05:52 <mvaittin> marex: Do I dare to ask ... Any progress with the critical-clocks?

06:31 <marex> mvaittin: dare to ask sboyd

06:52 <mvaittin> marex: I'll take that as no visible progress :(

06:53 <marex> sboyd: ^

06:54 <mvaittin> marex: I was just asking this because I see someone else who hit issues - and sent a patch to overcome this with of_clk_detect_critical()

06:55 <mvaittin> https://lore.kernel.org/lkml/5f34b6d6-c2dd-44f9-c1bc-fe1deb336334@gmail.com/

06:55 <marex> mvaittin: I know

10:53 <ardb> arnd: i did the experiment of making ALT_SMP() issue a 'brk' instruction throughout, and

10:53 <ardb> everything seems to work happily

10:54 <ardb> booted on a UP system, none of the BRKs are ever hit

15:55 <hushmoney> i have a board with an i2c gpio expander and an i2c mux, one channel of that i2c mux addresses two devices with the same address, but they are enabled with pins on the gpio expander. i'm inexperienced in this area so i'm struggling to write a device tree for this. am i on the right track with mux-controller and gpio-mux? not sure how to combine that with the i2c mux too

17:50 <hanetzer> is there a mechanism for getting a device from a device_node? I'm pretty sure I've asked before but I forget

18:07 <headless> hanetzer: of_find_device_by_node()?

18:13 <hanetzer> that returns a platform_device struct. there was a patch back in 2015 that put a `struct device' under `struct device_node' but I guess that never got merged

18:16 <headless> hanetzer: &pdev->dev points to your struct device :-)

18:16 <hanetzer> ah, so it does.

18:19 <hanetzer> seems to work. no complaints from the LSP but hey, we'll see at boot time :)

18:19 <headless> hanetzer: LSO?

18:19 <headless> *LSP

18:20 <hanetzer> language server protocol. basically a mechanism for shoving language awareness into various editors.

18:22 <hanetzer> so instead of having to do a 3x4 square of vim/emacs/foo x c/lua/rust/bar you have two separate 1x3 and 1x4 rects of vim-lsp/emacs-lsp/foo-lsp and c-lsp/lua-lsp/rust-lsp/bar-lsp.

18:23 <hanetzer> because making dedicated integration of x languages against y editors means x*y, but doing it this way ends up being x+y

18:26 <hanetzer> well, it compiles without warnings. a good sign.

19:23 <hanetzer> izzn't there some tool to turn func_name+0xXX/0xYY into a line number?

19:27 <rfs613> hanetzer: ksymoops ?

19:28 <hanetzer> ./scripts/decode_stacktrace.sh vmlinux < dmesg :D