narmstrong changed the topic of #linux-amlogic to: Amlogic mainline kernel development discussion - our wiki http://linux-meson.com/ - ml linux-amlogic@lists.infradead.org - Publicly Logged on https://irclog.whitequark.org/linux-amlogic
<lvrp16> Lyude: i get that message normally and I am able to get past it. what's ur boot args?
<lvrp16> Lyude: it could be the kernel stops passing information to that UART
<Lyude> lvrp16: oh hey! you're the open suse person
<Lyude> BOOT_IMAGE=/vmlinuz-4.19.0-rc6Lyude-Test-00321-gc1d84a1b42ef root=UUID=9a926ac9-5938-4539-bb79-72f87b36095f ro LANG=en_US.UTF-8
<lvrp16> huh?
<lvrp16> opensuse?
<lvrp16> not affiliated in any way
<Lyude> lvrp16: oh, whoops, must be mistaking you with someone else. thought I saw your nickname on an opensuse blog at some point
<Lyude> anyway: i'm fairly sure it's not the boot args, I've got another kernel using chewitt's default config working on here
<lvrp16> oh no, you need to add the /dev/ttyAML0
<Lyude> lvrp16: hm?
<lvrp16> console=ttyAML0,115200n8
<Lyude> nope, same thing. it should be noted I do actually get console output with this, and it finds ttyAML0 automatically
<Lyude> there is something I'm very suspecious of that I only noticed a moment ago
<Lyude> at 3.447171 look at the line below it, isn't that from the bootloader?
<lvrp16> [ 3.517152] mmc1: mmc_select_hs400 failed, error -84
<lvrp16> [ 3.526513] mmc1: error -84 whilst initialising MMC card
<lvrp16> seems like it didn't tune for hs400
<lvrp16> is the os on emmc or sd card?
<Lyude> lvrp16: sd card, but I also get that same message on the kernel that works
<Lyude> lvrp16: jfyi, this is the starting dmesg from the working kernel: https://paste.fedoraproject.org/paste/leYdALjZQPUOEUDjKeGiQQ
<lvrp16> compiled with the same config?
<lvrp16> rc7 broke something?
<Lyude> lvrp16: well the difference in configs is what I'm trying to figure out (also, I've unfortunately tried other rc versions with the same results)
<Lyude> hold on, let me post both kernel configs
<lvrp16> config or dmesg?
<Lyude> https://paste.fedoraproject.org/paste/bYpjal7K90qfJRByO0tHJQ - working kernel config from chewitt
<Lyude> oh! whoops
<Lyude> lvrp16: sorry, I meant to post this as the config that's not working https://paste.fedoraproject.org/paste/YUX5pKow2VPHrG8Nduvovw
<Lyude> as a note: the reason I'm trying to figure this out is so I can figure out what we need to add to fedora's default kernel config so it boots by default on these boards
<Lyude> starting to wonder if scpi has anything to do with this
<lvrp16> wow there's a lot of config differences, I would try enabling all the MESON and SCPI first and then doing bisects, don't you have a threadripper?
<lvrp16> XD
<Lyude> no you're thinking of hdkr
<lvrp16> oh
<lvrp16> borrow his machine :P
<Lyude> heh, I've still got a fast machine :P, just a little older
<lvrp16> i'm using a two core 6600u laptop...it takes me 2 hours to build a desktop image and 8 hours to build retropie...
<lvrp16> i feel like my 3d printer is faster than my computer
<Lyude> lvrp16: also just checked, I do have everything it lets me enable for meson enabled unfortunately
<Lyude> well minus sound, but I doubt that matters
<Lyude> lvrp16: do you know what the latest stable kernel that is supposed to work with this board is, 4.18?
<lvrp16> Lyude: I don't have any Khadas VIM boards. I couldn't tell you. They use an external MCU to control stuff so it's a little different than a typical S905X or S912
<lvrp16> loic from open suse might be a better resource since he has boards
<Lyude> h m
<Lyude> *hm
<Lyude> i'm almost convinced that BL31 message we're seeing is the board getting shut off
<chewitt> Lyude: you will need most (so probably easier to just grab all) of the Linux patches in that dir
<Lyude> chewitt: alright; i'm going to give that a shot. that still doesn't entirely explain why the other config you gave me worked though
<Lyude> i have a 4.18 kernel i'm about to try though, if that doesn't work (and I'm expecting it won't...) I'll go back and recompile with those patches and see what difference it makes
<Lyude> that being said: the weird [BL31]: tee size: 0 <-- that is coming up in that dmesg is definitely from the bootloader
<chewitt> i've only been using my images from SD cards not emmc
<Lyude> I mean so have I, the only thing I've done with the emmc is update uboot so I could get fedora loaded
<Lyude> but that message comes up half way through the kernel boot on the non-functional kernel, a couple lines before the kernel stops booting and the led turns odff
<Lyude> *off
<chewitt> that patch has been in my patch-set for a long time
<chewitt> (it's gone upstream and should appear in 4.20)
<chewitt> so unlikely to be the cause of problems
<Lyude> alright, adding all those patches in
<chewitt> the only one I haven't pushed to that branch yet is the patchwork link, which fixes (or works-around) a reboot problem on VIM2
<Lyude> that is definitely a thing I've hit
vdehors has quit [Quit: No Ping reply in 180 seconds.]
<chewitt> it was annoying the bejesus out of me for a while
<chewitt> our update process requires two reboots; one worked, one failed
<chewitt> each fail requires me to get up and walk across the room to push the reset button
<Lyude> hehehe, I have been doing that since I got this board
vdehors has joined #linux-amlogic
<chewitt> btw, that patch set is adapted in a couple of places for 4.19-rc7 so might not apply clean to 4.18
<Lyude> that's alright; 4.18 didn't work so I just went back to rc7
<Lyude> i had just been trying 4.18 to make sure I wasn't hitting some regression somewhere
<chewitt> ^ whack that in as well
<chewitt> I'm still not sure what that's for, but Oleg (who does the Armbian images) keeps adding it
<Lyude> 0x01080000 is the start address for the kernel i'm pretty sure
<chewitt> it's not needed in my images, but then i'm relying on ye olde 2015 vendor u-boot
<Lyude> alright; had to fix some merge conflicts but i've got a kernel with those patches built now
vagrantc has quit [Quit: leaving]
<Lyude> chewitt: same bug as before
<Lyude> this is immensely bizarre
<chewitt> so .. worth mentioning
<chewitt> in my images, the uart output stops mid-boot
<chewitt> but the kernel continues to boot
<Lyude> that's weird; on my working kernel it works the whole way through
<chewitt> [ 3.430992] Freeing unused kernel memory: 448K
<chewitt> ^ around that point in boot
<chewitt> i've no idea why, but our init script is a bit homebrew and possibly/probably the reason
<Lyude> chewitt: the thing is I'm quite sure the system just stops booting at this point, I don't know why the LED would go off otherwise
<chewitt> hmm
<Lyude> it honestly seems like somehow the CPU jumps to somewhere in the bootloader/early boot firmware, but I have no idea why
<Lyude> that's the only explanation I can think of which would explain why [BL31]: tee size: 0
<Lyude> comes up right before the kernel hangs, despite the fact we've already gone from uboot->grub->kernel
<Lyude> on the working kernel I see it output some information about dvfs being setup there instead
<chewitt> sadly.. boot stuff is one of my less experienced areas of knowledge
<chewitt> I normally poke narmstrong and co for inspiration on that stuff
<Lyude> narmstrong: ^^
<chewitt> it was the Baylibre 5th Birthday party last night so they might be a little slow on the uptake today
<Lyude> chewitt: well at the very least, those patches for the dtb made the emmc finally come up on my working kernel :)
<chewitt> seems to be a trend with Amlogic boards
<chewitt> people optimistically copy the spec and have the sdcard node supporting the highest-speeds
<chewitt> then I have to patch it back to reality or nothing boots
<chewitt> or there's some clock foo missing somewhere
_whitelogger has joined #linux-amlogic
chewitt has quit [Ping timeout: 246 seconds]
Darkmatter66_ has quit [Quit: ZNC 1.7.1 - https://znc.in]
Darkmatter66 has joined #linux-amlogic
pionen has joined #linux-amlogic
pionen has quit [Quit: leaving]
The_Coolest has quit [Read error: Connection reset by peer]
The_Coolest has joined #linux-amlogic
The_CooIest has joined #linux-amlogic
<narmstrong> Lyude: did it finally boot ?
<narmstrong> [BL31]: tee size: 0 is from the ATF bl31 resident code for the nvmem fuse management calls
<narmstrong> Nothing alarming
<narmstrong> I never booted the vim2 with the kernel as an EFI stub, this may lead to other issues
Barada has joined #linux-amlogic
Barada has quit [Client Quit]
Darkmatter66 has quit [Quit: ZNC 1.7.1 - https://znc.in]
_whitelogger has joined #linux-amlogic
_whitelogger has joined #linux-amlogic
xdarklight_ is now known as xdarklight
return0xe has quit [Remote host closed the connection]
Darkmatter66 has joined #linux-amlogic
return0e has joined #linux-amlogic
_whitelogger has joined #linux-amlogic
sputnik_ has quit [Read error: Connection reset by peer]
sputnik_ has joined #linux-amlogic
_whitelogger has joined #linux-amlogic
edcragg has quit [Quit: ZNC - http://znc.in]
edcragg has joined #linux-amlogic
<narmstrong> Lyude: the vim2 is boot tested on kernelci with defconfig, it’s worth trying !
Ivanovic has quit [Quit: Caught sigterm, terminating...]
Ivanovic has joined #linux-amlogic
Ivanovic has quit [Client Quit]
The_Coolest has quit [Quit: FOLD For The Cause!!!]
Ivanovic has joined #linux-amlogic
Ivanovic has quit [Client Quit]
Ivanovic has joined #linux-amlogic
Ivanovic has quit [Quit: Caught sigterm, terminating...]
Ivanovic has joined #linux-amlogic
Ivanovic_ has joined #linux-amlogic
Ivanovic has quit [Ping timeout: 252 seconds]
Ivanovic_ has quit [Client Quit]
Ivanovic has joined #linux-amlogic
Ivanovic has quit [Read error: No route to host]
Ivanovic has joined #linux-amlogic
sputnik__ has joined #linux-amlogic
sputnik_ has quit [Ping timeout: 252 seconds]
sputnik_ has joined #linux-amlogic
sputnik__ has quit [Ping timeout: 260 seconds]
<Lyude> narmstrong: alright, huh
<Lyude> I don't think it's the efi stub though, I'm using chewitt's working config for now since it seems like I can just enable the efi stuff on there and it works fine
<Lyude> also: does anyone have a workaround for the board not rebooting properly
AntonioND has joined #linux-amlogic
return0e has quit [Remote host closed the connection]
<narmstrong> Yep chewitt found a fix recently for that
<Lyude> narmstrong: awesome! I'll give it a shot in just a little bit
return0e has joined #linux-amlogic
jakogut has joined #linux-amlogic
<Lyude> narmstrong: btw, regarding the SCPI cpufreq backtraces with the kernel that I was seeing on the vim2 the other day, do you know if the workaround of disabling scpi actually has an impact on performance?
<xdarklight> Lyude: yes, changing the CPU frequency requires SCPI to work (CPU frequency changes are set via the SCPI protocol)
<Lyude> xdarklight: aw, guessing there aren't any workarounds for this yet/
<Lyude> *?
<narmstrong> Lyude: if you disable it you will stay at full cpu freq
<Lyude> oh
<Lyude> that's exactly what I would want :)
<narmstrong> Setting the performance cpufreq governor does the same, it leaves the max freq as boot default
<Lyude> narmstrong: jfyi; the config option I want to disable is ARM_SCPI_CPUFREQ right?
random_yanek has quit [Ping timeout: 252 seconds]
random_yanek has joined #linux-amlogic
<AntonioND> there is no way I can get timer interrupts
<AntonioND> there is something preventing them from reaching the gic, I think
<AntonioND> if I set the ISPENDR register bits manually I can trigger any SPI I want
<AntonioND> but I've enabled all timers, they are counting down, and I don't get any interrupt
<AntonioND> or any bit set in ISPENDR
<AntonioND> so either there is another register I have to set up, or there is another problem I don't know about
Darkmatter66_ has joined #linux-amlogic
<AntonioND> has anyone actually got timer interrupts working in the odroid C2? the timers are counting just fine, but no interrupts at all
Darkmatter66 has quit [Ping timeout: 276 seconds]
<Lyude> narmstrong: hm. those patches for the reboot issue don't seem to have changed anything on my board
<Lyude> also: does anyone know if the vim2 has an iommu?