Traditionally, firmware loading for Serial Engines (SE) in the QUP hardware
of Qualcomm SoCs has been managed by TrustZone (TZ). While this approach
ensures secure SE assignment and access control, it limits flexibility for
developers who need to enable various protocols on different SEs.
Add the firmware-name property to QUPv3 nodes in the device tree to enable
firmware loading from the Linux environment. Handle SE assignments and
access control permissions directly within Linux, removing the dependency
on TrustZone.
Signed-off-by: Viken Dadhaniya <viken.dadhaniya@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20250925042605.1388951-1-viken.dadhaniya@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Traditionally, firmware loading for Serial Engines (SE) in the QUP hardware
of Qualcomm SoCs has been managed by TrustZone (TZ). While this approach
ensures secure SE assignment and access control, it limits flexibility for
developers who need to enable various protocols on different SEs.
Add the firmware-name property to QUPv3 nodes in the device tree to enable
firmware loading from the Linux environment. Handle SE assignments and
access control permissions directly within Linux, removing the dependency
on TrustZone.
Signed-off-by: Viken Dadhaniya <viken.dadhaniya@oss.qualcomm.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Acked-by: Mukesh Kumar Savaliya <mukesh.savaliya@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20250924035409.3976652-1-viken.dadhaniya@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Traditionally, firmware loading for Serial Engines (SE) in the QUP hardware
of Qualcomm SoCs has been managed by TrustZone (TZ). While this approach
ensures secure SE assignment and access control, it limits flexibility for
developers who need to enable various protocols on different SEs.
Add the firmware-name property to QUPv3 nodes in the device tree to enable
firmware loading from the Linux environment. Handle SE assignments and
access control permissions directly within Linux, removing the dependency
on TrustZone.
Signed-off-by: Viken Dadhaniya <viken.dadhaniya@oss.qualcomm.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20250923161107.3541698-1-viken.dadhaniya@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
The BQ Aquaris X5 (Longcheer L8910) has a Himax HX852x-ES touchscreen,
which can now be described with the bindings recently added to linux-next.
Add it to the device tree to allow using the touchscreen.
Signed-off-by: Jonathan Albrieux <jonathan.albrieux@gmail.com>
Co-developed-by: Stephan Gerhold <stephan@gerhold.net>
Signed-off-by: Stephan Gerhold <stephan@gerhold.net>
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20250919-msm8916-l8910-touchscreen-v1-1-c46e56ec0a3b@gerhold.net
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Update min/max voltage settings for regulators below to align
with the HW specification
vreg_l3b_0p504
vreg_l6b_1p2
vreg_l11b_1p504
vreg_l14b_1p08
vreg_l16b_1p1
vreg_l17b_1p7
vreg_s1c_2p19
vreg_l8c_1p62
vreg_l9c_2p96
vreg_l12c_1p65.
While at it, remove RPMH regulator rails (listed below) as
these are not to be used on APPS, and any client accidently
voting on it can potentially cause issues.
vreg_s2b_0p876
vreg_s2c_0p752
vreg_s5c_0p752
vreg_s7c_0p752
vreg_s10c_0p752
vreg_l4b_0p752
vreg_l5b_0p752.
Signed-off-by: Rakesh Kota <rakesh.kota@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20250919-b4-rb3gen2-update-regulator-v1-1-1ea9e70d01cb@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Like on other platforms, if the OS does not support recovering the state
left by the bootloader it needs access to MDSS_CORE, so that it can
clear the MDSS configuration. Add a reference to the relevant reset.
This also fixes display init on Linux v6.17.
Signed-off-by: Luca Weiss <luca.weiss@fairphone.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20250919-sm6350-mdss-reset-v1-3-48dcac917c73@fairphone.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Implement spinlock support for SMP on UML/x86, leveraging x86's
spinlock implementation. In addition, to support SMP on CPUs that
do not support CX8, some additional work is required. However,
considering that such CPUs are already very outdated, and x86 is
also removing support for them [1], let's enable SMP support only
on CPUs that support CX8.
[1] https://lore.kernel.org/lkml/20250515085708.2510123-1-mingo@kernel.org/
Signed-off-by: Tiwei Bie <tiwei.btw@antgroup.com>
Link: https://patch.msgid.link/20251027001815.1666872-9-tiwei.bie@linux.dev
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Currently, asm/percpu.h is directly or indirectly included by
some assembly files on x86. Some of them (e.g., checksum_32.S)
are also used on um. But x86 and um provide different versions
of asm/percpu.h -- um uses asm-generic/percpu.h directly.
When SMP is enabled, asm-generic/percpu.h will introduce C code
that cannot be assembled. Since asm-generic/percpu.h currently
is not designed for use in assembly, and these assembly files
do not actually need asm/percpu.h on um, let's add the assembly
guard in asm-generic/percpu.h to fix this issue.
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: linux-arch@vger.kernel.org
Signed-off-by: Tiwei Bie <tiwei.btw@antgroup.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Link: https://patch.msgid.link/20251027001815.1666872-8-tiwei.bie@linux.dev
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
We are going to support SMP on UML/x86, so we can't hard code
the CPU and NUMA node in __vdso_getcpu() anymore. Let's just
remove it and let applications fall back to the syscall.
Suggested-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Tiwei Bie <tiwei.btw@antgroup.com>
Link: https://patch.msgid.link/20251027001815.1666872-7-tiwei.bie@linux.dev
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Add initial symmetric multi-processing (SMP) support to UML. With
this support enabled, users can tell UML to start multiple virtual
processors, each represented as a separate host thread.
In UML, kthreads and normal threads (when running in kernel mode)
can be scheduled and executed simultaneously on different virtual
processors. However, the userspace code of normal threads still
runs within their respective single-threaded stubs.
That is, SMP support is currently available both within the kernel
and across different processes, but still remains limited within
threads of the same process in userspace.
Signed-off-by: Tiwei Bie <tiwei.btw@antgroup.com>
Link: https://patch.msgid.link/20251027001815.1666872-6-tiwei.bie@linux.dev
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Define timers on a per-CPU basis to enable each CPU to have its
own timer. This is a preparation for adding SMP support.
Signed-off-by: Tiwei Bie <tiwei.btw@antgroup.com>
Link: https://patch.msgid.link/20251027001815.1666872-5-tiwei.bie@linux.dev
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
With SMP and NO_HZ enabled, the CPU may still need to sleep even
if the timer is disarmed. Switch to deciding whether to sleep based
on pending resched. Additionally, because disabling IRQs does not
block SIGALRM, it is also necessary to check for any pending timer
alarms. This is a preparation for adding SMP support.
Signed-off-by: Tiwei Bie <tiwei.btw@antgroup.com>
Link: https://patch.msgid.link/20251027001815.1666872-4-tiwei.bie@linux.dev
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Turn signals_enabled, signals_pending and signals_active into
thread-local variables. This enables us to control and track
signals independently on each CPU thread. This is a preparation
for adding SMP support.
Signed-off-by: Tiwei Bie <tiwei.btw@antgroup.com>
Link: https://patch.msgid.link/20251027001815.1666872-3-tiwei.bie@linux.dev
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Currently, initial_thread_cb() temporarily disables kmalloc when
it invokes the callback, allowing the callback to bypass kmalloc.
This is unnecessary for the current users of initial_thread_cb(),
and we should avoid memory allocations that are not under the
control of the UML kernel. Therefore, let's stop temporarily
disabling kmalloc in initial_thread_cb().
Signed-off-by: Tiwei Bie <tiwei.btw@antgroup.com>
Link: https://patch.msgid.link/20251027001815.1666872-2-tiwei.bie@linux.dev
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
The file-based iomem emulation was introduced to support writing
paravirtualized drivers based on emulated iomem regions. However,
the only driver that makes use of it is an example driver called
mmapper, which was written over two decades ago.
We now have several modern device emulation mechanisms, such as
vhost-user-based virtio-uml. Remove the file-based iomem emulation
support to reduce the maintenance burden.
Signed-off-by: Tiwei Bie <tiwei.btw@antgroup.com>
Link: https://patch.msgid.link/20251027054519.1996090-5-tiwei.bie@linux.dev
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Although UML_ROUND_UP() is defined in a shared header file, it
depends on the PAGE_SIZE and PAGE_MASK macros, so it can only be
used in kernel code. Considering its name is not very clear and
its functionality is the same as PAGE_ALIGN(), replace its usages
with a direct call to PAGE_ALIGN() and remove it.
Signed-off-by: Tiwei Bie <tiwei.btw@antgroup.com>
Link: https://patch.msgid.link/20251027054519.1996090-4-tiwei.bie@linux.dev
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Currently, host_task_size is a global variable, but it is only used
in linux_main() to compute stub_start and task_size. Make it a local
variable to limit its scope to where it is actually needed.
Signed-off-by: Tiwei Bie <tiwei.btw@antgroup.com>
Link: https://patch.msgid.link/20251027054519.1996090-2-tiwei.bie@linux.dev
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Apply WA_14024681466 to Xe3_LPG graphics IP versions from 30.00 to 30.05.
v2: (Matthew Roper)
- Remove stepping filter as workaround applies to all steppings.
- Add an engine class filter so it only applies to the RENDER engine.
Signed-off-by: Nitin Gote <nitin.r.gote@intel.com>
Link: https://patch.msgid.link/20251027092643.335904-1-nitin.r.gote@intel.com
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Now full support for the UMP buffers is available, it is possible to
read the SDCA HID descriptors from the device and pass them to
user-space. Add a helper function to process HID events from an SDCA
device.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
Link: https://patch.msgid.link/20251020155512.353774-20-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Some IRQs (FDL) require processing before the primary soundcard is
brought up, as the downloaded files could be firmware required for
operation of the audio functions of the device. Add a new helper
function which registers the required IRQs.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
Link: https://patch.msgid.link/20251020155512.353774-19-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Several of the UMP transactions in the FDL process should timeout if the
device does not respond within a certain time, add handling into the UMP
helpers and the FDL code to handle this.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
Link: https://patch.msgid.link/20251020155512.353774-18-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Add some completions and a helper function to allow other parts of the
system to wait for FDL to complete. The sdca_fdl_sync() function will
wait until it completes a full time out without a new FDL request
happening, this ensures that even parts requiring multiple rounds of FDL
should be fully downloaded before the driver boot continues.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
Link: https://patch.msgid.link/20251020155512.353774-17-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Hookup the XU IRQs required for the FDL process.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Maciej Strozek <mstrozek@opensource.cirrus.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
Link: https://patch.msgid.link/20251020155512.353774-16-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Some instances of the XU Entity have a need for Files to be downloaded
from the Host. In these XUs, there is one instance of a Host to Device
(Consumer) UMP, identified by the FDL_CurrentOwner Control. FDL Library
introduced here implements the FDL flow triggered by FDL_CurrentOwner
irq, which sends a file from SoundWire File Table (SWFT) or from the
firmware directory in specific cases, to the Device FDL UMP.
Currently only Direct method of FDL is implemented.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Maciej Strozek <mstrozek@opensource.cirrus.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
Link: https://patch.msgid.link/20251020155512.353774-15-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Add parsing of ACPI DisCo information specific to FDL (File DownLoad).
DisCo contains a list of File Sets which can be requested by the device
and within each of those a list of individual files to be downloaded to
the device. Optionally the contents of the files may also be present in
a special ACPI table, called SWFT (SoundWire File Table).
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Maciej Strozek <mstrozek@opensource.cirrus.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
Link: https://patch.msgid.link/20251020155512.353774-14-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Add helper functions for handling Universal Message Passing (UMP)
buffers on SDCA devices. These are generic mechanisms to pass blocks of
binary data between the host and the device, in both directions. They
are used for things like passing HID descriptors and the File Download
process.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
Link: https://patch.msgid.link/20251020155512.353774-13-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Parse the DisCo property to get the timeout for a Function Reset.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
Link: https://patch.msgid.link/20251020155512.353774-12-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Parse the DisCo properties for XU Entities.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
Link: https://patch.msgid.link/20251020155512.353774-11-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Whilst SDCA does specify an Access Mode for each Control, there is not a
1-to-1 mapping between that and ASoC's internal representation. Some
registers require being treated as volatile from the hosts perspective
even in their Access Mode is Read-Write. Add an explicit list of SDCA
controls that should be forced volatile.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
Link: https://patch.msgid.link/20251020155512.353774-10-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
In the future some IRQs (mostly the UMPs used during File
DownLoad) will need to run after the device has enumerated on the
bus but before the soundcard is actually constructed. As such
refactor more of the IRQ handling to use raw device and regmap
pointers, rather than accessing things through the component.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
Link: https://patch.msgid.link/20251020155512.353774-9-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Add a helper function to locate the sdca_interrupt structure for a given
SDCA IRQ, this makes the code a little more readable and will facilitate
some additions in the future.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
Link: https://patch.msgid.link/20251020155512.353774-8-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Currently there is a flag to indicate if an IRQ has been requested by
something outside the SDCA core, such that the core can skip requesting
that IRQ. However, it is simpler and more useful to always store the
allocated IRQ number. This will allow the core to see if the IRQ has
been requested, to perform additional operations on the IRQ, and
request IRQs in multiple phases.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
Link: https://patch.msgid.link/20251020155512.353774-7-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Store a copy of the device register map in the structure for the IRQ
handlers. This will allow the individual IRQ handlers access to the
device level register map if required.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
Link: https://patch.msgid.link/20251020155512.353774-6-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The SDCA HID code can't assume that the struct device it is passed is
the SoundWire slave device. HID is represented by a Function in SDCA and
will thus likely be implemented by a child driver. Update the code to
explicitly pass in the SoundWire slave device.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
Link: https://patch.msgid.link/20251020155512.353774-5-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
SDCA interrupt handling is a bit odd, the SDCA IRQ registers are
defined on a device level but the handling of the IRQ is at the
Function level. As such the regmap IRQ's PM runtime operations need to
be on the device itself to ensure those registers are available but an
additional runtime get is required for the Function child when the IRQ
is actually handled. Add the required manual PM runtime gets.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
Link: https://patch.msgid.link/20251020155512.353774-4-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Currently, the code assumes that the device that registered the
MBQ register map is the actual SoundWire slave device. This works
fine for all current users, however future SDCA devices will
likely be implemented with the SoundWire slave as a parent device
and separate child drivers with regmaps for each audio Function.
Update the regmap_init_sdw_mbq_cfg macro to allow these two
to be specified separately.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
Link: https://patch.msgid.link/20251020155512.353774-3-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The SDCA core processes two different struct device's at various times,
usually it is processing the struct device associated with an individual
SDCA Function driver. However, there are times that it is processing the
struct device associated with the actual SoundWire peripheral, usually
the parent of the Function device.
To ensure this distinction remains clear in the code, rename the variable
when handling the actual SoundWire device to sdev.
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
Link: https://patch.msgid.link/20251020155512.353774-2-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Commit 086d08725d ("remoteproc: create vdev subdevice with specific dma
memory pool") added an export for rproc_va_to_pa. However, since its
introduction, this symbol has not been used by any loadable modules. It
remains only referenced within remoteproc_virtio.c, which is always built
together with remoteproc_core.c.
As such, exporting rproc_va_to_pa is unnecessary, so remove the export.
No functional changes.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Acked-by: Andrew Davis <afd@ti.com>
Link: https://lore.kernel.org/r/20251016-rproc-cleanup-v3-v3-4-774083716e8a@nxp.com
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
There is no user of crc32.h, debugfs.h, of_reserved_mem.h, virtio_ids.h,
so remove from the included headers.
No functional changes.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Acked-by: Andrew Davis <afd@ti.com>
Link: https://lore.kernel.org/r/20251016-rproc-cleanup-v3-v3-3-774083716e8a@nxp.com
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Reordered the header includes in drivers/remoteproc/remoteproc_core.c
to follow alphabetical order to simplify future maintenance.
No functional changes.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Acked-by: Andrew Davis <afd@ti.com>
Link: https://lore.kernel.org/r/20251016-rproc-cleanup-v3-v3-2-774083716e8a@nxp.com
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
The variable ret is immediately assigned the return value of
mutex_lock_interruptible(), making its prior initialization to zero
unnecessary. Remove the redundant assignment
No functional changes.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Acked-by: Andrew Davis <afd@ti.com>
Link: https://lore.kernel.org/r/20251016-rproc-cleanup-v3-v3-1-774083716e8a@nxp.com
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
I recently got occasional build failures at -Os or -Oz that would always
involve waitpid(), where the assembler would complain about this:
init.s: Error: .size expression for waitpid.constprop.0 does not evaluate to a constant
And without -fno-asynchronous-unwind-tables it could also spit such
errors:
init.s:836: Error: CFI instruction used without previous .cfi_startproc
init.s:838: Error: .cfi_endproc without corresponding .cfi_startproc
init.s: Error: open CFI at the end of file; missing .cfi_endproc directive
A trimmed down reproducer is as simple as this:
int main(int argc, char **argv)
{
int ret, status;
if (argc == 0)
ret = waitpid(-1, &status, 0);
else
ret = waitpid(-1, &status, 0);
return status;
}
It produces the following asm code on x86_64:
.text
.section .text.nolibc_memmove_memcpy
.weak memmove
.weak memcpy
memmove:
memcpy:
movq %rdx, %rcx
(...)
retq
.section .text.nolibc_memset
.weak memset
memset:
xchgl %eax, %esi
movq %rdx, %rcx
pushq %rdi
rep stosb
popq %rax
retq
.type waitpid.constprop.0.isra.0, @function
waitpid.constprop.0.isra.0:
subq $8, %rsp
(...)
jmp *.L5(,%rax,8)
.section .rodata
.align 8
.align 4
.L5:
.quad .L10
(...)
.quad .L4
.text
.L10:
(...)
.cfi_def_cfa_offset 8
ret
.cfi_endproc
.LFE273:
.size waitpid.constprop.0.isra.0, .-waitpid.constprop.0.isra.0
It's a bit dense, but here's the explanation: the compiler has emitted a
".text" statement because it knows it's working in the .text section.
Then, our hand-written asm code for the mem* functions forced the section
to .text.something without the compiler knowing about it, so it thinks
the code is still being emitted for .text. As such, without any .section
statement, the waitpid.constprop.0.isra.0 label is in fact placed in the
previously created section, here .text.nolibc_memset.
The waitpid() function involves a switch/case statement that can be
turned to a jump table, which is what the compiler does with the .rodata
section, and after that it restores .text, which is no longer the
previous .text.nolibc_memset section. Then the CFI statements cross a
section, so does the .size calculation, which explains the error.
While a first approach consisting in placing an explicit ".text" at the
end of these functions was verified to work, it's still unreliable as
it depends on what the compiler remembers having emitted previously. A
better approach is to replace the ".section" with ".pushsection", and
place a ".popsection" at the end, so that these code blocks are agnostic
to where they're placed relative to other blocks.
Fixes: 553845eebd ("tools/nolibc: x86-64: Use `rep movsb` for `memcpy()` and `memmove()`")
Fixes: 12108aa8c1 ("tools/nolibc: x86-64: Use `rep stosb` for `memset()`")
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Add MSI doorbell support to reduce latency between PCI host and EP.
Before this change:
ping 169.254.172.137
64 bytes from 169.254.172.137: icmp_seq=1 ttl=64 time=0.575 ms
64 bytes from 169.254.172.137: icmp_seq=2 ttl=64 time=1.80 ms
64 bytes from 169.254.172.137: icmp_seq=3 ttl=64 time=8.19 ms
64 bytes from 169.254.172.137: icmp_seq=4 ttl=64 time=2.00 ms
After this change:
ping 169.254.144.71
64 bytes from 169.254.144.71: icmp_seq=1 ttl=64 time=0.215 ms
64 bytes from 169.254.144.71: icmp_seq=2 ttl=64 time=0.456 ms
64 bytes from 169.254.144.71: icmp_seq=3 ttl=64 time=0.448 ms
Also, change the data type of 'epf_ntb::db' to 'atomic_64' as multiple
doorbell interrupts might trigger at the same time.
Signed-off-by: Frank Li <Frank.Li@nxp.com>
[mani: added 'pci_epf_' prefix to doorbell IRQ name, minor code cleanups]
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Link: https://patch.msgid.link/20251015-vntb_msi_doorbell-v6-4-9230298b1910@nxp.com
Remove the 'allow-set-time' property from the rtc node because APPS
is prohibited from setting the hardware RTC time.
Signed-off-by: Tingguo Cheng <tingguo.cheng@oss.qualcomm.com>
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20250926-remove-rtc-allow-set-time-v1-1-76895a666939@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
SM6150 and QCS615 are two names for the same die, collectively known as
'talos'. Follow the example of other platforms and rename SM6150 to
talos.dtsi.
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20250923-rename-dts-v1-3-21888b68c781@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
The X1E80100 and several other similar names (X1E78100, X1E001DE) all
belong to the platform now known as 'hamoa'. Follow the example of
'lemans' and rename the x1e80100.dtsi to hamoa.dtsi and
x1e80100-pmics.dtsi to hamoa-pmics.dtsi.
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20250923-rename-dts-v1-2-21888b68c781@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
The QCS8300 and QCS8275 are two variants of the same die with no
difference visible to the Linux kernel, which are collectively named as
'monaco'. Rather than trying to using the name, which is not always
relevant, follow the example of 'lemans' and rename qcs8300.dtsi to
monaco.dtsi (and qcs8300-pmics.dtsi to monaco-pmics.dtsi).
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20250923-rename-dts-v1-1-21888b68c781@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
