<conchuod>
Esmil: p much for SoB I use my work email for anything remotely work related, but I split the email in maintainers/files depending on whether I get paid to look after it.
<conchuod>
Figure that's the best way to keep legal types at microchip off my back :)
Snuupy has quit [Read error: Connection reset by peer]
Snuupy has joined #riscv
courmisch has quit [Ping timeout: 246 seconds]
courmisch has joined #riscv
scruffyfurn has quit [Ping timeout: 246 seconds]
jmdaemon has quit [Ping timeout: 252 seconds]
scruffyfurn has joined #riscv
<Esmil>
conchuod: yeah, i think it's fine to keep it at my non-work email. i'll probably do the reviews from my work email for now, but then I don't need to change it if priorities change
tusko has quit [Remote host closed the connection]
tusko has joined #riscv
pedja has joined #riscv
tusko has quit [Remote host closed the connection]
tusko has joined #riscv
dor has joined #riscv
wingsorc has quit [Ping timeout: 252 seconds]
epony has joined #riscv
frkzoid has quit [Ping timeout: 260 seconds]
jimbzy has quit [Changing host]
jimbzy has joined #riscv
somlo has joined #riscv
prabhakarlad has quit [Quit: Client closed]
wingsorc has joined #riscv
Trifton has quit [Quit: Error: no route to host]
BootLayer has quit [Quit: Leaving]
marcj has joined #riscv
BootLayer has joined #riscv
jacklsw has joined #riscv
jack_lsw has joined #riscv
jack_lsw has quit [Client Quit]
jacklsw has quit [Ping timeout: 246 seconds]
<another|>
hmm.. do vector instructions require some priviledged mode? M? S?
<bjdooks>
don't think so
<bjdooks>
not sure if they rquire a csr bit set though to work
bauruine has quit [Remote host closed the connection]
jacklsw has joined #riscv
ldevulder has quit [Ping timeout: 252 seconds]
ldevulder has joined #riscv
prabhakarlad has joined #riscv
frkzoid has joined #riscv
jacklsw has quit [Quit: Back to the real life]
jacklsw has joined #riscv
jmdaemon has joined #riscv
meowray has quit [Changing host]
meowray has joined #riscv
vagrantc has joined #riscv
aburgess has quit [Ping timeout: 248 seconds]
frkzoid has quit [Read error: Connection reset by peer]
frkzoid has joined #riscv
jacklsw has quit [Read error: Connection reset by peer]
frkzoid has quit [Read error: Connection reset by peer]
<palmer>
you can't really have RISC-V without _some_ privilege mode. IIRC mstatus.VS controls access to the V state
davidlt has quit [Ping timeout: 252 seconds]
<muurkha>
palmer: in the sense that if you only have one mode it's M-mode, which is "privileged"?
<palmer>
ya, though kind of a moot point
prabhakarlad has joined #riscv
<palmer>
but I guess "do vector instructions require some privileged mode" is sort of a poorly formed question: the V stuff doesn't, but everything does ;)
<muurkha>
not everything, just some things. "do vector instructions require some privileged mode" is relevant to programs that don't require privileged mode because they're running under an OS
<muurkha>
it's true that the OS requires privileged mode!
<muurkha>
I mean it needs to handle traps, right?
jjido has joined #riscv
BootLayer has quit [Quit: Leaving]
frkazoid333 has joined #riscv
frkzoid has quit [Ping timeout: 255 seconds]
aburgess has joined #riscv
prabhakarlad has quit [Quit: Client closed]
frkzoid has joined #riscv
frkazoid333 has quit [Ping timeout: 246 seconds]
___nick___ has quit [Ping timeout: 252 seconds]
<palmer>
is anyone else getting watchdog/soft lockups? I just merged Linus' master
<bjdooks>
but if the OS can't save the vector state on task switch then there's no a lot of point in enabling vector instructions
zjason has quit [Read error: Connection reset by peer]
Trifton has quit [Quit: Error: no route to host]
<conchuod>
palmer: i dont really run qemu but I have seen no issues specific to 6.1-rcN so far
<palmer>
OK
<conchuod>
in hardware
<palmer>
do you run lockdep?
<conchuod>
I have:
<conchuod>
CONFIG_PROVE_LOCKING=y
<conchuod>
CONFIG_DEBUG_LOCK_ALLOC=y
<conchuod>
I sometimes have the selftests on too
<palmer>
CONFIG_DEBUG_LOCKING_API_SELFTESTS=y ? I just have all 3
<conchuod>
Yah, but I usually leave that off since it wrecks my boot time
<conchuod>
Either way, only lockup I ever see is the suspend thing and that's existed for 2 cycles now.
<palmer>
cool
crabbedhaloablut has quit [Ping timeout: 255 seconds]
crabbedhaloablut has joined #riscv
tusko has quit [Remote host closed the connection]
tusko has joined #riscv
<drmpeg>
6.1.0-rc4 is good here on Unmatched. Don't have any of that lockdep stuff on though.
tusko has quit [Remote host closed the connection]
tusko has joined #riscv
crabbedhaloablut has quit [Remote host closed the connection]
crabbedhaloablut has joined #riscv
jmd_ has joined #riscv
<another|>
btw: solved it by patching 5.18.19 with the v10 vector patchset
jmdaemon has quit [Ping timeout: 255 seconds]
<another|>
6.0+ with v12 is borken
Gravis has quit [Read error: Connection reset by peer]
Gravis has joined #riscv
jjido has quit [Quit: My laptop has gone to sleep. ZZZzzz…]
vagrantc has quit [Quit: leaving]
Gravis has quit [Client Quit]
Gravis has joined #riscv
<palmer>
drmpeg: thanks, hopefully it was just a suprious failure
<palmer>
another|: did you respend to the v12 patches? I think vineet and bjorn are working on them now
tusko has quit [Remote host closed the connection]
tusko has joined #riscv
Gravis has quit [Quit: Murdered]
rsalveti has quit [Quit: Connection closed for inactivity]
Gravis has joined #riscv
<conchuod>
Getting some nice oopses in network on my new machine 🙃
<conchuod>
Price of buying brand new hw I suppose.
<another|>
palmer: no, I'm not setup for email rn
<palmer>
OK, bjoto will probably see here
crabbedhaloablut has quit [Remote host closed the connection]
tusko has quit [Remote host closed the connection]
tusko has joined #riscv
crabbedhaloablut has joined #riscv
<conchuod>
For the vector stuff, what is the best tc to actually test it with?
<conchuod>
clang/llvm main?
<palmer>
what do you mean by "test"?
<palmer>
the kernel patches should be fine with pretty much any toolchain, but really testing it needs some userspace
<palmer>
we don't have a libc yet (vineet was working it last I checked), and there's not a whole lot that can be done without one
<conchuod>
I mostly just want to see what it breaks by compiling it in.
pecastro has quit [Remote host closed the connection]
<palmer>
then you should just need a V-capable assembler, any new binutils should be fine (LLVM had some older V specs for a while, not sure how to detect those)
<conchuod>
new = ?
<conchuod>
bjoto: not merged yet but tuxmake has added setting CROSS_COMPILE
<conchuod>
Sweet, thanks for doing what I should probably have googled myself...
<jrtc27>
if LLVM accepts -march=...v then it's the right v
<jrtc27>
pre-ratification requires a v0p7 etc and -menable-experimental-extensions
<palmer>
OK, that's easy then
<muurkha>
cool!
prabhakarlad has quit [Quit: Client closed]
<dh`>
one of these days I should really package netbsd's libc for experimental 3rd-party use
<muurkha>
steal musl's thunder!
JanC has quit [Read error: Software caused connection abort]
<dh`>
that would be hard given that musl's got an actual userbase
JanC has joined #riscv
<muurkha>
so does netbsd
<dh`>
also I can do some things but porting to linux would then take a bunch further work
<dh`>
yes, but not on linux...
<muurkha>
admittedly netbsd's user base on linux is pretty small!
<muurkha>
but glibc grumbling isn't a new phenomenon with docker, it goes back decades
<dh`>
glibc being crap goes back decades :-)
<muurkha>
heh
<dh`>
(it and the community dysfunction it represents were a substantial chunk of the reason I left the linux world and didn't look back)
seds has quit [Read error: Software caused connection abort]
seds has joined #riscv
<jrtc27>
~~you've heard of eglibc, get ready for fglibc~~
<muurkha>
do you mean ulrich being aggressive?
<dh`>
it was never any one thing; he was partly a symptom
tusko has quit [Ping timeout: 255 seconds]
tusko_ has joined #riscv
<dh`>
that is, the way glibc was maintained and the fact that a person with the ... leadership skills he had was able to be in charge of it for years reflect fundamental problems in the linux community
<dh`>
(most of which I think has not changed, in fact)
<dh`>
(but I haven't been present in ages so IDK)
tusko_ is now known as tusko
prabhakarlad has joined #riscv
tusko has quit [Remote host closed the connection]
tusko has joined #riscv
<drmpeg>
palmer: I did a build with the lockdep stuff on. It boots fine, but I get a bunch of warnings. https://www.w6rz.net/warning.txt
<drmpeg>
Same warning repeated 35 times.
<muurkha>
dh`: Stallman has his merits but he has never been a kind man; he has often been vicious, and I think that shaped the free software movement