- 06 Mar, 2019 2 commits
-
-
SergeyDegtyar committed
-
SergeyDegtyar committed
-
- 05 Mar, 2019 1 commit
-
-
SergeyDegtyar committed
-
- 04 Mar, 2019 1 commit
-
-
SergeyDegtyar committed
-
- 01 Mar, 2019 1 commit
-
-
'a'.
SergeyDegtyar committed
-
- 28 Feb, 2019 1 commit
-
-
Add tests for splitnets command Expand coverage for trace command Expand coverage for splice command (tests for options) Expand coverage for scc command (different loops) Expand coverage for rename command Change test for passes/techmap/deminout Update test for 'dff2dffe -direct' command Add test for 'dffsr2dff' command Expand coverage for iopadmap command
SergeyDegtyar committed
-
- 27 Feb, 2019 1 commit
-
-
testing problems: Issue #61 does not contain enough info. Issues #82 and #83 I skipped because of errors. --------------------- All of tasks will go to group regression, please change naming convention there so files are named according to issue number (for example issue_86 for directory and issue_86.ys script) directory should contain exact example from issue, or in case it was just generally described you can create your own example according to explanation. Idea for these tasks is not getting greater coverage but making sure old issues fixed are never broken again. Usually problem was failing of yosys to generate any kind of output, in most of cases bellow, so just checking if yosys did not assert any error is enough at least for these. Issues bellow are taken from last two pages on issue tracker. Some of these can really be trivial, and issue contains all info you need. Some maybe need more thinking to be sure that you understood issue correctly. In any case you are free to continue going trough the list of issues and select new from it to cover. I will see to crate some online spreadsheet that will have issue number, type (if it is something that needed to be fixed or was a new feature or some user error, and will have a status if it is covered by test or not. That way we can have better overview what is going on. So if you see that something is without a way to reproduce, or it was user error just skip it, and when we have a spreadsheet we can update there. 1. https://github.com/YosysHQ/yosys/issues/86 2. https://github.com/YosysHQ/yosys/issues/85 3. https://github.com/YosysHQ/yosys/issues/84 4. https://github.com/YosysHQ/yosys/issues/83 5. https://github.com/YosysHQ/yosys/issues/82 6. https://github.com/YosysHQ/yosys/issues/81 7. https://github.com/YosysHQ/yosys/issues/78 8. https://github.com/YosysHQ/yosys/issues/71 9. https://github.com/YosysHQ/yosys/issues/67 10. https://github.com/YosysHQ/yosys/issues/65 11. https://github.com/YosysHQ/yosys/issues/61 12. https://github.com/YosysHQ/yosys/issues/59 13. https://github.com/YosysHQ/yosys/issues/41 14. https://github.com/YosysHQ/yosys/issues/18
SergeyDegtyar committed
-
- 21 Feb, 2019 1 commit
-
-
SergeyDegtyar committed
-
- 20 Feb, 2019 1 commit
-
-
simple & misc =========== Note that some of commands you can not test with checking with testbench so those place in misc. 1. passes/cmds/add.cc Note that here you need to load some existing verilog and add additional wires, inputs or outputs 2. passes/cmds/blackbox.cc you could create design with sub module, execute blackbox and check if sub module is replaced with blackbox module. 3. passes/cmds/bugpoint.cc 4. passes/cmds/chformal.cc 5. passes/cmds/chtype.cc 6. passes/cmds/connect.cc Maybe can be covered together with add command 7.passes/cmds/connwrappers.cc 8. passes/cmds/design.cc missing covering -import option 9.passes/cmds/plugin.cc 10. passes/cmds/rename.cc rename parts of existing design 11. /passes/cmds/select.cc Lot of options is not used , so room to improve 12.passes/cmds/setattr.cc note there are 3 commands to cover here 13. passes/cmds/setundef.cc setting with one, anyseq, anyconst ... 14. passes/sat/assertpmux.cc 15. passes/sat/async2sync.cc 16. passes/sat/eval.cc 17. passes/sat/freduce.cc 18. passes/sat/miter.cc run with -assert option 19. passes/sat/sat.cc many options are not tested 20. passes/sat/sim.cc 21. passes/techmap/flowmap.cc
SergeyDegtyar committed
-
- 06 Feb, 2019 1 commit
-
-
Testing problems: 1. coverage_html/passes/equiv/equiv_make.cc.gcov.html - 164-227 are not covered; 2. coverage_html/passes/equiv/equiv_add.cc.gcov.html - is not covered ("ERROR: This command must be executed in module context!")
SergeyDegtyar committed
-
- 05 Feb, 2019 1 commit
-
-
backends ======= 1. backends/firrtl/firrtl.cc – skipped because of error: ERROR: Unclocked write port 0 on memory top.ram. Add model with memory block too, so you can cover line block starting at line 419 simple ========= + 1. passes/techmap/attrmap.cc There is example in help block create example with (* keep="true" *) attribute and run example + 2. passes/techmap/dff2dffe.cc execute with -unmap parameter + 3. passes/techmap/dff2dffs.cc + 4. passes/techmap/extract.cc Lot of variations of input parameters is not excersised ? 5. passes/techmap/extract_counter.cc This is specific to greenpak4 architecture, but idea is to have counter in design and call this after synth_greenpak + 6. passes/techmap/shregmap.cc. This also seams to be greenpak4 specific, but extracting shift register misc ======== 1. passes/techmap/insbuf.cc Testing problems: 1. backends/firrtl/firrtl.cc – skipped because of error: ERROR: Unclocked write port 0 on memory top.ram. Add model with memory block too, so you can cover line block starting at line 419 2.dff2dffe -direct - skipped: ERROR: Found error in internal cell \dffe.$procdff$47 ($dffe) at 3.insbuf -buf $_BUF_ in out - skipped: ERROR: Found error in internal cell
SergeyDegtyar committed
-
- 29 Jan, 2019 2 commits
-
-
SergeyDegtyar committed
-
MISC ===== Tasks in this group are just generating some output, so no traditional test bench could be made. Writing output from yosys and checking if result is expected is advised. 1.http://scratch.clifford.at/coverage_html/passes/tests/test_abcloop.cc.gcov.html This represents some internal tests for ABC loops 2. http://scratch.clifford.at/coverage_html/passes/tests/test_cell.cc.gcov.html note that you need to provide multiple tests in order to cover testing of all cell types. Also output should be checked if it contain "Fail" or "Error" in it 3. http://scratch.clifford.at/coverage_html/passes/cmds/torder.cc.gcov.html Use some verilog file, do proc and this should display you all cells in designed ordered. Require multiple different parameters to cover all. 4.http://scratch.clifford.at/coverage_html/passes/cmds/trace.cc.gcov.html loading verilog file and doing "trace proc" will create output with additional logs, not much to check from output to be honest 5. http://scratch.clifford.at/coverage_html/passes/cmds/write_file.cc.gcov.html just write to file and check if written output is fine 6. http://scratch.clifford.at/coverage_html/passes/cmds/stat.cc.gcov.html if you have liberty cell library files, try it with those specified. 7. http://scratch.clifford.at/coverage_html/passes/cmds/show.cc.gcov.html note that you would always need to set -viewer since you do not wish to call graphwiz. Need covering of various parameters on a reasonable complicated design to cover. 8. http://scratch.clifford.at/coverage_html/passes/cmds/scc.cc.gcov.html detection of logic loops, output needs to be tested in order to check if loop was detected on not. 9. http://scratch.clifford.at/coverage_html/passes/cmds/scatter.cc.gcov.html After doing synth on existing design, just run this step, can do opt_clean after and there should be lot of temporary and unused wires removed. 10. http://scratch.clifford.at/coverage_html/passes/cmds/rename.cc.gcov.html can load design and rename top modue, output should contain module with new name 11. http://scratch.clifford.at/coverage_html/passes/cmds/qwp.cc.gcov.html Use it on simple design since it takes a while on big ones. 12. http://scratch.clifford.at/coverage_html/passes/cmds/ltp.cc.gcov.html Check if returns expected value for your design 13. http://scratch.clifford.at/coverage_html/passes/cmds/edgetypes.cc.gcov.html it just outputs report, so not sure what would be valid to check, just execute for now. 14. http://scratch.clifford.at/coverage_html/passes/cmds/delete.cc.gcov.html load existing design and try removing inputs, outputs,... complete design 15. http://scratch.clifford.at/coverage_html/passes/cmds/cover.cc.gcov.html it just prints report, so nothing to check
SergeyDegtyar committed
-
- 27 Jan, 2019 1 commit
-
-
Miodrag Milanovic committed
-
- 25 Jan, 2019 3 commits
-
-
add tests for synth_anlogic and 'regression'; Megre commit: "Fix tests according to latest yosys and proper...
Miodrag Milanović committed -
Miodrag Milanović committed
-
Merge commit "Fix tests according to latest yosys and proper gitignore files". Make the same changes for 'regression'.
SergeyDegtyar committed
-
- 24 Jan, 2019 4 commits
-
-
Miodrag Milanovic committed
-
Miodrag Milanovic committed
-
1. Add tests for synth_anlogic command Now this tests are commented because of: + iverilog -o testbench ../testbench.v synth.v ../../common.v ../../../../../techlibs/common/simcells.v ../../../../../techlibs/anlogic/cells_sim.v ../../../../../techlibs/anlogic/cells_sim.v:20: error: Unable to bind wire/reg/memory `A' in `testbench.uut._09_' ../../../../../techlibs/anlogic/cells_sim.v:20: error: Unable to elaborate r-value: (INIT)>>(A) 2 error(s) during elaboration. 2. Add 'regression' test
SergeyDegtyar committed -
Add tests for "II architecture" and "III backends" and "IV frontends"
Miodrag Milanović committed
-
- 23 Jan, 2019 1 commit
-
-
IV frontends ============= For frontends idea is that you do not do read_verilog but some you are reading input in different format. Since most of formats yosys can also read and write you can use plain verilog write blif for example and use that as source for this coverage. For liberty files many examples can be found so that is also in list of tasks 1. http://scratch.clifford.at/coverage_html/frontends/blif/blifparse.cc.gcov.html 2. http://scratch.clifford.at/coverage_html/frontends/ilang/ilang_frontend.cc.gcov.html 3. http://scratch.clifford.at/coverage_html/frontends/json/jsonparse.cc.gcov.html 4. http://scratch.clifford.at/coverage_html/frontends/liberty/liberty.cc.gcov.html
SergeyDegtyar committed
-
- 15 Jan, 2019 1 commit
-
-
SergeyDegtyar committed
-
- 02 Jan, 2019 7 commits
-
-
Signed-off-by: Clifford Wolf <clifford@clifford.at>
Clifford Wolf committed -
I simple
Clifford Wolf committed -
Signed-off-by: Clifford Wolf <clifford@clifford.at>
Clifford Wolf committed -
Signed-off-by: Clifford Wolf <clifford@clifford.at>
Clifford Wolf committed -
First tests for pilot project
Clifford Wolf committed -
Clifford Wolf committed
-
5. clk2fflogic (104 - 144,180-195 are not covered) 9. memory_nordff(75-101 is not covered) 10. memory_unpack(91-108 is not covered) 12. hierarchy (the coverage from 44% increased to 61,3%)
SergeyDegtyar committed
-
- 19 Dec, 2018 1 commit
-
-
SergeyDegtyar committed
-
- 16 Dec, 2018 1 commit
-
-
Signed-off-by: Clifford Wolf <clifford@clifford.at>
Clifford Wolf committed
-
- 12 Dec, 2018 2 commits
-
-
SergeyDegtyar committed
-
SergeyDegtyar committed
-
- 21 Sep, 2018 1 commit
-
-
Signed-off-by: Clifford Wolf <clifford@clifford.at>
Clifford Wolf committed
-
- 06 Sep, 2018 1 commit
-
-
Signed-off-by: Clifford Wolf <clifford@clifford.at>
Clifford Wolf committed
-
- 05 Sep, 2018 2 commits
-
-
Signed-off-by: Clifford Wolf <clifford@clifford.at>
Clifford Wolf committed -
Signed-off-by: Clifford Wolf <clifford@clifford.at>
Clifford Wolf committed
-
- 28 Aug, 2018 1 commit
-
-
Signed-off-by: Clifford Wolf <clifford@clifford.at>
Clifford Wolf committed
-
- 22 Aug, 2018 1 commit
-
-
Signed-off-by: Clifford Wolf <clifford@clifford.at>
Clifford Wolf committed
-