On 9.01.2025 2:27 PM, Jens Glathe wrote: > On 09.01.25 12:13, Bartosz Golaszewski wrote: >> Well, that's weird. Are you 100% sure this commit is the culprit? I >> have no idea how this could happen, it's just a Kconfig tweak. > > That's what I thought, too. I merged qcom/for-next into my branch for > testing and adventure, compared / merged my dt's that were in both, and > then this showed up on the HP X14. On the Snapdragon Dev Kit all 5 eUSB2 > repeaters work regardless with the same build. The chip in question is > the SMB2360, a PMIC which also contains the eUSB2 repeater. > > jglathe@sdbox2:~/src/linux_ms_dev_kit$ git bisect log > git bisect start > # status: waiting for both good and bad commits > # bad: [a9059ebc60d5b845a95b25650eb24c1df5ed2132] arm64: dts: qcom: > x1e80100: Fix usb_2 controller interrupts > git bisect bad a9059ebc60d5b845a95b25650eb24c1df5ed2132 > # status: waiting for good commit(s), bad commit known > # good: [2bab385991536a07d6f4c0970fa665f9626db059] Ubuntu: rebase to > v6.13-rc6 > git bisect good 2bab385991536a07d6f4c0970fa665f9626db059 > # good: [2cc57105ca61b9fb3744adbf83bbc8b2c0c0d055] johan_defconfig: add > custom config > git bisect good 2cc57105ca61b9fb3744adbf83bbc8b2c0c0d055 > # good: [260e63d4eece13d5f908646b8b9e5af44a40d012] arm64: dts: qcom: > msm8996-xiaomi-gemini: Fix LP5562 LED1 reg property > git bisect good 260e63d4eece13d5f908646b8b9e5af44a40d012 > # good: [033bf6bd4e245be05553ed1ec774dd78c77922ec] arm64: dts: qcom: > sar2130p: correct sleep clock frequency > git bisect good 033bf6bd4e245be05553ed1ec774dd78c77922ec > # good: [76f8b9a9e0d9fdafa6103ea789865dd6897f3bd6] dt-bindings: clock: > qcom-rpmhcc: Add RPMHCC bindings for QCS615 > git bisect good 76f8b9a9e0d9fdafa6103ea789865dd6897f3bd6 > # good: [22507018ce51df4f63a5ad468608426bcd799e2b] clk: qcom: > gcc-mdm9607: Fix cmd_rcgr offset for blsp1_uart6 rcg > git bisect good 22507018ce51df4f63a5ad468608426bcd799e2b > # good: [8db2b9a883cbb85a189ad74c3fb3264cc54134b8] soc: qcom: > pmic_glink_altmode: simplify locking with guard() > git bisect good 8db2b9a883cbb85a189ad74c3fb3264cc54134b8 > # bad: [17f8cef45b6a390dd2216d28290c21a92c3e3ae1] Ubuntu: Update changelog > git bisect bad 17f8cef45b6a390dd2216d28290c21a92c3e3ae1 > # bad: [1c51ac09700c1722329672f6254dca78fbc32101] soc: qcom: rmtfs: > constify rmtfs_class > git bisect bad 1c51ac09700c1722329672f6254dca78fbc32101 > # bad: [d6c656205ae9eb32dce89df95faabe898029e668] soc: qcom: rmtfs: > allow building the module with COMPILE_TEST=y > git bisect bad d6c656205ae9eb32dce89df95faabe898029e668 > # first bad commit: [d6c656205ae9eb32dce89df95faabe898029e668] soc: > qcom: rmtfs: allow building the module with COMPILE_TEST=y > > I have put the tag up on github [1]. It _is_ odd. I think I've seen this issue before. I'd bet it's just timing. Konrad
