wakeup bit. So driver need to re-enable it
in resume function to keep the next wakeup capability. But this BIOS
clearing of wakeup bit doesn't decrement internal OS GPE reference count,
so this reenabling on every resume will cause reference count to overflow.
So first disable and reenable ACPI GPE bit using acpi_disable_gpe().
Fixes: 2e23a70eda ("HID: intel-ish-hid: ipc: finish power flow for EHL OOB")
Reported-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Closes: https://lore.kernel.org/lkml/CAAd53p4=oLYiH2YbVSmrPNj1zpMcfp=Wxbasb5vhMXOWCArLCg@mail.gmail.com/T/
Tested-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
When a reset notify IPC message is received, the ISR schedules a work
function and passes the ISHTP device to it via a global pointer
ishtp_dev. If ish_probe() fails, the devm-managed device resources
including ishtp_dev are freed, but the work is not cancelled, causing a
use-after-free when the work function tries to access ishtp_dev. Use
devm_work_autocancel() instead, so that the work is automatically
cancelled if probe fails.
Signed-off-by: Reka Norman <rekanorman@chromium.org>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Specify that both TGL and ADL don't support DMA cache snooping.
Signed-off-by: Ye Xiang <xiang.ye@intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
When OOB is disabled, FW will be power gated when system is in S3/S4/S5
which is the same behavior with legacy ISH FW.
When OOB is enabled, FW will always power on which is totally different
comparing to legacy ISH FW.
So NO_D3 flag is not enough to check FW's status after resume.
Here we can use IPC FW status register to check host link status.
If it is false, it means FW get reset after power gated, need go through
the whole initialization flow; If it is true, it means FW is alive, just
set host ready bit to let fw know host is up.
Co-developed-by: Wei Jiang <wei.w.jiang@intel.com>
Signed-off-by: Wei Jiang <wei.w.jiang@intel.com>
Signed-off-by: Even Xu <even.xu@intel.com>
Signed-off-by: Zhang Lixu <lixu.zhang@intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
The EHL (Elkhart Lake) based platforms provide a OOB (Out of band)
service, which allows wakup device when the system is in S5 (Soft-Off
state). This OOB service can be enabled/disabled from BIOS settings. When
enabled, the ISH device gets PME wake capability. To enable PME wakeup,
driver also needs to enable ACPI GPE bit.
Once wakeup, BIOS will clear the wakeup bit to identify wakeup is
successful. So driver need to re-enable it in resume function to
keep the next wakeup capability.
Since this feature is only present on EHL, we use EHL PCI device id to
enable this feature.
Co-developed-by: Najumon Ba <najumon.ba@intel.com>
Signed-off-by: Najumon Ba <najumon.ba@intel.com>
Signed-off-by: Even Xu <even.xu@intel.com>
Signed-off-by: Zhang Lixu <lixu.zhang@intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
PCI devices support two variants of the D3 power state: D3hot (main power
present) D3cold (main power removed). Previously struct pci_dev contained:
unsigned int d3_delay; /* D3->D0 transition time in ms */
unsigned int d3cold_delay; /* D3cold->D0 transition time in ms */
"d3_delay" refers specifically to the D3hot state. Rename it to
"d3hot_delay" to avoid ambiguity and align with the ACPI "_DSM for
Specifying Device Readiness Durations" in the PCI Firmware spec r3.2,
sec 4.6.9.
There is no change to the functionality.
Link: https://lore.kernel.org/r/20200730210848.1578826-1-kw@linux.com
Signed-off-by: Krzysztof Wilczyński <kw@linux.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
PCI_DEV_FLAGS_NO_D3 should not be used outside of PCI core.
Instead, we can use pci_save_state() to hint PCI core that the device
should stay at D0 during suspend. By doing so, PCI core will let the
upstream bridges also stays at D0 so the entire hierarchy is in the
correct power state as PCI spec mandates.
Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
The NO_D3 flag would be set if the ISH enter D0i3 in ish_suspend(),
The resume paths can be distinguished by checking the NO_D3 flag.
It's more reasonable than checking the FW status.
Signed-off-by: Zhang Lixu <lixu.zhang@intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
For suspend-to-idle, send suspend message and set N0_D3 flag to put
the ISH into D0i3 state.
For suspend-to-mem, disable the DMA bit before ISH entering D3, and
NO_D3 flag is cleared by default, then the ISH would enter D3.
Signed-off-by: Zhang Lixu <lixu.zhang@intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Currently, the NO_D3 flag is set in ish_probe(), and cleared in
ish_remove(). So even if the system goes into S3, ISH is still
in D0i3 state. It makes more sense that put ISH into D3 as system
goes into S3 and put ISH into D0i3 as system goes into suspend-to-idle.
I remove the NO_D3 setting in ish_probe(), so that ISH can enter
D3 state when system enters S3. Only set N0_D3 flag when system
enters the suspend-to-idle or platform specified, and clear it
when system resume.
When the ISH enters D3, the FW will check the DMA bit status.
If the DMA bit is set, the FW will reset automatically. So the
DMA bit need be clear before putting ISH into D3 state.
Signed-off-by: Zhang Lixu <lixu.zhang@intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Based on 1 normalized pattern(s):
this program is free software you can redistribute it and or modify
it under the terms and conditions of the gnu general public license
version 2 as published by the free software foundation this program
is distributed in the hope it will be useful but without any
warranty without even the implied warranty of merchantability or
fitness for a particular purpose see the gnu general public license
for more details
extracted by the scancode license scanner the SPDX license identifier
GPL-2.0-only
has been chosen to replace the boilerplate/reference in 263 file(s).
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Allison Randal <allison@lohutok.net>
Reviewed-by: Alexios Zavras <alexios.zavras@intel.com>
Cc: linux-spdx@vger.kernel.org
Link: https://lkml.kernel.org/r/20190529141901.208660670@linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently we are using one additional static variable and a spinlock to
prevent contention of writing IPC messages to ISH hardware, which is
not necessary. Once ISH is ready to accept new data, we can push new
data to hardware. This pushing of new data is already protected by
wr_processing_spinlock for contention, which is enough. So use this
spinlock to check both readiness for accepting new data and once ready
allow writing of ipc message from queue to ISH hardware.
While here, cleaned up some space after return.
Signed-off-by: Hong Liu <hong.liu@intel.com>
Tested-by: Hongyan Song <hongyan.song@intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Host driver should handle interrupt mask register earlier than wake up ish FW
else there will be conditions when FW interrupt comes, host PIMR register still
not set ready, so move the interrupt mask setting before ish_wakeup.
Clear PISR busy_clear bit in ish_irq_handler. If not clear, there will be
conditions host driver received a busy_clear interrupt (before the busy_clear
mask bit is ready), it will return IRQ_NONE after check_generated_interrupt,
the interrupt will never be cleared, causing the DEVICE not sending following
IRQ.
Since PISR clear should not be called for the CHV device we do this change.
After the change, both ISH2HOST interrupt and busy_clear interrupt will be
considered as interrupt from ISH, busy_clear interrupt will return IRQ_HANDLED
from IPC_IS_BUSY check.
Signed-off-by: Song Hongyan <hongyan.song@intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Now ish hid ipc only support sideband interrupt but on some platforms
they use MSI interrupt.
In order to make the interrupt type coverage all the scenario add
single MSI interrupt support, it can match all interrupt types.
Signed-off-by: Song Hongyan <hongyan.song@intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Nathan Chancellor
Zhang Lixu
Jiri Kosina
Chen Ni
Even Xu
Kai-Heng Feng
Srinivas Pandruvada
Reka Norman
Li Qiong
Ye Xiang
Thomas Weißschuh
Jiri Kosina
Lee Jones
You-Sheng Yang
Krzysztof Wilczyński
Song Hongyan
Thomas Gleixner
Hong Liu