Tuesday, 2023-08-01

*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC00:57
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-90-154-80-239.ip.moscow.rt.ru> has joined #libre-soc00:58
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-90-154-80-239.ip.moscow.rt.ru> has quit IRC04:57
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc04:57
*** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has quit IRC07:42
*** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has joined #libre-soc07:43
*** yambo <yambo!~yambo@098-127-130-021.biz.spectrum.com> has quit IRC08:21
*** yambo <yambo!~yambo@098-127-130-021.biz.spectrum.com> has joined #libre-soc08:34
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC08:56
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-90-154-80-239.ip.moscow.rt.ru> has joined #libre-soc08:57
*** programmerjake <programmerjake!~programme@2001:470:69fc:105::172f> has joined #libre-soc09:14
*** psydroid1 <psydroid1!~psydroid@user/psydroid> has joined #libre-soc09:46
*** Ryuno-KiAndrJaen <Ryuno-KiAndrJaen!~ryuno-kim@2001:470:69fc:105::14ed> has joined #libre-soc09:46
*** sadoon[m] <sadoon[m]!~sadoonalb@2001:470:69fc:105::3:5f7c> has joined #libre-soc09:46
*** nyanbinary[m] <nyanbinary[m]!~nikocondu@2001:470:69fc:105::3:871b> has joined #libre-soc09:46
*** vaino[m] <vaino[m]!~vainomatr@2001:470:69fc:105::3:5474> has joined #libre-soc09:46
*** octavius <octavius!~octavius@> has joined #libre-soc10:52
octaviuslkcl, red meeting10:52
lkcloctavius, thx11:01
lkclprogrammerjake, fantastic on the split_insns, will run and check later11:01
*** mx08_ <mx08_!~mx08@user/mx08> has quit IRC11:09
*** mx08 <mx08!~mx08@user/mx08> has joined #libre-soc11:10
octaviuslkcl, bug on documenting issues has been raised11:54
*** psydroid1 <psydroid1!~psydroid@user/psydroid> has quit IRC12:41
*** psydroid1 <psydroid1!~psydroid@user/psydroid> has joined #libre-soc12:43
octaviusApologies to all for spamming the chat, and not properly raising a bug report. My work on generating the external core+svp64 is in this bug https://bugs.libre-soc.org/show_bug.cgi?id=112712:48
octaviusAlso, for future reference, a guide to how to raise issues is in this bug: https://bugs.libre-soc.org/show_bug.cgi?id=112612:48
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-90-154-80-239.ip.moscow.rt.ru> has quit IRC13:11
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc13:12
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC13:22
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-90-154-80-239.ip.moscow.rt.ru> has joined #libre-soc13:23
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-90-154-80-239.ip.moscow.rt.ru> has quit IRC13:57
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-90-154-80-239.ip.moscow.rt.ru> has joined #libre-soc13:58
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-90-154-80-239.ip.moscow.rt.ru> has quit IRC14:32
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc14:32
lkclprogrammerjake, preliminary unit test runs fine (test_caller_svp64_ldst.py) i'm running the rest (all the test_caller_*)14:33
markos_programmerjake, there seems to be an extra . in av/minmax: av/minmax..mdwn14:48
lkclmarkos_, it's because the instruction *is* named "minmax."15:03
lkclprogrammerjake, the *entire* instruction needs to be split not the pseudocode.  split-insn.sh can therefore simply output from "#".15:03
lkclalthough it is much easier to use the parser.py that already exists which i suggested be used rather than rewrite from scratch an entirely new shell script15:04
markos_ah, right, I hope this doesn't confused filename parsers15:04
programmerjakethat's intentional since the mnemonic contains a . and that's how ikiwiki likes . in the page name15:07
markos_nm then15:09
programmerjakei explained why i split out the pseudocode -- see the latest bug comment15:14
programmerjakebasically, rfcs have a bunch of extra sections in there that the parser doesn't like, and other places we quote just the pseudocode, so just the pseudocode needs to be in a file of its own15:16
programmerjakeif you want the rest split out too i can do that today15:16
programmerjakebut it'll be to separate files from the pseudocode that still have [[!inline]] in them15:17
programmerjake> although it is much easier to use the parser.py that already exists which i suggested be used rather than rewrite from scratch an entirely new shell script15:19
programmerjakenot really since that discards lines as it goes, the shell script keeps them all which makes copying to the output much easier15:19
programmerjakei wasn't expecting the script to be this complex when i started...15:19
programmerjakea large part of the complexity is extracting the insn name so we know where to put the output file15:28
programmerjakeso, just looking for # is too simplistic15:29
lkclhmm it _shouldn't_ have discarded the lines... ahh i know: addw added the "comments" capability. drat15:56
lkclmmm that's annoying because it's effectively resulted in duplicating the entirety of the parser... in shell-script15:57
lkclwhich is unreadable.15:57
lkclshell-scripts always are15:57
lkclah well, if it "does the job" that's fine, it's a one-off15:57
*** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has quit IRC16:11
*** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has joined #libre-soc16:11
programmerjakewell, it's a one-time use script, so if it works we can delete it afterward16:22
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC16:22
programmerjakethough imho it's pretty readable16:22
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc16:22
programmerjakechanging the parser to accommodate rfc insn syntax is a much more complex task16:24
programmerjakeplus, isn't the spec in latex? so we need to support that...16:25
*** octavius_ <octavius_!~octavius@> has joined #libre-soc17:25
*** yambo <yambo!~yambo@098-127-130-021.biz.spectrum.com> has quit IRC17:25
*** octavius <octavius!~octavius@> has quit IRC17:28
*** yambo <yambo!~yambo@098-127-130-021.biz.spectrum.com> has joined #libre-soc17:38
sadoon[m]<lkcl> "sadoon, you can still log in..." <- Yes but I need to change my email address so that all emails also reach the new one + I'm letting go of the old domain in a few years and don't want it to be maliciously used :)17:54
sadoon[m]Ah, found a bugzilla way to change my email and add my real name, on it17:58
sadoon[m]Ok done17:59
sadoon[m]My email is sadoon (at) albader (dot) co btw17:59
*** octavius_ <octavius_!~octavius@> has quit IRC18:01
*** octavius <octavius!~octavius@> has joined #libre-soc18:01
programmerjakenice! just in case you forget: you need to also change it for the mailing lists18:18
sadoon[m]Do I just unsub the old one and sub the new one?18:19
programmerjakethat should work18:20
sadoon[m]done :)18:22
sadoon[m]Before attempting gdb I'll try building a kernel for POWER7 (using the menuconfig option) and trying to boot it with both power9 and power718:22
sadoon[m]Suspecting something fishy happening when a kernel is built for power7 on little endian18:23
sadoon[m]Also I've been building a 64k pagesize kernel because I did not change the default, afaik POWER7 doesn't use that but supports it18:25
sadoon[m]Because even the PowerMac G5 does afaik18:25
sadoon[m]It emulates iot18:25
sadoon[m]power7 kernel does not fail under power9 confirmed18:52
sadoon[m]the issue was disabling VSX and Altivec18:52
sadoon[m]still fails on power718:52
sadoon[m]in other words the kernel will not boot if I disable VSX and altivec, time to try one at a time18:52
sadoon[m]no vsx, power9 fails19:13
sadoon[m]power7 also fails, back to square one19:15
sadoon[m]I guess I can conclude now with some confidence that what Luke said re: the kernel emulating missing instructions to be not working19:16
sadoon[m]Because booting power7 with a default power9 kernel (4k pages) does not work19:16
sadoon[m]However I'm not 100% sure of that because VSX might not be emulated in that process19:17
markos_guys, I will be about 30 min late for tonight's meeting, fyi19:32
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC20:02
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-90-154-80-239.ip.moscow.rt.ru> has joined #libre-soc20:03
*** octavius <octavius!~octavius@> has quit IRC22:33
lkclsadoon[m], ah you managed to change it on bugzilla too, awesome23:23
lkclsadoon[m], yes, the ppc emulation present in the linux kernel is apparently there for gdb kernel single-stepping (?!?!) not for *actual* emulation sigh23:25

Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!