Friday, 2023-05-12

*** gnucode <gnucode!~gnucode@user/jab> has quit IRC00:20
*** gnucode <gnucode!~gnucode@user/jab> has joined #libre-soc00:20
programmerjakewell, my build server randomly crashed for the 3rd or 4th time, I found no indication of what went wrong in any of the logs, and the temperature logging I added said the temperature was fine...02:28
programmerjakeI'm thinking it's time to replace the motherboard and cpu with the ones from my old desktop ... even if piledriver is less likely to be able to be hacked/backdoored, having to constantly manually turn it on again is imho too much. also, the replacement system has more ram and is probably 2x faster.02:30
programmerjakelkcl: thank you for fixing the broken tests!02:47
*** gnucode <gnucode!~gnucode@user/jab> has quit IRC02:51
*** libredev <libredev!> has quit IRC09:03
*** leohoo_sdu[m] <leohoo_sdu[m]!~leohoosdu@2001:470:69fc:105::3:56ef> has joined #libre-soc09:29
*** libredev <libredev!> has joined #libre-soc10:02
*** ghostmansd[m] <ghostmansd[m]!> has quit IRC10:05
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc10:06
programmerjakethe internet is currently down here, so sorry if anyone needed the build server...11:09
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC11:59
*** ghostmansd[m] <ghostmansd[m]!> has joined #libre-soc11:59
*** ghostmansd <ghostmansd!> has joined #libre-soc12:02
*** ghostmansd[hexch <ghostmansd[hexch!> has joined #libre-soc12:03
*** ghostmansd <ghostmansd!> has quit IRC12:06
*** ghostmansd[pc] <ghostmansd[pc]!> has joined #libre-soc12:10
*** ghostmansd <ghostmansd!> has joined #libre-soc12:12
*** ghostmansd[hexch <ghostmansd[hexch!> has quit IRC12:14
*** ghostmansd[hexch <ghostmansd[hexch!> has joined #libre-soc12:15
*** ghostmansd[pc] <ghostmansd[pc]!> has quit IRC12:15
*** ghostmansd <ghostmansd!> has quit IRC12:19
*** ghostmansd[hexch is now known as ghostmansd13:43
ghostmansdlkcl, a quick question: why is self.insnlog established inside execute_one()? I suspect you circumvent it by using append mode for stream, but shouldn't we rather establish the stream before executing anything?14:29
*** octavius <octavius!~octavius@> has joined #libre-soc14:55
octaviusvaino[m], as promised, the video on Microwatt simulation :)15:11
ghostmansdprogrammerjake, lkcl, it seems that ISACaller is used not only by tests. I think nmutil.get_test_path is not a real option. How about creating a temporary file per ISA class instance?15:24
vaino[m]octavius: thanks, I'll check it out :)15:35
*** ghostmansd[m] <ghostmansd[m]!> has quit IRC15:36
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc15:37
octaviusghostmansd, lkcl here : just pass in the filename as an option to ISACaller15:41
octavius(put it at the end on the list of arguments)15:41
octaviusdefault to None15:41
octaviusyou get the  idea15:41
*** octavius <octavius!~octavius@> has quit IRC15:58
*** psydroid <psydroid!~psydroid@user/psydroid> has joined #libre-soc17:19
*** psydroid <psydroid!~psydroid@user/psydroid> has quit IRC18:04
programmerjakeghostmansd: just pass the filename into ISACaller and use nmutil.get_test_path in TestRunnerBase18:14
programmerjakeinside the with self.subTest block18:15
programmerjakeyou may need to add a function to ISACaller to close the old file and open a new file with passed-in filename18:16
*** leohoo_sdu[m] <leohoo_sdu[m]!~leohoosdu@2001:470:69fc:105::3:56ef> has quit IRC18:49
lkclghostmansd, eslf.insnlog is per-instruction, and i happen to know that execute_one() is the main "execute one instruction" entry-point20:42
lkcltherefore, at the *start* of execute_one() the (copy of) the disassembled instruction is placed as the first (and only) entry, overwriting whatever was in it previously20:44
lkclat various points within the "yield from" additional "stuff" will be appended20:45
lkcl(registers read, registers written)20:45
lkcland towards the end of execute_one you write out the entire contents of self.insnlog to the log file20:45
lkcl(only if self.insnlog_file has been initialised, obviously).20:45
lkclopen /tmp/inslog.txt is obviously intended as a proof-of-concept hack20:46
lkclto be entirely replaced, but i wanted you to at least know *where* the replacement code should go20:47
lkclas it's not immediately obvious20:47
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC23:32
*** ghostmansd[m] <ghostmansd[m]!> has joined #libre-soc23:32

Generated by 2.17.1 by Marius Gedminas - find it at!