<pabs3>
several other channels got hijacked too, #samba #osmocom #openmoko-cdevel #megaglest at minimum, but probably lots of them
<pabs3>
I guess any channel that mentions libera in any way
<palmer>
ya, so I guess if it's the freenode admins then there's not a lot we can do?
<palmer>
my Matrix gets all confused, and my irccloud forwards me to ##riscv
<pabs3>
right, except show up in the ##riscv channel and tell people who ask questions where the real channel is
<palmer>
OK
<pabs3>
yeah, the +f mode does that forwarding
<palmer>
OK, well, I'll idle in ##riscv in case anyone gets lost
<palmer>
aside from that I guess this is just functioning how the freenode folks want?
<xentrac>
pabs3: ugh
KREYREN has joined #riscv
KREYREN has quit [Changing host]
KREYREN has joined #riscv
<palmer>
that LWN thing kind of indicates they're wiping the slate clean on any channels that have a libera
<pabs3>
yep
<palmer>
OK, well, I guess I'll hang out in ##riscv and see if they end up banning me from the network or whatever
<palmer>
aside from that I'm not really sure of a better way to handle this sort of thing
<palmer>
if it was some regular IRC stuff then I could try to do op things to fix the channel, but I've probably lost them and even if I haven't it smells like that's not the goal
<palmer>
kind of a bummer, as IIUC a lot of people hadn't moved over yet
<pabs3>
yep, freenode-place-holder is the owner of both
<palmer>
OK
<palmer>
well, I guess there's nothing better we can do
<palmer>
maybe it's worth a post to sw-dev?
<xentrac>
well
<pabs3>
if anyone has an IRC client with a pingall command, that might help prod people to move?
<xentrac>
ensure that we don't rely on administratively centralized infrastructure in the future
<palmer>
IMO we've got way bigger problems than IRC with centralized administration ;)
<palmer>
one of my other freenode channels kicked me (and I assume everyone else) with a message about moving to libera
<palmer>
not sure if we can still do that, as I'm not even sure if anyone is in #riscv any more
<pabs3>
there are lots of folks in freenode #riscv, but no-one can send anything to channel or kick people as freenode staff took over
<palmer>
OK
<xentrac>
well, you can privmsg everybody one by one
<pabs3>
good idea
<xentrac>
maybe divide it up among different people
<palmer>
presumably this is a task for as cript?
<palmer>
does anyone know of a scheme for doing that sort of thing?
<xentrac>
and try to exclude the "freenode" staff
KREYREN has quit [Quit: Leaving]
<palmer>
OK, well, I can't even figure out how to get the list of users in there
<palmer>
so I give up
<xentrac>
/who *
<xentrac>
/names *
<palmer>
both of those say "no response" on irccloud
<palmer>
is that just an error over there?
<xentrac>
dunno, and I'm sure not gonna log back in to freenode to check
<xentrac>
I mean I would if I felt responsible for the channel but so far I'm just a random user
<palmer>
makes sense
<palmer>
hexchat doesn't give any response either
<palmer>
so I'm guessing they've somehow made that stop working, presumably to avoid this sort of tihng
<palmer>
if someone has the list I'm happy to privmsg everyone, I guess just email it to me
<palmer>
otherwise ##riscv is as good as it'll get -- IIUC everyone'll eventually end up there after a reconnect
<pabs3>
I just got a list, will privmsg you
<palmer>
anyone paying attention can probably figure out how to get here anyway
<palmer>
ah, thanks
<pabs3>
a surprisingly large amount of users
<palmer>
about as many as we have here, though
<xentrac>
they probably can't go full Tiananmen Square Massacre on people though
<pabs3>
here is about 30 less
<xentrac>
I mean if people get /killed every time they say "libera.chat" then other people will find the ambiance hostile
<pabs3>
allegedly one of the staff "wants to build a pan-irc surveillance apparatus" so that does seem like it could happen
<xentrac>
right, but what I mean is, presumably the objective of nuking the signposts is to keep people from deserting "freenode"
<xentrac>
Imperial Freenode
<xentrac>
but if they start blasting away at whoever says "we're now on perl -le '$_ = "mjcfsb.dibu"; y/a-z/za-y/; print'" then the people who see it happen are going to want to desert Imperial Freenode
<palmer>
I just got a global notice from freenode saying "[Global Notice] In the recent policy enforcement, some channels were erroneously included. We greatly apologize for the inconvenience. Please contact us in #freenode-services or contact-us@freenode.net. Thanks for your patience and choosing freenode!"
valentin has joined #riscv
valentin_ has quit [Read error: Connection reset by peer]
palmer___ has joined #riscv
palmer___ has quit [Client Quit]
riff-IRC has quit [Remote host closed the connection]
riff-IRC has joined #riscv
craigo has joined #riscv
valentin has quit [Read error: Connection reset by peer]
valentin_ has joined #riscv
valentin has joined #riscv
valentin_ has quit [Read error: Connection reset by peer]
<rm>
I'm just happy IRC still exists and even with these event's, it's a kind of continuity from 90s
<rm>
eventually the freenode story will be another paragraph by that link
<rm>
given it is/was one of the largest active networks
* rjek
nods
<rjek>
I think the best way to think about it is that Freenode changed its name to libera.chat
<xentrac>
yeah
valentin_ has quit [Read error: Connection reset by peer]
valentin has joined #riscv
rjek is now known as ewrx
ewrx is now known as rjek
SanchayanM has joined #riscv
KREYREN has joined #riscv
KREYREN has quit [Changing host]
KREYREN has joined #riscv
Sos has joined #riscv
SanchayanM has quit [Quit: SanchayanM]
SanchayanM has joined #riscv
wigyori has joined #riscv
KREYREN has quit [Quit: Leaving]
cousteau has joined #riscv
<cousteau>
Hi
jotweh has joined #riscv
<Sos>
hey
<Sos>
just got an update from the Nezha IGG
<Sos>
apparently the boards are ready to ship
<Sos>
woohoo
<cousteau>
PicoRV32 has a cpu_state register which is encoded as a one-hot register. Is there any reason for this, as far as the standard is concerned?
<cousteau>
(a one-hot register with 8 states = 8 bits)
<cousteau>
If the standard requires some way to access that register and says that it must be one hot then I can understand it, but otherwise I see no advantage of using a one hot over just storing it in a 3 bit reg
<cousteau>
Maybe I should ask in ##Verilog or similar...
ats_ has joined #riscv
<sorear>
if other people are definitely going to squat ##riscv then I have no remaining reason to be connected
ats has quit [Ping timeout: 252 seconds]
<pabs3>
several folks are
<cousteau>
Never mind; the Verilog was using some pragmas that might make one-hot efficient
aurel32 has quit [Quit: leaving]
<geist>
cousteau: interesting. i'd guess it's because some logic downstream of it that's using the states can more easily test against a one hot bit
<geist>
than some sort of more complex comparison
<geist>
ah
<cousteau>
geist: nowhere in the code were they comparing against single bits explicitly, hence my confusion
ats_ is now known as ats
<geist>
yah dunno
<geist>
i mean in general you can use a one hot for state machines nicely, since you just test against a single bit for any given state
<geist>
but if they're not actually using it that way it may just be a pattern that the designer used
<cousteau>
It just puts the cpu_state in a case statement (ie compares the whole thing with multiple constants). But I just noticed the case statement decorated with a full_case pragma, meaning "and other cases cannot ever happen, so go crazy with optimizations"
<sorear>
anyway sw-dev seems like a reasonable idea, i haven't been able to follow that in a while
<cousteau>
So maybe the comparison is being optimized as a "check this bit"
<geist>
yeah
<cousteau>
One thing I miss in Verilog is the ability to create "enumerate" constants without specifying their value or width, and let the synthesis tool choose the optimal representation
aurel32 has joined #riscv
<cousteau>
Now back to the real topic... What's this "squatting" I've kept hearing about after recent events?
<sorear>
one or more people are staying in ##riscv on freenode and telling people that all of the regulars moved to libera
<sorear>
probably with generous censoring and pantomime because they're doing server-side message filtering
<geist>
makes sense, because the original #riscv channel has been taken over by freenode
aurel32 has quit [Client Quit]
aurel32 has joined #riscv
<geist>
has happened another one i was in
<sorear>
and apparently staff has decided that *now* is the time to start enforcing the "no single-hash channels without official affiliation" rule, because they renamed the channel at the same time they took it over
SanchayanM has quit [Quit: SanchayanM]
SanchayanM has joined #riscv
<cousteau>
Wait, what happened to single-hash #riscv?
<cousteau>
Oh
<Sos>
#gentoo has set up a bot
<Sos>
that messages incomig users about the move
<Sos>
since they seized the channel too
SanchayanM has quit [Client Quit]
SanchayanMaity_ has joined #riscv
SanchayanMaity_ has quit [Client Quit]
<cousteau>
sorear: no matter how one attempts to see it. This is NOT ok
<cousteau>
I tried to "assume good faith" but it is now clear that they simply want to destroy freenode
<cousteau>
Or cannibalize it for who knows what purpose.
<cousteau>
So what exactly happened? Was #riscv on freenode abandoned/dropped, or did staff take it over, usurp it, and set a redirect to a channel controlled by them?
<cousteau>
Seriously, is there no legal way around this? This feels borderline illegal...
<Sos>
* Cannot join #riscv (Channel is invite only)
<cousteau>
But what about gentoo? It WAS an official channel
<Sos>
says "unofficial" in title
<cousteau>
The official one is the libera one. It is listed in gentoo.org. Maybe the freenode one BECAME unofficial after the Rasengularity
<sorear>
#riscv on freenode was already on shaky legal ground since AFAIK the foundation never grandfathered our right to use the name, but there are a lot of other orgs that *do* have grounds and this is going to be extremely popcorn
<cousteau>
Also "the gentoo namespace is registered to the Gentoo project"... you sure that one was taken over? It even has a Chanserv notice, I think
<cousteau>
Does gentoo have a strong enough legal representative base? Aren't the Linux foundation, FSF, etc doing anything about this?
<cousteau>
At some point freenode is gonna mess with the wrong entity and they're gonna get in deep trouble
<gruetzkopf>
naah, we're just gonna leave them with zero users
<gruetzkopf>
much less work
<gruetzkopf>
(and zero sponsored servers)
<cousteau>
Hard to get everyone to move out
<cousteau>
The latter seems more feasible though :)
<cousteau>
FSF Europe seems to have already officially moved out
<sorear>
i wonder how many of the sponsored servers are going to stick around for years due to bureaucratic inertia
<cousteau>
Lol, and #fsfe on freenode has been usurped and double-hashed as well, despite of being an official channel
<cousteau>
With official affiliation
<rjek>
They've shat the bed, here
TwoNotes has joined #riscv
valentin is now known as valentin_
valentin_ is now known as valentin__
valentin__ is now known as valentin
<gordonDrogon>
it's not nice all ways round - my Linux User Group had the move from # to ## last night too. It was as official as it gets but by merely mentioning the move to libera it got ##'d with all op's de-op'd ..
<gordonDrogon>
s/move/forced move/
<cousteau>
Yeah... Is there even an official policy that forbids mentioning other networks?
<gordonDrogon>
there is probably some catch-all about op's forcing users to move or something ... at least that's what I've gained from m LUG mailing list.
jotweh has quit [Ping timeout: 265 seconds]
craigo has quit [Read error: No route to host]
Gravis has quit [Ping timeout: 244 seconds]
<Sos>
cousteau, they have introsuced that rule few days ago
<cousteau>
Did they, or did they just rephrase the ToS in a way that could be interpreted that way?
jotweh has joined #riscv
<gordonDrogon>
personally, I don't think it matters - the actions appear somewhat erratic which does not make for a comfortable place/community.
<cousteau>
Anyway, I suspect it should be possible to force freenode to close channels named like projects they don't represent, on grounds of trademark infringement
dilfridge is now known as hephaistos
wingsorc has joined #riscv
<wingsorc>
is #riscv on freenode now abandonned?
<cousteau>
Usurped, more like
<wingsorc>
oh?
<gordonDrogon>
I just tried to join it - got bounced to ##
<cousteau>
From what I could gather, it set a notice saying that it had been moved here, and doing so makes the staff on that network to block the channel and redirect to ##riscv
<gordonDrogon>
17 users, so pretty much abandoned now.
<cousteau>
And this is how you kill IRC.
<gordonDrogon>
well - one network, but that will cause fall-out overall.
<gordonDrogon>
lack of trust, etc.
<cousteau>
gordonDrogon: make that 16; I just joined there to confirm the disaster was real
<cousteau>
Yeah. That dude wanted to "make IRC popular again" and did so by killing its biggest player.
<gordonDrogon>
almost ironic.
rah has joined #riscv
sh1r4s3 has joined #riscv
elastic_dog has quit [Quit: elastic_dog]
elastic_dog has joined #riscv
<sorear>
we've known this was coming for about four years, not much of a trust thing imo
<sorear>
what is frustrating is that so many channels weren't able to move until now
Raito_Bezarius has quit [Quit: free()]
Raito_Bezarius has joined #riscv
<cousteau>
I sure didn't know
Raito_Bezarius has quit [Remote host closed the connection]
Raito_Bezarius has joined #riscv
jn__ is now known as jn_
jn_ is now known as jn
h2t has quit [Ping timeout: 264 seconds]
geist has quit [Ping timeout: 264 seconds]
geist has joined #riscv
meowray has quit [Ping timeout: 252 seconds]
h2t has joined #riscv
meowray has joined #riscv
_whitelogger has joined #riscv
connojd has quit [Ping timeout: 264 seconds]
connojd has joined #riscv
sorear has quit [Ping timeout: 264 seconds]
geist2 has quit [Ping timeout: 264 seconds]
geist2 has joined #riscv
elms has quit [Ping timeout: 264 seconds]
meowray has quit [Ping timeout: 264 seconds]
arnd has quit [Ping timeout: 252 seconds]
connojd has quit [Ping timeout: 264 seconds]
JSharp has quit [Ping timeout: 264 seconds]
sorear has joined #riscv
meowray has joined #riscv
connojd has joined #riscv
elms has joined #riscv
arnd has joined #riscv
JSharp has joined #riscv
craigo has joined #riscv
choozy has joined #riscv
<NishanthMenon>
wingsorc, if you get a chance to drop into #openocd, we will appreciate it :)
adjtm has quit [Quit: Leaving]
adjtm has joined #riscv
elastic_dog has quit [Quit: #linux]
elastic_dog has joined #riscv
adjtm has quit [Quit: Leaving]
adjtm has joined #riscv
<choozy>
pabs3, see you here on the other side :p
<pabs3>
:)
adjtm has quit [Client Quit]
adjtm has joined #riscv
<pabs3>
Sos: I wonder if that Gentoo bot also works if you invite it to other channels?
adjtm has quit [Read error: Connection reset by peer]
TMM_ has joined #riscv
crest has joined #riscv
elastic_dog has quit [Quit: elastic_dog]
elastic_dog has joined #riscv
adjtm_ has quit [Quit: Leaving]
adjtm has joined #riscv
elastic_dog has quit [Quit: elastic_dog]
elastic_dog has joined #riscv
jotweh has quit [Remote host closed the connection]
jotweh has joined #riscv
<xentrac>
sorear: I didn't have any idea until last week
<xentrac>
cousteau: what HDLs do allow you to let the synthesis tool choose the value or width of enumerated constants?
<sorear>
hmm
guerby has quit [Quit: Leaving]
<sorear>
well if I find out that libera has become owned by a for-profit i'll make sure to say something here
<cousteau>
VHDL, for instance. I think.
<cousteau>
At least you don't explicitly assign them. Maybe they're implicitly assigned sequentially.
<cousteau>
type state_t is (ST_IDLE, ST_BUSY, ST_READY);
<cousteau>
And then Xilinx has an option to make them one-hot or not
elastic_dog has quit [Quit: elastic_dog]
elastic_dog has joined #riscv
<xentrac>
huh, yeah, I thought they were implicitly assigned sequentially in VHDL
guerby has joined #riscv
<cousteau>
Hm, that's the way it is in C. Now I need to check
<cousteau>
I don't think they have a numeric significance at all anyway
elastic_dog has quit [Client Quit]
<gordonDrogon>
random question - would the best way to learn RISC-V assembler be on a Linux board or some arduino-style embeded board? (or via a book!) I did a lot of sparc and i860 risc in years gone by fwiw...
elastic_dog has joined #riscv
<gordonDrogon>
or I guess I could find an emulator ...
<xentrac>
if "best" means "easiest" then probbly qemu-user
<xentrac>
or qemu-system if you want to write a kernel
olofj has joined #riscv
<cousteau>
xentrac: random PDF I found says they're automatically encoded as 000 001 and so on (not sure if that's official though)
<cousteau>
But in any case it DOES make them a little more abstract
<xentrac>
hmm
<cousteau>
Ideally I'd look it up on the standard, but I don't have it here
<TwoNotes>
gordonDragon Read all the ISP specs first. Then play with qemu or get a Longnan Nano or similar
<TwoNotes>
And get all the GNU development tools. assembler, loader, etc
<xentrac>
attribute enum_encoding of t_state : type is "one-hot"; -- encoding style of the enumerated type
<gordonDrogon>
xentrac, TwoNotes thanks. yes, might read the fine manuals then get a little board. I find working on 'real' hardware easier than an emulator in-general, although to better understand the last assembler I learned, I wrote my own emulator (pdp8, so relatively easy!)
<xentrac>
RV64I is probably easier than PDP-8
<xentrac>
to write an emulator for
<xentrac>
RV64G is probably a bit harder
<gordonDrogon>
sticking to 32-bit for now though, but...
<xentrac>
I think that holds a fortiori for RV32I
awordnot has quit [Quit: WeeChat 1.9.1]
awordnot has joined #riscv
choozy has joined #riscv
Narrat has joined #riscv
jotweh has quit [Ping timeout: 264 seconds]
wingsorc has quit [Remote host closed the connection]
hephaistos is now known as dionysos
wingsorc has joined #riscv
fabs has joined #riscv
<TwoNotes>
I used to program IBM 360s in assembler. RV32 felt very familiar.
<TwoNotes>
Ive programmed PDP8s too. That was harder
<sorear>
reasonable ISAs with their hexdump-friendly instruction sets ;)
<gordonDrogon>
I've programmed i860's. ... Lost too many braincells on that one... Actually it was OK as long as you avoided the floating point/dual instruction mode stuff...
<joev>
I'm futzing around making a RV32/64/128 emulator as a side-side-project to refamiliarize myself with C, and the I instructions are a breeze, but for the F/D/Q (floating point) instructions, I don't know if I can trust C to handle the IEEE754 cases properly on the underlying hardware, so I may end up writing Yet Another FP Library if it can keep my interest.
<sorear>
C is a bit generous yeah
<sorear>
spike and qemu use (different versions of) the berkeley softfloat lib
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.]
<dh`>
my solution to that problem has just been to not support floating point
<dh`>
but for my emulator's goals (basically it's a teaching platform) FP isn't particularly necessary
Narrat has joined #riscv
stefanct has quit [Quit: quit]
<gordonDrogon>
my interest in RISC-V is as a platform to support my little 'retro' style OS thing - which runs BCPL. I'm almost tempted to write an emulator in BCPL to run the er. BCPL system under ... It might be a little slow though...
<gordonDrogon>
it currently runs on a 65x816 at 16Mhz...
stefanct has joined #riscv
<joev>
I, too, hang out on #6502 :)
peeps[zen] is now known as peepsalot
<gordonDrogon>
joev, so you'll know about my shenanigans then ;-)
<gordonDrogon>
for others (if interested!) the bcpl compiler generates a bytecode that's then interpreted - my interpreter creates the 32bit environment on a 16-bit processor with an 8-bit memory bus and 64K banks of RAM. Such is the joy of the 65816...
sh1r4s3 has quit [Ping timeout: 272 seconds]
<gordonDrogon>
so my 'retro' part was initially the hardware (6502/65816) but then I looked at what might be the next thing - ARM is the obvious choice, but I don't like being obvious, so there are limited retro/period CPUs available, so jump to the future with RISC-V ...
<sorear>
FYI: the #libera-matrix people are currently saying 2-4 days and existing plumbed rooms will be migrated
<sorear>
if there are any [m] stragglers we may be able to point them to #riscv:libera.chat now
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.]
<dh`>
I disconnected for good after last night's shenanigans
<geist>
yah agreed
vagrantc has quit [Quit: leaving]
<jrtc27>
I'm still in, I'll leave once my znc has to reconnect and thus I get booted to all the ## channels
<jrtc27>
atm it's a select, time-limited club :P
<sorear>
well you can't send a message to the # channel
<jrtc27>
boo
<sorear>
so the whole reason to be there is to tell people in the ## that the regulars are all somewhere else
<sorear>
in particular (a) matrix users, who could not connect to libera.chat until a few minutes ago; (b) people not online enough to know that most freenode channels have moved to libera or oftc
<sorear>
(I can see people testing the bridge, it may not be ready for serious use yet)
<sorear>
i don't expect to get back to 228 users any time soon but 126 is a bit alarming, many of my other channels are at 75% or more
<dh`>
#coq is much smaller here than it used to be over there, but most of those people have never said a word in the past year+
<geist>
looks like a pretty good chunk of the #osdev channel has made it over
<jrtc27>
tends to be smaller when people have lost their excitement
* jrtc27
hides
<geist>
though there was a small spinoff of folks that wanted to stay behind
<dh`>
yeah, I've seen some of that, the reaction is "more ircop drama? no reason for me to care" but most of them have done a double-take after seeing the actual situation
<sorear>
looks like 128 unique nicks in the freenode channel this year counting only non-ACTION messages
<cousteau>
jrtc27: lol
<cousteau>
I was wondering what happened if a channel got repounded when you were still in it
<cousteau>
sorear: I still fail to see how preventing channel owners from telling their "public" that the channel is no longer official is even legal
<xentrac>
are you talking about civil or criminal law, and in what jurisdiction?
<sorear>
AFAIK trademark infringement is always civil and the entity Lee bought, Freenode Ltd, operates under UK law
<xentrac>
that doesn't mean freenode is only subject to UK jurisdiction
<geist>
oh i think the Korean Monarchy would say otherwise
* geist
rolls eyes
<sorear>
if they're subject to multiple jurisdictions that only hurts them because they only have to be liable in one
<xentrac>
right
<dh`>
it is unlikely that there's actually an enforceable contract between freenode and any project organization
<sorear>
IDK if they're liable in any but given the whole "#rhel takeover" thing we may find out
<xentrac>
but I was curious which one cousteau was thinking of. maybe your inference that they were thinking of trademark infringement is correct, sorear
<dh`>
since there's no money involved and (AIUI, IANAL) forming a contract usually requires money
<dh`>
trademark infringement is a very thin proposition here
<xentrac>
dh`: are you suggesting that cousteau was talking about a breach-of-contract tort rather than a trademark infringement tort?
<xentrac>
forming a contract, including a synallagmatic contract, does *not* require money under US law or UK law