00:37 <leland[m]> T14s snapdragon on a good sale right now it looks like

00:43 <HdkR> "sale" :)

00:49 <leland[m]> Haha yeah.....

01:10 <erebion[m]> I've got Secure Boot working with Unified Kernel Images :D

01:11 <erebion[m]> So yeah, this indeed works without bricking anything.

01:46 <robclark> dummy regulator msgs are ok.. that dates back to days when genpd functionality was exposed to drivers as regulators.. probably some code we could remove these days.. the "fb0: Framebuffer is not in virtual address space" stuff, not sure what that is about offhand or if it matters

02:19 <erebion[m]> Battery life does not seem the way it should be, just around six or seven hours, something must be missing for this X13s... 🤔

02:20 <erebion[m]> Also, suspend seems to only blank the screen and not actually suspend

02:34 <steev> we don't hit the deepest sleep states, it's known

02:34 <steev> we should be hitting s2idle in suspend though, not just blanking

05:39 <JensGlathe[m]> Framebuffer not in VA space is a warning, introduced that the graphics driver isn't using some memory mapping correctly afair

05:39 <JensGlathe[m]> it works regardless (for now)

08:12 <kuruczgy[m]> ahoneybun: You probably have to build the install ISO for yourself, I haven't seen any upstream support for booting the ISO with a device tree.

08:12 <kuruczgy[m]> But I believe if you configure it with the right initrd kernel modules + right device tree you should be able to get it booting on the X13s

08:12 <kuruczgy[m]> I have a repo with the necessary nixpkgs patches to do exactly that, but unfortunately for the x1e80100 and not sc8280xp: https://github.com/kuruczgy/x1e-nixos-config/

09:11 <SpieringsAE> for some reason it seems impossible to find the lenovo t14s with 64gb ram here in europe, the sale makes it actually reasonable, but only in the US :(

09:12 <SpieringsAE> is there already fwupd support for arm devices? I guess not

09:13 <SpieringsAE> it does run on my pinebook pro, but there just isnt any supported firmware

11:47 <ahoneybun[m]> <kuruczgy[m]> "ahoneybun: You probably have..." <- > <@kuruczgy:matrix.org> ahoneybun: You probably have to build the install ISO for yourself, I haven't seen any upstream support for booting the ISO with a device tree... (full message at <https://matrix.org/oftc/media/v1/media/download/AZHZn1xRIphru0JwlU2vtAE0B04Uqn-A4lQmVs8aPT0BKngaFBfXqLCVTSQQqrHJtXeLuq2WAAvLBnyOh6ppbX5CeSXjkytwAG1hdHJpeC5vcmcvWnl4anZqa2lOZ0p1aWd1VlZteFVBeEJ3>)

11:48 <ahoneybun[m]> The device tree file is in upstream now it's just telling it that it's there or something

12:52 <erebion[m]> Tried using the Pinephone dock with the X13s, lsusb shows this as a "billboard", but I don't get a video output. Is there anything missing kernel side for video output or am I just missing some sort of firmware or udev rule for that?

12:53 <erebion[m]> <SpieringsAE> "for some reason it seems..." <- That was also the case for the X13s around a year ago. Maybe Lenovo first want to sell the less interesting option or whatever. I did not buy an X13s for a long time as 16 GB of RAM just is not enough for me. And 32 GB was not available. Maybe it will be soon-ish.

13:13 <erebion[m]> Where does the X13s store its Bluetooth MAC?

13:16 <Jasper[m]> <erebion[m]> "Tried using the Pinephone dock..." <- USB C alternate mode stuff is still a bit messy iirc, I've had issues upstream with some ports being stuck to single functions

13:17 <Jasper[m]> You can try flipping the cable over, trying the other ports

13:17 <Jasper[m]> erebion[m]: I think the latest theory is that tz is involved? I don't remember completely

13:18 <Jasper[m]> I've also heard it's supposed to be in DPP

13:21 <\[m]> <SpieringsAE> "for some reason it seems..." <- I've actually called their sales twice and got a direct answer - but if it's not available that will probably be their answer. they might have a view on timelines

13:30 <erebion[m]> <Jasper[m]> "You can try flipping the cable..." <- Already tried that

13:30 <erebion[m]> <Jasper[m]> "I think the latest theory is..." <- timezone? o.O

13:30 <Jasper[m]> Trustzone

13:31 <erebion[m]> Oh, that makes more sense

13:48 <\[m]> I know there's like repackaging companies for getting stuff from the usa though? might add significant costs though

13:48 <Jasper[m]> Superbuy will work iirc

14:38 <\[m]> I am planning to get an x1e laptop like around release date of 6.13 kernel - that's a good timing as I understood?

14:41 <Jasper[m]> I'm just waiting for them to drop in price or show up on second hand sites a lot more

15:35 <macc24> there will probably be a lot of them in outlets

15:44 <erebion[m]> <\[m]> "I know there's like repackaging..." <- I've had good experience with shipito.com

15:44 <erebion[m]> Also, I know some nerds get their devices via people bringing them to Chaos Communication Congress, as we have people from everywhere coming there :D

16:34 <JosDehaes[m]> Jasper: I'm willing to sell my yoga 7x

16:35 <Jasper[m]> JosDehaes[m]: NU AL

16:35 <Jasper[m]> That's quick

16:35 <JosDehaes[m]> a bit disappointed in the device

16:35 <Jasper[m]> Can imagine

16:36 <JosDehaes[m]> it's not even faster than my M1 MBA

16:37 <JosDehaes[m]> so, no, the X elite is not a macbook killer

16:37 <macc24> JosDehaes[m]: wtf

16:37 <macc24> did you benchmark them both?

16:38 <JosDehaes[m]> yes, in a benchmark that matters to me, compiling big rust projects

16:38 <JosDehaes[m]> as that's my job

16:39 <JosDehaes[m]> even though it has 12 P cores and the M1 only has 4+4

16:39 <Jasper[m]> JosDehaes[m]: On Linux?

16:39 <Jasper[m]> Can imagine there's some performance scaling issues for now

16:39 <JosDehaes[m]> linux on the yoga, macos on the M1

16:39 <macc24> do you have cpufreq stuff enabled?

16:40 <JosDehaes[m]> I use johan_defconfig

16:40 <macc24> are you using a patched kernel or not?

16:40 <JosDehaes[m]> should it be faster?

16:40 <Jasper[m]> JosDehaes[m]: Yeah okay, it should be faster in every way possible hahahaha

16:40 <macc24> you need patches for cpufreq on x elite

16:40 <macc24> otherwise it's running on boot clocks(slow)

16:40 <JosDehaes[m]> oh, haven't seen those then

16:40 <Jasper[m]> So I'd indeed recheck your kernel builds

16:41 <JosDehaes[m]> can you point me to the patches?

16:42 <JosDehaes[m]> I'll benchmark it again, and may need to revisit my offer Jasper 😁

16:42 <Jasper[m]> JosDehaes[m]: Hahahah no worries

16:42 <JosDehaes[m]> anyway, planning on getting an M4 max, which will for sure blow it out of the water (at 5 times the cost of course)

16:43 <Jasper[m]> Yeah M4's quicker, haven't seen by how much though

16:43 <Jasper[m]> iirc ipc on the X elite cores is a lot better going off the die pics someone made

16:43 <JosDehaes[m]> it's not even close to anything else on the market

16:44 <JosDehaes[m]> and benchmarked on a thin ipad

16:44 <JosDehaes[m]> so imagine how fast that will be in a macbook pro with active cooling

16:51 <Jasper[m]> <JosDehaes[m]> "it's not even close to anything..." <- I'm mostly saying that it's quite a bit bigger than the X Elite on a smaller TSMC node

16:52 <Jasper[m]> Sure it's quicker, but I'm sure if qualcomm was planning a bigger die it'd be quicker too

16:52 <leland[m]> I think my next device will be from ampere

16:53 <Jasper[m]> I kinda hope Nvidia and qualcomm's (possible) desktop stuff will trickle down to consumers

16:54 <leland[m]> Can't believe Qualcomm didn't have great Linux support on release for the new laptops

16:57 <Jasper[m]> I can

16:57 <JosDehaes[m]> macc24: is this the relevant patch? https://lore.kernel.org/all/20240612124056.39230-6-quic_sibis@quicinc.com/

16:57 <Jasper[m]> But they could have made it a lot easier, or at least planned it better

17:00 <JosDehaes[m]> no that seems already in

17:00 <Jasper[m]> @leland some parts of that issue are not directly Qualcomm's fault

17:00 <Jasper[m]> JosDehaes[m]: Just to be sure, you're running the latest kernel Johan built?

17:01 <Jasper[m]> And you have a corresponding recent dt for the yoga?

17:03 <Jasper[m]> Jasper[m]: Well, you may have built it, but I'm specifically talking about his tree

17:04 <macc24> <JosDehaes[m]> "macc24: is this the relevant..." <- yes, however, you need the entire series

17:05 <macc24> <leland[m]> "Can't believe Qualcomm didn't..." <- that's the standard for arm chips

17:05 <Jasper[m]> macc24: Even then they overdelivered

17:05 <Jasper[m]> Which is still not much

17:08 <HdkR> macc24: Something to note is that 3.4Ghz is quite a bit slower than the 4.2Ghz of the higher end SKUs still. So it'll still be a lot closer to M1/M2 than expected, just more cores on average :)

17:08 <HdkR> Would have been nice if Lenovo shipped something higher than x1e-80-100, but eh.

17:35 <robclark> I mean, tbf qcom's support for linux has been a _lot_ better than apple's ;-)

17:35 <robclark> the missing bits are really things only the OEMs could have supplied, for the most part (although I think we are still waiting for some speedbin bits, like for gpu)

17:38 <robclark> as far as speed, it (my x1e78100 7x) is faster at compiling stuff than the new zen5 laptop parts (comparing phoronix benchmarks to local runs of same benchmark) which is apparently still faster than LL.. the apple SoCs aren't really relevant to me (not interested in running macos), but I would be interested to see some M1/2/3/4 benchmarks with 4k page size (I guess 5-10% of the perf advantage is down to apple controlling

17:38 <robclark> the OS and being able to do things like use larger page size)

17:41 <macc24> <HdkR> "macc24: Something to note is..." <- yeah... i see quite often the cpu frequency pegged at 3.4ghz and i sometimes wonder how it compares to faster skus

17:47 <pstef> robclark: would you be willing to run geekbench 6 preview and share the result?

17:47 <robclark> I think there already should be GB6 results from linux uploaded

17:48 <robclark> if not, I can run it later, but I think GB results aren't hard to find

17:54 <pstef> I was only able to find https://www.cnx-software.com/2023/10/31/qualcomm-snapdragon-x-elite-soc-benchmarks-in-windows-and-linux/

17:55 <JosDehaes[m]> <macc24> "yes, however, you need the..." <- it seems everything except drivers/mailbox/Kconfig is already in johan's 6.12-rc1 tree

17:57 <robclark> I did run GB6 when I first got linux running on it, and pretty sure the results were uploaded.. IIRC it was similar to the results from windows

18:04 <JosDehaes[m]> ok, the entire cpufreq series is in johan's tree

18:21 <HdkR> Yea, I also ran geekbench and bytemark and the numbers were roughly comparable between Linux and Windows on x1e-80-100. Higher end SKUs would take a hit since CPU turbo isn't wired up?

18:22 <HdkR> bytemark scales linerally with clockspeed on the platform, so it's nice to see where 4.2/4.3Ghz /would/ land it

18:24 <HdkR> linearly

18:40 <spawacz> I can't play the sound via headphones plugged to x13s. The internal speakers work but there's no sound when I plug the headphones. I tinkered in pavucontrol to no avail

18:46 <spawacz> reboot helped. But the balance is weird. The playback volume jumps up and down strangely

18:48 <spawacz> no, it sounds like there are short breaks without any sound

18:54 <spawacz> Oct 05 20:52:54 x13s pulseaudio[939]: ALSA woke us up to write new data to the device, but there was actually nothing to write. Most likely this is a bug in the ALSA driver 'snd_soc_sc8280xp'. Please report this issue to the ALSA developers.

19:18 <Jasper[m]> <robclark> "the missing bits are really..." <- Or venus :^)

19:18 <Jasper[m]> <JosDehaes[m]> "ok, the entire cpufreq series is..." <- Okay, dumb question, but did you replace the DT you're booting with a new one aswell?

19:18 <steev> venus is just waiting on the firmware to be submitted (we have no idea what the hold up is)

19:18 <Jasper[m]> steev: Yeah, that's what I mean

19:19 <JosDehaes[m]> Jasper[m]: yes, of course, I have a script that does the building, initrd and devicetree stuff automatically so I can't forget.

19:19 <JosDehaes[m]> I did retest it just now, and the yoga IS a bit faster than the M1

19:19 <Jasper[m]> Alright

19:20 <Jasper[m]> Is there a firmware based performance mode setting maybe?

19:20 <JosDehaes[m]> 1m15 for the yoga and 1m42 for the M1

19:20 <Jasper[m]> I remember some laptops having it, but idk if it's windows driver driven or something in the firmware

19:20 <JosDehaes[m]> in windows our rust code doesn't even compile (rocksdb issue)

19:21 <JosDehaes[m]> and in WSL2 it's MUCH slower than in linux

19:21 <JosDehaes[m]> so can't use that benchmark 😅

19:21 <JosDehaes[m]> to compare with windows I mean

21:48 <erebion[m]> X13s: ath11k_pci 0006:01:00.0: firmware: failed to load ath11k/WCN6855/hw2.1/firmware-2.bin (-2)

21:48 <erebion[m]> Is that obtainable somewhere?

21:49 <JensGlathe[m]> no, standard error message. it loads board-2.bin and amss.bin, m3.bin.

21:56 <erebion[m]> Huh, I thought that would explain my network connection being somewhat slow

21:58 <erebion[m]> Compiling the kernel on the X13s - for porting Mobian to the Pixel 3a - is definitely a lot more fun than on the X270

22:08 <erebion[m]> Maybe 3 or 4 minutes instead of closer to ~1.5 hours

22:53 <erebion[m]> Huh, today the Pinephone dock just works on the X13s

22:53 <erebion[m]> Seems I did something that makes it work \o/

22:58 <erebion[m]> And I get 2.5k video \o/