08:51 <f00b4r0> https://openwrt.org/docs/guide-user/base-system/uci#file_syntax is this the closest thing we have to a grammar definition for uci config files? :)

08:55 <dwfreed> probably

08:56 <dwfreed> it's at least fully specified, so you could make an ABNF from it

08:57 <f00b4r0> I'm planning on a quick LALR validator, courtesy of flex/bison :)

08:57 <dwfreed> heh

08:58 <dwfreed> there is one thing not noted there; 'package' lines are only useful on import; you wouldn't normally use them in the actual /etc/config files

08:59 <f00b4r0> *nod*

09:11 <mrkiko> Ansuel: thanks a lot for the VDSL fix, I admit i should still understand how the reasoning goes for the macros part... I was worried we lost the VDSL part this time

13:38 <KGB-2> https://tests.reproducible-builds.org/openwrt/openwrt_kirkwood.html has been updated. (100.0% images and 99.9% packages reproducible in our current test framework.)

17:36 <mrnuke> robimarko: When debugging QCA clock issues, do you have a method to go from dmesg error to the casue of the problem?

17:58 <PaulFertser> mrnuke: what do you mean, what kind of error?

18:02 <mrnuke> PaulFertser: A variety of "status stuck at 'off'", "can't find a configuration for rate 125000000"

18:03 <mrnuke> I should probable mention that I am using a frankenstein tree for IPQ95xx

18:10 <PaulFertser> mrnuke: an obvious first idea would be to: git grep "find a configuration for rate"

18:13 <PaulFertser> mrnuke: and if it's not obvious where it's called from you can add WARN_ON_ONCE(1) to see the backtrace at that moment

18:17 <mrnuke> I get plenty of bactraces. They point to drivers/clk/qcom/clk-rcg2. But that's just a loop parsing the data structures grom other gcc-* and nss-* clock drivers. And I suspect something in those structures might be incorrect

18:17 <mrnuke> For fun, this is one example: https://gist.github.com/mrnuke/225375c1cd144d9aa7cb632e5e9fd78f

18:30 <PaulFertser> mrnuke: well see drivers/clk/qcom/clk-branch.c:87 and check manually if there's a configuration you need present in the source or wherever it's fetched from.

18:31 <PaulFertser> But this is not missing rate, this is just clock not enabled, it tries to enable the clock but apparently checks and waits and it's still not enabled so "nss_cc_port5_rx_clk status stuck at 'off'"

18:45 <mrnuke> I don't understand the reason for it being or not being enabled. I remember there being a workaround on older IPQ SoC, where .fw_name was used along with .name, but I don't understand that one either

20:24 <philipp64> jow: I tried to address your review comments where I understood them, but a couple weren't clear to me. Can you please update https://github.com/openwrt/libubox/pull/6 ?