* psydroid
would like to interject that Latin is taught in schools in certain (European) countries today and Sanskrit is also taught in certain (Indian) schools today
<geertu>
mps: Teaching Latin is still active (but optional) in high school here, too.
<mps>
geertu: I think is also active (and obligatory) in high school also here but I didn't checked so can't say for sure
Reinhilde is now known as MelMalik
jmdaemon has quit [Ping timeout: 250 seconds]
conordooley49 has joined #riscv
conordooley49 is now known as conchuod
conchuod has quit [Client Quit]
conordooley55 has joined #riscv
<mmind00>
palmer: out of curiosity, did you get a chance to try the updated svpbmt series? Especially on those special boards - with alternatives now behind a kconfig symbol again, I do hope they should really be unaffected now
cousteau has joined #riscv
<cousteau>
Sup
<cousteau>
Out of curiosity, are there any standard instructions (ratified, frozen, or draft) that aren't 32 or 16 bits? E.g. a 64-bit standard instruction (the ones with opcode 0111111)?
<cousteau>
I expected V extension instructions to be long (64 bits), but they seem to be 32
handsome_feng has joined #riscv
ivii has joined #riscv
bauruine has quit [Remote host closed the connection]
eroux has joined #riscv
jacklsw has quit [Quit: Back to the real world]
<la_mettrie>
courmisch: you said on 14th day that CEC is "entirely" missing from VisionFive tree. This thread seems to say it's just a matter of "enabling" (does it mean setting it from kernel config when building?) https://forum.rvspace.org/t/hdmi-cec-support/227/3
bauruine has joined #riscv
littlebobeep has quit [Ping timeout: 240 seconds]
eroux has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Noisytoot has joined #riscv
Noisytoot has quit [Read error: Connection reset by peer]
eroux has joined #riscv
Noisytoot has joined #riscv
handsome_feng has quit [Quit: Connection closed for inactivity]
drmpeg has quit [Ping timeout: 250 seconds]
drmpeg has joined #riscv
eroux has quit [Ping timeout: 256 seconds]
aerkiaga has joined #riscv
<la_mettrie>
i wonder why polarfire is so dominating board on the riscv kernel list
<conchuod>
what does that mean la_mettrie?
<conchuod>
"Why are a large number of the emails on the mailing list related to polarfire"?
<la_mettrie>
approximately
<conchuod>
The answer to that, at least right now, is that the clock driver has a (silly) omission which needs to be fixed so there's a bunch of patches for that.
<conchuod>
And I guess there just aren't too many other boards with production silicon (although i use that term quite loosely for mpfs) & we are getting to the point that things are becoming ready to actually upstream.
<conchuod>
Currently cleaning up the other drivers we have so they can be sent upstream, so you'll have to endure more emails from me soon!
<conchuod>
Does that answer your question la_mettrie ?
<la_mettrie>
yes
BOKALDO has quit [Quit: Leaving]
<drmpeg>
Can you still buy a PolarFire?
<drmpeg>
Crowd Supply says Jan 31, 2023
<conchuod>
I honestly have no idea what the story is with that side of things.
<drmpeg>
Ok.
<conchuod>
Mouser says the same about delivery for icicle kits, obviously my acquisition process is slightly different so I'd have to ask.
<conchuod>
"slightly"
<conchuod>
I think the PolarBerry devices were meant to ship around now, wonder if they arrived.
Andre_H has joined #riscv
<drmpeg>
I know someone that has one. So they did ship.
<conchuod>
I need to get an updated dt sorted for that board.
Andre_H has quit [Client Quit]
<drmpeg>
I remember now. That was the one where Crowd Supply bought all the boards to fund the project.
<conchuod>
I didn't even know it was on crowd supply until just now
avoidr has quit [Quit: leaving]
<drmpeg>
Yeah, my friend was one of the 8 backers, and we thought it wasn't going to happen. Then right at the end, Crowd Supply pushed it over the goal.
<conchuod>
What is your friend running on it?
<drmpeg>
Nothing. It's probably still in the box.
<drmpeg>
She runs a non-profit research organization and just bought it in case someone was interested.
<conchuod>
I'd be curious as to what it actually shipped with. The (current) vendor tree for mpfs is based on 5.12 (although that should change shortly) I'd expect PolarBerry to predate even that.
zkrx has quit [Quit: zkrx]
conchuod has quit [Quit: Client closed]
prabhakarlad has quit [Quit: Client closed]
BOKALDO has joined #riscv
<palmer>
mmind00: sorry, I'm still tied up with other stuff -- looks like I dropped the ball on Alex's patches, then I have to go Guo's, then yours ;)
mps has quit [Ping timeout: 250 seconds]
jmdaemon has joined #riscv
mps has joined #riscv
hexteex has joined #riscv
<courmisch>
la_mettrie: I don't know.
wingsorc has joined #riscv
avoidr has joined #riscv
seninha has joined #riscv
jacklsw has joined #riscv
<conordooley>
palmer: It's pretty likely that I am going to go from one to four device trees in the coming weeks, at what point does it become easier for me to collect the changes and send you a PR rather than you tracking it?
<mmind00>
palmer: no worries, but good to know it's on the radar :-)
<palmer>
conordooley: it really doesn't make that much of a different to me if you're sending patches or a PR, I moved over to this "merge my own branches" flow because in theory it was more like PRs -- we talked about it last year at Plumbers, but nobody's sent a PR yet
<palmer>
so if you're all setup with kernel.org and GPG and such then you're more than welcome to send PRs
<palmer>
(also, I'm about to respond to your v3 ;))
avoidr has quit [Quit: leaving]
vagrantc has joined #riscv
<palmer>
mmind00: yep, I didn't forget about it
<palmer>
you can run the riscv-systems-ci tests yourself if you want, they're super ugly but everything's sort of in there
zjason``` is now known as zjason
<palmer>
if not no big deal, I'm going to run them either way ;)
<palmer>
(also, I'm supposed to set up some CI this quarter to make this saner but who knows)
<mmind00>
palmer: sounds interesting ... where do I need to look for those?
<palmer>
github/palmer-dabbelt/riscv-systems-ci
<palmer>
it's a mess because it's just sort of my dumping ground to move my working directory between machines
<palmer>
I keep meaning to re-write it to be sane, but I never have time ;)
<palmer>
(I think sparse might be on in there now, but still broken on master so you might need to juggle stuff a bit)
<mmind00>
palmer: hehe, I know that type of dumping-ground well ... my build-scripts are similar (for dumping into a kernel/u-boot/etc tree and have it do something for my boardfarm automatically)
<palmer>
Yep, just how it goes
<palmer>
It always works enough to get stuff done, I guess I just need to break it or something
EchelonX has joined #riscv
<geertu>
palmer: Shouldn't you merge your own patches using b4, so you'll get nice Link: tags added (cfr. Documentation/maintainer/configure-git.rst)?
avoidr has joined #riscv
<palmer>
probably, but I haven't gotten b4 working right yet
<palmer>
I guess maybe that's a project for today, though, as everyone else says it's great ;)
<palmer>
(I'm unfortunately stuck in the stone ages when it comes to email)
<courmisch>
Esmil: you didn't look at the Tensilica VPU device driver, per chance?
<courmisch>
Esmil: looks to me that it's poking the device registers, programming DMA and handling interrupts from userspace :O
<conordooley>
heh, been spending some time making our CI actually do something more than bootrr, should try and run your riscv-ci as part of it too.
<conordooley>
I do not have kernel.org acc palmer, figured it'd be easier to wait til a bunch of people show up down the street in september to get the GPG side of that done
<palmer>
conordooley: OK, so if you don't have all the kernel.org stuff set up yet then it's probably more work to make PRs than patches
<palmer>
there's a map of folks who will meet to sign keys and such, if you want to do it faster
<conordooley>
Ye, there's a couple of the at91 microchip guys that could do it for me.
<conordooley>
Depends on how soon I end up actually sending the other DTs p much.
<conordooley>
I assume that since you have an MC proposal, that you're coming here palmer?
<palmer>
"here" is Dublin/Plumbers?
<conordooley>
Yes, sorry
<palmer>
ya, that's the plan -- aside from any COVID craziness I should be there, and that's impossible to predict
avoidr has quit [Quit: leaving]
avoidr has joined #riscv
<conordooley>
on the subject of DTs actually, I have a bit of an issue - although maybe its more of a kzrk/rob thing. I have a -fabric.dtsi which represents the contents of the reference FPGA design.
<conordooley>
There's a planned change to the addresses of the blocks in there, which obviously breaks the DT's backwards compatability etc
avoidr has quit [Client Quit]
<conordooley>
Not sure if I am "allowed" to change that, or if it being the FPGA fabric makes it an exception to the rule since no "real" user will ever use that fabric.dtsi
<palmer>
TBH, I'm not really sure -- the rules kind of don't work for FPGAs (same as that thread about the DTs for that FPGA-based IP core eval kit)
<palmer>
I guess you could call it a different version of the board or something?