<davidlt[m]>
JH7110 ISA should be: RV64IMAFDCSU_Zicsr_Zifencei_Zba_Zbb_Sscofpmf
<davidlt[m]>
GCC 13 build is still alive,
jcajka has joined #fedora-riscv
davidlt has quit [Remote host closed the connection]
davidlt has joined #fedora-riscv
Kevinsadminaccou has quit [Quit: You have been kicked for being idle]
davidlt has quit [Ping timeout: 252 seconds]
davidlt has joined #fedora-riscv
<davidlt[m]>
neil: booted VF2 V1.3B!
masami has joined #fedora-riscv
<davidlt[m]>
I booted from GH images instead of a custom distro. New GH release was posted few hours ago.
<davidlt[m]>
There is no support for HTTPs on wget, no sftp support too, and no netcat too :)
<davidlt[m]>
Time to test USB with USB 2.0 drive to transfer new firmware files.
<davidlt[m]>
Both Ethernet ports work. Reset button works. UART0 works (3.3V). Fan header is 5V (tested with multi-meter).
<davidlt[m]>
There is no SBI command in U-Boot, reset/poweroff is not wired too.
<davidlt[m]>
All USBs seem to work with USB 2.0 storage drive.
<davidlt[m]>
Booted into a new firmware.
<davidlt[m]>
Kernel crashes with eMMC module.
<davidlt[m]>
I think M.2 slot screw is M1 or M1.5
<davidlt[m]>
It's a bit sad there is no USB FTDI chip.
<davidlt[m]>
NVMe seems to work too.
<davidlt[m]>
There are reports about IO timing out on NVMe thus probably some issues exist.
<davidlt[m]>
Looking in dmesg there are definitely issues, but it's early days. Several months and tons of user reporting issues/testing will iron out most of the issues.
davidlt_ has joined #fedora-riscv
davidlt has quit [Remote host closed the connection]
davidlt_ has quit [Remote host closed the connection]
<davidlt[m]>
So far I didn't. I disconnected it for now to avoid spending too much time on it today :)
<cwt[m]>
I think i/o timeouts is general issue on the Linux kernel, people got this on the latest kernel on amd64 too.
<neil>
davidlt[m]! nice :D i need to find an SD card...
Kevinsadminaccou has joined #fedora-riscv
<somlo>
davidlt[m]: just got the greg-kh auto-email about litex uart irq support having made it into the `tty-testing` branch of `tty-next`
<somlo>
so I now know I don't need to harass anyone anymore, the patches are indeed "percolating" :D
<davidlt[m]>
Cool :)
<davidlt[m]>
So it seems Carlos is fine with me pulling in binutils 2.40. Nick will land it in Rawhide soon after F38 is branched and I will pick up the build once it is in a proper tag.
<davidlt[m]>
Our mass rebuild should probably start soon after the upstream one finishes.
<davidlt[m]>
(not to mention long GCC build times for us)
<davidlt[m]>
I expect Jakub to bump GCC 13 at least once a week (usually every several days IIRC).
<davidlt[m]>
Keeping up will be a bit difficult.
<cwt[m]>
Do you put any fan or heatsink on your VF2?
<davidlt[m]>
So far I haven't seen anything too bad in builds failed during mass rebuild with GCC 13.
<davidlt[m]>
don't know yet, but the cooler the better and I expect to use some of these boards in the build farm. That means high-loads for a long time.
<cwt[m]>
davidlt[m]: I put only tiny Raspberry Pi heatsinks on CPU, RAM, and USB without fan. make -j4 bump temp up to 65C, should I worry?
<davidlt[m]>
Probably not. I don't have max T for JH7110, but it's probably in datasheet.
<davidlt[m]>
For FU740 (which is not perfect) we set max to 80C, but you really don't want to be close to that.
<davidlt[m]>
IIUC max is 125C from their datasheet.
<davidlt[m]>
I am not sure if individual chips are validated for stability under high temps.
<davidlt[m]>
Probably not a problem if you are just using it, but of a problem if you have a constant high load and you are in hot environment.
<cwt[m]>
I plan to put 2 boards together on a gigabit switch to make a cluster then try to stick a 5 inch fan for the whole stack.
<davidlt[m]>
Upstream support is unknown. I would expect this to fall into Alibaba/T-HEAD and community efforts.
<cwt[m]>
I am very interested indeed! but now I really want to know about the "optimized toolchain" mentioned in the CoreMarks graph.
<davidlt[m]>
There are a lot of T-HEAD vendor extensions. Those have been merged into GCC 13 and binutils 2.40.
<davidlt[m]>
I bet is that extension + march/mtune were set. Maybe other generic riscv arch specific tweaks.
<somlo>
huh... it just occurred to me, I have 24 host cores (old, 2014 vintage, but still) -- so I could have *three* riscv fedora qemu machines building *three* kernels, speculatively, in "parallel" (for my config option "bisect" adventure)
<somlo>
I'm frankly ashamed it took me this long to think about it :D
<davidlt[m]>
somlo: that's a good thinking ;)
<davidlt[m]>
I do a lot of parallel stuff because things take too long.
<davidlt[m]>
That's way I have way too much tabs, tmux sessions, etc.
<somlo>
I get to *speculate* on which flags might be useful to test, and then earlier builds will determine whether my speculation was right -- almost like the real thing, except at comparatively geological time scales ;P
<cwt[m]>
Time to sleep for me now. Good night.
<davidlt[m]>
cwt: good night.
<somlo>
later! :)
dtometzki is now known as dti
davidlt has quit [Ping timeout: 268 seconds]
jcajka has quit [Quit: Leaving]
tg has joined #fedora-riscv
davidlt has joined #fedora-riscv
dti is now known as dtometzki
iooi has quit [Read error: Connection reset by peer]
iooi has joined #fedora-riscv
iooi has quit [Read error: Connection reset by peer]