Commit Graph

182 Commits

Author SHA1 Message Date
Ross Thompson
63afd95ad3 Fixed bugs in boot and new flash card merge. Works with arty a7 now. 2023-07-22 15:52:25 -05:00
Ross Thompson
ab6ef5bb58 At least it simulates and gets through fpga elaboration. 2023-07-21 18:40:26 -05:00
Ross Thompson
a89a1e675c Merge branch 'boot' into mergeBoot
Merges Jacob's new sdc controller into wally.
2023-07-21 17:43:45 -05:00
Ross Thompson
d04d2afed2 Modified the LSU/IFU and caches to improve critical path. Arty A7 went from 15 to 17Mhz. I believe we can push all the way to 20+Mhz with relatively little effort. Along the way I'm fixing up the scripts build the linux images for the flash card. 2023-07-21 13:06:27 -05:00
Jacob Pease
380d96b359 Working new boot process. Buildroot package for sdc. 2023-07-20 14:15:59 -05:00
Ross Thompson
2752e5de4c Fixed a bunch of timing constraints for the arty a7 board. 2023-07-19 17:08:16 -05:00
Ross Thompson
97a16f75dc Fixed typo in fpga top for arty a7. 2023-07-19 11:37:29 -05:00
Ross Thompson
e4d6a9f8c6 Removed all old configuration files. 2023-07-19 10:28:54 -05:00
Ross Thompson
b756b248b4 Wow. The newest version of Vivado does not like the enums as parameters.
The solution is simple.  I changed the type to logic [31:0] and defined macros for the branch predictor types as 32 bit integers.
2023-07-18 15:07:10 -05:00
Ross Thompson
6a2b752fc0 Updated arty a7 fpga top. 2023-07-17 15:55:57 -05:00
Jacob Pease
b3aaa87cba Modified bootloader to access GUID partitions. SDC interrupt to PLIC.
Since writing an SD card image generation script, the bootloader
needed to be altered to access individual binaries from specific
partitions. A new file, gpt.c with it's header gpt.h, have been added
to the bootloader to facilitate this.

The SDC has been added to the device tree for the VCU108
board. Additionally the SDC interrupt signal was added to the PLIC
node in the device tree. The PLIC itself  was modified to accept the
SDC interrupt signal.
2023-07-14 13:36:44 -05:00
Ross Thompson
7aecd72c35 Fpga does not correctly boot linux. I think the solution here is to revert out all substantive changes except for parameterization and then add them back in one at a time. This is necessary because the parameterization is not completed in one contiguous group of commits. 2023-06-22 12:55:49 -05:00
Ross Thompson
a8f11dcad0 FPGA updates. 2023-06-20 11:11:34 -05:00
Ross Thompson
af187d96ca Updated fpga wave config. 2023-06-19 12:28:30 -05:00
Ross Thompson
1a23f1360f Updated fpga wally wrapper to work with the ILA. 2023-06-19 12:15:48 -05:00
Ross Thompson
0423d7df82 I think the fpga is building again, but the debugger script needs to be updated. For some reason the nets are not present despite being marked debug. 2023-06-16 17:00:27 -05:00
Ross Thompson
443c568994 Vivado requires an intermediate wrapper file for parameterization. 2023-06-16 16:30:14 -05:00
Ross Thompson
c44d4321fb FPGA synthesis is broken. This commit moves closer to fixing the issues causes by parameterization. 2023-06-16 15:40:13 -05:00
Jacob Pease
40f81d5da6 The Vivado-RISC-V SDC works. Wally is now booting through it. 2023-05-26 15:42:33 -05:00
Ross Thompson
6907f0ccc1 FPGA makefile update. 2023-04-25 16:24:26 -05:00
Ross Thompson
f22e6d0e48 Updated fpga Makefile to work with both the Arty and VCU platforms. 2023-04-21 11:08:35 -05:00
Ross Thompson
b13fe870cf Yeah We boot linux on the arty a7! 2023-04-19 11:17:33 -05:00
Ross Thompson
1fec535b32 Fixed the reset for Arty A7 and now partially boots. Copies flash card to dram.
but the data is wrong.
2023-04-19 10:35:18 -05:00
Ross Thompson
224bf74530 Found the first issue. the axi clock converter was stuck in reset because the polarity was reversed. 2023-04-18 17:45:41 -05:00
Ross Thompson
367bd0f8dc More debug stuff. 2023-04-18 16:00:10 -05:00
Ross Thompson
668e69fdc9 Added more signals to debugger in hopes I can figure out why the mig is not responding. 2023-04-18 15:51:52 -05:00
Jacob Pease
2839f4f41a AHB triggers write, but AXI side doesn't update. 2023-04-18 15:23:22 -05:00
Ross Thompson
3588c53e66 It's almost working. 2023-04-18 14:24:59 -05:00
Ross Thompson
deb0bfc24d Improved constraints and set ddr3 voltage to correct 1.35V. This voltage is only for synthesis. However I'm concerned because the gui did not let me select 1.35V. 2023-04-17 20:05:59 -05:00
Ross Thompson
777bec2e24 Fixed timing constraint issue. 2023-04-17 19:53:43 -05:00
Ross Thompson
b2b30936be Found the DDR3 memory is not ready when issuing the first store. 2023-04-17 19:33:13 -05:00
Ross Thompson
fbbba0e5c2 Finally we are building the fpga and can view the ila. we are getting out of reset, but we are stuck at PCM = 10b8. 2023-04-17 18:39:25 -05:00
Ross Thompson
2cbaa5c27b Dang. Looks like the reset button on the arty a7 is actually resetn. I wish they'd named it that way. 2023-04-17 16:37:18 -05:00
Ross Thompson
480562e53e Yay! We now have a functional ila and the uart connection on the pc side works. However the CPU is stuck in reset. Not really sure what's going on there. 2023-04-17 16:00:02 -05:00
Ross Thompson
b0f0fb1da7 Adding in the ILA to the arty a7. 2023-04-17 14:54:10 -05:00
Ross Thompson
30d017c258 Lowered arty a7 clock frequency to 15Mhz to meet timing. can probalby go faster. 2023-04-17 12:16:31 -05:00
Ross Thompson
fe692dacce Finally got the arty a7 to build. 2023-04-17 11:54:22 -05:00
Ross Thompson
4ad33d7acc OMG. the ddr3 has it's own mmcm (pll) which had incorreclty specified the input clock period as 3000 ps rather than 6000 ps so the pll was running at twice the speed. I speed the whole weekend on this. :( 2023-04-17 11:10:19 -05:00
Ross Thompson
5591b447d6 Fixed more issues with arty a7 constarints. 2023-04-16 13:25:02 -05:00
Ross Thompson
f4734c0d1b Found and fixed the major architecture issue with the mig 7 used in the arty a7 board.
mig 7 is completely different from the ddr4 mig in that the internal pll does not general the required clocks. An external mmcm is required to general two inputs clocks and the required user clock.
2023-04-15 11:13:28 -05:00
Ross Thompson
2f8359e6cc Realized we need a separate mmcm when using the mig 7 for ddr3 rather than the ddr4 mig. Go figure. 2023-04-14 18:02:16 -05:00
Ross Thompson
d967e05c20 Finally fixed the ddr3 mig script to work correclty. 2023-04-14 11:41:51 -05:00
Ross Thompson
777edb0fcd Progress on arty a7 board. 2023-04-13 17:57:12 -05:00
Ross Thompson
4563b650bf Fixed more bugs in the ila debug constraints. 2023-04-11 14:32:53 -05:00
Ross Thompson
e490ab09cf Updated to help debut Jacob's crossbar woes. 2023-04-11 14:22:42 -05:00
Ross Thompson
6c07a2e595 Fixed sum bugs with arty a7 ila script. 2023-04-11 10:00:06 -05:00
Ross Thompson
c4e5b8db49 Updates for arty a7. 2023-04-10 17:02:19 -05:00
Ross Thompson
5bcb0f6ace Fixed syntax errors in arty7 top level. 2023-04-10 16:08:40 -05:00
Ross Thompson
0700202001 Added more support for Arty A7 board. 2023-04-10 16:01:17 -05:00
Ross Thompson
9d9c2b170d Finally building ddr3 xilinx ip from script. 2023-04-10 14:36:33 -05:00