PaulFertser has quit [Remote host closed the connection]
fakuivan has joined #openwrt-devel
fakuivan has quit [Remote host closed the connection]
fakuivan has joined #openwrt-devel
minimal has quit [Quit: Leaving]
iocampomx has joined #openwrt-devel
rua has quit [Quit: Leaving.]
tSYS has quit [Quit: *squeak*]
tSYS has joined #openwrt-devel
<iocampomx>
I'm personalizing the /etc/banner file, how can I get the "%D %V, %C" text replaced with the corresponding values?
iocampomx has quit [Ping timeout: 480 seconds]
Tusker has joined #openwrt-devel
<Tusker>
heya guys, I am trying to debug some kernel modules on openwrt, and wondering how to get the DEBUG output working? it has #ifdef DEBUG, but how do I get debug to actually print somewhere ?
iocampomx has joined #openwrt-devel
iocampomx has quit [Read error: No route to host]
Tusker is now known as Guest2110
Guest2110 has quit [Ping timeout: 480 seconds]
iocampomx has joined #openwrt-devel
iocampomx has quit [Read error: No route to host]
iocampomx has joined #openwrt-devel
<fioriceta>
KanjiMonster: might try to figure out the original bootloader
<fioriceta>
probably the best approach imo
<fioriceta>
I think Borromini looked at the format for the rtl switches
tidalf has quit [Remote host closed the connection]
<fioriceta>
there seems to be a way to download an image over the console
<fioriceta>
control flow's all fucked, time for the other board...
iocampomx has joined #openwrt-devel
<iocampomx>
Hi there! I'm working on the network configuration of the router, and trying to define the wan interface, the upstream router is offering an IP via DHCP, but the OpenWRT isn't getting it.
<iocampomx>
Any high-level advice on how to setup network interfaces?
<fioriceta>
iocampomx: what version of openwrt is this?
<fioriceta>
does `ps` show the dhcp client running?
<iocampomx>
v23.05.0-rc3
<iocampomx>
Yes, the dhcp client seems to be running /usr/sbin/odhcpd
<fioriceta>
no, that is the DHCP server, anything else?
<fioriceta>
I think udhcpcd is the client
<fioriceta>
*udhcpc
<iocampomx>
Yes, udhcpc -p /var...
<iocampomx>
-i eth0.2
<fioriceta>
hmm
<fioriceta>
would suggest a packet capture (tcpdump), could try turning off the fw for a moment
<iocampomx>
Does the config file look good?
<fioriceta>
yeah
<iocampomx>
The router has 5 ports, port 0 is WAN, port 1-4 are LAN. I tested that physically and reviewed the logread
<fioriceta>
hmm, odd
<fioriceta>
can you manually assign an IP and see if the interface works for eth0.2?
<fioriceta>
e.g. ip addr add dev eth0.2 192.168.1.1/24
<iocampomx>
Sure, let me try
<fioriceta>
oh, yeah, what target is this?
<fioriceta>
`ubus call system board`
<iocampomx>
It's a cnPilot R190W, I'm building my own image for it
torv has quit [Remote host closed the connection]
<fioriceta>
is this something ath79 related?
<iocampomx>
But if you're wondering about the chip, it's an MT7628An
<fioriceta>
ah
torv has joined #openwrt-devel
<iocampomx>
rt3050 is the ethernet chipset I think
<fioriceta>
check to see if the interfaces work, I guess, I don't see support for this in mainline
<iocampomx>
LAN interfaces work well. If I connect my computer with a network cable, I'm able to get an IP address and communicate with the router properly via LAN.
<fioriceta>
oh, yeah, you could get rid of the br-lan stuff as you have swconfig stuff
<fioriceta>
but hmm
<fioriceta>
I don't know, run a packet dump and see
<iocampomx>
What is the basic network config file I can use to start from zero?
<fioriceta>
that one is fine
<iocampomx>
If I do swconfig dev switch0 show, I can see 6 ports. And 2 VLANs.
<iocampomx>
VLAN1: 1 2 3 4 5t. VLAN2: 0 5t. Am I missing port 6 somewhere?
<fioriceta>
yeah, 6 ports is expected cause one of them is going to be for the CPU
<fioriceta>
are you able to ping 192.168.1.1 from br-lan?
<fioriceta>
oh, *192.168.11.1
<iocampomx>
Do you mean, within the router, doing: ping -I br-lan 192.168.11.1? If so, yes.
<fioriceta>
no, no, from one of the untagged ports
<fioriceta>
I assume you have a device tree for this, yeah, post that, too
goliath has joined #openwrt-devel
<iocampomx>
It's basically this one: https://git.openwrt.org/?p=openwrt/openwrt.git;a=blob;f=target/linux/ramips/dts/mt7628an.dtsi;hb=HEAD
<iocampomx>
I didn't include anything else other than the partitions.
<fioriceta>
I don't have any experience with mtk stuff, better to wait for someone else
<iocampomx>
Sounds good. thanks! I will try tomorrow!
<fioriceta>
well, just make sure to idle in here
KGB-2 has quit []
KGB-2 has joined #openwrt-devel
rua has joined #openwrt-devel
iocampomx has quit [Ping timeout: 480 seconds]
mrkiko has quit [Quit: leaving]
gladiac has joined #openwrt-devel
rua has quit [Quit: Leaving.]
<lu_zero>
anybody with a bpi-r3 can confirm if in rc4 the recovery partition is again usable?
<robimarko>
hitech95: Yeah, that should be the other way around
tidalf has joined #openwrt-devel
<hitech95>
robimarko, yea yesterday I've forked U-boot and backported the patch after seeing your message.
<robimarko>
Thanks for that
<hitech95>
no problem. Today we are testing again on a set of HW the other patches before updating an open PR we have then we can properly merge.
<hitech95>
Uboot mailing list rejected the patch email and is pending a moderator so we will see what I have to do over there too!
<robimarko>
Probably as you are sending something for the first time
<hitech95>
Yea for sure!
<hitech95>
PS. we have 2 PR open one for the SPI fix and one for device support in uboot :D the last one is a bit stale right now as we are testing the latest fixes.
<hitech95>
And we are working in steps so we dont send a mega PR with a lot of stuff changed :D
<robimarko>
Every patch should be properly formatted
tidalf has quit [Ping timeout: 480 seconds]
<hitech95>
robimarko, got it, so the quick way is just to clone the package repo and then commit and generate a patch from it?
<robimarko>
Yes, that would be the proper/easiest way
tidalf has joined #openwrt-devel
tidalf has quit [Ping timeout: 480 seconds]
tidalf has joined #openwrt-devel
kwz has quit [Ping timeout: 480 seconds]
<hitech95>
ok, I've reworked the patch to add the headers
<kabel>
Ansuel: omg, I sent v2 of the patches, fixing the typo Simon noticed, and forgot to add your Reviewed-by tags. Could you please send them again? sorry
Danct12 has quit [Remote host closed the connection]
<hitech95>
hey the node memory in dts ho wis calculated? reg = <0x40000000 0x20000000>; Means 512MB of ram right?
<fioriceta>
you don't need that, kernel should autoprobe it
<robimarko>
No, there is no autoprobing
<fioriceta>
oh?
<robimarko>
Usually its up to U-Boot to properly set it
<robimarko>
hitech95: Are you sure its just using 1 for address and size cells?
<robimarko>
Cause most modern ARM cores use more than 32 bits for memory
gladiac has joined #openwrt-devel
<fioriceta>
robimarko: hmm, maybe I was thinking of mips, then
<hitech95>
robimarko, I dont know, the refbrd has 2GB while our board has 1GB. But we are passing 0x20000000 for now (sometimes we have issues related to this) So I wanted to be sure that the value we are calculating are ok
<hitech95>
And yes I'm taliking about uboot, on the kernel dts it is correctly set at 0x40000000
<robimarko>
Well, in your case its just <address size>
<robimarko>
So, you are passing 512MiB as size
<hitech95>
ok, now the question is why sometimes there are issues if we pass the full 1G. The memory node is also set in the kernel dts so over there the memory is correct
<robimarko>
You sure there isnt some reserved memory not being defined or?
slh64 has quit [Remote host closed the connection]
<hitech95>
robimarko, not sure, the bpi-r3 pass full 2GB in there so those addresses should be all free to map the ram.
<hitech95>
it also seems to be board dependent
<hitech95>
so for now we are bootstrapping with "only" 512MB
<hitech95>
Now I have to wait for the daniel review
<robimarko>
Well, it is board dependant size they use different RAM size
<hitech95>
robimarko, we are not sure at the moment that is why the PR only uses 512 for now.
<hitech95>
We have to open up more devices and see the chips
<robimarko>
Make sure to point that out in the PR
<hitech95>
Yea for sure.
<robimarko>
Cause if DRAM init suceeded, then it should work
tidalf has quit [Ping timeout: 480 seconds]
filimonic has quit [Remote host closed the connection]
gladiac is now known as Guest2133
gladiac has joined #openwrt-devel
gladiac has quit []
Guest2133 has quit [Ping timeout: 480 seconds]
<hitech95>
robimarko, dumb question, I've worked on some patches and a frind on others. how can we maintain the commit history when submitting a PR?
<hitech95>
Generally a single commit is required for the same context.
<robimarko>
You can use co-developed by
<robimarko>
And also, by patch you mean commit includes patch files or?
<fioriceta>
tbh, I think that patchwork is far better than github
<fioriceta>
this reliance on M$ is concerning
<robimarko>
Again, can you stop just popping up in every single question?
<fioriceta>
it's not intentional
<robimarko>
I dont know how humans can write non-intentionally so
rua has joined #openwrt-devel
<fioriceta>
also, it's fucking IRC, just add me to your ignore list
<fioriceta>
this isn't a forum where stuff is structured
<fioriceta>
I'm not reading corporate nonsense >.>
<hitech95>
I mean right now we have different of commit from me and my friend for the same context (ex target mediatek same router). He worked on basic support while I finalized and cleanedup the work.
<hitech95>
as you can see if I shqsh everything into a couple of commits the patch will be huge and hard to follow.
<fioriceta>
just squash it
<hitech95>
I could squash it but then there is no reference on who whorked on what. and some commit add support for stuff that is not !00% basic so we are not even sure if those will be accepted.
<robimarko>
Basically just squash the device support commits
<robimarko>
You can add co-developed-by for his contribution
<robimarko>
Depends on who authored the most
<robimarko>
If he did the initial support then usually you would keep him as the author and add your signed off by
<robimarko>
You can also include a small changelog of what your part was
<hitech95>
got it
<hitech95>
YEa basically he made the board boot I've done all the chores :D I'vve put myself as co-dev and co-author
<fioriceta>
openwrt needs a higher barrier to entry
<robimarko>
Can you clarify as you keep repeating that every day
<fioriceta>
it's not self-explanatory?
<robimarko>
No, because why should there be any barrier for entry?
<fioriceta>
maybe your time isn't worth much if you're having to answer basic questions...
<fioriceta>
99% of the case people cannot read
<fioriceta>
don't know why it seems to originate from the US
<robimarko>
Ok, Il bite, what is your issue with the US?
<fioriceta>
well, vendors release this shit that is a clusterfuck to unfuck
<fioriceta>
DRM-ridden shit, secboot shit, etc., I could go on
<fioriceta>
I guess one example would be st stuff, you can reuse the IC, but be forced to wipe flash
<fioriceta>
US vendors literally give zero fucks about anything, it is insanity, and then, I mean, look at how people live in that country
<robimarko>
If you have suck a view, why do you use HW from US vendors then?
<fioriceta>
not many good alternatives
<fioriceta>
(too expensive, impossible to source, etc)
rua is now known as Guest2141
rua has joined #openwrt-devel
<robimarko>
And how is blaming the US going to help you?
<fioriceta>
well, I don't know, they keep fucking with china, so it's only going to fuck over the supply of cheap stuff
<robimarko>
I mean, if you are not from the US I dont see how you supply of what is basically e-waste at the time of manufacturing impacting you
<robimarko>
Cause lets be honest all of that wonderfull no-name barelly works generic electronics is just e-waste
<fioriceta>
well, I bought one of those ralink dongles
<fioriceta>
worked for 5mins, then died
<robimarko>
Well, that is exactly my point
<fioriceta>
I think I ought to just source e-waste in bulk tbh
<robimarko>
Perfect, that would really help China and the rest of the world
<robimarko>
I am sure you can buy it in bulk
Guest2141 has quit [Remote host closed the connection]
<fioriceta>
lol, is that sarcastic...
<fioriceta>
but I don't know, nothing that there are thousands of these things out there, completely unusable because of some DRM shit
<fioriceta>
s/nothing/knowing/
<robimarko>
I dont know what to tell you other than go out in nature, meditate or whatever relaxes you as being frustrated because some vendor locked something up is not going to help anybody
<linusw>
talking about e-waste: I need help to get my ixp4xx branch merged, is there any way to achieve that other than aspiring to become a committer?
<robimarko>
Anybody with commit rights can merge it
<KanjiMonster>
robimarko: the question was who will merge it, not who can merge it ;-)
<robimarko>
Answer is kind of the same
<robimarko>
I can give it a high-level review if that helps somebody pick it
<robimarko>
Yeah, I am going to go through it and probably ask some questions
<Ansuel>
I can allready say that all the kmods addition can be merged right away
<robimarko>
Wait, I have some questions there
<Ansuel>
the hdparm script... i don't like it do we really need it?
<Ansuel>
also think package/firmware/ixp4xx-microcode/Makefile should also be split in a separate commit ?
<fioriceta>
KanjiMonster: what 6362 board do you happen to have?
<fioriceta>
trying to locate the jtag pinout, I think these 4 resistors are it, but hmm...
<KanjiMonster>
fioriceta: Neufbox 6
<fioriceta>
yeah, just saw that on your wiki page
<KanjiMonster>
which is horribly outdated
<fioriceta>
hmm
<fioriceta>
your board only has jtag via test pads
<robimarko>
linusw: Gave your series a quick pass
<KanjiMonster>
I also don't have any jtag adapters. well technically I have two, but I don't have any drivers for it (BCM9EJTAG, which is a bcm6328 based device with a pcie conntected fpga)
<KanjiMonster>
from a time when Broadcom was more leasurily with its employees sending hardware to random people
<Ansuel>
robimarko about the bisect problem... it's either we ignore the bisect problem and put the commit after patch 5 or we put everything in patch 5 (but I would love to keep them split)
<robimarko>
I would also prefer to split it as much as possible
jetm has joined #openwrt-devel
<KanjiMonster>
which PR?
<linusw>
robimarko: I'll go over your comments and answer/send v3
<robimarko>
That would be great, we can then at least get the RTC kmods merged
<linusw>
there is one thing though: the apex boot loader probs need a file to be uploaded/imported and I have no idea how to achieve that
<Ansuel>
robimarko if we want them split then also package/firmware/ixp4xx-microcode/Makefile should be split from patch5
<linusw>
I refer to an external DL URL right now
<Ansuel>
linusw can we redistribute it?
<robimarko>
linusw: Can you point to it?
<KanjiMonster>
the arm-magic deletion could also be applied right now (and may very well be patch 1)
<linusw>
KanjiMonster: I can move the patch first, but if someone want to apply it right off that works too, I'll just rebase.
<fioriceta>
KanjiMonster: NB6 uses SPI NOR flash?
<Ansuel>
linusw I will take care of cherry-pick it
<linusw>
Ansuel: thanks a lot
<fioriceta>
I need a cfe dump for SPI NOR, if you don't mind
<robimarko>
linusw: That is the source tarball?
<Ansuel>
wait that tar is just source of the thing to be compiled?
<KanjiMonster>
fioriceta: yes
<linusw>
robimarko: yes
<fioriceta>
I want to try to solder on a spare chip onto my board, then remove the nand stra
<Ansuel>
and you need to add your changes on top of it?
<fioriceta>
*strap
<KanjiMonster>
and it's safely tucked away in a box somewhere
<robimarko>
Then its fine, you are fetching it properly
<linusw>
Ansuel: there were so many changes needed to Apex, both from our old patch stack, from the abandoned project, and from myself, that I just forked it.
<Ansuel>
linusw that is find but I wonder if it's better to use git to clone it instead of downloading from release?
<Ansuel>
also I would change the default branch to master (you have main empty)
<Ansuel>
or swap them if nobody use it
<robimarko>
Ansuel: If the tarball if reproducible from GH why not just download it
<Ansuel>
robimarko i don't like using github direct url
<Ansuel>
they like to change like stair in harry potter
<robimarko>
Yeah, I remember the last time
<linusw>
Ansuel: haha :D good point
<robimarko>
Nothing was reproducible
<Ansuel>
(and once we have that merged we will download the tar from your mirror and never have the overhead of a slow clone)
<linusw>
I thought the way things work is OpenWrt downloads a copy of the tar.gz and reference it by hash, then if it updates upstream, then someone need to pick the new file into the file cache.
<linusw>
At least that's how U-Boot Makefiles looked so...
<linusw>
but I'm happy to do it any way
<Ansuel>
just taste but if the source comes from a git source i always prefer git instead of plain download
<linusw>
OK can you point me to a Makefile that does this and I will use that NP
<robimarko>
PKG_SOURCE_VERSION is the commit hash you want to checkout
<linusw>
OK no problem
<linusw>
and the PKG_MIRROR_HASH...?
<robimarko>
PKG_MIRROR_HASH is the sha256 hash of the packaged tarball that buildsystem makes after downloading via git
<robimarko>
You can have buildsystem fix it for your
tidalf has joined #openwrt-devel
<Ansuel>
sha256 dl/name of the package if you want it manually
<robimarko>
Just put whatever or something like that into PKG_MIRROR_HASH
<robimarko>
And then: make path-to-your-package/download
<robimarko>
make path-to-your-package/check FIXUP=1
<robimarko>
That will do it for you
<Ansuel>
yep
<linusw>
hehe yeah a bit of chicken and egg problem there I see, so hash whatever will be fixed by the build system eventually?
<linusw>
Ah FIXUP=1 clever
<linusw>
I fix
<hitech95>
Yes, just received the clamp to flash the SPI nand, and my board is up and running! Now I can test my crazy stuff
<KanjiMonster>
robimarko: replied also at the ml, no bisectability issue with patch 1 :)
<robimarko>
Perfect
<KanjiMonster>
@FOO is just verbatim taking over without any lookups or anything else, so you can use any (future/non existing) symbol you like
fioriceta has quit [Quit: quit]
<Ansuel>
linusw if that's the case then please please split also package/firmware/ixp4xx-microcode/Makefile from patch 5
<linusw>
KanjiMonster: makes sense we do that in the kernel all the time.
<linusw>
Ansuel: no problem I fix.
<KanjiMonster>
hitech95: have you sent the mediatek spi clock fix upstream?
<hitech95>
KanjiMonster, yes but the email is pendign moderation approval
<KanjiMonster>
hitech95: great. Maybe put that in the commit message (just that it was sent upstream, not that it is pending approval), so you don't get that asked more often ;)
<KanjiMonster>
linusw: you still need to account for the header of the integrated switch (6 bytes), so the correct buffer size would be mtu + ETH_HLEN + ETH_FCS_LEN + VLAN_HLEN + 6
<KanjiMonster>
linusw: also I'm currently tempted to say we should stick to 1532, because that is what's guaranteed to work, at least with the integrated switch
<linusw>
KanjiMonster: but I need at least 1536 to make this device work .... due to the external Marvell switch. Otherwise I can't fit the DSA tags etc on a regular 1500 byte packet.
<linusw>
It certainly works on the BCM6328 at least because they manufactured a few million of these devices.
<linusw>
(Inteno XG6846 used by a majority of Swedish ISP:s for fiber interface)
<linusw>
the default MTU is 1506 in any case so bigger sizes need to be asked for, but if you wanna keep it to 1536 that works for me.
<hitech95>
Hey dumb question I have the SPI bus dump of a proslic FXS codec and I've done some reverse engineering on it.
<hitech95>
Is someone interested on the dump?
tidalf has quit [Ping timeout: 480 seconds]
<KanjiMonster>
linusw: buffer size would be 1532 + 4 (VLAN_LEN) + 6 (BROADCOM_HEADER), which would leave enough room for the 8 byte DSA tag
valku has joined #openwrt-devel
goliath has quit [Quit: SIGSEGV]
<KanjiMonster>
linusw: 1532 would be more than enough for 1518 (normal ethernet frame) + 8 (marvell dsa tag), it even leaves 6 bytes space. Note that I don't include the vlan tag and the broadcom header tag in there, these are on top of the 1532. (so the rx buffer length is 1542)
rua has quit [Ping timeout: 480 seconds]
scr4tchy has quit [Quit: :]]
mentalow has joined #openwrt-devel
rua has joined #openwrt-devel
KGB-1 has joined #openwrt-devel
xavifr has quit [Ping timeout: 480 seconds]
robimarko has quit [Read error: Connection reset by peer]
Danct12 has joined #openwrt-devel
<hurricos>
procd question: where is the enablement state of each procd service stored?
rmilecki has quit [Quit: Konversation terminated!]
<KanjiMonster>
hurricos: enabled (start at boot) or active (running)? former is just a symlink in /etc/rc.d/
<hurricos>
enasbled. /etc/rc.d? Really?
<hurricos>
wtf
<hurricos>
you're reight.
<hurricos>
right*. I don't know why my find -mmin didn't catch it.
<hurricos>
Thanks.
<hurricos>
probably because I type -f'd.
rmilecki has joined #openwrt-devel
<hurricos>
I'm looking to preserve *disablement* status of certain services after a sysupgrade, and I realize now why this is infeasible here, unless I symlink /etc/init.d/65olrd6 to /dev/null or somesuch
<hurricos>
... Let's try that.
<KanjiMonster>
some services have support a enabled uci option to work around that
<hurricos>
Many do, it's true. In my experience olsrd6 is too crufty to have that :(
<Mangix>
Ansuel: porting my archer c7v2 to use qca8k again. How to I verify that the second CPU port is used?
<Ansuel>
Mangix in the ipq806x there is a script to change cpu port
<Ansuel>
if the cpu port works, after changing the port you should have traffic
<Ansuel>
no traffic = cpu port doesn't work
<Mangix>
I must be out of the loop. Does DSA default to round robin assignment?
<Ansuel>
nope still everything to cpu1 and manual assign with ip tools
<Mangix>
so it's not like that turris patch...OK
<Ansuel>
nope but won't change a thing as you can set that from userspace (the idea of upstream is to create a lag of the 2 cpu port) (and that is also doable)
hitech95 has quit [Read error: Connection reset by peer]
<Mangix>
Ansuel: by LAG you mean teaming/bonding the two CPU ports?
<Ansuel>
yes
<Mangix>
I remember KanjiMonstersuggesting that as well. Somewhat weird that it's not default.
<Ansuel>
i prefer giving full gigabit to wan port and shared gigabit to the rest
<jetm>
linusw: Thanks!
schmars[m] has quit [Ping timeout: 480 seconds]
tohojo has quit [Ping timeout: 480 seconds]
magnusk[m] has quit [Ping timeout: 480 seconds]
Tapper has quit [Ping timeout: 480 seconds]
skorpy[m] has quit [Ping timeout: 480 seconds]
aparcar has quit [Ping timeout: 480 seconds]
goliath has quit [Quit: SIGSEGV]
<Mangix>
Ansuel: now to figure out how to add LED definitions to dts. I forgot all of this already
<Ansuel>
ipq806x have 2 device that show how
<Ansuel>
there is a commit dedicated to that
<Mangix>
yeah. i forgot how the one on the archer is configured.
<slh>
the tl-wr1043nd v1 (8/32) comes with RTL8366RB, iirc I've read years ago that realtek_smi should cover it as well
<Mangix>
oh jesus this 01_leds modification looks like a mess
<Mangix>
Ansuel: uhhh how do I run this? I'm doing ~/devstuff/openwrt/scripts/ar8327-initvals.py . . right now
<Ansuel>
create a directory
<Ansuel>
put dts there
<Ansuel>
script -i path to directory
<Mangix>
ah ok
<Mangix>
Ansuel: so that scripts shows me: Custom rule for phy0123 led 0 applied using qca,led_rules, blink-auto, tx-blink, rx-blink, link-10M, link-100M, link-1000M
<Mangix>
I assume ucidef_set_led_netdev "wan-port" "WAN-PORT" "qca8k-0.0:00:green:wan" "wan" "link-10 link-100 link-1000 tx rx" maps to that
<Mangix>
that's going to be fun going through all of these devices
<Mangix>
oh this is interesting. on some devices, led is set to mode rule, but no custom rule.
<Mangix>
wonder what the default rule is.
<slh>
tl-wdr3600/ tl-wdr4300 (ar8327n) were fine with PR#4622, I was using it for a long time (until the rebasing became too much of a problem)
<Mangix>
slh: then stay tuned :)
<slh>
don't worry, I won't go away (at least no longer than a day/ night) ;)
<Mangix>
hrm firstboot is taking a while. I may need to open this up and use serial :(.
<Ansuel>
Mangix consider that most of the time oem randomly set value but actually it's just default
<Ansuel>
i might consider improving the script to report when default value are used
<Mangix>
not working. need serial to figure out what's going on