13:16 <danylo> On sm8650 with 6.12 or msm-next kernel I get the following context fault every ~20 minutes, which brings down wifi (though wifi restarts after it, but on msm-next the restart fails after a couple of times):

13:16 <danylo> [ +0.106531] mhi mhi0: Requested to power ON

13:16 <danylo> [ +0.080833] arm-smmu 15000000.iommu: FSR = 00000402 [Format=2 TF], SID=0x1401

13:16 <danylo> [ +0.000006] arm-smmu 15000000.iommu: FSYNR0 = 004e0013 [S1CBNDX=78 WNR PLVL=3]

13:16 <danylo> [Dec23 12:39] arm-smmu 15000000.iommu: Unhandled context fault: fsr=0x402, iova=0xffc80000, fsynr=0x4e0013, cbfrsynra=0x1401, cb=25

13:16 <danylo> [ +0.232469] mhi mhi0: Power on setup success

13:16 <danylo> [ +0.107930] mhi mhi0: Wait for device to enter SBL or Mission mode

13:16 <danylo> [ +0.497968] ath12k_pci 0000:01:00.0: chip_id 0x2 chip_family 0x4 board_id 0x3d soc_id 0x40170200

13:16 <danylo> Then many ath12k_pci errors, then "ieee80211 phy0: Hardware restart was requested"

13:34 <konradybcio> danylo any chance you could bisect this?

13:35 <danylo> konradybcio I don't have a good idea where it started

13:35 <danylo> And the issues may happen after 20 min of runtime

13:37 <konradybcio> well the error basically tells us 'something bad happened on pcie0' :/

13:43 <danylo> konradybcio doesn't sound nice....

13:44 <aklimov> any chances it can be caused by pcie power management?

13:53 <konradybcio> i wouldn't bet on that, but i wouldn't rule it out either

13:55 <danylo> konradybcio Anything I could try to check if it's a power management issues?

13:56 <aklimov> danylo: can you please try to use pcie_port_pm=off pcie_aspm=off in cmdline and check if anything changes?

13:56 <konradybcio> try pcie_aspm.policy=performance in cmdline

13:56 <konradybcio> or that ^

13:56 <danylo> ok, will try

14:40 <danylo> konradybcio aklimov With "pcie_port_pm=off pcie_aspm=off" now I got "[ +0.011437] arm-smmu 15000000.iommu: FSYNR0 = 007e0013 [S1CBNDX=126 WNR PLVL=3]", but otherwise it looks the same.

14:59 <konradybcio> right, that still only tells us "something went wrong" :(

15:37 <danylo> And sometimes it fails to recover and says that it "failed to send WMI_START_SCAN_CMDID" in a loop, fun