Commit Graph

3777 Commits

Author SHA1 Message Date
Udi Finkelstein eb40278a16 Turned a few member functions into const, esp. dumpAst(), dumpVlog(). 2017-09-30 07:37:38 +03:00
Udi Finkelstein 72a08eca3d Resolved classical Bison IF/THEN/ELSE shift/reduce conflict using the textbook solution
(Oreilly 'Flex & Bison' page 189)
2017-09-30 06:39:07 +03:00
Clifford Wolf c5b204d8d2 Add first draft of eASIC back-end 2017-09-29 17:53:43 +02:00
Clifford Wolf e64b9d5a4d Fix synth_ice40 doc regarding -top default 2017-09-29 17:52:57 +02:00
Clifford Wolf dbfd8460a9 Allow $size and $bits in verilog mode, actually check test case 2017-09-29 11:56:43 +02:00
Clifford Wolf 637a02eb5c Merge pull request #425 from udif/udif_dollar_bits
Add $bits() and $size()
2017-09-29 11:39:36 +02:00
Clifford Wolf 29f8acf095 Merge pull request #421 from stephengroat/osx-travis
Add osx tests using brew bundle
2017-09-28 14:45:47 +02:00
Stephen 57b3c34e69 delete bad backslash 2017-09-27 16:52:20 -07:00
Stephen 1ba06cefcc forgot to install bundles 2017-09-27 16:51:50 -07:00
Stephen Groat de0797f073 Add osx tests using brew bundle 2017-09-27 16:49:03 -07:00
Clifford Wolf 30396270a2 Increase maximum LUT size in blifparse to 12 bits 2017-09-27 15:27:42 +02:00
Udi Finkelstein e951ac0dfb $size() now works correctly for all cases!
It seems the issues was that AST_MULTIRANGE is converted into a multirange_dimensions[] array on the AST_MEMORY node directly.
2017-09-26 20:34:24 +03:00
Udi Finkelstein 6ddc6a7af4 $size() seems to work now with or without the optional parameter.
Multidimensional arrays still don't work. I suspect the problem is that the array is flattened into a 1D array before $size() is evaluated.
2017-09-26 19:18:25 +03:00
Clifford Wolf 91d9c50bb3 Parse reals as string in JSON front-end 2017-09-26 14:37:03 +02:00
Clifford Wolf 660473a485 Merge branch 'vlogpp-inc-fixes' 2017-09-26 14:02:57 +02:00
Clifford Wolf 2c04d883b1 Minor coding style fix 2017-09-26 13:50:14 +02:00
Clifford Wolf cb1d439d10 Merge branch 'master' of https://github.com/combinatorylogic/yosys into combinatorylogic-master 2017-09-26 13:48:13 +02:00
Udi Finkelstein 7e391ba904 enable $bits() and $size() functions only when the SystemVerilog flag is enabled for read_verilog 2017-09-26 09:19:56 +03:00
Udi Finkelstein 2dea42e903 Added $bits() for memories as well. 2017-09-26 09:11:25 +03:00
Udi Finkelstein 17f8b41605 $size() now works with memories as well! 2017-09-26 08:36:45 +03:00
Udi Finkelstein 64eb8f29ad Add $size() function. At the moment it works only on expressions, not on memories. 2017-09-26 06:25:42 +03:00
Clifford Wolf 2cc09161ff Fix ignoring of simulation timings so that invalid module parameters cause syntax errors 2017-09-26 01:52:59 +02:00
combinatorylogic 64ca0be971 Adding support for string macros and macros with arguments after include 2017-09-21 18:25:02 +01:00
Clifford Wolf 143c0abd33 Merge pull request #413 from azonenberg/extract-reduce-tweaks
Added support for off-chain loads in extract_reduce
2017-09-16 11:31:37 +02:00
Andrew Zonenberg 2b65b65d70 Added missing "break" 2017-09-15 17:54:52 -07:00
Andrew Zonenberg 7b3966714c Implemented off-chain support for extract_reduce 2017-09-15 13:59:18 -07:00
Andrew Zonenberg 3404934c9c extract_reduce now only removes the head of the chain, relying on "clean" to delete upstream cells. Added "-allow-off-chain" flag, but it's currently ignored. 2017-09-15 13:59:05 -07:00
Clifford Wolf 5c4ea1366f Merge branch 'master' of github.com:cliffordwolf/yosys 2017-09-15 21:28:16 +02:00
Clifford Wolf 76c11d7454 Update ABC to hg rev cd6984ee82d4 2017-09-15 21:25:59 +02:00
Clifford Wolf ce78717e36 Merge pull request #412 from azonenberg/reduce-fixes
extract_reduce: Fix segfault on "undriven" inputs
2017-09-14 22:36:25 +02:00
Robert Ou ab1bf8d661 extract_reduce: Fix segfault on "undriven" inputs
This is easily triggered when un-techmapping if the technology-specific
cell library isn't loaded. Outputs of technology-specific cells will be
seen as inputs, and nets using those outputs will be seen as undriven.
Just ignore these cells because they can't be part of a reduce chain
anyways.
2017-09-14 12:54:44 -07:00
Clifford Wolf 498526cc0b Merge pull request #411 from azonenberg/counter-extraction-fixes
Various improvements and bug fixes to extract_counter
2017-09-14 21:44:26 +02:00
Clifford Wolf b0b2f3fe29 Merge pull request #410 from azonenberg/opt_demorgan
Added "opt_demorgan" pass (fixes #408)
2017-09-14 21:42:34 +02:00
Andrew Zonenberg 66e8986ae7 Minor changes to opt_demorgan requested during code review 2017-09-14 10:35:25 -07:00
Andrew Zonenberg 367d6b2194 Fixed bug where counter extraction on non-GreenPAK devices incorrectly handled parallel counter output 2017-09-14 10:27:10 -07:00
Andrew Zonenberg c8f2f082c6 Added support for inferring counters with reset to full scale instead of zero 2017-09-14 10:26:43 -07:00
Andrew Zonenberg 122532b7e1 Added RESET_TO_MAX parameter to $__COUNT_ cell. Cannot yet be extracted. 2017-09-14 10:26:32 -07:00
Andrew Zonenberg 0484ad666d Added support for inferring counters with active-low reset 2017-09-14 10:26:21 -07:00
Andrew Zonenberg a84172b23b Initial support for extraction of counters with clock enable 2017-09-14 10:26:10 -07:00
Andrew Zonenberg c4a70a8cc3 Fixed typo in comment. Fixed bug where extract_counter would create up counters when it meant to create down counters. 2017-09-14 10:25:51 -07:00
Andrew Zonenberg 6da5d36968 Initial version of opt_demorgan is functioning for AND/OR gates. Not the prettiest results for bus inputs, but this can be improved 2017-09-12 18:47:46 -07:00
Clifford Wolf f9d023c53f Add src attribute to extra cells generated by proc_dlatch 2017-09-09 10:18:08 +02:00
Clifford Wolf 13eb47c692 Add src arguments to all cell creator helper functions 2017-09-09 10:16:48 +02:00
Clifford Wolf 7d41c5e177 Further improve extract_fa (but still buggy) 2017-09-02 16:39:17 +02:00
Clifford Wolf 2f75240e36 Merge pull request #406 from azonenberg/coolrunner-techmap
Coolrunner techmapping improvements
2017-09-02 13:43:51 +02:00
Clifford Wolf 3331c281db Merge pull request #405 from azonenberg/gpak-refactoring
Gpak refactoring
2017-09-02 13:43:36 +02:00
Robert Ou 5f65e24ccb coolrunner2: Finish fixing special-use p-terms 2017-09-01 07:22:16 -07:00
Robert Ou fa04366f38 coolrunner2: Generate a feed-through AND term when necessary 2017-09-01 07:22:01 -07:00
Robert Ou 6775177171 coolrunner2: Initial fixes for special p-terms
Certain signals can only be controlled by a product term and not a
sum-of-products. Do the initial work for fixing this.
2017-09-01 07:21:51 -07:00
Robert Ou 7f08be4304 coolrunner2: Fix mapping of flip-flops 2017-09-01 07:21:39 -07:00