#asahi-dev · 2024-04-28 10:24
jannauhttps://github.com/AsahiLinux/linux/commit/72909db13b2e258b69eae8e33b5af02376829f82 fixes or at least almost fixes the kwin stuttering#asahi-dev · 2024-04-28 10:21
chadmed_thanks :)#asahi-dev · 2024-04-28 10:21
jannauchadmed: for macsmc-hwmon + dts: Signed-off-by: Janne Grunau <j@jannau.net>#asahi-dev · 2024-04-28 09:53
chadmed_huh... rust just caused an oops...#asahi-dev · 2024-04-28 09:47
mpsah, thanks#asahi-dev · 2024-04-28 09:46
chadmednot the other thing ;) #asahi-dev · 2024-04-28 09:46
chadmedsigned-off-by#asahi-dev · 2024-04-28 09:45
mpswhat means acronym 's-o-b'#asahi-dev · 2024-04-28 09:44
chadmedjannau: could i have your s-o-b for the hwmon driver?#asahi-dev · 2024-04-28 00:41
tpw_rulesjannau: it seems u-boot only fails to boot the kernel when going through bootflow. bootmenu works fine so i've advised users to use that. might also be good to update the installer packages too, at least the uefi-only one. not sure if you have that power#asahi-dev · 2024-04-27 16:03
jannaukwin 6 reduced render rate fixed by deleting code: "1 file changed, 1 insertion(+), 21 deletions(-)"#asahi-dev · 2024-04-27 12:20
chadmedand all we have to do in probe is collect whichever ones were created into the array of pointers#asahi-dev · 2024-04-27 12:20
chadmedyeah that's what i was thinking of trying tomorrow, then we can just roll creating it into populate_sensors#asahi-dev · 2024-04-27 12:20
jannauthat would get rid of all the pointer arithmetic#asahi-dev · 2024-04-27 12:19
jannauthen you can fill out each of the *_info structs and afterwards write pointers to the present sensor types into channel_infos[]#asahi-dev · 2024-04-27 12:17
jannaustruct hwmon_channel_info chip_info; struct hwmon_channel_info temp_info; ...#asahi-dev · 2024-04-27 12:16
jannaui.e. add `struct hwmon_channel_info *channel_infos[NUM_SENSOR_TYPES + 1 /*chip*/ + 1 /* terminating NULL */];` and#asahi-dev · 2024-04-27 12:14
jannauthe hwmon_channel_info would be simpler if we move the channel info array and the channel info into `struct macsmc_hwmon` and only allocate the config arrays dynamically#asahi-dev · 2024-04-27 12:09
chadmedj475#asahi-dev · 2024-04-27 12:05
jannauwhich device(s) miss TW0P?#asahi-dev · 2024-04-27 11:48
chadmedthere _has_ to be a better way to do that hwmon_channel_info stuff... surely...#asahi-dev · 2024-04-27 11:47
chadmedalso picked jannau's commits refactoring the devicetree stuff but that's broken and needs work (TW0P doesn't exist on every machine for example)#asahi-dev · 2024-04-27 11:47
chadmedtook some time to refactor various parts of the hwmon driver and rebase it atop 6.8-3: https://github.com/chadmed/linux/tree/hwmon-rebase#asahi-dev · 2024-04-27 11:40
jannauM2 Ultra needs https://github.com/AsahiLinux/m1n1/pull/389 for setup on the second die (untested but likely to work)#asahi-dev · 2024-04-27 11:39
jannauM1/M2 Ultra DP/HDMI audio support pushed to https://github.com/jannau/linux/tree/asahi-wip-6.8#asahi-dev · 2024-04-27 10:02
chadmeddoesnt look like anything related to asahi-audio. those speaker routing props logs are from the volume limiter and are expected to be there#asahi-dev · 2024-04-27 09:55
jannauchadmed: someone reporting a wireplumber crash on f40: https://discussion.fedoraproject.org/t/wireplumber-crash-in-fedora-40-no-sound-through-speaker/114578#asahi-dev · 2024-04-26 07:02
chadmed+1 #asahi-dev · 2024-04-26 07:01
jannauI'm not proposing to merge that version. there is a legitimate development use for fan speed control. It makes CPU core temp key mapping easier#asahi-dev · 2024-04-26 06:46
chadmedwe cannot guarantee that messing with the fan settings is safe yet so we should not allow it to be done in production systems at all until we can#asahi-dev · 2024-04-26 06:46
chadmedlook at how many people ignored the repeated warnings about prematurely enabling their speakers#asahi-dev · 2024-04-26 06:46
chadmedi still dont think we should allow the setting of fan speeds even behind a parameter btw#asahi-dev · 2024-04-26 06:43
chadmedyeah of course it is i rewrote parts of the driver and this version has no fan support because im not up to that yet#asahi-dev · 2024-04-26 06:43
leioline number is like 20 off as well#asahi-dev · 2024-04-26 06:42
chadmedoh duh put it in a block#asahi-dev · 2024-04-26 06:42
leiocase _SMC_KEY("flt "): {#asahi-dev · 2024-04-26 06:41
leiowhere there is a block after case label#asahi-dev · 2024-04-26 06:41
leiothat diff doesn't match the code I'm looking at from jannau's branch#asahi-dev · 2024-04-26 06:41
chadmedclang 18 btw#asahi-dev · 2024-04-26 06:40
chadmedhttps://github.com/chadmed/linux/commit/306ee43fdb600acf45f73e257096afaa07335f9b#asahi-dev · 2024-04-26 06:40
chadmeddrivers/hwmon/macsmc-hwmon.c:103:3: warning: label followed by a declaration is a C23 extension [-Wc23-extensions]#asahi-dev · 2024-04-26 06:27
leioAdding -Wc2x-extensions only adds a warning from include/linux/path.h code, nothing for macsmc-hwmon.c#asahi-dev · 2024-04-26 06:24
leio-Wall is there though#asahi-dev · 2024-04-26 06:24
leiono -Wc2x-extensions passed though#asahi-dev · 2024-04-26 06:23
leiono warnings here from touch drivers/hwmon/macsmc-hwmon.c && make LLVM=1 LLVM_IAS=1 -j25 #asahi-dev · 2024-04-26 06:19
chadmedyup, have for years even on amd64#asahi-dev · 2024-04-26 06:19
leioyou run your kernel from clang builds?#asahi-dev · 2024-04-26 06:19
leioooh, LTOing options#asahi-dev · 2024-04-26 06:19
leiomeh, why not CC=clang, too obvious? :D#asahi-dev · 2024-04-26 06:18
chadmednote that youll probably get asked about some config items