* Value 24 was wrongfully set for parameter LENGTH_WIDTH, because
it is not among the valid values, which are 28, 29, ..., 34. Set '28'
to be the default value
* Vivado Tcl somehow didn't accept the old expression set for
calculating the HBM_SEGMENTS_PER_MASTER parameter, so it was changed
accordingly to work. Dropped "expr", ".0" and "int ()" parsing and now
it works
Signed-off-by: Iulia Moldovan <Iulia.Moldovan@analog.com>
* SPI Engine: fix early sdi data clear
In case an SPI read was immediately followed by a cs assert, the sdi
register was being cleared one cycle too soon, so that the data being
passed on was always 'b0.
Signed-off-by: Laez Barbosa <laez.barbosa@analog.com>
Previous level-based trigger could cause issues in some low
sampling rate setups. This commit changes it to edge-based.
Signed-off-by: Laez Barbosa <laez.barbosa@analog.com>
The lack of the create_xgui_files causes Vivado to exit with an error
when running multiple Vivado instances (parallel make case)
Signed-off-by: Jorge Marques <jorge.marques@analog.com>
Set validation ranges for DATA_WIDTH and NUM_OF_CS for the expected
min/max values in the verilog source code.
Also, fix swapped description for CPHA in the documentation.
Signed-off-by: Carlos Oliveira <caosjr8@gmail.com>
Previously when issuing a load_config, each pwm channel
was stopped in its tracks and waited for an external sync,
if that was active, or load_config release.
The desired behaviour is to wait for the pwm channels to finish
their events from the current period, before a new aligned start.
Also, the first positive edge of each pulse was initiated only
in the second pwm channel period.
This niche behaviours have not affected any functionality in the
long term alignments for current setups.
Signed-off-by: AndreiGrozav <andrei.grozav@analog.com>
When leaving the offset equal to zero for a pwm
channel. That pwm channel was not waiting for all
channels to get in sync after a load config.
Signed-off-by: AndreiGrozav <andrei.grozav@analog.com>
Drop shell for loops in favor of makefile pattern rules,
so make can run targets in parallel using -j.
This doesn't affect Vivado's own settings.
As a benchmark, 12th Gen Intel(R) Core(TM) i9-12900H 5GHz(max):
$ make -C projects/adrv9009/zcu102/ clean-all
$ time make -C projects/adrv9009/zcu102/ -j$CORES lib
CORES=1:
real 9m27.223s
user 9m2.556s
sys 0m32.358s
CORES=8:
real 1m54.639s
user 16m26.512s
sys 1m2.317s
i.e. about 5 times faster to build IP core dependencies.
Signed-off-by: Liam Beguin <liambeguin@gmail.com>
Signed-off-by: Jorge Marques <jorge.marques@analog.com>
1. Simplify the control logic by adding a state machine.
The improvements are on code readability and reliability.
2.Add a flush feature which can be used to clean the data from the DMA fifo.
This is useful when the DMA is programmed in cyclic mode and
data transmission is stopped by dma_transfer_suspend flag
The software intervention is reduced at setting the flag(dma_flush_en).
Flushing can also be done when activating the raw value with dma_flush_en active.
3. Add raw value support. Through this changes a user can set
the dac output to a fixed predefined value in the following two cases:
1. direct, without using the dma.
2. with dma, as a hold value. The fixed value will be kipped after a cyclic
buffer is stopped by axi_dac_interpolate, through dma_transfer_suspend
register/signal.
The raw value ca be set and transmitted independently on each channel.
The predefined value is stored in reg 0x19(0x64). For more details se
the documentation available at
https://wiki.analog.com/resources/fpga/docs/axi_dac_interpolate
axi_dac_interpolate - Remove last sample hold control
axi_ad9963 - Remove last sample hold control and set as default the
last sample hold functionality plus code optimization changes.
This commit introduces a different interface to submit transfers, using
DMA descriptors.
The structure of the DMA descriptor is as follows:
struct dma_desc {
u32 flags,
u32 id,
u64 dest_addr,
u64 src_addr,
u64 next_sg_addr,
u32 y_len,
u32 x_len,
u32 src_stride,
u32 dst_stride,
};
The 'flags' field currently offers two control bits:
- bit 0: if set, the transfer will complete after this last descriptor
is processed, and the DMA core will go back to idle state; if cleared,
the next DMA descriptor pointed to by 'next_sg_addr' will be loaded.
- bit 1: if set, an end-of-transfer interrupt will be raised after the
memory segment pointed to by this descriptor has been transferred.
The 'id' field corresponds to an identifier of the descriptor.
The 'dest_addr' and 'src_addr' contain the destination and source
addresses to use for the transfer, respectively.
The 'x_len' field contains the number of bytes to transfer,
minus one.
The 'y_len', 'src_stride' and 'dst_stride' fields are only useful for
2D transfers, and should be set to zero if 2D transfers are not
required.
To start a transfer, the address of the first DMA descriptor must be
written to register 0x47c and the HWDESC bit of CONTROL register must
be set. The Scatter-Gather transfer is queued similarly to the simple
transfers, by writing 1 in TRANSFER_SUBMIT.
The Scatter-Gather interface has a dedicated AXI-MM bus configured for
read transfers, with its own dedicated clock, which can be asynchronous.
The Scatter-Gather reset is generated by the reset manager to reset the
logic after completing any pending transactions on the bus.
When the Scatter-Gather is enabled during runtime, the legacy cyclic
functionality of the DMA is disabled.
Signed-off-by: Ionut Podgoreanu <ionut.podgoreanu@analog.com>
Fixed wrong behaviour on chip select instruction:
- previously, a sleep time happened before the chip select change
- the intended behaviour was for another sleep time, of equal amount, to happen after the chip select change as well
- additionally, the counter logic implementation was creating an additional factor of 2 on the sleep time
All of the above points were fixed. The changes introduced also fix another issue where the sleep instruction was likewise happening with a duration larger than intended by a factor of 2
Signed-off-by: Laez Barbosa <laez.barbosa@analog.com>
The axi_ad4858 IP core is design as the HDL interface for the AD4858 ADC.
Features:
- AXI based configuration
- LVDS and CMOS support
- Configurable number of active data lines (CMOS - build-time configurable)
- Oversampling support
- Supports packet formats 0,1,2 or 3
- CRC check support
- Real-time data header access
- Channel based raw data access(0x0408)
- Xilinx devices compatible
Documentation at https://wiki.analog.com/resources/fpga/docs/axi_ad4858
The custom interface IP for AD3552R DAC has more operation capabilities:
- 8b register read/write SDR/DDR
- 16b register read/write SDR/DDR
- data stream SDR/DDR ( clk_in/8 or clk_in/4 update rate)
- selectable input source : DMA/ADC/TEST_RAMP
- data out clock(SCLK) has clk_in/8 frequency when the converter is configured and clk_in/2 when the converter is in stream mode
- the IP reference clock (clk_in) can have a maximum frequency of 132MHz
- the IP has multiple device synchronization capability when the DMA is set as an input data source
Signed-off-by: PopPaul2021 <Paul.Pop@analog.com>
The DDS for each channel was synchronized by the main channel.
One problem with this aporoach is that when a user sets a DDS that
is not from the main channel the sinchronization does not happend.
This behavior is not user friendly in IIO-Oscilloscope or within other
configuration methods.
This commit keeps all channels in sync by triggering the sync on all
channels from each individual channel.
Add support for DDS phase width greather than 16.
The software should read the DDS phase data width register and configure
the DDS init and increment registers accordingly, otherwise the obtained
DDS output frequency will not be the desired one for DDS phase width
different than 16.
DDS_incr = (f_out * 2^(phase_width) * clkratio)/f_if
* Removed the commercial JESD license and put the ADIBSD or GPL v2 like
for other Verilog files
Signed-off-by: Iulia Moldovan <Iulia.Moldovan@analog.com>
Due to nets being optimized at IP-level during the no-OOC synthesis flow,
constraints related to req_clk (request clock) were not being applied,
causing the design to not meet timing.
The fix considers the synchronous modes, appending the possible resulting
req_clk's names after the synthesis flow.
Due to grounded signals in the DMA_TYPE_SRC != DMA_TYPE_STREAM_AXI config.,
sync_rewind is removed during synthesis, even so, constraints were
trying to be applied to those nets.
To resolve this, sync_rewind block was moved to inside the generate.
Vivado seems to properly suppress "Empty list" warnings when the circuit does not exist because of a generate rule.
Signed-off-by: Jorge Marques <jorge.marques@analog.com>
Signed-off-by: Ionut Podgoreanu <ionut.podgoreanu@analog.com>
docs/regmap/adi_regmap_adc.txt:
- add softspan to regmap
library/common/up_adc_channel.v
- update copyright year header
- add softspan to regmap
library/common/up_adc_common.v
- update minor version
Signed-off-by: John Erasmus Mari Geronimo <Johnerasmusmari.Geronimo@analog.com>
Iulia Moldovan
PIoandan
LBFFilho
IstvanZsSzekely
Jorge Marques
caosjr
AndreiGrozav
Alin-Tudor Sferle
Liam Beguin
Ionut Podgoreanu
Stanca Pop
PopPaul2021
Jem Geronimo
Jem Geronimo