| *** tpb has joined #tomu | 00:00 | |
| twoolie | although checking the head of hacker-foboot-v2.0.3.dfu shows that it doesn't have the magic string for a properly formatted dfu file... | 00:00 |
|---|---|---|
| *** asante has joined #tomu | 00:10 | |
| *** asante has joined #tomu | 00:10 | |
| xobs | twoolie: the dfu prefix is a nonstandard ST-only extension. The DFU standard only specifies a suffix, which those files have. You can check it with the `dfu-suffix` command. | 00:41 |
| twoolie | ah, cool! thanks :) is dfu-suffix from the dfu-utils package? | 00:52 |
| CarlFK | seems so: dfu-suffix ... sudo apt install dfu-util | 00:53 |
| xobs | twoolie: I do know you can strip off the 16-byte dfu suffix with `dfu-util`, or even just `dd`. I also know that output file is created by https://github.com/im-tomu/foboot/blob/master/booster/make-booster.c which just copies the file to the start and sticks `booster` after it. | 01:00 |
| tpb | Title: foboot/make-booster.c at master · im-tomu/foboot · GitHub (at github.com) | 01:00 |
| xobs | (Note: it also patches up one of the reset vectors, but if you're looking for top.bin then you largely can ignore that, unless you're doing something exotic like manually flashing it) | 01:01 |
| *** emeb has quit IRC | 01:16 | |
| *** rohitksingh has quit IRC | 01:34 | |
| *** CarlFK has quit IRC | 01:40 | |
| *** rohitksingh has joined #tomu | 02:06 | |
| *** squelch has joined #tomu | 02:18 | |
| *** squelch has quit IRC | 02:52 | |
| *** alexhw has quit IRC | 03:27 | |
| *** alexhw has joined #tomu | 03:29 | |
| *** squelch has joined #tomu | 03:49 | |
| *** CarlFK has joined #tomu | 03:56 | |
| *** xkapastel has joined #tomu | 04:33 | |
| *** rohitksingh has quit IRC | 04:45 | |
| *** rohitksingh has joined #tomu | 04:46 | |
| *** cdmatter has quit IRC | 05:44 | |
| *** cdmatter has joined #tomu | 05:45 | |
| *** squelch has quit IRC | 08:07 | |
| *** GNUtoo has quit IRC | 08:36 | |
| *** GNUtoo has joined #tomu | 08:37 | |
| *** CarlFK has quit IRC | 08:39 | |
| *** xkapastel has quit IRC | 09:22 | |
| *** rohitksingh has quit IRC | 10:04 | |
| *** xkapastel has joined #tomu | 10:12 | |
| *** alexhw has quit IRC | 13:57 | |
| *** alexhw has joined #tomu | 13:59 | |
| *** xkapastel has quit IRC | 14:02 | |
| *** im-tomu has left #tomu | 14:47 | |
| *** squelch has joined #tomu | 15:46 | |
| *** emeb has joined #tomu | 15:48 | |
| emeb | trying to build foboot. Gets most of the way through but then errors out: -s -f /tmp/yosys-abc-GTDyII/abc.script 2>&1 | 16:24 |
| emeb | ERROR: ABC: execution of command "/usr/local/bin/yosys-abc -s -f /tmp/yosys-abc-GTDyII/abc.script 2>&1" failed: return code 134. | 16:24 |
| emeb | I *think* I've got a fairly recent yosys: Yosys 0.9+1706 (git sha1 cb7bc6a1, clang 8.0.0 -fPIC -Os) | 16:24 |
| tnt | You need to track down what the error message is ... | 16:25 |
| tnt | You can also check in dmesg if it segfaulted. | 16:26 |
| emeb | dmesg appears clean | 16:27 |
| tnt | then it probably prints something that get absorbed by the build system of foboot ... gotta track down what it prints. | 16:27 |
| emeb | ah. ran that yosys cmd from cmdline and got: yosys-abc: src/opt/lpk/lpkCore.c:549: int Lpk_ResynthesizeNodeNew(Lpk_Man_t *): Assertion `nGain >= 1 - p->pPars->fZeroCost' failed. | 16:28 |
| emeb | Aborted (core dumped) | 16:28 |
| tnt | Well ... I'd say rebuilt yosys and hope something got fixed .... | 16:29 |
| emeb | :) | 16:29 |
| tnt | I know abc had some issue not long ago with assertion failure, although that was another one, not sure if it's related. | 16:30 |
| emeb | ok, thanks. I've been out of this stuff for a while and wasn't sure how volatile yosys is ATM. I built it about a month ago so I'll try again. | 16:35 |
| emeb | Nope - rebuilt from scratch w/ latest yosys from yosyshq github. still gets same error. | 16:44 |
| emeb | so I guess foboot is currently borken | 16:44 |
| *** squelch has quit IRC | 16:56 | |
| tnt | Well, it's more a yosys/abc issues. It should never assert, no matter the input. | 16:56 |
| emeb | true. for want of a nail... | 16:58 |
| tnt | You can try the binary toolchain from xobs. It's most likely what he used to build foboot binary releases. | 16:58 |
| tnt | so it might be a version that happens to not bug. | 16:58 |
| emeb | worth a try | 16:59 |
| daveshah | emeb: there's a workaround here https://github.com/YosysHQ/yosys/issues/1675#issuecomment-582541724 | 17:05 |
| tpb | Title: ABC: execution of command "..." failed: return code 134. · Issue #1675 · YosysHQ/yosys · GitHub (at github.com) | 17:05 |
| daveshah | it's a known issue in abc | 17:05 |
| emeb | cool - thx daveshah | 17:05 |
| *** im-tomu has joined #tomu | 17:07 | |
| * emeb wonders how xobs got around that if it's a known thing in abc... | 17:07 | |
| daveshah | It seemingly appeared from some recent ish Yosys change | 17:08 |
| emeb | that workaround succeeds if I hand-edit the autogenerated abc.script but I'm unsure of how to insert the workaround into the automated foboot build process since it creates a new script every time w/o the workaround. | 17:30 |
| emeb | gah - even grepping through the whole foboot build hierarchy I can't find an instance of that evil "dc2" string. How is the abc.script created? | 17:47 |
| *** squelch has joined #tomu | 18:12 | |
| *** rohitksingh has joined #tomu | 19:14 | |
| *** im-tomu has left #tomu | 19:55 | |
| *** im-tomu has joined #tomu | 19:59 | |
| *** CarlFK has joined #tomu | 20:26 | |
| emeb | hmm... ok - that's the wrong approach. workaround is to give yosys a different abc script with that scratchpad command. | 20:40 |
| emeb | However, inserting that into the self.nextpnr_yosys_template in the icestorm.py source causes a whole raft of new errors to be spewed. | 20:41 |
| *** rohitksingh has quit IRC | 21:03 | |
| *** rohitksingh has joined #tomu | 21:36 | |
| *** squelch has quit IRC | 22:32 | |
| *** rohitksingh has quit IRC | 22:47 | |
| *** squelch has joined #tomu | 23:22 | |
| *** squelch has quit IRC | 23:32 | |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!