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
1,167 Commits 7 Branches 1 Tag
Commit Graph

1167 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Andreas Olofsson
f9910d6094 Fixing depth parameter in fifo 2016-04-14 15:54:23 -04:00
Andreas Olofsson
8faa91bc52 Instantiating oh_memory_ram in dp/sp memories 2016-04-14 10:30:19 -04:00
Andreas Olofsson
709f91c306 Using DEPTH as specifier for memory 
- More natural design interface (than AW)
 2016-04-14 10:29:49 -04:00
Andreas Olofsson
a08b5d55b5 Adding generic RAM module 
- one read, one write port
- needed something simpler for modeling, maximuze code reuse
- _sp/_dp are really there so that you can build designs that will abstract away chip/fpga details
 2016-04-14 10:28:30 -04:00
Andreas Olofsson
d730e46e0b Fixing oh_dsync instantiation 
- to add new reset pin (can be tied to 1'b1 sometimes)
 2016-04-13 20:54:12 -04:00
Andreas Olofsson
70340040ce Fixing killer bugs in async fifo! 
- Adding reset on dsync (sometimes there is no clock)
- Separated reset for wr/rd
- One of the sync clocks was wrong (found by review)
 2016-04-13 20:50:48 -04:00
Andreas Olofsson
42c7f4ed0d Improving dsync 
- adding ability to drive random delay values on each bit in multibit sync
- adding async reset (need was found during integration testing)
 2016-04-13 20:48:37 -04:00
Andreas Olofsson
a6e1b22f9f Adding async reset to debuncer. 
- Can't guarantee that there will be a clock at startup.
- All IO modules with suspect clocking situation should have async reset on all key signals.
- To test, then you turn off clock and look for "x propagation".
 2016-04-13 20:47:22 -04:00
Andreas Olofsson
0432f22c67 Simplifyinf bin2gray interface 
- Using gray/bin as inputs was cute, but not useful when actually instantiating and reading the code. in/out works better.
 2016-04-13 20:46:02 -04:00
Andreas Olofsson
3ccd72016e Merge pull request #61 from olajep/gpio-driver 
GPIO driver
 2016-04-13 09:56:37 -04:00
Ola Jeppsson
510b6669a8 GPIO: Add raw register access API 
Add gpio_reg_write() and gpio_reg_read() functions.

Signed-off-by: Ola Jeppsson <ola@adapteva.com>
 2016-04-13 15:31:47 +02:00
Ola Jeppsson
137994f13a GPIO: Drop "oh" prefix in file names and API 
Signed-off-by: Ola Jeppsson <ola@adapteva.com>
 2016-04-13 15:30:16 +02:00
Andreas Olofsson
ff7018b4ac Making all significant GPIO registers readable for now 
- Would really prefer if this could be handled in the driver..
 2016-04-13 09:02:30 -04:00
Andreas Olofsson
33f3af9f79 Making GPIO_DIR readable 2016-04-13 09:00:24 -04:00
Ola Jeppsson
f22b833908 GPIO: Add GPIO driver 
Signed-off-by: Ola Jeppsson <ola@adapteva.com>
 2016-04-13 13:24:29 +02:00
Ola Jeppsson
94b73eb2d0 GPIO: Remove old driver template 
Signed-off-by: Ola Jeppsson <ola@adapteva.com>
 2016-04-13 13:17:32 +02:00
Andreas Olofsson
30555c5375 Adding nreset to command reg 
- Not 100% sure on this. Ideally I think you would want the ss signal to be the "reset" of all this logic, but the command register gets too tricky.
- Is there an issue in having a async active low reset pin on the this interface. Considering that this is only a problem for the remote fetch logic (which will definitely have a nreset signal) I think this is ok...
 2016-04-12 09:42:52 -04:00
Andreas Olofsson
24957b73d3 Making interrupt 0 non-maskable. 2016-04-11 20:36:58 -04:00
Andreas Olofsson
e6ed2f2855 Cleanup register addresses of PIC 2016-04-11 20:32:42 -04:00
Andreas Olofsson
6c0ce1ff00 Adding README for PIC 2016-04-11 20:31:30 -04:00
Andreas Olofsson
191f3b4db8 Fixing gpio initial use case example 2016-04-11 19:27:23 -04:00
Andreas Olofsson
7ab30d9f9b Initial thoughts on a gpio driver 2016-04-11 15:46:54 -04:00
Andreas Olofsson
101fb7de66 Copyright transfer 
- All OH! code transferred from Adapteva to non-profit Parallella Foundation
 2016-04-11 12:05:29 -04:00
Andreas Olofsson
1a790eaf24 Preparing for safer clock setting changes 2016-04-11 12:04:30 -04:00
Andreas Olofsson
f25aa0ff45 Cleanup after common lib refactoring 
- clockdiv takes no parameter
 2016-04-11 12:03:44 -04:00
Andreas Olofsson
2688bc5aa4 Refactoring common library 
- Updating interfaces to 2005 style
- Adding license pointers to all files
 2016-04-11 12:01:59 -04:00
Andreas Olofsson
5b8328826e Making rsync interface scalar 
- Who ever uses a vectorized reset....
(if you do, use generate...)
 2016-04-11 11:59:18 -04:00
Andreas Olofsson
fef2bdbe08 Making latch interface generic 
- Should look like a standard cell gate
 2016-04-11 11:58:54 -04:00
Andreas Olofsson
5304fd2df2 Fixing compilation errors 2016-04-11 11:13:08 -04:00
Andreas Olofsson
01a74a3db9 Removing ce from oddr 
- Better to turn off clock if you want to save power
- The CE was actually increasing the power by N x by making every data signal a clock :-)
 2016-04-11 10:30:02 -04:00
Andreas Olofsson
f62d7c0975 Vectorizing csa modules 2016-04-11 09:34:57 -04:00
Andreas Olofsson
d2fbbcf341 Enhancing GPIO block 
- implementing interrupt handling
- edge/level support
- moving to single register for "DIR", avoids confusion
- adding ILATCLR for clearing interrupts
- changing to N parameter away from clumsy 64 bit chunk
- if you want longer GPIO, instantiate more vectors
- up to N==AW supported by this model
- trying out more modern Verilog features
 2016-04-10 23:24:26 -04:00
Andreas Olofsson
233a4150dd Fixing remote fetch spi feature 
- Added missing reset
- Using rising edge of ss (after nreset deasserted) as indicator (and command reg is set to fetch)
- Seems like it should be robust across all clk/sclk frequencies. Can use ss high pause to control timing.
-Assumption, reset is asserted until ss has settled on to "high" at powerup...seems like this is a must not just for fetch.
 2016-04-10 20:56:28 -04:00
Andreas Olofsson
e22a32853b Merge branch 'master' of github.com:parallella/oh 2016-04-07 19:08:44 -04:00
Andreas Olofsson
7fc5d7da1c Cleanin up wait signal 
-There were some corner case problems when driving non safe interfaces like the SPI...can't always have the luxuary of a fifo
 2016-04-07 19:07:44 -04:00
Andreas Olofsson
f610a31d42 Cleanup spi readme 2016-04-07 19:05:01 -04:00
Andreas Olofsson
534f847592 Adding SPI examples 2016-04-07 19:03:11 -04:00
Andreas Olofsson
dd238b5356 SPI now with working emesh packet transfers working 
- (see test_basic.emf for example)
- basically works by pushing consecutive bytes into FIFO (104 bits)
- need to make sure the FIFO does not go empty
 2016-04-07 18:46:26 -04:00
Andreas Olofsson
f5104a62ea Launching data on falling edge on default spi mode 
(as it should be...not sure how I missed this)
 2016-04-07 18:44:41 -04:00
Andreas Olofsson
f1babdb680 Fixing spi master pushback bug 2016-04-07 18:44:24 -04:00
Andreas Olofsson
952657dbd5 Reomving redundant emode feature from spi 2016-04-07 18:44:05 -04:00
Andreas Olofsson
0a1690de94 Fixing epiphany spi fetch 
- adding a separate serializer for the remote fetch (cleaner)
- fixing the load signal to load on all bytes while ss low
- removing reset from command register
- review sync for spi...feels ok to use the rising edge of ss to sync to clock
(note, there must be a free running clk for remote fetch)
- adding wait pushback on remote fetch
 2016-04-07 18:39:39 -04:00
Andreas Olofsson
42649ef2c1 SPI slave fixup 
- removing emode (was redundant)
- fixing user regs issue
- adding status bit to poll for returned data on fetch
 2016-04-07 18:38:29 -04:00
Andreas Olofsson
542fec92e9 Completing test for "epiphany" remote read over spi 2016-04-07 18:37:39 -04:00
Andreas Olofsson
ea3c38e189 Adding shutdown signal for macro 
- Needed for proper power management, should not be lumped with config
 2016-04-06 11:53:03 -04:00
Andreas Olofsson
1b512cdbde Removing asic config dependancy 
- No plans for dual ported memory use for now, if this changes and there needs to be differentiation, then this should probably be a parameter rather than define. (since you might want to choose between regs based and macro based depending on size of memory).
 2016-04-06 11:51:03 -04:00
Andreas Olofsson
83c3caf1c4 Adding back resets 
- Was a little overzealous removing resets in module. Note that these problems popped in recently, were no in the original silicon.
 2016-04-06 11:50:09 -04:00
Andreas Olofsson
8c224d6771 Fixing memory macro instantiation 
- too many variations (polarity etc) on memory interface, stay with generics
(repair + config)
- hookup magic happens under the hood
 2016-04-05 22:43:29 -04:00
Andreas Olofsson
1c1a4dae66 Merge branch 'master' of github.com:parallella/oh 2016-04-05 16:10:50 -04:00
Andreas Olofsson
2bc7c82271 Driving realistic power ramp in startup 2016-04-05 16:10:25 -04:00