**** BEGIN LOGGING AT Tue Aug 02 02:59:57 2022 Aug 02 07:27:46 Morning! Aug 02 07:28:37 Herrie: yesterday I also analyzed the dmesg output we get with a classic webOS bootup... and there's actually nothing specific regarding usb hw Aug 02 07:28:48 and no /dev/usb0 either Aug 02 07:29:02 (but that one is related to usb gadget) Aug 02 07:31:38 Morning! Aug 02 07:32:13 Tofe: Ah OK. Maybe worth comparing with our old uImage to see what it shows there? Aug 02 07:32:44 Though it Aug 02 07:32:58 's probably similar, since it uses a similar kernel, just Android based Aug 02 07:34:09 It might be worth poking some of the pmOS guys who are familiar with mainlining in their matrix or wherever they are nowadays to see if they have some thoughts Aug 02 07:34:14 They probably ran into this before Aug 02 07:37:42 well, our diagnostic is a bit fragile: we don't actually really know how much of the USB hw is detected/powered/activated Aug 02 07:43:12 Tofe: True, but they still might have been at this point before and have some useful pointers on how to get USB hw working or some hints Aug 02 07:43:23 It's worth to ask I'd say Aug 02 07:43:30 Qcom often does things quite specific Aug 02 08:23:12 Tofe: Came across this one, quite a good reference on device trees: https://elinux.org/Device_Tree_Usage Aug 02 08:23:39 And this DSI panel driver generator: https://github.com/msm8916-mainline/linux-mdss-dsi-panel-driver-generator Aug 02 11:16:18 Morning Aug 02 12:34:02 o/ Aug 02 17:44:25 Herrie: I see you indicated earlier that TP has an ISP1763: but isn't it a Chipidea instead ? or maybe the two things are not the same components and both are needed ? I'm a bit lost now :p Aug 02 17:46:15 Tofe: Well according to the schematics the ISP1763A is a USB controller, connected to the APQ8060 on the left and PCI-e connector to the right Aug 02 17:46:36 HS USB from APQ8060 connects to MicroUSB Aug 02 17:47:08 ah, ok Aug 02 17:47:20 also, I saw this little patch: https://patchwork.kernel.org/project/linux-omap/patch/20220210131028.1404851-1-aford173@gmail.com/ might be already merged Aug 02 17:47:26 Which in turn is connected to the MAX8903B Switch Charge and in turn connects to the battery & PM8058 PMIC Aug 02 17:48:28 Tofe: Well this ISP1763 patch is for OMAP, but could be modified to be used for us I guess at some point Aug 02 17:48:50 But seeing it connects the PCI-e (which is used on Opal/Topaz 4G, it's not that urgent right now I guess Aug 02 17:49:07 HS USB from APQ8060 should solve the USB connection for us Aug 02 18:38:31 Tofe: https://github.com/torvalds/linux/blob/master/drivers/clk/qcom/Kconfig#L152 and/or https://github.com/torvalds/linux/blob/master/drivers/clk/qcom/Kconfig#L175 might be what we need as well Aug 02 18:40:01 Seems we don't have the IPQ one set (https://github.com/shr-distribution/meta-smartphone/blob/master/meta-hp/recipes-kernel/linux/linux-hp-tenderloin/defconfig#L3807) but the 8660 one is set: https://github.com/shr-distribution/meta-smartphone/blob/master/meta-hp/recipes-kernel/linux/linux-hp-tenderloin/defconfig#L3810 Aug 02 18:40:30 we'll need CONFIG_CHARGER_MAX8903 too Aug 02 18:44:06 Tofe: Yes, but that should be for charging, not for the port itself to work I would say? Aug 02 18:44:22 I.e. to route current from USB port via the MAX8903->Battery, no ? Aug 02 18:44:44 It shouldn't be needed for getting an USB connection itself to work at least that's my understnading Aug 02 18:44:48 I might be wrong of course Aug 02 18:48:20 Frankly I don't know, I agree with you, but it's worth a test Aug 02 18:56:55 nope, all this didn't work Aug 02 18:57:54 The problem with the MSM8660 and APQ8060 is that it's quite obscure, so documentation is hard to find Aug 02 18:58:23 For MSM8960 it's a bit better and it should be fairly similar since it's the successor Aug 02 19:10:59 Tofe: I'm asking around in #postmarketos-mainline on oftc network, see if there's any input there Aug 02 19:29:54 Ah, maybe I should join that network too Aug 02 19:30:06 I never bothered subscribing there Aug 02 19:31:28 ah there's a webchat, great Aug 02 19:33:18 The most recent changes are in https://github.com/Tofee/shr-linux/commits/tenderloin/5.19/mainline Aug 02 19:33:48 and the defconfig didn't change much lately Aug 02 19:35:03 Herrie: and I think one good way to describe our main issue is that /sys/class/udc/ is empty Aug 02 20:28:06 Tofe: Ah got distracted by a phone call and laptop disconnected from the webchat Aug 02 20:28:11 Did anything happen there since you joines? Aug 02 20:28:13 joined Aug 02 20:29:07 well, he pointed me to his patches https://patchwork.kernel.org/project/linux-arm-msm/list/?series=664213 but nothing new Aug 02 20:29:38 Herrie: https://paste2.org/BJbbxK1p Aug 02 20:31:01 in the meantime, I found a little mistake in the regulator for usb (the voltage is slightly different it seems), so testing that now Aug 02 20:34:35 nope Aug 02 20:56:44 Tofe :( Aug 02 20:59:09 right I'm wondering if the "gsbi6_serial" which activates "qcom,msm-uartdm" driver is related to usb or not Aug 02 20:59:14 +now Aug 02 20:59:16 https://github.com/Tofee/shr-linux/commits/tenderloin/5.19/mainline feel free to send me PR for shr-linux whenever needed Aug 02 20:59:35 (looking at https://github.com/Tofee/shr-linux/blob/tenderloin/3.4/halium-9.0/arch/arm/mach-msm/hp/tenderloin/board-tenderloin.c#L3235 ) Aug 02 20:59:53 JaMa: well, sure, but it doesn't work that well anyway Aug 02 21:00:29 not worse than the one we already have, right? :) Aug 02 21:00:44 did you merge the final 5.19 into it already? Aug 02 21:01:38 JaMa: nope, I'm staying on the current 5.19; it's not worse, but not better either Aug 02 21:01:51 but yes I can always PR, won't hurt Aug 02 21:02:31 done Aug 02 21:05:17 ah sorry, I was thinking about hammerhead sources and based on git log was surprised how many commits you've added on top of 5.19-rc7 Aug 02 21:06:15 at least will get notification when you push something interesting Aug 02 21:06:23 ok :) Aug 02 21:07:14 Herrie: msm_otg seems related to the 8058_s1 regulator too, but I can't find it in the dts https://github.com/Tofee/shr-linux/blob/tenderloin/3.4/halium-9.0/arch/arm/mach-msm/hp/tenderloin/board-tenderloin-regulator.c#L109 Aug 02 21:30:02 Tofe: Which one you cannot find in DTS? Aug 02 21:30:06 8058? Aug 02 21:31:09 the vddcx relationship between msm_otg and the regulator 8058_s1 Aug 02 21:31:23 but maybe it's just a voltage to setup, that's all Aug 02 21:32:30 I'm trying one last thing on the dts before I give up for tonight Aug 02 21:33:55 Tofe: Question is if we need OTG for simple USB connection already Aug 02 21:34:03 Isn't that more "advanced" stuff? Aug 02 21:34:24 it seems to be a classic way of handling this msm usb stuff in dts Aug 02 21:34:34 but I agree it's a bit weird Aug 02 21:35:17 ok, last hack didn't work, so it's all for tonight Aug 02 21:40:00 Tofe: OK Aug 02 21:40:04 Yeah here also bedtime Aug 02 21:40:30 https://github.com/Tofee/shr-linux/commit/307026439da0d90b05e7fda8a06964885a399415 today's hacks Aug 02 21:40:43 just fyi :) Aug 02 21:44:42 Tofe: OK Aug 02 22:31:55 Tofe: Saw this one, thought it might be useful, serial output over USB: https://wiki.postmarketos.org/wiki/Mainlining_FAQ#How_to_get_serial_output_over_USB.3F_.28USB_serial_gadget.29 **** ENDING LOGGING AT Wed Aug 03 02:59:57 2022