<ynezz>
fda: in other words, PR is not currently possible
<fda>
great, thanks. i have local AM format
<fda>
well, i have no mail in this vm... i hope just append the patch works
<PaulFertser>
fda: you can send it yourself and see if "git am" is happy with it
<PaulFertser>
fda: but it's also very much desired for Patchwork to pick it up so probably a mime attachment is not appropriate
<PaulFertser>
fda: you can use "git format-patch" and then "git send-email <thatpatch>" from a host where you do have a functional MTA.
<fda>
thx. my linux devices have no mail password, so they could only send local mails and no forward... i have to test a little
<PaulFertser>
It doesn't have to run Linux, git send-email should work the same on other kernels too.
<fda>
seems i have to use send-mail directly ... as html mails are not accepted :D
<PaulFertser>
mime attachments and html is orthogonal
<fda>
mailinglists are always complicated
<fda>
can i add a comment to an yet existing patch at patchwork.ozlabs.org by login (i have none) or only by mail?
<aiyion>
I toyed around with a commit of karlp from yesterday; and found a working combination of how to use simple-mfd and a gmac section that works with the onion omega.
<aiyion>
I applied my mac-adress again and now I have proper lan without ghosting interfaces.
<Tusker>
good job :)
<aiyion>
One thing that bugs me though is how long it takes br-lan to go to work.
<aiyion>
After 15 Seconds the boot console becomes rather silent;
<aiyion>
after 74 seconds jffs finds the end of the fie system;
<aiyion>
after 83 seconds br-lan becomes ready.
<aiyion>
That's an awful lot of time and feels like something timesout after a minute or so, which then kicks br-lan.
<aiyion>
This si the boot log I'm talking about; as well as the console activation after everything is done:
<aiyion>
Tusker: thanks, trying to run luci right now; will test for gpios of expansion boards afterwords.
<aiyion>
Well, at least luci works out of the box :)
Rentong has quit [Ping timeout: 480 seconds]
<fda>
aiyion: for me it takes only so long after sysupgrade
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
abiliomarques_ has joined #openwrt-devel
abiliomarques has quit [Read error: Connection reset by peer]
<aiyion>
that's good to know, I'll keep an eye on it.
Tusker has quit [Quit: Time wasted on IRC: 12 hours 1 minute 29 seconds]
Grommish has quit [Read error: Connection reset by peer]
hitech95 has joined #openwrt-devel
<hitech95>
Hi guys I have a question about the network configuration stack. this is about MAP! Is there a way sto store additional information about the interface? Right now there is no wat to get esay mapped infos like public ip on such interface.
abiliomarques_ has quit [Remote host closed the connection]
abiliomarques_ has joined #openwrt-devel
<aiyion>
I poked around the gpios of the board and found in addtition to whats already in ar71xx three gpios that control an rgb led on the dock-expansion.
<aiyion>
Is there info on how to reflect rgb leds in dts?
<aiyion>
I'd go with ignoring they're part of one led, as they're controlled using three different gpios
abiliomarques__ has joined #openwrt-devel
abiliomarques_ has quit [Read error: Connection reset by peer]
<abiliomarques__>
I was tying a performance issue with OpenWRT in a MIPS, compared to the original Qualcomm code... managed to narrow the issue to musl ... somehow the compiler is issuing the instruction RDHWR, which is not supported by a lot of MIPS, and instead emulated by the kernel
<abiliomarques__>
I went to check my other MIPS based devices running OpenWRT, and it seems all of them suffer from it. Has anyone seen this before?
<abiliomarques__>
to be specific, is an RDHWR to the UserLocal register
abiliomarques has joined #openwrt-devel
abiliomarques__ has quit [Read error: Connection reset by peer]
abiliomarques_ has joined #openwrt-devel
abiliomarques has quit [Read error: Connection reset by peer]
<PaulFertser>
abiliomarques_: interesting, what is your testcase? Seems to be quite a specific workload.
<abiliomarques_>
PaulFertser: first noticed slower I/O... though it was the flash driver, but doing dd from /dev/zero to /dev/null was also slow... kept looking, found that issuing syscalls was slower... profiled the kernel, found it calling a lot to handle_ri_int ... with some printk, it showed it was executing 0x7Cyye83b, which maps with RDHWR
<abiliomarques_>
basically musl uses it in a lot of places
<PaulFertser>
abiliomarques_: amazing work
<abiliomarques_>
rdhwr with a rd=29
<abiliomarques_>
musl is using this for __pthread_self()
<abiliomarques_>
it seems they expect the kernel to emulate it ... the original uclibc didn't seem to use this at all
<aiyion>
Is using github-PRs preferred over mailinglist for adding device support these days?
Tapper has quit [Ping timeout: 480 seconds]
valku has joined #openwrt-devel
<PaulFertser>
abiliomarques_: sounds like a musl missing optimisation opportunity. I suggest you discuss that on their mailing list, and if you find a workaround probably OpenWrt can pick it up till upstream releases a new version with that incorporated.
<PaulFertser>
aiyion: why do you think it might be so?
<hitech95>
jow: Is there a way to customize the LUCI interface render status for a specific protocol? (both in luci-mod-status & luci-mod-network). I'm looking a way to show available ports in case of MAP-T
<aiyion>
subjective hint of a colleague, that GH was usually faster.
<aiyion>
I take that as a now?
<aiyion>
*no
<PaulFertser>
aiyion: faster for OpenWrt new device support specifically?
<aiyion>
no i general I think.
<aiyion>
*in.
<PaulFertser>
aiyion: then I do not understand the statement. In general for all the projects that accept patches both via "github pull requests" and the mailing list?
abiliomarques_ has quit [Ping timeout: 480 seconds]
<aiyion>
C. said "use github, usually my contributions get done faster that way". My question was if provided, thats a larger trend, GH might be preferred over ML.
abiliomarques has joined #openwrt-devel
<aiyion>
It was not my intention to offend.
<PaulFertser>
aiyion: I know some core devs do not look at github pull requests at all. Some are quite active about it. I think all the core devs are looking at the regular patches. But I can't tell you what way is faster currently.
<aiyion>
thx
<PaulFertser>
aiyion: I wonder what your C. says about projects that do not use github at all :)
<aiyion>
mh
<PaulFertser>
aiyion: so have you found a way to properly enable the ethernet and mdio without ghost interfaces?
<abiliomarques>
PaulFertser: I've been having connectivity issues ... did you say anything else since my last sentence?
nitroshift has quit [Quit: Gone that way --->]
<PaulFertser>
abiliomarques_: sounds like a musl missing optimisation opportunity. I suggest you discuss that on their mailing list, and if you find a workaround probably OpenWrt can pick it up till upstream releases a new version with that incorporated.
<aiyion>
PaulFertser: the relevant part was "compatible = "syscon", "simple-mfd";" for eth0 and the gmac portion for eth1
<aiyion>
It's not 100% it, the id in the bootlog is a bunch of zeros this way; but it does work.
<PaulFertser>
Probably something's buggy with the existing bindings and/or drivers.
Rentong has joined #openwrt-devel
Rentong has quit [Remote host closed the connection]
minimal has joined #openwrt-devel
Rentong has joined #openwrt-devel
Rentong has quit [Ping timeout: 480 seconds]
danitool has joined #openwrt-devel
<aiyion>
openwrt demands clear names for patches, right?
Rentong has joined #openwrt-devel
Rentong has quit [Remote host closed the connection]
Rentong has joined #openwrt-devel
mkresin has quit [Quit: leaving]
mkresin has joined #openwrt-devel
mkresin_ has joined #openwrt-devel
mkresin_ has quit []
Rentong has quit [Ping timeout: 480 seconds]
<abiliomarques>
PaulFertser: looking a bit deeper in the code, rdhwr $3, $29 seem to be emulated by the kernel way faster than to any other register ... mentioned that to the maintainer of that part of Musl, we're doing an experiment right now
goliath has quit [Quit: SIGSEGV]
ecloud_ has quit [Ping timeout: 480 seconds]
ecloud has joined #openwrt-devel
<karlp>
aiyion: glad to hear it's working! re boot time, is that just first boot? that's always a lot longer with jffs2, but secondary boots should be pretty quick, but it's never going to be as quick as an emmc or anyting...
<aiyion>
appears to be just the first yes.
<karlp>
aiyion: re: your leds on an expansion, remember that the dts should onle be for the module, you can't try and have all combinations of shields/docks in the base dts.
<aiyion>
Just realized I never migrated my mailinglist participation from lede devel to openwrt devel...
<karlp>
device tree overlays are the "normal" solution to this sort of thing....
<aiyion>
huh.
<aiyion>
well the gpios match for both the powerdock and expansion dock;
<aiyion>
but I might remove it as well.
Rentong has joined #openwrt-devel
rejoicetreat has quit [Ping timeout: 480 seconds]
rejoicetreat has joined #openwrt-devel
<abiliomarques>
PaulFertser: performance is now the same as uclibc
hitech95 has quit [Remote host closed the connection]
<PaulFertser>
abiliomarques: awesome :)
<abiliomarques>
is good, all my routers will be faster now :)
<abiliomarques>
just by accident
<abiliomarques>
who do I keep posted about the patch being available?
<PaulFertser>
abiliomarques: and by your skills and willingness to dig deep!
<PaulFertser>
abiliomarques: I think for OpenWrt the best would be to have a ticket on bugs.openwrt.org regarding that.
<abiliomarques>
I will try to keep an eye on musl, the maintainer there seems quite responsive and positive about the change
<abiliomarques>
it wouldn't have been possible without the help of a lot of people
floof58 has quit [Ping timeout: 480 seconds]
Tapper has joined #openwrt-devel
goliath has joined #openwrt-devel
Rentong has quit [Remote host closed the connection]
rejoicetreat has quit [Ping timeout: 480 seconds]
goliath has quit [Quit: SIGSEGV]
floof58 has joined #openwrt-devel
DragoonAethis has joined #openwrt-devel
rejoicetreat has joined #openwrt-devel
rejoicetreat has quit [Ping timeout: 480 seconds]
minimal has quit []
rejoicetreat has joined #openwrt-devel
Borromini has joined #openwrt-devel
goliath has joined #openwrt-devel
rejoicetreat has quit [Ping timeout: 480 seconds]
slh has quit [Quit: leaving]
slh64 has quit [Quit: gone]
Tapper has quit [Ping timeout: 480 seconds]
Tapper has joined #openwrt-devel
<nick[m]12>
blocktrron_: do you have an idea what is "RxBadByte" and it is very high on a ipq40xx soc (fritzbox 7530)
slh has joined #openwrt-devel
jlsalvador2 has joined #openwrt-devel
jlsalvador has quit [Read error: Connection reset by peer]
jlsalvador2 is now known as jlsalvador
KGB-0 has quit [Quit: KGB-0]
KGB-0 has joined #openwrt-devel
Borromini has quit [Quit: Lost terminal]
jlsalvador2 has joined #openwrt-devel
jlsalvador has quit [Read error: Connection reset by peer]
<mrkiko>
Hello all! Quick ath79 question - who uses the clock-frequency property in the &ref node? And how can I discover how to define it for a device?
<mrkiko>
I see it in different .dts* files in ath79, but can't find where it's used on the code
abiliomarques has quit [Ping timeout: 480 seconds]