08:54 <Bioxvirizm-x13s[m]> If I understand correctly, mouse twitch is somehow related to ERROR:gbm_wrapper.cc(293)] Failed to export buffer to dma_buf: No such file or directory (2). This error is present in chronium logs and I found in the Internet that it is somehow connected with gfx::BufferFormats supported for allocation and texturing

10:49 <HdkR> Coolio, plugged in a 5gbit adapter to the Lenovo and was able to get 3.47gbit/s with iperf3

10:51 <HdkR> Just have to build in the aqc111 driver since it was disabled by default

10:52 <jhovold> HdkR: good to hear!

10:53 <HdkR> oop, but it is crash happy now

10:55 <jhovold> not so good, any hints about why/where?

10:55 <HdkR> Had a log but I need to restart it so I can copy it off

10:57 <HdkR> [ 240.620101] arm-smmu 15000000.iommu: Unhandled context fault: fsr=0x402, iova=0xfc7ff000, fsynr=0x200001, cbfrsynra=0x820, cb=3

10:57 <HdkR> [ 240.620126] xhci-hcd xhci-hcd.0.auto: WARNING: Host System Error

10:57 <HdkR> That's all it had

10:58 <HdkR> and then stuff was locking up

11:15 <HdkR> Back to gigabit until that gets solved I guess

13:04 <jhovold> HdkR: ok, thanks for the log

15:00 <exeat> Here's a keyboard oddity on the x13s: Win+LeftShift+{Left,Up,Down} all work as expected but Win+LeftShift+Right does nothing (tested with xev, showkey, hid-recorder). Win+RightShift+Right does work.

15:01 <HdkR> n-key rollover strikes again

15:08 <HdkR> jhovold: I guess another piece of information is that it is an Aquantia AQtion agc111 device. Which might be handy.

15:12 <jhovold> HdkR: indeed, and are you seeing this with my 6.3.1 wip branch or steev's?

15:13 <steev> i don't think i have anything that would affect usb here that you don't

15:13 <jhovold> I'm running iperf3 here with a Lenovo gigabit adapater (realtek) at 940 Mbits/s without any such hickups

15:13 <jhovold> steev: me neither, but you had quite a few patches on top last time I looked

15:14 <HdkR> This is with steev's 6.3.y branch

15:14 <jhovold> more likely related to different device and driver

15:15 <jhovold> HdkR: how reproducable is it? Do you hit it every time you run iperf? both directions?

15:15 <steev> i suppose the llcc or pcie stuff on top might

15:17 <HdkR> Good question, I haven't tried to reproduce

15:26 <HdkR> Hm, only reproduces when running Steam? wtf

15:30 <HdkR> Wonder if I need to fully load the CPU or something

15:34 <HdkR> `aqc111 1-1:1.0 enx002427884751: Link Speed 1000, USB 2` That's also interesting. Looks like the USB ports don't always come back at full speed?

15:34 <HdkR> Unplug and replug `aqc111 2-1:1.0 enx002427884751: Link Speed 5000, USB 3`

15:35 <HdkR> Might explain the abysmal performance of devices if that happens

15:45 <HdkR> hah, consistently reproduces by running Steam. That's a sucky repro

15:48 <HdkR> Oh!

15:48 <HdkR> There's a hung task!

15:54 <HdkR> https://gist.github.com/Sonicadvance1/ad5aa41fa014a7049b06e829eeb44183 That's a pretty spicy bug

15:56 <HdkR> Maybe need to throw some fio tests at it or something

16:58 <steev> jhovold: what is the holdup on letting qcom pci/e be modular?

23:13 <steev> and linus has tagged 6.4 rc1 :)