<enyc>
hrrm generally winning wit 23.05-rc2 alas I guess no xrx200 / HHv5a image being produced yet I can add to my testing especially wrt DSA.
tSYS has quit [Quit: *squeak*]
tSYS has joined #openwrt-devel
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
<Christophe[m]1>
What's the general approach to change kernel defaults? I'm thinking about BPF event support and possibly some security related things.
<slh>
if that's not exposed as config symbols, make kernel_menuconfig
madwoota has quit [Remote host closed the connection]
madwoota has joined #openwrt-devel
aiyion has quit [Remote host closed the connection]
aiyion has joined #openwrt-devel
Borromini has quit [Ping timeout: 480 seconds]
schwicht has joined #openwrt-devel
schwicht has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
schwicht has joined #openwrt-devel
schwicht has quit []
schwicht has joined #openwrt-devel
schwicht has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
schwicht has joined #openwrt-devel
schwicht has quit []
<rsalvaterra>
Dilemma: if we can't have both, which should we choose, lto or mips16? Intuitively, I'd say the latter, as it generates tighter code (and mips32r2 devices are almost certainly space-constrained)…
schwicht has joined #openwrt-devel
schwicht has quit []
schwicht has joined #openwrt-devel
robimarko has joined #openwrt-devel
schwicht has quit []
schwicht has joined #openwrt-devel
schwicht has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
robimarko has quit [Remote host closed the connection]
robimarko has joined #openwrt-devel
<hauke>
rsalvaterra: isn't it a compiler bug if this does not work together?
<hauke>
I think mips16 does not work with big functions because it can not do very long relative jumps or they are buggy in GCC
<hauke>
probably lto generates such functions
<hauke>
maybe create a ticket for GCC for an application which compiled with GCC 12 and does not compile with GCC 13 when using mips16 + LTO
<rsalvaterra>
hauke: I don't know how GCC 12 behaves, though. I've just been playing lately with globally enabling lto and gc-sections, seeing what breaks in my builds and massage the respective makefiles accordingly.
<rsalvaterra>
However, I do agree that if lto doesn't play well with mips16, it's a compiler bug.
<hauke>
rsalvaterra: ok this is happening when you activate lto for all applications
<rsalvaterra>
Yes. Some fallout is obviously expected at this time.
<hauke>
I think we should just deactivate LTO when mips16 is used for some appliations, maybe add an additional flag
<hauke>
and create a ticket for GCC
<hauke>
but I think there are not so many people caring about mips at GCC any more
<rsalvaterra>
Indeed. Also unfortunate, but not unexpected. It's old.
<f00b4r0>
mips16 maybe. Mips in general I'm not so sure. GCC is e.g. the primary compiler for Loongson
<Mangix>
173 PRs open on packages. how did this get out of hand?
<hauke>
Mangix: I am currently trying to keep the open pull request for main at ~350 ;-)
<robimarko>
hauke: mt7621 slipped through with CONFIG_FRAME_WARN being set in kernel config
<robimarko>
As it was added yesterday
<Mangix>
I had the packages ones at ~80 at one point.
<hauke>
robimarko: I fixed it
<Mangix>
rsalvaterra: ./github.com/nextdns/nextdns/run.go:175:(.text+0x4352cc): relocation R_MIPS_26 against `a local symbol' cannot be used when making a shared object; recompile with -fPIC <-- I guess this is what you meant.
<robimarko>
hauke: great
<robimarko>
I am trying to review stuff I have understanding off, so PR-s dont go stale
<hauke>
I normally order the PRs by last updated and then look at the last 2 weeks
<hauke>
a Tested-by tag or something similar in a pull request also helps
FLD has quit [Ping timeout: 480 seconds]
FLD has joined #openwrt-devel
srslypascal has joined #openwrt-devel
srslypascal has quit []
floof58 has quit [Ping timeout: 480 seconds]
minimal has joined #openwrt-devel
floof58 has joined #openwrt-devel
floof58 has quit [Ping timeout: 480 seconds]
floof58 has joined #openwrt-devel
schwicht has joined #openwrt-devel
schwicht has quit []
schwicht has joined #openwrt-devel
schwicht has quit []
schwicht has joined #openwrt-devel
schwicht has quit []
schwicht has joined #openwrt-devel
<Christophe[m]1>
I'm happy to help with testing or anything else that needs to be done?
schwicht has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
schwicht has joined #openwrt-devel
robimarko has quit [Quit: Leaving]
schwicht has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
schwicht has joined #openwrt-devel
schwicht has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
schwicht has joined #openwrt-devel
goliath has joined #openwrt-devel
schwicht has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Christophe[m]1>
"ath10k_pci ... mac flush" messages render the kernel log unusable for me. Is it just me? https://ibb.co/1vjKQ10
ptudor has joined #openwrt-devel
ptudor has quit [Ping timeout: 480 seconds]
goliath has quit [Quit: SIGSEGV]
dgcampea has quit [Remote host closed the connection]
dgcampea has joined #openwrt-devel
tidalf has joined #openwrt-devel
zkrx has quit [Ping timeout: 480 seconds]
danitool has joined #openwrt-devel
zkrx has joined #openwrt-devel
goliath has joined #openwrt-devel
fakuivan has quit [Remote host closed the connection]
fakuivan has joined #openwrt-devel
ptudor has joined #openwrt-devel
ptudor has quit [Ping timeout: 480 seconds]
dgcampea has quit [Remote host closed the connection]
<Habbie>
-O1 realises the value doesn't matter and turns it into a countdown
djfe is now known as Guest4743
djfe has joined #openwrt-devel
<Mangix>
rsalvaterra: probably supposed to be a sleep of some kind
<Habbie>
also my guess
<Habbie>
terrible if true :)
<rsalvaterra>
Yeah, that's what I thought, but no, you *never* busy-loop like that. :P
<Mangix>
some more info: the breed bootloader guy removed overclocking on NAND devices for stability reasons. apparently it's recommended to do it in the kernel.
<rsalvaterra>
So, if the patch is working, that line is useless.
<rsalvaterra>
Alas, my RM2100 has a NAND chip. :P
<Mangix>
which means this patch is the way to overclock :P
<rsalvaterra>
Oh, I misunderstood. Nice.
Guest4743 has quit [Ping timeout: 480 seconds]
<rsalvaterra>
Applied to my tree, minus the loop. :P