Not sure this is requered as we do not mangle with domains, but let's
make sure this cache is clear.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
A surprising issue: while sys_set_bit() on VTD_GCMD_REG work on
apollo_lake, it seems to create border effect on elkhart_lake: it may
reset some previous bit to 0.
So switching to full write of the register at once, which works for
both.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
VT-D will only be enabled if MSI multi-vector will also be requested,
on x86_64, for all boards BUT QEMU ones.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
There is no point to enable VT-D if PCIE MSI multi-vector support is not
requested, as VT-D is uniquely being used to remap MSI multi-vector
based interrupts.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Those MMIO registers might be mapped in RAM and thus it's
really important to keep storage space for such mapping.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
QI operation are not supposed to last forever.
It's not an actual timeout based on clock, but a dummy counter instead.
That's because system clock might not have beed initialized yet, since
VT-D's init comes first (and that same init will use QI...)
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Depending on a VT-D capability, it might be necessary to flush objects
from the cache.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Looks like it's mandatory to invalidate the Interrupt Entry Cache in
VT-D and the only way to do so is to enable Queued Interface.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
IR faults are non-recoverable, so it's good to know why. Thus let's
handle the fault event and print the fault.
Other faults are printed as well.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Looks like a source of fault if pieces of IRTE are written instead of
the whole 64 bits block it belongs to.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
If VT-D's interrupt remapping is in place, all IOAPIC RTEs need to get
remapped as well (or then they will be simply blocked).
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Adding a way to pre-install the irg/vector on an irte and a way to get
an irte based on irq/vector.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
This will not only be used by MSI remapping but by all relevant
interrupts.
Fix also IRTE settings:
- handle x2apic for destination id
- destination mode is always logical (as for IOAPIC)
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
In XAPIC mode, it's possible to tell VT-D to let interrupt in compatible
format passing through.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
SHV bit depends on the number of vectors allocated.
If it's facing a multi-vector MSI array, it will set the bit.
If not the bit must be 0.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Refactor to handle this case. This is valid only when MSI multi-vector
feature is enabled.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
The is meant to fix a chicken & egg issue with MSI interrupt remapping.
Currently, drivers first connect the irq (by-passing any possible MSI
remapping), so the IRQ ends-up being remapped at the IOAPIC level which
is not what we want.
So adding a dedicated function to properly handle this case. This is
valid only for runtime dynamic IRQ connection obviously.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
And since it does yet another round of pcie_get_cap() on PCIE_MSI_CAP_ID
and PCIE_MSIX_CAP_ID, let's factorize that into a utility function and
change the relevant places to use that function instead.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
The host reassembles fragmented advertising reports from the controller.
Non-complete advertising reports from different advertisers may not be
interleaved. If non-complete advertising reports from an advertiser
is received while advertising reports from another advertiser is
reassembled, an error message is logged and the advertising report is
discarded. Future scan results may be incomplete.
Advertising reports from legacy PDUs or complete extended advertising
reports may be interleaved as these do not require reassembly.
If the controller sends more advertising data than fits in the
reassembly buffer, the data is truncated. Further advertising reports
from the advertiser are discarded until the final complete advertising
report is received and discarded.
Signed-off-by: Herman Berget <herman.berget@nordicsemi.no>
Enable the emulated GPIO driver by default if instances exist in
devicetree. This removes the need to manually enable `CONFIG_GPIO_EMUL`
when `CONFIG_GPIO` is enabled.
Signed-off-by: Jordan Yates <jordan.yates@data61.csiro.au>
General refactoring to clean up and futureproof this driver.
Remove false dependency on CONFIG_CAVS_ICTL. This requires the CAVS
interrupt mask API, but doesn't touch the interrupt controller driver.
Remove a racy check for simultaneous interrupts. This seems to have
been well intentioned, but it's needless: the spinlock around the
last_count computation guarantees that colliding interrupts will
correctly compute elapsed ticks (i.e. the last will compute and
announce zero ticks, which is correct and expected). And this opened
a tiny window where you could incorrectly ignore a just-set timeout.
Factor out the specific registers used (there are only five) into
pointer-valued macros instead of banging them directly.
Unify interrupt initialization for main and auxiliary cores.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Instead of using maskstatus to see if an interrupt has fired,
use finalstatus instead. It has been observed that some
controllers do not update maskstatus correctly with incoming
interrupts, but finalstatus works fine.
FYI, the DW driver in Linux is also using finalstatus.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Recent changes in this function made the code now compatible
with the use of LL API to perform this configuration.
Use it to simplify the driver.
Signed-off-by: Erwan Gouriou <erwan.gouriou@linaro.org>
Fix hard fault due to NULL being passed as argument into
rssi_handler, which now derives the gsm_modem struct pointer from
the argument.
Signed-off-by: Yong Cong Sin <yongcong.sin@gmail.com>
Disabling clock may conflict with concurrent usage of GPIOs and pins in
AF mode. This can be improved once a proper clock control API is in
place.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
Add GPIO driver for Gigadevice SoCs. The driver supports devices with
the AF and AFIO models.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
Add initial support for the GigaDevice External Interrupt Controller.
This driver is required to manage GPIO interrupts. Only EXTI lines 0 to
15 are supported for now (no LVD, RTC, etc.). Driver can be extended in
the future to add support for extra EXTI lines.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
This change provides support for 4 IO read (via 4READ command) and
program (via 4PP). Flash memory pins SIO[0123], CLK and CS are used.
All of them are controlled by stm32 QSPI IP block.
The instruction code for fast reading as well as number of latency
cycles required are read from SFDP structure provided by flash memory.
The number of required read latency cycles when performing reading
is the sum of SPI-NOR memory mode bits and wait states (also named
as 'dummy cycles').
It also has been assumed that memory, which supports fast read (4READ),
also will support fast programming (4PP command) as this information is
not available in SFDP.
One also need to enable the QUAD IO support in board's device tree by
defining 'spi-bus-width = <4>' property. It is required as it may
happen that not all QSPI dedicated pins are used (for example only two
of them are available).
Signed-off-by: Lukasz Majewski <lukma@denx.de>
For proper operation of QUADSPI IP block, both sides of communication;
NOR flash memory and STM32 controller need to support 4 IO
transmission.
After this change the QSPI stm32 driver is able to program NOR flash
memory to switch itself to use all 4 IO lines (SIO[0123]) to transmit
and receive data.
The QE bit (in SPI-NOR's Status Register) is the non-volatile one, so
setting it is done only once (at first boot of the device).
Signed-off-by: Lukasz Majewski <lukma@denx.de>
This commit provides define for reading the control register (CR) on
the flash memory.
In that register the information about 4B addressing is stored.
Signed-off-by: Lukasz Majewski <lukma@denx.de>
Some flash memories connected to QUADSPI IP block on stm32[fh]7 devices
require proper reset pulse before configuration.
This patch adds two new properties - the 'reset-gpios' phandle,
which allows specifying GPIO pin for RESETn pulse and
'reset-gpios-duration', which provides the time (in ms) for reset
duration.
Signed-off-by: Lukasz Majewski <lukma@denx.de>
The 'flash-id' property defines the number of QSPI's dedicated GPIO
bank (and flash ID), which will be used to communicate with flash
memory.
For example, on stm32h7xx it is possible to use 'quadspi_bk1_*' and
'quadspi_bk2_*' set of pins, so one may need to select between them
when required.
By default - pins from 'quadspi_bk1_*' bank are used, so the
'flash-id = <2>;' property, when your use case (e.g. PCB design)
requires it, forces usage of 'quadspi_bk2_*' pins.
Signed-off-by: Lukasz Majewski <lukma@denx.de>
The stm32 HAL code for qspi (SHA1: 5c8275071ec1cf160bfe8c18bbd93):
stm32cube/stm32h7xx/drivers/src/stm32h7xx_hal_qspi.c implicitly
depends on the stm32h7xx_hal_mdma.c driver as it uses unconditionally
'HAL_MDMA_Abort_IT()' function.
To avoid build breaks - no matter if QSPI IP block works with DMA
or interrupts - the CONFIG_USE_STM32_HAL_MDMA shall be defined.
Signed-off-by: Lukasz Majewski <lukma@denx.de>
In npcx_itim_evt_isr, it updates the cyc_sys_announced variable and
then calls sys_clock_announce() to update the kernel curr_tick variable.
If an ISR handler with higher priority preempts the timer ISR after the
sys_clock_announce is updated and before the sys_clock_announce() is
called, it will read the wrong time when calling k_uptime_get() because
the cyc_sys_announced and the curr_tick are not synchronized.
The commit fixes the problem by raising the timer's interrupt priority
to the highest one (i.e. 1 in npcx's configuration).
This commit also moves the computation of the delta cycle inside the
spinlock in sys_clock_elapsed() to prevent another potential racing
condition.
Signed-off-by: Jun Lin <CHLin56@nuvoton.com>
Signed-off-by: Wealian Liao <WHLIAO@nuvoton.com>
Convert the CAN loopback driver from being configured via Kconfig to
multi-instance configured via devicetree.
Signed-off-by: Henrik Brix Andersen <hebad@vestas.com>
gsm_ppp_stop should cancel possible running work items.
Signed-off-by: Yong Cong Sin <yongcong.sin@gmail.com>
Signed-off-by: Wouter Cappelle <wouter.cappelle@crodeon.com>
The 'work' argument of a delayable work's handler should be
converted to delayable work before passing into the CONTAINER_OF
macro.
Signed-off-by: Yong Cong Sin <yongcong.sin@gmail.com>
