**** BEGIN LOGGING AT Sat Mar 14 03:02:45 2020 Mar 14 15:49:39 build #316 of arc770/generic is complete: Success [build successful] Build details are at http://buildbot.openwrt.org/master/images/builders/arc770%2Fgeneric/builds/316 Mar 14 16:08:10 jow: thanks for that pkgconf fix Mar 14 16:08:23 jow: feel free to add my Acked-by Mar 14 16:45:56 when an ethernet LED is hardware controlled but can be returned to standard GPIO control, is there a preference to give control back to the system or leave it hardwired (porting a new device)? Mar 14 16:56:48 ynezz: can you reply with acked-by? Thne I can simply grab the mbox from patchwork Mar 14 16:57:43 * f00b4r0 scratches head. Same gpio appears to be tied to both an LED and a button ?_? Mar 14 16:59:14 f00b4r0: mikrotik? they seem to be fond of doing that Mar 14 16:59:54 KanjiMonster: aye. The reset button shares gpio with LED Mar 14 17:00:01 not sure how to deal with that Mar 14 17:00:15 (not even sure how that can even work) Mar 14 17:02:03 don't ask me. iirc the same issue is on the 493, except that the button shares its line with the clk gpio of the (bitbanged) mdio bus IIRC Mar 14 17:02:16 o_O Mar 14 17:02:27 that sounds even more brain damaged Mar 14 17:07:29 Are you saying the same physical GPIO pin on the device is switched between input or output mode? Or are there 2 physical pins, one GPI one GPO that are called GPI'n' / GPO'n' ? Mar 14 17:09:20 it would appear the be the former. At least from what I can tell from the linux interface Mar 14 17:10:50 meanwhile I have yet another RB piece of *** supported (although via ar71xx, which is sure to displease a few ;P) Mar 14 17:21:31 ldir: yes output => clock signal (or LED), input => button Mar 14 17:24:39 KanjiMonster: do you how it is handled in owrt? I can't seem to find the relevant code in mach-rb4xx.c ? (i might be blind) Mar 14 17:24:53 f00b4r0: it isn't Mar 14 17:25:00 makes sense Mar 14 17:32:06 f00b4r0: the least dirty solution I can think of is creating a virtual gpio chip driver that takes one gpio as input and exports it as one gpi and one gpo line Mar 14 17:32:51 KanjiMonster: that could work, at least on paper, but I don't see how sharing the _reset_ button GPIO with anything isn't bound to wreck havoc :) Mar 14 17:34:26 f00b4r0: it seems to work for mikrotik ;p Mar 14 17:35:21 KanjiMonster: ack. They weren't kind enough to provide me with their latest kernel patch (they sent me an archive from 2017 claiming it was their last), so I can't tell how they do it. Besides their code makes me cringe ;P Mar 14 17:37:00 for all I know they could have implemented some ugly polling Mar 14 17:37:44 that how gpio buttons are implemented when you have no interrupt for the gpio available Mar 14 17:38:01 true Mar 14 21:26:13 sxtsq-2nd support patch sent. I hope it won't be too big an offense that it's an ar71xx patch :^) **** ENDING LOGGING AT Sun Mar 15 03:02:57 2020