<Entei[m]>
<khem> "one of the valid tunes in oe-..." <- Yes I used that. Maybe I am misunderstanding where to define the tune though.
<Entei[m]>
In my `build/conf/local.conf` I have `MACHINE = "qemuriscv64" `like before, and in the openembedded core directory `meta/conf/machine/include/riscv/arch-riscv.inc` I added the line the `DEFAULTTUNE = "riscv64nc" ` . In `meta/conf/machine/include/riscv/tune-riscv.inc`, I added `riscv64nc` to `AVAILTUNES` variable.
kscherer has quit [Quit: Konversation terminated!]
davidinux has quit [Ping timeout: 246 seconds]
davidinux has joined #yocto
jimbo has quit [Quit: Going offline, see ya! (www.adiirc.com)]
Fanfwe has quit [Ping timeout: 246 seconds]
abelloni has quit [Ping timeout: 246 seconds]
abelloni has joined #yocto
Fanfwe has joined #yocto
starblue has quit [Ping timeout: 248 seconds]
starblue has joined #yocto
sakoman has joined #yocto
nemik has quit [Ping timeout: 255 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 268 seconds]
nemik has joined #yocto
jclsn has quit [Ping timeout: 255 seconds]
jclsn has joined #yocto
Thorn has joined #yocto
Fanfwe has quit [Ping timeout: 268 seconds]
Fanfwe has joined #yocto
amitk has joined #yocto
camus has quit [Ping timeout: 246 seconds]
camus has joined #yocto
sakoman has quit [Quit: Leaving.]
rob_w has joined #yocto
aak-rookie has joined #yocto
Thorn has quit [Ping timeout: 255 seconds]
<aak-rookie>
hi, im working with imx8mq(coral-dev-board). i can run image in SD mode, but when i switch it to EMMC mode, i get an error. log: https://pastebin.com/tt8T39zY
<aak-rookie>
i tried to indicate the important places by putting "----".
olani- has quit [Remote host closed the connection]
florian has joined #yocto
olani- has joined #yocto
<Entei[m]>
> <@dai_enkai:matrix.org> Yes I used that. Maybe I am misunderstanding where to define the tune though.
<Entei[m]>
> In my `build/conf/local.conf` I have `MACHINE = "qemuriscv64" `like before, and in the openembedded core directory `meta/conf/machine/include/riscv/arch-riscv.inc` I added the line the `DEFAULTTUNE = "riscv64nc" ` . In `meta/conf/machine/include/riscv/tune-riscv.inc`, I added `riscv64nc` to `AVAILTUNES` variable.
<Entei[m]>
Now I even tried making making a custom machine config called `qemuriscv64nc.conf` which uses the same config as `qemuriscv64.conf` but adds `DEFAULTTUNE= "riscv64nc"`. However it gives the error that could not find BSP definition of qemuriscv64nc/standard.
<LetoThe2nd>
Entei[m]: usually creating a custom machine that requires the original one and then overrides is a good approach.
leon-anavi has joined #yocto
amitk_ has quit [Ping timeout: 248 seconds]
<LetoThe2nd>
Entei[m]: modifying anything in an upstream layer is a bad idea, though. so i'd say 1) create your own layer, 2) create a machine there that is your desired name, but essentially just replicates the original via require+OVERRIDES 3) build 4) then work from there.
aak-rookie has quit [Quit: Client closed]
amitk_ has joined #yocto
<Entei[m]>
<LetoThe2nd> "Entei: modifying anything in..." <- But that's exactly what I am doing right now. Made my own machine config. It gave out errors regarding `COMPATIBLE_MACHINE`. Added my machine name to couple of linux kernel bb files which had the variable `COMPATIBLE_MACHINE`, and started building. Only to get error at the very end of process....
<LetoThe2nd>
Entei[m]: well then don't do it ;-)
<Joel63>
Good morning! What recipe provides the dd utility? The one I get is from busybox and it doesn't support the iodirect flag.
<Entei[m]>
So....what's the next possible solution to the problem? I see couple of tunes in the riscv-tunes including riscv64nf, riscv64nc, so it should be possible.
<LetoThe2nd>
Entei[m]: the key is that your custom machine needs to build upon the original machine. and you can always bbappend the kernel to extend compatible machine.
aak-rookie has joined #yocto
tilman[m] has quit [Quit: You have been kicked for being idle]
<KareemZarka[m]>
hey , so i submitted a test with [PATCH] meta/lib/oeqa/selftest/cases/wic ,i got a feedback that this test case fails on auto-builds with an error that indicates that mtools package in not installed there .. what can i do about this ? whats the right way of solving this issue ? can i add a dependency to core-image-minimal ``` DEPENDS = "mtools" or in local.conf?
<KareemZarka[m]>
````
<KareemZarka[m]>
* hey , so i submitted a test with \[PATCH\] meta/lib/oeqa/selftest/cases/wic ,i got a feedback that this test case fails on auto-builds with an error that indicates that mtools package in not installed there .. what can i do about this ? whats the right way of solving this issue ? can i add a dependency to core-image-minimal \`\`\` DEPENDS = "mtools" or in local.conf?
<rburton>
bitbake mtools-native before trying to use the binaries
<KareemZarka[m]>
* test with \[PATCH, * [PATCH\] meta/lib/oeqa/selftest/cases/wic, * = "mtools"``` or
<Entei[m]>
<LetoThe2nd> "Entei: https://github.com/..."; <- Well I followed it verbatim, still getting the same error saying could not locate BSP definition for qemuriscv64nc/standard
<LetoThe2nd>
Entei[m]: "it says". who or what is "it"? a recipe? the parser?
<Entei[m]>
Oops sorry for being vague. Give me a min, I'll need to copy through a usb drive. I am doing this on a different machine on a different network.
<KareemZarka[m]>
KareemZarka[m]: would that be a legit solution ?
<Entei[m]>
Tried the other conf you sent, still the same error...
<Entei[m]>
Unable to find much on internet how to tune the machine config
<LetoThe2nd>
Entei[m]: no, you're on the wrong track. it is not about a problem with the tunes. it is that the kernel recipe that you are using has quite a bit of magic for the qemu-machines.
<LetoThe2nd>
Entei[m]: plus, where does 5.19 come from?
<Entei[m]>
LetoThe2nd: Not too sure tbh. I am on Langdale
<LetoThe2nd>
Entei[m]: what layers are involved?
<LetoThe2nd>
Entei[m]: ok, langdale has 5.15 and 5.19.
<Entei[m]>
meta, meta-poky, meta-yocto-bsp, meta-riscv. The riscv layer was added manually, others were defaults from poky
<LetoThe2nd>
Entei[m]: I'm unfortunately not very experienced with the linux-yocto internals, but again, *guessing* it has something to do with KMACHINE/KBRANCH.
<LetoThe2nd>
maybe somebody else can chip in here.
<Entei[m]>
LetoThe2nd: Yeah I am trying to find a branch in yocto kernel repository right now.
amitk_ has quit [Ping timeout: 255 seconds]
camus has quit [Quit: camus]
kroon has joined #yocto
starblue has quit [Ping timeout: 246 seconds]
starblue has joined #yocto
leonanavi has joined #yocto
leon-anavi has quit [Ping timeout: 252 seconds]
seninha has joined #yocto
dgriego has quit [Ping timeout: 252 seconds]
rob_w has quit [Remote host closed the connection]
dgriego has joined #yocto
brazuca has joined #yocto
kroon has quit [Quit: Leaving]
d-fens_ has joined #yocto
<d-fens_>
re, any idea how to give non root users acces to the gpio on a raspi ?
<d-fens_>
# Include the device files in the package
<d-fens_>
FILES_${PN} += "/dev/gpiomem /dev/mem"
<d-fens_>
spoiler , it doesn't work, chgrp: cannot access '/build/tmp/work/cortexa7t2hf-neon-vfpv4-poky-linux-gnueabi/firmware/0.1+gitAUTOINC+798afc4d43-r0/image/dev/mem': No such file or directory
<Joel17>
Folks, how do I find out what recipes provide fastboot and adb?
kscherer has joined #yocto
<rburton>
Joel17: layers.openembedded.org is the usual search tool. they're both android things though, so you might need to write recipes if you want to use them.
<JaMa>
there are recipes for these android-tools
<rburton>
the summary should mention them so they're searchable :)
<Joel17>
Looks like android-tools is a recipe. Thanks!
<rburton>
oh god i hate cmake and llvm so much
<rburton>
"quick cleanup", I said
<JaMa>
more than nodejs this week? :)
<rburton>
oh yes
Joel17 has quit [Quit: Client closed]
NTShetty has joined #yocto
shivamurthy has quit [Excess Flood]
shivamurthy has joined #yocto
<RP>
rburton: I will be shocked when you find something you like ;-)
<rburton>
haha
rfuentess has quit [Remote host closed the connection]
aak-rookie has quit [Quit: Client closed]
<landgraf>
rburton: do you like rburton ? :)
brazuca has joined #yocto
louis_ has quit [Quit: Lost terminal]
NTShetty has quit [Ping timeout: 260 seconds]
seninha has quit [Remote host closed the connection]
vladest has quit [Remote host closed the connection]
vladest has joined #yocto
invalidopcode1 has quit [Remote host closed the connection]
invalidopcode1 has joined #yocto
mason has joined #yocto
<mason>
Hi all! I'm new to Yocto and hitting an issue I don't yet understand. I'm adding a kernel patch to an existing layer, but my build is erroring out because my patch file isn't being found. I'm saying FILESEXTRAPATHS:prepend := "${THISDIR}/${PN}:" and SEC_URI += "file://001-my.patch" and it's not finding my patch file, where 001-my.patch is in this same directory next to linux-intel_5.15.bbappend. Is this
<mason>
something obvious to you folks that know the system better?
<JPEW>
It needs to be in a subdirectory named "linux-intel"
<mason>
aha, my subdir that this stuff is in was "linux"
<JPEW>
Leave the bbappend where it is
<mason>
JPEW: Do I want "linux-intel" to contain bbappend, and also contain a "linux-intel" dir with the patchfile?
<RP>
there is something disturbing with a test suite reporting INFO: ... unexpected success
<mason>
Alright, so "linux" with bbappend and a linux-intel directory, patch in the latter? Trying that.
<mason>
JPEW: You rock. Thank you so much! What's the best stuff to read about this so I can learn to fish on my own?
<mason>
JPEW: Build's already getting further than it was, so you've nailed it. THANK YOU.
<mason>
I'm going to start lurking here as Yocto is going to be an increasingly large part of my life moving forward.
<JPEW>
`FILESEXTRAPATHS` tells bitbake what extra paths to search: `${THISDIR}` is the directory that the bbappend file lives in (recipes-kernel/linux), and `${PN}` is the name of the recipe (linux-intel)
<mason>
Aha, I dind't know what ${PN} represented.
<JPEW>
Therefore, when you say FILESEXTRAPATHS:prepend := "${THISDIR}/${PN}:" you are telling bitbake to look for files in "recipes-kernel/linux/linux-intel"
<JPEW>
mason: Ya, PN is the name of the recipe, and PV is the version (5.15 in your case)
<mason>
JPEW: Hrm. One of my tries, I thought I'd just set "${THISDIR}" and that failed. I'd think that'd not fail, but maybe I'm misremembering.
<JPEW>
mason: `FILESEXTRAPATHS:prepend := "${THISDIR}/${PN}:"` is the recommended thing to do unless you have a good reason to do otherwise
<JaMa>
mason: and you should always use BPN in this instead of PN (or multilib will hunt you in sleep later)
<mason>
kk, I'm fine with convention
<JPEW>
(maybe you forgot the trailing ":")
<JPEW>
JaMa: Ah, right
<JPEW>
I need update my vim plugin
<mason>
JPEW: Ah! I bet I did. I'll try it.
<JaMa>
or native
<mason>
JaMa: Where can I read about how PN differs from BPN? None of my co-workers are a ton more familiar with Yocto than me and we're all using it henceforth.
<RP>
JaMa, JPEW: Did one of you use icecc? I'm curious about your thoughts on that network flag patch
<JPEW>
mason: It's related to multilib (e.g. having multiple different architectures installed at the same time, like i686 & x86-64), but I'm not familiar with it.
<JPEW>
We used to
<JPEW>
I'll look
<JaMa>
RP: some people at @LGE still use it and I've already replied to ML
<JaMa>
since the network was enabled I've overlayed that .bbclass in our layers not to enable the network globaly and asked people who use it to enable it only for recipes where it matters (e.g. chromium)
<JPEW>
RP: Ya, the inline version from JaMa is better :)
<RP>
JaMa: I see you did reply, thanks
<RP>
Well, it is "better" and even more parse overhead
<RP>
JaMa: anoynmous python gets run once, the other at every variable expansion event
<JaMa>
because it's 4 times the expression?
<JaMa>
ah :(
<RP>
JaMa: right and we have to rerun it any time a setVar occurs and nukes the expand cache
<JaMa>
than I'm happy with Jose's version and I shouldn't even ask for that network flag evaluation :/ sorry
<RP>
JaMa: your version is neater and I can see why people like it, it just won't perform as well :/
<JPEW>
RP: Someone wanted it for export too right?
<RP>
JPEW: people have wanted it for various flags but yes
<JaMa>
RP: it also predates Jose's version quite a bit :)
<JPEW>
Ah, this is the `FOO[export] = ""` still enables the export problem
<JaMa>
as I did it as well as expanding network varflag as soon as I've noticed the original icecc.bbclass
<JaMa>
before realizing what I'm getting into :)
florian has quit [Quit: Ex-Chat]
<RP>
JaMa: its a tricky one :/
florian_kc has quit [Ping timeout: 252 seconds]
ptsneves has quit [Ping timeout: 252 seconds]
Herrie has quit [Ping timeout: 255 seconds]
mckoan is now known as mckoan|away
Herrie has joined #yocto
leonanavi has quit [Quit: Leaving]
<JaMa>
mason: it's not only multilib, PN-native and nativesdk-PN have the same behavior, so unless you really need to have e.g. different patch for native version of that recipe, you should just use BPN in FILESEXTRAPATHS to be safe
<mason>
JaMa: Alright, noted.
<RP>
tlwoerner: FWIW, "qemu-system-x86_64 tmp/deploy/images/qemux86-64/build-appliance-image-qemux86-64.wic.vmdk -enable-kvm" nearly works so I think we can boot these images under runqemu with some small tweaks
<tlwoerner>
RP: ok
<RP>
tlwoerner: I feel a bit bad pushing discussion on some of this, I am really happy to see this getting some tlc!
<tlwoerner>
it'll be nice to be able to test pseudo proactively
<RP>
tlwoerner: amongst other uses, yes!
<RP>
"./scripts/yocto_testresults_query.py regression-report 4.2_M1 4.2_M2 " on master-next is quite cool
<RP>
It doesn't fully work yet but is getting there
invalidopcode1 has quit [Remote host closed the connection]
invalidopcode1 has joined #yocto
<mrybczyn[m]>
rburton: sakoman: the new nvd fetcher posted as rfc
amitk_ has joined #yocto
azcraft has quit [Remote host closed the connection]