02:02
iooi has quit [Quit: iooi]
05:18
cyberpear has joined #fedora-riscv
05:50
jcajka has joined #fedora-riscv
07:07
<
davidlt[m] >
We passed 1'400'000 tasks on our Koji instance!
09:10
iooi has joined #fedora-riscv
13:15
<
PrathamPatel[m] >
Now that some components of VF2/JH7110 have been merged in -next, I wonder what is yet to be done for minimal support from the mainline kernel
13:16
<
PrathamPatel[m] >
I recall asking this question to you davidlt but can't search for it in this room
13:16
<
davidlt[m] >
All what's needed is already approved for v6.4 IIRC.
13:16
<
PrathamPatel[m] >
Do you mind re-iterating? 😅
13:16
<
PrathamPatel[m] >
Oh okay, compiling -next as we speak then :D
13:16
<
PrathamPatel[m] >
I will update on how this goes
13:17
<
davidlt[m] >
v6.4 should boot, but DT that landed lacks a few DT nodes that have drivers merged in v6.3 IIRC. That should get enabled in DTS in v6.5.
13:17
<
PrathamPatel[m] >
What drivers are affected by this devicetree-related issue?
13:18
<
davidlt[m] >
This was needed for systems that use older toolchains, or have a different default ISA spec selected at compile time.
13:18
<
PrathamPatel[m] >
Still needed for CXX though
13:18
<
davidlt[m] >
It's not an issue. It's just you couldn't enable them until the main DTS landed :)
13:18
<
davidlt[m] >
But now it's basically to late for the merge window.
13:18
<
PrathamPatel[m] >
That's what I meant, like which drivers are affected from working as intended because of DTs?
13:19
<
davidlt[m] >
It's pretty much common for the drivers to land earlier and then get enabled in N+1 kernel version.
13:19
<
PrathamPatel[m] >
Ah okay
13:19
<
davidlt[m] >
I don't recall, check kernel log on what was merged early. There is 3-4 bits IIRC.
13:20
zsun has joined #fedora-riscv
13:20
<
davidlt[m] >
The more interesting bits are PCIe and DFVS.
13:20
<
davidlt[m] >
QSPI (I don't think it got approval yet).
13:20
<
davidlt[m] >
By default v6.4 will run at 1.0GHz IIRC.
13:20
<
PrathamPatel[m] >
QSPI is merged for 6.4
13:21
<
davidlt[m] >
Oh, cool.
13:21
<
davidlt[m] >
I haven't been looking into the details.
13:21
<
PrathamPatel[m] >
The only thing taking long is u-boot upstreaming
13:21
<
PrathamPatel[m] >
That's stuck on 'under review'
13:21
<
davidlt[m] >
Yeah, but it needs to be aligned with kernel stuff.
13:21
<
davidlt[m] >
E.g. all the DT stuff and compatible strings.
13:22
<
PrathamPatel[m] >
Sorry, I'm new to this low-level area of things 😅
13:22
<
PrathamPatel[m] >
Oh, is that so?
13:22
<
davidlt[m] >
Basically U-Boot updates their DT from the upstream kernel source. They have DTSi (an overlay), which adds additional information.
13:22
<
davidlt[m] >
For example DDR init configuration, etc.
13:23
<
davidlt[m] >
There are other smaller problems.
13:23
<
davidlt[m] >
For example U-Boot does not properly parse riscv isa string.
13:23
<
davidlt[m] >
like _zba_zbb
13:24
<
PrathamPatel[m] >
Ah, okay
14:08
jcajka has quit [Quit: Leaving]
14:20
iooi has quit [Quit: iooi]
15:13
iooi has joined #fedora-riscv
15:54
zsun has quit [Quit: Leaving.]
17:45
iooi has quit [Quit: iooi]
19:02
iooi has joined #fedora-riscv