<csharper2005>
I noticed that tx power with OpenWrt + mt76 driver 8-10dB less than with OEM firmware
<csharper2005>
Device 7602+7612. Both bands affected. Is it possible to increase tx power somewhere in driver code or define tx power delta in DTS for example?
pmanolov has quit [Ping timeout: 480 seconds]
<PaulFertser>
csharper2005: how are you measuring the power?
<csharper2005>
PaulFertser: WiFi analyzer app on my phone.
<PaulFertser>
csharper2005: the limits used by OpenWrt are all taken from the factory calibration data.
<PaulFertser>
csharper2005: probably your device has an additional amplifier which is not enabled?
<csharper2005>
PaulFertser: yeah. Limits are stored in calibration data. Looks like the mediatek and opensource driver handle this data in a different manner. There are no any additional amplifiers on the board. So I need to adapt the calibration data for OpenwpWrt (increase tx pwr) or find a way to make a hack in mt76 driver.
<csharper2005>
I would prefer the second option
<PaulFertser>
csharper2005: is that a board you made and calibrated yourself?
<PaulFertser>
csharper2005: btw, I don't know about mediatek but for QCA there was an additional value "antenna gain" (or something) as part of the calibration data, and it was taken into account when limiting. Also make sure you're not limited by your country regulations, that's something OpenWrt honors too.
<csharper2005>
PaulFertser: I have already read mediatek eeprom guide. Really, antenna gain value exists in eeprom. I tried to change temporarily the countries. Chinese 22 dB in OpenWrt is less than European 20 dB in OEM firmware. :)
<PaulFertser>
csharper2005: I suggest you hang around here, someone who has more knowledge about mediatek power limiting might have an idea.
<PaulFertser>
csharper2005: is antenna gain value reasonable in your case?
pmanolov has joined #openwrt-devel
csharper2005 has quit [Ping timeout: 480 seconds]
csharper2005 has joined #openwrt-devel
csharper2005 has quit [Read error: Connection reset by peer]
danitool has joined #openwrt-devel
Borromini has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
mx has joined #openwrt-devel
pmanolov has quit [Read error: Connection reset by peer]
mx is now known as Guest467
rua has joined #openwrt-devel
csharper2005 has joined #openwrt-devel
csharper2005 has quit [Read error: Connection reset by peer]
csharper2005 has joined #openwrt-devel
owrt-2203-builds has quit [Quit: buildmaster reconfigured: bot disconnecting]
owrt-2203-builds has joined #openwrt-devel
owrt-snap-builds has quit [Quit: buildmaster reconfigured: bot disconnecting]
owrt-snap-builds has joined #openwrt-devel
owrt-1907-builds has quit [Quit: buildmaster reconfigured: bot disconnecting]
owrt-1907-builds has joined #openwrt-devel
owrt-2102-builds has quit [Quit: buildmaster reconfigured: bot disconnecting]
<ynezz>
it's always reproducible on a fresh buildworker, when I spinup new VPS and thus there is no build dir, it happens
<ynezz>
any other consecutive build is then OK
<ynezz>
so if I add 10 new buildworker VPSes, we get 10 build failures like the ones above, then the builds continue working just fine
<zorun>
I had noticed that also, but couldn't figure it out
csharper2005 has quit [Read error: Connection reset by peer]
csharper2005 has joined #openwrt-devel
<csharper2005>
PaulFertser: as a quick fix I changed few values in eeprom and got +- the same power as in stock fw. Thanks for your ideas.
<PaulFertser>
csharper2005: can you please share your findings on a github ticket?
Tapper has joined #openwrt-devel
<csharper2005>
PaulFertser: do you mean some existing ticket or suggest me to create the new one?
<PaulFertser>
csharper2005: a new one
<csharper2005>
PaulFertser: ah, ok. During the investigation I found the similar problems and tickets here: https://github.com/openwrt/mt76/issues/633 maybe I'll share my findings there
<PaulFertser>
csharper2005: good idea, it looks appropriate indeed
csharper2005 has quit [Read error: Connection reset by peer]
csharper2005 has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
<T-Bone>
ynezz: looking at current config, I see there's a bit of cleanup that could be done now that buildbot has been updated. See for instance the comment above gitclean/gitfresh :)
T-Bone is now known as f00b4r0
<f00b4r0>
sadly the explanation link for the gitverify step is a 404
<f00b4r0>
ynezz: may be "normal" but line 16 of the gitfresh step says: "stat: No such file or directory: /builder/malta_be/build/.buildbot-patchedprogram finished with exit code 2"
csharper2005 has quit [Read error: Connection reset by peer]
<f00b4r0>
ynezz: so I can reproduce the problem locally
<f00b4r0>
hmm
owrt-2203-builds has quit [Quit: buildmaster reconfigured: bot disconnecting]
owrt-2203-builds has joined #openwrt-devel
owrt-snap-builds has quit [Quit: buildmaster reconfigured: bot disconnecting]
owrt-snap-builds has joined #openwrt-devel
owrt-1907-builds has quit [Quit: buildmaster reconfigured: bot disconnecting]
owrt-1907-builds has joined #openwrt-devel
owrt-2102-builds has quit [Quit: buildmaster reconfigured: bot disconnecting]
<f00b4r0>
ynezz: I think the failure is normal: it's trying to assert we are on a branch or on a tag, and we checkout a commit (2d03f27f0f0768e25f3b00fb5b4f2974144c66e3) that is on neither.
<f00b4r0>
I'm not clear yet why it would work on second run though.
<f00b4r0>
ynezz: should I keep investigating or have you lost interest? ;P
<ynezz>
f00b4r0: multitasking between 22.03 and IPA brewing :P I would like to get this fixed, indeed
<f00b4r0>
ynezz: as far as I can tell it's already fixed. Could it just have been the buildbot lagging behind your last commit b8f076c9a4a7574bef7d22d6df292dd144737103 ?
<ynezz>
IIRC it has started happening when we bumped from Debian 9 to Debian 10, probably some new Git version behaviour/exit code?
<f00b4r0>
the last failures seem to be "proper" failures (in pkginstall)
<ynezz>
yes, thats due to the b8f076c9a4a7574bef7d22d6df292dd144737103
<ynezz>
but the previous failures are related to that gitverify step
<f00b4r0>
oh i see
<f00b4r0>
git checkout openwrt-22.03 && git rev-parse --abbrev-ref HEAD
<f00b4r0>
git checkout -f b8f076c9a4a7574bef7d22d6df292dd144737103 && git rev-parse --abbrev-ref HEAD
<f00b4r0>
HEAD
<f00b4r0>
in the case where it fails, you can see that gitfresh checks out a specific commit; and that's what causes the problem
<f00b4r0>
where it doesn't fail, gitfresh checks out the branch and there it works
<f00b4r0>
the problem comes from the gitfresh step. Not sure what causes that thouhg
<f00b4r0>
ynezz: does this make sense?
<f00b4r0>
gitverify works as intended. gitfresh did not, for some reason.
csharper2005 has joined #openwrt-devel
Tapper has quit [Ping timeout: 480 seconds]
<ynezz>
f00b4r0: it starts making sense, ther is following `git checkout -f openwrt-22.03 && git branch --set-upstream-to origin/openwrt-22.03 || rm -fr .git`
csharper2005 has quit [Read error: Connection reset by peer]
csharper2005 has joined #openwrt-devel
<f00b4r0>
oh right, that's in gitcheckout
<ynezz>
so that `rm -fr .git` apparently runs even if that checkout succeeds
<f00b4r0>
but that only happens *after* the git folder has been checked out
<f00b4r0>
no, that command does not run on the first start of the builder
<f00b4r0>
*that* is why it does nothing for the first run, and only for the second ones
<ynezz>
ah
<f00b4r0>
the phase1 config really needs a thorough cleanup ;P
<ynezz>
still it's probably wrong, shellcheck complains "SC2015 (info): Note that A && B || C is not if-then-else. C may run when A is true."
<f00b4r0>
C may run if A is true and B is false, yes
<f00b4r0>
but that's ok. I think the issue is that this command is only run after the second call of the builder
<f00b4r0>
which is why only on the second runs it does try to checkout the branch instead of the head commit
<stintel>
neggles: never got it to work (didn't really try either)
<f00b4r0>
so either the gitverify step is not needed (and having access to the original email would help assess that), or we need to review the logic here
<f00b4r0>
ynezz: truth is, I want to clean up this whole phase1 thing, but my incentive is when I start my buildbot again. Right now I'm still waiting for fiber (and I need to replace failing SAS drives), so incentive currently is slightly lower ;P
csharper2005 has quit [Read error: Connection reset by peer]
csharper2005 has joined #openwrt-devel
Borromini has quit [Quit: Lost terminal]
csharper2005 has quit [Read error: Connection reset by peer]
<f00b4r0>
ynezz: basically the current setup only "works" as intended when switchtag is executed, which is not the case here since we're not on a tag. The quick fix is to rerun the gitcheckout step immediately before the gitverify one, but that's a hack. This really needs some cleanup so we interfere less with buildbot's Git() routines (the main cause for this problem).
<f00b4r0>
ynezz: HTH, I have to run some Sunday errands :)
csharper2005 has joined #openwrt-devel
csharper2005 has quit [Read error: Connection reset by peer]
csharper2005 has quit [Read error: Connection reset by peer]
csharper2005 has joined #openwrt-devel
csharper2005 has quit [Read error: Connection reset by peer]
csharper2005 has joined #openwrt-devel
<csharper2005>
PaulFertser: 5 GHz was a challenge but I also won. lol. 28 dBm in OpenWrt is equal 20 dBm in stock fw. I'll comment the issue a bit later.
robimarko has joined #openwrt-devel
<csharper2005>
PaulFertser: By the way, do you know how to upstream patches to Linux?
rua has quit [Ping timeout: 480 seconds]
<PaulFertser>
csharper2005: yes, usually you can just use scripts/get-maintainers.pl and it'll tell you all the addresses that should be CCed on patches.
<csharper2005>
PaulFertser: ok, thanks. The patch needs to be tested, right? I have a patch related to OpenWrt only. How to test it on actual linux-next kernel?
<csharper2005>
Just check that Linux compiles successfully, without errors and warnings?
<PaulFertser>
csharper2005: hm, it's better to test it, yes, but it might be problematic with linux-next, I see how. I suggest you wait for someone who has a better idea than me.
<PaulFertser>
csharper2005: if the patch is OpenWrt-specific no need to test it on linux-next.
rua has joined #openwrt-devel
csharper2005 has quit [Read error: Connection reset by peer]
csharper2005 has joined #openwrt-devel
ephemer0l has quit [Ping timeout: 480 seconds]
<csharper2005>
PaulFertser: good news. It's openwrt-specific mtd parser for netgear-sercomm devices.
<PaulFertser>
csharper2005: in this case just follow the OpenWrt wiki on submitting patches.
Znullptr has quit [Quit: Connection closed for inactivity]
<Slimey>
im trying to tack down what it is but something i selected in make menuconfig makes successful builds that are missing kernel mods
srslypascal is now known as Guest487
srslypascal has joined #openwrt-devel
Guest487 has quit [Ping timeout: 480 seconds]
Tapper has joined #openwrt-devel
<Slimey>
it looks like nmap-full can someone confirm with ath79 target?
csharper2005 has quit [Read error: Connection reset by peer]
csharper2005 has joined #openwrt-devel
<neoraider>
Is it possible that our build system currently doesn't automatically pick up changes to the sources of the various image loaders (lzma-loader, relocate), so you need to clean to ensure a rebuild?
csharper2005 has quit [Ping timeout: 480 seconds]
csharper2005 has joined #openwrt-devel
<neoraider>
And if that's the case, any ideas how to fix it? I'm not too familiar with that part of the image build code
dangole has quit [Ping timeout: 480 seconds]
<mangix>
ramips 5.15 kernel. oh man
<mangix>
Something tells me I shouldn't try this
Larhzu has quit [Quit: leaving]
Larhzu has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
<Grommish>
mangix: Any hints on creating a config-5.15 stub for a target? When I try, I get kconfig.pl errors
rua has joined #openwrt-devel
<tmn505>
neoraider: yes, it's not rebuilt, it bit me three weeks ago. I don't have final solution, but You can remove loader-* from $(KDIR) before each rebuild.
<mangix>
Grommish: nope
pmanolov has quit [Quit: WeeChat 3.0]
csharper2005 has quit [Read error: Connection reset by peer]
csharper2005 has joined #openwrt-devel
csharper2005 has quit [Read error: Connection reset by peer]
csharper2005 has joined #openwrt-devel
csharper2005 has quit [Read error: Connection reset by peer]
csharper2005 has joined #openwrt-devel
Tapper has quit [Ping timeout: 480 seconds]
<dansan>
Hello all! I've got a bunch of OpenWRT build trees and I'm trying to optimize my system with a tiered cache. I moved my entire /home filesystem an LVM volume an HDD raid5 array and then ran it through a 1T lvmcache volume and while the reads are great, it's horribly slow when I'm building OpenWRT and there's a lot of writes -- especially since I have CONFIG_DEBUG enabled (-g3). Is anybody else using bcache or any other tiered cache in their build
<dansan>
setup?
<dansan>
I've been using all of these ugly bind mounts and symlinks in my /home filesystem to put some things on SSD and others on HDD (it won't all fit on my SSDs), so I figured a tiered cache is what I need to clean this mess up and still have good build performance
<neoraider>
tmn505: thanks for the confirmation. I guess a simple solution would be to always force rebuilding the loaders, as their build time is fairly insignificant
GNUmoon2 has quit [Ping timeout: 480 seconds]
<neoraider>
There, fixed (at least in my staging tree) https://git.openwrt.org/?p=openwrt/staging/neoraider.git;a=commitdiff;h=6654fab7486c46a964436039d981d9b92d24269f;hp=02fed0a10a2cad56086b7d3759dcdba052abc001