mmind00 changed the topic of #linux-rockchip to: Rockchip development discussion | public log at https://libera.irclog.whitequark.org/linux-rockchip
<naoki> Kwiboo: May I use "FIT at 2MiB and U-Boot proper at 8MiB" change from your ramboot branch?
<naoki> ^for my DFU in SPL
psydroid has quit [Ping timeout: 252 seconds]
<naoki> oops, it doesn't work with "rockchip: ROCKCHIP_COMMON_STACK_ADDR improvements"...
psydroid has joined #linux-rockchip
stikonas has quit [Quit: Konversation terminated!]
lvrp16 has quit [Ping timeout: 248 seconds]
lvrp16 has joined #linux-rockchip
repk has quit [Ping timeout: 248 seconds]
repk has joined #linux-rockchip
linkmauve has quit [Ping timeout: 248 seconds]
hexdump0815 has quit [Ping timeout: 265 seconds]
hexdump0815 has joined #linux-rockchip
vagrantc has quit [Quit: leaving]
kevery1 has joined #linux-rockchip
kevery has quit [Read error: Connection reset by peer]
kevery1 is now known as kevery
hexdump0815 has quit [Ping timeout: 265 seconds]
Daanct12 has joined #linux-rockchip
System_Error has quit [Remote host closed the connection]
hexdump0815 has joined #linux-rockchip
System_Error has joined #linux-rockchip
System_Error has quit [Remote host closed the connection]
System_Error has joined #linux-rockchip
System_Error has quit [Remote host closed the connection]
System_Error has joined #linux-rockchip
kevery1 has joined #linux-rockchip
kevery has quit [Ping timeout: 244 seconds]
kevery1 is now known as kevery
franoosh has joined #linux-rockchip
ldevulder has joined #linux-rockchip
raster has joined #linux-rockchip
naoki has quit [Quit: naoki]
naoki1 has joined #linux-rockchip
naoki1 is now known as naoki
franoosh has quit [Remote host closed the connection]
franoosh has joined #linux-rockchip
kevery1 has joined #linux-rockchip
kevery has quit [Ping timeout: 248 seconds]
kevery1 is now known as kevery
eballetbo has joined #linux-rockchip
System_Error has quit [Ping timeout: 264 seconds]
System_Error has joined #linux-rockchip
linkmauve has joined #linux-rockchip
paulk-bis has quit [Quit: WeeChat 3.0]
paulk has joined #linux-rockchip
<Kwiboo> naoki: yes, I pushed an updated version at https://github.com/Kwiboo/u-boot-rockchip/commits/rk3xxx-2025.04 before I went to sleep, did not manage to get the ramboot out before that, want to test on an older armv7 soc before I send it out, should probably hit ML later today :-)
franoosh has quit [Ping timeout: 252 seconds]
<naoki> Maybe I should check with older SoCs too. I'm sure there's a Radxa ROCK (RK3188) and a ROCK2 (RK3288) somewhere. ;)
<diederik> naoki: Please document the *why* in your patch descriptions (like Kwiboo did)
<naoki> hmm
<diederik> What you put in the series cover letter does NOT get commited to git (and each commit should describe why that's an improvement)
<Kwiboo> naoki: ahh, you already sent it out, will take a look later, yesterday I spent too much time trying to unbrick my rk3228a box that I managed to brick just 1-2 hours after it arrived the other day :-)
<naoki> I once bricked something within minutes of opening the box...
<wens> at least it didn't go up in a puff of smoke ...
paulk has quit [Ping timeout: 268 seconds]
paulk has joined #linux-rockchip
paulk has joined #linux-rockchip
sfo has quit [Remote host closed the connection]
fleg has quit [Write error: Broken pipe]
<Kwiboo> hehe, my box did make some funny noises when I tried to ground different pins/resistors, but the box seem to have survived my trail-and-error and I finally got it to go to maskrom mode
<Kwiboo> at least I finally have a rockchip soc that pre-dates the rk3288 :-)
<dsimic> the collection is now complete :)
<robmur01> Nah, 3228 is still newer than 3288 - older is the Cortex-A9 stuff ;)
<maz> PL310 FTW!
sfo has joined #linux-rockchip
fleg has joined #linux-rockchip
franoosh has joined #linux-rockchip
naoki has quit [Ping timeout: 252 seconds]
<Kwiboo> robmur01: hehe, true, more correct may have been to say it is my first older rk soc using cortex-a7 :-)
System_Error has quit [Remote host closed the connection]
System_Error has joined #linux-rockchip
<tlwoerner> dsimic: thanks for the review. interestingly enough i did notice the exact same discrepancy that you mention in your comments
<tlwoerner> however, since the rk3328 table was obviously using 4096-<table value> i simply continued with the same pattern
psydroid2 has joined #linux-rockchip
psydroid2 has quit [Excess Flood]
<dsimic> tlwoerner: anytime... yes, something strange is going on, and I suspect one of the TRMs to be wrong
<dsimic> because 1024 as the "offset" value is simply too small... we'll see, I'll do some more testing, and the patch is fine in the meantime
<dsimic> qschulz: Kwiboo: tlwoerner: good news, Radxa responed with "just send the list of boards you need"... woohoo! :)
* qschulz prepares his shopping list
<tlwoerner> dsimic: nice!
<tlwoerner> speaking of Radxa... when yocto switched from the 6.6.y kernel to the 6.12.x kernel, i lost PCIe on my rock-pi-4a
<tlwoerner> i have the radxa penta-sata hat on a rock-pi-4a and it stopped working with the update
<dsimic> tlwoerner: I saw people complaining about that, and I tried going through the offending patches a few times, to no avail
<dsimic> any chances, please, to send a complete kernel log from the Penta-hat-equipped board when it has no PCIe working?
<tlwoerner> someone else reported something similar on the mailing list in Jan also with an rk3399-based device, but that fix was specific to the dts of their specific device (rk3399-gru-chromebook)
<qschulz> tlwoerner: yeah that was a mistake when renaming DT nodes IIRC
<dsimic> yes, I saw that back then
<tlwoerner> okay, i'll grab a log at some point. it's my only setup, and it's a "production" device
<dsimic> qschulz: I'm so happy that I'll get the boards needed for the SoC binning project... and of course, for the RK3308 thermal stuff :)
<tlwoerner> dsimic: my point being, ask for a penta-sata hat too! :-)
<dsimic> tlwoerner: I see, we'll need that log to be able to investigate further
<tlwoerner> (so you can test PCIe stuff)
<dsimic> tlwoerner: hmm, not a bad idea at all... maybe not a Penta hat, but an M.2 extender hat, which is smaller but good enough for testing
naoki has joined #linux-rockchip
Daanct12 has quit [Quit: WeeChat 4.5.1]
naoki has quit [Client Quit]
raster has quit [Quit: Gettin' stinky!]
<Kwiboo> does rk3399 also have pcie issue on longterm kernel?, I do know that rk356x has broken pcie, sata and usb3 on v6.1.123+, v6.6.69+ and v6.12.8+, but that combphy issue should not affect rk3399
<qschulz> Kwiboo: as far as I've been told PCIe is simply broken inside RK3399 silicon, only PCIe gen1 is stable., Don't know how much of that is truth though
<jakllsch> i mean, it's got quirks, but it's not broken..
<jakllsch> oh, i guess it was broken for a few kernel versions..
<robmur01> we should really be better at differentiating "kernel version" and "DTB version"... :P
<Kwiboo> qschulz: look like rk3399-gru still is a little bit broken in torvalds tree after regulator node rename, I can still see aliases being added to the empty pp3300 node name, other boards was also affected of same or different issue?
<qschulz> Kwiboo: "aliases being added to the empty pp3300 node name"?
<qschulz> (I only followed from far, I sadly don't have much time to spend on kernel dev aside from DT stuff for our products :/)
<Kwiboo> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arch/arm64/boot/dts/rockchip/rk3399-gru-scarlet.dtsi#n112 <<-- here is one example, pp3300_s0 is being added as an alias to the pp3300 regulator, now that pp3300 node no longer exists, mening the alias point to an empty node
<Kwiboo> guessing one of those alias/label could be used elsewhere in a supply prop or similar and no longer be a phandle to the regulator
<qschulz> Kwiboo: https://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip.git/commit/?h=v6.14-armsoc/dtsfixes&id=2f9eb5262e63396a315c7da34a6c80c5d335df9f
<Kwiboo> great, that is probably a commit that should be pick for u-boot v2025.04 to fix current broken v6.13 DTs for gru
franoosh has quit [Ping timeout: 252 seconds]
<qschulz> Kwiboo: not yet in devicetree-rebasing though :/
<qschulz> mmind00: would you have some idea on when https://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip.git/commit/?h=v6.14-armsoc/dtsfixes&id=2f9eb5262e63396a315c7da34a6c80c5d335df9f would land into an rc of 6.14?
<qschulz> Kwiboo: rc2 released in U-Boot... wondering if we have time to wait for this patch to land in devicetree-rebasing or if we need to go the hackish way and fix it with /delete-node/ in rk3399-whatever-u-boot.dtsi
<Kwiboo> not fully sure the impact, but at least one of those labels not pointing to a 1v8 regulator are used for io-domain, so io-domain driver in u-boot may no longer change io-domain voltage
<mmind00> qschulz: with the amount of "interesting" fixes in that branch, and the two still on the list, I guess I'll do a tag+pull to armsoc this week ... in in master some time next week maybe?
<qschulz> mmind00: ok that'd work out for U-Boot 2025.04 then :)
<tlwoerner> wrt PCIe: yes, gru got fixed, but rock-pi-4a is still broken on the 6.12.11 kernel. 6.6.74 works for me
<qschulz> tlwoerner: would help tremendously if you can do a git bisect I guess
<qschulz> but that's "a few" commits to bisect :)
<tlwoerner> my painfully slow build machine is still churning out a 6.12.y build for me :-)
<qschulz> tlwoerner: is there any change in DT between 6.6.74 and 6.12.11?
<tlwoerner> bisecting along 2 separate branches is not easy (or sometimes not possible?)
<qschulz> decompile the DT from both and diff them to check
<tlwoerner> qschulz: yes, lots, everywhere
<qschulz> tlwoerner: oof :/
<qschulz> tlwoerner: let git bisect figure things out for you?
<tlwoerner> i'll try bisecting. good thing i got rauc working :-)
<qschulz> Another test could be to use 6.6 DT with 6.12 kernel
<tlwoerner> okay, i can try that
<qschulz> this is something that we should actively support
<qschulz> if not, then we need to fix it anyway
<qschulz> but at least it could indicate it's not necessarily a driver issue
<tlwoerner> qschulz: by the way, any comment on the meta-rockchip patch to add Jonas' therm monitoring for rk3308?
<qschulz> tlwoerner: yes, don't :D ?
<tlwoerner> no prob. i can carry it in my project-specific layers
<qschulz> At least have some patch sent upstream instead of just "pending"
<qschulz> but I don't think it's any hackier than when we had a linux-next recipe
<qschulz> so up to you
<qschulz> tlwoerner: BTW, I had a quick look at using this ddrbin_tool.py and I'm hesitating on the implementation a bit
<qschulz> The script takes a conf file as input
<qschulz> I can either have bitbake variables and generate the conf file myself
<qschulz> very bitbakey, i like it, but limits what you can put in this conf file (and there are MANY options)
<qschulz> also kinda user-friendly
<qschulz> the other option is rely on the user passing a conf file (via SRC_URI)
<qschulz> this I like because I have nothing to do except calling the script on it
<tlwoerner> support both workflows?
<qschulz> oof
<qschulz> what happens if I have a file AND a variable that conflict?
<tlwoerner> they yocto kernel tooling allows users to use the in-kernel defconfigs or specify their own
<tlwoerner> if the user says "used the in-kernel" but also provides a SRC_URI:defconfig it warns and uses the ... defconfig (?i think?)
<qschulz> yeah but they are all pointing at a defconfig
<tlwoerner> in any case, implement the one you like best
<qschulz> we don't autogenerate the defconfig based on some variable
<tlwoerner> something is better than nothing
<qschulz> e.g. we don't have KERNEL_WIFI ?= "1" means I enable all the wifi drivers in defconfig
<qschulz> yeah, and we can always extend it later on, or simplify it
<tlwoerner> yep
<qschulz> let's do instead of think eh
<qschulz> or rather "overthink"
<qschulz> thinking is nice, overthinking not
<tlwoerner> it's a balance
<qschulz> and my brain clearly doesn't agree with that :D
raster has joined #linux-rockchip
System_Error has quit [Remote host closed the connection]
System_Error has joined #linux-rockchip
dsimic has quit [Ping timeout: 252 seconds]
dsimic has joined #linux-rockchip
franoosh has joined #linux-rockchip
<tlwoerner> ooh, the pcie driver is crashing during probe on my 6.12.11 kernel
<tlwoerner> do you want to see the working 6.6.74 log?
<qschulz> tlwoerner: this is 6.12.11 kernel + 6.12.11 dt?
<tlwoerner> yes
<qschulz> (I'm not planning on looking into it, but considering what we discussed before, not giving the information may mislead other people)
erg_ has joined #linux-rockchip
franoosh has quit [Read error: Connection reset by peer]
<qschulz> tlwoerner: there doesn't seem to be anything out of the ordinary between v6.6 and v6.12, and v6.12 and master for the pcie host controller
<robmur01> tlwoerner: what's the actual oops/panic there? (the splat seems to be missing its head)
<robmur01> if it's an SError while probing config space (as I suspect), it may well be a matter of timing relative to regulators and resets
<tlwoerner> robmur01: i don't look at oops/panics often, but i was thinking the exact same thing (that a line or two was missing)
<tlwoerner> that's a copy of /var/log/messages from that boot
<tlwoerner> it looks like it's missing, i could reboot and try again
<robmur01> at worst you can likely bodge it with a massive regulator-enable-ramp-delay on vpcie3v3-supply, like gru-scarlet does
macromorgan has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
macromorgan has joined #linux-rockchip
System_Error has quit [Remote host closed the connection]
vagrantc has joined #linux-rockchip
stikonas has joined #linux-rockchip
System_Error has joined #linux-rockchip
raster has quit [Quit: Gettin' stinky!]
stikonas has quit [Ping timeout: 276 seconds]
stikonas has joined #linux-rockchip
naoki has joined #linux-rockchip
stikonas has quit [Quit: Konversation terminated!]
stikonas has joined #linux-rockchip
erg_ has quit [Remote host closed the connection]
kevery1 has joined #linux-rockchip
kevery has quit [Ping timeout: 248 seconds]
kevery1 is now known as kevery
<Kwiboo> naoki: I have pushed a small update for ramboot to https://github.com/Kwiboo/u-boot-rockchip/commits/rk3xxx-2025.04/ tested on rk322x and had to moved the RAM payload offset to speed up loading on the older cortex-a7 socs
<Kwiboo> will send patches tomorrow once I have tested on rk3576