FPGA/oh
1
0
Fork 0
mirror of https://github.com/aolofsson/oh.git synced 2025-01-17 20:02:53 +08:00
Code Issues Projects Releases Wiki Activity
1,487 Commits 7 Branches 1 Tag
Commit Graph

322 Commits

Author SHA1 Message Date
Andreas Olofsson
7d524d0f68 Changing axi interface <--> elink protocol 
-Now consistant with packet, access, wait protocol
 2015-05-19 22:08:41 -04:00
Andreas Olofsson
8d3cbf8257 Clean axi_elink module 
-Clocks included inside for easy integration
-Another version might have the clocks and reset as inputs instead
 2015-05-19 22:07:16 -04:00
Andreas Olofsson
451a1fa925 MILESTONE: Bursts working!!! 
-Fairly clean minimalist design
-Complete redesign
-Need to do random read/write testing to make sure
-Speed?
 2015-05-18 15:38:30 -04:00
Andreas Olofsson
41f97e45ff Converting to synchronous reset 2015-05-17 23:00:53 -04:00
Andreas Olofsson
4fb6e7407c Integrating idelay elements in erx_io 2015-05-16 22:06:40 -04:00
Andreas Olofsson
017b72c37a MILESTONE: "almost final" block works! 
-bursting disabled, otherwise it looks good!
 2015-05-15 15:32:43 -04:00
Andreas Olofsson
d052da4ec9 Speed optimization 
-adding IDDR/ODDR blocks in IO
-still need to add the IDELAY controller and blocks
 2015-05-15 15:31:01 -04:00
Andreas Olofsson
9d3ecd0d06 Clock block cleanup 
-Moving TXCLK to CCLK domain, "feels cleaner"
-Also makes TXCLK independent of RXCLK
-So complete solution has a PLL and an MMCM
-Feeling is that the PLL for RXCLK could eventually be removed
-Need to add more state for locking clocks for receiver at end?
REVIEW!!!
 2015-05-15 15:28:19 -04:00
Andreas Olofsson
d2dcc15c52 Reset and clock cleanup 
-In the default mode we now have 7 input clocks to basic elink
-This is too many, need to simplify, not reasonable!
-But with all the knobs on the MMCM, performance will be great...
-WIP on bursting...
 2015-05-14 22:31:42 -04:00
Andreas Olofsson
77e210e7c2 Synchronous exit from reset 
-Async asert, sync deassert
-Haven't used this before.. (review?)
 2015-05-14 22:28:41 -04:00
Andreas Olofsson
d3d8f3f759 Clock and reset integration 
-moving clock and reset outside basic elink
-adding idelay reference clock
-separate and synchronized reset for each domain
-adding proper reset to fifo_cdc  (per domain, not asynch)
 2015-05-14 22:26:05 -04:00
Andreas Olofsson
1d848fe1d5 Making 2015-05-14 22:24:42 -04:00
Andreas Olofsson
5c690d38a1 Adding reset hardware state machine 
-concept developed by Gunnar Hillerstrom (VHDL)
-takes the driver out of the picture
-some people will run this bare metal
-more deterministic
-cuts down on hw/sw development issues
 2015-05-14 22:22:33 -04:00
Andreas Olofsson
a094f835c9 Rename (clarity) 
-file contains chip id, reset, and clocks
-base config driven by sys_clk
 2015-05-14 22:21:05 -04:00
Andreas Olofsson
befc18f368 MILESTONE: read/write works with all new RX/TX IO logic! 
-Fixes issue with back to back transactions!
-Read/writes work!!
-Needs more verification/analysis...
 2015-05-14 00:00:12 -04:00
Andreas Olofsson
4245c16b0d Adjusting phases for clocks 
-This is because the PLL model does not account for the input clock
-Actually a big dangerous cheat...
-How to model a PLL more accurately?
 2015-05-13 23:57:33 -04:00
Andreas Olofsson
6ba45155fd Integrating clock approach change 
-clocks moved outside elink
-new packet interface format between protocol and io block
 2015-05-13 23:29:18 -04:00
Andreas Olofsson
b8c699fb22 Complete redesign 
-Communication with IO is with "packet format"
-No need to invent a 64 bit format just for stupid OSERDES
 2015-05-13 23:28:06 -04:00
Andreas Olofsson
af1f8a03eb Complete redesign 
-junking the old logic
-not needed with new IO approach
 2015-05-13 23:27:35 -04:00
Andreas Olofsson
0214df5804 Complete redesign of erx io logic 
-Building from primitives
-Work in progress, not quite complete
 2015-05-13 23:26:41 -04:00
Andreas Olofsson
9dcde59979 Complete redesign of erx 
-Giving up on ISERDES. No freaking proper documentaion and no open source simulation model.
-Rewriting io module with primitives.
-Looks like most of the logic disappears...
-Still work in progress
 2015-05-13 23:24:54 -04:00
Andreas Olofsson
89d54f4ed8 More flexible clock solution 
-moving clock block outside elink
-driving all key clocks into erx/etx
 2015-05-13 23:23:23 -04:00
Andreas Olofsson
624d0e6134 Reorg cleanup 
-renamed disty for consistency (there is an arbiter there now)
-adding missing ID to etx/erx
-New org working!
 2015-05-12 07:41:48 -04:00
Andreas Olofsson
81db0b7582 Completing elink hierarchy change 
-splits out clock domains
-makes the core portion a clean/reusable module with defined interface
 2015-05-10 23:38:08 -04:00
Andreas Olofsson
d2b4dabc58 Moving chipid back to clocks 
-unnatural for it to be in etx
(link has nothing to do with epiphany id)
 2015-05-10 23:35:41 -04:00
Andreas Olofsson
a627ecae7b Removing testmode, bad idea 
-Should be input to fifo or etx_core
 2015-05-10 23:35:04 -04:00
Andreas Olofsson
eaadfc6465 Adding etx/erx core modules 
-Single clock domain
-Super-light...
-Need to be able to remove internal feature as well
-(MMU/DMA should be optional but on by default..)
 2015-05-10 23:06:52 -04:00
Andreas Olofsson
fa374e666a Cleanup 2015-05-09 08:57:49 -04:00
Andreas Olofsson
eb3051ea93 Cleaning up logic to fit new access/packet interface 
(pre-debug)
 2015-05-09 08:56:51 -04:00
Andreas Olofsson
ab26378a99 Adding elink with axi interfaces 2015-05-09 08:52:55 -04:00
Andreas Olofsson
a52fa86edb Fixing instances errors from fpga synthesis 2015-05-08 20:55:31 -04:00
Andreas Olofsson
1f6c18a764 Using fifo_cdc instead of fifo_async 2015-05-07 23:45:36 -04:00
Andreas Olofsson
c51f8f3dc9 Adding clock buffer 2015-05-07 23:44:39 -04:00
Andreas Olofsson
ba32323306 Cleaning up clocks 
-moving to "real" Xilinx PLL instantiation
-one PLL for CCLK one for LCLK
-removing clock dividers, can't work at speed, put inside model
-configuration needs to be done differently
-removing pll_bypass signal, can't work with logic
-clocks should be done with hard macro primitives (no logic)
 2015-05-06 12:23:15 -04:00
Andreas Olofsson
d8b5fa78ef Adding emesh as basic building block 2015-05-05 21:38:41 -04:00
Andreas Olofsson
d0439f871f Adding example design for FPGA 2015-05-05 21:37:17 -04:00
Andreas Olofsson
c843fc5fe0 Renaming for my sanity (etx/erx split) 2015-05-05 14:56:35 -04:00
Andreas Olofsson
a3cfa17b06 Removing old module 2015-05-04 22:38:28 -04:00
Andreas Olofsson
de74f8accc Removed synchronizer, not needed 2015-05-04 22:34:14 -04:00
Andreas Olofsson
b2846c5312 MILESTONE: Read/write works back and forth 
-Pipeline looks good, now need to test clk1>>clk2 and clk2>clk1
-Still not completely happy with reset (using async for now)
 2015-05-04 17:13:51 -04:00
Andreas Olofsson
570fbffd7f Baking in the IO wait signal into rd/wr wait 
-Separate waits for rd/wr wait
-Adding wait to protocol block as well
-io_wait always goes through
-using active frame signal to select/clear data for output
 2015-05-04 17:10:32 -04:00
Andreas Olofsson
dcf72537e4 Separate rd/wr stalls 2015-05-04 17:09:50 -04:00
Andreas Olofsson
0aba754b7e Cleanup 2015-05-04 10:54:42 -04:00
Andreas Olofsson
72aff72558 MILESTONE: register read/write working! 
-Bullet proof clock domain crossings!
 2015-05-04 10:49:17 -04:00
Andreas Olofsson
bb8f5f861b Moving the read response to separate group (not register) 
Why?
1.) This will allow for support of multiple outstanding channels at some point
2.) Much easier to debug, can tag transactions in test bench
 2015-05-04 10:41:42 -04:00
Andreas Olofsson
e9d6794833 Blocking TX outgoing transcations on LINKID match 2015-05-04 10:41:14 -04:00
Andreas Olofsson
25b0b188ff Implementing register readback on read response channel 2015-05-04 10:40:43 -04:00
Andreas Olofsson
6907d39490 Making readback work 
-Simplifying logic for rx_en
-Reaback data was incorrectly pipelined (one too many)
 2015-05-04 10:39:01 -04:00
Andreas Olofsson
b63de8b1d8 Filter the txwr access 
We don't want reset/clock transaction to propagate through etx!
 2015-05-04 10:37:27 -04:00
Andreas Olofsson
1ee720fc67 Organization changes 
-dma with packet format
-using the fifo_cdc block
 2015-05-03 23:29:32 -04:00