philipp64|work has quit [Ping timeout: 240 seconds]
philipp64|work_ is now known as philipp64|work
<hgl>
i'm trying to understand how stopping an interface in LuCI actually works. but from the js code in interface.js, it seems all it does is to update the button UI? i wonder where is the ubus calling code to actually stop the interface?
vdl has quit [Read error: Connection reset by peer]
vdl has joined #openwrt-devel
_whitelogger has joined #openwrt-devel
ds_shadof has joined #openwrt-devel
Guest40023 is now known as lep-delete
lep-delete is now known as Guest40023
Guest40023 is now known as lep-delete
decke has joined #openwrt-devel
<damex>
mangix: is 'older' ath10k cards considered good? i salvaged qca9880-br4a (oversize, 3 antenna and mpci-e connector) from old archer c2 router a while ago ;p
<damex>
although it is 5ghz only ;/
<damex>
c7 v2 *
<pkgadd>
it's not quite as good as more modern wave2 variants (e.g. qca9984), QCA wave1 chipsets as a whole seem to be plagued more by firmware issues - but it's still a very usable wireless chipset, *if* you extracted its ART partition from the donor device. without ART, it only serves as a paperweight
<damex>
oh, i haven't saved its ART partition :(
<pkgadd>
then you could just as well nail it to the wall, as decoration
<damex>
so ART have calibrations? can't we do it in software?
<pkgadd>
no
<damex>
i just have ath9k 9380 and it does not require anything external (okay, ath10k needs firmware blobs) - so ath10k is not just external blob but also vendor calibrations?
<pkgadd>
'normal' PCIe cards (e.g. ones that are meant to go into notebooks or the aftermarket) ship their calibration (which is unique to your individual device, determined at production time) on a tiny EPROM on the card, but on routers the vendors are saving the pennies for that and include this vital data in the router's main flash storage instead - lose it and it's done
<pkgadd>
ar9380 needs the same kind of device specific calibration data
goliath has joined #openwrt-devel
<damex>
pkgadd: hmm.. it seem to perform pretty well as a client device (laptops and such) or ap on pci-e slot. used it for years this way. just compile ath9k and you're free to go
<damex>
could it be present on device?
<pkgadd>
06:07:50 < pkgadd> 'normal' PCIe cards (e.g. ones that are meant to go into notebooks or the aftermarket) ship their calibration (which is unique to your individual device, determined at production time) on a tiny EPROM on the card, but on routers the vendors are saving the pennies for that and include this vital data in the router's main flash storage instead - lose it and it's done
<damex>
so calibration is on tiny eeprom then ... i guess
<pkgadd>
yes
<pkgadd>
it is on 'normal' PCIe/ mini-PCIe cards, but not in the one you salvaged from an archer c7-v2
<pkgadd>
(very modern netbook-style notebooks and tablets might also be in this can, storing the calibration data in the main system's UEFI flash efivar backend, but that's less common. in router saving this tiny EEPROM is the norm, not the exception)
<damex>
is there a way to find out if that card have a tiny eeprom or require external one by examining it physically?
<damex>
can't tell. there is a tiny chip on the left side that resemble a eeprom and it is not marked as such
<pkgadd>
sure, just look for a small spi-nor chip on the wireless card
<pkgadd>
that's not a guarantee that it's used for this purpose, but a strong indication
<pkgadd>
but in case of the oversized mini-PCIe-like card of the archer c7-v2, it's known where the calibration data resides - in the ART partition of the main router board, not the wireless card itself
<pkgadd>
unless you dig out the donor router from the boneyard again - and extract this vital data from its main flash (with an external spi-nor writer, if you need to), the card is useless
<damex>
(i should have kept the backup of that spi flash)
<pkgadd>
(and even then you'll have a little fun of uploading the calibration at module load time correctly, something general purpose linux distributions aren't usually equipped to do (owl-loader) - doesn't mean that it can't be done, if your have the calibration data, but it's custom hackery)
Darkmatter66 has joined #openwrt-devel
<pkgadd>
needless to mention that windows drivers won't give you the option to load this calibration from non-EPROM location anyways
<damex>
well, router along with its spi flash is long dead
<damex>
for 3 years already or even 4
<pkgadd>
well, then it's time to let the card follow suit - it's done
<damex>
TIL not all pci-e wireless cards created equal
<pkgadd>
router manufacturers buy these card in quantities of tens of thousands to hundreds of thousands, saving a few cents by using one (larger) flash chip for everything (which is needed anyways) and omitting the 64 KB EPROM makes financial sense. re-using the card is not intended - and it's oversized/ needing more power than normal and can't be used in normal PC hardware anyways
dedeckeh has joined #openwrt-devel
nitroshift has joined #openwrt-devel
<damex>
it does not look/sound good. dead spi render every other piece of hardware useless (that is actually external to the board)
<damex>
so you're forced to buy a new hardware instead of reusing -for-parts-
<pkgadd>
since when do manufacturers care about that? they see the device as one unit, in sealed condition
<damex>
pkgadd: well, some of them do care. lenovo used to ship parts for their laptops if needed (as a replacement in the field aka FRU) but as for the routers... yeah, they don't care about it :(
poljar has quit [Remote host closed the connection]
poljar has joined #openwrt-devel
<gch9812135>
damex: Those pcie cards on routers can't be used in a normal minipcie slot even if you have a copy of the caldata. They use a NC pin to supply +5v for PAs.
gch9812135 is now known as gch981213
Borromini has joined #openwrt-devel
Ycarus has joined #openwrt-devel
<pkgadd>
thanks for clarifying that further (I knew that they're oversized, needing more power and need more ventilation than a normal card, but didn't know about those additional voltage pins)
<PaulFertser>
Haha, nice to know. Though getting another 5 V line shouldn't be too problematic
<damex>
shouldn't caldata be read only like ROM? ;/
<PaulFertser>
gch981213: do you also know if trying to use ath10k from another same model device is totally impractical?
<PaulFertser>
damex: it is sometimes called Flash ROM :)
gch9812132 has joined #openwrt-devel
rsalvaterra has joined #openwrt-devel
gch981213 has quit [Read error: Connection reset by peer]
gch9812132 is now known as gch981213
<PaulFertser>
I meant to ask about using ath10k caldata from another device, is that not a good idea to even try?
valku has quit [Quit: valku]
<Tapper>
Hi what hapend to the kernel bump I saw on the ML? The pr was closed but it's not in master.
<svanheule[m]>
Borromini: pong
<nitroshift>
Tapper, what kernel bump? major version or minor one?
<Tapper>
miner
<Tapper>
If there is nothing to test I wate for a kernel bump to build a new bin and flash.
<Tapper>
I don't flash everry bump but I know that there is a bump on the way so mite as well wate for it. lol
<pkgadd>
[PATCH] kernel: bump 5.4 to 5.4.53 was the last bump on the mailing list, master it at 5.4.69 (and kernel.org on 5.4.70)
<nitroshift>
pkgadd, thanks, thought there was a pr for a kernel major version bump
gch981213 has quit [Read error: Connection reset by peer]
gch9812134 is now known as gch981213
<damex>
how is the process of merging device support (new device) works? let's say it works well, all adjustments are done. do we need someone (beside person who open pull request) to test changes for it to be merged?
<nitroshift>
pkgadd, thanks
<PaulFertser>
damex: no, if your code looks sane and there're no more comments regarding it, the support is just merged.
<lemmi>
damex: did you have any more progress with your er-4? i have a er-12 sitting here which i managed to boot from usb, but it also has a switch chip in there that need proper config
<lemmi>
damex: on the er-12 the last 2 rj45 ports and the 2 sfp ports work. the first 8 ports hangig on the switch don't
dopje_ has joined #openwrt-devel
danitool has joined #openwrt-devel
<damex>
lemmi: i wonder... how does the all ports connected internally? 7130 have just two SGMII (QSMII?) coming from SOC
<damex>
also interesting if they use the same vsc8504 or opt for something else
<lemmi>
damex: i'm a little lost on this as well, which is why i'm currently bidding on an er-4 in the hopes it's a little easier to get going
dopje has quit [Ping timeout: 258 seconds]
<damex>
lemmi: can you show output of uboot's printenv, mii device and mdio list ?
<pkgadd>
to the best of my knowledge v5.9 hasn't been formally declared LTS yet - yes, it's supposed to be the last stable kernel of the year and v5.10 will probably only arrive in mid january, but... close enough
<lemmi>
damex: i'll have the er-12 on the bench this weekend again
<lemmi>
i'm meeting a couple of people, maybe we can figure out something useful
<enyc>
pkgadd: hrrm but ... openwrt-20 to remain 5.4 ?
aszeszo has quit [Quit: aszeszo]
<pkgadd>
enyc: yes
<damex>
is there an easy way to test device target on a new kernel ?
<pkgadd>
no
<Borromini>
enyc: you don't just switch to another kernel in a jiffy
feriman has joined #openwrt-devel
psnsilva has quit [Remote host closed the connection]
cyrozap has quit [Ping timeout: 260 seconds]
cyrozap has joined #openwrt-devel
cyrozap is now known as Guest30462
gch9812130 has joined #openwrt-devel
gch981213 has quit [Read error: Connection reset by peer]
gch9812130 is now known as gch981213
aszeszo has joined #openwrt-devel
<Tapper>
pkgadd sorry mate It was not on the ML it was on one of the emails I get from git.
<Tapper>
It's has bin puled into the staging tree of @blocktrron
dedeckeh has quit [Remote host closed the connection]
kubrickdave_ has joined #openwrt-devel
kubrickdave has quit [Ping timeout: 264 seconds]
Nick_Lowe has joined #openwrt-devel
<enyc>
Borromini: ok yes that fits with my guesses
dedeckeh has joined #openwrt-devel
<damex>
PaulFertser: https://github.com/openwrt/openwrt/pull/3482 i am just looking toward merging this one. that is why i ask about what is needed to merge changes. there does not seem much left to do (i am using that target daily) beside making someone else test it. currently @pepe2k helping with testing the target and then hopefully we could merge.
<PaulFertser>
damex: yeah. Please keep in mind that not all maintainers like to work with "github pull requests", most prefer the mailing list.
Borromini has quit [Ping timeout: 240 seconds]
nitroshift has quit [Quit: Gone that way --->]
Borromini has joined #openwrt-devel
gch9812130 has joined #openwrt-devel
gch981213 has quit [Read error: Connection reset by peer]
gch9812130 is now known as gch981213
am0rphis has joined #openwrt-devel
svanheule[m] has quit [Quit: killed]
agb[m] has quit [Quit: killed]
JuniorJPDJ has quit [Quit: killed]
fblaese has quit [Quit: killed]
nick[m] has quit [Quit: killed]
decke[m] has quit [Quit: killed]
WiredLife has quit [Quit: killed]
shalzz has quit [Quit: killed]
voltagex has quit [Quit: killed]
sielicki has quit [Quit: killed]
aparcar[m] has quit [Quit: killed]
pgwipeout[m] has quit [Quit: killed]
olmari has quit [Quit: killed]
pavlix has quit [Quit: killed]
Nick_Lowe has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
JuniorJPDJ has joined #openwrt-devel
Nick_Lowe has joined #openwrt-devel
adrianschmutzler has joined #openwrt-devel
gch9812136 has joined #openwrt-devel
gch981213 has quit [Ping timeout: 240 seconds]
gch9812136 is now known as gch981213
dedeckeh has quit [Remote host closed the connection]
<damex>
i see there is support for raspberrypi4. am i looking at something wrong or there is no support for upgrading its bootloader? all you need to do is place correct binary with sha next to it on a fat32 partition you boot from
<PaulFertser>
If rpi-eeprom is not mentioned in the whole git tree then it's likely it's just not there.
<damex>
yeah, it is not
<damex>
what could be preffered way to have it in a tree?
<PaulFertser>
Creating a package.
<damex>
so package for 'stable version of rpi-eeprom' and package for 'stable version of vl805 usb3.0 controller' ?
<damex>
and just bump versions as they update them?
<stintel>
utils/bcm27xx-eeprom/Makefile
<stintel>
is in the packages feed
Nick_Lowe has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Nick_Lowe has joined #openwrt-devel
Borromini has quit [Quit: leaving]
<damex>
nice, thanks, i can just it on my rpis then ;)
<rsalvaterra>
Right, this is awkward… I want to build a specific package, but for some reason it's not working. :/
<rsalvaterra>
So, I manually cloned the packages feed and pointed to it in feeds.conf.
<rsalvaterra>
(It's the only feed I have.)
<rsalvaterra>
Now, I want to build a specific package from that feed, but I not being able to specify the correct path (make V=s <path>/{clean,compile}).
<rsalvaterra>
The normal build works just fine, of course.
gch9812130 has joined #openwrt-devel
gch981213 has quit [Ping timeout: 244 seconds]
gch9812130 is now known as gch981213
<stintel>
it's not necessarily <path>/{clean,compile}
<stintel>
I believe it is package/PKG_NAME/...
<PaulFertser>
Yes, and before that you need scripts/feeds install -a
<PaulFertser>
And select the package you need with =m in config.
<rsalvaterra>
stintel: Hey! :) I tried lots of "paths", but nothing worked… It's probably something trivial and obvious. :P
<stintel>
been there done that ;-)
<stintel>
what package are we talking about
<rsalvaterra>
When building an image, this is the output I get (for that package)…
luke-jr has quit [Read error: Connection reset by peer]
luke-jr has joined #openwrt-devel
Nick_Lowe has joined #openwrt-devel
luke-jr has quit [Read error: Connection reset by peer]
luke-jr has joined #openwrt-devel
<Grommish>
rsalvaterra: the ./package directory is wierd.. its package/feeds/packages/<pkgname> or package/<core>/<pkgname> They don't sub-divide like the feeds/... structure does
<Grommish>
and yes, I know it's 4 hours late, but you're just getting into the good part of the day where you are ;p
feriman has quit [Ping timeout: 240 seconds]
mwarning has quit [Quit: Leaving.]
Grommish has quit [Ping timeout: 272 seconds]
Grommish has joined #openwrt-devel
<karlp>
just make package/<pkgname/{clean,compile} works.
<gch981213>
PaulFertser: do you mean using calibration data from another same model device? That would be the best option if you've lost the original caldata. When producing routers, vendors only calibrate one board and the result is used for the same batch of device produced.
Darkmatter66 has quit [Ping timeout: 240 seconds]
gch9812137 has joined #openwrt-devel
Darkmatter66 has joined #openwrt-devel
<PaulFertser>
gch981213: yes, that's what I meant. Thanks for clarifying!
<PaulFertser>
damex: ^^
gch981213 has quit [Ping timeout: 260 seconds]
gch9812137 is now known as gch981213
Immanuel has quit [Quit: Connection reset by reptilians]
Immanuel has joined #openwrt-devel
silverwhitefish has joined #openwrt-devel
Borromini has joined #openwrt-devel
merbanan has quit [Ping timeout: 240 seconds]
Borromini has quit [Quit: Lost terminal]
merbanan has joined #openwrt-devel
feriman has joined #openwrt-devel
Ycarus has quit [Quit: Ycarus]
Nick_Lowe has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
dangole has joined #openwrt-devel
dangole has quit [Remote host closed the connection]
<Hauke>
aparcar[m]: did you also add CONFIG_F2FS_FS_XATTR=y to the generic config?
dottedmag has joined #openwrt-devel
<Hauke>
when I build CONFIG_TARGET_mvebu_cortexa53_DEVICE_globalscale_espressobin=y this is missing
<dottedmag>
I've got a Ubiquity USG-Pro 4 device that is quite similar to their EdgeRouter Pro (the only difference I can see is 4 Ethernet ports instead of 8). target/linux/octeon/base-files/lib/preinit/01_sysinfo matches on /proc/cpuinfo "system type", and they seem to be identical for both devices, though Wiki page for ERP seems to have a wrong boot log pasted. What would be a good way to distinguish them further?
silverwhitefish has quit [Remote host closed the connection]
silverwhitefish has joined #openwrt-devel
lnslbrty has quit [Quit: ~-AmIsraelChai-~]
am0rphis has quit [Remote host closed the connection]
silverwhitefish has quit [Read error: Connection reset by peer]
am0rphis has joined #openwrt-devel
aszeszo has quit [Quit: aszeszo]
dangole has joined #openwrt-devel
Nick_Lowe has joined #openwrt-devel
Nick_Lowe has quit [Client Quit]
Nick_Lowe has joined #openwrt-devel
aszeszo has joined #openwrt-devel
<dottedmag>
OTOH, is there any reason to distinguish two devices that differ only in number of Ethernet ports?
Nick_Lowe has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
goliath has quit [Quit: SIGSEGV]
dangole has quit [Remote host closed the connection]
<aparcar[m]>
Hauke: seems missing, please readd
<aparcar[m]>
Odd, I guess it was missing in the config and so wasn't covered via the sed replacement.