| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
transports:
Transports build on I/O service and implements flow control and
sequence number checking.
The rx streamer subclass extends the streamer implementation to connect
it to the rfnoc graph. It receives configuration values from property
propagation and configures the streamer accordingly. It also implements
the issue_stream_cmd rx_streamer API method.
Add implementation of rx streamer creation and method to connect it to
an rfnoc block.
rfnoc_graph: Cache more connection info, clarify contract
Summary of changes:
- rfnoc_graph stores more information about static connections at the
beginning. Some search algorithms are replaced by simpler lookups.
- The contract for connect() was clarified. It is required to call
connect, even for static connections.
|
|
|
|
|
|
| |
The valid bit helps prevent placeholder defaults from being
propagated through the graph. Values that are not valid will
not be forwarded.
|
|
|
|
| |
Implement uhd::rfnoc::rfnoc_graph::enumerate_*_connections()
|
|
|
|
|
| |
This will call init_props() on every block after the device
initialization is complete, but before control returns to the user.
|
|
|
|
|
|
| |
During construction of the rfnoc_graph, flush and reset each block in
each motherboard we need to enumerate. This will ensure that each
block is in a clean state when we construct it's block controller.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
During registration, blocks must now specify which clock they are using
for the timebase (i.e., for timed commands) and for the ctrlport (this
is used to determine the length of sleeps and polls). For example, the
X300 provides bus_clk and radio_clk; typically, the former is used for
the control port, and the latter for the timebase clock.
Another virtual clock is called "__graph__", and it means the clock is
derived from property propagation via the graph.
The actual clocks are provided by the mb_iface. It has two new API
calls: get_timebase_clock() and get_ctrlport_clock(), which take an
argument as to which clock exactly is requested. On block
initialization, those clock_iface objects are copied into the block
controller.
The get_tick_rate() API call for blocks now exclusively checks the
timebase clock_iface, and will no longer cache the current tick rate in
a separate _tick_rate member variable. Block controllers can't manually
modify the clock_iface, unless they also have access to the
mb_controller (like the radio block), and that mb_controller has
provided said access.
This commit also adds the clock selection API changes to the DDC block,
the Null block, and the default block.
|
|
|
|
|
| |
An immutable clock means it has a locked frequency, and attempts to
change the frequency will cause an exception to be thrown.
|
|
|
|
|
|
|
|
| |
On destruction, the rfnoc_graph will call shutdown() on all blocks. This
allows a safe de-initialization of blocks independent of the lifetime of
the noc_block_base::sptr.
Also adds the shutdown feature to null_block_control.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
These changes add APIs to instantiate the new transports. However,
only the control/management transport is currently implemented. It
uses the chdr_ctrl_xport.
Also update the mgmt_portal to use an ephemeral reference to the
shared transport, to indicate that it has no ownership of the
transport's memory.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
chdr_ctrl_xport is a dumb-pipe transport for RFNoC control transactions
and management frames.
Also remove the I/O service's check on num_recv_frames and num_send_frames.
The transports may request additional virtual channels, so the send_io_if
and recv_io_if may not reserve additional frames, as they are shared with
a previously-allocated instance.
Note: this uses a mutex to force sequentual access to the
chdr_ctrl_xport. This is supposed to go away when the multi threaded
xport is done.
|
| |
|
|
|
|
|
|
|
|
| |
Applying clang format for upcoming changes.
clang-format -i --style=file host/lib/usrp/cores/gpio_atr_3000.cpp
clang-format -i --style=file \
host/lib/include/uhdlib/usrp/cores/gpio_atr_3000.hpp
|
|
|
|
|
|
| |
The existing implementation assumes registers are spaced 4 bytes apart.
In the current radio block design, all backward compatible registers are
spaced 8 bytes apart. This adds a feature to configure that offset.
|
|
|
|
|
|
| |
This is a single atomic allocator in the global uhd::rfnoc namespace, so
that devices can allocate themselves device IDs without having to confer
with the rest of the RFNoC infrastructure.
|
|
|
|
|
| |
This converts from register_iface& to wb_iface::sptr. Useful for
connecting new block controllers with older interfaces.
|
|
|
|
|
|
| |
Following the changes in RX frontend controls, TX frontend register
offsets are now arguments to the factory function. They default to 4,
which is what the register offset was in the file before these changes.
|
|
|
|
|
|
|
|
| |
Changing how we calculate RX frontend register addresses to allow for
different register offsets. The register addresses are now calculated
in a manor similar to how gpio_atr_300_impl does register address
calculations, which is to allow a reg_offset to be passes in at
construction. The current default is reg_offset=4.
|
|
|
|
|
| |
Removes the requirement for a wb_iface, and also the requirement for
regs to be 4 addresses apart.
|
|
|
|
|
| |
This API lets blocks decide if their current topology is OK for them,
and make decisions based on their topology.
|
| |
|
|
|
|
|
|
|
|
|
| |
The mb_controller is an interface to hardware-specific functions of the
motherboard. The API works in two ways:
- The user can request access to it, and thus interact directly with the
motherboard
- RFNoC blocks can request access to it, if they need to interact with
the motherboard themselves.
|
|
|
|
|
| |
The management looks at the transport endianness from the packet
factory to determine if the byte_swapper in the FPGA needs to be enabled
|
| |
|
| |
|
|
|
|
|
| |
- Fleshed out mb_iface
- Managers currently only export ctrl APIs. Data APIs TBD
|
|
|
|
|
|
| |
- chdr_ctrl_endpoint can manage multiple dest EPIDs
- Moved from both_xports_t to a special defs in rfnoc_common
- Changed data-structures where appropriate
|
|
|
|
|
|
| |
- Moved chdr_packet and chdr_types from rfnoc/chdr to rfnoc and updated
all references
- Moved non-CHDR definitions to rfnoc_common.hpp
|
| |
|
|
|
|
| |
This contains both_links_t
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The inline_io_service connects transports to links without any
worker threads. Send operations go directly to the link, and recv
will perform the I/O as part of the get_recv_buffer() call.
The inline_io_service also supports muxed links natively. The receive
mux is entirely inline. There is no separate thread for the
inline_io_service, and that continues here. A queue is created for
each client of the mux, and packets are processed as they come in. If
a packet is to go up to a different client, the packet is queued up
for later. When that client attempts to recv(), the queue is checked
first, and the attempts to receive from the link happen ONLY if no
packet was found.
Also add mock transport to test I/O service APIs. Tests I/O service
construction and some basic packet transmision. One case will also
uses a single link that is shared between the send and recv transports.
That link is muxed between two compatible but different transports.
|
|
|
|
|
|
|
|
|
|
|
| |
Split the transport into three layers to allow for greater flexibility
in scheduling algorithms. The io_service will make queues on behalf of
the transport and take responsibility for scheduling data transfers
through the links.
The transport layer is the explicit handler for flow control. This
enables the possibility of a scheduling layer in between, so flow
control may be offloaded on the same thread as the link.
|
|
|
|
|
|
|
|
|
|
| |
New interface aimed to replace zero_copy_if for new code, including new
RFNoC development and redesign of streamer objects.
Generic implementation of send and receive transport interfaces to allow
reuse by various transport types. Derived classes implement
transport-specific functions that are invoked by the base classes
through CRTP.
|
|
|
|
|
| |
The inteface provides a mechanism for users of clocks to query
information such as the running status or rate
|
|
|
|
|
|
| |
- Add support for new backend iface with max_async_msgs and mtu
moved to after the noc ID
- Fixed offsets for block info registers
|
|
|
|
|
|
|
|
| |
This structure represents information about a graph edge. Required by
detail::graph and rfnoc_graph.
graph_edge_t::to_string() will now provide a textual representation of
the edge.
|
|
|
|
|
| |
All USRP device impls that are RFNoC devices will need to derive from
this (instead of device3).
|
|
|
|
|
|
| |
- noc_block_base now has a ctor defined
- The registry stores factory functions to the individual Noc-Block
implementations
|
|
|
|
|
|
|
|
|
|
| |
- Adding client_zero class, which gathers information about our device
form the global registers on port 0 of the RFNoC backend registers.
- adding unit tests to exercise client_zero
- mock_reg_iface class: adding fake register_iface so we can run
unit tests in software only
Co-authored-by: Martin Braun <martin.braun@ettus.com>
|
|
|
|
|
|
|
|
| |
- Added new register_iface class that translates high-level
peek/poke calls into CHDR control payloads
- Added new chdr_ctrl_endpoint class that emulates a control
stream endpoint in SW. It can create and handle multiple
register interfaces
|
|
|
|
|
|
|
|
| |
- Added action_info class
- Allow to send actions from node to node
- Allow to post actions into nodes
- Allow to set default forwarding policies
- Added unit tests
|
|
|
|
|
|
|
| |
- The management portal is the interface for the framework
to allow discovering the data topology, setup routes between
stream endpoints and configure streamers
- Use a zero_copy_if and the mgmt_paylod to send/recv packets
|
|
|
|
|
|
|
| |
- Moved packet interface code from public to private include
- Split packet interface into two files: payload paring and packet iface
- Added support for all CHDR packet types
- Added more test cases to unit test
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Adds a detail::graph_t class, which handles the propagation
- Adds methods to node_t to aid with propagation
- Adds unit tests
- Adds dynamic property forwarding:
Nodes are now able to forward properties they don't know about by
providing a forwarding policy. A good example is the FIFO block which
simply forwards most properties verbatim.
- node: Temporarily disabling consistency check at init
|
|
|
|
|
| |
This is a storage for the noc_block_base derivatives. It supports
finding blocks.
|
|
|
|
|
| |
Its purpose is to provide a device-agnostic back-channel interface into
the device guts for all rfnoc_graph devices.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Adds the following classes:
- uhd::rfnoc::node_t, the base class for RFNoC nodes
- uhd::rfnoc::node_accessor_t, a class to access private properties
- uhd::rfnoc::res_source_info, a struct that identifies where properties
come from
- uhd::rfnoc::property_t, and property_base_t (its parent)
- uhd::rfnoc::prop_accessor_t, a class to access properties
Add always dirty property (dirtifier).
Also adds unit tests for properties.
|
| |
|
|
|
|
|
|
| |
Benchmarks show that using C++ chrono features beats
uhd::get_system_time(), and the latter is simply not appropriate unless
a uhd::time_spec_t is required.
|
| |
|