jellydonut has quit [Read error: Connection reset by peer]
<OGBoobyBlack>
Github is the worst thing that ever happened to git
<jrtc27>
it's the best thing that happened to git
<jrtc27>
it made it easily available to the masses for free
jjido has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
jacklsw has joined #riscv
radu242 has joined #riscv
<OGBoobyBlack>
no, now idiots just think git is github
elastic_dog is now known as Guest1530
elastic_dog has joined #riscv
rurtty has quit [Ping timeout: 268 seconds]
zjason has joined #riscv
Trifton has joined #riscv
<muurkha>
so do many people who are smarter than you, OGBoobyBlack :)
<muurkha>
they just haven't learned the difference yet — nobody is born knowing it
raym has quit [Ping timeout: 248 seconds]
<muurkha>
I have my reservations about whether we really want Microsoft to own the namespace of most free software projects, but I also think GitHub has provided a lot of very useful service
vagrantc has quit [Quit: leaving]
raym has joined #riscv
<somlo>
the pareto curve of freedom vs. convenience :)
davidlt has joined #riscv
<OGBoobyBlack>
git is great
<OGBoobyBlack>
github is wack
KombuchaKip has joined #riscv
cwebber has quit [Read error: Connection reset by peer]
<conchuod>
Maz had a suggestion for how to rewrite things to support both of the interrupt models that we haven't had the chance to do yet.
<conchuod>
I am a bit sick which is my excuse for not remembering: but iirc we use ~that v6 in our vendor tree because without a chained flow there were missed interrupts w/ CONFIG_PM=y
<conchuod>
and thanks Geert
eroux has joined #riscv
jmdaemon has quit [Ping timeout: 268 seconds]
raym has quit [Ping timeout: 264 seconds]
raym has joined #riscv
<Esmil>
conchuod: i don't want to be the LED guy, but this works for me :) http://sprunge.us/Xbkdtu
<Esmil>
(on the 22.10 prerelease)
jjido has joined #riscv
<geertu>
Esmil: meh!
<conchuod>
Esmil: I was actually thinking about it the other day when I looked the one for unleashed. The design I have at home has the LEDs hooked up to PWMs so couldn't test and then just forgot...
<conchuod>
By all means be the LED guy :)
aerkiaga has joined #riscv
<conchuod>
s/design I have at home/FPGA design I have on my Icicle at home/
<Esmil>
cool. it's actually quite nice when you're running from the sd-card to have a LED light up on activity
<Esmil>
kind of like when your hdd used to say that sound
jjido has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
xypron has quit [Ping timeout: 240 seconds]
xypron has joined #riscv
indy_ has quit [Ping timeout: 264 seconds]
zjason` has joined #riscv
<conchuod>
I like the heartbeat LEDs
<conchuod>
Esmil: you get a visionfive2 yet?
<Esmil>
conchuod: yes, i have an early version
zjason has quit [Ping timeout: 268 seconds]
<conchuod>
I bought one, hopefully it'll not get bogged down..
<conchuod>
Did they actually release docs for it this time? I asked Hal if there were any but they've been a bit spotty with their replies to that thread.s
<Esmil>
i have a 7110 TRM preview and they were very happy to receive the errors i found so at least they have someone working on the docs
<Esmil>
don't know when they'll be released
<conchuod>
If/when Hal sends v2 I'll ask him for it again.
<geertu>
conchuod: Heartbeat rulez
<geertu>
Esmil: BTW, the only reason I didn't hook up the LEDs was the lack of GPIO support
<Esmil>
geertu: ah, yeah. i can wait until that's added if that's better
<Esmil>
i just wanted to see it all work :)
<geertu>
Esmil: I've applied GPIo v6 and your patch. and will give it a try next week
* geertu
is an LED guy, too
<Esmil>
:))
<Esmil>
geertu: maybe use this version instead. the pwm header isn't needed: http://sprunge.us/5Rhawa
jjido has joined #riscv
<geertu>
Esmil: thanks, replaced
zjason` is now known as zjason
cwebber has joined #riscv
bauruine has quit [Remote host closed the connection]
zjason has quit [Remote host closed the connection]
zjason has joined #riscv
Andre_H has quit [Quit: Leaving.]
jacklsw has joined #riscv
Yaku has quit [Ping timeout: 248 seconds]
jjido has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
<prabhakarlad>
has anyone tried to run an executable which is compiled -fno-pie -static flags?
<palmer>
prabhakarlad: I do somewhat regularly, but always in QEMU
<palmer>
prabhakarlad: you're sure you're not getting hit by that __clear_cache() bug I forgot to fix?
<prabhakarlad>
palmer: to rule out any cache related issue I did run the tests by disabling both L1 and L2 caches and was still able to see the issue.
<palmer>
and that disable the icache too?
<prabhakarlad>
yep.
<palmer>
then sound like this isn't running into a fence.i related issue, as described in that mail
<palmer>
unless I'm missing something?
<palmer>
prabhakarlad: looks like you're building the binaries with clang? I don't think I've run a static/no-PIE clang-output binary
<prabhakarlad>
Experiments done by Ulrich (in the link) were done by caches enabled. That was later I did try out with caches disable and was still able to see an issue.
<palmer>
OK
<palmer>
can you put the exact binary that's failing somewhere public? Then we can remove any possible toolchain-related issues when we poke around inside it
<prabhakarlad>
sure, I'll just compile simple hello world application and share it.
<prabhakarlad>
btw ldconfig was built with the toolcahin provided by meta-riscv.
<palmer>
I just want to reduce variables, so the exact binary that fails for you gets rid of any toolchain-related differences
<palmer>
not usually all that fun to pass around binaries, but I think it this case it does actually simplify things
<palmer>
OK. I don't think I've seen one of those trying to run upstream stuff before, and last I looked they had a bunch of custom stuff that was needed to work sanely
Yaku has quit [Ping timeout: 268 seconds]
<prabhakarlad>
that was me trying to upstreaming ;)
<prabhakarlad>
basically the core doesn't have any issue, the IO-coherence port is disabled on the SoC so due to which IP blocks requiring DMA arent working.
<palmer>
ya, I remember you sending those bits. IIRC there's other userspace stuff too, like custom jump table optimizations and such
<palmer>
it was a while ago though, and I'm not sure what the new core looks like. Do you have a manual for it?
<prabhakarlad>
I havent come across any userspace optimizations.
vagrantc has joined #riscv
<palmer>
OK, I'd only seen microcontroller stuff. The extensions there would sort of weak havoc in Linux land, so hopefully they're not in these cores
<palmer>
conchuod: I'm just reading the GCC docs as I'm checking out a copy on the new laptop, but maybe "folks" is a GCC-ism? https://gcc.gnu.org/gitwrite.html
<muurkha>
a GCC-ism? I think the FSF has always tended towards using short, common words instead of longer and more recondite ones whenever possible
<palmer>
OK, so maybe an FSF-ism, then? Conor had pointed out I use "folks" a lot, I'm not sure where I picked it up from
<conchuod>
lol pointing out the awkward questions i ask
<conchuod>
palmer: speaking of gnu, is that tools thing invite only or can random folks show up?
<palmer>
what tools thing?
<conchuod>
idk you mentioned it in that sparse thread
<muurkha>
e.g. https://www.gnu.org/philosophy/shouldbefree.en.html doesn't say "folks" but it says " The only way to judge is to see who is helped and who is hurt by recognizing owners of software, why, and how much. In other words, we should perform a cost-benefit analysis on behalf of society as a whole, taking account of individual freedom as well as production of material goods."
<conchuod>
Or, it was mentioned in the issue you linked in that thread
<palmer>
conchuod: can you give me a pointer? if I said something more than like 5 minutes ago I can't remember it ;)
<muurkha>
it could have said, "The appropriate method for evaluation is to determine to whom fall the benefits and negative externalities induced by recognizing owners of software, ..."
<palmer>
it's just a Zoom meeting that happens sometimes. In theory 8am PST every other Thursday, but it's canceled somewhat frequently. There's usually a post to at least the GCC mailing list about it, if you can't find it I can try and dig it up (we've just got it in a Rivos calendar entry, but I don't think we're set up to share those outside)
<palmer>
NP. That link doesn't make Google find the posts for me, but they should be around somewhere public
<palmer>
I was thinking we should have some Linux one too, but I wanted to go to Carlos' glibc patchwork meeting to see what they're doing and of we can just copy it. It's Monday 6am PST, though, and I've yet to get up in time
rurtty has quit [Quit: Leaving]
jamtorus has joined #riscv
jacklsw has quit [Read error: Connection reset by peer]
jellydonut has quit [Ping timeout: 260 seconds]
<conchuod>
palmer: they have a weekly meeting about patchwork? or a weekly meeting to run through what is on their patchwork?
<palmer>
IIUC it's a weekly "get patches triaged" meeting, and since they use patchwork that essentially means sorting through the patches there. I'd bet that involves a lot of talk about stuff that's under review, as it's sometimes easier to do that real-time.
jobol has quit [Quit: Leaving]
<conchuod>
Ye, idk that sounds like a good idea?
<palmer>
ya, generally seems worthwhile to me
<palmer>
but I wanted to go and see an example of it working
<conchuod>
Prob something between 0900 and lunchtime Pacific is the best window?
<palmer>
actually I think earlier is better, for Asia
<palmer>
the global meetings are usually 7am-8am PST
<conchuod>
Ye I guess it has to be miserable time wise for someone
<palmer>
yep
<palmer>
so just medium-bad for everyone is as good as it gets
<conchuod>
But 0700 is better than 0200 or w/e
<palmer>
yep, 7am PST is 10pm Taiwan, so it's about as good as it gets
<palmer>
I can usually make 7am meetings, but 6am is pushing it
<conchuod>
idk, sounds like a good idea anyway.
<palmer>
yep, so my rough plan was to go to Carlos' meeting until I get a handle on how to run something like this, and then just blast linux-riscv with an invite ;)
DoubleJ has joined #riscv
wingsorc has quit [Quit: Leaving]
indy has quit [Ping timeout: 252 seconds]
indy has joined #riscv
justHaunted is now known as ffffffradiofree
ffffffradiofree is now known as jusFffffffradiof
nvmd has joined #riscv
prabhakarlad has quit [Quit: Client closed]
davidlt has quit [Ping timeout: 248 seconds]
wingsorc has joined #riscv
raym has quit [Remote host closed the connection]
raym has joined #riscv
___nick___ has joined #riscv
BootLayer has quit [Quit: Leaving]
Narrat has quit [Quit: They say a little knowledge is a dangerous thing, but it's not one half so bad as a lot of ignorance.]