Rev |
Log message |
Author |
Age |
Path |
31 |
Initial network is now working. Adding CPU control files to repository. |
dgisselq |
2878d 08h |
/. |
30 |
Network transmit and MIG memory both work now, though the clock speed has
been dropped to 80.125MHz. |
dgisselq |
2878d 09h |
/. |
29 |
Here's the memory pin list, necessary for running Xilinx's Memory Interface
Generator. |
dgisselq |
2906d 05h |
/. |
28 |
Including the updates and corrections from the wbuart32 project. |
dgisselq |
2906d 05h |
/. |
27 |
Bus changes ... |
dgisselq |
2906d 06h |
/. |
26 |
Adjusted the timing comments. |
dgisselq |
2906d 06h |
/. |
25 |
The memory now works. However, the core speed has been lowered to 81.25MHz
to do this. The top level file is no longer fasttop.v, but toplevel.v. |
dgisselq |
2914d 14h |
/. |
24 |
Here are the updates from the first (failed) attempt to try to integrate
the DDR3 SDRAM controller onto this board. |
dgisselq |
2933d 09h |
/. |
23 |
Includes settings necessary for the Arty to load from flash builds, and to
reconfigure itself later. |
dgisselq |
2943d 08h |
/. |
22 |
A useful script for programming the device, given that the current device
program includes a valid comms interface. |
dgisselq |
2943d 09h |
/. |
21 |
Removed the OLED controller one additional clock from the bus. This was
necessary to maintain the 200MHz clock speed, especially given the growing
fanout of the device bus. |
dgisselq |
2943d 09h |
/. |
20 |
Lots of bug fixes: After turning on XIP, and running in XIP mode, leaving XIP
mode turns it back off again, necessitating a new write to the VCon register.
Further, XIP mode starts in extended SPI mode, and only transfers in QSPI
mode for data. Finally, two new commands have been created: enabling the
SPI memory reset, and actually resetting the SPI memory. In general, these
are all better--as the EQSPI flash controller now works with these changes,
whereby it didn't really work without them before. |
dgisselq |
2943d 09h |
/. |
19 |
Creates an LED mask portion of writing to the LED's register. Only those
bits specified in the mask (bits [7:4]) will be adjusted in the LED
register on a write. Hence to set all on, set the LED register to 0x0ff,
all off, 0x0f0, or to set LED 0 to on while leaving the others unchanged,
set it to 0x011. |
dgisselq |
2943d 09h |
/. |
18 |
The device can now program and configure itself, allowing bootstrap possibilities.
` |
dgisselq |
2943d 09h |
/. |
17 |
Fixed the misaddressed I/O peripherals in the fastio peripheral group. In
particular, UART and GPS were misaddressed. This has now been fixed, so these
peripherals (should) match the spec. Further, the default speed of the two
UARTs has been adjusted to 115200/8N1 for the aux UART, and 9600/8N1 for the
GPS UART. The Aux UART transmitter also passes testing, so at least that one
works. |
dgisselq |
2945d 08h |
/. |
16 |
This returns the scope to its original configuration, where the decode()
function is not allowed to have any side effects. |
dgisselq |
2945d 09h |
/. |
15 |
These files were changed/perfected as part of getting the EQSPI flash driver
up and running. They currently match the hardware, as far as I can tell, and
the test bench (_tb.cpp file) shows that the Verilog modules work (with the
simulated hardware) as designed. |
dgisselq |
2945d 09h |
/. |
14 |
All changes made as a part of getting the flash driver up and running. Today,
it successfully programmed a configuration into the Arty via wbprogram, so ...
that's been my last/best test. |
dgisselq |
2945d 09h |
/. |
13 |
Here are the updates necessary to get the initial Quad SPI flash driver working.
This includes in particular the bus interconnect (in fastmaster.v), and the
final hookup to the external wires (in fasttop.v). |
dgisselq |
2945d 10h |
/. |
12 |
A work in progress. This checkin contains fixes for the flash (mostly),
together with some logic reductions facilitating full speed (200MHz) build. |
dgisselq |
2946d 12h |
/. |