04:26 <travmurav[m]> craftyguy: does earlycon=efifib do anything?

05:55 <craftyguy> travmurav: ya with earlycon set the kernel boots and I get output all the way until "legacy bootconsole [efifb0] disabled" then it's stuck/no output

05:56 <travmurav[m]> craftyguy: hm can you try adding keep_bootcon?

06:01 <craftyguy> ya got some more info, mind if I send to you on matrix? (photos of screen are easier to share there 😅)

06:01 <travmurav[m]> yeah go ahead

06:05 <travmurav[m]> (on the photos I saw a mostly normal boot log, then cpu0 gets stuck and linux detects rcu stall trying to wake it up forever)

07:56 <JosDehaes[m]> <alexVinarskis[m]> "Did not test yet, don't have..." <- USB-C DP works on all 3 ports on yoga (with Jens Glathe 's kernel which has the LTTPR patches). Can hotplug many times.

08:47 <JensGlathe[m]> BT/powerseq patch for the Yoga Slim 7X is on the [list](https://lore.kernel.org/all/20250426-slim7x-bt-v1-1-d68f961af886@oldschoolsolutions.biz/), would appreciate a tested-by konradybcio thanks for the hint re qcp

09:22 <konradybcio> actually we now have pcieport pwrseq driver so IIUC we should resolve both mounting cases

09:48 <anthony25> JensGlathe[m]: thanks a lot! I'm going to check, I think I was already using this

11:01 <albsen[m]> I've been experimenting with the different settings to mask ram on the t14s 64gb oled. I believe the only way to mask correctly is using grub with the following command badram 0x8800000000,0xf800000000 (take not of the comma, the 2nd value is the mask). in addition one can / should also add mem=31G to the kernel params but that kernel param by itself doesn't mask the same memory as far as I can tell by running sudo cat

11:01 <albsen[m]> /proc/iomem.

11:01 <albsen[m]> * I've been experimenting with the different settings to mask ram on the t14s 64gb oled. I believe the only way to mask correctly is using grub with the following command badram 0x8800000000,0xf800000000 (take note of the comma, the 2nd value is the mask). in addition one can / should also add mem=31G to the kernel params but that kernel param by itself doesn't mask the same memory as far as I can tell by running sudo cat

11:01 <albsen[m]> /proc/iomem.

11:05 <albsen[m]> Jens Glathe: I been using your kernel build `6.15.0-rc2-jg-0-qcom-x1e` + the ram masking and haven't seen any crashes in 3 days of using it as my main machine. previously I got a crash within a few hours of using it. the newer kernel builds didn't work and crashed immediately again.

11:07 <albsen[m]> one more observation is, the password dialog during boot isn't visible to me anymore since 6.12 (or so), the screen shows grub, then after selecting the kernel shows the first few lines and then goes dark. I then enter the password with a black screen and if the password is correct it will continue to boot into multi user. this is all with the oled dtb (or with the non-oled dtb)

11:33 <JensGlathe[m]> @albsen is trying slbounce to boot to full ram usage feasible? There is the adsp „lite“ firmware loaded which can be used with a patch from sgerhold

11:45 <albsen[m]> Jens Glathe: I don't know how to do that. is there any documentation I can look at? I'm assuming this is so that I can gain access to the other 32gb ram, right? its actually not that much of a pressing issue to me, as up until 6.15.0 the machine kept crashing pretty much immediately.

12:01 <JensGlathe[m]> This is actually for the 64GB issue mainly. There apears to be a bug in Gunyah which can't handle the upper 32GB for \<reasons\>. Booting the kernel into EL2 replaces Gunyah, but it usually kills adsp and cdsp since they rely on the HYP firmware to load their firmware. At boot, adsp appears to be running some "lite" firmware already with partial functionality. This can be re-used on EL2 with the patch von sgerhold. The other

12:01 <JensGlathe[m]> (positive) effects are: full use of the 64GB RAM since Linux controls the smmuv3, and /dev/kvm. You will likely loose sound and maybe dp altmode. Take a look at [slbounce](https://github.com/TravMurav/slbounce) for further reading.

13:17 <tobhe_> albsen[m]: yeah, badram/cutmem works, the additional mem= is not needed since the grub option propagates to Linux

13:45 <albsen[m]> tobhe_: so an unknown to me reason the system only stabilized once I set both. could also be coincidence as the badram setting I used before was wrong.

17:01 <JosDehaes[m]> Jens Glathe: bluetooth was already working since a long time on yoga 7x?

17:03 <JensGlathe[m]> ... upstream?

17:04 <JensGlathe[m]> same with lots of x1 Purwa stuff, Vivobook 15, Acer Swift Go 14, ThinkBook 16

17:32 <tstachecki> Bluetooth can be fixed on a lot of the x1e80100 by cherry picking a change from the CRDs dtb

17:32 <tstachecki> I just got it working on the SL7 this morning

17:33 <alexVinarskis[m]> careful, most laptops so far have wcn7850, but some come with wcn6855 instead.

17:37 <tstachecki> kudos to everyone involved btw, at this point with John Hovolds patches, I have a 6

17:38 <tstachecki> 6.12* kernel with working suspend to ram even now. the only thing holding me back from a daily driver is figuring out the spi over hid stuff for the touchpad

17:39 <JensGlathe[m]> Microsoft is sort of an issue, doing things differently

17:39 <tstachecki> it's a shame but I'll admit the haptic touchpad is really nice when it works

17:40 <JensGlathe[m]> Has somebody experience with a DC dimming backlight panel? Breaking my fingers on it

17:43 <JensGlathe[m]> normal panel-simple and panel-edp say aux dp backlight dimming is not supported, and pwm configturation shows a control but dowsn't reaxt (naturally).

17:52 <konradybcio> tstachecki: i'm slowly poking at the touchpad fwiw, no commitments on the date though..

18:02 <spawacz> How to use hwdec on x13s with mpv? I'm using Johan's kernel, ffmpeg supports it it seems (ffplay -v debug -vcodec h264_v4l2m2m /tmp/foo.mp4 works flawlessly) but whatever I try to pass to mpv, it falls back to software decoding

18:07 <tstachecki> konradybcio I got the haptics firing, but I need qspi for geni I think :( I don't know how to address "invalid proto 9" from the geni module when it loads

18:08 <tstachecki> it does vibrate when I touch the pad now after playing with the reset gpio

18:17 <exeat> spawacz: mpv --hwdec=v4l2m2m-copy

18:25 <spawacz> exeat: that's it, thanks!

23:19 <Ariadne> is there any documentation on this slbounce thing?

23:41 <tobhe> Ariadne: https://github.com/TravMurav/slbounce/blob/main/theory_of_operation.md

23:41 <Ariadne> yep read that already.

23:42 <Ariadne> i will try it later :)