00:22 <robclark> mmediouni[m]: we should have GMU voting for gpu bandwidth now.. although I guess that is new in recent kernels

00:23 <robclark> I don't think el2 vs el1 should factor in.. but maybe new vs old kernel?

00:35 <HdkR> Would be nice if there was an ability /somewhere/ to just lock memory clocks to max, or specific frequency.

00:36 <HdkR> That way if I'm concerned about GMU voting shenanigans not happening, I can lock it to max, or even query it.

01:10 <valpackett> if anyone's wondering about cpuidle on next.. https://lore.kernel.org/all/20251121010756.6687-1-val@packett.cool/

01:11 <valpackett> (we need to fix the values on hamoa, but still, the check was done way too harshly imo)

02:03 <McCak[m]> Are there any update for GPU suppport on Snapdragon X? I just saw Snapdragon X2 news and nothing for Snapdragon X

02:04 <HdkR> McCak[m]: X1E Has been supported for quite a while.

02:04 <McCak[m]> yes but i meant non elite verson like Plus and Base

02:07 <HdkR> https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35404 Should be supported since June

02:08 <McCak[m]> Available on Ubuntu ARM64 iso or need my own compiling?

02:08 <HdkR> That one I'm not sure about

02:08 <McCak[m]> last time i tried on july - september, its still missing

03:18 <valpackett> definitely should be available in current ubuntu isos

04:01 <steev> https://www.phoronix.com/news/Dell-Pro-Max-With-Qualcomm-NPU

04:04 <HdkR> ooo, big AI numbers

04:05 <HdkR> Absolutely wacking huge amount of PCB space to stick two in to that form factor.

04:54 <mmediouni[m]> more hexagon v68

05:28 <steev> i appreciate that its *starting* weight for the 18 inch is 7.17 lbs

05:29 <steev> about time we get away from laptops that weigh nothing

05:38 <longptr> Funny, I love the light weight laptops, I wish the X14s wasn't as heavy and was lighter like X13s. The Thinkpad X1 Carbons are light-weight and have a bunch of usable ports. I think the X1 Carbon Aura is like 2.14 lbs?

05:54 <steev> sorry, that was heavily sarcastic - a friend just priced one of them out and she was able to get it to almost 10 1/2 lbs

06:11 <longptr> Heh

06:13 <steev> which reminds me, when i was at bsides lv earlier this year, i was talking to a group of people and they were asking me about the qualcomm laptops, and i was like, my only complaint is the battery life

06:22 <longptr> steev: Out of curoisity, what kind of battery life do you see? My measured power shows X13s ~5W playing back video, etc

06:22 <steev> longptr: ~8 hours

06:23 <longptr> Hrm, yeah that seems lower than expected. I have a Lunar Lake laptop I've been testing and doing stuff like ssh to remote machines, web use, I'd project upwards of 12 hrs

06:23 <steev> oh right, sorry i got distracted - so that's what i told them, anywhere from 8-10 hours or so of usage, and they were all like.... we get like 3 hours max with ours

06:23 <steev> well, for kali work, i was doing a lot of compiling of packages :)

06:23 <longptr> Oh, 8 hrs of actual compiling and building aka "real work", that's impressive :)

06:24 <steev> the c630 would get quite a bit more

06:24 <steev> admittedly it wasn't as powerful but still

06:58 <valpackett> steev: oh since you can change the topic, maybe the parenthesis should start with "X Elite" something something and not "NovaGo TP370QL" :D

06:58 <steev> tbh, i'd probably strip all the specific ones out

06:59 <valpackett> yeah

07:05 <valpackett> btw re: power, i think i might've just gotten "somewhere" with the little "enable PM thingy" changes now.. (unless i'm screwing up the measurements)

07:06 <valpackett> i think .use_rpm = true in LPASS clk got me -0.24W in screen-on idle (3.86 → 3.62W)

07:08 <valpackett> and `.pm = &snd_soc_pm_ops` in sound/soc/qcom/x1e80100.c kinda saves 0.36W in suspend (???) (seems like way too much tho)

07:18 <steev> you mean when the card is suspended? or the device itself

07:40 <valpackett> in deep sleep, as calculated by the script from https://gitlab.com/Linaro/arm64-laptops/linux/-/issues/19

08:04 <steev> ah

09:28 <steveej[m]> on that note, my x13s doesn't go to sleep anymore. drained it in my backback yesterday while thinking it was

09:29 <steveej[m]> also i'm getting blue screens lately sometimes when plugging in the external monitor. just dark blue color on both screens. i can still switch to a different VT and reboot the machine

09:31 <travmurav[m]> fwiw I have something similar - underflow blue on rand(internal, external) monitor after dpms, cleared by unplugging external monitor and doing dpms again

09:31 <travmurav[m]> (with a 4k display)

09:33 <steveej[m]> i think these logs are related:... (full message at <https://matrix.org/oftc/media/v1/media/download/ARs78YaW9W76jm8yaBi-XGRKhnMcf3W4xs8fcZhzcO9kFouIlNb1Lds6vmNvtZuCv7Kug3Sl-u79btkw_PuTsuhCea7du1_wAG1hdHJpeC5vcmcvSkdZU09lWUpjRlFsT1ZERHp4em15SVRt>)

09:34 <travmurav[m]> lol can't reproduce it anymore now, only ~100ms of underflow on the internal screen after waking both displays

09:41 <travmurav[m]> perhaps there was some fix in 6.17 or laptop kernel on which el2 tree is based, was avoiding dpms with external screen connected so idk when exactly it got fixed but I think the error spam was the same

09:43 <steveej[m]> i'm not running on el2 yet. don't have much bandwidth the next months to experiment with my machine

09:50 <travmurav[m]> actually I guess what I had back then was slightly different, this was on clean-ish 6.16 with 4 lane patches on top, so I've disregarded it as my mistake, but fwiw... (full message at <https://matrix.org/oftc/media/v1/media/download/AWdAVq3zOAOHwVBLecL0TbF87R3POxDBLqkqo9NkT7k7lH-kUIQZfa67J-HSL9pqAbUn4gEPl-cuhubY_NpNm05Cea7et7ygAG1hdHJpeC5vcmcvcEpGVFR1SUVNQWhHY1FMRGZUc2tRVXdI>)

09:51 <travmurav[m]> and some other times dpu was hitting iommu faults, drawing a beautiful rolling black bar over the screen

11:32 <neobrain[m]> Anyone else updated their X13s to Mesa 25.3.0 and seeing occasional crashes in Firefox now?

14:56 <serdeliuk[m]> anyone know if those laptops with qualcomm have a cmos battery? if something goes wrong with bios settings and the setup interface is not available to load bios defaults, how one can reset the settings, there should be a battery to replace, but looking at a teardown for zenbook A14 i am not able to spot one.....

14:57 <serdeliuk[m]> s/replace/remove/

15:01 <krzk> abelvesa: I think you were rebasing tree on rc6 and dropp the "arm64: dts: qcom: x1e78100-t14s: Add audio playback over DisplayPort" patch

15:02 <krzk> Any reason? It rebases cleanly

15:15 <gwolf> steev: (re: bettery life you were mentioning ~9hr ago): I haven't really measured yet the X13s' battery life in "real" use, but I _feel_ it can suspend without fully draining for somewhat longer than the C630.

15:35 <\[m]> Val Packett are you going over every possible config and disabling / enabling / changing values then measuring? that sounds like a lot of work? I can do some, if you tell me what to do ? in terms of going over many values etc

15:36 <\[m]> (is matrix mentioning broken on irc - let me go look)

15:36 <jdb[m]> The drain is pretty high on X13s, much higher than I was expecting, especially knowing that these SoC share a lot with the smartphone SoCs

15:37 <\[m]> (it's not)

15:37 <jdb[m]> Is hibernate / suspend-to-disk supposed to work or in the works for the near future?

16:51 <anthony25> serdeliuk[m]: I'm not sure if laptops have a CMOS battery, I can be wrong but I think that unplugging the laptop battery for some time can reset the BIOS settings

17:34 <mmediouni[m]> <serdeliuk[m]> "anyone know if those laptops..." <- Nah. Some desktop-y machines have a CMOS battery but that's to not trip SPU anti-replay mitigation too often

17:43 <steveej[m]> on the x13s i have this issue where on high system load my USB audio device gets lost and wireplumber/pulseaudio don't recover from that easily while i'm in a call. is this a known issue with high load and USB devices?

19:25 <konradybcio> steevej what kind of high system load can be used to reproduce?

20:42 <steveej[m]> konradybcio: compiling rust projects while on a jitsi call does it for me regularly.

20:44 <steveej[m]> maybe the specific set up is important. the USB audio receiver is 2 hubs deep.

20:44 <steveej[m]> laptop -> USB-C -> hub -> USB-A -> hub -> USB-A -> audio dongle

20:47 <longptr> USB hub might be the issue

20:54 <longptr> Can you get rid of one hub? USB hubs tend to have weird issues and strange corner cases.

21:20 <HdkR> I do love that USB has a max tree depth of 5 so you end up hitting some really wacky problems there when the final device tends to have its own internal hub :D