ChanServ changed the topic of #armlinux to: ARM kernel talk [Upstream kernel, find your vendor forums for questions about their kernels] | https://libera.irclog.whitequark.org/armlinux
scosu has joined #armlinux
scosu_ has quit [Ping timeout: 272 seconds]
apritzel has quit [Quit: Leaving]
macromorgan has joined #armlinux
Nact has quit [Quit: Konversation terminated!]
amitk has joined #armlinux
CrashTestDummy2 has joined #armlinux
CrashTestDummy has quit [Ping timeout: 272 seconds]
CrashTestDummy3 has joined #armlinux
CrashTestDummy2 has quit [Ping timeout: 250 seconds]
frieder has joined #armlinux
frieder has quit [Ping timeout: 272 seconds]
CrashTestDummy2 has joined #armlinux
CrashTestDummy3 has quit [Ping timeout: 258 seconds]
sszy has joined #armlinux
guillaume_g has joined #armlinux
CrashTestDummy has joined #armlinux
Pali has joined #armlinux
CrashTestDummy2 has quit [Ping timeout: 265 seconds]
iivanov has joined #armlinux
frieder has joined #armlinux
matthias_bgg has joined #armlinux
<Misanthropos> ukleinek, I tried the stock kernel 5.14-rc3 no patches just config to OXNAS and no v6/v7 but OX810SE enabled
<ukleinek> Misanthropos: and which device tree?
<Misanthropos> ukleinek, none
<ukleinek> Misanthropos: there is your problem
<Misanthropos> well I used the def config you mentioned
<ukleinek> this is correct, but not enough to yield a booting kernel
<Misanthropos> so which dt should I use and where do I get it?
<arnd> I see that openwrt only includes dts files for ox820 devices not ox810
<ukleinek> Misanthropos: give me a few minutes, I'll make a patch for you
<ukleinek> hmm, if ox810 doesn't have dt support those few minutes won't do it
<arnd> ox810se-wd-mbwe.dts in the mainline kernel is the only ox810 based one that I can find
<ukleinek> ah right, that's good enough
<Misanthropos> thats how generated uImage: LOADADDR=0x48008000 make ARCH=arm CROSS_COMPILE=armv5te-softfloat-linux-gnueabi- uImage
<ukleinek> Misanthropos: that gives you an uImage without dt though
<Misanthropos> I have to read someting about that because I have no clue about what and why this dt is good for
<Xogium> er well
<Misanthropos> well I have a clue about it describing registers
<ukleinek> Misanthropos: the dt is hardware description. It tells the kernel where to find what kind of device
<Xogium> it describes the hardware so linux and u-boot knows what kind of hardware you even have
<Xogium> there is no bios on arm to do that kind of stuff
<Xogium> arm is kinda more or less bare metal bootloader
<Xogium> a tiny rom before, but that's about it
<Misanthropos> ok - thanks. Now I know why I need it.
<Xogium> sometimes I wish there were actual powerhouses like an equivalent or better of the ryzen but for arm
<Xogium> I had to blindly try to turn off secure boot in my stupid x86_64 bios a few days ago. It was not fun, at all
<ukleinek> Xogium: why do you need to do that, today's Linux should cope well enough with secure boot?!
<Xogium> because archlinux does not
<Xogium> :p
<ukleinek> Xogium: I spotted your error :-)
<Xogium> for some reason, secure boot was added in their iso back in 2013, then removed in 2016, and was never added again ever since
<Misanthropos> so could I generate a uImage with dt using make?
<Misanthropos> or do I have to do your steps from abover cat.. mkimage?
<ukleinek> Misanthropos: the kernel Makefile doesn't support that, so you have to do it manually
<Misanthropos> and it has to be a .dtb file?
<ukleinek> Misanthropos: https://paste.debian.net/1206310/
<ukleinek> oh no, this is wrong, do s/ttyS1/ttyS0/ and s/uart1/uart0/ on that one
<ukleinek> ah, and s/serial1/serial0/
<Misanthropos> any more :D?
<ukleinek> Misanthropos: I think that's it
<ukleinek> Misanthropos: than make should give you arch/arm/boot/dts/ox810se-zyxel-nsa210.dtb that you can append after your zImage
<Misanthropos> ok got it
<Misanthropos> are you sure I need zImage not Image?
<Misanthropos> Image Name:
<Misanthropos> Load Address: 48008000
<Misanthropos> Image Type: ARM Linux Kernel Image (uncompressed)
<Misanthropos> Created: Mon Aug 2 10:37:53 2021
<Misanthropos> Data Size: 3889118 Bytes = 3797.97 KiB = 3.71 MiB
<Misanthropos> Entry Point: 48008000
<ukleinek> positive: use zImage
<ukleinek> Misanthropos: did you see my query?
<Misanthropos> should the uImage be compressed?
<Xogium> depends if your bootloader supports that, I'd think. But, probably not
<Misanthropos> right. I just checked what make produced and thats uncompressed either
<ukleinek> Misanthropos: the easiest approach is to use zImage and claim that to be uncompressed for U-Boot
geertu has quit [Quit: leaving]
* ukleinek grumbles about register defines without prefix. Today: UART_TX
wwilly has quit [Remote host closed the connection]
frieder has quit [Ping timeout: 272 seconds]
frieder has joined #armlinux
shailangsa has quit [Ping timeout: 268 seconds]
Nact has joined #armlinux
iivanov has left #armlinux [Leaving...]
iivanov has joined #armlinux
wwilly has joined #armlinux
shailangsa has joined #armlinux
frieder has quit [Ping timeout: 256 seconds]
frieder has joined #armlinux
<arnd> everything merged into soc.git, geertu wins the race for the earliest submission of v5.15 content
<ukleinek> ..ooOO(Is this a content, can I already today lock for v5.16? :-)
<ukleinek> s/content/contest/
<arnd> ukleinek: there was one submission left from before the merge window that didn't make it in at the time, and I did not count that here but considered it the last submission for 5.14
<arnd> so if you want to compete, maybe write a script to send your submissions once -rc1 is tagged
<ukleinek> arnd: the challenge for me will be to become a platform maintainer until v5.16-rc1 :-)
<arnd> I think we have a couple of orphaned platforms, you could take over vt8500 if you like
<ukleinek> without having looked at the platform details: I don't
<arnd> or oxnas ;-)
<ukleinek> arnd: I shortly considered that, but I don't have such a platform
jlinton has joined #armlinux
macromorgan has quit [Read error: Connection reset by peer]
<Misanthropos> ukleinek, you can have mine :-D
sszy has quit [Ping timeout: 272 seconds]
<ukleinek> Misanthropos: you're giving up?
<maz> I think PXA could do with some lovin'...
macromorgan has joined #armlinux
frieder has quit [Remote host closed the connection]
<ukleinek> arnd: do you see something obvious why we're out of memory here?: http://sprunge.us/hq1kco
<ukleinek> arnd: "19020K reserved" looks too much, but that should not explain why there isn't enough space to squeeze in a sh
<ukleinek> ah, this isn't outofmem any more
<ukleinek> it dies with sigill, which means wrong binary format maybe?
<ukleinek> Misanthropos:
<ukleinek> ^
<arnd> ukleinek: is this the oxnas machine?
<Misanthropos> arnd, yes
<arnd> which user space is it? debian armel or the original OS?
<ukleinek> arnd: an initrd from the debian installer with some components stripped to make it smaller
<arnd> ok, and it's the armel version, not armhf, right?
<ukleinek> right
<ukleinek> kirkwood is armv5te, too, isn't it?
<arnd> yes, but it's not an arm926
<ukleinek> I would matter that the busybox binary and the libs are the same still
<ukleinek> s/matter/expect/
* ukleinek cannot type today
<arnd> maybe try the debian stable version? I wonder if a bug could have made it into testing unnoticed. It's all built on armv8 hardware now, so there is a small chance that some config script ended up building user code for something later than arm926-softmmu
<arnd> or maybe you stripped out too much and it's missing some library
* ukleinek is positive to not have stripped a dependency of /bin/sh
<arnd> ok
<ukleinek> mostly lib/modules, var/lib/dpkg and things that obviously belong to the installer
<Misanthropos> ukleinek, ok on it
<ukleinek> if you want do the following before mkimage wrapping: zcat initrd.gz | cpio-filter --exclude lib/modules | gzip -c > initrd-stripped.gz
<ukleinek> arnd: if busybox is indeed broken in testing for armel, this means we maybe should have dropped it ...
<ukleinek> well, or people only update after we call it stable
<arnd> ukleinek: it probably isn't actually broken, that was just the first explanation that I could think of
<ukleinek> is there an easy way to make the oops more helpful, e.g. dump the instruction that resulted in the signal?
<ukleinek> there is CONFIG_VFP=y, can this hurt?
<ukleinek> CONFIG_AEABI is not set
<ukleinek> Misanthropos: please enable that one
* ukleinek wonders if there is still anybody out there not using AEABI
<arnd> I think rmk has some strongarm machines on oabi
headless has joined #armlinux
wwilly has quit [Remote host closed the connection]
matthias_bgg has quit [Quit: Leaving]
wwilly has joined #armlinux
jlinton has quit [Quit: Client closed]
torez has joined #armlinux
<Misanthropos> so after lots of unfruitful trying something came up by accident: if booted the linux in the flasrom and reboot. the u-boot procedure tftpboot tftboot 0xaddress ramdisk and then bootm 0xkernel 0xramdisk goes through.. but only then
<Misanthropos> my guess is some hardware component got initialized by the OEM firmware which still is active when booting into our kernel
<Misanthropos> otherwise the system hangs at /init
nsaenz has quit [Remote host closed the connection]
_whitelogger has joined #armlinux
pg12_ has joined #armlinux
pg12 has quit [Ping timeout: 252 seconds]
bps has quit [Ping timeout: 252 seconds]
<Misanthropos> so the question is: what could be missing for init to continue?
headless has quit [Quit: Konversation terminated!]
torez has quit [Quit: torez]
guillaume_g has quit [Quit: Konversation terminated!]
nsaenz has joined #armlinux
iivanov has quit [Remote host closed the connection]
iivanov has joined #armlinux
iivanov has quit [Ping timeout: 272 seconds]
wwilly has quit [Ping timeout: 272 seconds]
shailangsa has quit [Ping timeout: 272 seconds]
Pali has quit [Ping timeout: 272 seconds]