sorear changed the topic of #riscv to: RISC-V instruction set architecture | https://riscv.org | Logs: https://libera.irclog.whitequark.org/riscv | Matrix: #riscv:catircservices.org
Jackneill has quit [Ping timeout: 276 seconds]
jmdaemon has joined #riscv
jacklsw has joined #riscv
Tenkawa has quit [Quit: Was I really ever here?]
jacklsw has quit [Ping timeout: 255 seconds]
averymt has joined #riscv
dh` has quit [Ping timeout: 252 seconds]
agentcasey has quit [Quit: Lost terminal]
hightower3 has joined #riscv
hightower2 has quit [Ping timeout: 252 seconds]
vagrantc has quit [Quit: leaving]
averymt has quit [Read error: Connection reset by peer]
averymt has joined #riscv
jmd_ has joined #riscv
jmdaemon has quit [Ping timeout: 256 seconds]
heat has quit [Read error: Connection reset by peer]
heat has joined #riscv
averymt has quit [Ping timeout: 255 seconds]
averymt has joined #riscv
mlw has joined #riscv
ntwk has quit [Read error: Connection reset by peer]
heat has quit [Ping timeout: 276 seconds]
davidlt has joined #riscv
mlw has quit [Ping timeout: 255 seconds]
mlw has joined #riscv
jacklsw has joined #riscv
<courmisch> sorear: not sure, but I guess 0.7.1
BootLayer has joined #riscv
davidlt has quit [Ping timeout: 255 seconds]
crabbedhaloablut has joined #riscv
psydroid2 has joined #riscv
Jackneill has joined #riscv
shamoe has quit [Quit: Connection closed for inactivity]
Leopold has quit [Quit: No Ping reply in 180 seconds.]
Leopold has joined #riscv
dh` has joined #riscv
dh` has joined #riscv
TMM_ has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
TMM_ has joined #riscv
Stat_headcrabed has joined #riscv
davidlt has joined #riscv
dzaima[m] has joined #riscv
<dzaima[m]> contains vdot.vv, so not 1.0
Laurence_b has joined #riscv
Stat_headcrabed has quit [Ping timeout: 255 seconds]
mlw has quit [Ping timeout: 256 seconds]
mlw has joined #riscv
jfsimon has quit [Remote host closed the connection]
jrjsmrtn has joined #riscv
jrjsmrtn_ has quit [Ping timeout: 255 seconds]
<courmisch> so the T-Head kernels set the V bit in HWCAP if the DT has it, regardless of actual kernel support
<courmisch> WhatCouldPossiblyGoWrong
<courmisch> I don't remember who asked for the FFmpeg perfs https://paste.debian.net/1298564/
<courmisch> sorry the layout sucks
ntwk has joined #riscv
Laurence_b has quit [Ping timeout: 250 seconds]
Laurence_b has joined #riscv
fusta has joined #riscv
<conchuod> courmisch: meanwhile their "first gen" stuff putting there is actively incorrect to begin with.
<conchuod> "there" being in the dt
fusta has quit [Quit: Leaving]
<courmisch> conchuod: oh I know
<courmisch> conchuod: fun getting a SIGILL even though you checkef HWCAP
<courmisch> conchuod: did the conditional blacklisting of v from DT get merged?
shamoe has joined #riscv
ntwk has quit [Read error: Connection reset by peer]
ntwk has joined #riscv
ntwk has quit [Client Quit]
ntwk has joined #riscv
Andre_Z has joined #riscv
Laurence_b has quit [Quit: Client closed]
pabs3 has quit [Ping timeout: 256 seconds]
pabs3 has joined #riscv
heat has joined #riscv
<conchuod> wdym?
<courmisch> conchuod: wasn't there a patch to ignore V in DT for a certain vendor?
jmd_ has quit [Ping timeout: 256 seconds]
cwittlut_ is now known as cwittlut
<unlord> courmisch: this is on k230?
<conchuod> Oh ye I did send one of those I think
<courmisch> unlord: it's on all T-Head kernels
<unlord> with 0.7.1?
<courmisch> K230 has V in HWCAP even if Linux is running on the little core
<courmisch> and TH1520 has V in HWCAP even though it's 0.7.1
<courmisch> a year ago somebody already noted that I broke FFmpeg on Allwinner D1 (C906), and I'm pretty sure that was the reason
<courmisch> well there are ways to check for 0.7.1
<courmisch> but that's not how the official kernel ABI works
<unlord> huh
<courmisch> both functions return true on RVV 1.0.0, false on 0.7.1, and cause SIGILL without V
monoidog has joined #riscv
<courmisch> left as an exercise to the reader: how do they work...
<courmisch> I'll let you think about it
jacklsw has quit [Ping timeout: 264 seconds]
EchelonX has joined #riscv
<unlord> I think I understand what they do, but I'm not sure why
ntwk has quit [Quit: ntwk]
___nick___ has joined #riscv
___nick___ has quit [Client Quit]
___nick___ has joined #riscv
ntwk has joined #riscv
<conchuod> I dunno, not much "we" can do if vendors ship crappy kernels.
<conchuod> Or use incorrect devicetrees.
<courmisch> conchuod: you can't fix vendor kernels based on old kernel versions? shocking!
<conchuod> I mean, we could litter userspace or whatever with hacks, but I don't think that is right.
<courmisch> I agree, and I no longer maintain my FFmpeg for RVV 0.7.1 branch
<courmisch> But I have this fear that users will install distro with vendor kernels on such devices and start whining that stuff crashes
<courmisch> I mean, it's one thing to not provide optimisation for 0.7.1, and it's another thing to crash
<gurki> its a quite volatile situation for rv imho. theres a lot of boards with 0.7.1 right now, and if these misbehave people will move away from rv
<gurki> im aware none of this strictily is the sw guys fault
<gurki> strictly*
<gurki> (what im trying to say? you are correct. but we still need to somehwat need to stuff not die horribly on these boards ...)
<gurki> i butchered that one, but i hope you get my point
<courmisch> and approach is to use only riscv_hwprobe() and skip HWCAP entirely
<courmisch> that one will just fail on affected vendor kernels
<courmisch> but then my favorite pet peeve: where is glibc's support
<courmisch> s/^and/one/
monoidog has quit [Ping timeout: 256 seconds]
ntwk has quit [Ping timeout: 256 seconds]
davidlt has quit [Ping timeout: 264 seconds]
Jackneill has quit [Ping timeout: 268 seconds]
uzix is now known as mahk
mlw has quit [Ping timeout: 255 seconds]
TMM_ has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
TMM_ has joined #riscv
heat has quit [Remote host closed the connection]
heat has joined #riscv
Tenkawa has joined #riscv
danielinux has quit [Changing host]
danielinux has joined #riscv
aredridel has quit [Quit: The Lounge - https://thelounge.chat]
aredridel has joined #riscv
crabbedhaloablut has quit [Ping timeout: 255 seconds]
crabbedhaloablut has joined #riscv
___nick___ has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
___nick___ has joined #riscv
___nick___ has quit [Ping timeout: 256 seconds]
vagrantc has joined #riscv
crabbedhaloablut has quit [Ping timeout: 256 seconds]
crabbedhaloablut has joined #riscv
<sorear> riscv_hwprobe is strictly a hardware/kernel check and will crash if V is disabled at runtime with the prctl
vgtw_ has quit [Quit: ZNC - https://znc.in]
psydroid2 has quit [Quit: KVIrc 5.0.0 Aria http://www.kvirc.net/]
Jackneill has joined #riscv
BootLayer has quit [Quit: Leaving]
vgtw has joined #riscv
gordonDrogon has quit [Ping timeout: 264 seconds]
gordonDrogon has joined #riscv
prabhakarlad has quit [Ping timeout: 250 seconds]
jmdaemon has joined #riscv
prabhakarlad has joined #riscv
notgull has quit [Ping timeout: 255 seconds]
notgull has joined #riscv
sakman_ has joined #riscv
sakman has quit [Ping timeout: 264 seconds]
prabhakarlad has quit [Quit: Client closed]
prabhakar has quit [Quit: Connection closed]
prabhakar has joined #riscv
prabhakarlad has joined #riscv
Andre_Z has quit [Quit: Leaving.]
ntwk has joined #riscv
whysix_ has joined #riscv
crabbedhaloablut has quit []
tlwoerner_ has quit [Quit: Leaving]
tlwoerner has joined #riscv
monoidog has joined #riscv
monoidog has quit [Remote host closed the connection]
wingsorc has quit [Read error: Connection reset by peer]
Jackneill has quit [Ping timeout: 276 seconds]