Media pipe is missing the scheduling component, so move it after the dai
definition and add the scheduling component.
Signed-off-by: Jaska Uimonen <jaska.uimonen@intel.com>
Compiler options has to be changed in order to compute
the literal block address correctly.
Some literals must be declared explicitly and loaded
indirectly to make sure that compiler does not optimize
the out to another shared region.
Signed-off-by: Marcin Maka <marcin.maka@linux.intel.com>
The i.MX8 platform doesn't have an SSP clock. We had a fake define for
this platform due to prior inflexibility in SOF itself. Now that said
lack of flexibility is gone, this patch removes all such references.
Signed-off-by: Paul Olaru <paul.olaru@nxp.com>
Some platforms do not have a SSP clock, so a dummy define for CLK_SSP
and related macros shouldn't be required to exist in order for the
firmware to build.
Signed-off-by: Paul Olaru <paul.olaru@nxp.com>
It didn't work correctly if multiple cores write to "sw regs"
allocated on the same cache line.
Signed-off-by: Karol Trzcinski <karolx.trzcinski@linux.intel.com>
Enables DW-DMA channel linear link position counter based
on the selected peripheral connection. It can be used
to retrieve timestamping information on platforms supporting it.
Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Cleanups definitions of GPDMA shim registers for all cAVS platforms
and extends them by adding support for channel linear link position
control register.
Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
After disabling tracing code have compilation error because of
implicit declaration of platform_trace_point function.
After calling through trace_point it is replaced with empty
define after disabling trace.
Signed-off-by: Karol Trzcinski <karolx.trzcinski@linux.intel.com>
Xtensa ABI requires the stack to always be aligned to a multiple of 16
bytes. This commit aligns the stacks used for interrupt handling.
Signed-off-by: Paul Olaru <paul.olaru@nxp.com>
Create abstract interface for clocks and move platform specific
clocks code to platform folders.
It was also necessary to move SSP clocks declarations to
SSP driver code to decouple clocks logic from mandatory
SSP clocks.
Signed-off-by: Janusz Jankowski <janusz.jankowski@linux.intel.com>
The changes fix command unbalanced groupings and remove discrepancy
between parameters in function declaration & definition of
set_power_gate_for_memory_address_range() function.
Signed-off-by: Lech Betlej <lech.betlej@linux.intel.com>
The description says that any LX6 Xtensa architecture revision DSP
should have this enabled, and the particular DSP in the i.MX8 does fit
this description.
Signed-off-by: Paul Olaru <paul.olaru@nxp.com>
This change adds stream map and channel map structures
used for channel re-routing and stream aggregation.
Signed-off-by: Slawomir Blauciak <slawomir.blauciak@linux.intel.com>
Reverts all changed media pipelines to again have 4000 ms
scheduling period. Change was introduced in order to workaround
firmware scheduling limitation, but with pipelines scheduled
separately we can revert it back.
Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
This patch enables separate scheduling of connected
pipelines. It allows for different scheduling periods
between connected pipelines and flexible controlling
of such topologies.
Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Uses pipeline_task in DMA domains to check, if tasks should be even
registered on interrupts. Non registrable tasks are the one,
whose pipelines are not the owners of scheduling component e.g.
host pipelines connected to mixer pipeline. Such tasks are still
added to the scheduler's list and executed, but they are driven
by the pipelines with scheduling component.
Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Adds definition of pipeline_task. It's a task type registered
by pipelines and used in DMA scheduling domains.
Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Removes SOF_SCHEDULE_FLAG_SYNC and SOF_SCHEDULE_FLAG_ASYNC.
They are no longer used nor needed.
Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Changes ll scheduler implementation to use synchronous property
of domain to check if given task should be scheduled synchronously
or asynchronously. It makes sense, since it's heavily dependent
on particular domain functionality.
Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Adds new synchronous field to indicate if tasks registered
to this domain should be scheduled synchronously or asynchronously.
Signed-off-by: Tomasz Lauda <tomasz.lauda@linux.intel.com>
Instead of taking the name from an array of strings (which is allowed by
GCC) I compose the name directly inside the IRQSTR_CASCADE_TMPL_DECL
macro. GCC is able to detect that the name is a compile time constant
and allows the old code, XCC does not and requires this change.
Signed-off-by: Paul Olaru <paul.olaru@nxp.com>
This allows the SGN macro to no longer generate a warning on the "-1"
case and thus allows building with XCC. GCC doesn't seem to care about
this one. No actual change in the logic itself.
Signed-off-by: Paul Olaru <paul.olaru@nxp.com>
PADC means to have ESAI pad samples with zeros instead of sign-extending
them when they are smaller than the (32-bit) container. This doesn't
make senseon capture and indeed that specific bit is hardware reserved.
So don't use it.
Signed-off-by: Paul Olaru <paul.olaru@nxp.com>
Generally comment cleanups but also moving the "|" operator to be at
the end so that the various options in dai_update_bits are aligned.
Signed-off-by: Paul Olaru <paul.olaru@nxp.com>
The configurations will be mostly similar to those used for playback,
but while the ESAI is master on playback it must be slave on capture
as a workaround due to hardware limitations. These differences must be
hardcoded due to the topology not specifying these configurations for
each direction of the link.
Signed-off-by: Paul Olaru <paul.olaru@nxp.com>
Currently this just returned a zero and wasn't able to correctly
interpret the direction of the channel.
Signed-off-by: Paul Olaru <paul.olaru@nxp.com>
The direction was already configured correctly in the component during
dai_new, and the IPC data is shared for both instances of the DAI
component (playback and capture).
Signed-off-by: Paul Olaru <paul.olaru@nxp.com>
Use block expression instead of do {} while(0) macro
in spin_try_lock in order to return its value in a
safer way.
Signed-off-by: Bartosz Kokoszko <bartoszx.kokoszko@linux.intel.com>
Jaska Uimonen