Friday, 2022-08-12

*** tplaten <tplaten!> has quit IRC03:54
*** tplaten <tplaten!> has joined #libre-soc04:08
ghostmansdlkcl, we don't support VLE and SPE2, do we?04:34
ghostmansdNah, we don't.04:36
ghostmansd  { "libresoc",(PPC_OPCODE_PPC | PPC_OPCODE_ISEL | PPC_OPCODE_6404:36
ghostmansd        | PPC_OPCODE_ALTIVEC | PPC_OPCODE_VSX | PPC_OPCODE_SVP64),04:36
programmerjakevle and spe2 aren't in the powerisa spec. v3.1b, though iirc we had plans for something similar to vle before -- 16-bit instructions -- I came up with a scheme to make them work in little-endian mode04:40
programmerjakealso, why does libresoc include altivec and svx in the code you just posted, seems incorrect04:41
lkclnor POWER4-POWER8 either04:45
lkclor G4.04:45
lkclISEL yes.04:46
lkclif POC_OPCODE_POWER9 is what's behind "-mpower9" and that means the 214 instructions of the SFFS Compliancy Level, then that's what is implemented04:48
lkclwhich is Power ISA 3.004:48
lkclif POWER8 means "Power ISA 2.08" which includes some instructions *not present* in Power ISA 3.0 then we definitely don't support those04:49
lkcland can't04:49
lkclthe Power ISA EULA is only available for v3.0 and v3.104:49
lkclas in: we cannot release a commercial ASIC product that is compliant with only v2.08 or v2.0704:50
lkclbottom line is, i suspect those flags will have to be reduced to:04:50
lkclPPC_OPCODE_ISEL (if that is for the isel instruction)04:50
lkcland that's probably it04:51
lkclstrictly speaking binutils should have a PPC_OPCODE_SFFS_COMPLIANCY_LEVEL04:53
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc06:32
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC06:46
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc06:47
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC06:51
*** lx0 <lx0!~lxo@gateway/tor-sasl/lxo> has quit IRC07:04
*** lx0 <lx0!~lxo@gateway/tor-sasl/lxo> has joined #libre-soc07:06
programmerjakeghostmansd: your latest commit broke:
ghostmansdseems strange, since I don't recall committing anything to master07:54
ghostmansdI recently committed to binutils branch...07:55
programmerjakeit's on the binutils branch07:55
programmerjakegitlab tests commits to any branch07:55
ghostmansdwhy for God's sake does it run binutils branch?07:55
ghostmansdplease drop it, it makes no sense to test it, at all07:55
programmerjakebecause it's there and useful to test07:55
ghostmansdit's not, it's WiP07:55
ghostmansdstill, a good catch07:56
programmerjakeimho it should be tested, if you think otherwise, feel free to ignore it. it's more work to disable it anyway...07:57
ghostmansdI thought for a whilem let's keep it07:58
ghostmansdis master also broken?07:58
programmerjakeyes, but only cuz some of the tests try to use soc when they shouldn't:
ghostmansdI see that the master fails with `E   ModuleNotFoundError: No module named 'soc'`08:00
ghostmansdOK, got it08:00
programmerjakeopenpower-isa has no soc dependency listed in, so therefore we should test it without soc08:00
programmerjakealso cuz it would be a circular dependency if soc was listed08:01
*** jn <jn!~quassel@user/jn/x-3390946> has quit IRC08:56
*** jn <jn!~quassel@2a02:908:1065:960:20d:b9ff:fe49:15fc> has joined #libre-soc09:06
*** jn <jn!~quassel@user/jn/x-3390946> has joined #libre-soc09:06
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc09:51
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC09:54
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc09:56
lkclsigh yes, we had a new contributor, lack of experience they introduced a circular dependency, by moving some code from soc to openpower-isa which still depended on imports from soc10:57
lkcli didn't catch it at the time and haven't had time to fix it.10:57
lkclghostmansd, opcode name/mask looks brilliant10:59
lkclkinda fascinating/exciting to see that evolve11:05
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC11:05
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc11:05
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC11:12
*** ghostmansd[m] <ghostmansd[m]!> has joined #libre-soc11:13
*** ghostmansd <ghostmansd!> has quit IRC12:15
*** ghostmansd[m] <ghostmansd[m]!> has quit IRC12:29
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc12:29
*** ghostmansd <ghostmansd!~ghostmans@> has joined #libre-soc12:56
*** ghostmansd <ghostmansd!~ghostmans@> has quit IRC13:10
*** ghostmansd <ghostmansd!~ghostmans@> has joined #libre-soc13:11
ghostmansd> <programmerjake> ghostmansd: your latest commit broke:
ghostmansdThis happens if we glob("*.csv") instead of iterating over the fixed set set of names, or, well, if we don't check the names.13:37
ghostmansdIdeally this piece of code should use new Database, but let's move with the smaller steps. I'll add an intermediate commit which skips insndb.csv file.13:38
ghostmansdprogrammerjake, I've signed up for salsa repository and I'm going to join the project there to launch the pipelines; could you approve the account, please?13:47
ghostmansdAh, well, it's triggered automatically. OK, perhaps I won't need this.13:48
ghostmansdFuck there are two *.csv globs. In one file. With the same filter. Copied and pasted.13:52
*** ghostmansd <ghostmansd!~ghostmans@> has quit IRC14:00
lkclghostmansd[m], niiice. that's clearly me being dumb14:09
lkclbtw i'm currently going through spec and openpower-isa removing LD-ST-with-shift14:09
lkclafter Paul spotted it as an anomaly14:09
*** ghostmansd <ghostmansd!~ghostmans@> has joined #libre-soc14:20
ghostmansd[m]I'll decouple it14:21
ghostmansd[m]Already in binutils branch14:21
*** ghostmansd <ghostmansd!~ghostmans@> has quit IRC14:39
*** ghostmansd <ghostmansd!~ghostmans@> has joined #libre-soc14:58
ghostmansdWould the testing in CI be sufficient to assume the code works?15:23
*** ghostmansd <ghostmansd!~ghostmans@> has quit IRC15:37
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC16:05
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc18:03
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC18:39
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc18:39
*** Guest16 <Guest16!~Guest16@2402:d000:8114:342c:fd7d:f033:dccb:b516> has joined #libre-soc19:48
Guest16Hi everyone, is this project active?19:48
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC19:52
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc20:12
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC20:16
*** tplaten <tplaten!> has quit IRC20:20
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc20:25
programmerjake<ghostmansd> "Would the testing in CI be..." <- only if .gitlab-ci.yml is modified to ensure all your code is run and properly tested (e.g. building binutils), note that ccache is used to avoid huge compile times20:50
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC20:52
programmerjakeif you want email notifications when your commits fail ci, I can add you to the system I built for that...gitlab itself will only ever notify me since I own the ssh key used to push to the gitlab mirrors20:53
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc21:13
programmerjakeghostmansd: if you want email notifications ^ -- in case you missed it due to ping timeout21:42
lkclGuest16, yes it is.

Generated by 2.17.1 by Marius Gedminas - find it at!