<davidlt[m]>
I haven't had too much time lately and there is a bit of storage space related issues :)
<davidlt[m]>
I could cook up a new one with an updated kernel if that helps, would take some days
<defolos>
well, I have a starfive visionfive on my desk and would like to test things 😉
<davidlt[m]>
That board is not supported directly by this Koji
<davidlt[m]>
There are no proper upstream support and might not even get one
<davidlt[m]>
StarFive V2 (with 7110 SoC) most likely will be supported
<defolos>
oh that's a bummer
<defolos>
if you have a minute, why might it never get upstream support?
<davidlt[m]>
There is some work going on using alternative framework to get 7100 working, but that's months in WIP
<davidlt[m]>
Because the way the chip was made, everything is not cache coherent, I mean peripherals
<davidlt[m]>
The current code has plenty of manual cache flushing, etc. in various drivers IIRC
<tekkamanninja>
yes, 7110 will be support
<tekkamanninja>
davidlt[m]: ping
<tekkamanninja>
need your help
<davidlt[m]>
yes
<tekkamanninja>
do you know what is the best practice of update python and perl on Linux distro, like fedoar
<davidlt[m]>
Just update python, macros, etc. and then rebuild anything that depends on python (most of that will be imported from the main Koji).
<davidlt[m]>
There might be a few cyclic dependencies that could require manual fixing.
<tekkamanninja>
we will need to update python and perl for Fedora 36,
<tekkamanninja>
from f33
<davidlt[m]>
Perl will be a bit more problematic, it always requires more work.
<davidlt[m]>
There are lots of cyclic dependencies.
<tekkamanninja>
for example, we thought we should build perl , and build his module and rpmbuild manually and install manually, then setup rpm build env , am I correct?
<tekkamanninja>
davidlt[m]: ^^^
<davidlt[m]>
Don't you have Koji?
<tekkamanninja>
using new version of perl with old module will break the env of rpmbuild
<davidlt[m]>
Just make sure all macros are up to date and just rebuild Perl (most likely in bootstrap mode).
<davidlt[m]>
Never had this problem before and we had a few ABI breaking changes with Perl.
<davidlt[m]>
As long as you follow thing somewhat properly it will work fine.
<tekkamanninja>
bootstrap mode , OK, thanks :-)
<tekkamanninja>
let me have a try
<tekkamanninja>
Great thanks
zsun has joined #fedora-riscv
<Esmil>
davidlt[m]: that's not quite true. in the visinofive kernel all the cache flushing spread around drivers has been replaced by a non-coherent dma implementation
<Esmil>
the work is in make that acceptable for upstream
<Esmil>
*making the non-coherent dma implementation acceptable for upstream
<tekkamanninja>
Esmil: 7110 maybe doesn't need non-coherent, hope they have fixed the problem
<tekkamanninja>
hope I understand correctly,
<tekkamanninja>
davidlt[m]: ^^^
<Esmil>
tekkamanninja: yes, that's why Michael tells me
<Esmil>
*that's what
<tekkamanninja>
:-)
<tekkamanninja>
Micheal was here last night , I guess :-)
zsun has quit [Quit: Leaving.]
zsun has joined #fedora-riscv
zsun has quit [Client Quit]
tekkamanninja_ has joined #fedora-riscv
tekkamanninja has quit [Ping timeout: 256 seconds]