| *** tpb has joined #yosys | 00:00 | |
| *** atk has joined #yosys | 00:00 | |
| *** craigo has quit IRC | 00:00 | |
| *** emeb has left #yosys | 00:15 | |
| *** vidbina_ has quit IRC | 00:55 | |
| *** lf has quit IRC | 00:55 | |
| *** lf_ has joined #yosys | 00:55 | |
| *** craigo has joined #yosys | 01:57 | |
| *** Degi_ has joined #yosys | 03:07 | |
| *** Degi has quit IRC | 03:07 | |
| *** Degi_ is now known as Degi | 03:07 | |
| *** citypw has joined #yosys | 03:35 | |
| *** _whitelogger has quit IRC | 03:42 | |
| *** _whitelogger has joined #yosys | 03:44 | |
| *** citypw has quit IRC | 04:23 | |
| *** futarisIRCcloud has joined #yosys | 04:50 | |
| *** FFY00_ has quit IRC | 04:54 | |
| *** FFY00_ has joined #yosys | 05:00 | |
| *** kraiskil has joined #yosys | 05:52 | |
| *** s_frit has quit IRC | 05:55 | |
| *** s_frit has joined #yosys | 05:56 | |
| *** kraiskil has quit IRC | 06:46 | |
| *** emeb_mac has quit IRC | 07:03 | |
| *** s_frit has quit IRC | 07:46 | |
| *** s_frit has joined #yosys | 07:47 | |
| *** s_frit has quit IRC | 08:19 | |
| *** s_frit has joined #yosys | 08:20 | |
| *** s_frit has quit IRC | 08:24 | |
| *** s_frit has joined #yosys | 08:24 | |
| *** shivampotdar has joined #yosys | 08:32 | |
| *** futarisIRCcloud has quit IRC | 08:59 | |
| *** wiizzard has quit IRC | 09:00 | |
| *** vidbina_ has joined #yosys | 09:10 | |
| *** jakobwenzel has joined #yosys | 09:34 | |
| *** vidbina_ has quit IRC | 10:04 | |
| *** FFY00_ has quit IRC | 10:06 | |
| *** FFY00_ has joined #yosys | 10:07 | |
| gatecat | somlo: done | 10:19 |
|---|---|---|
| *** craigo has quit IRC | 11:10 | |
| *** craigo has joined #yosys | 11:14 | |
| *** FFY00_ has quit IRC | 11:19 | |
| *** FFY00_ has joined #yosys | 11:20 | |
| *** jakobwenzel has quit IRC | 11:27 | |
| *** FFY00_ has quit IRC | 11:46 | |
| *** FFY00_ has joined #yosys | 11:47 | |
| *** kraiskil has joined #yosys | 11:54 | |
| *** kraiskil has quit IRC | 12:10 | |
| *** jakobwenzel has joined #yosys | 12:41 | |
| *** vidbina_ has joined #yosys | 12:59 | |
| *** jakobwenzel has quit IRC | 13:03 | |
| somlo | gatecat: thanks! | 13:20 |
| somlo | I requested a Fedora RPM build of yosys (at commit 9cdc6b5, as of yesterday). That involves building on a variety of architectures (686, x86_64, arm, ppc, s390x). | 13:48 |
| somlo | it builds everywhere, and passes tests on all but s390x, where tests/techmap/shiftx2mux.ys fails with | 13:49 |
| somlo | "ERROR: Called with -verify and proof did fail!" | 13:49 |
| somlo | currently I'm working on peeling off all the fedora-specific layers to try and create the most straightforward reproducer possible for a bug report (probably no way around involving qemu-s390x-static :) | 13:50 |
| somlo | in the mean time, any insight into the shiftx2mux.ys test much appreciated -- I'm not ready (yet) to simply turn off the "%check" (`make test`) portion of the build on s390x to paper over the issue -- on the off chance this failure might be highlighting an actual problem... | 13:51 |
| *** lethalbit has quit IRC | 13:57 | |
| *** lethalbit has joined #yosys | 13:59 | |
| gatecat | Thanks for looking into this - I'll raise this on the Yosys HQ slack too | 14:14 |
| mwk | ... oh yay | 14:19 |
| mwk | .... I wonder what happened to that cursed s390x VM I had laying around | 14:20 |
| somlo | mwk: after turning yosys #9cdc6b5 into a .src.rpm, I can reproduce it with `mock -r fedora-rawhide-s390x yosys*.src.rpm` -- which involves containers and emulation via qemu-s390x-static, i.e. "usermode" emulation) | 14:23 |
| mwk | any chance of getting some reproducer I could run with just plain qemu? | 14:25 |
| mwk | but anyway | 14:25 |
| somlo | that's what I'm trying to do, "peel off" all the fedora-specific crap before filing an issue on github | 14:25 |
| somlo | wouldn't want to inflict the RPM crap on anyone who isn't a Fedora nut :D | 14:26 |
| mwk | as far as I can see, that sat -verify call basically verifies the validity of a techmap + abc/abc9 on a design | 14:26 |
| mwk | the actual shiftx2mux part shouldn't be relevant to this failure | 14:26 |
| mwk | so the question is | 14:26 |
| somlo | abc bug? | 14:27 |
| mwk | is one of these passes somehow broken on s390x, or is the sat itself broken | 14:27 |
| mwk | not impossible | 14:27 |
| mwk | but anyway | 14:27 |
| mwk | I'd suggest modifying the test to dump rtlil at three points in the script | 14:27 |
| mwk | and uploading the resulting files somewhere so we can compare with known-good execution | 14:27 |
| mwk | please add: `write_rtlil gold.il` right before `design -save gold` | 14:28 |
| mwk | `write_rtlil gate1.il` right after first `select -assert-count 16 t:$lut` | 14:28 |
| mwk | `write_rtlil gate2.il` right after second `select -assert-count 16 t:$lut` | 14:28 |
| mwk | and upload the three il files that appear (or two if it crashes earlier) | 14:29 |
| mwk | also could you upload the full execution log of that test? I'm not even sure *which* sat call crashes | 14:29 |
| somlo | mwk: https://pastebin.com/kYmRATps -- I'll get back to you here once I collect the output. And yeah, I'll dig around for the execution log as well. Thanks! | 14:31 |
| tpb | Title: diff --git a/tests/techmap/shiftx2mux.ys b/tests/techmap/shiftx2mux.ysindex f7 - Pastebin.com (at pastebin.com) | 14:31 |
| mwk | the execution log is just the .err or .log file next to the test | 14:32 |
| somlo | got it, thx! | 14:32 |
| *** emeb has joined #yosys | 15:12 | |
| *** kraiskil has joined #yosys | 15:44 | |
| *** kraiskil has quit IRC | 15:56 | |
| *** vidbina_ has quit IRC | 15:58 | |
| *** vidbina_ has joined #yosys | 17:08 | |
| *** kraiskil has joined #yosys | 17:50 | |
| *** craigo has quit IRC | 18:10 | |
| *** kraiskil has quit IRC | 18:39 | |
| *** danvet has joined #yosys | 18:43 | |
| *** kraiskil has joined #yosys | 18:47 | |
| somlo | mwk: http://mirror.ini.cmu.edu/yosys/yosys-shiftx2mux-test.tar.gz; I ran the build on both x86_64 (successfully) and on s390x (where it never made it to the second gate) | 19:41 |
| somlo | mwk: happy to open an issue on github if you prefer, just LMK | 19:48 |
| mwk | well, sat doesn't pass here either on those | 19:54 |
| mwk | so... abc or techmap bug | 19:54 |
| mwk | somlo: is s390x the only big endian architecture you tested? | 20:02 |
| mwk | actually hmmmmm | 20:04 |
| somlo | mwk: the full list is armv7hl, i686, x86_64, aarch64, ppc64le, s390x | 20:06 |
| mwk | and that means a no | 20:06 |
| mwk | I just pulled up a ppc64 machine, let's try it | 20:06 |
| mwk | er, I mean, that means a yes; s390x is the only BE one on that list | 20:07 |
| mwk | yep | 20:07 |
| somlo | yeah, ppc64le sounds like the "le" bit is for little | 20:07 |
| mwk | the test fails | 20:07 |
| mwk | okay, this means I have a repro machine | 20:08 |
| somlo | oh, so it's an endianness thing | 20:08 |
| mwk | yes | 20:08 |
| mwk | not a good sign, is it | 20:08 |
| somlo | thank goodness I don't have to peel off the fedora container layers, if you can reproduce on your end :) | 20:08 |
| somlo | but yeah, sorry about the less than good news | 20:09 |
| mwk | and I just confirmed that it is, indeed, abc and not techmap | 20:09 |
| mwk | gods, I do *not* want to dig into abc sources to find a fucking endianness bug | 20:09 |
| somlo | I have mixed feelings of sadness and relief at clearly not being qualified to tackle this myself :D | 20:10 |
| somlo | at least not in any useful time frame... | 20:11 |
| mwk | and it's a dangerous one, too | 20:11 |
| mwk | as in, it silently mangles synthesized logic | 20:11 |
| mwk | given this issue, yosys should be considered unusable on big endian machines | 20:11 |
| mwk | at least for synthesis flows | 20:11 |
| somlo | ouch... | 20:12 |
| somlo | but at least we know, which is arguably better than the alternative | 20:12 |
| mwk | yes | 20:12 |
| mwk | please file a bug, this needs to be fixed | 20:12 |
| somlo | will do | 20:13 |
| cr1901_modern | oh shit ._. | 20:23 |
| somlo | mwk: issue #2645 | 20:26 |
| *** kraiskil has quit IRC | 20:35 | |
| *** emeb_mac has joined #yosys | 20:52 | |
| *** jakobwenzel has joined #yosys | 21:12 | |
| *** jakobwenzel has quit IRC | 21:31 | |
| *** vidbina_ has quit IRC | 22:03 | |
| *** jakobwenzel has joined #yosys | 22:07 | |
| mwk | somlo: the culprit seems to be the mfs2 pass in abc; I can make the testsuite pass by removing mfs2 from the default script in passes/techmap/abc.cc | 22:11 |
| *** jakobwenzel has quit IRC | 22:28 | |
| jix | mwk: I found the bug (see slack) | 22:29 |
| mwk | oh ffs | 22:32 |
| mwk | somlo: given the underlying mess, you may want to just fail the build on big-endian systems | 22:48 |
| mwk | this is just a case of someone baking a little-endian assumption deep into the code, fixing this is going to involve a *lot* of work | 22:50 |
| somlo | mwk: thanks, I'll ask in #fedora-devel how to go about "unsupporting" a package on one of their "supported" architectures (the irony is that I couldn't care less about s390x for myself or anyone I know :) | 22:53 |
| somlo | mwk: the baked-in assumption, is it in upstream abc? If so, any chance one of you kind folks can open a bug about it in upstream abc that I can reference to support my case with #fedora-devel ? | 22:55 |
| mwk | yes, upstream abc | 22:56 |
| sorear | are you asking a political question or a question about rpmbuild? | 22:56 |
| sorear | on second thought it's been several years since I did this and I'm not sure I could explain how | 22:57 |
| *** danvet has quit IRC | 22:57 | |
| sorear | but there's a way to stick architecture allowlists/denylists in spec files and a bunch have them | 22:57 |
| somlo | sorear: it's technically possible to specify "this package should not be built on this arch" (and quite a simple one-liner in the rpm spec file) | 22:59 |
| sorear | yeah but saying "it's easy" without details is not useful, so I'm not doing that | 23:00 |
| somlo | but (without having crossed all my t's and dotted all my i's) I also assume that if we had yosys and abc packaged for all arches before, stopping that on arch "foo" would be easier if I could point at some good reason :) | 23:00 |
| somlo | besides, having the breakage documented and on record is a good thing on general principle, nevermind "packaging politics"... | 23:01 |
| somlo | I'd do it, but (as I might have mentioned earlier) this is WAAAAY deeper and darker voodoo than I am currently qualified to handle -- all I did was step on the landmine, and I don't have the vocabulary to describe it in technical detail :) | 23:03 |
| mwk | somlo: you can link to https://github.com/YosysHQ/yosys/issues/2645 | 23:08 |
| mwk | I added a short summary of the problem | 23:09 |
| somlo | mwk, thanks! I also have to contact the principal fedora maintaner of abc, and having that to point at helps | 23:10 |
| *** richbridger has quit IRC | 23:49 | |
| *** richbridger has joined #yosys | 23:50 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!