Monday, 2022-08-01

*** mx08 <mx08!~mx08@user/mx08> has quit IRC00:55
*** mx08 <mx08!~mx08@user/mx08> has joined #libre-soc00:56
*** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has quit IRC03:48
*** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has joined #libre-soc03:48
*** yambo <yambo!~yambo@> has joined #libre-soc05:48
*** ghostmansd[m] <ghostmansd[m]!> has quit IRC07:17
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc07:18
*** toshywoshy <toshywoshy!> has quit IRC07:38
*** toshywoshy <toshywoshy!> has joined #libre-soc07:38
*** markos <markos!> has quit IRC07:50
*** philpax_ <philpax_!> has quit IRC07:50
*** philpax_ <philpax_!> has joined #libre-soc07:51
*** markos <markos!> has joined #libre-soc07:51
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC09:29
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc09:53
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC10:00
*** octavius <octavius!> has joined #libre-soc10:03
*** ghostmansd[m] <ghostmansd[m]!> has joined #libre-soc10:09
toshywoshyFYI, NLNet proposals have been submitted, I will update the wiki later to reflect the same information as submitted11:10
*** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has quit IRC11:54
*** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has joined #libre-soc11:55
lkcltoshywoshy, magic thank you12:15
lkcltoshywoshy, btw can you send a link to any public information to FundingBox, soon? so they can put it into their blahblah "this is a real person make the EU happy" documentation?12:33
octaviuslkcl, for the NGI Pointer team foundation statement, NLnet doesn't count as an EU grant, does it?12:48
octaviusreason I ask is point 2.b) of your example statement12:50
*** octavius <octavius!> has quit IRC13:29
*** octavius <octavius!> has joined #libre-soc13:30
*** openpowerbot <openpowerbot!> has quit IRC13:33
lkcloctavius, i have no idea what 2.b) is13:40
lkcljust send them a link to a CV and give them some reassurance that you exist.13:40
lkclit's the EU-audit equivalent of "are you a robot"13:41
lkclhello i exist (i am not a fake identity trying to scam the EU out of grant money) i have been working on Libre-SOC for NN months, here's some links online where you can confirm i do in fact exist13:42
lkcli.e. you're not a fake person on a fake project trying to fleece the EU. like the fake farms which took so much EU Grant Money for such a long time that no Chartered Accountancy Firm will sign off on the EU's accounts.13:43
*** openpowerbot <openpowerbot!> has joined #libre-soc13:47
*** openpowerbot <openpowerbot!> has quit IRC14:03
*** openpowerbot <openpowerbot!> has joined #libre-soc14:04
octaviuslkcl, sent them an email with the form and linkedin14:21
octaviusalso I copied out the text of that pdf doc you sent me into .odt format. If it will be useful in the future, I can upload it as a template14:22
*** octavius <octavius!> has quit IRC14:44
ghostmansdlkcl, whay should we do with entries like `0b1000000000,,,,,,,,,,,,,,,,,,,,,,,mcrxr,X,`?14:45
ghostmansdI'm trying to wrap get_svp64_csv et al. so that we have some nice pure-Python abstractions atop, so that we could re-use this stuff in sv_binutils and power_decoder14:46
ghostmansdHeck I cannot even recall how it works in sv_binutils now. I should've hit this, since sv_binutils also use get_svp64_csv.14:53
ghostmansdAh yeah. We don't grab "unit" field in sv_binutils, that's why. And other fields happen to be set by get_svp64_csv magic. Sigh.14:57
ghostmansdlkcl, it would be great if you could guide me through PowerDecoder's opcode calculations but w/o all these Signal/eq/does_something stuff. :-) The tree_analyze is a difficult beast, I'm still struggling to undestand how it works.15:09
ghostmansdLet's assume that for any given instruction, I know the following: 1) the major opcode pattern (Subdecoder.pattern); 2) whether this pattern is with dashes (Subdecoder.opint); 3) which Subdecoder.bitsel to use; 4) which Subdecoder.suffix to use.15:11
ghostmansdThat is, assuming I have an instruction entry from certain CSV via get_svp64_csv, and I know the Subdecoder in advance... do we have a simpler way to calculate the opcode without diving into Signal and its friends?15:14
ghostmansdThe only simple and obvious thing is extra.csv: it uses the whole range (all 32 bits) and there '-' symbols are simply translated as zeroes, nothing special. I don't know what on Earth OP_SIM_CONFIG is, though: couldn't find anything in binutils.15:28
*** ghostmansd[m] <ghostmansd[m]!> has quit IRC16:10
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc16:11
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC17:08
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc17:11
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC17:29
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc17:30
*** octavius <octavius!> has joined #libre-soc17:34
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC17:34
*** ghostmansd[m] <ghostmansd[m]!> has joined #libre-soc17:35
*** ghostmansd[m] <ghostmansd[m]!> has quit IRC18:30
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc18:31
lkclghostmansd, treat it as a binary number, masked by bitsel.18:42
lkclyou shift it up by the start-point of bitsel18:43
lkcland use the *full* range of start-end of bitsel as the mask18:43
lkclyou would create a mask of (1<<(5-0)-1)<<018:44
lkclyou create a mask of18:44
lkcland that value 0b1000000000 you simply shift up by 2018:45
lkclyes, i described how to split out the entirety of create_pdecode() into a completely separate module18:45
lkclignore extra.csv. its contents cannot be SVP64-prefixed anyway18:47
*** choozy <choozy!> has joined #libre-soc18:52
lkclif you encounter a mask-string, that's a leeetle more complex but not by much18:52
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC19:35
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc19:35
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC20:28
*** ghostmansd[m] <ghostmansd[m]!> has joined #libre-soc20:29
*** tplaten <tplaten!> has joined #libre-soc20:34
*** tplaten <tplaten!> has quit IRC21:07
*** choozy <choozy!> has quit IRC22:06
*** lkcl- <lkcl-!> has joined #libre-soc22:41
*** tinybronca[m] <tinybronca[m]!~tinybronc@2001:470:69fc:105::2:1af6> has quit IRC22:48
*** lkcl <lkcl!> has quit IRC22:48
*** tinybronca[m] <tinybronca[m]!~tinybronc@2001:470:69fc:105::2:1af6> has joined #libre-soc23:03
programmerjakeintrinsics handling idea: lazily generate them, so even if there's 10M, we only pay for what we use which is very likely to be much smaller. patch for clang risc-v v which does that:

Generated by 2.17.1 by Marius Gedminas - find it at!