According to the coding guidelines "dynamic allocation is not allowed".
This commit removes handling invalid DMA capable buffers by allocating
temporary buffer in a valid memory region, considering them as errors.
Signed-off-by: Lucas Tamborrino <lucas.tamborrino@espressif.com>
Add GDMA support for esp32s3.
Remove suspend/resume since they are optional and do
the same as start/stop.
Fix possible null pointer derreference.
Signed-off-by: Lucas Tamborrino <lucas.tamborrino@espressif.com>
Make use of positive status values in the DMA callback to pass
info to the DMA client after a successful DMA operation.
A completed DMA transfer uses the status 0 while a reached
water mark uses the status 1.
Signed-off-by: Cyril Fougeray <cyril.fougeray@worldcoin.org>
CAVS platforms are not fully integrated with zephyr. Some of the
registers are still programed from SOF side. This feature can be enabled
for those platforms later when integration is fully done.
Signed-off-by: Tomasz Leman <tomasz.m.leman@intel.com>
Adding function that is allowing to release ownership of the DMA. When
DSP is no longer using dma instance it ownership can be released.
Signed-off-by: Tomasz Leman <tomasz.m.leman@intel.com>
This patch is refactoring device power flow routine. DMA instance will
not be power up after kernel initialization if device power management
is enabled. Power menager will power up device and its power domain if
device will be in use and disable it when it's no longer needed.
The DMA disabling part has not yet been implemented.
Signed-off-by: Tomasz Leman <tomasz.m.leman@intel.com>
The address increment logic was looking at both source
and destination parameters together. Separate them so
we can set one.
Signed-off-by: Mahesh Mahadevan <mahesh.mahadevan@nxp.com>
CFG register uses fields that are not defined in Synopsys databook of
Designware AHB DMA Controller.
Since current Zephyr code uses this driver only for the
intel_adsp_gpdma driver I assume that those fields are specific to
this DMA which is not the standard Designware one.
This patch allows to use either the standard Designware register or
the Intel one.
Signed-off-by: Sylvain Chouleur <schouleur@graimatterlabs.ai>
Add a second instance of DMA to the stm32 dma driver from the DTS
That can be found in the stm32H5 serie, where 2 GPDMA with 8 channels
are available. Rebuilding with Macro.
Use the dma_channels property of the dedive tree to count the nb
of dma channels: 16 or 8 (like in stm32h5).
Signed-off-by: Francois Ramu <francois.ramu@st.com>
Unify the drivers/*/Kconfig menuconfig title strings to the format
"<class> [(acronym)] [bus] drivers".
Including both the full name of the driver class and an acronym makes
menuconfig more user friendly as some of the acronyms are less well-known
than others. It also improves Kconfig search, both via menuconfig and via
the generated Kconfig documentation.
Signed-off-by: Henrik Brix Andersen <hebad@vestas.com>
Because the DMA driver allows multiple start and stop calls for the same
instance and the same channel, we cannot rely on the error codes
returned by these functions to notify the device's power manager that a
device is still in use.
Signed-off-by: Tomasz Leman <tomasz.m.leman@intel.com>
Failures are seen with SOF digital mic capture test cases on Intel cAVS2.5
platforms if the SAR/DAR/CTL writes are skipped.
This reverts commit 08d9efb202.
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
After disabling the channel, it is recommended to poll with timeout to
ensure that the channel has actually been disabled. Without this,
reconfiguring the DMA again while the channel is active could lead to
unexpected behavior and/or DMA underruns.
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
The WAIT_FOR() function returns the value of checked expression. So fix
the return value check to log and return the timeout error when checking
if the FIFO is empty during stop.
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Add a helper function to dump the GPDMA SHIM and channel registers for
debug along with a config option to enable it.
Co-developed-by: Tom Burdick <thomas.burdick@intel.com>
Singed-off-by: Tom Burdick <thomas.burdick@intel.com>
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Add the loop to wait until the stream is really disable after disabling
and also cleared the TCIF flag.
This is a specifity of the dma of type V1.
Signed-off-by: Francois Ramu <francois.ramu@st.com>
Extends dmamux driver to support DMAMUX 2,
which supports the BDMA on STM32H7 devices.
Co-authored-by: Jeroen van Dooren <jeroen.van.dooren@nobleo.nl>
Signed-off-by: Hein Wessels <heinwessels93@gmail.com>
Wrap calls to DMA drivers through fops to allow
different dmamuxes to point to different types of
dma with different function calls.
In preperation to add support for BDMA and DMAMUX2.
Co-authored-by: Jeroen van Dooren <jeroen.van.dooren@nobleo.nl>
Signed-off-by: Hein Wessels <heinwessels93@gmail.com>
The Microchip XEC family of microcontrollers includes a
simple DMA block implementing multiple channels. DMA supports
memory to memory, memory to peripheral, and peripheral to
memory transfers. Peripheral support is limited by each
chip to I2C and SPI controllers. DMA hardware does not support
scatter-gather or linked transactions.
Signed-off-by: Jay Vasanth <jay.vasanth@microchip.com>
Allow DMA channel to be stopped even if it is in suspended state.
The current code in dw_dma_stop() supports this, but the recently
added early check for dw_dma_is_enabled() results in DMA stop
sequence to be skipped if the channel was in suspended state.
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
The conditional definitions are not correct as "chan_data"
is used without any include guards, but defined only if
CONFIG_DMA_DW_HW_LLI or CONFIG_DMA_DW_SUSPEND_DRAIN is defined.
Correct the declaration to match usage.
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
On a repeated dma_start, the stm3u5 dma driver should return
immediately if it is already started.
On a repeated dma_stop, the stm3u5 dma driver should return
immediately if it is already stopped.
This like the dma driver does for other stm32 series.
Signed-off-by: Francois Ramu <francois.ramu@st.com>
The DMA API contract specifies that start/stop may be called multiple
times. Prior to adding power management this was perfectly fine as it was.
In adding power management, there are additional side effects that can
cause issues. Instead check the state of the channel prior to start/stop
and do nothing if already in the desired state.
Signed-off-by: Tom Burdick <thomas.burdick@intel.com>
The hda driver uses -EPIPE to signal xrun, as proposed in the alsa lib
https://www.alsa-project.org/alsa-doc/alsa-lib/pcm.html. This commit
changes the xrun error code in dw dma driver from -ENODATA to -EPIPE to
make it consistent across drivers.
Signed-off-by: Adrian Warecki <adrian.warecki@intel.com>
To support the repeated start/stop, the stm32 dma driver is
returning 0 if the channel is already started/stopped.
This is not done at the dmamux level if any.
Signed-off-by: Francois Ramu <francois.ramu@st.com>
Use dma_slot for peripheral request instead of linked_channel.
This is a more suitable usage as described in dma_config.
Signed-off-by: TOKITA Hiroshi <tokita.hiroshi@gmail.com>
Enable Zephyr device runtime power management mechanisms in Intel HDA
DMA driver. This allows Zephyr to track usage reference for power domain
gating.
PM action handler is currently empty and will be fill out later.
Signed-off-by: Tomasz Leman <tomasz.m.leman@intel.com>
Enable Zephyr device runtime power management mechanisms in Intel GP DMA
driver. This allows Zephyr to track usage reference for power domain
gating.
Signed-off-by: Tomasz Leman <tomasz.m.leman@intel.com>
Adds an attribute that describes the maximum number of scatter gather
blocks that the DMA will accept. Useful in cases where DMA is used as
the abstraction for moving data around (e.g. SoF audio streams) and
variances of the hardware need to inform decisions about how to use
the DMA API.
Co-authored-by: Adrian Bonislawski <adrian.bonislawski@intel.com>
Co-authored-by: Jaska Uimonen <jaska.uimonen@intel.com>
Signed-off-by: Tom Burdick <thomas.burdick@intel.com>
This function is responsible for checking if combination of msize,
mburst and FIFO level is allowed. Possible combinations can be found in
ST documentation, eg. Table 36. FIFO threshold configurations, RM0402
9.3.13 FIFO chapter.
Previously there was no 'break' or '__fallthrough' in msize switch which
caused compilation errors. Since we are confirming that combination is
correct, 'break' statements should be used.
Besides of introducing missing 'break' statements, this patch moves
'return false' from switch to the end of the function. This makes code
shorter and easier to understand, because we have only correct
combinations.
Signed-off-by: Patryk Duda <pdk@semihalf.com>
Add the resume API function for the dma driver of the stm32U5 serie.
That completes the suspend API function.
Controlling the SUSPF bit of the GPDMA CR register is enough
to suspend/resume the channel.
Signed-off-by: Francois Ramu <francois.ramu@st.com>
There is no need to reset the channel else DMA config is lost and
channel should be enabled again in case of resume.
Signed-off-by: Francois Ramu <francois.ramu@st.com>
Reduce logging verbosity in status call from INF to DBG as the function
can be called any given time i.e. eveything is fine even if the read and
write pointers don't differ.
Signed-off-by: Jaska Uimonen <jaska.uimonen@linux.intel.com>
