warpme has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
warpme has joined #u-boot
monstr has joined #u-boot
jfsimon1981 has quit [Remote host closed the connection]
jfsimon1981 has joined #u-boot
jfsimon1981 has quit [Max SendQ exceeded]
jfsimon1981 has joined #u-boot
jfsimon1981 has quit [Remote host closed the connection]
jfsimon1981 has joined #u-boot
mmu_man has quit [Ping timeout: 256 seconds]
Peng_Fan has quit [Quit: Connection closed for inactivity]
naoki has quit [Quit: naoki]
xypron has joined #u-boot
mmu_man has joined #u-boot
Jah has joined #u-boot
rockosov has joined #u-boot
<Jah>
Hi together! I am trying to port uboot based on imx7 from 2017 to 2024. Uboot was from imx repository in which custom board and config were designed. Now i have pulled the new uboot and copied the config and board to it as a test without changing anything in config file or board directory. When I try to build it, it takes around 20 mins to crash
<Jah>
the terminal. I have checked the ram usage, it swallows almost 24 Gb of ram and gets so bloated eventually to kill the terminal. I am using yocto to build. Does anyone have any clue how should I analyze it? Help would be greatly appreciated.
<marex>
Jah: start the board port from scratch, there are so many changes between 2017..2024 versions and between downstream fork and mainline one that it is not worth trying to upport
<marex>
Jah: as for why do you OOM your machine, build u-boot by hand with make -j1 and use some ps aux
<marex>
Jah: which exact version are you building and from which repository ?
<Jah>
I am new to building uboot so thought I would erradicate error as it comes instead of making it from scratch. I am using 2024.04 from imx as well
<Jah>
marex if you suggest that it's not a good idea, I'll try to start it from scratch
<derRichard>
how many jobs are you creating when you run make?
<Jah>
derRichard I am trying to compile only uboot
<derRichard>
if you just run "make -j" (without limiting the number of jobs), it can overload your system
apritzel has joined #u-boot
<Jah>
derRichard CPU is at 85-90% moreover however RAM gets bloated to its limits eventually killing the terminal. Is there a way to asses it, or should I just refrain from upport as suggested by marex and start from scratch
pivi has quit [Server closed connection]
pivi has joined #u-boot
Jah has quit [Ping timeout: 256 seconds]
Jah has joined #u-boot
<xypron>
Jah: make -j$(nproc)
<quinq>
yocto is for building Linux, not for building U-Boot
<quinq>
U-Boot will never take 24Gb of ram to build
<quinq>
ditch yocto, build U-Boot normally
<Jah>
xypron quinq thanks alot for your suggestion. I'll try to give it a go to compile it through make, hoping to have a better result with it.
* derRichard
notes that yocto does not need 24gb of ram to build
<Jah>
derRichard I doubted indeed, however didn't have clue how to find out why is it swallows so much of ram. Other packages works fine even linux 6.6 was buildable
<apritzel>
Hi, do I get this correctly that v2024.10 will be released with the (current) Linux v6.9 based dts/upstream directory? As the next sync will only happen when -next opens, after -rc1?
Stat_headcrabed has joined #u-boot
tgamblin has quit [Server closed connection]
tgamblin has joined #u-boot
memset has quit [Ping timeout: 260 seconds]
memset has joined #u-boot
zibolo has joined #u-boot
mmu_man has quit [Ping timeout: 264 seconds]
mmu_man has joined #u-boot
tgamblin_ has joined #u-boot
tgamblin has quit [Ping timeout: 245 seconds]
tgamblin_ is now known as tgamblin
frieder has quit [Ping timeout: 246 seconds]
joeskb7 has joined #u-boot
<marex>
Jah: if you are not building from u-boot upstream, but NXP downstream fork with 300000+ lines of changes, do ask NXP for support
<marex>
the NXP downstream stuff makes no sense on mx7 , simply use upstream 2024.07 and be done with it
frieder has joined #u-boot
MrCryo has joined #u-boot
<derRichard>
where can i find a (git) repo of the old ppcboot code?
dsimic has quit [Ping timeout: 258 seconds]
<derRichard>
IOW, a pre u-boot history
dsimic has joined #u-boot
goliath has quit [Quit: SIGSEGV]
<derRichard>
ah, sourceforce offers a backup via rsync
<derRichard>
but rsync stopps here after a few seconds, so the repo is not complete :-(
<Tartarus>
Going for me
<Tartarus>
Neither done nor hung
<derRichard>
hm, maybe i got a bad mirror here in austria
<Tartarus>
... spoke too soon. Hangs for you after ppcboot/board/musenki/Attic/dc2114x.c,v ?
<Tartarus>
ah, wait, going again
<Tartarus>
If it doesn't finish for you I'll put a tar of the rsync up for you shortly and git conversion and mirror on my TODO list
<derRichard>
anyway, u-boot's dlmalloc.c contains a funny bug and i'd like to understand where it comes from. the original dlmalloc.c from doug lea does not have it
<derRichard>
Tartarus: would be great!
<derRichard>
you can also run git cvsimport and push the git repo somewhere
sauce has quit [Server closed connection]
sauce has joined #u-boot
<derRichard>
i managed to rsync just common/dlmalloc.c,v, the bug seems to predate also cvs :-(
<Tartarus>
Ouch
<Tartarus>
And yeah, I think it's been since early git days since I converted something to git, so I'm going to be optimistic it's easy and correct enough these days
<derRichard>
i'll send a mail soon. barebox is also affected. it's an integer overflow in malloc itself. i.e. you get less memory than you request.
<derRichard>
you can trigger it using unbound allocations like squashfs/ext4 and likely other drivers do.
<apritzel>
Tartarus: thanks, v6.10 sounds much better, though in general the update policy leads to quite some lags, which significantly delays new board support when using dts/upstream
mmu_man has quit [Ping timeout: 246 seconds]
<apritzel>
for instance the Linux v6.11 DT changes are all done by now, just awaiting to be merged in the next two weeks, but would only be part of U-Boot v2025.01
<apritzel>
for the manual sync in the last years I typically copied Linux -rc1 DT files into U-Boot up to U-Boot's -rc2 or -rc3, which gives a much quicker turnaround
mmu_man has joined #u-boot
frieder has quit [Remote host closed the connection]
cbmuser has quit [Server closed connection]
cbmuser has joined #u-boot
eballetbo has quit [Server closed connection]
eballetbo has joined #u-boot
tpw_rules has quit [Server closed connection]
stefanro has quit [Ping timeout: 256 seconds]
tpw_rules has joined #u-boot
Stat_headcrabed has joined #u-boot
MrCryo has quit [Remote host closed the connection]
ConorDooley is now known as conchuod
slow99 has quit [Server closed connection]
slow99 has joined #u-boot
Hypfer6 has quit [Read error: Connection reset by peer]
Hypfer6 has joined #u-boot
<Tartarus>
apritzel: You're allowed to cherry-pick
<Tartarus>
So boards don't have to lag.
goliath has joined #u-boot
<apritzel>
I see, so that's commits from the upstream repo, so that a later merge will skip them?
<Tartarus>
Yeah
<Tartarus>
The dts script has a cherry-pick option
<Tartarus>
Use that, submit those as a patch to apply, next re-sync works as expected