crabbedhaloablut has quit [Remote host closed the connection]
crabbedhaloablut has joined #riscv
<dramforever__>
Esmil: Just curious, but I see that starfive-tech/linux lists NVDLA as ready, but I couldn't find any description of anyone actually using it. Is there like a 'smoke test script' you used to try it out or something?
<dramforever__>
I know there's a 'umd' in https://github.com/nvdla/sw but I'm clueless about what the heck is going on
<dramforever__>
It does seem to compile fine (with a few #include changes)
<palmer>
I was hoping to send a PR today but I'm stuck on this lm90 stringop-overread build issue
memoryleak has quit [Ping timeout: 252 seconds]
frkzoid has joined #riscv
<conchuod>
I won't claim to understand what that is... Sounds "fun" though
frkazoid333 has quit [Ping timeout: 244 seconds]
memoryleak has joined #riscv
jacklsw has quit [Read error: Connection reset by peer]
enthusi has joined #riscv
<palmer>
it's just a build failure on rv32 allmodconfig with Linus' master, but we're triggering some fortify-string craziness so it's a headache
loggervicky1 has joined #riscv
loggervicky has quit [Ping timeout: 268 seconds]
loggervicky1 is now known as loggervicky
<conchuod>
I see. Didn't see anything about that but it's prob gone to some list I don't subscribe to.
<conchuod>
Can't say I'd be much help with fortify-string stuff, but I am p much always around palmer so if you need any help with the janitorial side of things lmk...
<conchuod>
You always seem to be flat out so I always feel guily asking things of you hah
<conchuod>
guilty*
<palmer>
no, it's OK, it's not right to let thus stuff slide
<palmer>
if you want to help, one good thing to do would be to test linux-next and find build/boot issues, I don't think anyone's doing that with a test suite as aggressive as mine is
<conchuod>
I do build and boot -next every day (although with a limited set of configs & pretty much just on polarfire)
<palmer>
Ok, great
<conchuod>
Plan on our side is to increase coverage of what we are testing once booted, but that takes time :/
matt__ has joined #riscv
<palmer>
I've found that it's mostly rv32 allmodconfig/allyesconfig that fails, it looks we're pretty much the only 32-bit platform for a lot of drivers these days
<palmer>
so if you have bandwidth for one more thing to test, that'd be awesome -- even just build testing is great, it's mostly just build bugs in drivers that crop up
<conchuod>
I should just chuck your riscv-system stuff into my daily upstream-next build
<palmer>
if that's easy that'd be awesome
<palmer>
another option would be to move all those configs I have over to upstream configs, that'd make it easier for other folks to run them
<conchuod>
One or two "bells and whistles" configs would be nice.
frkzoid has quit [Ping timeout: 255 seconds]
<palmer>
I guess to start nobody else seems to be testing all*config, and those are generally the first ones to get tested
BootLayer has quit [Quit: Leaving]
<conchuod>
allmodconfig is one that I added today but it is failing for me.
<conchuod>
Prob some sort of -Werror thing, I was on the way out the door when I saw it fail
<conchuod>
I forget, riscv-systems-ci just needs to be pointed at a kernel tree right?
<conchuod>
And if it passes, it'll just return 0?
<drmpeg>
I have a big Ubuntu config that I use for Linux stable. It's about 2/3 of allyesconfig.
<conchuod>
Ah I think the allmodconfig failure was -Werror + C=1
<conchuod>
Either way, I'll throw your riscv-systems-ci into my daily runs :)